d 1076 acta universitatis ouluensis university of …jultika.oulu.fi/files/isbn9789514263286.pdfthe...

ABCDEFG

UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND

A C T A U N I V E R S I T A T I S O U L U E N S I S

S E R I E S E D I T O R S

SCIENTIAE RERUM NATURALIUM

HUMANIORA

TECHNICA

MEDICA

SCIENTIAE RERUM SOCIALIUM

SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR

Professor Mikko Siponen

University Lecturer Elise Kärkkäinen

Professor Hannu Heusala

Professor Olli Vuolteenaho

Senior Researcher Eila Estola

Information officer Tiina Pistokoski

University Lecturer Seppo Eriksson


Publications Editor Kirsti Nurkkala

ISBN 978-951-42-6327-9 (Paperback)ISBN 978-951-42-6328-6 (PDF)ISSN 0355-3221 (Print)ISSN 1796-2234 (Online)

U N I V E R S I TAT I S O U L U E N S I S

MEDICA

ACTAD

D 1076

ACTA

Ville Saarela

OULU 2010

D 1076

Ville Saarela

STEREOMETRIC PARAMETERS OF THE HEIDELBERG RETINA TOMOGRAPH IN THE FOLLOW-UP OF GLAUCOMA

FACULTY OF MEDICINE,INSTITUTE OF CLINICAL MEDICINE,DEPARTMENT OF OPHTHALMOLOGY,UNIVERSITY OF OULU

ABCDEFG





HUMANIORA

TECHNICA

MEDICA


SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR












MEDICA

ACTAD

D 1076

ACTA

Ville Saarela

OULU 2010

D 1076

Ville Saarela



A C T A U N I V E R S I T A T I S O U L U E N S I SD M e d i c a 1 0 7 6

VILLE SAARELA


Academic dissertation to be presented with the assent ofthe Faculty of Medicine of the University of Oulu forpublic defence in Auditorium 5 of Oulu UniversityHospital, on 12 November 2010, at 12 noon

UNIVERSITY OF OULU, OULU 2010

Copyright © 2010Acta Univ. Oul. D 1076, 2010

Supervised byProfessor Anja Tuulonen

Reviewed byDocent Päivi PuskaProfessor Hannu Uusitalo

ISBN 978-951-42-6327-9 (Paperback)ISBN 978-951-42-6328-6 (PDF)http://herkules.oulu.fi/isbn9789514263286/ISSN 0355-3221 (Printed)ISSN 1796-2234 (Online)http://herkules.oulu.fi/issn03553221/

Cover designRaimo Ahonen

JUVENES PRINTTAMPERE 2010

Saarela, Ville, Stereometric parameters of the Heidelberg Retina Tomograph in thefollow-up of glaucoma Faculty of Medicine, Institute of Clinical Medicine, Department of Ophthalmology, Universityof Oulu, P.O.Box 5000, FI-90014 University of Oulu, Finland Acta Univ. Oul. D 1076, 2010Oulu, Finland

AbstractGlaucoma is a progressive neuropathy of the optic nerve. It causes degeneration of ganglion cellaxons resulting in defects in the retinal nerve fibre layer (RNFL) and characteristic changes in theoptic nerve head (ONH). The Heidelberg Retina Tomograph (HRT) is a confocal scanning laserimaging device, which measures the topography of the ONH and the adjacent RNFL. To quantifythe measurements of the ONH topography, various stereometric parameters are calculated.

The change in the stereometric parameters of the HRT was studied in 34 eyes withglaucomatous progression in RNFL photographs and 34 eyes without progression. The change inonly one stereometric parameter, the cup shape measure, showed a statistically significantcorrelation with the progression of the RNFL defect. An optimised change in the best three-parameter combination had 77% sensitivity and 79% specificity for progression.

The change in the stereometric parameters was compared in 51 eyes with glaucomatousprogression in stereoscopic ONH photographs and 425 eyes without progression. The parametershaving the best correlation with progression include cup:disc area ratio, vertical cup:disc ratio, cupvolume and rim area. The parameter with the largest area under the receiver operatingcharacteristics curve (0.726) was the vertical cup:disc ratio. A change of 0.007 in the verticalcup:disc ratio had a sensitivity of 80% and a specificity of 65% for progression.

The factors having the most significant effect on the sensitivity and specificity of thestereometric parameters for progression were the reference height difference and the meantopography standard deviation, indicating image quality. The change in image quality and age alsoshowed a consistent, but variably significant influence on all parameters tested.

Exercise was associated with an increase in variance in 17 of the 18 stereometric parameters. In conclusion, the change in the stereometric parameters provides useful information on ONH

topography, especially when image quality is excellent. However, the evaluation of glaucomatousprogression should not rely solely on the stereometric parameters of the HRT.

Keywords: glaucoma, Heidelberg Retina Tomograph, optic nerve head, progression,retinal nerve fibre layer

To my family

7

Acknowledgements

This study was carried out in the Department of Ophthalmology, University of Oulu, during the years 2003–2010.

I wish to express my sincere gratitude to my supervisor, Professor Anja Tuulonen, M.D., Ph.D., Head of the Department of Ophthalmology for her advice and guidance during this work, and to Professor Juhani Airaksinen, M.D., Ph.D., for introducing me to the world of glaucoma imaging and teaching me how to write scientifi c articles.

I wish to express my warmest gratitude to my co-authors Aura Falck, M.D., Ph.D., for her support and for taking the time to evaluate glaucomatous damage in hundreds of optic nerve heads, and Elina Ahvenvaara, M.D., for her energy and effectiveness in the prospective study.

I am very grateful to Professor Hannu Uusitalo, M.D., Ph.D., and Docent Päivi Puska, M.D., Ph.D., for carefully reviewing this thesis and providing such constructive criticism and valuable advice.

I also wish to thank Risto Bloigu for his advice and for always having the time for my questions concerning statistical analysis, and Anna Vuolteenaho, M.A., for revising the English of this thesis.

I wish to express my gratitude to Jukka Veijola for his enthusiasm and expertise in imaging and for swiftly solving so many of the software and imaging-related problems, to Maritta Runtti and Tarja Laurila for all their help and support, and to Pauli Hyytinen, Sylvi Savolainen and Päivi Pehkonen-Karioja for their work in imaging the patients.

I want to thank all my colleagues in the Department of Ophthalmology during these years. Especially I wish to thank Pasi Hägg, Jukka Nevalainen, Johanna Liinamaa, Katri Stoor and Katja Kemppainen for their support and Hannu Alanko for his constructive criticism and advice. I also wish to thank the personnel of the Department of Ophthalmology for the warm and supportive atmosphere.

I am very grateful to my relatives and friends for their valuable encouragement and interest. A special thank you goes to Terhi, Timo, Kaisa, Mika, Raku and Miia for their support.

I wish to express my deepest gratitude to my parents Anja and Mikko for their love and support.

Finally, and most of all, I wish to express my loving thanks to my wife Elina for her love, encouragement and continuous support during these years and my

8

sons Paavo, Niilo and Juho for bringing me so much happiness and reminding me of the real life outside the world of research.

This work was supported fi nancially by grants from the De Blindas Vänner – Sokeain Ystävät Foundation, Glaukooma Tukisäätiö LUX foundation, the Eye Foundation, the Eye and Tissue Bank Foundation, the Evald and Hilda Nissi Foundation and KEVO funding of Oulu University Hospital.

Oulu, October 2010 Ville Saarela

9

Abbreviations

AGIS Advanced Glaucoma Intervention Study CIGTS Collaborative Initial Glaucoma Treatment StudyCLM Contour Line ModulationECC Enhanced Corneal CompensationEGPS European Glaucoma Progression StudyEMGT Early Manifest Glaucoma TrialGCP Glaucoma Change ProbabilityGDx Imaging device using scanning laser polarimetryGPA Glaucoma Progression AnalysisGPS Glaucoma Probability ScoreHRT Heidelberg Retina Tomograph IOP Intraocular Pressure LDF Linear Discriminant FunctionMD Mean DeviationMRA Moorfi elds Regression Analysis OCT Optical Coherence TomographyOHTS Ocular Hypertension Treatment Study ONH Optic Nerve Head POD Proper Orthogonal DecompositionRNFL Retinal Nerve Fibre LayerROC Receiver Operating CharacteristicsSAP Standardised Automated Perimetry SIM Statistical Image MappingTCA Topographic Change AnalysisTSD Topography Standard DeviationVCC Variable Corneal Compensation

11

List of original publications

This thesis is based on the following articles, which are referred to in the text by their Roman numerals:

I Saarela V & Airaksinen PJ (2008) Heidelberg Retina Tomograph Parameters of the Optic Disc in Eyes With Progressive Retinal Nerve Fibre Layer Defects. Acta Ophthalmol 86: 603–8.

II Saarela V, Falck A, Airaksinen PJ & Tuulonen A (2010) The sensitivity and specifi city of Heidelberg Retina Tomograph parameters to glaucomatous progression in disc photographs. Br J Ophthalmol 94: 68–73.

III Saarela V, Falck A, Airaksinen PJ & Tuulonen A (2010) Factors affecting the sensitivity and specifi city of the Heidelberg Retina Tomograph parameters to glaucomatous progression in disc photographs. Acta Ophthalmol [Epub ahead of print] doi: 10.1111/j.1755–3768.2010.01881.x

IV Saarela V, Ahvenvaara E & Tuulonen A Variability of Heidelberg Retina Tomograph parameters during exercise. Manuscript.

13

Contents

Abstract Acknowledgements 7Abbreviations 9List of original publications 11Contents 131 Introduction 152 Review of the literature 17

2.1 Summary of the literature search ............................................................ 172.2 Glaucoma ................................................................................................ 18

2.2.1 Epidemiology, risk factors and consequences of glaucoma ......... 182.2.2 Course and management of glaucoma ......................................... 19

2.3 Diagnosis of glaucomatous damage ....................................................... 192.3.1 Stereoscopic optic nerve head photography ................................ 212.3.2 Heidelberg Retina Tomograph ..................................................... 23 2.3.2.1 Reference plane defi nition ............................................. 25 2.3.2.2 Stereometric optic nerve head parameters ..................... 27 2.3.2.3 Moorfi elds regression analysis ...................................... 29 2.3.2.4 Glaucoma probability score .......................................... 30 2.3.2.5 Topographic change analysis ......................................... 31 2.3.2.6 Statistical image mapping and proper orthogonal decomposition................................................................ 31 2.3.2.7 Measurement variability of the Heidelberg Retina Tomograph ..................................................................... 32 2.3.3 Retinal nerve fi bre layer photography .......................................... 33 2.3.4 Scanning laser polarimetry .......................................................... 352.3.5 Optical coherence tomography .................................................... 362.3.6 Standardised automated perimetry ............................................... 36

2.4 Diagnosing glaucoma with the Heidelberg Retina Tomograph .............. 372.5 Detecting glaucomatous progression with the Heidelberg Retina

Tomograph .............................................................................................. 383 Aims of the research 414 Subjects and methods 43

4.1 Subjects ................................................................................................... 434.2 Stereoscopic ONH photography ............................................................. 454.3 Scanning laser ophthalmoscopy ............................................................. 45

14

4.4 Retinal nerve fi bre layer photography .................................................... 464.5 Examinations during exercise ................................................................ 464.6 Statistical analysis ................................................................................... 464.7 Ethical considerations ............................................................................. 47

5 Results 495.1 Correlation of the change in the stereometric parameters with

progression of the RNFL defects (I) ....................................................... 495.2 Correlation of the change in the stereometric parameters with

progression in stereoscopic ONH photographs (II) ................................ 505.3 Factors affecting the sensitivity and specifi city of the stereometric parameters to glaucomatous progression in ONH photographs (III) ..... 525.4 The effect of exercise on the stereometric ONH parameters (IV) .......... 525.5 The effect of image quality on agreement for detecting glaucomatous

progression .............................................................................................. 556 Discussion 57

6.1 Linear discriminant functions ................................................................. 576.2 Correlation between the stereometric ONH parameters with

progression in RNFL photographs (I) ..................................................... 586.3 Correlation between the stereometric parameters with progression in

ONH photographs (II) ............................................................................ 586.4 Factors affecting the sensitivity and specifi city to progression in ONH

photographs ( III) ..................................................................................... 586.5 Variability of ONH topography during exercise (IV) ............................. 596.6 Comparison of methods for detecting glaucomatous progression

using the HRT ......................................................................................... 606.7 Signifi cance of change ............................................................................ 636.8 Detection of glaucomatous progression ................................................. 636.9 Practical implications ............................................................................. 64

7 Conclusions 67References 69Original publications 87

15

1 Introduction

Glaucoma is the second leading cause of blindness in the world (Quigley & Broman 2006). The visual impairment due to glaucoma is irreversible. In Finland, there are 80000 patients diagnosed with glaucoma. The prevalence increases with age, being up to 10% in a population over 80 years of age (Mitchell et al. 1996). The course of glaucoma is typically a slowly progressing one, but the rate of progression varies (Airaksinen et al. 1992, Heijl et al. 2002). Since the progression of glaucomatous damage can be prevented or delayed with treatment (Leske et al. 2003, Maier et al. 2005), it would be important to identify the patients with a high rate of progression.

Glaucoma causes degeneration of ganglion cell axons. The structural damage can be seen as defects in the retinal nerve fi bre layer and as topographic changes in the optic nerve head. The corresponding functional damage can be detected in the visual fi elds (Quigley et al. 1982, Airaksinen et al. 1985). However, glaucomatous changes in the optic nerve head, the retinal nerve fi bre layer and the visual fi elds may manifest themselves at different times during follow-up (Tuulonen et al. 2003).

Glaucomatous damage in the optic nerve head can be detected with ophthalmoscopy, fundus photography and scanning laser ophthalmoscopy. Both ophthalmoscopy and the evaluation of photographs are subjective methods and subject to high variation between observers (Lichter 1976, Varma et al. 1992). A quantitative analysis of optic nerve head topography can be achieved with scanning laser ophthalmoscopy.

The Heidelberg Retina Tomograph is a confocal scanning laser imaging device, which measures the topography of the optic nerve head and the adjacent retinal nerve fi bre layer (Zinser et al. 1989). To quantify the measurements of the optic nerve head topography, various stereometric parameters are calculated (Burk et al. 1990). The change in stereometric optic nerve head parameters may be used for detecting glaucomatous progression. The aim of the research was to study how the change in the stereometric parameters correlates with glaucomatous progression. Factors that may affect the correlation with progression were also evaluated.

17

2 Review of the literature

2.1 Summary of the literature search

A literature search in PubMed was performed on 2nd of March 2010. After the exclusion of case reports and articles not related to the topic or using irrelevant methodology, 40 articles written in English were included. An additional 16 articles were identifi ed in the references of the 40 articles and the Cochrane database. The specifi cs of the search are given in Figure 1.

PubMed search with keywords “glaucoma” and “progression”: n = 1668

n = 1572

Keywords “Heidelberg Retina Tomograph or scanning laser ophthalmoscopy or optic nerve head topography or optic disc

topography” added: n = 96 n = 17

Included based on language (English): n = 79 n = 19

Included based on title: n = 60 n = 17

Included based on abstract: n = 43 n = 3

Included based on article: n = 40

Review of the references of the 40 accepted articles and the Cochrane database

n = 16

56 articles dated 1994 – 2010 accepted based on literature search

Figure 1. Flow chart summarising the literature search.

18

2.2 Glaucoma

Glaucoma is a progressive neuropathy of the optic nerve. It causes degeneration of ganglion cell axons resulting in defects in the retinal nerve fi bre layer (RNFL) and characteristic changes in the optic nerve head (ONH). The corresponding functional damage can be detected in the visual fi elds (Quigley et al. 1982, Airaksinen et al. 1985).

2.2.1 Epidemiology, risk factors and consequences of glaucoma

The vast majority of glaucoma patients in a Caucasian population have open-angle glaucoma, primary open-angle glaucoma being the most common (Wensor et al. 1998, Quigley & Broman 2006). The risk factors for open-angle glaucoma include raised intraocular pressure (IOP), age, exfoliation, family history of glaucoma, myopia and decreased perfusion pressure with age (Ringvold et al. 1991, Sommer et al. 1991b, Klein et al. 1992, Dielemans et al. 1994, Tielsch et al. 1994, Hirvelä et al. 1995, Tielsch et al. 1995, Ekström 1996, Mitchell et al. 1996, Nemesure et al. 1996, Mitchell et al. 1999). Patients with relatively thin central corneas may have an increased risk of developing glaucoma (Gordon et al. 2002, Kass et al. 2002). Small central corneal thickness may also cause underestimation of IOP (Doughty & Zaman 2000).

The estimated prevalence of open-angle glaucoma is 0.2–1.0% in people between 40 and 50 years of age in a predominantly Caucasian population. The prevalence increases with age, being 2–10% in a population over 80 years (Tielsch et al. 1991, Mitchell et al. 1996, Reidy et al. 1998).

In 2010 there are approximately 61 million people suffering from glaucoma worldwide, 8.4 million of them bilaterally blind (Quigley & Broman 2006). In Finland, there are 80000 patients treated for glaucoma. It is the second most common cause of visual impairment in the elderly after macular degeneration (Finnish Register of Visual Impairment 2008). The visual impairment due to glaucoma is irreversible.

In addition to visual deterioration, glaucoma also imposes an economical burden on both the patient and society. Expenses arise from treatment, follow-up and visual disability due to glaucoma (Vaahtoranta-Lehtonen et al. 2007). The severity of glaucomatous damage correlates to the increased overall cost of illness (Fiscella et al. 2009).

19

2.2.2 Course and management of glaucoma

The course of open-angle glaucoma is typically a slowly progressing one (Tuulonen et al. 2003). There is progression of glaucomatous damage despite treatment (Migdal et al. 1994, AGIS Investigators 2000). With treatment the visual impairment due to glaucoma can be delayed by several years. It has been estimated to take a mean time of 23 years without treatment and 35 years with treatment for glaucomatous damage to progress from mild visual fi eld damage to at least unilateral blindness (Burr et al. 2007). The estimate is based on randomised controlled trials of treatment vs. non-treatment. However, the rate of progression varies between patients (Airaksinen et al. 1992, Collaborative Normal Tension Glaucoma Study Group 1998a, 1998b, Heijl et al. 2002).

Risk factors for glaucomatous progression include exfoliation, age, bilaterality of glaucoma, thin central corneal thickness, optic disc haemorrhages, high IOP at baseline and follow-up and advanced glaucomatous damage at baseline (Airaksinen et al. 1981, Leske et al. 2003, Chauhan et al. 2008). Fluctuation of IOP may also increase the risk of progression (Asrani et al. 2000, Caprioli & Coleman 2008), although no correlation was found in the Early Manifest Glaucoma Trial (Bengtsson & Heijl 2005).

Currently, the only effective treatment for glaucoma is to lower IOP (Tuulonen et al. 2003, Burr et al. 2007). The onset of glaucoma is prevented or delayed by IOP reducing treatment in patients with ocular hypertension (Kass et al. 2002). The progression of glaucomatous damage is also delayed by treatment in both patients with normal tension glaucoma (Collaborative Normal Tension Glaucoma Study Group 1998a, 1998b) and patients with high initial IOP (AGIS Investigators 2000, Leske et al. 2003, Maier et al. 2005).

IOP can be reduced with medication, laser treatment or surgery. The IOP reduction is greatest with surgical treatment (AGIS Investigators 1998, Lichter et al. 2001, Burr et al. 2004). In patients with high initial IOP, it is superior to medical and laser treatment in reducing visual fi eld progression (Migdal et al. 1994). Cataract surgery also modestly lowers IOP (Pohjalainen et al. 2001), and the effect is sustained for several years (Falck et al. 2009).

2.3 Diagnosis of glaucomatous damage

There are several different diagnostic criteria for glaucoma in the scientifi c literature. The diagnosis is based on the evaluation of the ONH, RNFL and visual fi elds. A

20

meticulous clinical examination including IOP measurements, biomicroscopy and gonioscopy is needed to differentiate between different types of glaucoma. Even in recent diagnostic studies, there is variation in the defi nition of glaucomatous damage (Burr et al. 2007).

Albrecht von Graefe reported in 1855 that glaucoma may cause excavation of the ONH (von Graefe 1855). Since then, the morphology of the ONH has been examined using several methods. The clinical evaluation of the ONH may be done using direct ophthalmoscopy, binocular ophthalmoscopy or ONH photography (Jonas et al. 1999). Direct ophthalmoscopy has the disadvantage of missing stereopsis. With ophthalmoscopy, no permanent record is left if the appearance of the ONH is not drawn by the examiner. Both ophthalmoscopy and the evaluation of ONH photographs are subjective methods and as such subject to high variation even between expert observers (Lichter 1976, Varma et al. 1992). Several classifi cation systems for glaucomatous ONH damage have been proposed (Armaly 1969, Read & Spaeth 1974, Jonas et al. 1988, Spaeth et al. 2002, Henderer 2006). A quantitative analysis of ONH topography can be achieved with scanning laser ophthalmoscopy.

The RNFL can be evaluated using green light on ophthalmoscopy. Documentation of the RNFL is achieved with high-resolution black-and-white RNFL photography using a green or blue fi lter (Airaksinen & Nieminen 1985). The interpretation of the images is subjective, but semiquantitative scoring of the defects has been presented (Airaksinen et al. 1984, Quigley et al. 1993). A quantitative analysis of the thickness of the RNFL can be obtained with scanning laser polarimetry or optical coherence tomography (OCT).

Visual fi eld testing can be performed using static or kinetic perimetry. In static perimetry the visual object is stationary and in kinetic perimetry it is moving. In static perimetry the sensitivity threshold for seeing is determined for each point of measure in a test grid. Partly due to the pattern of glaucomatous visual fi eld defects, kinetic perimetry may be of less value in the diagnostics of glaucoma (Heijl et al. 1980, Beck et al. 1985).

In frequency doubling perimetry rapidly fl ickering stimuli are presented in the visual fi eld. Due to frequency doubling illusion, twice as many bars as actually exist are seen in a normal fi eld. A higher contrast level has to be used in the abnormal regions of the visual fi eld for the illusion to occur (Cello et al. 2000).

Standardised automated perimetry (SAP) is considered the current reference standard for detecting glaucomatous visual fi eld defects (Foster et al. 2002, Tuulonen et al. 2003). The two most commonly used instruments are the Humphrey

21

and Octopus perimeters. Static perimetry is performed to the central visual fi eld, usually limited to the 24-degree or 30-degree area, with test points 6 degrees apart. The reference standard is white-on-white perimetry (white testpoints on a white background), but in short-wavelength automated perimetry blue testpoints on a yellow background are used. Short-wavelength automated perimetry measures the blue cone system of the visual pathway and may detect very early loss of sensitivity in the visual fi elds (Johnsson et al. 1993). However, it is more diffi cult and time-consuming and has a higher test-retest variability compared to white-on-white perimetry (Blumenthal et al. 2003, Burr et al. 2007).

2.3.1 Stereoscopic optic nerve head photography

Stereoscopic ONH photography has been shown to be effective in evaluating glaucomatous changes (Airaksinen et al. 1984, Quigley et al. 1992). It is a well-established documentation of basic clinical examination that has remained unchanged for decades. The validity of the method has been verifi ed in large randomised controlled clinical trials such as the Ocular Hypertension Treatment Study (OHTS), the European Glaucoma Prevention Study (EGPS) and the Early Manifest Glaucoma Trial (EMGT) (Gordon & Kass 1999, Leske et al. 1999, Miglior et al. 2002b). These studies show that by standardising the evaluation of stereoscopic ONH photographs, it can be performed in a reproducible manner (Zeyen et al. 2003, Parrish et al. 2005).

The stereoscopic ONH photographs are obtained in either a simultaneous or a sequential manner. In simultaneous stereophotography a camera with a beam-splitting prism may be used to capture the images with a fi xed relative angle (Donaldson et al. 1980). The advantage of a fi xed stereobasis is the reproducibility of the stereo-effect and the stereometric measurements. With sequential photography the stereo-effect is achieved by lateral shift of a normal fundus camera with a built-in stopper keeping the relative angle constant.

In order to see the three-dimensional appearance of the ONH, both images in a stereo pair are viewed simultaneously. This may be achieved by convergent or divergent view of adjacent images with the right and the left eye fi xating on different images. Other methods include fl icker glasses rapidly alternating the image between the eyes (Morgan et al. 2005) and a stereoviewer, a method also used with stereo slides (Armaly 1969, Reus et al. 2010).

Stereoscopic ONH photography is a qualitative method and there is variability in the detection of glaucomatous damage and the detection of progression using

22

Figure 2. Stereoscopic ONH photographs taken ten years apart. Progression of the

glaucomatous damage can be seen in the superior and inferior quadrants.

photography (Tuulonen et al. 2003). In a recent study, general ophthalmologists from countries around Europe evaluated a set of ONH photographs of glaucoma patients and healthy controls. The study showed moderate accuracy (80.5%) for detecting glaucoma. The average intra-observer agreement was good (κ = 0.7). The ophthalmologists participating from Finland showed the greatest specifi city (93.2%) but low sensitivity (69.3%) for detecting glaucomatous damage (Reus et al. 2010).

23

The inter-observer variability in assessing stereoscopic ONH photographs tends to be larger than intra-observer variability (Tielsch et al. 1988, Varma et al. 1992, Zeyen et al. 2003). The inter-observer agreement for detecting progression may be improved by training (Zeyen et al. 2007). However, there is considerable inter-observer variability in evaluating progression even among glaucoma experts (Azuara-Blanco et al. 2003, Jampel et al. 2009). In the EGPS, observers evaluating progression were experienced ophthalmologists masked for the temporal sequence of the stereo photographs. The inter-observer agreement of the fi rst evaluations of the observers was moderate, with kappa-values ranging from 0.45 to 0.60 (Zeyen et al. 2003). Knowledge of the temporal sequence has been shown to infl uence the determination of progression. In a setting where three observers evaluated a set of stereo photographs for progression with a knowledge of the chronology, the kappa values were 0.68 +/- 0.05. When the same set was evaluated masked for sequence, the kappa values were 0.30 +/- 0.05 (Altangerel et al. 2005).

Both colour photographs and black-and-white photographs may be used in stereoscopic ONH photography. There is evidence suggesting that with black-and-white photographs both intra-observer and inter-observer agreement of progression may be higher (Zeyen et al. 2009). A variety of methods have been presented for standardising the evaluation of stereoscopic ONH photographs in order to diminish the variability of assessments (Henderer et al. 2003, Medeiros et al. 2005, DeLeon-Ortega et al. 2006). The masked evaluation of stereoscopic ONH photographs by glaucoma specialists is still considered the reference standard for ONH evaluation (Lin et al. 2007).

2.3.2 Heidelberg Retina Tomograph

The Heidelberg Retina Tomograph (HRT) is a confocal scanning laser imaging device, which uses a 670 nm diode laser to measure the topography of the ONH and the adjacent RNFL (Zinser et al. 1989). The device scans a 15º angle of the retinal surface in vertical and horizontal directions in equidistant focal planes. Each plane consists of a 384 x 384 pixel array with a lateral resolution of 10 μm. The number of scanned focal planes depends on the maximum depth of the cup from the highest point on the rim. There are from 16 up to 64 focal planes scanned depending on the depth of the cup. The stack of scanned planes is aligned and reassembled and the depth of the maximal refl ectance intensity of each pixel is calculated to form a three-dimensional topography image.

24

The device automatically performs three scans, and the mean of the three scanned images is used as the fi nal topography image. The standard deviation of the three scanned height measurements of each pixel is calculated. The mean of these standard deviations is the topography standard deviation (TSD). If the three scanned images are uniform, the TSD is low and the quality of the topography image is high.

There have been three commercially available generations of Heidelberg Retina Tomograph. The original HRT is also referred to as HRT I or HRT classic. The device relies heavily on operator input. In addition to the 15º scan, also 10º or 20º scans can be taken with HRT I. The resolution of the 10º scan is the same as in HRT II. The lateral resolution is 10 to 20 μm / pixel and longitudinal resolution is 62 to 128 μm / plane. The image acquisition time for HRT I is around 1.4 seconds (Fingeret 2005).

Compared to its predecessor, HRT II has more automated features, such as averaging of the scans, serial scans, fi ne focus and scan depth. The operator is still required to manually draw the contour line defi ning disc margin onto the topography image (Fingeret 2005). The imaging head and the light source are different in HRT I and HRT II. The images of HRT I cannot be analysed with HRT II software.

The HRT 3 introduced several changes in the imaging software (Strouthidis & Garway-Heath 2008). The imaging head is the same for HRT II and HRT 3, making these two generations of the device backwards compatible. The HRT 3 software is also able to analyse HRT I images, but since the images are acquired with different types of imaging heads, the results of HRT I and HRT 3 are not comparable as such. Compared to HRT II a slight scaling error was corrected and the drawing of the contour line was automatised (Strouthidis & Garway-Heath 2008). Also larger and ethnicity-specifi c, as yet unpublished, normative databases and improvements in the image alignment algorithms were introduced. The size of the imaging hardware was reduced so that HRT 3 is portable. The lateral image resolution of HRT II and HRT 3 is 10 μm/pixel and longitudinal resolution is 62 μm/plane. The image acquisition time is typically 1.0 seconds (Fingeret 2005).

Improvements in the image alignment technique have been introduced during the continuous development of HRT. A subpixel-based image alignment algorithm improved image quality and detection of glaucomatous damage using a stereometric parameter measuring cup shape (Burk & Rendon 2001). The image alignment algorithm introduced in HRT 3 software is based on a landmarking technique used in face recognition (Capel 2004). It has been reported to decrease pixel-to-pixel variability, and marginally improve the repeatability of stereometric measures compared to HRT II (Bergin et al. 2008).

25

2.3.2.1 Reference plane defi nition

To quantify stereometric rim and cup parameters in ONH topography, a reference plane is defi ned (Burk et al. 1990). The reference plane is parallel to the retinal surface. It needs to be stable so that the parameters change only when true structural changes in the ONH occur (Breusegem et al. 2008). Within the disc margin, the retinal surface located above the reference plane is defi ned as rim and below the reference level as cup. The reference plane used in HRT I is referred to as the 320 reference plane. It is set 320 μm posterior to a reference ring located in the image periphery.

With the introduction of HRT II, the reference plane defi nition was changed. In HRT II and HRT 3, the reference plane lies 50 μm posterior to the temporal disc margin. It is fi xed upon the most stable part of the disc margin, which is on the papillomacular bundle 4º to 10º below the horizontal meridian (Burk et al. 2000). This reference plane defi nition is referred to as the standard reference plane.

The 320 reference plane and the standard reference plane were compared with an individually determined reference level fi xed on the ring of Eschnig (Tuulonen et al. 1994) and a reference level fi xed on mean height of the 1º disc margin above the horizontal meridian (Airaksinen 1994). The standard reference plane gave the most reliable results for the diagnosis of glaucoma (Vihanninjoki et al. 2002). However, the thickness of the papillomacular bundle may be decreased due to glaucoma. This may cause underestimation of glaucomatous damage by the standard reference plane (Chen et al. 2001).

Another reference plane defi nition is based on measurements of RNFL thickness using the OCT (Park & Caprioli 2002). It may improve the diagnosis of early glaucomatous damage in eyes with tilted discs. An experimental reference plane that always lies beneath the contour line has been presented. It is based on the mean height of the contour line and the lowest 5% of these height values (Tan & Hitchings 2003). This reference plane defi nition showed the best agreement with subjective rim area estimates from topography images (Tan et al. 2004b).

There may be increased variability in rim area measurements with the standard reference plane compared to the 320 reference plane (Strouthidis et al. 2005b). The residual standard deviation from a regression slope of rim area measurements obtained from patients with progressing glaucomatous damage was highest with the standard reference plane (Poli et al. 2008). Hence, while the standard reference plane may be superior in diagnosing glaucomatous damage, the 320 reference plane may be more stable and have higher repeatability of rim area measurements during follow-up.

26

The Moorfi elds reference plane uses the defi nition of the standard reference plane in the baseline images. During follow-up the height difference of the standard reference plane from the reference ring used in 320 reference plane is kept constant (Poli et al. 2008). The Moorfi elds reference plane and the 320 reference plane have shown better repeatability of rim area measurements compared to the standard reference plane (Asaoka et al. 2009). The correlation of rim area change with identifi ed visual fi eld progression was higher with the Moorfi elds reference plane compared to the 320 reference plane (Asaoka et al. 2009).

Figure 3. A three-dimensional HRT image of the ONH. The margin of the ONH is defi ned

by the contour line. The cross-sectional image below demonstrates the position of the

standard reference plane. The reference plane is needed to distinguish between cup

and rim.

27

2.3.2.2 Stereometric optic nerve head parameters

To quantify the measurements of the ONH topography, various stereometric parameters are calculated (Burk et al. 1990). The disc area is defi ned as the area within a contour line (Figure 3). The contour line is transferred from the baseline image to the follow-up images, keeping the disc area constant. The modulation of the contour line in the temporal to superior and temporal to inferior octants as well as the maximal elevation, depression and height variation along the contour line are calculated. In addition, the mean thickness and the cross-sectional area of the RFNL along the contour line are given.

A reference plane is defi ned to distinguish between cup and rim (Figure 3). The topography within the contour line and above the reference plane is defi ned as rim and below the reference plane as cup (Burk et al. 1990). The mean height of the reference plane from the peripapillary retinal surface is calculated. The mean and maximum depth of the cup is measured. The area and volume of the cup and rim are calculated as well as cup:disc and rim:disc area ratios. The cup:disc ratios in the vertical and horizontal directions are given.

The cup shape measure is the third moment of the frequency distribution of the depth values within the disc margin. It is obtained by dividing the square of the third moment of the distribution by the third power of the second moment. It may have values between -1 and 1 (Burk et al. 1990, Zinser et al. 1990). The value is typically negative in normal eyes with small shallow cups and frequently small depth values. It is less negative or positive in deep cups with steep slopes and frequently high depth values (Uchida et al. 1996).

[Σ (Xi - X)3/n ] 2

Cup shape measure = ____________________ [Σ (X

i - X)2/n ] 3

where X is the mean depth, Xi is depth of each pixel, and n is the number of pixels

within the disc margin.Before the introduction of HRT 3, the inter-operator difference in drawing

the contour line could affect the stereometric ONH parameters. However, the agreement on the stereometric parameters between masked observers drawing the contour line has been shown to be good (Hatch et al. 1999). In fact, the image-acquisition-induced variability of the parameters has been reported to be higher than the operator-induced variability (Miglior et al. 2002a). There is a difference in

28

the parameters calculated with different generations of the device. The parameters calculated using HRT 3 are smaller than those calculated with HRT II, due to a corrected scaling error (Gabriele et al. 2008).

The stereometric ONH parameters have been used for diagnosing glaucomatous damage (Wollstein et al. 1998). The cup shape measure has shown high diagnostic precision for glaucomatous damage (Mikelberg et al. 1995, Uchida et al. 1996, Vihanninjoki et al. 2000, Kiriyama et al. 2003). It has also been reported to be the stereometric parameter with the best correlation with glaucomatous progression in the visual fi elds (Harju & Vesti 2001, Philippin et al. 2006). The cup shape measure is independent of the reference plane.

Rim area is a stereometric ONH parameter that can be used in the detection of both glaucomatous damage and progression (Iester et al. 1997a, Lan et al. 2003, Tan & Hitchings 2004a, Strouthidis et al. 2006, Fayers et al. 2007). It is a reference plane-dependent parameter and subject to variability due to the fl uctuation of the reference height (Tan et al. 2003). However, the variability of the rim area has been reported to be smaller than the variability of the cup shape measure (Jampel et al. 2006). Small rim area has been shown to be among the most predictive parameters for future development of glaucoma in ocular hypertension, along with mean contour height and mean cup depth (Zangwill et al. 2005). Both rim area and the cup shape measure have shown high correlation with glaucomatous damage in the visual fi elds (Iester et al. 1997a). Global and sectoral rim area measurements have been shown to correlate with visual fi eld indices (Lan et al. 2003).

The cup:disc area ratio and the vertical cup:disc ratio are both among the parameters with relatively high accuracy for diagnosing glaucoma (DeLeon-Ortega et al. 2006, Ferreras et al. 2008, Bozcurt et al. 2010). An advantage of rim area and cup:disc ratio parameters is that they are familiar to clinicians. The vertical cup:disc ratio can also be estimated with ophthalmoscopy and stereoscopic ONH photography. The photographic evaluation and HRT measurement of the cup:disc area ratio may be used interchangeably in the risk analysis of the OHTS (Medeiros et al. 2007). Both rim area and cup:disc area ratio have shown statistically signifi cant change before confi rmed visual fi eld changes in ocular hypertensives converting to glaucoma (Kamal et al. 1999).

Using a combination of parameters rather than a single parameter improves the precision of the diagnostic test (Mikelberg et al. 1995, Iester et al. 2002, Ferreras et al. 2008). Linear discriminant functions taking into account several stereometric ONH parameters have been described (Mikelberg et al. 1995, Iester et al. 1997b, Iester et al. 1997c, Bathija et al. 1998, Burk et al. 1999, Mardin et al. 1999, Ferreras

29

et al. 2008). The discrimination between glaucoma patients and controls is not as good in other populations compared with the original population the functions are derived from (Ford et al. 2003, Ferreras et al. 2008).

Table 1. The linear discriminant functions for diagnosing glaucoma with stereometric

ONH parameters.

Study Linear discriminant function

Mikelberg et al. 1995 (-13.079 × ( cup shape measure + ( 0.001981 × ( 50 - age ) ) ) ) + (10.99 ×

volume above the reference plane) - (7.245 × height variation along contour

line) - 2.662

Iester et al. 1997c (10.068 × inferior area below reference) - (7.018 × inferior effective area)

+ (4.181 × nasal mean height contour) + (3.1 × temporal mean height

contour) - (2.081 × superior peak height contour) + (6.094 × cup shape

measure) - (11.048 × rim volume) + 1.828

Bathija et al. 1998 (-4.37 × cup shape measure) - (5.57 × height variation along contour line)

+ (11.78 × mean retinal nerve fi bre layer thickness) + (1.85 × rim area) -

3.722803

Burk et al. 1999 (4.197 × contour line height difference temporal - temporal superior) +

(5.642 × contour line height difference temporal - temporal inferior) - (3.885

× temporal superior cup shape measure) - 0.974

Mardin et al. 1999 (0.3 × rim area) + (3.7 × rim volume) + (4.3 × retinal nerve fi bre layer) - (3.7

× cup shape) - (3.1 × cup volume) - (0.9 × cup area) - 2.77

Ferreras et al. 2008 (9.41 × cup shape measure) - (8.00 × contour line modulation temporal-

superior) - (4.07 × rim area) + 8.23

2.3.2.3 Moorfi elds regression analysis

Moorfi elds regression analysis (MRA) is a method for detecting glaucomatous damage with the HRT. The MRA analyses the regression of the logarithm of the global and six sectoral rim areas to the matching disc areas and compares the results to a normative database (Wollstein et al. 1998). It defi nes these areas as normal, borderline and outside normal limits based on the 95% and 99.9% confi dence intervals. The method accurately discriminates between healthy controls and early glaucoma patients diagnosed using stereoscopic ONH photography (Wollstein et al. 1998) or visual fi elds (Ford et al. 2003, Miglior et al. 2003).

The diagnostic specifi city of the method is decreased by increasing size of the ONH in a non-glaucomatous population (Hawker et al. 2006). The diagnostic accuracy of MRA has been reported to be similar to stereoscopic ONH photographs

30

(Reus et al. 2007). As a screening method for glaucomatous damage, it has been reported superior to standardised automated perimetry and frequency doubling perimetry (Robin et al. 2005). It has shown potential to be a valid screening method also in high-risk populations (Harasymowycz et al. 2005). In the OHTS the development of glaucomatous damage was increased when baseline MRA showed at least one sector outside normal limits (Zangwill et al. 2005). However, the positive predictive value was low.

2.3.2.4 Glaucoma probability score

The glaucoma probability score (GPS) provides an automated interpretation of ONH topography. It is based on a three-dimensional model of a normal and a glaucomatous ONH. The model is constructed by combining the horizontal and vertical curvature of the RNFL with the steepness, size and depth of the cup (Swindale et al. 2000). The analysis is performed both globally and in six sectors. The outcome is determined by the sector with the highest probability score for glaucoma (Coops et al. 2006). The values range from 0 to 1 and represent the probability of glaucomatous damage. Values between 0.28 and 0.64 are considered borderline (Coops et al. 2006). The method is independent of the contour line and the reference plane, which reduces the sources of variability. The reproducibility of GPS is high, but an increase in variability is seen related to age, diminished image quality and diagnosis of glaucoma (Taibbi et al. 2009).

The diagnostic accuracy of GPS is comparable to MRA (Coops et al. 2006, Burgansky-Eliash et al. 2007, Ferreras et al. 2007, Reddy et al. 2009). The sensitivity is reported to be higher, but the specifi city lower with GPS (Harizman et al. 2006, Zangwill et al. 2007). Based on likelihood ratios, an abnormal MRA is most likely to confi rm the presence of a glaucomatous disc, while a normal GPS is most likely to confi rm a normal disc (Zangwill et al. 2007). Both methods have been reported to have higher sensitivity related to increasing disc size (Zangwill et al. 2007). With very large discs the diagnostic accuracy of the methods decreases (Hoesl et al. 2009). High GPS values have been shown to predict which glaucoma suspects develop glaucomatous changes in the visual fi elds and the ONH (Alencar et al. 2008). The linear regression of GPS scores may be used for determining glaucomatous progression with high specifi city, especially in the region of values representing low probability of glaucoma (GPS < 0.30) and very high probability of glaucoma (GPS > 0.78) (Strouthidis et al. 2010).

31

2.3.2.5 Topographic change analysis

The topographic change analysis (TCA) is a method for detecting glaucomatous progression. It estimates the probability that a difference in surface height occurs between baseline and follow-up images by chance alone. The analysis is performed in 4 x 4 pixel clusters called superpixels. The baseline for the comparison is constructed from two or preferably three images. Progression is identifi ed when change becomes larger than measurement variability (Chauhan et al. 2000). The method takes into account local variability. In areas of high variability, a large height change is needed to reach signifi cance. TCA is independent of the contour line and the reference plane (Chauhan et al. 2001).

The volume and area of a cluster of changing adjacent superpixels is calculated. The sensitivity and specifi city of different TCA cluster area and volume parameters for progression detected with photography and visual fi elds has been evaluated (Bowd et al. 2009). The specifi city was low for apparently non-progressing eyes. The agreement between stereoscopic ONH photographs and the TCA on detecting progression has been reported to be only fair (Kourkoutas et al. 2007).

2.3.2.6 Statistical image mapping and proper orthogonal decomposition

Statistical image mapping (SIM) is an analysis of the topographic height of each pixel compared with a statistical image generated by permutations of the measurements of pixel height (Patterson et al. 2005). It analyses the signifi cance level of change at each pixel using only the patient’s own imaging data. In addition to the single pixel change analysis, the signifi cance of the size of the changing cluster is also calculated. In a simulation model data set SIM showed better diagnostic precision for change than TCA (Patterson et al. 2005). The performance of the two methods was similar when compared in a real patient data set, but the sensitivity for progression found with stereoscopic ONH photographs was poor (O’Leary et al. 2008).

A new framework using a proper orthogonal decomposition (POD) method for detecting glaucomatous changes in ONH topography has been introduced. The mathematical method constructs a baseline subspace that contains all possible topographies, using observed baseline variability. The follow-up topographies are compared with the topographies in the baseline subspace for structural and geometrical similarity. If the baseline subspace accurately describes the follow-up topography, change is considered minimal (Balasubramanian et al. 2009). When applied to a clinical data set POD was slightly better in differentiating between progressing and

32

non-progressing glaucomatous ONH when compared to TCA. The difference between the methods was not statistically signifi cant (Balasubramanian et al. 2010). POD and SIM are not available in HRT software.

2.3.2.7 Measurement variability of the Heidelberg Retina Tomograph

The HRT has been reported to image ONH topography in a reproducible manner (Mikelberg et al. 1993, Rohrschneider et al. 1994, Verdonck et al. 2002). The test-retest variability of measurements is highest along the cup border. The variability increases with age and is higher in glaucoma patients compared to healthy controls (Chauhan et al. 1994). The variability of topography measurements and stereometric ONH parameters infl uence the detection of glaucomatous damage and especially the detection of glaucomatous progression (Artes & Chauhan 2005, Strouthidis & Garway-Heath 2008).

There are several factors that infl uence variability. The scan resolution is limited by the optical properties of the eye. The scanning laser passes through the tear fi lm, cornea, anterior chamber, lens and vitreous. All these media produce aberrations degrading the obtained topography image (Artal et al. 2001). Misalignment of the patient and the scanner (Orgul et al. 1996) and movement of the eye or head during image acquisition contribute to measurement noise. Changes in the distance between the eye and the scanner cause magnifi cation changes (Tan et al. 2004a). Cataract and pupil size may infl uence the obtained image (Zangwill et al. 1997). Variability due to cardiac cycle (Chauhan & McCormick 1995) and menstrual cycle (Yucel et al. 2005) has been related to changes in ocular blood fl ow.

Time separation between imaging does not seem to affect measurement variability (Chauhan & MacDonald 1995). The long-term fl uctuation of stereometric ONH parameters has been estimated to be similar to standardised automated perimetry (SAP) (Funk & Mueller 2003). In a longitudinal study, the relative variability and progression detection rate of rim area change with HRT was similar to the pattern standard deviation of SAP (Jampel et al. 2006).

Changes in IOP and cerebrospinal fl uid pressure may alter ONH topography (Parrow et al. 1992, Morgan et al. 2002). The changes in ONH topography related to IOP changes mostly occur with marked IOP changes after glaucoma surgery (Irak et al. 1996, Lesk et al. 1999), but there may be changes after medical treatment of IOP as well (Bowd et al. 2000). IOP-reducing medication may cause an increase in rim area. This effect has been reported to last at least a year (Tan & Hitchings 2004b). The decrease in cupping due to treatment has been identifi ed as a

33

protective factor against glaucomatous progression (Harju et al. 2008). However, no signifi cant change in the stereometric ONH parameters or with TCA was identifi ed in a group of glaucoma patients in whom IOP was elevated by discontinuing topical medication (Nicolela et al. 2006).

Poor visual acuity, age and high astigmatism have been reported to increase the variability of stereometric parameters (Sihota et al. 2002). Strouthidis and co-workers have concluded that different factors affecting image quality, including lens opacifi cation, age and astigmatism could be appropriately summarised by the topography standard deviation (TSD) (Strouthidis et al. 2005a). Image quality shows consistent infl uence on the variability of the stereometric ONH parameters (Strouthidis et al. 2005a). Follow-up rates may be adjusted according to image quality and different rates of rim area loss (Owen et al. 2006). The threshold of signifi cant change may be adjusted according to image quality (Fayers et al. 2007). The increasing severity of glaucomatous damage has been reported to decrease the repeatability of rim area. In the same study, the variability of vertical cup:disc ratio was not affected by the severity of the disease (DeLeon et al. 2007).

There is inter-observer variability in drawing the contour line (Wollstein et al. 1998). The placement of the contour line also affects the reference height of the standard reference plane. The contour line is automatically transferred from the baseline image to the follow-up images by the software. A major source of variability of the stereometric ONH parameters is the inter-test variability of the reference height (Tan et al. 2003, Strouthidis et al. 2005a, Breusegem et al. 2008, Lin et al. 2009). Improving the stability of the reference plane would decrease variability of the reference plane-dependent parameters. Repetitive testing and short inter-test intervals may improve change detection despite variability (Weinreb et al. 1993, Owen et al. 2006).

2.3.3 Retinal nerve fi bre layer photography

The retinal nerve fi bre layer (RNFL) may be visualised with red-free or green light, which does not penetrate the RNFL and is refl ected by the superfi cial layers of the retina (Behrendt & Wilson 1965). RNFL photographs can be obtained with a wide-angle fundus camera using a blue or green exciter fi lter with high-contrast fi lm (Airaksinen & Nieminen 1985). Currently, the images are captured digitally and may undergo image-processing to improve RNFL visibility and contrast (Tuulonen et al. 2000, Hwang et al. 2006). The RNFL bundles are seen as silvery striations, and with meticulous examination of the images, both localised and diffuse defects may

34

be detected (Tuulonen & Airaksinen 1991). The visibility of the RNFL is improved by a large pupil, relatively dark and evenly distributed fundus pigmentation and clear optical media (Teesalu & Airaksinen 1998). Especially yellow-brown coloration of the ageing lens decreases the quality of RNFL images (Siik et al. 1997).

Figure 4. Progression of superior and inferior retinal nerve fi bre layer defects detected

with RNFL photography. The corresponding changes in the optic nerve head can be

seen in Figure 2. The changes in the RNFL are substantially more prominent.

35

RNFL photography has been reported to be an effective diagnostic screening method for glaucoma (Quigley et al. 1980, Airaksinen et al. 1984, Wang et al. 1994, Niessen et al. 1997). Abnormal RNFL changes may also appear in other conditions affecting the optic nerve (Hoyt 1976) as well as in eyes considered normal (Airaksinen et al. 1984). The specifi city for glaucomatous damage has been estimated to be from 83% to 97% (Quigley et al. 1980, Airaksinen et al. 1984, Wang et al. 1994). There are several reports of RNFL damage, detected with photography, preceding both visual fi eld and ONH damage (Airaksinen & Alanko 1983, Sommer et al. 1991a, Tuulonen et al. 1993, Zeyen & Caprioli 1993, Quigley et al. 1994, Katz et al. 1997). There are no reports of a defect, once documented with RNFL photography, diminishing or disappearing.

The RNFL thickness measurements with OCT and scanning laser polarimetry are subject to variability (Iacono et al. 2006, Vizzeri et al. 2009a). A strong spatial correlation between diminished RNFL thickness measurements and localised defects in RNFL photography has been reported (Hwang et al. 2006). However, only 60 to 80% of localised defects found with RNFL photography could be identifi ed using OCT, scanning laser polarimetry or HRT (Windisch et al. 2009). RNFL and ONH photography every one to two years together with visual fi eld examinations every year has been recommended in the Finnish evidence-based guideline for open-angle glaucoma to reach a very good level in glaucoma follow-up (Tuulonen et al. 2003).

2.3.4 Scanning laser polarimetry

The RNFL thickness can be measured using scanning laser polarimetry. The most commonly used instrument is the GDx. The device uses a polarised laser beam that passes through the RNFL, which is a birefringent structure. A phase shift that correlates with RNFL thickness is measured from the refl ected light (Sehi et al. 2007). The anterior segment of the eye also has birefringent properties that have to be compensated for. Originally this compensation was fi xed, but the current generations of the device use variable cornea compensation (VCC) or enhanced cornea compensation (ECC) that allow for the compensation to be eye-specifi c. The individual compensation has improved the ability to detect glaucomatous damage (Brusini et al. 2005).

Unreliable measurements using scanning laser polarimetry may be caused by media opacities, ocular surface diseases, peripapillary atrophy and prior refractive surgery (Hoh et al. 1998). Especially GDx-VCC has a problem with atypical birefringerence patterns, caused by an attempted compensation of poor signal-to-noise ratio by the device (Bagga et al. 2005). In case an atypical birefringerence

36

pattern appears, the results of the scan are severely altered and do not match the expected retardation distribution based on RNFL anatomy.

The diagnostic accuracy of the GDx has been reported to be high, with areas under the ROC curve of the best parameters above 0.90 (Essock et al. 2005, Mai et al. 2007). A guided progression analysis method has been developed for the GDx. The sensitivity for glaucomatous progression detected with visual fi elds and ONH photography was only 50% (Alencar et al. 2010).

2.3.5 Optical coherence tomography

The optical coherence tomography (OCT) is an imaging technique based on low-coherence interferometry. It measures the time delay difference between the light refl ected from the retinal layers and a reference light (Huang et al. 1991). The method provides cross-sectional images of ocular structures and is used to quantify RNFL thickness. The axial resolution of time-domain OCT is around 10 μm (Schumann et al. 1995). OCT scans are affected by eye movements, media opacities and poor signal-to-noise ratio. Time-domain OCT has been used in most published studies. The recent introduction of spectral-domain OCT has made image acquisition faster, which improves resolution and reduces artefacts due to eye movement.

The diagnostic accuracy for detecting glaucomatous damage with time domain OCT is relatively good. The areas under the ROC curve of the best RNFL parameters have been reported to vary from 0.79 to 0.91 when detecting glaucomatous damage that has been verifi ed in the visual fi elds (Hong et al. 2007, Parikh et al. 2007, Pueyo et al. 2007). Progression rates have been reported to be higher when using OCT compared to visual fi elds (Wollstein et al. 2005).

2.3.6 Standardised automated perimetry

Standardised automated perimetry is considered the reference standard for detecting the presence and progression of glaucomatous visual fi eld defects (Foster et al. 2002, Tuulonen et al. 2003). The defi nition of a glaucomatous visual fi eld defect varies, even between major clinical trials (AGIS Investigators 1994, Schulzer 1994, Leske et al. 1999, Musch et al. 1999, Miglior et al. 2002b, Vesti et al. 2003, Keltner et al. 2005). There are features considered typical for a glaucomatous defect. These include a depression in adjacent test points or a severe depression of one point, localised nature of the defect, a difference across the horizontal meridian and reproducibility of the defect (AGIS Investigators 1994, Schulzer 1994, Leske et al. 1999, Musch et al. 1999, Miglior et al. 2002, Keltner et al. 2005).

37

There is also variation in the defi nition of glaucomatous progression (Schulzer 1994, AGIS Investigators 1998, Chauhan et al. 1999, Musch et al. 1999, Heijl et al. 2002, Vesti et al. 2003). Substantial differences in progression rates using different methods have been reported (Vesti et al. 2003). Variability of the visual fi elds increases progression rates of some methods while decreasing progression rates of others (Vesti et al. 2003). A glaucoma progression analysis (GPA), based on the criteria developed for the EMGT, has been introduced for the Humphrey perimeter (Heijl et al. 2003). The GPA software considers the random variability of visual fi eld analyses from two baseline fi elds and identifi es progressive visual fi eld loss exceeding the normal level of test-retest variability. The software identifi es test points with a signifi cant change from baseline.

2.4 Diagnosing glaucoma with the Heidelberg Retina Tomograph

The quality of reporting diagnostic studies using the HRT is suboptimal (Shunmugam & Azuara-Blanco 2006). For example, more than 90% of the studies did not report how many of the participants who fulfi lled the inclusion criteria did or did not undergo HRT examination or the test used as the reference standard (Shunmugam & Azuara-Blanco 2006).

The methods to detect glaucomatous ONH damage using the HRT include the sectoral and global stereometric ONH parameters, the linear discriminant functions calculated from these parameters, MRA and GPS. The sensitivity and specifi city for glaucoma varies depending on the study population, the defi nition of glaucoma and the diagnostic method. The sensitivity for glaucoma has been reported to vary from 56% to 81% when calculated at 80% specifi city; from 39% to 75% at 90% specifi city and from 38% to 67% at 95% specifi city, when using predetermined methods of analysing glaucomatous damage with HRT (Ford et al. 2003, Medeiros et al. 2004, Zangwill et al. 2007, Ferreras et al. 2008, Hoesl et al. 2009).

A bagging classifi cation tree approach using stereometric parameters has shown 82% sensitivity at 89% specifi city (Mardin et al. 2003). A method using the mean contour height at 36 sectors of the ONH has shown 86% specifi city at 90% sensitivity (Zangwill et al. 2004). The results of these studies have not been confi rmed in independent populations. The discrimination between glaucoma patients and controls has been reported to be lower in independent populations compared to original studies describing the diagnostic methods (Ford et al. 2003).

The diagnostic studies are affected by the varying defi nition of glaucoma (Miglior et al. 2005) and the classifi cation of glaucoma suspects as either glaucomatous or

38

healthy (Miglior et al. 2003). The diagnostic accuracy of the HRT is affected by the size of the ONH. The sensitivity tends to be lower and the specifi city higher with small discs (Ford et al. 2003). The diagnostic accuracy is decreased by unusually small and very large discs (Iester et al. 1997b, Hoesl et al. 2009).

The diagnostic accuracy of the HRT has been reported similar to OCT, GDx and expert evaluation of stereoscopic ONH photographs (Greaney et al. 2002, Medeiros et al. 2004, DeLeon-Ortega et al. 2006, Badala et al. 2007, Reus et al. 2007). Using a combination of RNFL and ONH topography measurements has been shown to improve the sensitivity to 91% at a specifi city of 96% (Badala et al. 2007). However, the results were not calculated using predefi ned diagnostic criteria and they have not been confi rmed in independent populations.

2.5 Detecting glaucomatous progression with the Heidelberg Retina Tomograph

There are several studies evaluating the ability of the HRT to detect glaucomatous progression. However, there is a lack of both a golden standard for progression and a generally agreed method of detecting progression using the HRT. TCA is a method designed for detecting progression, but the statistically signifi cant number, area and volume of changing adjacent superpixels is not clear. As a result, several defi nitions for progression have been used (Table 2). A cluster of at least 20 depressing superpixels confi rmed by three follow-up exams has been used based on the work of Chauhan and co-workers (Chauhan et al. 2001). It is based on empirical data and the statistical and clinical signifi cance of this defi nition for progression is not known.

The reference standards used in the studies assessing the accuracy of HRT to detect progression include different visual fi eld criteria, expert evaluation of stereoscopic ONH photographs or both (Table 2). The concordance between methods for detecting progression is quite low (Table 2). There may be several reasons for this discrepancy. In the presence of advanced glaucomatous damage in the visual fi elds or the ONH, progression within the remaining capacity for change may not reach the criteria of signifi cant change (Hudson et al. 2007). Progression in the visual fi elds may be covered by learning effect (Hudson et al. 2007). The measurement variability of the methods may differ or there may be temporal dissociation between measurable progression (Strouthidis & Garway-Heath 2008). There are reports of structural changes detected with HRT that are followed by functional changes in the visual fi elds (Kamal et al. 1999, Chauhan et al. 2009b). The correlation between the changes in the visual fi elds and structural glaucomatous progression, evaluated with photography, has been reported to be only 35% on average (Tuulonen et al. 2003).

39

Ta

ble

2.

Stu

die

s e

va

lua

tin

g t

he

de

tec

tio

n o

f g

lau

co

ma

tou

s p

rog

res

sio

n u

sin

g H

RT

wit

h p

red

efi

ne

d c

rite

ria

fo

r p

rog

res

sio

n.

Stu

dy

Popula

tion (

n)

Refe

rence

sta

ndard

HR

TC

rite

ria for

pro

gre

ssio

nS

ensi

tivity

Specifi c

ityF

ollo

w-u

p

time

Chauhan e

t al.

2001

Gla

uco

ma

(77)

SA

P

(4 lo

catio

ns)

IT

CA

(20 s

uperp

ixels

)88%

40%

5.5

years

Nic

ole

la e

t al.

2003

Gla

uco

ma

(48)

SA

P

(3 lo

catio

ns)

IT

CA

(20 s

uperp

ixels

)74%

32%

6.5

years

Art

es

& C

hauhan 2

005

Gla

uco

ma

(84)

SA

P

(1 lo

catio

n)

IT

CA

(6%

of d

isc

are

a)

57%

69%

7.4

years

Art

es

& C

hauhan 2

005

Gla

uco

ma

(84)

SA

P (

3 lo

catio

ns)

IT

CA

(18%

of dis

c are

a)

27%

86%

7.4

years

Str

outh

idis

et al.

2006

Ocu

lar

hyp

ert

ensi

on

(198)

SA

P

(1 lo

catio

n)

IS

ect

ora

l rim

are

a r

egre

ssio

n

(hig

h v

ariabili

ty)

37%

69%

6.0

years

Str

outh

idis

et al.

2006

Ocu

lar

hyp

ert

ensi

on

(198)

SA

P

(1 lo

catio

n, c

onfi r

med x

3)

IS

ect

ora

l rim

are

a r

egre

ssio

n (

low

variabili

ty)

18%

89%

6.0

years

Phili

ppin

et al.

2006

Ocu

lar

hyp

ert

ensi

on

(109)

SA

P

(2 lo

catio

ns

or

MD

change)

IC

up s

hape m

easu

re r

egre

ssio

n44%

59%

10 y

ears

Phili

ppin

et al.

2006

Ocu

lar

hyp

ert

ensi

on

(109)

SA

P

(2 lo

catio

ns

or

MD

change)

IR

im a

rea r

egre

ssio

n22%

67%

10 y

ears

Kourk

outa

s et al.

2007

Gla

uco

ma (

54)

ON

H s

tere

o-p

hoto

gra

phs

IIT

CA

(20 s

uperp

ixels

)70%

64%

2.6

years

Bow

d e

t al.

2009

Health

y / G

lauco

ma

(267)

SA

P a

nd /or

photo

gra

phs

IIT

CA

(are

a ≥

0,0

36 m

m2)

78%

80%

3.9

years

Viz

zeri e

t al.

2009b

Gla

uco

ma a

nd s

usp

ect

(237)

ON

H s

tere

o-p

hoto

gra

phs

IIT

CA

(eva

luatio

n b

y exp

erience

d

obse

rvers

)

50%

85%

4.6

years

Viz

zeri e

t al.

2009b

Gla

uco

ma a

nd s

usp

ect

(237)

ON

H s

tere

o-p

hoto

gra

phs

IIT

rend a

naly

sis

of st

ere

om

etr

ic

para

mete

rs

13%

89%

4.6

years

Viz

zeri e

t al.

2009b

Gla

uco

ma a

nd s

usp

ect

(237)

ON

H s

tere

o-p

hoto

gra

phs

IIM

RA

(eva

luatio

n b

y exp

erience

d

obse

rvers

)

31%

89%

4.6

years

Str

outh

idis

et al.

2010

Ocu

lar

hyp

ert

ensi

on

(198)

SA

P

(1 lo

catio

n)

IG

PS

(posi

tive r

egre

ssio

n s

lope)

31%

92%

6.0

years

The s

ensi

tivity

and s

pecifi c

ity v

alu

es

are

pre

sente

d in

a s

um

mary

rece

iver

opera

ting c

hara

cterist

ics

plo

t (F

igure

5).

HR

T =

Heid

elb

erg

Retin

a T

om

ogra

ph, S

AP

= S

tandard

ised A

uto

mate

d P

erim

etr

y, M

D =

Mean D

evi

atio

n, T

CA

= T

opogra

phic

Change A

naly

sis,

MR

A =

Moorfi

eld

s R

egre

ssio

n a

naly

sis,

G

PS

= G

lauco

ma P

robabili

ty S

core

40

There is relatively poor concordance between progression detected with stereoscopic ONH photography and HRT (Table 2). Agreement on assessments within a method as well as measurement variability affects the evaluation of progression (Strouthidis & Garway-Heath 2008). It has been suggested that the evaluation of ONH using HRT and the stereoscopic ONH photographs assess different aspects of ONH change (O’Leary et al. 2008, Vizzeri et al. 2009b). It has been reported that the HRT using TCA performs at least as well as experts evaluating ONH photographs in determining progression (Chauhan et al. 2009a).

Figure 5. The summary receiver operating characteristics plot for the detection of glau-

comatous progression using HRT. The sensitivities and specifi cities are reported in

studies using predefi ned criteria for progression (Table 2). The reference standards for

progression include standardised automated perimetry (black diamonds), stereoscopic

ONH photography (black circles) or both (black triangle).

41

3 Aims of the research

The purpose of the present study was to assess the stereometric optic nerve head parameters of the HRT in the follow-up of glaucoma. The objective was to determine how the change in the stereometric parameters correlates to glaucomatous progression (I, II) and the factors that may affect the correlation with progression (III, IV). More specifi cally, the aims of the research were:

1. to investigate how change in the stereometric ONH parameters correlates to the small topographic changes associated with progression of the RNFL defects (I)

2. to assess the accuracy of the stereometric ONH parameters for detecting glaucomatous progression found with stereoscopic ONH photography (II)

3. to evaluate the factors affecting the sensitivity and specifi city of the stereometric parameters to glaucomatous progression in ONH photographs (III)

4. to determine whether exercise affects the stereometric parameters representing ONH topography (IV)

43

4 Subjects and methods

4.1 Subjects

The data collection for the retrospective case-control study (I) was performed in the following manner. From the subjects monitored at Oulu University Hospital Glaucoma Clinic, we selected those who had had at least three consecutive yearly examinations with the HRT and successful RNFL photography at each visit. Patients with a fundus disease that disturbed the interpretation of RNFL photographs as well as subjects with unsuccessful HRT examinations or without open-angle glaucoma were excluded. Progression of the RNFL defect was identifi ed in 34 eyes of 29 patients. 34 non-progressive eyes matched for age and diagnosis were selected to serve as controls (Table 3).

Table 3. Demographic details (I).

Eyes with progression Controls

Number of eyes (% right) 34 (55.9) 34 (47.1)

Diagnosis (poag/ntg/pex)* 16 / 15 / 3 16 / 15 / 3

Mean age (yrs) 60.8 (SD 10.2) 60.8 (SD 10.7)

Mean time between fi rst and third visits (years) 2.03 (SD 0.29) 2.06 (SD 0.25)

Number of patients (% male) 29 (55.1%) 22 (22.7%)

*poag = primary open-angle glaucoma, ntg = normal tension glaucoma, pex = glaucoma with pseudo-

exfoliation

Mean age and mean time between visits are calculated for eyes.

For the retrospective follow-up study (II, III) all eyes with at least 18 months of follow-up with HRT and stereoscopic ONH photographs were reviewed. A total of 1030 eyes of 532 patients met the initial follow-up criteria. The photographs were approved for image quality by three masked evaluators. The HRT images had topography standard deviations (TSD) of < 36 μm; the TSD values could not vary more than 10 μm between visits, and image alignment was to be perfect. Figure 6 presents the fl ow chart summarising study exclusion.

Progression in the stereoscopic ONH photographs was assessed by three glaucoma specialists in a masked fashion. A two-out-of-three agreement was used for progression and three-out-of-three agreement for non-progression. 476 eyes of 342 patients were included in the fi nal analysis. The demographic details are given in Table 4.

44

Figure 6. Flow chart summarising study exclusion (II).

45

Table 4. Demographic details (II).

Variable Value

Mean age (range), years 62 (13–86)

Mean follow-up (range), months 28 (18–56)

Mean disc area (range), mm2 2.4 (1.2–4.7)

Male-to-female ratio 147:195

Laterality (right-to-left) 249:227

35 healthy volunteers with no systemic or ocular disorders were recruited for the prospective study (IV). Subjects between 18 and 40 years of age with spherical equivalents between -7 and +5 diopters in the study eye (right eye) were accepted. 5 subjects did not achieve the predetermined 30 mmHg increase in systolic blood pressure during exercise and were excluded. The demographic details are given in Table 5.

Table 5. Demographic details (IV).

Variable Value

Mean age (range), years 27 (20–36)

Mean spherical equivalent (range), diopters - 1.4 (-6.0 – +4.5)

Male-to-female ratio 12:18

Disc area (range), mm2 1.91 (1.26–2.75)

Cup:disc area ratio (range) 0.15 (0.00–0.37)

4.2 Stereoscopic ONH photography

The stereoscopic ONH photographs were obtained with a Canon CF-60 UVi funduscamera with a Canon EOS 1 D MK II, 8 Mpixel Digital SLR CMOS camera (Canon Inc, Tokyo, Japan). The patients’ pupils were dilated for the photography. The stereo-effect was obtained by lateral shift of the camera. The photographs are digital black-and-white images in stereo-pairs. When both pictures of a pair are viewed simultaneously, a three-dimensional image of the ONH can be seen.

4.3 Scanning laser ophthalmoscopy

The Heidelberg Retina Tomograph II with software version 1.6 was used to obtain three-dimensional topographic images of the ONH. A contour line was drawn at the inner edge of the scleral ring on the fi rst image, and it transferred automatically onto the next images. All contour lines were drawn by an experienced photographer

46

trained for this purpose. The stereometric ONH parameters are calculated from the topographic image. In the prospective study (IV), software version 3.0 was used to recalculate stereometric ONH parameters and the drawing of the contour lines was performed automatically by the software.

4.4 Retinal nerve fi bre layer photography

The RNFL photographs were taken with a monochromatic blue interference fi lter (495 nm). The present technique is refi ned from a well-documented technique (Airaksinen & Nieminen 1985, Tuulonen et al. 2000) used in our glaucoma clinic for more than two decades. The RNFL photographs are digital images acquired with a Canon CF-60 UVi fundus camera with a Canon EOS 1 D MK II, 8-Mpixel Digital SLR CMOS camera (Canon Inc., Tokyo, Japan) using Neacapture software (Neagen Oy, Oulu, Finland) and processed with Adobe Photoshop CS (Adobe Systems Inc., San Jose, CA, USA). The RNFL photographs were taken and processed by three photographers, each with more than 10 years of experience in fundus photography in our clinic.

4.5 Examinations during exercise

In the prospective study (IV), the HRT examinations and the measurements of heart rate, blood pressure and IOP were obtained twice before exercise (baseline and pre-exercise). The ONH imaging took place during the measuring of blood pressure, in the short period between systolic and diastolic pressure measurements. IOP was measured using the Goldmann applanation tonometer. Exercise was carried out using an exercise bike with an adjustable workload. Heart rate and blood pressure were monitored. After an increase of at least 30 mmHg in systolic blood pressure was achieved, the examinations were performed for the third time (mid-exercise). Exercise was then discontinued and the subject was monitored until blood pressure normalised. The examinations were subsequently repeated (post-exercise).

4.6 Statistical analysis

Sensitivity for progression is defi ned as the proportion of eyes with progression in the reference standard (RNFL or ONH photography) that have a positive test. If the change in the parameter or linear discriminant function value is above the cut-off value, the test is positive. Specifi city is defi ned as the proportion of eyes

47

with no progression in the reference standard that have a negative test. A plot of sensitivity versus one minus specifi city (a receiver operating characteristic curve) is used to determine the quality of a diagnostic test. The areas under the receiver operating characteristic (ROC) curve were compared using MedCalc, Version 11.2 (MedCalc Software, Mariakerke, Belgium) when calculating the effect of image quality on the ROC curves for detecting glaucomatous progression (Results 5.5). A detailed description of the other statistical methods used can be found in the original publications (I–IV).

4.7 Ethical considerations

Approval for all studies was obtained in accordance to the guideline of the Ethical Committee of the Northern Ostrobothnia Hospital District. In the prospective study (IV), all subjects signed an informed consent form explaining the study protocol. The studies followed the tenets of the Declaration of Helsinki.

49

5 Results

5.1 Correlation of the change in the stereometric parameters with progression of the RNFL defects (I)

The change in only one stereometric ONH parameter, the cup shape measure, showed a statistically signifi cant (p = 0.049) correlation with the progression of the RNFL defect. A change of 0.015 in the cup shape measure had a sensitivity of 53% and specifi city of 74% for progression.

The best combination of two parameters found with multiple logistic regression analysis was ∆ [(-7.219 × maximum cup depth) + (21.639 × linear cup:disc area ratio)]. It showed a statistically signifi cant correlation with progression of the RNFL defect (p = 0.009). A change of 0.1372 in the two-parameter combination value had 71% sensitivity and 74% specifi city.

Figure 7. The area under the receiver operating characteristics (ROC) curve calculated

for the best three parameter combination [(-7.121 × maximum cup depth) + (-3.801 ×

horizontal cup:disc ratio) + (28.091 × linear cup:disc area ratio)] :0.753 ; the best two

parameter combination [(-7.219 × maximum cup depth) + (21.639 × linear cup:disc area

ratio)] :0.724 and the cup shape measure :0.617.

The three-parameter combination with the best correlation with progression was ∆ [(-7.121 × maximum cup depth) + (-3.801 × horizontal cup:disc ratio) + (28.091 × linear cup:disc area ratio)]. The correlation with progression was statistically signifi cant (p = 0.007). A change of 0.1602 in the three-parameter combination

50

value had 77% sensitivity and 79% specifi city. The ROC curves of the cup shape measure, the two-parameter and the three-parameter combination are presented in Figure 7.

5.2 Correlation of the change in the stereometric parameters with progression in stereoscopic ONH photographs (II)

The cup:disc area ratio had the best correlation with progression in stereoscopic ONH photographs (p < 0.0005). The cut-off value optimised for sensitivity and specifi city was 0.005 for change in the cup:disc area ratio. Sensitivity was 75% and specifi city 56% at the cut-off. Sensitivity was 28% at a fi xed specifi city of 90%.

Figure 8. The receiver operating characteristics (ROC) curves calculated for the linear

discriminant function, the cup:disc area ratio, and the vertical cup:disc ratio. The areas

under the ROC curves were 0.694, 0.666 and 0.724, respectively.

The parameter with the largest area under the ROC curve (0.726) was the vertical cup:disc ratio. A change of 0.007 in the vertical cup:disc ratio had a sensitivity of 80% and a specifi city of 65%. Sensitivity was 22% at a fi xed specifi city of 90%.

51

Figure 9. Box plot demonstrating the changes in the linear discriminant function com-

pared between the groups of non-progressors and progressors. The cut-off values are

set at 90% specifi city, and at the value optimised for sensitivity and specifi city. The box-

es represent inter-quartile range (IQR) between the 75th and 25th percentiles. The out-

liers (°) are values between 1.5 IQRs and 3 IQRs from the end of a box. Values of more

than 3 IQRs from the end of a box are defi ned as extreme (*).

The linear discriminant function with the best correlation with progression was [(12.241 × cup:disc area ratio) + (3.540 × mean cup depth) - (2.146 × horizontal cup:disc ratio) + (27.486 × average variability)]. An optimised change in the linear discriminant function value was 0.34 with a sensitivity of 65% and specifi city of 69%. Sensitivity was 31% at a fi xed specifi city of 90%. With the cut-off at 90% specifi city, the kappa value for inter-method agreement was 0.211 with stereoscopic ONH photography. This indicates only fair agreement. The results are summarised in Figures 8 and 9.

52

5.3 Factors affecting the sensitivity and specifi city of the stereometric parameters to glaucomatous progression in ONH photographs (III)

The factors having the most signifi cant effect on the sensitivity and specifi city of the stereometric ONH parameters were the reference height difference and the mean topography standard deviation (TSD), indicating image quality. These two factors combined accounted for 21% of the increase in misjudging progression (r2) when calculated for the cup:disc area ratio. The change in the TSD and age also showed a consistent, but variably signifi cant infl uence on all parameters tested. The sensitivity and specifi city improved when there was little change in the reference height, the image quality was good and stable, and the patients were younger.

The sensitivity and specifi city of the vertical cup:disc ratio was improved by a large disc area and high baseline cup:disc area ratio. The rim area showed a better sensitivity and specifi city for progression with a small disc area and low baseline cup:disc area ratio. Gender, the laterality of the eye, the diagnosis of glaucoma, and the duration of follow-up had no statistically signifi cant effect. The results are summarised in Table 6.

5.4 The effect of exercise on the stereometric ONH parameters (IV)

Exercise was associated with an increase in variance in 17 of the 18 stereometric ONH parameters. The increase in variance was statistically signifi cant in eight parameters including rim area, cup:disc area ratio and cup shape measure. There was no statistically signifi cant change in image quality.

53

Ta

ble

6.

C

orr

ela

tio

n c

oe

ffi c

ien

ts a

nd

sta

tis

tic

al

sig

nifi

ca

nc

e o

f fa

cto

rs a

na

lys

ed

fo

r a

ffe

cti

ng

th

e s

en

sit

ivit

y a

nd

sp

ec

ifi c

ity

of

ste

reo

me

tric

ON

H p

ara

me

ters

.

Fact

or

Ste

reom

etr

ic O

NH

para

mete

rS

tatis

tical m

eth

od

Vert

ical c

up:d

isc

ratio

Cup:d

isc

are

a r

atio

Cup v

olu

me

Rim

are

aLin

ear

dis

crim

inant

funct

ion

Change in

refe

rence

heig

ht

.296

**.3

91

**.3

64

**.3

58

**.3

04

**P

ears

on C

orr

ela

tion

<.0

005

<.0

005

<.0

005

<.0

005

<.0

005

p-v

alu

e

Mean T

SD

.268

**.3

04

**.1

71

*.3

08

**.2

43

**P

ears

on C

orr

ela

tion

.001

<.0

005

.046

<.0

005

.003

p-v

alu

e

Age

.187

*.2

31

**.1

22

.163

*.0

97

Pears

on C

orr

ela

tion

.018

.001

.157

.021

.240

p-v

alu

e

Change in

TS

D.1

87

*.0

66

.137

.106

.194

*P

ears

on C

orr

ela

tion

.018

.350

.113

.134

.018

p-v

alu

e

Dis

c are

a-.

217

**-.

085

.121

.215

**.0

16

Pears

on C

orr

ela

tion

.006

.226

.159

.002

.845

p-v

alu

e

Base

line c

up:d

isc

are

a r

atio

-.232

**.0

57

.348

**.1

78

*.0

95

Pears

on C

orr

ela

tion

.003

.422

<.0

005

.011

.251

p-v

alu

e

Base

line r

im a

rea

.017

-.123

-.213

*.0

30

-.075

Pears

on C

orr

ela

tion

.834

.080

.013

.669

.361

p-v

alu

e

Tim

e b

etw

een v

isits

.088

-.072

-.101

-.128

-.074

Pears

on C

orr

ela

tion

.270

.305

.240

.071

.370

p-v

alu

e

Gender

Mann –

Whitn

ey

U

.239

.860

.855

.926

.203

p-v

alu

e

Late

ralit

y (r

ight/le

ft)

Mann –

Whitn

ey

U

.726

.846

.762

.892

.835

p-v

alu

e

Dia

gnosi

s of gla

uco

ma

Kru

skal –

Walli

s

.922

.239

.961

.493

.517

p-v

alu

e

TS

D, to

pogra

phy

standard

devi

atio

n

**. C

orr

ela

tion is

sta

tistic

ally

sig

nifi

cant. P

-valu

e ≤

0.0

1.

*. C

orr

ela

tion is

sta

tistic

ally

sig

nifi

cant. P

-valu

e ≤

0.0

5.

54

Parameter Variance of change

before exercise

(pre-exercise –

baseline)

Variance of change

during exercise

(mid-exercise –

baseline)

Statistical

signifi cance

(P-value)

Cup area 0.510 1.294 0.0080*

Rim area 0.510 1.294 0.0080*

Cup:disc area ratio 0.0930 0.243 0.0066*

Rim:disc area ratio 0.0930 0.243 0.0066*

Cup volume† 0.0369 0.0323 0.36

Rim volume 0.983 2.373 0.011*

Mean cup depth 0.109 0.201 0.027*

Maximum cup depth 0.637 1.389 0.020*

Height variation contour 1.059 1.744 0.086

Cup shape measure 0.303 0.910 0.0024*

Mean RNFL thickness 0.398 0.576 0.16

RNFL cross sectional area 9.598 13.715 0.17

Horizontal cup:disc ratio 1.350 2.251 0.091

Vertical cup:disc ratio 0.451 0.759 0.087

Maximum contour elevation 0.370 0.548 0.12

Maximum contour depression 0.528 0.648 0.30

CLM temporal-superior 0.377 0.642 0.080

CLM temporal-inferior 0.460 0.575 0.28

*The change is statistically signifi cant (p < 0.05). † There was decrease in the variance in one

parameter. The decrease was not statistically signifi cant.

Table 7. The variance of the change (×103) from baseline of the stereometric ONH pa-

rameters before and during exercise. The statistical signifi cance of the change in vari-

ance is calculated.

IOP decreased and blood pressure increased during exercise, resulting in an increase in mean ocular perfusion pressure. The absolute change in rim area, cup area and rim volume showed a statistically signifi cant correlation with change in mean ocular perfusion pressure. The results are summarised in Tables 7 and 8.

55

Parameter Pearson correlation Statistical signifi cance

(P-value)

Cup area 0.194 0.033*

Rim area 0.194 0.033*

Cup:disc area ratio 0.172 0.052

Rim:disc area ratio 0.172 0.052

Rim volume 0.202 0.028*

Mean cup depth 0.014 0.447

Maximum cup depth 0.128 0.114

Cup shape measure 0.071 0.253

The correlation is calculated for absolute changes from baseline before, during and after exercise.

*The change is statistically signifi cant (p < 0.05).

Table 8. The correlation between the change in mean ocular perfusion pressure and the

stereometric ONH parameters with statistically signifi cant increase in variance. The sta-

tistical signifi cance of the correlation is calculated.

5.5 The effect of image quality on agreement for detecting glaucomatous progression

The eyes in the retrospective follow-up study (II) were divided into three subgroups based on the mean image quality of the HRT examinations. The ROC curves of three stereometric parameters (cup:disc area ratio, rim area and cup volume) were compared between the three subgroups. With better image quality, there was a considerable improvement in the areas under the ROC curve calculated for progression detected with stereoscopic ONH photographs (Figure 10). The areas under the ROC curves of each parameter were compared between images of different quality (A and C). The difference was statistically signifi cant for cup volume with a p-value of 0.016.

With excellent image quality, the area under the ROC curve was best for the change in cup volume. A cut-off value of 0.030 mm3 was determined for cup volume based of the ROC curve. With this cut-off sensitivity was 86% and specifi city 90% for progression in stereoscopic ONH photographs. The inter-method agreement (kappa) for progression detected with photographs was 0.501. The inter-observer agreement between the masked evaluators of the stereoscopic ONH photographs for progression was moderate, with kappa values ranging from 0.403 to 0.510.

56

A

B

CFigure 10. The effect of image quality on the receiver operating characteristics (ROC)

curves for detecting glaucomatous progression in stereoscopic ONH photographs.

A. Excellent image quality, mean topography standard deviation (TSD) < 11 μm (n = 75).

The area under the ROC curve for rim area is 0.773, for cup:disc area ratio 0.784 and for

cup volume 0.861.

B. Very good image quality, 11 μm ≤ TSD ≤ 16 μm (n = 232). The area under the ROC

curve for rim area is 0.712, for cup:disc area ratio 0.725 and for cup volume 0.703.

C. From very good to acceptable image quality, 16 μm < TSD ≤ 34 μm (n = 169). The area

under the ROC curve for rim area is 0.537, for cup:disc area ratio 0.548 and for cup vol-

ume 0.560.

57

6 Discussion

6.1 Linear discriminant functions

When diagnosing glaucomatous damage with HRT, it is more accurate to use a combination of the stereometric parameters rather than just one parameter (Mikelberg et al. 1995, Ferreras et al. 2008). The correlation with progression in the RNFL was also improved by a linear discriminant function combining several parameters. (I)

The linear discriminant function [(-7.121 × maximum cup depth) + (-3.801 × horizontal cup:disc ratio) + (28.091 × linear cup:disc area ratio)] had the highest sensitivity (77%) and specifi city (79%) for glaucomatous progression in RNFL photographs. (I) The linear discriminant function [(12.241 × cup:disc area ratio) + (3.540 × mean cup depth) - (2.146 × horizontal cup:disc ratio) + (27.486 × average variability)] had the best correlation with progression in stereoscopic ONH photographs (II). It has been suggested that the parameters, which provide good sensitivity and specifi city in recognising the presence or absence of glaucomatous damage, may be different from those recognising progression of the disease (Anderson et al. 2000). The parameters improving diagnostic precision in several of the reported LDFs include the cup shape measure, rim area, rim volume, the height of the contour line, and the height variation along the contour line (Table 1). The parameters in the LDFs recognising progression are different from those recognising the presence or absence of glaucoma.

The enlargement of the cup:disc area ratio is a classic sign of glaucomatous damage. Increase in the cup:disc ratio in the vertical direction was highly correlated with progression in stereoscopic ONH photographs. Both LDFs calculated for detecting progression (I, II) included a cup:disc area ratio parameter with a large positive coeffi cient and a horizontal cup:disc ratio parameter with a small negative coeffi cient. This further implies that if one focuses on the cupping occurring away from the horizontal plane, it enhances the correlation with progression, regardless of the reference standard used. This may partly be explained by the fact that the nasal area and the papillomacular bundles located temporally are the last parts of the ONH that are affected by glaucomatous damage (Quigley et al. 1982, Caprioli 1989).

58

6.2 Correlation between the stereometric ONH parameters with progression in RNFL photographs (I)

The glaucomatous changes in the ONH anatomy result from the loss of ganglion cell axons, which is typically fi rst seen in the retinal nerve fi bre layer (Airaksinen & Alanko 1983, Sommer et al. 1991a). However, it is yet to be established whether the changes in the ONH are measurable at the stage when the loss of axons fi rst becomes visible in the RNFL. Only one stereometric ONH parameter, the cup shape measure, showed a barely statistically signifi cant (p = 0.049) correlation with progression detected with RNFL photography. Reaching statistical signifi cance is affected by limited sample size. There may also be temporal dissociation in measurable progression between methods. The small changes in topography may become undetectable due to variability in the stereometric parameters, especially when analysing an event of progression, instead of performing trend analysis.

6.3 Correlation between the stereometric parameters with progression in ONH photographs (II)

The progression in the stereoscopic ONH photographs had a 2-out-of-3 agreement criterion and for non-progression a 3-out-of-3 criterion. Both the stereoscopic photographs and the HRT measure change in the ONH. There is no potential time delay in measurable progression between the methods as is the case when comparing changes in the ONH and the RNFL.

Despite strict inclusion criteria for image quality and alignment excluding 46% of the original study population, there was a considerable lack in concordance between the two methods (Figure 9). The change in most parameters had a statistically signifi cant correlation with progression. Reaching statistical signifi cance is improved by the fairly large sample size. However, the sensitivity and specifi city values of the parameters leave much to be desired, as can be seen from the ROC curves (Figure 8).

6.4 Factors affecting the sensitivity and specifi city to progression in ONH photographs (III)

The factors that infl uence the short-term variability of the stereometric parameters include inter-test reference height difference, image quality, lens opacifi cation, age, poor visual acuity and astigmatism (Sihota et al. 2002, Strouthidis et al. 2005a,

59

Breusegem et al. 2008). Strouthidis and co-workers concluded that different factors affecting image quality, including lens opacifi cation, age and astigmatism could be appropriately summarised by topography standard deviation (Strouthidis et al. 2005a).

The factors affecting the sensitivity and specifi city of the stereometric ONH parameters to glaucomatous progression in stereoscopic ONH photographs are essentially the same as those factors related to short-term variation of the parameters (Strouthidis et al. 2005a). This suggests that improving the methods of controlling the variability of the stereometric parameters would improve the agreement between stereometric ONH parameters of the HRT and stereoscopic ONH photographs for progression.

The rim area showed better sensitivity and specifi city for progression with a small disc area and low baseline cup:disc area ratio. The sensitivity and specifi city of the vertical cup:disc ratio was improved by a large disc area and high baseline cup:disc area ratio. This is in concordance with the results of DeLeon-Ortega and co-workers. They found that rim area variability increases with disease severity, but the variability of the vertical cup:disc ratio remains stable, even at stages of severe glaucomatous visual fi eld loss (DeLeon-Ortega et al. 2007). Therefore it may be optimal to follow the change of different parameters depending on the size of the ONH and the stage of glaucomatous damage.

6.5 Variability of ONH topography during exercise (IV)

Changes during the cardiac cycle as well as changes in IOP may result in variation of ONH topography (Chauhan & McCormick 1995, Azuara-Blanco et al. 1998). There may be a change in vessel position and calibre during a cardiac cycle, but also a forward and backward shift of the ONH, independent of the vessels (Chauhan & McCormick 1995). Blood pressure is increased and IOP is decreased during exercise, resulting in increase in mean ocular perfusion pressure. Study IV shows some association between increasing variability in ONH topography and exercise-induced increase in mean ocular perfusion pressure.

The exercise-induced decrease in IOP is also seen in the elderly (Erä et al. 1993), a major age group in glaucoma follow-up. Climbing stairs or walking to the offi ce may be physically demanding for the elderly. ONH imaging should be done after a suffi cient time to rest, in order to avoid increased variance in the stereometric ONH parameters.

60

6.6 Comparison of methods for detecting glaucomatous progression using the HRT

Studies evaluating the detection of glaucomatous progression with the HRT use different analysis methods and different cut-off values to defi ne progression (Table 2). One of the methods used is the regression of certain stereometric parameters such as rim area. This is a trend analysis, which requires more HRT examinations than an event analysis, but is less affected by variation. The detection rates for progression using trend analysis and event analysis are different (Fayers et al. 2007). A trend analysis is the method of choice for detecting progression when the examinations are performed at fairly short intervals. However, short intervals may not be possible with limited resources and a high number of glaucoma patients. In the Finnish evidence-based guideline for open-angle glaucoma, a good level of follow-up included ONH or RNFL imaging every two years (Tuulonen et al. 2003). With such intervals, an event analysis for determining progression would be preferable.

The topographic change analysis (TCA) is regarded as an event analysis, but it requires three or preferably four examinations with the HRT. The method identifi es the location and the direction, but not the statistical signifi cance of change. Several different defi nitions of progression using the TCA have been reported (Table 2). A possible method to diminish the effect of variability on results is repetitive imaging on each visit (Weinreb et al. 1993). This would also allow the use of TCA for detecting change between each visit.

The accuracy for detecting progression in the present study (I, II) is similar to other studies (Figure 11, Figure 12). As expected, the correlation between methods appears to be higher in studies using ONH photography as a reference standard compared to visual fi elds (Figure 5). The mean follow-up time of the present study was two years and one month (I) and two years and four months (II). Two years is the recommended time between imaging in the Finnish evidence-based guideline (Tuulonen et al. 2003). However, the relatively short follow-up time is a limitation of the present study compared to other published studies. (Table 2).

61

Figure 11. A summary receiver operating characteristics plot for the detection of glauco-

matous progression using HRT, with the sensitivity and specifi city values of single stere-

ometric ONH parameters for progression. The results of the present study are displayed

with open symbols and the results of other studies (Table 2) with black symbols.

Progression in RNFL photographs (I) is the reference standard for the change in cup

shape measure (open diamond). Progression in the stereoscopic ONH photographs (II)

is the reference standard for the change in cup:disc area ratio (open square) and verti-

cal cup:disc ratio (open triangle). In addition to sensitivity and specifi city at the opti-

mised cut-off, sensitivities at 90% specifi city are also plotted.

The sensitivity and specifi city values of studies using the regression of single stere-

ometric parameters to detect progression are displayed for comparison. The reference

standards include progression in the visual fi elds (black diamonds) or the stereoscopic

ONH photographs (black circle).

62

In the present study progression was analysed by calculating change in the stereometric ONH parameters between two points in time (I, II). It is an event analysis and relates to the clinician’s question ‘has there been progression in the ONH since the last visit’. As an event analysis it is very sensitive to variation of the stereometric parameters and to factors inducing variability such as decrease in image quality (Figure 10). Regardless of the reference standard and the method used for analysing progression with the HRT, high concordance between methods to detect progression has not been reported (Table 2).

Figure 12. A summary receiver operating characteristics plot for the detection of glau-

comatous progression with HRT using stereoscopic ONH photographs as a reference.

The results of the present study are displayed with open symbols and the results of

other studies (Table 2) with black symbols.

The optimised sensitivity and specifi city values of the cup:disc area ratio (open square)

and vertical cup:disc ratio (open triangle) for progression are given. Sensitivities at 90%

specifi city are also displayed (II). The studies detecting ONH progression with the HRT,

using stereoscopic ONH photographs, with (black triangle) and without visual fi elds

(black circles), as a reference standard are plotted for comparison.

63

6.7 Signifi cance of change

Even if the change detected with an imaging method reaches statistical signifi cance, the clinical signifi cance of the change remains unknown. There are no studies demonstrating the effi cacy of a treatment strategy based on imaging alone. Whether the change is clinically signifi cant and warrants change in treatment is still decided by the clinician treating glaucoma. The subsequent treatment may depend on the severity and prior course of the disease, IOP, age and life expectancy of the patient and the safety and effi cacy of the remaining treatment options (Tuulonen et al. 2003, European Glaucoma Society 2008).

Even when using the best LDFs in the present study, with sensitivity and specifi city values close to 80%, more than one fi fth of the patients showing no progression in ONH photographs would be classifi ed as progressors and more than one fi fth of the patients with progression in ONH photographs would be classifi ed as stable. When detecting progression the specifi city of the diagnostic test should be high in order to avoid overtreatment and subjecting stable patients to the inconvenience and risks related to additional medication or surgery. When screening for glaucoma, the specifi city of the diagnostic test should be more than 96% in order for the screening to be cost-effective (Vaahtoranta-Lehtonen et al. 2007). Also during follow-up, specifi cities below 90% are of limited value. In the present study, sensitivities at 90% specifi city were below 50%, thus identifying less than half of the real progressors. Combining the results of the ONH, RNFL and visual fi elds improves the accuracy of detecting glaucomatous damage (Tuulonen et al. 2003, Badala et al. 2007, Strouthidis & Garway-Heath 2008, Zhu et al. 2009). Even if the clinical decisions are based on several methods for detecting progression (Figure 13), the results are still subjectively interpreted by the clinician treating glaucoma.

6.8 Detection of glaucomatous progression

There is no golden standard for detecting glaucomatous progression. Progression may be identifi ed at different times during follow-up in the ONH, RNFL and visual fi elds. There is both a temporal dissociation and a difference in detection rates using different methods of analysis (Vesti 2003, Fayers et al. 2007). Figure 13 summarises some of the methods used for detecting glaucomatous progression.

64

6.9 Practical implications

Reducing the variability of the stereometric parameters improves the accuracy of detecting progression. Repetitive imaging and allowing suffi cient time to rest before the HRT examination diminishes variability. Image quality should be the best possible, and mydriatics and artifi cial tears should be used if necessary. When images of different quality were compared, the correlation of the change in the stereometric parameters with progression in stereoscopic ONH photographs was considerately improved with images of excellent quality. With excellent image quality, the inter-method agreement for cup volume was at the same level as the inter-observer agreement between the evaluators of the stereoscopic ONH

Figure 13. Detection of glaucomatous progression. Comparisons performed in the

present study are indicated by wide double-headed arrows. Detailed descriptions of

the methods may be found in the references: 1. RNFL photography (Airaksinen et al.

1984), 2. GDx progression methods (Alencar et al. 2010), 3. Stereometric parameters of

the HRT (Burk et al. 1990), 4. TCA (Chauhan et al. 2000), 5. TCA parameters (Bowd et al.

2009), 6. GPS (Swindale et al. 2000, Strouthidis et al. 2010) 7. GPA (Heijl et al. 2003), 8.

GCP (Heijl et al. 1991), 9. AGIS (AGIS Investigators 1994), 10. CIGTS (Musch et al. 1999),

11. Pointwise linear regression (Wild et al. 1997).

65

photographs. This would challenge the view that evaluation of ONH using the HRT and the stereoscopic ONH photographs assesses different aspects of ONH change (O’Leary et al. 2008, Vizzeri et al. 2009b). However, images of excellent quality cannot be obtained from all patients. In the retrospective follow-up study (II) only 8% had mean TSD values below 11 μm, indicating excellent quality.

The detection of glaucomatous damage or progression should not rely on any single method (Tuulonen et al. 2003, Badala et al. 2007, Strouthidis & Garway-Heath 2008). There may be a considerable temporal dissociation in progression detected in the ONH, RNFL and the visual fi elds. Confi rming progression may require repeated examinations (Tuulonen et al. 2003, Owen et al. 2006). Examining the RNFL, ONH and the visual fi elds on each follow-up visit improves diagnostic precision and allows implementing the 2-out-of-3 rule for progression (Tuulonen et al. 2003). The change in the stereometric parameters provides useful information on ONH topography, especially with excellent image quality. Based on the results of the current study the stereometric ONH parameters that should be used in follow-up include cup:disc area ratio, vertical cup:disc ratio, rim area, cup volume and cup shape measure. Especially when treating patients with advanced glaucomatous damage or large ONHs, change in the vertical cup:disc ratio should be monitored. With small ONHs and early stages of glaucomatous damage, change in the rim area should receive special attention. However, the evaluation of glaucomatous progression in the ONH should not rely solely on the stereometric parameters of the HRT.

67

7 Conclusions

1. The change in only one out of 23 stereometric ONH parameters, the cup shape measure, showed a statistically signifi cant correlation with the progression of the RNFL defect. An optimised change in the best three-parameter combination had 77% sensitivity and 79% specifi city for progression.

2. The parameter with the best correlation with progression in stereoscopic ONH photography was the cup:disc area ratio. The sensitivity of the cup:disc area ratio was 28% at 90% specifi city. The parameter with the largest area under the ROC curve (0.726) was the vertical cup:disc ratio with a sensitivity of 22% at a specifi city of 90%. The sensitivity of a linear discriminant function [(12.241 × cup:disc area ratio) + (3.540 × mean cup depth) - (2.146 × horizontal cup:disc ratio) + (27.486 × average variability)] was 31% at a specifi city of 90%. With the cut-off at 90% specifi city, the kappa value for inter-method agreement was 0.211 with stereoscopic ONH photography, indicating only fair inter-method agreement.

3. The factors having the most signifi cant effect on the sensitivity and specifi city of the stereometric ONH parameters were the reference height difference and the mean topography standard deviation (TSD) indicating image quality. The change in the TSD and age also showed a consistent, but variably signifi cant infl uence on all parameters tested.

4. Exercise was associated with an increase in variance in 17 of the 18 stereometric ONH parameters. The increase in variance was statistically signifi cant in eight parameters including rim area, cup:disc area ratio and cup shape measure.

69

References

AGIS Investigators (1994) Advanced Glaucoma Intervention Study. 2. Visual fi eld test scoring and reliability. Ophthalmology 101: 1445–1455.

AGIS Investigators (1998) The Advanced Glaucoma Intervention Study (AGIS): 4. Comparison of treatment outcomes within race: seven-year results. Ophthalmology 105: 1146–1164.

AGIS Investigators (2000) The Advanced Glaucoma Intervention Study (AGIS): 7. The relationship between control of intraocular pressure and visual fi eld deterioration. Am J Ophthalmol 130: 429–440.

Airaksinen PJ (1994) Quantitation of optic disc and retinal nerve fi ber layer examination for the diagnosis and follow-up of glaucoma. Unpublished data, presented at The Gullstrand Meeting on Glaucoma. Uppsala. May 26.–27.

Airaksinen PJ & Alanko HI (1983) Effect of retinal nerve fi bre loss on the optic nerve head confi guration in early glaucoma. Graefes Arch Clin Exp Ophthalmol 220: 193–196.

Airaksinen PJ, Drance SM, Douglas GR, Mawson DK & Nieminen H (1984) Diffuse and localized nerve fi ber loss in glaucoma. Am J Ophthalmol 98: 566–71.

Airaksinen PJ, Drance SM, Douglas GR, Schulzer M & Wijsman K (1985) Visual fi eld and retinal nerve fi bre layer comparisons in glaucoma. Arch Ophthalmol 103: 205–207.

Airaksinen PJ, Mustonen E & Alanko HI (1981) Optic disc haemorrhages precede retinal nerve fi bre layer defects in ocular hypertension. Acta Ophthalmol Scand 59: 627–641.

Airaksinen PJ & Nieminen H (1985) Retinal nerve fi bre layer photography in glaucoma. Ophthalmology 92: 877–879.

Airaksinen PJ, Tuulonen A & Alanko HI (1992) Rate and pattern of neuroretinal rim area decrease in ocular hypertension and glaucoma. Arch Ophthalmol 110: 206–210.

Alencar LM, Bowd C, Weinreb RN, Zangwill LM, Sample PA & Medeiros FA (2008) Comparison of HRT-3 glaucoma probability score and subjective stereophotograph assessment for prediction of progression in glaucoma. Invest Ophthalmol Vis Sci 49: 1898–906.

Alencar LM, Zangwill LM, Weinreb RN, Bowd C, Vizzeri G, Sample PA, Susanna R Jr & Medeiros FA (2010) Agreement for Detecting Glaucoma Progression with the GDx Guided Progression Analysis, Automated Perimetry, and Optic Disc Photography. Ophthalmology 117: 462–70.

Altangerel U, Bayer A, Henderer JD, Katz LJ, Steinmann WC & Spaeth GL (2005) Knowledge of chronology of optic disc stereophotographs infl uences the determination of glaucomatous change. Ophthalmology 112: 40–3.

Anderson DR, Chauhan B, Johnson C, Katz J, Patella VM & Drance SM (2000) Criteria for progression of glaucoma in clinical management and in outcome studies. Am J Ophthalmol 130: 827-9.

Armaly MF (1969) The optic cup in the normal eye. I. Cup width, depth, vessel displacement, ocular tension and outfl ow facility. Am J Ophthalmol 68: 401–7.

Artal P, Guirao A, Berrio E & Williams DR (2001) Compensation of corneal aberrations by the internal optics in the human eye. J Vis 1: 1–8.

70

Artes PH & Chauhan BC (2005) Longitudinal changes in the visual fi eld and optic disc in glaucoma. Prog Retin Eye Res 24: 333–54.

Asaoka R, Strouthidis NG, Kappou V, Gardiner SK & Garway-Heath DF (2009) HRT-3 Moorfi elds reference plane: effect on rim area repeatability and identifi cation of progression. Br J Ophthalmol 93: 1510–3.

Asrani S, Zeimer R, Wilensky J, Gieser D, Vitale S & Lindenmuth K (2000) Large diurnal fl uctuations in intraocular pressure are an independent risk factor in patients with glaucoma. J Glaucoma 9: 134–142.

Azuara-Blanco A, Harris A, Cantor LB, Abreu MM & Weinland M (1998) Effects of short term increase of intraocular pressure on optic disc cupping. Br J Ophthalmol 82: 880–3.

Azuara-Blanco A, Katz LJ, Spaeth GL, Vernon SA, Spencer F & Lanzl IM (2003) Clinical agreement among glaucoma experts in the detection of glaucomatous changes of the optic disk using simultaneous stereoscopic photographs. Am J Ophthalmol 136: 949–50.

Badalà F, Nouri-Mahdavi K, Raoof DA, Leeprechanon N, Law SK & Caprioli J (2007) Optic disk and nerve fi ber layer imaging to detect glaucoma. Am J Ophthalmol 144: 724–32.

Bagga H, Greenfi eld DS & Feuer WJ (2005) Quantitative assessment of atypical birefringence images using scanning laser polarimetry with variable corneal compensation. Am J Ophthalmol 139: 437–46.

Balasubramanian M, Bowd C, Weinreb RN, Vizzeri G, Alencar LM, Sample PA, O’Leary N & Zangwill LM (2010) Clinical evaluation of the proper orthogonal decomposition framework for detecting glaucomatous changes in human subjects. Invest Ophthalmol Vis Sci 51: 264–71.

Balasubramanian M, Zabic S, Bowd C, Thompson HW, Wolenski P, Iyengar SS, Karki BB & Zangwill LM (2009) A framework for detecting glaucomatous progression in the optic nerve head of an eye using proper orthogonal decomposition. IEEE Trans Inf Technol Biomed 13: 781–93.

Bathija R, Zangwill L, Berry CC, Sample PA & Weinreb RN (1998) Detection of early glaucomatous structural damage with confocal scanning laser tomography. J Glaucoma 7: 121–7.

Beck RW, Bergstrom TJ & Lichter PR (1985) A clinical comparison of visual fi eld testing with a new automated perimeter, the Humphrey Field Analyzer, and the Goldmann perimeter. Ophthalmology 92: 77–82.

Behrendt T & Wilson LA (1965) Spectral Refl ectance Photography of the Retina. Am. J. Ophthalmol 59: 1079–1088.

Bengtsson B & Heijl A (2005) Diurnal IOP fl uctuation: not an independent risk factor for glaucomatous visual fi eld loss in high-risk ocular hypertension. Graefes Arch Clin Exp Ophthalmol 243: 513–8.

Bergin C, Garway-Heath DF & Crabb D (2008) Evaluating the effect of the new alignment algorithm for longitudinal series of Heidelberg retina tomography images Acta Ophthalmol 86: 207–214.

71

Blumenthal EZ, Sample PA, Berry CC, Lee AC, Girkin CA, Zangwill L, Caprioli J & Weinreb RN (2003) Evaluating several sources of variability for standard and SWAP visual fi elds in glaucoma patients, suspects, and normals. Ophthalmology 110: 1895–902.

Bowd C, Balasubramanian M, Weinreb RN, Vizzeri G, Alencar LM, O’Leary N, Sample PA & Zangwill LM (2009) Performance of confocal scanning laser tomograph Topographic Change Analysis (TCA) for assessing glaucomatous progression. Invest Ophthalmol Vis Sci 50: 691–701.

Bowd C, Weinreb RN, Lee B, Emdadi A & Zangwill LM (2000) Optic disk topography after medical treatment to reduce intraocular pressure. Am J Ophthalmol 130: 280–6.

Bozkurt B, Irkec M & Arslan U (2010) Diagnostic accuracy of Heidelberg Retina Tomograph III classifi cations in a Turkish primary open-angle glaucoma population. Acta Ophthalmol 88: 125–130.

Breusegem C, Fieuws S, Stalmans I & Zeyen T (2008) Variability of the standard reference height and its infl uence on the stereometric parameters of the Heidelberg retina tomograph 3. Invest Ophthalmol Vis Sci 49: 4881–5.

Brusini P, Salvetat ML, Parisi L, Zeppieri M & Tosoni C (2005) Discrimination between normal and early glaucomatous eyes with scanning laser polarimeter with fi xed and variable corneal compensator settings. Eur J Ophthalmol 15: 468–76.

Burgansky-Eliash Z, Wollstein G, Bilonick RA, Ishikawa H, Kagemann L & Schuman JS (2007) Glaucoma detection with the Heidelberg retina tomograph 3. Ophthalmology 114: 466–71.

Burk RO, Noack H, Rohrschneider K & Volcker HE (1999) Prediction of glaucomatous visual fi eld defects by reference plane independent three-dimensional optic nerve head parameters. In: Wall M, Wild JM, eds. Perimetry Update 1998/1999: Proceedings of the XIIIth International Perimetric Society Meeting, Gardone Riviera (BS), Italy, September 6–9, 1998. The Hague, Kugler: 463–74.

Burk RO & Rendon R (2001) Clinical detection of optic nerve damage: measuring changes in cup steepness with use of a new image alignment algorithm. Surv Ophthalmol 45: S297–S303.

Burk ROW, Rohrschneider K, Völcker HE & Zinser G (1990) Analysis of three-dimensional optic disc topography by laser scanning tomography. In: Nasemann, JE & Burk, ROW, eds. Laser Scanning Ophthalmoscopy and Tomography. München, Quintessenz: 161–176.

Burk R, Vihanninjoki K, Bartke T, Tuulonen A, Airaksinen PJ, Völcker HE & König JM (2000) Development of the standard reference plane for the Heidelberg Retina Tomograph. Graefes Arch Clin Exp Ophthalmol 238: 375–384.

Burr J, Azuara-Blanco A & Avenell A (2004) Medical versus surgical interventions for open angle glaucoma. Cochrane Database Syst Rev. Article No.: CD004399. DOI: 10.1002/14651858.CD004399.pub2.

Burr JM, Mowatt G, Hernández R, Siddiqui MA, Cook J, Lourenco T, Ramsay C, Vale L, Fraser C, Azuara-Blanco A, Deeks J, Cairns J, Wormald R, McPherson S, Rabindranath K & Grant A (2007) The clinical effectiveness and cost-effectiveness of screening for open angle glaucoma: a systematic review and economic evaluation. Health Technol Assess 11: 1–190.

72

Capel D (2004) Image mosaicing and superresolution,1st edn. London, Springer-Verlag.Caprioli J (1989) Correlation of visual function with optic nerve and nerve fi ber layer

structure in glaucoma. Surv Ophthalmol 33 (Suppl): 319–30.Caprioli J & Coleman AL (2008) Intraocular pressure fl uctuation: A risk factor for visual

fi eld progression at low intraocular pressures in the Advanced Glaucoma Intervention Study. Ophthalmology 115: 1123–9.

Cello KE, Nelson-Quigg JM & Johnson CA (2000) Frequency doubling technology perimetry for detection of glaucomatous visual fi eld loss. Am J Ophthalmol 129: 314–322.

Chauhan BC, Blanchard JW, Hamilton DC & LeBlanc RP (2000) Technique for detecting serial topographic changes in the optic disc and peripapillary retina using scanning laser tomography. Invest Ophthalmol Vis Sci 41: 775–782.

Chauhan BC, House PH, McCormick TA & LeBlanc RP (1999) Comparison of conventional and high-pass resolution perimetry in a prospective study of patients with glaucoma and healthy controls. Arch Ophthalmol 117: 24–33.

Chauhan BC, Hutchison DM, Artes PH, Caprioli J, Jonas JB, LeBlanc RP & Nicolela MT (2009a) Optic disc progression in glaucoma: comparison of confocal scanning laser tomography to optic disc photographs in a prospective study. Invest Ophthalmol Vis Sci 50: 1682–91.

Chauhan BC, LeBlanc RP, McCormick TA & Rogers JB (1994) Test-retest variability of topographic measurements with confocal scanning laser tomography in patients with glaucoma and control subjects. Am J Ophthalmol 118: 9–15.

Chauhan BC & McCormick TA (1995) Effect of the cardiac cycle on topographic measurements using scanning laser tomography. Graefe’s Arch Clin Exp Ophthalmol 233: 568–572.

Chauhan BC, McCormick TA, Nicolela MT & LeBlanc RP (2001) Optic disc and visual fi eld changes in a prospective longitudinal study of patients with glaucoma: comparison of scanning laser tomography with conventional perimetry and optic disc photography. Arch Ophthalmol 119: 1492–1499.

Chauhan BC & MacDonald CA (1995) Infl uence of time separation on variability estimates of topographic measurements with confocal scanning laser tomography. J Glaucoma 4: 189–93.

Chauhan BC, Mikelberg FS, Balaszi AG, LeBlanc RP, Lesk MR & Trope GE; Canadian Glaucoma Study Group (2008) Canadian Glaucoma Study: 2. risk factors for the progression of open-angle glaucoma. Arch Ophthalmol 126: 1030–6.

Chauhan BC, Nicolela MT & Artes PH (2009b) Incidence and rates of visual fi eld progression after longitudinally measured optic disc change in glaucoma. Ophthalmology 116: 2110–8.

Chen E, Gedda U & Landau I (2001) Thinning of the papillomacular bundle in the glaucomatous eye and its infl uence on the reference plane of the Heidelberg retinal tomograph. J Glaucoma 10: 386–389.

Collaborative Normal Tension Glaucoma Study Group (1998a) Comparison of glaucomatous progression between untreated patients with normal tension glaucoma and patients with therapeutically reduced intraocular pressure. Am J Ophthalmol 126: 487–497.

73

Collaborative Normal Tension Glaucoma Study Group (1998b) The effectiveness of intraocular pressure reduction in the treatment of normal tension glaucoma. Am J Ophthalmol 126: 498–505.

Coops A, Henson DB, Kwartz AJ & Artes PH (2006) Automated analysis of heidelberg retina tomograph optic disc images by glaucoma probability score. Invest Ophthalmol Vis Sci 47: 5348–55.

DeLeon-Ortega JE, Arthur SN, McGwin G Jr, Xie A, Monheit BE & Girkin CA (2006) Discrimination between glaucomatous and nonglaucomatous eyes using quantitative imaging devices and subjective optic nerve head assessment. Invest Ophthalmol Vis Sci 47: 3374–80.

DeLeon-Ortega JE, Sakata LM, Kakati B, McGwin G Jr, Monheit BE, Arthur SN & Girkin CA (2007) Effect of glaucomatous damage on repeatability of confocal scanning laser ophthalmoscope, scanning laser polarimetry, and optical coherence tomography. Invest Ophthalmol Vis Sci 48: 1156–63.

Dielemans I, Vingerling JR, Wolfs RC, Hofman A, Grobbee DE & de Jong PT (1994) The prevalence of primary open-angle glaucoma in a populationbased study in The Netherlands. The Rotterdam Study. Ophthalmology 101: 1851–5.

Donaldson D, Prescott R & Kennedy S (1980) Simultaneous stereoscopic fundus camera utilizing a single optical axis. Invest Ophthalmol Vis Sci 19: 289–97.

Doughty MJ & Zaman ML (2000) Human corneal thickness and its impact on intraocular pressure measures: a review and meta-analysis approach. Surv Ophthalmol 44: 367–408.

Ekström C (1996) Prevalence of open-angle glaucoma in central Sweden. The Tierp glaucoma survey. Acta Ophthalmol Scand 74: 107–112.

Erä P, Pärssinen O, Kallinen M & Suominen H (1993) Effect of bicycle ergometer test on intraocular pressure in elderly athletes and controls. Acta Ophthalmol Scand 71: 301–7.

Essock EA, Zheng Y & Gunvant P (2005) Analysis of GDx-VCC polarimetry data by Wavelet-Fourier analysis across glaucoma stages. Invest Ophthalmol Vis Sci 46: 2838–47.

European Glaucoma Society (2008) Terminology and guidelines for glaucoma. 3rd edition. Savona, Italy, Dogma.

Falck A, Hautala N, Turunen N & Airaksinen PJ (2009) A four-year prospective study on intraocular pressure in relation to phacoemulsifi cation cataract surgery. Acta Ophthalmol (Epub ahead of print) doi: 10.1111/j.1755-3768.2009.01790.x

Fayers T, Strouthidis NG & Garway-Heath DF (2007) Monitoring glaucomatous progression using a novel Heidelberg Retina Tomograph event analysis. Ophthalmology 114: 1973–80.

Ferreras A, Pablo LE, Larrosa JM, Polo V, Pajarín AB & Honrubia FM (2008) Discriminating between normal and glaucoma-damaged eyes with the Heidelberg Retina Tomograph 3. Ophthalmology 115: 775–781.

Ferreras A, Pajarin AB, Polo V, Larrosa JM, Pablo LE & Honrubia FM (2007) Diagnostic ability of Heidelberg Retina Tomograph 3 classifi cations: glaucoma probability score versus Moorfi elds regression analysis. Ophthalmology 114: 1481–7.

74

Fingeret M (2005) Using the Heidelberg Retina Tomograph II (HRT II): Image acquisition and accessing the data. In: Fingeret M, Flanagan JG, Liebmann JM (eds) The Essential HRT Primer. San Ramon, California: Jocoto advertising Inc: 11–30.

Finnish Register of Visual Impairment (2008) Annual Statistics 2008.Fiscella RG, Lee J, Davis EJ & Walt J (2009) Cost of illness of glaucoma: a critical and

systematic review. Pharmacoeconomics 27: 189–98.Ford BA, Artes PH, McCormick TA, Nicolela MT, LeBlanc RP & Chauhan BC (2003)

Comparison of data analysis tools for detection of glaucoma with the Heidelberg Retina Tomograph. Ophthalmology 110: 1145–50.

Foster PJ, Buhrmann R, Quigley HA & Johnson GJ (2002) The defi nition and classifi cation of glaucoma in prevalence surveys. Br J Ophthalmol 86: 238–42.

Funk J & Mueller H (2003) Comparison of long-term fl uctuations: laser scanning tomography versus automated perimetry. Graefes Arch Clin Exp Ophthalmol 241: 721–4.

Gabriele ML, Wollstein G, Bilonick RA, Burgansky-Eliash Z, Ishikawa H, Kagemann LE & Schuman JS (2008) Comparison of parameters from Heidelberg Retina Tomographs 2 and 3. Ophthalmology 115: 673–7.

Gordon MO, Beiser JA, Brandt JD, Heuer DK, Higginbotham EJ, Johnson CA, Keltner JL, Miller JP, Parrish RK 2nd, Wilson MR & Kass MA (2002) The ocular hypertension treatment study. Baseline factors that predict the onset of primary open-angle glaucoma. Arch Ophthalmol 120: 714–720.

Gordon MO & Kass MA (1999) The Ocular Hypertension Treatment Study: design and baseline description of the participants. Arch Ophthalmol 117: 573–83.

Greaney MJ, Hoffman DC, Garway-Heath DF, Nakla M, Coleman AL & Caprioli J (2002) Comparison of optic nerve imaging methods to distinguish normal eyes from those with glaucoma. Invest Ophthalmol Vis Sci 43: 140–5.

Harasymowycz P, Papamatheakis D, Fansi AK, Gresset J & Lesk MR (2005) Validity of screening for glaucomatous optic nerve damage using confocal scanning laser ophthalmoscopy (Heidelberg Retina Tomograph II) in high-risk populations: a pilot study. Ophthalmology 112: 2164-71.

Harizman N, Zelefsky JR, Ilitchev E, Tello C, Ritch R & Liebmann JM (2006) Detection of glaucoma using operator-dependent versus operator-independent classifi cation in the Heidelberg retinal tomograph- III. Br J Ophthalmol 90: 1390–2.

Harju M, Saari J, Kurvinen L & Vesti E (2008) Reversal of optic disc cupping in glaucoma. Br J Ophthalmol 92: 901–5.

Harju M & Vesti E (2001) Scanning laser ophthalmoscopy of the optic nerve head in exfoliation glaucoma and ocular hypertension with exfoliation syndrome. Br J Ophthalmol 85: 297–303.

Hatch WV, Flanagan JG, Williams-Lyn DE, Buys YM, Farra T & Trope GE (1999) Interobserver agreement of Heidelberg retina tomograph parameters. J Glaucoma 8: 232–7.

Hawker MJ, Vernon SA & Ainsworth G (2006) Specifi city of the Heidelberg Retina Tomograph’s diagnostic algorithms in a normal elderly population: the Bridlington Eye Assessment Project. Ophthalmology 113: 778–85.

75

Heijl A, Drance SM & Douglas R (1980) Automatic perimetry (COMPETER). Ability to detect early glaucomatous fi eld defects. Arch Ophthalmol 98: 1560–1563.

Heijl A, Leske MC, Bengtsson B, Bengtsson B, Hussein M & the EMGT Group (2003) Measuring visual fi eld progression in the Early Manifest Glaucoma Trial. Acta Ophthalmol Scand 81: 286–293.

Heijl A, Leske MC, Bengtsson B, Hyman L, Bengtsson B, Hussein M & the EMGT Group (2002) Reduction of intraocular pressure and glaucoma progression. Results from the Early Manifest glaucoma Trial. Arch Ophthalmol 120: 1268–1279.

Heijl A, Lindgren G, Lindgren A, Olsson J, Asman P, Myers S & Patella M (1991) Extended empirical statistical package for evaluation of single and multiple fi elds: Statpac 2. In: Mills RP, Heijl A, eds. Perimetry Update 1990/1. New York: Kugler and Ghedini: 303–315.

Henderer JD (2006) Disc damage likelihood scale. Br J Ophthalmol 90: 395–6.Henderer JD, Liu C, Kesen M, Altangerel U, Bayer A, Steinmann WC & Spaeth GL (2003)

Reliability of the disk damage likelihood scale. Am J Ophthalmol 135: 44–8.Hirvelä H, Tuulonen A & Laatikainen L (1995) Intraocular pressure and prevalence of glaucoma

in elderly people in Finland: a population-based study. Int Ophthalmol 18: 299–307.Hoesl LM, Mardin CY, Horn FK, Juenemann AG & Laemmer R (2009) Infl uence of

glaucomatous damage and optic disc size on glaucoma detection by scanning laser tomography. J Glaucoma 18: 385–9.

Hoh ST, Greenfi eld DS, Liebmann JM, Maw R, Ishikawa H, Chew SJ & Ritch R (1998) Factors affecting image acquisition during scanning laser polarimetry. Ophthalmic Surg Lasers 29: 545–51.

Hong S, Ahn H, Ha SJ, Yeom HY, Seong GJ & Hong YJ (2007) Early glaucoma detection using the Humphrey Matrix Perimeter, GDx VCC, Stratus OCT, and retinal nerve fi ber layer photography. Ophthalmology 114: 210–215.

Hoyt WF (1976) Ophthalmoscopy of the retinal nerve fi bre layer in neuro-ophthalmologic diagnosis. Aust J Ophthalmol 4: 14–34.

Huang D, Swanson EA, Lin CP, Schuman JS, Stinson WG, Chang W, Hee MR, Flotte T, Gregory K, Puliafi to CA & Fujimoto JG (1991) Optical coherence tomography. Science 254: 1178–1181

Hudson CJ, Kim LS, Hancock SA, Cunliffe IA & Wild JM (2007) Some dissociating factors in the analysis of structural and functional progressive damage in open-angle glaucoma. Br J Ophthalmol 91: 624–8.

Hwang JM, Kim TW, Park KH, Kim DM & Kim H (2006) Correlation between topographic profi les of localized retinal nerve fi ber layer defects as determined by optical coherence tomography and red-free fundus photography. J Glaucoma 15: 223–8.

Iacono P, Da Pozzo S, Fuser M, Marchesan R & Ravalico G (2006) Intersession reproducibility of retinal nerve fi ber layer thickness measurements by GDx-VCC in healthy and glaucomatous eyes. Ophthalmologica 220: 266–71.

Iester M, Mardin CY, Budde WM, Jünemann AG, Hayler JK & Jonas JB (2002) Discriminant analysis formulas of optic nerve head parameters measured by confocal scanning laser tomography. J Glaucoma 11: 97–104.

76

Iester M, Mikelberg FS, Courtright P & Drance SM (1997a) Correlation between the visual fi eld indices and Heidelberg Retina Tomography parameters. J Glaucoma 6: 78–82.

Iester M, Mikelberg FS & Drance SM (1997b) The effect of optic disc size on diagnostic precision with the Heidelberg retina tomograph. Ophthalmology 104: 545–8.

Iester M, Parfi tt CM, Swindale NV & Mikelberg FS (1997c) Sector-based analysis of Heidelberg Retina Tomograph (HRT) parameters in normal and glaucomatous eyes [Abstract]. Invest Ophthalmol Vis Sci 38(Suppl): S835.

Irak I, Zangwill L, Garden V, Shakiba S & Weinreb RN (1996) Change in optic disk topography after trabeculectomy. Am J Ophthalmol 122: 690–5.

Jampel HD, Friedman D, Quigley H, Vitale S, Miller R, Knezevich F & Ding Y (2009) Agreement among glaucoma specialists in assessing progressive disc changes from photographs in open-angle glaucoma patients. Am J Ophthalmol 147: 39–44.

Jampel HD, Vitale S, Ding Y, Quigley H, Friedman D, Congdon N & Zeimer R (2006) Test-retest variability in structural and functional parameters of glaucoma damage in the glaucoma imaging longitudinal study. J Glaucoma 15: 152–7.

Johnson CA, Adams AJ, Casson EJ & Brandt JD (1993) Blue-on-yellow perimetry can predict the development of glaucomatous visual fi eld loss. Arch Ophthalmol 111: 645–50.

Jonas JB, Budde WM & Panda-Jonas S (1999) Ophthalmoscopic evaluation of the optic nerve head. Surv Ophthalmol 43: 293–320.

Jonas JB, Gusek GC & Naumann GO (1988) Optic disc morphometry in chronic primary open-angle glaucoma. II. Correlation of the intrapapillary morphometric data to visual fi eld indices. Graefes Arch Clin Exp Ophthalmol 226: 531–8.

Kamal DS, Viswanathan AC, Garway-Heath DF, Hitchings RA, Poinoosawmy D & Bunce C (1999) Detection of optic disc change with the Heidelberg retina tomograph before confi rmed visual fi eld change in ocular hypertensives converting to early glaucoma. Br J Ophthalmol 83: 290–4.

Kass MA, Heuer DK, Higginbotham EJ, Johnson CA, Keltner JL, Miller JP, Parrish RK, Wilson MR & Gordon MO (2002) The Ocular Hypertension Treatment Study. A randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma. Arch Ophthalmol 120: 710–713.

Katz J, Gilbert D, Quigley HA & Sommer A (1997) Estimating progression of visual fi eld loss in glaucoma. Ophthalmology 104: 1017–1025.

Keltner JL, Johnson CA, Levine RA, Fan J, Cello KE, Kass MA & Gordon MO (2005) Normal visual fi eld test results following glaucomatous visual fi eld end points in the Ocular Hypertension Treatment Study. Arch Ophthalmol 123: 1201–6.

Kiriyama N, Ando A, Fukui C, Nambu H, Nishikawa M, Terauchi H, Kuwahara A & Matsumura M (2003) A comparison of optic disc topographic parameters in patients with primary open angle glaucoma, normal tension glaucoma, and ocular hypertension.Graefes Arch Clin Exp Ophthalmol 241: 541–5.

Klein BEK, Klein R, Sponsel WE, Franke T, Cantor LB, Martone J & Menage MJ (1992) Prevalence of glaucoma. The Beaver Dam Eye Study. Ophthalmology 99: 1499–1504.

77

Kourkoutas D, Buys YM, Flanagan JG, Hatch WV, Balian C & Trope GE (2007) Comparison of glaucoma progression evaluated with Heidelberg retina tomograph II versus optic nerve head stereophotographs. Can J Ophthalmol 42: 82–88.

Lan YW, Henson DB & Kwartz AJ (2003) The correlation between optic nerve head topographic measurements, peripapillary nerve fi bre layer thickness, and visual fi eld indices in glaucoma. Br J Ophthalmol 87: 1135–41.

Lesk MR, Spaeth GL, Azuara-Blanco A, Araujo SV, Katz LJ, Terebuh AK, Wilson RP, Moster MR & Schmidt CM (1999) Reversal of optic disc cupping after glaucoma surgery analyzed with a scanning laser tomograph. Ophthalmology 106: 1013–8.

Leske MC, Heijl A, Hussein M, Bengtsson B, Hyman L, Komaroff E & EMGT Group (2003) Factors for glaucoma progression and the effect of treatment: the early manifest glaucoma trial. Arch Ophthalmol 121: 48–56.

Leske MC, Heijl A, Hyman L & Bengtsson B (1999) Early Manifest Glaucoma Trial: design and baseline data. Ophthalmology 106: 2144–2153.

Lichter PR (1976) Variability of expert observers in evaluating the optic disc. Trans Am Ophthalmol Soc 74: 532–72.

Lichter PR, Musch DC, Gillespie BW, Guire KE, Janz NK, Wren PA, Mills RP & the CIGTS Study Group (2001): Interim clinical outcomes in the Collaborative Initial Glaucoma Treatment Study comparing initial treatment randomized to medications or surgery. Ophthalmology 108: 1943–1953.

Lin D, Leung CK, Weinreb RN, Cheung CY, Li H & Lam DS (2009) Longitudinal evaluation of optic disc measurement variability with optical coherence tomography and confocal scanning laser ophthalmoscopy. J Glaucoma 18: 101–6.

Lin SC, Singh K, Jampel HD, Hodapp EA, Smith SD, Francis BA, Dueker DK, Fechtner RD, Samples JS, Schuman JS & Minckler DS; American Academy of Ophthalmology; Ophthalmic Technology Assessment Committee Glaucoma Panel (2007) Optic nerve head and retinal nerve fi ber layer analysis: a report by the American Academy of Ophthalmology. Ophthalmology 114:1937–49.

Mai TA, Reus NJ & Lemij HG (2007) Diagnostic accuracy of scanning laser polarimetry with enhanced versus variable corneal compensation. Ophthalmology 114: 1988–93.

Maier PC, Funk J, Schwarzer G, Antes G & Falck-Ytter YT (2005) Treatment of ocular hypertension and open angle glaucoma: meta-analysis of randomised controlled trials. BMJ 331: 134–6.

Mardin CY, Horn FK, Jonas JB & Budde WM (1999) Preperimetric glaucoma diagnosis by confocal scanning laser tomography of the optic disc. Br J Ophthalmol 83: 299–304.

Mardin CY, Hothorn T, Peters A, Jünemann AG, Nguyen NX & Lausen B (2003) New glaucoma classifi cation method based on standard Heidelberg Retina Tomograph parameters by bagging classifi cation trees. J Glaucoma 12: 340–6.

Medeiros FA, Zangwill LM, Bowd C, Sample PA & Weinreb RN (2005) Use of progressive glaucomatous optic disk change as the reference standard for evaluation of diagnostic tests in glaucoma. Am J Ophthalmol 139: 1010–18.

78

Medeiros FA, Zangwill LM, Bowd C, Vasile C, Sample PA & Weinreb RN (2007) Agreement between stereophotographic and confocal scanning laser ophthalmoscopy measurements of cup/disc ratio: effect on a predictive model for glaucoma development. J Glaucoma 16: 209–14.

Medeiros FA, Zangwill LM, Bowd C & Weinreb RN (2004) Comparison of the GDx VCC scanning laser polarimeter, HRT II confocal scanning laser ophthalmoscope, and stratus OCT optical coherence tomograph for the detection of glaucoma. Arch Ophthalmol 122: 827–37.

Migdal C, Gregory W & Hitchings R (1994) Longterm functional outcome after early surgery compared with laser and medicine in open-angle glaucoma. Ophthalmology 101: 1651–1657.

Miglior S, Albe E, Guareschi M, Rossetti L & Orzalesi N (2002a) Intraobserver and interobserver reproducibility in the evaluation of optic disc stereometric parameters by Heidelberg Retina Tomograph. Ophthalmology 109: 1072–7.

Miglior S, Guareschi M, Albe E, Gomarasca S, Vavassori M & Orzalesi N (2003) Detection of glaucomatous visual fi eld changes using the Moorfi elds regression analysis of the Heidelberg retina tomograph. Am J Ophthalmol 136: 26–33.

Miglior S, Guareschi M, Romanazzi F, Albe E, Torri V & Orzalesi N (2005) The impact of defi nition of primary open-angle glaucoma on the cross-sectional assessment of diagnostic validity of Heidelberg retinal tomography. Am J Ophthalmol 139: 878–87.

Miglior S, Zeyen T, Pfeiffer N, Cunha-Vaz J, Torri V & Adamsons I; European Glaucoma Prevention Study Group (2002b) The European glaucoma prevention study design and baseline description of the participants. Ophthalmology 109: 1612–21.

Mikelberg FS, Parfi tt CM, Swindale NV, Graham SL, Drance SM & Gosine R (1995) Ability of the Heidelberg retina tomograph to detect early glaucomatous visual fi eld loss. J Glaucoma 4: 242–7.

Mikelberg FS, Wijsman K & Schulzer M (1993) Reproducibility of topographic parameters obtained with the Heidelberg Retina Tomograph. J Glaucoma 2: 101–103.

Mitchell P, Hourihan F, Sandbach J & Wang J (1999) The relationship between glaucoma and myopia. The Blue Mountains Eye Study. Ophthalmology 106: 2010–2015.

Mitchell P, Smith W, Attebo K & Healey PR (1996) Prevalence of open-angle glaucoma in Australia. The Blue Mountains Eye Study. Ophthalmology. 103: 1661–9.

Morgan JE, Sheen NJ, North RV, Choong Y & Ansari E (2005) Digital imaging of the optic nerve head: monoscopic and stereoscopic analysis. Br J Ophthalmol 89: 879–84.

Morgan WH, Chauhan BC, Yu DY, Cringle SJ, Alder VA & House PH (2002) Optic Disc Movement with Variations in Intraocular and Cerebrospinal Fluid Pressure. Invest Ophthalmol Vis Sci 43: 3236–3242.

Musch DC, Lichter PR, Guire KE & Standardi CL (1999) The Collaborative Initial Glaucoma Treatment Study: study design, methods, and baseline characteristics of enrolled patients. Ophthalmology 106: 653–662.

Nemesure B, Leske MC, He Q, Mendell N & the Barbados Eye Study Group (1996) Analysis of reported family history of glaucoma. A preliminary investigation. Ophthalmic Epidemiol 3: 135–141.

79

Nicolela MT, McCormick TA, Drance SM, Ferrier SN, LeBlanc RP & Chauhan BC (2003) Visual fi eld and optic disc progression in patients with different types of optic disc damage: a longitudinal prospective study. Ophthalmology 110: 2178–2184.

Nicolela MT, Soares AS, Carrillo MM, Chauhan BC, LeBlanc RP & Artes PH (2006) Effect of moderate intraocular pressure changes on topographic measurements with confocal scanning laser tomography in patients with glaucoma. Arch Ophthalmol 124: 633–640.

Niessen AGJE, Langerhorst CT, Geijssen HC & Greve EL (1997) Design of low cost glaucoma screening. Doc Ophthalmol 93: 293–315.

O’Leary N, Mansberger SL, Twa MD, Fortune BA, Lloyd MJ, Cioffi GA, Johnson CA, Garway-Heath DF & Crabb DP (2008) Glaucomatous Progression in Series of Stereo-Paired Photographs and Heidelberg Retinal Tomography Images. Invest Ophthalmol Vis Sci 49: E-Abstract 5431.

Orgül S, Cioffi GA, Bacon DR & Van Buskirk EM (1996) Sources of variability of topometric data with a scanning laser ophthalmoscope. Arch Ophthalmol 114: 161–4.

Owen VM, Strouthidis NG, Garway-Heath DF & Crabb DP (2006) Measurement variability in Heidelberg Retina Tomograph imaging of neuroretinal rim area. Invest Ophthalmol Vis Sci 47: 5322–5330.

Parikh RS, Parikh S, Sekhar GC, Kumar RS, Prabakaran S, Babu JG & Thomas R (2007) Diagnostic capability of optical coherence tomography (Stratus OCT 3) in early glaucoma. Ophthalmology 114: 2238–43.

Park KH & Caprioli J (2002) Development of a novel reference plane for the Heidelberg retina tomograph with optical coherence tomography measurements. J Glaucoma 11: 385–91.

Parrish RK, Schiffman JC, Feuer WJ, Anderson DR, Budenz DL, Wells-Albornoz MC, Vandenbroucke R, Kass MA & Gordon MO; Ocular Hypertension Treatment Study Group (2005) Test-retest reproducibility of optic disk deterioration detected from stereophotographs by masked graders. Am J Ophthalmol 140: 762–4.

Parrow KA, Shin DH, Tsai CS, Hong YJ, Juzych MS & Shi DX (1992) Intraocular pressure-dependent dynamic changes of optic disc cupping in adult glaucoma patients. Ophthalmology 99: 36–40.

Patterson AJ, Garway-Heath DF, Strouthidis NG & Crabb DP (2005) A new statistical approach for quantifying change in series of retinal and optic nerve head topography images. Invest Ophthalmol Vis Sci 46: 1659–1667.

Philippin H, Unsoeld A, Maier P, Walter S, Bach M & Funk J (2006) Ten-year results: detection of long-term progressive optic disc changes with confocal laser tomography. Graefes Arch Clin Exp Ophthalmol 244: 460–4.

Pohjalainen T, Vesti E, Uusitalo RJ & Laatikainen L (2001) Phacoemulsifi cation and intraocular lens implantation in eyes with open-angle glaucoma. Acta Ophthalmol Scand 79: 313–316.

Poli A, Strouthidis NG, Ho TA & Garway-Heath DF (2008) Analysis of HRT images: comparison of reference planes. Invest Ophthalmol Vis Sci 49: 3970–5.

80

Pueyo V, Polo V, Larrosa JM, Ferreras A, Pablo LE & Honrubia FM (2007) Diagnostic ability of the Heidelberg retina tomograph, optical coherence tomograph, and scanning laser polarimeter in open-angle glaucoma. J Glaucoma 16: 173–177.

Quigley HA, Addicks EM & Green WR (1982) Optic nerve damage in human glaucoma. III. Quantitative correlation of nerve fi bre loss and visual fi eld defect in glaucoma, ischaemic neuropathy, papilloedema, and toxic neuropathy. Arch Ophthalmol 100: 135–146.

Quigley HA & Broman AT (2006) The number of people with glaucoma worldwide in 2010 and 2020. Br J Ophthalmol 90: 262–7.

Quigley HA, Enger C, Sommer A, Scott R & Gilbert D (1994) Risk factors for the development of glaucomatous visual fi eld damage in ocular hypertension. Arch Ophthalmol 112: 644–649.

Quigley HA, Katz J, Derick RJ, Gilbert D & Sommer A (1992) An evaluation of optic disc and nerve fi ber layer examinations in monitoring progression of early glaucoma damage. Ophthalmology 99: 19–28.

Quigley HA, Miller NR & George T (1980) Clinical evaluation of nerve fi ber layer atrophy as an indicator of glaucomatous optic nerve damage. Arch Ophthalmol 98: 1564–71.

Quigley HA, Reacher M, Katz J, Strahlman E, Gilbert D & Scott R (1993) Quantitative grading of nerve fi ber layer photographs. Ophthalmology 100: 1800–7.

Read RM & Spaeth GL (1974) The practical clinical appraisal of the optic disc in glaucoma: the natural history of cup progression and some specifi c disc-fi eld correlations. Trans Am Acad Ophthalmol Otolaryngol 78: 255–74.

Reddy S, Xing D, Arthur SN, Harizman N, Dorairaj S, Ritch R & Liebmann JM (2009) HRT III glaucoma probability score and Moorfi elds regression across the glaucoma spectrum. J Glaucoma 18: 368–72.

Reidy A, Minassian DC, Vafi dis G, Joseph J, Farrow S, Wu J, Desai P & Connolly A (1998) Prevalence of serious eye disease and visual impairment in a north London population: population based, cross sectional study. BMJ 316: 1643–6.

Reus NJ, de Graaf M & Lemij HG (2007) Accuracy of GDx VCC, HRT I, and clinical assessment of stereoscopic optic nerve head photographs for diagnosing glaucoma. Br J Ophthalmol 91: 313–8.

Reus NJ, Lemij HG, Garway-Heath DF, Airaksinen PJ, Anton A, Bron AM, Faschinger C, Holló G, Iester M, Jonas JB, Mistlberger A, Topouzis F & Zeyen TG (2010) Clinical Assessment of Stereoscopic Optic Disc Photographs for Glaucoma: The European Optic Disc Assessment Trial. Ophthalmology 117: 717–23.

Ringvold A, Blika S, Elsås T, Guldahl J, Brevik T, Hesstvedt P, Hoff K, Høisen H, Kjørsvik S & Rossvold I (1991) The Middle-Norway Eye Screening Study. II. Prevalence of simple and capsular glaucoma. Acta Ophthalmol Scand 69: 273–280.

Robin TA, Muller A, Rait J, Keeffe JE & Taylor HR (2005) Performance of community-based glaucoma screening using frequency doubling technology and Heidelberg retinal tomography. Ophthalmic Epidemiol 12: 167–78.

81

Rohrschneider K, Burk RO, Kruse FE & Völcker HE (1994) Reproducibility of the optic nerve head topography with a new laser tomographic scanning device. Ophthalmology 101: 1044–9.

Schulzer M (1994) Errors in the diagnosis of visual fi eld progression in normal-tension glaucoma. Ophthalmology 101: 1589–94.

Schuman JS, Hee MR, Puliafi to CA, Wong C, Pedut-Kloizman T, Lin CP, Hertzmark E, Izatt JA, Swanson EA & Fujimoto JG (1995) Quantifi cation of nerve fi ber layer thickness in normal and glaucomatous eyes using optical coherence tomography. Arch Ophthalmol 113: 586–596.

Sehi M, Guaqueta DC, Feuer WJ & Greenfi eld DS; Advanced Imaging in Glaucoma Study Group (2007) Scanning laser polarimetry with variable and enhanced corneal compensation in normal and glaucomatous eyes. Am J Ophthalmol 143: 272–279.

Shunmugam M & Azuara-Blanco A (2006) The quality of reporting of diagnostic accuracy studies in glaucoma using the Heidelberg retina tomograph. Invest Ophthalmol Vis Sci 47: 2317–23.

Sihota R, Gulati V, Agarwal HC, Saxena R, Sharma A & Pandey RM (2002) Variables affecting test-retest variability of Heidelberg Retina Tomograph II stereometric parameters. J Glaucoma 11: 321–8.

Siik S, Airaksinen PJ, Tuulonen A & Nieminen H (1997) Infl uence of lens autofl uorescence on retinal nerve fi ber layer evaluation. Acta Ophthalmol Scand 75: 524–7.

Sommer A, Katz J, Quigley HA, Miller NR, Robin AL, Richter RC & Witt KA (1991a) Clinically detectable nerve fi ber atrophy precedes the onset of glaucomatous fi eld loss. Arch Ophthalmol 109: 77–83.

Sommer A, Tielsch JM, Katz J, Quigley HA, Gottsch JD, Javitt J & Singh K (1991b) Relationship between intraocular pressure and primary open-angle glaucoma among white and black Americans. The Baltimore Eye Survey. Arch Ophthalmol 109: 1090–1095.

Spaeth GL, Henderer J, Liu C, Kesen M, Altangerel U, Bayer A, Katz LJ, Myers J, Rhee D & Steinmann W (2002) The disc damage likelihood scale: reproducibility of a new method of estimating the amount of optic nerve damage caused by glaucoma. Trans Am Ophthalmol Soc 100: 181–5.

Strouthidis NG, Demirel S, Asaoka R, Cossio-Zuniga C & Garway-Heath DF (2010) The Heidelberg Retina Tomograph Glaucoma Probability Score Reproducibility and Measurement of Progression. Ophthalmology 117: 724–729.

Strouthidis NG & Garway-Heath DF (2008) New developments in Heidelberg retina tomograph for glaucoma. Curr Opin Ophthalmol. 19: 141–8.

Strouthidis NG, Scott A, Peter NM & Garway-Heath DF (2006) Optic disc and visual fi eld progression in ocular hypertensive subjects: detection rates, specifi city, and agreement. Invest Ophthalmol Vis Sci 47: 2904–10.

Strouthidis NG, White ET, Owen VM, Ho TA & Garway-Heath DF (2005a) Factors affecting the test-retest variability of Heidelberg retina tomograph and Heidelberg retina tomograph II measurements. Br J Ophthalmol 89: 1427–32.

82

Strouthidis NG, White ET, Owen VM, Ho TA, Hammond CJ & Garway-Heath DF (2005b) Improving the repeatability of Heidelberg retina tomograph and Heidelberg retina tomograph II rim area measurements. Br J Ophthalmol 89: 1433–7.

Swindale NV, Stjepanovic G, Chin A & Mikelberg FS (2000) Automated analysis of normal and glaucomatous optic nerve head topography images. Invest Ophthalmol Vis Sci 41: 1730–42.

Taibbi G, Fogagnolo P, Orzalesi N & Rossetti L (2009) Reproducibility of the Heidelberg Retina Tomograph III Glaucoma Probability Score. J Glaucoma 18: 247–52.

Tan JCH, Garway-Heath DF & Hitchings RA (2003) Reasons for rim area variability in scanning laser tomography. Invest Ophthalmol Vis Sci 44: 1126–1131.

Tan JC & Hitchings RA (2004a) Optimizing and validating an approach for identifying glaucomatous change in optic nerve topography. Invest Ophthalmol Vis Sci 45: 1396–403.

Tan JC & Hitchings RA (2003) Reference plane defi nition and reproducibility in optic nerve head images. Invest Ophthalmol Vis Sci 44: 1132–7.

Tan JC & Hitchings RA (2004b) Reversal of disc cupping after intraocular pressure reduction in topographic image series. J Glaucoma 13: 351–5.

Tan JC, Poinoosawmy D, Fitzke FW & Hitchings RA (2004a) Magnifi cation changes in scanning laser tomography. J Glaucoma 13: 137–41.

Tan JC, White E, Poinoosawmy D & Hitchings RA (2004b) Validity of rim area measurements by different reference planes. J Glaucoma 13: 245–50.

Teesalu P & Airaksinen PJ (1998) Clinical and photographic assessment of the retinal nerve fi bre layer. Ophthalmol Clin N Am 11: 297–309.

Tielsch JM, Katz J, Quigley HA, Miller NR & Sommer A (1988) Intraobserver and interobserver agreement in measurement of optic disc characteristics. Ophthalmology 95: 350–6.

Tielsch JM, Katz J, Sommer A, Quigley HA & Javitt JC (1994) Family history and risk of primary open-angle glaucoma. Baltimore Eye Survey. Arch Ophthalmol 112: 69–73.

Tielsch JM, Katz J, Sommer A, Quigley HA & Javitt JC (1995) Hypertension, perfusion pressure and primary open-angle glaucoma. A population-based assessment. Arch Ophthalmol 113: 216–221.

Tielsch JM, Sommer A, Katz J, Royall RM, Quigley HA & Javitt J (1991) Racial variations in the prevalence of primary open-angle glaucoma. The Baltimore Eye Survey. JAMA 266: 369–74

Tuulonen A & Airaksinen PJ (1991) Initial glaucomatous optic disc and retinal nerve fi ber layer abnormalities and their progression. Am J Ophthalmol 111: 485–90.

Tuulonen A, Airaksinen PJ, Erola E, Forsman E, Friberg K, Kaila M, Klemetti A, Mäkelä M, Oskala P, Puska P, Suoranta L, Teir H, Uusitalo H, Vainio-Jylhä E & Vuori ML (2003) The Finnish evidence-based guideline for open-angle glaucoma. Acta Ophthalmol Scand 81: 3–18.

83

Tuulonen A, Alanko H, Hyytinen P, Veijola J, Seppänen T & Airaksinen PJ (2000) Digital imaging and microtexture analysis of the nerve fi ber layer. J Glaucoma 9: 5–9.

Tuulonen A, Lehtola J & Airaksinen PJ (1993) Nerve fi ber layer defects with normal visual fi elds. Do normal optic disc and normal visual fi eld indicate absence of glaucomatous abnormality? Ophthalmology 100: 587–97.

Tuulonen A, Vihanninjoki K, Airaksinen PJ, Alanko H & Nieminen H (1994) The effect of reference levels on neuroretinal rim area and rim volume measurements in the Heidelberg Retina Tomograph (HRT) [Abstract]. Invest Ophthalmol Vis Sci 35: 1729.

Uchida H, Brigatti L & Caprioli J (1996) Detection of structural damage from glaucoma with confocal laser image analysis. Invest Ophthalmol 37: 2393–401.

Vaahtoranta-Lehtonen H, Tuulonen A, Aronen P, Sintonen H, Suoranta L, Kovanen N, Linna M, Läärä E & Malmivaara A (2007) Cost effectiveness and cost utility of an organized screening programme for glaucoma. Acta Ophthalmol Scand 85: 508–18.

Varma R, Steinmann WC & Scott IU (1992) Expert agreement in evaluating the optic disc for glaucoma. Ophthalmology 99: 215–21.

Verdonck N, Zeyen T, Van Malderen L & Spileers W (2002) Short-term intra-individual variability in Heidelberg retina tomograph II. Bull Soc Belge Ophtalmol 286: 51–7.

Vesti E, Johnson CA & Chauhan BC (2003) Comparison of different methods for detecting glaucomatous visual fi eld progression. Invest Ophthalmol Vis Sci 44: 3873–9.

Vihanninjoki K, Burk RO, Teesalu P, Tuulonen A & Airaksinen PJ (2002) Optic disc biomorphometry with the Heidelberg retina tomograph at different reference levels. Acta Ophthalmol Scand 80: 47–53.

Vihanninjoki K, Teesalu P, Burk RO, Läärä E, Tuulonen A & Airaksinen PJ (2000) Search for an optimal combination of structural and functional parameters for the diagnosis of glaucoma: multivariate analysis of confocal scanning laser tomograph, blue-on-yellow visual fi eld and retinal nerve fi ber layer data. Graefes Arch Clin Exp Ophthalmol 238: 477–81.

Vizzeri G, Bowd C, Medeiros FA, Weinreb RN & Zangwill LM (2009a) Effect of signal strength and improper alignment on the variability of stratus optical coherence tomography retinal nerve fi ber layer thickness measurements. Am J Ophthalmol 148: 249–255.

Vizzeri G, Weinreb RN, Martinez de la Casa JM, Alencar LM, Bowd C, Balasubramanian M, Medeiros FA, Sample P & Zangwill LM (2009b) Clinicians agreement in establishing glaucomatous progression using the Heidelberg retina tomograph. Ophthalmology 116: 14–24.

Von Graefe A (1855) Notiz über die Lage der Ziliarfortsätze bei Ausdehnung der Sklera. von Graefes Arch Ophthalmol 2: 242–250.

Wang F, Quigley HA & Tielsch JM (1994) Screening for glaucoma in a medical clinic with photographs of the nerve fi ber layer. Arch Ophthalmol 112: 796–800

Weinreb RN, Lusky M, Bartsch DU & Morsman D (1993) Effect of repetitive imaging on topographic measurements of the optic nerve head. Arch Ophthalmol 111: 636–8.

84

Wensor MD, McCarty CA, Stanislavsky YL, Livingston PM & Taylor HR (1998) The prevalence of glaucoma in the Melbourne Visual Impairment Project. Ophthalmology 105: 733–9.

Wild JM, Hutchings N, Hussey MK, Flanagan JG & Trope GE (1997) Pointwise univariate linear regression of perimetric sensitivity against follow-up time in glaucoma. Ophthalmology 104: 808–815.

Windisch BK, Harasymowycz PJ, See JL, Chauhan BC, Belliveau AC, Hutchison DM & Nicolela MT (2009) Comparison between confocal scanning laser tomography, scanning laser polarimetry and optical coherence tomography on the ability to detect localised retinal nerve fi bre layer defects in glaucoma patients. Br J Ophthalmol 93: 225–30.

Wollstein G, Garway-Heath DF & Hitchings RA (1998) Identifi cation of early glaucoma cases with the scanning laser ophthalmoscope. Ophthalmology 105: 1557–63.

Wollstein G, Schuman JS, Price LL, Aydin A, Stark PC, Hertzmark E, Lai E, Ishikawa H, Mattox C, Fujimoto JG & Paunescu LA (2005) Optical coherence tomography longitudinal evaluation of retinal nerve fi ber layer thickness in glaucoma. Arch Ophthalmol 123: 464–70.

Yucel I, Akar M, Durukan A, Akar Y, Taskin O, Dora B & Yilmaz N (2005) The effect of the menstrual cycle on the optic nerve head analysis of migrainous women. Curr Eye Res 30: 163–169.

Zangwill LM, Chan K, Bowd C, Hao J, Lee TW, Weinreb RN, Sejnowski TJ & Goldbaum MH (2004) Heidelberg retina tomograph measurements of the optic disc and parapapillary retina for detecting glaucoma analyzed by machine learning classifi ers. Invest Ophthalmol Vis Sci 45: 3144–51.

Zangwill L, Irak I, Berry CC, Garden V, de Souza Lima M & Weinreb RN (1997) Effect of cataract and pupil size on image quality with confocal scanning laser ophthalmoscopy. Arch Ophthalmol 115: 983–90.

Zangwill LM, Jain S, Racette L, Ernstrom KB, Bowd C, Medeiros FA, Sample PA & Weinreb RN (2007) The effect of disc size and severity of disease on the diagnostic accuracy of the Heidelberg Retina Tomograph Glaucoma Probability Score. Invest Ophthalmol Vis Sci 48: 2653–60.

Zangwill LM, Weinreb RN, Beiser JA, Berry CC, Cioffi GA, Coleman AL, Trick G, Liebmann JM, Brandt JD, Piltz-Seymour JR, Dirkes KA, Vega S, Kass MA & Gordon MO (2005) Baseline topographic optic disc measurements are associated with the development of primary open-angle glaucoma: the Confocal Scanning Laser Ophthalmoscopy Ancillary Study to the Ocular Hypertension Treatment Study. Arch Ophthalmol 123: 1188–97.

Zeyen TG & Caprioli J (1993) Progression of disc and fi eld damage in early glaucoma. Arch Ophthalmol 111: 62–5.

Zeyen T, Crabb DP, Pfeiffer N & Miglior S (2009) The Agreement in the Evaluation of Optic Disc Glaucomatous Change Assessing Color versus Black and White Stereo Photos Invest Ophthalmol Vis Sci 50: E-Abstract 2244.

85

Zeyen T, Miglior S, Pfeiffer N, Cunha-Vaz J & Adamsons I; European Glaucoma Prevention Study Group (2003) Reproducibility of evaluation of optic disc change for glaucoma with stereo optic disc photographs. Ophthalmology 110: 340–4.

Zeyen T, Stalmans I, Vandewalle E & European Glaucoma Prevention Study Group (2007) Agreement Between Non-Expert Ophthalmologists in the Assessment for Glaucomatous Change in Serial Optic Disc Stereo Photos Before and After a Training Session. Invest Ophthalmol Vis Sci 48: E-Abstract 3340.

Zhu H, Crabb DP, Anderson D, Fredette MJ & Garway-Heath DF (2009) Combining Structural and Functional Measurements to Improve Reproducibility of Follow Up Data in Glaucoma. Invest Ophthalmol Vis Sci 50: E-Abstract 2572.

Zinser G, Harbarth U & Schröder H (1990) Formation and analysis of three-dimensional data with the laser tomographic scanner LTS. In: Nasemann, JE & Burk, ROW, eds. Laser Scanning Ophthalmoscopy and Tomography. München, Quintessenz: 243–52.

Zinser G, Wijnaendts-van-Resandt R, Dreher A, Weinreb RN, Harbarth U, Schröder H & Burk ROW (1989) Confocal scanning laser tomography of the eye. Proc SPIE 1161: 337–344.

87

Original publications

I Saarela V & Airaksinen PJ (2008) Heidelberg Retina Tomograph Parameters of the Optic Disc in Eyes With Progressive Retinal Nerve Fibre Layer Defects. Acta Ophthalmol 86: 603–8.

II Saarela V, Falck A, Airaksinen PJ & Tuulonen A (2010) The sensitivity and specifi city of Heidelberg Retina Tomograph parameters to glaucomatous progression in disc photographs. Br J Ophthalmol 94: 68–73.

III Saarela V, Falck A, Airaksinen PJ & Tuulonen A (2010) Factors affecting the sensitivity and specifi city of the Heidelberg Retina Tomograph parameters to glaucomatous progression in disc photographs. Acta Ophthalmol [Epub ahead of print] doi: 10.1111/j.1755-3768.2010.01881.x

IV Saarela V, Ahvenvaara E & Tuulonen A Variability of Heidelberg Retina Tomograph parameters during exercise. Manuscript.

Reprinted with permission from John Wiley and Sons Ltd (I, III) and BMJ Publishing Group Ltd (II).

Original publications are not included in the electronic version of the dissertation.


Book orders:Granum: Virtual book storehttp://granum.uta.fi/granum/

S E R I E S D M E D I C A

1060. Vähäkangas, Pia (2010) Kuntoutumista edistävä hoitajan toiminta ja senjohtaminen pitkäaikaisessa laitoshoidossa

1061. Saari, Anne (2010) Autonomic dysfunction in multiple sclerosis and optic neuritis

1062. Eriksen, Heidi (2010) Carboxyterminal telopeptide structures of type I collagen invarious human tissues

1063. Gäddnäs, Fiia (2010) Insights into healing response in severe sepsis from aconnective tissue perspective. A longitudinal case-control study on woundhealing, collagen synthesis and degradation, and matrix metalloproteinases inpatients with severe sepsis

1064. Alaräisänen, Antti (2010) Risk factors and pathways leading to suicide with specialfocus in schizophrenia. The Northern Finland 1966 Birth Cohort Study

1065. Oja, Paula (2010) Significance of customer feedback. An analysis of customerfeedback data in a university hospital laboratory

1066. Mustonen, Erja (2010) The expression of novel, load-induced extracellular matrixmodulating factors in cardiac remodeling

1067. Mäkinen, Johanna (2010) Systematic search and evaluation of published scientificresearch. Implications for schizophrenia research

1068. Jartti, Pekka (2010) Computed tomography in subarachnoid haemorrhage. Studieson aneurysm localization, hydrocephalus and early rebleeding

1069. Rantala, Aino (2010) Susceptibility to respiratory tract infections in young men:the role of inflammation, mannose-binding lectin, interleukin-6 and their geneticpolymorphisms

1070. Tanskanen, Päivikki (2010) Brain MRI in subjects with schizophrenia and in adultsborn prematurely. The Northern Finland 1966 Birth Cohort Study

1071. Abass, Khaled M. (2010) Metabolism and interactions of pesticides in human andanimal in vitro hepatic models

1072. Luukkonen, Anu-Helmi (2010) Bullying behaviour in relation to psychiatricdisorders, suicidality and criminal offences. A study of under-age adolescentinpatients in Northern Finland

1073. Ahola, Riikka (2010) Measurement of bone exercise. Osteogenic features ofloading

1074. Krüger, Johanna (2010) Molecular genetics of early-onset Alzheimer's disease andfrontotemporal lobar degeneration

ABCDEFG





HUMANIORA

TECHNICA

MEDICA


SCRIPTA ACADEMICA

OECONOMICA

EDITOR IN CHIEF

PUBLICATIONS EDITOR












MEDICA

ACTAD

D 1076

ACTA

Ville Saarela

OULU 2010

D 1076

Ville Saarela



d 1076 acta universitatis ouluensis university of …jultika.oulu.fi/files/isbn9789514263286.pdfthe...

Documents