A Tactile Resonance Sensor System for Detection of Prostate Cancer ex vivo – Design and Evaluation on Tissue Models and Human Prostate

Anders P Åstrand

Department of Applied Physics and Electronics

Centre for Biomedical Engineering and Physics

Industrial Doctoral School

Umeå 2014

Department of Applied Physics and Electronics

Centre for Biomedical Engineering and Physics

Industrial Doctoral School

Umeå University

SE-901 87 Umeå, Sweden

This work is protected by Swedish copyright legislation (Act 1960:729)

© 2014 by Anders P Åstrand

ISBN: 978-91-7601-006-8

ISSN: 1653-6789:6

Electronic version http://umu.diva-portal.org/

Printed by: VMC-KBC

Umeå, Sweden 2014

Dedicated to my son Tobias

i

Table of Contents

Table of Contents i Abstract iii 1. Original papers v 2. Abbreviations vii 3. Introduction 1

3.1 General introduction 1 3.2 Aims 2

4. Anatomy, physiology and tactile resonance sensors 3 4.1 Prostate anatomy and pathology 3 4.2 Methods for diagnosis, in brief 4 4.3 Resonance sensors and theory 6 4.4 Tactile sensors in medical applications 8 4.4.1 Piezoelectric tactile resonance sensors 8 4.4.2 Other types of tactile sensors 8 4.5 Soft tissue silicone phantoms and biological tissue 9

5. Materials and methods 11 5.1 The piezoelectric resonance sensor 11 5.2 The measurement system 12 5.3 Technical setup 13 5.4 Preparation of soft tissue phantoms made of silicone 14 5.5 Preparation of biological tissue phantoms 15 5.6 Human prostate samples 16 5.7 Measurement set-up 17 5.7.1 Measurements on soft tissue silicone phantoms 18 5.7.2 Measurements on biological tissue phantoms 19 5.7.3 Measurements on excised human prostate tissue 20 5.8 Maximum estimated errors and overall accuracy 20 5.9 Statistics 21

6. Results and discussion 23 6.1 Instrument performance and accuracy 23 6.2 Angular dependency 24 6.3 Velocity dependency 26 6.4 Depth sensitivity 27 6.5 The clamping force 28 6.6 Measurements on prostate gland ex vivo 29

7. General Summary and Conclusion 33 7.1 Conclusion 33

8. Acknowledgements 35 9. References 37 Summary of papers 45

ii

iii

Abstract

Background

The most common form of cancer among males in Europe and the USA is

prostate cancer, PCa. Surgical removal of the prostate is the most common

form of curative treatment. PCa can be suspected by a blood test for a

specific prostate antigen, a PSA-test, and a digital rectal examination, DRE

where the physician palpates the prostate through the rectum. Stiff nodules

that can be detected during the DRE, and elevated levels of PSA are

indications for PCa, and a reason for further examination. Biopsies are taken

from the prostate by guidance of a transrectal ultrasound. Superficial cancer

tumours can indicate that the cancer has spread to other parts of the body.

Tactile resonance sensors can be used to detect areas of different stiffness in

soft tissue. Healthy prostate tissue is usually of different stiffness compared

to tissue with PCa.

Aim

The general aim of this doctoral thesis was to design and evaluate a flexible

tactile resonance sensor system (TRSS) for detection of cancer in soft human

tissue, specifically prostate cancer. The ability to detect cancer tumours

located under the surface was evaluated through measurements on tissue

phantoms such as silicone and biological tissues. Finally measurements on

resected whole prostate glands were made for the detection of cancer

tumours.

Methods

The sensor principle was based on an oscillating piezoelectric element that

was indented into the soft tissue. The measured parameters were the change

in resonance frequency, Δf, and the contact force F during indentation. From

these, a specific stiffness parameter fF was obtained. The overall

accuracy of the TRSS was obtained and the performance of the TRSS was

also evaluated on tissue models made of silicone, biological tissue and

resected whole human prostates in order to detect presence of PCa. Prostate

glands are generally spherical and a special rotatable sample holder was

included in the TRSS. Spherically shaped objects and uneven surfaces call for

special attention to the contact angle between the sensor-tip and the

measured surface, which has been evaluated. The indentation velocity and

iv

the depth sensitivity of the sensor were evaluated as well as the effect on the

measurements caused by the force with which spherical samples were held in

place in the sample holder. Measurements were made on silicone models

and biological tissue of chicken and pork muscles, with embedded stiff

silicone nodules, both on flat and spherical shaped samples. Finally,

measurements were made on two excised whole human prostates.

Results

A contact angle deviating ≤ 10° from the perpendicular of the surface of the

measured object was acceptable for reliable measurements of the stiffness

parameter. The sensor could detect stiff nodules ≤ 4 mm under the surface

with a small indentation depth of 0.4 to 0.8 mm.

Measurements on the surface of resected human prostate glands showed

that the TRSS could detect stiff areas (p < 0.05), which were confirmed by

histopathological evaluation to be cancer tumours on, and under the surface.

Conclusions

A flexible resonance sensor system was designed and evaluated on soft tissue

models as well as resected whole prostate glands. Evaluations on the tissue

models showed that the TRSS can detect stiffer volumes hidden below the

surface on both flat and spherical samples. The measurements on resected

human prostate glands showed that PCa could be detected both on and

under the surface of the gland. Thus the TRSS provides a promising

instrument aimed for stiffness measurements of soft human tissue that could

contribute to a future quantitative palpation method with the purpose of

diagnosing cancer.

v

1. Original papers

This doctoral thesis is based on the following papers, which are referred to

by Roman numerals. The papers are reprinted with permission from the

publisher when applicable

I. Åstrand AP, Jalkanen V, Andersson BM and Lindahl OA. A flexible

sensor system using resonance sensor technology for soft tissue

stiffness measurements – evaluation on silicone. Presented at the

15th Nordic Baltic Conference on Biomedical Engineering and

Medical Physics, Aalborg, Denmark, 16 June 2011. IFMBE

Proceedings 34 21-24. www.springerlink.com.

II. Åstrand AP, Jalkanen V, Andersson BM and Lindahl OA. Stiffness

measurements on spherical surfaces of prostate models using a

resonance sensor. Presented at the World Congress 2012 on Medical

Physics and Biomedical Engineering, Beijing, China, 30 May 2012.

IFMBE Proceedings 39 1401-1404. www.springerlink.com.

III. Åstrand AP, Jalkanen V, Andersson BM and Lindahl OA. A flexible

resonance sensor instrument for measurements of soft tissue –

evaluation on silicone models. Journal of Medical Engineering &

Technology 2012: 37 (3) 185-196.

DOI: 10.3109/03091902.2013.773097.

IV. Åstrand AP, Jalkanen V, Andersson BM and Lindahl OA. Detection

of stiff nodules embedded in soft tissue phantoms, mimicking cancer

tumours, using a tactile resonance sensor. Journal of Biomedical

Science and Engineering. Accepted 2014.

V. Åstrand AP, Andersson BM, Jalkanen V, Ljungberg B, Bergh A and

Lindahl OA. 2014: The first study on whole human prostate ex vivo

using a tactile resonance sensor for cancer detection. Manuscript.

vi

Peer-reviewed conference publications of relevance, but not

included in the thesis:

Åstrand AP, Jalkanen V, Lindahl OA and Andersson BM. Mätutrustning för

kontrollerad mätning och karatärisering av vävnad. Medicinteknik-dagarna i

Umeå, Sweden , 7 October 2010.

Åstrand AP, Jalkanen V, Lindahl OA and Andersson BM.

Ansättningsvinkelns betydelse vid mätning med piezoelektriska

resonanssensorer på vävnadsmodeller av silikon. Medicinteknikdagarna i

Linköping, Sweden, 12 October 2011.

Other publications:

Åstrand AP. A Flexible Resonans Sensor System for Detection of Cancer

Tissue – Evaluation on Silicone. Licentiate Thesis, 2012. ISBN 978-91-7459-

477-5

vii

2. Abbreviations

BPH Benign prostatic hyperplasia

CT Computer tomography

DRE Digital rectal examination

GS Gleason score

PCa Prostate cancer

MP Measurement position

MRI Magnetic resonance imaging

PLL Phase-locked loop

PSA Prostate specific antigen

PZT Lead zirconate titanate (referring to the piezoelectric element)

SD Standard deviation

TNM Tumour, Nodule, Metastasis

TRSS Tactile Resonance Sensor System

TRUS Transrectal ultrasound TURP Transurethral resection of the prostate

α Contact angle between sensor and object

αr Rotation angle

αsm Angle for sensor movement

d Distance to the surface above the centre of a stiff inclusion

D Diameter

viii

ε Error

Δf Frequency shift

f Loaded resonance frequency

F Contact force

FC Clamping force

Itot Total indentation - indetation depth during measurement

IZ Indentation - indetation depth of interest

n Number of measurements

vi Indentation velocity

fF Stiffness parameter

1

3. Introduction

3.1 General introduction

Research in new and improved methods for early detection of prostate

cancer (PCa) is necessary as PCa is the most common form of cancer among

males in Europe and in the USA. There were about 417000 new cases of PCa

diagnosed in Europe in 2012, and nearly 92000 deaths were reported [1].

The American Cancer Society estimates that more than 238000 new cases

will be diagnosed with PCa in 2013 and that close to 30000 men will die

from PCa [2]. The latest report from the Swedish National Board of Health

(Socialstyrelsen) states that PCa stood for 32.2% of the new cases of males

diagnosed with cancer in 2011, reporting 9663 new cases and 2382 deaths

due to PCa [3].

Common diagnostic methods for detecting PCA are digital rectal

examination (DRE), where the physician palpates the prostate from the

rectum and a blood test for a prostate specific antigen (PSA) [3]. However,

an elevated PSA-level can be due to other causes than PCa and the use of this

alone can lead to over-diagnosis [4]. A transrectal ultrasound (TRUS) guided

systematic needle biopsy is necessary for final diagnosing [5]. The outcomes

of these tests are important for decisions regarding what treatments are

thought to give the best results in each case.

When performing different kinds of open surgery, palpation of areas with

suspected cancer can be used for assessing areas of stiffness and other

abnormalities of the tissue. Earlier studies have reported that tumours in

soft tissue usually are stiffer compared to healthy tissue [6, 7, 8, 9]. An

artificial tactile sensor based on a piezoelectric resonance sensor can be of

aid for the physician when performing the otherwise general palpation of the

tissue [10]. Palpation of soft tissue has been performed during different

thoracotomies, where undetected malignant pulmonary nodules were found

during lobectomies, metastatic tumours has been found on patients with

colorectal cancer, tumours that had not been detected with prior helical

computed tomographies. [11, 12, 13, 14] For PCa diagnosis, the results of the

DRE are subjective and dependant on the physicians’ experience [15]. In a

study where palpation for detecting texture abnormalities in thoracic

paraspinal region were evaluated, they concluded that palpation to assess

2

differences in texture was complex and subjective depending on the

examiners’ experience [16]. For this reason, an objective method and

quantitative parameter related to the prostate tissue might be useful to

physicians.

The tactile resonance sensor system designed and evaluated in this

study, as well as in previous studies [III, IV] provides an instrument aimed

for stiffness measurements on resected whole human prostates.

3.2 Aims

The overall aim was to develop a sensor system for possible detection of

cancer in soft human tissue, specifically prostate cancer.

- To design a flexible resonance sensor system with the possibility to

measure on flat and spherical objects.

- To determine the performance of the system and maximum errors.

- To evaluate how the system parameters are affected by variations in

contact angle and indentation velocity, using flat soft tissue silicone

phantoms.

- To determine the depth sensitivity of the resonance sensor system,

using flat soft tissue silicone phantoms and biological tissue with

embedded harder silicone inclusions.

- To evaluate the functionality of the flexible resonance sensor system

on spherical samples of soft tissue silicone phantoms and biological

tissue with embedded stiff silicone inclusions.

- To perform measurements on whole human prostates ex vivo in

order to find stiffer areas that can indicate the presence of possible

cancer tumours and evaluate the findings against the results from

the histopathological analysis.

3

4. Anatomy, physiology and tactile resonance sensors

4.1 Prostate anatomy and pathology

The prostate is a gland, about 30 to 40 mm in diameter, which is located

below the urinary bladder, figure 1. The prostate encircles a part of the

urethra as it leaves the urinary bladder. The prostate gland consists of

glandular tissue and fibromuscular stroma which is surrounded by and

wrapped in a thick blanket of smooth muscle fibres called the capsule [17,

18].

Figure 1. Illustration of the prostate anatomy. Modified from Wikipedia.

(http://en.wikipedia.org/wiki/prostate).

4

The prostate produces prostatic fluid, a secretion that stands for about 25 %

of the volume of semen and promotes the mobility and viability of the

sperms [18]. The prostate grows until the age of 30, after which it shows a

decrease in the growth rate. After the age of 45-50 it is common that

enlargements start to occur, so called benign prostatic hyperplasia (BPH).

An enlarged prostate can cause pain and problems to urinate, as the swelling

of the prostate constricts the urethra [17]. The condition can be corrected

with a transurethral resection of the prostate (TURP) where some of

constricting parts of the prostate tissue is surgically removed trough the

urethra.

Prostate cancer (PCa) is an adenocarcinoma, a malignant and

metastasizing cancer that often originates from one of the secretory glands.

To metastize, the malignant cancer cells must migrate from the primary

tumour, penetrate and circulate through the blood stream to reach target

tissues such as bone marrow, the lymphatic system and lungs [17, 19]. For

this reason it is important to diagnose PCa at an early stage.

4.2 Methods for diagnosis, in brief

The most common screening methods for PCa are a blood test for a prostate

specific antigen (PSA) and a digital rectal examination (DRE), where the

physician palpates the prostate through the rectum. The aim of the palpation

through the rectal wall is to detect stiff areas or nodules on the prostate, as it

has been shown that cancer tumours in general are stiffer than the

surrounding healthy tissue [7, 8, 9, 20, 21]. In the case that the result from

the DRE is positive for PCa, transrectal ultrasound (TRUS) guided biopsies

are used for the final diagnosis [22]. With TRUS it is possible to determine

the volume of the prostate to determine the PSA density, but the main use

for this method is to guide the needles used for biopsies in different areas of

the prostate. Biopsies are usually obtained from six areas; left and right of

the apex, the midgland and the base, a total of 10 to 12 cores are common

practice [23]. However, for the early stages of PCa when the tumours might

be small, the needles can miss existing nodules which results in false-

negative determinations [24, 25]. The imaging method could be improved by

colour Doppler technology and a contrast enhanced technology with micro

bubbles, which increases the rate of prostate cancer detection [23, 25, 26].

Computer tomography (CT) produces images by the use of X-rays to

create images. CT is not considered an effective method for imaging and

diagnosing PCa as it lacks when it comes to distinguishing the prostate

anatomy from, for example nearby muscle tissue, the bladder wall etc. which

5

makes it difficult to detect cancers within the prostate gland. However, CT

can be useful as an imaging method for high risk patients’ with advanced

local disease [23, 26]. As prostate cancer can metastasize to the bone CT has

been used to monitor bone metastases, but for this purpose Magnetic

resonance imaging (MRI) is a better alternative [23, 27].

MRI can be used for visualizing cancer and determine its volume, extent

and location. MRI uses a strong magnetic field and by analysing the

resonance signal that is generated by the hydrogen nuclei and the protons in

the tissue when the magnetic field is disturbed by pulsing radio waves,

images can be created. MRI have an important role for patients where there

is a discrepancy between PSA-tests and the result from the biopsies taken

with the aid of TRUS, as such discrepancy might indicate missed cancer

tumours [28]. MRI is considered as a good method for imaging the prostate

anatomy and locations of PCa. The MRI has a high resolution for soft tissue

imaging, and can also be used for staging of the PCa [23, 26]. The method

should be avoided earlier than 4 to 8 weeks after biopsies [29, 30] as

haemorrhages, which are common after biopsies, can mimic cancer and

therefore limit the detection of actual PCa [26].

Before any decisions regarding the choice of treatment, the severity of the

PCa is classified by a classification system called the Gleason Score (GS), a

scale ranging from 1 to 5 with under groups. The two most descripting scores

are added which results in a GS with a maximum of 10, which would indicate

a severe condition with a bad prognosis [31, 32]. After surgery or a radical

prostatectomy, the tumour and possible tumour spreading, are classified by

a pathological classification system; tumour, nodule, metastasis (TNM)

staging system [33].

The different options for treatment vary depending on the medical

condition and age of the patient, as well as the grading and the stage of the

cancer. About half of the patients diagnosed with PCa are treated by radical

prostatectomy or radiation therapy which can be administered as external

beam radiation och by implants of radioactive seeds into the prostate

(brachytherapy). Hormonal treatment may be given before or after surgery.

All of these treatments may have an effect on the quality of life for the

patient, as side effects such as urinary and erectile difficulties are common

complications. For some patients with an early diagnose and a slow growing

PCa, an alternative is active monitoring with planned regular PSA-tests and

biopsies. Hormonal treatment is also an alternative for early detected PCa,

which may be given in order to temporarily stop or slow down the growth of

a tumour [2, 3].

6

4.3 Resonance sensors and theory

The theory behind the characteristics of resonance sensors, like the one used

in this study, can approximately be described by a model of vibrations in a

finite rod which has been reported by Omata et al [34] and Jalkanen et al

[35]. The rod is set to vibrate with its resonance frequency, f0, and when the

tip of the sensor comes into contact with an object, a change in the resonance

frequency occurs which can be described with

0

0

2 lZ

Vf x

1

where V0 stands for the wave velocity in the rod, βx is the reactance part of

the impedance, l is the length of the rod and Z0 is the acoustic impedance of

the rod. The impedance of the object with which the rod comes into contact

can be written

xxx iZ 2

where αx is the resistance. The reactance part of the impedance can be

describes as

x

xx

km 3

where mx is the inertia term, ω is the angular frequency and kx is the term for

the hardness (stiffness) of the surface. To calculate mx and kx, for these case

of a spherical tip, also used in this study, the following equations can be used

23

23

11

1

4S

v

amx

4

21

221 1

2S

v

Ek x

5

7

where the Poisson’s ratio is denoted ν (for tissue samples ν = 0.495 is

commonly used [36]) and a11 is a coefficient that depends on Poisson’s ratio

[37]. ρ is the density of the contacted object and S is the surface of the

contact area. E is the elastic modulus (Young’s modulus). The theoretical

model of the finite rod was studied by Jalkanen et al [35] with a Venustron®

resonance sensor system and it was found that as mx >> (kx/ω), the term for

the surface stiffness, can be omitted. Thus with equations 1 to 4 it is possible

to write

23Sf 6

which is valid if Poisson’s ratio is assumed to be constant, and the specific

rod is oscillating at a constant frequency. Since the contact force F between

the sensor and the contacted object can be written as [35].

23ESF 7

Eq / gives that the surface contact area S is depending on F, and by using this

expression in equation 6 the following is valid

E

Ff

8

By introducing a stiffness sensitive parameter, fF , equation 8 can be

written as

E

f

F

9

The stiffness parameter fF was derived from the measured Δf and F in

the work by Jalkanen et al [35], where they verified a theoretical model in

measurements on human prostate tissue. The stiffness parameter was found

valid, as the variations in density were small and mostly non-significant. In

this work, the absolute value, fF , of the stiffness parameter was used.

8

4.4 Tactile sensors in medical applications

4.4.1 Piezoelectric tactile resonance sensors

Resonance measurement systems are based on the principle of a

piezoelectric element oscillating at its resonance frequency, and the change

in the frequency that occurs when the element comes into contact with an

object. Piezoelectric resonance sensors have been used as tactile sensors in

different fields of medical research [6, 34]. Measurements with resonance

sensors have been made to study the differences in stiffness and elasticity of

the skin in order to detect oedema and lesions [38, 39]. In other studies,

measurements have been made to detect lymph nodes containing metastases

[40] and stiffness of the liver that can indicate liver fibrosis [41]. Jalkanen et

al. [9, 42], has reported measurements on slices of resected prostate tissue ex

vivo, which demonstrates the possibility to differentiate between healthy

tissue and malignant tumours. A piezoelectric resonance sensor in an

applanation tonometer has also been used by Hallberg et al. [43] and Eklund

et al. [44], to measure the intraocular pressure of the eye, as an elevated eye

pressure can be an indicator for glaucoma.

4.4.2 Other types of tactile sensors

There are many studies reporting of different experimental devices for

measuring stiffness in soft tissue. Force measurements in order to detect stiff

nodules during large deformations have been reported by Kim et al [45],

showing that they could detect cancer lesions in human prostates at a depth

of 3.4 mm using an 3 mm indentation depth. They have also reported on

measurements for characterization of tissue stiffness using large

deformations on porcine liver [46]. Rolling indenter probes measuring

indentation forces to detect inclusions of stiff nodules have also been used on

silicone phantoms and porcine kidney tissue [47, 48]. Their results showed

that they could detect stiff nodules at a depth of 22 mm with an indentation

depth of 6 mm. One interesting type of tactile sensing system for very soft

human tissue (brain) has been reported by Tanaka et al [49]. The sensor had

a tip made of a latex balloon with 4 mm in diameter that could expand. The

balloon was connected via a surgical line through an endoscope to a syringe

with which the pressure was controlled by a water column and a pressure

sensor.

During minimal invasive surgery, MIS, different solutions to enhance the

tactile feedback to the surgeon have been reported. One common type of

9

technical solution seems to be force sensors added to graspers and other

instruments used during laparoscopic, and robot assisted surgery. Samur et

al [50] developed a robotic indenter to measure soft tissue in the abdominal

region during MIS. They report on successful measurements made on adult

pigs in vivo. Similar instruments were evaluated on bovine liver and porcine

lungs ex vivo with implanted nodules, and it was suggested that such

instruments could give the surgeons the haptic feeling that is often lost

during MIS [51]. A very different approach to the issue of lost haptic

information during MIS was reported by Beccani et al [52]. Their solution

was a wireless tactile sensor to be used to palpate for lumps in soft tissue

during laparoscopic surgery. Their prototype was 60 mm long with a

diameter of 15 mm and was held by a standard laparoscopic grasper,

operated by the surgeon. With this wireless device they could localize

tumours and distribute a stiffness map in real time.

4.5 Soft tissue silicone phantoms and biological tissue

Silicones of different stiffness have been found to be very suitable for

evaluating tactile sensor techniques regarding soft human tissue

characterization as soft tissue phantoms for preliminary research due to the

similarities regarding mechanical properties [42, 53, 55, I]. Silicone is an

elastomer which can be regarded as incompressible with a mechanical

strain-stress relationship that can be described using hyperelastic models

[55, 56]. The silicone can be cast in different geometries and it is possible to

embed small volumes of different stiffness to mimic stiffness variations due

to disease, i.e. cancer nodules. Silicones are also more stable over time,

compared to biological materials which make them appropriate to store for

repeated measurements at a later date.

Biological tissue is more difficult to use as models as its properties can

show changes due to environmental conditions, such as degrading or

dehydrating during the time of measurements. Its normal properties are

different when measurements are made ex vivo, compared to in vivo, due to

the physiological differences and the change in supporting tissue structures.

Indentation measurements on human liver show a difference in elastic

properties between normal and diseased tissue [57]. Mazza et al [58] have

reported of increased stiffness in human liver ex vivo, compared to in vivo.

The fact that ex vivo measurements often are made in laboratory

environment with well-defined boundaries and conditions, can also be an

advantage for the outcome of the measurements.

10

In contrast to silicone, biological tissue does not have a linear stress-strain

relationship [58]. Most biological soft tissues can be described as

hyperelastic, but they also have viscoelastic properties when they are subject

to deformation [36, 56]. Viscoelasticity is a combination of elasticity and

viscosity, and the ratio between these two can vary depending on the type of

biological tissue [59]. Viscosity is the resistance that a certain material has

against changing when subjected to a force as it resists shear and strain

when stress is applied. Elasticity is the property of a material to temporarily

change shape during the time when it is subjected to a force and retract

when the force is unloaded [60]. There are other properties that can

characterise viscoelastic materials, such as hysteresis which is the difference

in behaviour of loading and unloading when subjected to a force. The

mechanical behaviour of healthy and malignant prostate tissue has been

described with viscoelastic models by Phipps el al. [61].

11

5. Materials and methods

5.1 The piezoelectric resonance sensor

The resonance sensor used in this measurement system consists of a

piezoelectric element. A piezoelectric element changes its shape when an

electric field is applied. If this electric field is of a sinusoidal type, the

piezoelectric element will start to oscillate. The resonance sensor used was

divided into two different piezoelectric elements, a driving element and a

pick-up, which is separated by a non-conducting section, figure 2. The signal

from the pick-up section was constantly transferred to a feedback circuit

where a phase-shift circuit ensured that a zero phase condition between the

pick-up and the drive signal was established. This kept the resonance sensor

oscillating at its resonance frequency.

fo f

Healthy tissueCancerous tissue

The frequency shift f

f = f - f0

Pickup element

Driving element

Phase shift

circuit

PEEK tip

Non conducting

section

Force sensor

Sensor movement

Figure 2. A schematic illustration of the piezoelectric resonance sensor in a

measurement situation.

12

When the driving section of the resonance sensor comes into contact with an

object, a change in frequency will occur, f = f - f0 where f0 is the free

(unloaded) resonance frequency of the piezoelectric element, and f is the

loaded resonance frequency when in contact with the object. Contact with a

soft object that can be deformed will result in a f with a negative sign, and a

positive sign will be the result from contact with a hard object.

5.2 The measurement system

The tactile resonance sensor system (TRSS) was developed and used for

these studies. The TRSS was equipped with three motorized translation

stages that were controlled by a computer and a LabView® (National

Instruments, Austin, Texas, USA) program. Two translation stages were used

for horizontal movements, X and Y. A third translation stage for vertical

movements, Z, was mounted on a rotational stage to enable movements at

different contact angles, α, between the sensor and the measured object [38,

III], figure 3.

Stepper

motorHorizontal translation

stages, X and Y

Platform with flat

sample

Stepper motor

Vertical translation

stage, Z

Sensor head

Rotational stage

Electronic circuits, I/O

connector block, etc.

Computer with DAQ

Figure 3. A schematic side view of the basic components of the TRSS in position

for measurements on a flat sample. The movements of the translation stages are

controlled from a desk-top computer.

13

5.3 Technical setup

The three translations stages for the TRSS enabled 3-D movement’s

necessary for the flexibility of the instrument. The translation stages for

horizontal movements (M4424; Parker Hannifin Corporation, Daedal

Division, Irwin, PA, USA) were equipped (in house) with stepper motors

(17HD1008; Moon’s, Shanghai, P.R. China) and connected to the

micrometre screws of the translation stages. The motorized translation stage

(VT-21, Micos, Irvine, CA, USA) for the vertical movements was controlled

by a stage controller (Pollux Box; Micos GmbH, Eschbach, Germany). All

three motorized translation stages had a resolution of 2.5 μm. The rotational

translation stage (NT55-030; Edmund Optics, York UK) with a resolution of

0.5° made it possible to adjust the angle of the movement of the vertical

translation stage ± 90°.

The sensor in the TRSS used for measuring f was a piezoelectric

cylindrical element (Morgan Electro Ceramics, Bedford, Ohio, USA) made of

lead zirconate titanate (PZT). The element was 15 mm long with an outer

diameter of 5 mm and an inner diameter of 3 mm. A hemispherical tip with

the radius of 2.5 mm made of polyether-ether-ketone (PEEK) was glued at

the end of the sensor which came into contact with the measured object. The

PZT-element and a preloaded force sensor (PS-05KC; Kyowa, Tokyo, Japan)

were mounted inside a cylindrical aluminium casing, which will be referred

to as the sensor head, figure 4. The force sensor and the PZT-element were

calibrated before each study to get valid constants for the voltage-to-force,

and voltage-to-frequency conversions. The signals from the force sensor and

the PZT-element were collected by a data acquisition card (NI6036E;

National Instruments, Austin, Texas, USA), and transferred to a computer at

a sampling rate of 1 kHz.

The TRSS can be used for measurements on both flat and spherical

objects, which can be rotated in a special holder. Due to the, often uneven,

and spherical shape of the prostate and the importance of making

measurements perpendicular towards the surface, the sensor was fitted to a

rotational device which allows to be varied. The rotatable holder, figure 4A,

for spherical objects during measurements was equipped with a cantilever

strain gauge (CEA-06-240UZ-120; Measurements Group Inc, Raleigh, NC,

USA) for measuring the clamping force FC and a sensor for recording the

rotation angle, αr [III].

14

1

4

2

3

7

8

9

10A B5 6

Figure 4. The TRSS A) The sensor head B): 1) The vertical and the rotational

stages. 2) Sensor head. 3) Fixture for spherical samples. 4) Rotation angle sensor. 5)

Cantilever strain gauge. 6) Translation stages for horizontal movements. 7) The

platform for flat samples. 8) Force sensor inside. 9) PZT-element. 10) PEEK tip.

5.4 Preparation of soft tissue phantoms made of silicone

For the studies on soft tissue phantoms, different hardness and shapes were

used in [I – V]. The silicone used was a two-component type (Wacker SilGel

612; Wacker-Chemie GmbH, Germany), which has been used in earlier

studies [37, 43, 55, 62, III].

Table 1. The mixing ratios and hardness values for the different silicone mixes used

in papers I – V. The Shore-000 scale was used only in papers IV and V.

Hardness Mixing ratio A:B

Shore-000 value

Cone penetration value (mm x 10-1)

Used in paper

Softest 4:3.75 16 251 I 4:3.50 23,5 221 I 4:3.30 33 188 IV, V 4:3.20 39 163 I, II, III 4:3.00 51 126 I 4: 2.50 82 49 I, II, III

Hardest 4: 2.22 88 33 IV, V

15

All flat silicone samples were cast in standard Petri dishes (diameter 87

mm and height 13 mm). By letting the silicone cure on a horizontal surface

and in room temperature, the surfaces came out even and glossy. For the

purpose of evaluating Δf at different hardness, five homogeneous samples

with the height 13 mm were used.

To study at what depth into soft silicone the TRSS could detect an

embedded stiffer nodule, silicone samples with different height were cast, all

with a stiff hemisphere of silicone (diameter 8 mm, height 4 mm), and a

nodule centred at the bottom of the Petri dish, figure 5A. The same

combination of soft silicone, and hard silicone for the nodules were used in

both paper IV and V, see table 1. The distance from the top of the nodule to

the surface of the silicone, d, varied from 1 to 9 mm between the different

samples.

The TRSS had a rotatable holder to facilitate measurements on spherical

objects. The possibility to perform reliable measurements on spherical

shaped samples needed to be evaluated, as well as how nodules could be

detected under a curved surface. For this purpose soft spherical silicone

samples were cast, using the inside of a table tennis ball (diameter 40 mm)

as a mould. Small spheres of stiff silicone with different diameters (2.5, 4

and 6 mm) were placed inside the mould during the casting procedure,

figure 5B. The embedded small stiff spheres could then mimic cancer

tumours.

d

A B

d

Figure 5. Tissue models made of soft silicone with embedded stiff nodules

mimicking cancer tumours at the distance d from the surface. A) Flat sample in a

Petri dish. B) Spherical sample with nodules.

5.5 Preparation of biological tissue phantoms

The biological tissue phantoms for the flat samples were made of chicken

muscle tissue [IV]. Commercially available chicken breasts were sliced into

different thickness, 5 to 11 mm. Silicone nodules of the same type as for the

16

silicone samples mimicked cancer tumours were used. An incision of the

same size as a nodule was made on the bottom of each slice of chicken

muscle, figure 6A, to make room for the nodule without causing a bulge on

the surface of the slice. Finally the slice was pinned on a polystyrene foam

plate. The nodules were placed with d= 1 to 7 mm under the surface of the

slices.

For the spherical tissue phantoms, pieces of porcine tenderloin were used

instead of chicken muscles because of the volume needed to get the right size

of spherical shaped samples [V]. The pieces were cut approximately into a

spherical shape with a diameter D = 40 mm. All visible membranes were

carefully removed, and pieces with visible tendons were avoided to get as

homogeneous tissue samples as possible. A spherical nodule with the

diameter 6 mm of stiff silicone was inserted under the surface of the sample;

this was done from the side, not to disturb the measurement area, figure 6B.

The nodule was placed at d = 3 ± 0.5 mm under the surface.

d

B

d

A

Figure 6. Cross-section side views of two biological tissue samples with embedded

silicone nodules at the distance d from the surface. A) A flat sample made of

chicken muscle tissue. B) A spherical sample made of porcine muscle tissue.

5.6 Human prostate samples

For the studies in paper V, two prostate glands were obtained with written

consent from the patients, aged 70 and 72 years old. Both were undergoing a

radical prostatectomy at the University Hospital in Umeå. Ethical approval

was given by the Ethics Committee at Umeå University (Dnr 03-423). The

younger man’s prostate gland weighed 60.1 g and had the approximate

diameter D = 50 mm. The other prostate gland weighed 40.6 g and had the

approximate diameter D = 40 mm. Before measurements, the pathologist

dyed the surface of the prostate gland according to their standard, figure 7.

The dorsal aspects (backside) of the prostates were marked with yellow

17

colour, the left anterior (front) red and the right anterior sides were marked

green.

A B C

Figure 7. A resected prostate from a radical prostatectomy. A) Fresh from surgery.

B) The dorsal side of the prostate gland, dyed yellow. The apex is pointing down. C)

The anterior side of the prostate gland seen from the front, dyed green and red.

5.7 Measurement set-up

The setup of the TRSS used in the studies referred to in this thesis has been

presented recently [III - V, 63]. The measured parameters were the applied

force, F, when the sensor was indented into the measured object and the

change in the resonance frequency, f. Jalkanen et al. [35], has shown a

theoretical model that relates the measured stiffness trough F and f to the

Young’s elastic stiffness modulus, as described in 4.3. For evaluating the

stiffness measurements, a parameter called the stiffness parameter fF

was introduced. This is found as an important parameter as it is a function of

the indentation depths, IZ. In the studies presented here, the TRSS has been

evaluated regarding how the measured parameters F, f and fF are

affected by variations in contact angles, and indentation velocities, vi.

Each measurement lasted four seconds, during which the sensor was at a

standstill for two seconds after the pre-set indentation depth, Itot, was

reached, before retracting from the sample. With the pre-set indentation

velocity, i, and the sampling rate of 1 kHz, any indentation depth, IZ, could

be calculated, and hence data of interest could be retrieved from the logged

data at any depth IZ ≤ Itot. The stiffness parameter, fF , was derived by

linear regression for each IZ. All measurements were made at normal room

temperatures, 20 to 22 °C.

For the measurements made on the different spherical samples, the

necessary force, the clamping force FC, used to hold the sample in place

18

between the two concave discs, see figure 4A, were logged as well as the

rotational angle αr. To prevent the biological samples, that were wet and

slippery, from gliding in the rotatable sample holder, pieces of emery cloth

(P80, KK114F, VSM® Abrasives Limited, GB-Milton Keynes, UK ) were

glued to the inside of the concave discs.

5.7.1 Measurements on soft tissue silicone phantoms

The flat silicone samples were mainly used to evaluate the basic functions of

the TRSS, such as the ability to differentiate between silicones of different

hardness. The dependencies of the measured parameters on α and i were

studied using flat silicone samples with two different hardness, see table 1,

and with the height 13 mm. The angle of the sensor movement, α, was

adjusted with the rotationa stage. The deviation in α from the perpendicular

(α = 0°) towards the surface were chosen in steps of 1° from α = 0° to 11°,

and thereafter in steps of 2° up to α = 35°. Five different i were used (i = 1

to 5 mm/s) for that part of the study and Itot was set t0 1 mm [III]. For

papers [IV, V], the indentation velocity was i = 4 mm/s for all

measurements.

The nodules in the spherical silicone models were place under and near

the surface. They were tinted to be visible in the otherwise transparent

silicone. They were aimed as phantoms to mimic a prostate gland with

cancer tumours. The measurements on the spheres were made in different

measurement positions (MPs) with the sensor head movement in different

angles, αsm, but always perpendicular to the surface at the point of contact,

figure 8.

19

0°-10°-20°

10° 20°

O - 360°

αr

αsm

1

2

3

4αr

αsm

A B

Figure 8. A) Illustration of a silicone sample in the rotatable holder, illustrating αsm

and αr. B) A picture of a porcine sample in the rotatable holder. 1) Rotational stage

for αsm. 2) USB-microscope. 3) Cantilever strain gauge. 4) Angle sensor for αr.

The MPs were chosen directly on the nodules, αsm = 0° and at the positions

αsm = 10° and αsm = 20° axially to the left and right of the vertical. The

difference in angle (10°) between each MP was equivalent to a distance of 3.5

mm on the surface of the sphere [V]. For every row of measurement points in

the axial direction, the sphere was rotated approximately αr = 60°, but

making sure to find the correct position for the nodules that was placed at

approximately every αr = 120°. For the measurements on all silicone

samples, there were six measurements made at every MP. The clamping

force, FC, for the silicone spheres were approximately 500 ± 50 mN during

the stiffness measurements.

5.7.2 Measurements on biological tissue phantoms

The biological tissue models of chicken breast muscle were cut in a semi

defrosted state with an electric slicing machine to get even surfaces.

Measurements were made with a vertical sensor movement, αsm = 0°,: one

MP aimed over the nodule wiht another two to the left and two to the right of

the nodule with a distance of 5 mm in between them [IV].

Spherical tissue models were made of porcine muscle from tenderloin.

The measurements on the porcine tissue were made at αsm = 0°, and the

20

tissue was rotated in steps of αr = 5° within an interval of ± 40° from the

nodule. The distance between each MP was approximately 1.75 mm on the

surface of the samples. For all measurements on the biological tissue models,

there was only one measurement made in each point, i.e. several samples of

both chicken and porcine muscles were used [V]. The spherical shaped

porcine samples were subjected to FC = 425 ± 39 mN.The tissue samples

were kept moist by spraying them with saline solution (Sodium Chloride 9

mg/ml, Fresenius Kabi AG, DE-61346 Bad Homburg, Germany) about every

5 minutes, using an atomizer.

5.7.3 Measurements on excised human prostate tissue

For the measurements on the resected prostate glands, the rotatable holder

for spherical objects was used. The two prostates were different in weight

and size, resulting in a higher FC for Prostate 1 (FC = 760 ± 99 mN),

compared to Prostate 2 which was held in place by FC = 460 ± 85 mN. The

measurements were made with a sensor movement in the vertical direction,

αsm = 0°. A total of 36 to 40 measurement points were made as the prostate

specimens were rotated in steps of 10° between each measurement, i.e. only

one measurement in every MP. By doing so, all MP were in a straight line

around the circumference of the prostate gland. Points along this line were

marked with black dye. The reason for this was to guide the pathologist when

cutting the slices and to ensure that the correct slice was used for

comparison with the measurements. The prostate was cut in the horizontal

plane in o.5 cm thick slices and were dehydrated before being embedded in

paraffin. The slices were then cut in thin slices (5 μm) before they were

stained with haematoxylin-eosin and examined by light microscope.

5.8 Maximum estimated errors and overall accuracy

The estimated errors, ε, for the measurement parameters in each study has

been reported [III, IV, V], taking into consideration the specific settings for

each measurement situation. For the indentation depth, the IZ varies

between the different papers, but a general ε for the different IZ was 2 – 4%.

The errors for Δf, F and fF were all < 1.5%. The strain gauge for the

clamping force had an estimated error of 10% due to mechanical friction.

The rotational stage for measuring αsm and the angle sensor for αr both had a

ε = 0.5°. For measurements made on uneven surfaces, for example spherical

21

biological tissue samples and the prostate gland, photographs were taken

during the contact phase which was later studied for incorrect α.

5.9 Statistics

In paper III, the statistical tests were done with a one-tailed t-test to test if

the measured parameters F, Δf and fF deviated from their respective

values at α = 0° with the null hypothesis that the average relative deviation

with respect to the measurements with α = 0° was equal to 10% against the

alternative hypothesis that it was ≤ 10%. Wilcoxon’s signed rank test was

used for the silicone discs in paper IV, and a non-parametric Kruskal-Wallis

method followed by a multiple comparison test for the differences between

the different slices of chicken muscle. In paper V, the statistical tests for the

measurements on silicone were performed with one-way ANOVA and

Tamhane’s post hoc comparison test. The differences between groups of data

from the prostate measurements were done with a Mann – Whitney –

Wilcoxon test.

The software used for the tests were IBM SPSS Statistics 21, and

Microsoft Excel 2010. The measured values were presented as mean ±

standard deviation (SD) and p-values of ≤ 0.05 were considered significant

for all statistical tests.

22

23

6. Results and discussion

The TRSS has been evaluated systematically step-by-step with good results

regarding performance and the different measurement parameters.

However, it must be clarified that the absolute values of the measurements

most likely is unique for this TRSS. The size and geometry of the sensor tip,

the free (unloaded) resonance frequency and the size and type of the PZT-

element are only a few important components that interact uniquely for this

TRSS.

The stiffness measurements on the chicken and porcine muscle tissues,

[IV, V] were only performed once in each MP, as soft biological tissue ex vivo

can be permanently deformed after indentations, and repeated

measurements in the same MP would be affected by this [64]. During the

first measurements on the flat silicone models in Petri dishes, it was noticed

that there was a 0.7 mm protruding rim on the underside. That resulted in

that the bottom could give after during indentation due to the applied force

F. The rim had to be removed to assure full contact between the Petri dish

and the measurement platform. The conditions surrounding the nodules in

the flat silicone models and the spherical samples were different. In fle flat

models, the nodules were supported from underneath, as the y were placed

on the rigid measurement platform. The nodules in the spherical silicone

samples were surrounded by soft silicone in all directions.

6.1 Instrument performance and accuracy

The TRSS was evaluated on five flat silicone samples regarding the ability to

distinguish between different hardness, see table 1 [I]. Figure 9 shows the

linear relationship between Δf and the cone penetration values which is a

method for measuring hardness in soft materials according to DIN ISO

standard 2137.

24

Figure 9. Measurements of Δf (mean ± SD, n = 10) at constant F = 40 mN on five

silicone samples with different hardness. For most Δf, the standard deviations are too

small to be visible as error bars.

The results from figure 8 show a linear correlation between Δf and the cone

penetration values, (p < 0.05, R2 = 0.99). The standard deviations are small,

about 1 to 3 % of the measured values. The higher values for the softer

silicone samples are in agreement with previous studies on the same type of

silicone [42]. The results show that the TRSS could differentiate between

materials with different hardness, which was important for the stiffness

measurements that lay ahead.

6.2 Angular dependency

Two silicone samples with different hardness, see table 1, and two different

indetation depths were used for evaluating the angular dependency of Δf, F

and fF on different α [III]. The dependency of F on α showed a weak

dependency, with a relative deviation due to α, figure 10. It was on average ≤

10% for α up to about 10° (one-tailed t-test, p < 0.05). However, the stiffer

silicone showed stable values of F with respect to α up to about 20° [III]. The

25

dependency of Δf on α showed to be very similar to the dependency of

fF on α, figure 10, with a relatve deviation due to α that on average

was ≤ 10 % for angles up to 10° (one-tailed t-test, p < 0.05). For all other α,

the angular dependency of fF were significant (p > 0.05) [III].

Figure 10. Measurements made on silicone samples with two different hardness at

IZ = 0.20 mm and 0.38 mm for α between 0° and 35°. The figure shows the relative

deviation of fF to fF (α=0°), (mean ± SD, n = 6). The vertical line was

drawn as a guide for the eye.

Measurements on both flat and spherical shaped tissue samples were

especially prone to get the contact angle α ≠ 0° due to the uneven surfaces.

By studying the photographs that were taken at the time of each

measurement, such deviations were possible to identify. Other studies [10,

65] support the findings in paper [III] that the precision in Δf decreases as α

increase, due to the change in contact area between the sensor and the

measured sample. For variations in α ≤ 10°, the measured Δf and F and thus

the fF were not significantly affected [III].

26

6.3 Velocity dependency

The velocity, with which the sensor moved during indentation, could be set

through the LabView® program. In the first papers [I, II], the standard

setting of i, = 5 mm/s was used. In paper III the possible effect that

different i could have on the measured parameters Δf, F and fF were

evaluated. The evaluation of fF as a function of i was made on flat

silicone samples, two different hardness, and α = 0°, figure 11.

Figure 11. Measurements of fF (mean ± SD, n = 10) on flat silicone samples

with two different hardnessfor different i, and at two different IZ.

The results showed that there was a dependency of fF on i, and that it

was increasing with higher i. The relative difference in fF for the

increase in i was more evident for the softer silicone sample compared to

the stiffer. The evaluation also showed results that indicated a lower

dependency for F on i, compared to the dependency for Δf. However, a low

i showed higher accuracy in Itot and Iz since the first sampled data point

showing contact with the surface was more distinct. For the measurements

in [IV, V] the indentation velocity used were chosen to i = 4 mm/s. The

reason was that the silicone samples used were of the softer kind to better

27

mimic biological tissue, and at the same time minimize some of the viscous

effects in the ex vivo biological tissues.

6.4 Depth sensitivity

The ability of the TRSS to detect stiff volumes embedded under a layer of

softer silicone or biological tissue was evaluated on flat silicone samples of

different thickness with embedded stiff hemisphere of silicone [IV]. Figure

12 shows the measurements of fF on ten different silicone samples.

Nine with with embedded hemispheres with the distance from the top of the

hemisphere to the surface of the soft silicone ranging from d = 1 t0 9 mm.

Figure 12. The fF (mean ± SD, n = 6) as a function of d with the sensor

centred above the embedded stiff silicone hemisphere, IZ = 0.4 and 0.8 mm is

shown.

28

The statistical tests of significant differences between fF measured at

the position of the hemisphere, and the average fF of the surrounding

soft silicone were performed using Wilcoxon signed rank test. From figure 11

and the statistical analysis, it was evident that the stiff hemisphere could be

detected at d ≤ 4 mm (p < 0.05). However, the statistical analysis also

showed that the hemisphere, with some exceptions, could be detected with a

weak significance at d = 5 to 9 mm (p < 0.05).

The same type of silicone hemispheres was used for the biological tissue

samples, chicken breast muscles, where a small incision was made on the

underside to make room for the hemisphere. The thicknesses of the

defrosted slices of chicken muscle were, as mentioned earlier, cut in a semi

defrosted state to get an even surface and equal thicknesses. However, the

thawed slices turned out slightly unequal in thickness so for that reason, the

slices were divided into three groups depending on their thickness. This

resulted in six slices for each group with d = 1.5 ± o.5 mm, d = 3.5 ± o.5 mm

and d = 6.5 ± o.5 mm when the nodules were in place. The measured

fF from a point centred over the hemisphere were significantly

different (p < 0.05) compared to the measurements of the surrounding

tissue.

In general, the depth sensitivity for this TRSS agrees with the estimated

sensing depth presented by Jalkanen et al [66]. In that study, the depth

sensitivity was extrapolated from a measured stiffness parameter and the

histological data of the prostate tissue and the results showed an estimated

sensing depth of 3.5 to 5.5 mm for IZ = 1 mm.

6.5 The clamping force

The spherical shaped samples were all subjected to an axially clamping force,

FC when mounted in the rotatable holder [V]. The effect on fF caused

by FC was evaluated with forces up to five or six times higher compared to

what was used later during the other stiffness measurements on spherical

samples. The evaluation on silicone samples showed that fF did not

depend on FC (p > 0.05). The fF for porcine muscle tissue showed a

weak but significant dependence (p < 0.05) on FC. The fF increased

with FC, most noticeable at FC > 800 mN. For this reason, FC was chosen to

be in the range of 500 ± 50 mN for the further studies. For Prostate 1, FC

was higher (FC = 760 ± 99 mN), due to its size and weight.

29

6.6 Measurements on prostate gland ex vivo

In previous works [9, 35, 42, 53, 66, 67], the stiffness measurements were

performed on slices of human prostate tissue with PCa, whereas in this study

the measurements were made on the whole human prostate gland including

the capsule. The previous measurements on other spherical shaped samples

with the rotatable sample holder showed to have been important as the time

frame for measurements on human prostates were very narrow so the

measurements on prostates was quite stressful. The measurements on the

two resected prostate glands took place about 1 hour after that they had been

surgically removed. The prostate glands were mounted in the rotatable

sample holder of the TRSS with the urethra axially. This was done to enable

measurements in a line around the prostate. For the measurements, the

prostate was rotated in steps of approximately 10° between each

measurement, as there was only enough time for a limited number of MPs

due to the short exsess time ( which was 1 hr including transport between the

location of the TRSS and the pathology department).

After the measurements with the TRSS, the prostates were transported

back to the pathologist for routine processing and examination. A tissue slice

of Prostate 1 embedded in paraffin, stained at the periphery in order to locate

the left anterior (red), the right anterior (green) and the dorsal (yellow) parts

of the prostate is shown in figure 13AB. Due to the large size of that

particular prostate gland, the slice had to be cut in a half in this phase, a left

and a right part. Note that they are placed correctly next to each other in

figure 12, even though the mid-section does not look as a straight cut, which

maybe could be expected.

30

A B C D(2) (3)(1)

Figure 13. Prostate 1 in a paraffin block is shown in figure A) and B). At the

periphery, the colours from the staining are visible. Arrow (1) points at punch holes

in the tissue slice where tissue specimens have been taken out after the stiffness

measurements. C) and D) Photomicrographs of the corresponding tissue slice.

Arrows (2) and (3) are aimed at small areas in both the left and right dorsal parts of

the prostate with cancer tumours.

The measurements on Prostate 2 were performed in the same way as for

Prostate 1. A large area of the anterior part of the prostate was marked as

cancer tumours, on the left anterior a Gleason score of 7 (3+4) was noticed,

figure 14.

A B(1) (2)

Figure 14. Prostate 2 is shown in A) and B). A) The tissue slice embedded in

paraffin. B) The photomicrograph of the corresponding tissue slice. Arrow (1) is

pointing at the Gleason staged cancer tumour, and arrow (2) is aimed at the anterior

area containing cancer tumours.

31

Figure 15 shows the results from the stiffness measurements on both

prostates. The areas with higher stiffness parameter values, fF , were

mainly located in the anterior parts of the prostate, except for Prostate 1

where higher values were noticed in the dorsal aspect as well. The anterior

parts consist of mainly smooth muscle and connective tissue, which usually

is stiffer. The dorsal aspect consists more of glandular tissue which is soft.

The dorsal aspect of Prostate 1 had small multiple areas of tumour tissue

encircled in figure 13CD which correspond to the stiffness measurements

and was found ti be significantly different (p < 0.05) to the rest of the

measuremnts on the dorsal side of theprostate [V]. On Prostate 2, figure 14B,

the dorsal aspect showed low stiffness values, but a large area in the anterior

were marked for presence of tumours by the pathologist.

Figure 15. The measurements on prostate 1 (P1) and prostate 2 (P2) are shown.

fF was measured in a line around the prostates,along the great circle,

perpendicular to the urethra. The colours in the figure are the same as how the

pathologist dyed the different areas of the prostate glands.

32

33

7. General Summary and Conclusion

7.1 Conclusion

In this doctoral thesis, the aim was to design and evaluate a flexible tactile

resonance sensor system for the possibility to detect cancer in soft human

tissue with a focus on prostate cancer.

The work started with measurements on flat samples, soft tissue models

made of silicone to evaluate the errors and overall accuracy of the measured

parameters Δf, F and fF which were estimated to be < 1.5% of the

measured values. The effect of a contact angle deviating from a

perpendicular indentation direction and the impact of different indentation

velocity on the measured parameters were evaluated. Contact angles

deviating < 10° from the perpendicular to the surface of the measured object

did not significantly affecte the measurement parameters. Flat silicone and

chicken muscle tissue samples with embedded stiffer volumes (nodules)

were used to investigate the ability to detect stiffer volumes at different

depth under the surface of the tissue phantoms. The results showed that

embedded stiff nodules with certainty could be detected down to a depth of ≤

4 mm under the surface of the measured object, with a small indentation

depth of 0.4 to 0.8 mm into the surface.

Spherical shaped soft tissue phantoms made of silicone and porcine

muscle tissue were used to evaluate the measurement procedures when

using the rotatable holder for spherical objects. Measurements have been

made on resected human prostates and after comparing the measurements

with the histopathological evaluation, it was shown that the TRSS stiffness

parameter could distinguish areas with cancer from normal prostate tissue

(p < 0.05).

In conclusion the work presented here involving the TRSS has shown that

the resonance sensor technique has the potential to contribute to the future

development of new techniques and routines for diagnosis and detection of

prostate cancer tumours, for example during open surgery procedures and

thus become an important clinical application in the future.

34

35

8. Acknowledgements

I would like to express my sincere gratitude to the persons who have

supported me in this work.

First of all, I would like to thank my supervisor Dr. Britt Andersson for

introducing me to the world of a PhD student. Her encouragement and

willingness to discuss problems that I have come across on the way have

been invaluable. Her positive attitude and ability to see things from a

different angle have been of great importance, as has also her help by being a

co-author on my papers.

I also want to thank Professor Olof Lindahl and Dr. Ville Jalkanen in their

role of co-supervisors. They have always been helpful and willing to help me

with obstacles that I have stumbled on. They have also been very important

co-authors on my papers.

Furthermore, I would like to express my gratitude to Sven Elmå, who helped

with the manufacturing of many mechanical components as well as Johan

Pålsson, who helped with the software LabView®, and Lars Karlsson who

helped with the electronics. Morgan Nyberg at Department of Engineering

Sciences and Mathematics, Luleå University of Technology, has assisted with

the measurements and logistics for the measurements on the excised

prostate gland.

I wish to express my gratitude to Kerstin Almroth, Research Assistant at the

Department of Surgical and Perioperative Sciences, Urology and Andrology

at Umeå University, Wanzhong Wang, Pathologist, Pernilla Andersson,

Research Assistant, Birgitta Ekblom, Research Assistant and Susanne

Gidlund, Research Assistant, at the Department of Clinical Pathology at

Umeå University Hospital.

The study was supported by The Industrial Doctoral School at Umeå

University, BioResonator AB, and by grants from Objective 2 North Sweden-

EU Structural Fund.

And last but not least, I would like to thank family and friends who have

supported me through this time.

36

37

9. References

1. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, Rosso S, Coebergh

JWW, Comber H, Forman D and Bray F 2013 Cancer incidence and

mortality patterns in Europe: Estimates for 40 countries in 2012

European Journal of Cancer 49 1374-1403

2. American Cancer Society (21/01/2014) at http://www.cancer.org

3. The Swedish National Board of Health and Welfare (Socialstyrelsen)

(21/01/2014) at http://www.socialstyrelsen.se

4. Schröder FH, Hugosson J, Roobol MJ, Tammela TLJ, Ciatto S,

Nelen V, Kviatkowski M, Lujan M, Lilja H, Zappa M, Denis LJ,

Recker F, Berenguer A, Määttänen L, Bangma CH, Aus G, Villers A,

Rebillard X, van der Kwast T, Blijenberg BG, Moss SM, de Koning

HJ, Auvinen A, and E.R.S.P.C. Investigators 2009 Screening and

prostate-cancer mortality in randomized European study N Engl J

Med. vol. 360 no. 13 1320-1328

5. Candefjord S, Ramser K, Lindahl OA 2009 Technologies for

localization and diagnosis of prostate cancer J. Med. Eng. Tech.

33 585-683

6. Lindahl OA, Constantinou CE, Eklund A, Murayama Y, Hallberg P

and Omata S 2009 Tactile resonance sensors in medicine J. Med.

Eng. Tech. 33 263-273

7. Gwilliam JC, Pezzementi Z, Jantho E, Okamura AM, Hsiao S 2010

Human vs. robotic tactile sensing: detecting lumps in soft tissue

IEEE Haptics Symposium 2010 (Waltham, MA) 21-28. ISBN 978-1-

4244-6821-8

8. Tiwana MI, Redmond SJ, Lovell NH 2012 A review of tactile sensing

technologies with applications in biomedical engineering Sensors

and Actuators A179 17-31

38

9. Jalkanen V, Andersson BM, Bergh A, Ljungberg B and Lindahl OA

2006 Resonance sensor measurements of stiffness variations in

prostate tissue in vitro – a weighted tissue proportion model

Physiol. Meas. 27 1373-86

10. Sakai N, Tatsuta M, Yano H, Iishi H, Ishiguro S 2000 Diagnosis of

the extent of gastric cancers by a new endoscopic ultrasonic tactile

sensor Gastrointest Endosc 51 69-73

11. Chung C-C. Hsieh C-C. Lee H-C. Wu M-H. Huang M-H. Hsu W-H

and Hsu H-S. 2010 Accuracy of helical computed tomography in the

detection of pulmonary colorectal metastases. The Journal of

Thoracic and Cardiovascular Surgery 141 no. 5 1207-1212.

12. Cerfolio RJ. and Bryant AS. 2008. Is palpation of the non-resected

pulmonary lobe(s) required for patients with non-small cell lung

cancer? A prospective study. The Journal of Thoracic and

Cardiovascular Surgery 135 no. 2 261-268.

13. Cerfolio RJ. McCarty T. and Bryant AS. 2009. Non-imaged

pulmonary nodules discovered during thoracotomy for

metastasectomy by lung palpation. European Journal of Cardio-

Thoracic Surgery 35 786-791

14. Detterbeck FC. and Egan TM. 2004. Thorascopy using a substernal

handport for palpation. Ann. Thorac. Surg. 78 1031-1036.

15. Ahn B-M, Kim J, Ian L, Rha K-H and Kim H-J 2010 Mechanical

property characterization of prostate cancer using minimaly

motorized indenter in an ex vivo indentation experiment Urology

vol 76 no 4 1007-1011

16. Paulet T. and Fryer G. 2009. Inter-examiner reliability of palpation

for tissue texture abnormality in the thoracic paraspinal region.

International Journal of Osteopathic medicine 12 92-96

17. Martini F 2004 Fundamentals of Anatomy & Physiology. 6th ed.

Pearson Benjamin Cummings, New York

18. Tortora G, Grabowski SR 1996 Principles of Anatomy and

Physilology. 8th ed. Harper Collins College Publishers, New York

39

19. Taichman RS, Cooper C, Keller ET, Pienta KJ, Taichman NS and

McCauley LK 2002 Use of the Stromal Cell-derived Factor-1/CXCR4

Pathway in Prostate cancer Metastasis to Bone Cancer Res 62 1832-

1837

20. Bayar DP, Mostofi FK and the veterans Administration Cooperative

Urological research Group 1972 Carcinoma of the prostate:

prognostig evaluation of certain pathological features in208 radical

prostatectomies Cancer 30 5-13

21. Krouskop TA, Wheeler TM, Kallel F, Garra BS and Hall T 1998

Elastic moduli of breast and prostate tissues under compression

Ultrason imaging 20 260-274

22. Zhang M, Nigwekar P, Castaned B, Hoyt K, Joseph JV, di SAnt

Agnese A, Messing EM, Strang JG, Rubens D and Parker KJ 2008

Quantative characterization of viscoelastic properties of human

prostate correlated with histology Ultrasound in Med & Biol vol 34

no 7 1033-1042

23. Hricak H. Choyke PL. Eberhardt SC. Leibel SA. And Scardino PT.

2007 Imaging prostate cancer: A multidicplinary perspective

Radiology vol. 243 no. 1 28-53

24. Feleppa EJ. Porter CR. Ketterling J. Lee P. Dasgupta S. Urban S and

Kalisz A. 2004 Recent developments in tissue-type imaging (TTI) for

planning and monitoring treatment of prostate cancer Ultrasonic

Imaging 26 163-172

25. Carey BM. 2005 Imaging for prostate cancer Clinical Oncology

17 553-559

26. Ravizzini G. Turkbey B. Kurdziel K and Choyke PL. 2009 New

horizons in prostate cancer imaging European Journal of Radiology

70 212-226

27. Taoka T. Mayr NA. Lee HJ. Yuh WTC. Simonson TM. Rezai K. and

Berbaum KS. 2001 Factors influencing visiulalization of vertebral

metatstases on MR imaging versus bone scintigraphy. AJR 176

1525-1530

40

28. Raz O. haider M. Trachtenberg J. Leibovici D. and Lawrentschuk N.

2010 MRI for men undergoing active surveillance or with rising PSA

and negative biopsies. Nature reviews, Urology 7 543-550

29. Qayyum A. Coakley FV. Lu Y. Olpin JD. Wu L. Yeh BM. Carroll PR.

And Kurhanewicz J. 2004 Organ-confined prostate cancer: Effect of

Prior transrectal biopsy on endorectal MRI and MR spectroscopic

imaging American Journal of Roentgenology 183 1079- 1083

30. Ikonen S. Kivisaari L. Vehmas T. Tervahartiala P. Salo JO. Taari K.

and Rannikko S. 2001 Optimal timing of post-biopsy MR imaging of

the prostate Acta Radiologica 42 70-73

31. Fine SW, Amin MB, Berney DM, Bjartell A, Egevad L, Epstein JI,

Humphrey PA, Mag-Galluzzi C, Montironi R and Steif C 2012 A

contemporary update on pathology reporting for prostate cancer:

Biopsy and radical prostatectomy specimens European Urology 62

20-39

32. Delahunt B, Miller RJ, Srigley JR, Evans AJ and Samaratunga H

2012 Gleason grading: past, present and future Histopathology 60

75-85

33. Srigley, JR, Humphrey PA, Amin MB, Chang SS, Egevad L, Epstein

JI, Grignon DJ, McKiernan JM, Montironi R, Renshaw AA, Reuter

VE and Wheeler TM 2009 Protocol for the examination of

specimens from patients with carcinoma of the prostate gland Arch

pathol Lab Med 133 1568-1576

34. Omata S, Terunuma Y 1992 New tactile sensor like the human hand

and its applications Sens Actuat. 35 9-15

35. Jalkanen V, Andersson BM, Bergh A, Ljungberg B and Lindahl OA

2008 Explanatory models for a tactile resonance sensor system –

elastic and density-related variations of prostate tissue in vitro

Physiol. Meas. 29 729-45

36. Fung YC 1993 Biomechanics – Mechanical properties of living

tissues. New York, Springer-Verlag

37. Gladwell GML and Kleesattel C 1968 The contact – impedance

meter-2 Ultrasonics 6 244-252

41

38. Lindahl OA, Ängquist K-A and Ödman S 1991 Impression technique

for the assessment of oedema – Technical improvement and

methodological evaluation of a new technique Med. & Biol. Eng. &

Comput. 29 591-7

39. Sasai S, Zhen Y-X, Suetake T, Tanita Y, Omata S and Tagami H 1999

Palpation of the skin with a robot finger: an attempt to measure skin

stiffness with a probe loaded with a newly developed tactile vibration

sensor and displacement sensor Skin Res. Tech. 5 237-46

40. Miyaji K, Furuse A, Nakajima J, Kohno T, Ohtsuka, T, Yagyu k, Oka

T and Omata S 1997 The stiffness of lymph nodes containing lung

carcinoma metastases – a new diagnostic parameter measured by a

tactile sensor Cancer 80 1920-25

41. Kusaka K, Harihara Y, Torzilli G, Kubota K, Takayama T, Makuuchi

M, Mori M and Omata S 2000 Objective evaluation of liver

consistency to estimate hepatic fibrosis and functional reserve for

hepatectomy J. Am. Coll. Surg. 191 47-53

42. Jalkanen V, Andersson BM, Bergh A, Ljungberg B and Lindahl OA

2006 Prostate tissue stiffness as measured with a resonance sensor

system: a study on silicone and human prostate tissue in vitro Med.

Biol. Eng. Comput. 44 593-603

43. Hallberg P, Lindén C, Lindahl OA, Bäcklund T and Eklund A 2004

Applanation Resonance Tonometry for Intraocular Pressure in

Humans Physiol. Meas. 25 1053-1065

44. Eklund A, Hallberg P, Lindén C and Lindahl OA 2003 An

Applanation Resonator Sensor for Measuring Intraocular Pressure

Using Combined Continuous Force and Area Measurement Invest.

Ophtalmol. Vis. Sci. 44 no. 7 3017-3024

45. Kim Y. Ahn B. Lee JW. Rha KH and Kim J. 2013 Local property

characterization of prostate glands using inhomogeneous modelling

based on tumor volume and location analysis Med. Biol. Eng.

Comput 51 197-205

42

46. Ahn B. and Kim J. 2010 Measurement and characterization of soft

tissue behaviour with surface deformation and force response under

large deformations Medical Image Analysis 14 138-148

47. Liu H. Sangpradit K. Li M. Dasgupta P. Althoefer K. and Seneviratne

LD. 2013 Inverse finite-element modeling for tissue parameter

identification using a rolling indentation probe Med. Biol. Eng

Comput 52 17-28

48. Sangpradit K. Liu H. Dasgupta P. Althoefer K. and Seneviratne LD.

2011 Finite-element modeling of soft tissue rolling indentation IEEE

Transactions on Biomedical engineering vol. 58 no. 12 3319-3327

49. Tanaka Y. Yu Q. Duomoto K. Sano A. Hayashi Y. Fujii M. Kajita Y.

Mizuno M. Wakabayashi T. and Fujimoto H. 2010 development of a

real time tactile sensing system for brain tumor diagnosis Int J

CARS 5 359-367

50. Samur E. Sedef M. Basdogan C. Avtan L and Duzun O. 2007 A

robotic indenter for minimally invasive measurement and

charaterization of soft tissue response Medical image Analysis

11 361-373

51. Perri MT. Trejos AL. Naish MD. Patel RV. and Malthaner RA. 2010

New tactile sensing sytem for minimally invasive surgical tumour

localization Int J Med Robotics Comput Assist Surg 6 211-220

52. Beccani M. Di Natali C. Sliker LJ. Schoen JA. Rentschler ME.

Valdastri P. 2014 Wireless tissue palpation for inoperative detection

of lumps in the soft tissue IEEE Transactions on Biomedical

engineering vol. 61 no. 2 353-361

53. Lindberg P, Andersson BM, Bergh A, Ljungberg B and Lindahl OA

2006 Prostate cancer detection with an improved resonance sensor

system: parameter evaluation in a silicone model and on human

prostate tissue in vitro Med Bio Eng Comput. 44 1053-59

54. Eklund a, Bergh A, Lindahl OA 1999 A catheter tactile sensor for

measuring hardness of soft tissue: measurement in a silicone modela

and in vitro human prostate model Med. Biol. Eng. Comput. 37 618-

624

43

55. Bergström JS and Boyce MC 1998 Constitutive modelling of the

large strain time-dependent behaviour of elastomers J Mech Phys

Solids 46 931-954

56. Martins PALS, Natal Jorge RM and Ferreira AJM 2006 A

comparative study of several material models for prediction of

hyperelastic properties: Application to silicone-rubber and soft

tissue Strain 42 135-147

57. Carter FJ. Frank TG. Davies PJ. McLean D. and Cuscheri A 2001

Measurements and modelling of compliance of human and porcine

organs. Medical Image Analysis 5 231-236

58. Mazza E. Nava A. Hahnloser D. Jochum W and Bajka M. 2007 The

mechanical response of human liver and its relation to histology: An

in vivo study Medical Inage Analysis 11 663-672

59. McKee CT, Last JA, Russell P, and Murphy CJ 2011 Indentation

versus tensile measurements of Young’s modulus for soft biological

tissues Tissue Engineering, part B vol 17 no 3 155-164

60. Sedef M. Samur E and Basdogan C. 2006 Real-time finite-element

simulation of linear viscoelastic tissue behaviour based on

experimental data Virtual and Augmented Reality Supported

Simulators , IEEE Computer Society nov/dec 58-68

61. Phipps S, Yang THJ, Habib FK, Reuben RL and McNeill SA 2005

Measurements of tissue mechanical characteristics to distinguish

between benign and malignant prostatic tissue Urology 66 447-50

62. Lindahl OA and omata S 1995 impression technique for the

assessment of oedema: comparison with a new tactile sensor that

measures physical properties of tissue Med. Biol. Eng. Comput.

33 27-32

63. Åstrand AP 2012 A Flexible Resonance Sensor Sytem for Detection

of Cancer Tissue – Evaluation on Silicone. Lic Thesis, Chapter 4, 5-8.

Umeå University, Sweden ISBN 978-91-7459-477-5

64. El Sayed T. Mota A. Fraternali F. and Ortiz M. 2008 A variational

constitutive model for soft biological tissues Journal of

Biomechanics 41 1458-1466

44

65. Murayama Y 2005 Development of a micro tectile sensor and its

biomedical applications. PhD thesis, chapter 5. Nihon University,

Japan.

66. Jalkanen V, Andersson BM, Bergh A, Ljungberg B and Lindahl OA

2007 Spatial variations in prostate tissue histology measured by a

tactile resonance sensor. Physiol. Meas. Vol 28 no.10 1267-1281

67. Jakanen V, Andersson BM, Bergh A, Ljungberg B and Lindahl OA

2013 Indentation loading response of a resonance sensor –

discriminating prostate cancer and normal tissue. J Med Eng

Technol vol. 37 no. 7 416-423

45

Summary of papers

Paper I

A flexible measurement system using piezoelectric resonance sensor

technology developed for measurements on both flat and spherical objects is

described. The newly developed measurement system can distinguish

between five flat silicone discs with different stiffness through measurements

of the frequency shift and the applied force. All measurements were made

perpendicular to the surface of the silicone disc and the indentation velocity

of the sensor was 5 mm/s.

The evaluation of the measurement system showed that the measured

frequency shift from the measurements on the different silicone mixtures

was in compliance with the given hardness measurements provided by the

silicone manufacturer and according to the DIN ISO standard 2137. A

linearly correlation with the cone penetration depth used in the standardwas

found (R2 = 0.99, p < 0.05). The paper reports that the measurement system

could distinguish between silicones of different stiffness with repeatability.

The function of the measurement system is further evaluated on spherical

objects and presented in Paper II.

Paper II

The measurement system described in Paper I is further evaluated, but in

this study the measurements were made on spherical objects. Spherical soft

tissue models with a diameter of 40 mm were made of a semi-soft silicone.

Initially, measurements were made with the sensor movement vertical

towards each measuring point along the axis of rotation, disregarding that

the contact angle varied for each point. Secondly, the measurements were

made ensuring that the contact angle was kept perpendicular to the surface

of the same measuring points as in the first measurement series. All

measurements were made with the indentation velocity of 5 mm/s. The two

measurement series were compared. A systematical deviation in the

frequency shift and applied force were noticed between the vertical

measurements compared to the measurements made perpendicular to the

surface of the silicone. The measurements with a contact angle that deviated

from the perpendicular were underestimated. However the applied force

seemed to be less sensitive for the deviating contact angles. The conclusion

was that the measurement system could be used for measurements on

46

spherical objects, and the importance of making measurements with the

sensor movement perpendicular to the surface of the measured object was

pointed out. Further evaluation of the angular dependency of the frequency

shift and applied force is reported in Paper III.

Paper III

Further measurements were made on flat soft tissue silicone models. In this

study, to evaluate the dependence of the frequency shift, applied force and

the stiffness parameter on the contact angle, the indentation velocity for

different impression (indentation) depths. The overall error of the resonance

sensor measurement system was also determined. In this study, two soft

tissue silicone models of different stiffness (hardness) were used.

Measurements were made with the indentation velocities ranging from 1

mm/s to 5 mm/s, and contact angles varying up to 35° from a line

perpendicular towards the surface (0°).

The results showed a velocity dependence for the stiffness parameter, as a

higher contrast between the silicones with different stiffness at higher

indentation velocity was noticed. The three measured parameters: frequency

shift, applied force and the stiffness parameter, all showed to be dependent

on the contact angle. One conclusion from the study was that deviations in

the contact angle from the perpendicular can be accepted up to ± 10°. Larger

deviation in the contact angle resulted in increased errors which lead to an

underestimation of the frequency shift. This could give a wrong indication

for cancer.

The maximum errors for the measured parameters were generally small

compared to the magnitude of the measured values, except for the relative

error in the impression (indentation) depth, which was 12.5% when the

highest indentation velocity was used. Continued work and measurements

on biological tissue are reported in Paper IV.

Paper IV

The focus in this study was to evaluate at what depth a stiff nodule of silicone

could be detected in flat discs of silicone and in biological tissue. A

hemisphere of silicone with a diameter of 8 mm was embedded at different

depths ranging from 1 to 9 mm in a surrounding softer silicone. The same

type of hemispheres were used for measurements on chicken muscle tissue

that were used as phantoms for biological tissue. The measurements were

made perpendicular to the surface of the measured objects and at a

indentation velocity of 4 mm/s. For depths of up to 4 mm of soft silicone

47

covering the stiff silicone hemisphere, the response of fF clearly

indicated the presence of the nodule. For depths exceeding 5 mm the

nodules could be detected using statistical analysis that showed that also

stiffer nodules at depths from 5 to 9 mm could be distinguished with

significance. The measurements on the biological tissue showed that it was

clear that the sensor could detect the stiff silicone hemisphere up to a depth

of 4 mm (p < 0.05, n = 24).

The results are promising for further studies for detection of prostate

cancer. In Paper V, the work is continued on spherical phantoms made of

silicone and biological tissue mimicking a prostate gland.

Paper V

In this paper, the tactile resonance sensor system (TRSS) was evaluated for

measurements on spherical objects, using the rotatable sample holder of the

TRSS. The evaluation was taken in three steps. At first measurements were

made on spherical soft tissue phantoms made of silicone with embedded stiff

nodules at different depths under the surface of the sphere. The next step

was to use spherical shaped biological tissue models of porcine muscle with

embedded stiff nodules before the las phase of the study that included

measurements on resected prostate glands with cancer tumours.

The spherical samples were mounted in a rotatable holder where they were

held in place by two spring loaded concave discs in a way that the sample

could be rotated around its own horizontal axis. The movement of the sensor

was kept perpendicular to the surface of the samples at all times. For the

measurements on the silicone spheres, the measurement positions were

chosen directly over the nodule αsm = 0° and ± 10° and ± 20° to the left and

right. The reason for this was to evaluate if the sensor could detect a nearby

stiff nodule when the sensor movement was angle towards the surface and a

short distance, 3,5 and 7 mm on the side of the curved surface.

In the study, two resected prostate glands from patients undergoing a

radical prostatectomy were obtained for measurements: Ethical approval

was given by the Ethical Committee at Umeå University (Dnr 03-423). From

the measurements it was concluded that the TRSS could detect stiffer areas

on the prostate glands, areas that were confirmed to be prostate cancer

through a histopathological evaluation.

48

Resonance Sensor Lab

1. P. Hallberg, Applanation Resonance Tonometry for

Intraocular Pressure Measurements, Dissertation 2006, ISSN 1653-6789:1, ISBN 91-7264-061-8

2. V. Jalkanen, Resonance Sensor Technology for

Detection of Prostate Cancer, Licentiate thesis 2006, ISSN 1653-6789:2, ISBN 91-7264-153-3

3. P. Lindberg, Development of a Resonance Sensor

System Designed for Prostate cancer Detection, Licentiate thesis 2007, ISSN 1653-6789:3, ISNB 978-91-7264-432-8

4. V. Jalkanen, Tactile Sensing of Prostate Cancer – A Resonance sensor Method Evaluated Using Human Prostate Tissue in Vitro, Dissertation 2007, ISSN 1653-6789:4, ISBN 978-91-7264-441-8

5. A P. Åstrand, A Flexible Resonance Sensor System For Detection of Cancer Tissue – Evaluation on Silicone, Licentiate thesis 2012, ISSN 1653-6789:5 ISBN 978-91-7459-477-5

6. A P. Åstrand, A tactile Resonance Sensor System for Detection of Prostate Cancer ex vivo – Design and Evaluation on Tissue Models and Human Prostate, Dissertation 2014, ISSN 1653-6789:6, ISBN 978-91-7601-006-8

Department of Applied Physics and

Electronics

Centre for Biomedical Engineering and

Physics

Industrial Doctoral School

Umeå University, SE-901 87 Umeå

www.tfe.umu.se

ISSN 1653-6789:6

ISBN 978-91-7601-006-8