Elizabeth A. Krupinski, PhDArizona Telemedicine

Program

Ovitt , et al.Intravenous angiography using digital video subtraction: x-ray imaging system.AJR 135(6):1141-4, 1980.

An x-ray imaging system, using digital subtraction techniques, has been developed. The system requires: (1) high output generation equipment; (2) an image intensif ier capable of receiving high output exposures, 1 mR (2.58 X 10(-7) C/kg) at the face of the intensif ier, without loss of either contrast or resolution; (3) a precision digital video camera; (4) processing computer with sufficient storage capacity; and (5) digital image storage. With this system it is possible to visualize the major arteries after intravenous contrast injection.

Capp et al.The digital radiology department of the future.Rad Clin N Am 23(2):349-55, 1985.

The boom in microelectronics, including cost-effectiveness, hasnow allowed us to consider the use of these objects to storedigital images. There remains much research, development, And clinical evaluation to be done in receptor technology. Further improvements in image processing, optical laser disk storage, & optical transmission and further commercial Development of display technology must take place. All of These developments are occurring simultaneously . Within 5 to 10 years, radiology departments wil l most l ikely be totallyelectronic, probably cost-effective, and, i t is hoped, morediagnostically accurate.

Oldest established TM application Well integrated in numerous settings Facil itated by co-evolution PACS Few to no reimbursement issues Only interventional radiology currently

less amenable to teleradiology applications

Litt le/no differences between teleradiology & on-site radiology

ACR-NEMA development DICOM Continual updates of DICOM Development of standards & practice

guidelines that explicit ly include teleradiology

http://medical.nema.org/ http://deckard.duhs.duke.edu/~samei/tg18.htm

Key is the human-computer interface

Series of observer performance studies designed to optimize the digital reading room environment

Performance metrics◦ Diagnostic accuracy (ROC)◦ Search eff iciency (eye position)

Human Visual System Modeling

Softcopy display parameters◦ Luminance◦ Calibration (tone scale)◦ Type of phosphor◦ CRT vs LCD◦ MTF◦ Viewing angle◦ Number of displays◦ Ambient l ighting◦ Compression◦ Role of color

P45 P104

Spatial & contrast resolut ion l imitations require radiologists to search images. The UVF is about 2.5 deg radius. Probabil i ty of target detection fal ls off as a function of target eccentricity from axis of gaze.

Total viewing time shorter Time to first hit shorter Total t ime on lesion shorter Fewer returns to lesion Total path length shorter Overall = more EFFICIENT

TaskTask

ExperienceExperience

Med StudentResidentPathologist

Display & InterfaceDisplay & Interface

20% fixations fell outside diagnostic image

Hospital/MC Clinics Mobile van◦ Mammo

Dedicated◦ PET cl inic◦ THH◦ UASA

Public Health Battlef ield Hand-held

Medical Imaging Consultants -> ? ◦Data acquisit ion & archiving

RadWorks (GE) -> Siemens/Fuji◦ Viewing station

35% of department’s reading volume 25% department’s income ◦Reading only & reading + archiving◦ $/case & $/set volume

68% sites using AHSC hub for TM services use TR service

TR typically 1 s t service requested◦ 79% of sites with TR use only TR◦ 21% started with TR & added

services TR specialty with most volume

Time from mammography to consult with oncologist ~ 28 days◦ Screening mammography◦Diagnostic mammography◦ Biopsy◦ Pathology processing & report◦Oncology consultation

THIS IS TOO LONG!

Even Worse in RuralEven Worse in Rural

- DS3 (45 Mbps) backbone- ATM protocol

- T1 (1.5 Mbps) links- 65 direct link sites- ~ 85 with affiliated - NARBHA- DOC- IHS

- RT & SF applications- ~ 55 sub-specialties- Teleradiology core app.

Started in 2001 to rural sites 7/28 telerad sites send mammo Mostly use GE system Directly to TBC for reading Some archive some do not Contracts specify 30-45 min TAT > 26,000 telemammography

MammographySurgical

ConsultationBreast Care Biopsy

Rapid tissue processing

Virtual slide scanning

Laboratory report

Telemedicine

Clinic

Digital Mammography

(Teleradiology)

Telepathology

Cancer specialistTeleconsultation

UltraClinic Model

Milestone Medical Systems RHS-1-30 Milestone Medical Systems RHS-1-30 Vacuum HistoprocessorVacuum Histoprocessor

Quick Processing

DMetrixTM -40 Slide Scanner DMetrixTM -40 Slide Scanner SystemSystem

TelepathologyTelepathology

Lab Process MinutesGrossing 3Tissue Processing 58Embedding 13Cutting 10Stain/Dry/Coverslip 32Scanning (2 Slides) 13LM Interpretation 6Telepathology 14

Telepathology report sent S&F to oncologist

Oncologist connects RT videoconference to rural location

Discuss pathology results If necessary discuss treatment

options and plan of action

MI = prevention, detection, diagnosis, treatment & therapy

Acquisit ion & display technology continually changes

Clinician shortages are not easing Rapidly expanding types & number images◦ Multi-modality & fusion complimentary

information sources is becoming common◦ Anatomy & function gross & molecular levels◦ Merging specialt ies

(Quon et al. J Cl in Oncol; 2005; 23:1664-1673)

Clarke et al. Validation if Tumor Burden Measurements Using 3D Histopathology. In: Digital Mammography 2008. Springer-Verlag.

Image Display, Analysis & Processing are key l inks in the imaging chain

Need to present data to the cl inician in the most eff icient & informative manner

Taking into account perceptual & cognit ive capabil i t ies of human observer

Ultimate goal = facil i tate decision-making process & enhance patient care

Related goal = improve workflow & the reading environment

Stereo vs Tradit ional

- Az 0.85 to 0.94

-23% increase TPs

-105% increase calcs

- 46% decrease FPsGetty et al. Stereoscopic Digital Mammography: Improved Accuracy of Lesion Detection in Breast Cancer Screening.In: Digital Mammography 2008. Springer-Verlag.

50 DR chest images (PA) : 1/2 solitary pulmonary nodule verif ied CT; 1/2 nodule free

6 radiologists (3 sr residents, 3 board-cert if ied)

3MP Barco color medical-grade display (Barco Coronis MDCC-3120-DL) vs COTS color 2MP monitor (Dell 2405)

Calibrated (DICOM GSDF ) to luminance corresponding to backlight aging 1-year t ime ◦ Max Barco 500 cd/m 2 min 0.77 cd/m 2

◦ Max Dell COTS 342 cd/m 2 min 0.376 cd/m 2

0

0.2

0.4

0.6

0.8

1RO

C A

z

Read

er 1

Read

er 2

Read

er 3

Read

er 4

Read

er 5

Read

er 6

Mea

n

Medical Grade

COTS

F = 4.1496, p = 0.0471Sensit ivity = 0.91 vs 0.86Specif icity = 0.93 vs 0.92

05

1015202530354045

Mea

n Vi

ewin

g Ti

me

(sec

)

Read

er 1

Read

er 2

Read

er 3

Read

er 4

Read

er 5

Read

er 6

Mea

n

Medical Grade

COTS

F = 3.38, p = 0.067

0

0.5

1

1.5

2

2.5

3

3.5

4

Mea

n Ti

me

Firs

t H

it (s

ec)

TP FN FP

COTS

Medical Grade

0

1

2

3

4

5

6

7

8

Cum

ulat

ive

Dw

ell (

sec)

TP FP FN TN

COTS

Medical Grade

No significant differences

TN significantly different

Firs t (Re feren ce ) Im ag e Or igin al sR GB Im ag e o n Displa y with S hif ted P r im ar ies

S-CIELAB Delta-E Image Map RMS Delta-E = 3.3997

200 400 600 800 1000

100

200

300

400

500

600

7005

10

15

20

25

30

35

40

Carpal tunnel syndrome Elbow & shoulder (cubital tunnel) Neck, back & shoulder strains Computer vision syndrome◦ Eye strain◦ Dry eyes◦ Glaucoma◦ Headaches◦ Corneal erosion and abrasions◦ Contact lens problems

Reader Fatigue

BLURRED VISION

Variable How long correlation How many correlation

Blurred vision R = 0.344 p = 0.0113 R = 0.422 p = 0.0015

Eyestrain R = 0.429 p = 0.0012 R = 0.475 p = 0.0003

Difficulty focus R = 0.384 p = 0.0042 R = 0.446 p = 0.0007

Headache R = 0.235 p = 0.0899 R = 0.432 p = 0.0011

Neck strain R = 0.384 p = 0.0042 R = 0.549 p < 0.0001

Shoulder strain R = 0.250 p = 0.0711 R = 0.469 p = 0.0003

Back strain R = 0.304 p = 0.0265 R = 0.424 p = 0.0014

General fatigue R = 0.471 p = 0.0003 R = 0.642 p < 0.0001

TR has made a signif icant impact on patient care over the past 20 years

Advances in technology wil l further change MI & interpretation of medical data by more clinicians

Costs can increase & decrease Optimizing observer accuracy while

maintaining eff iciency & comfort are crit ical to continued success

THANK YOU!