Digital Pathology Symposium

Digital Imaging

NE England ‘Proof of Concept’ Trial

David Bottoms, NESCN Diagnostics Project Manager

Dr Kaushik Dasgupta, Consultant Pathologist

Pathology Digital Imaging for Primary Diagnosis

“Are we nearly there yet ?”

Aims

• Scanner quality/ reliability• Case management software • Quality of the images in comparison to traditional

microscopy• Pathologist Workstations • Speed• Imaging tools – annotation/ measurements• Trust to Trust image sharing• Access from home• Algorithmic Analysis

Objectives

Outcome Views :-

• Does it improve the quality of diagnosis/ provide better outcomes ?

• Does it speed up diagnosis ?

• What are the constraints ?

• What would be the impact from an IM&T perspective ?

• Is it ‘affordable’ ?

Scope of Involvement

Digital Pathology deployments by phase

Durham North Tees South Tees

Phase 1

Phase 2

Northumbria

Newcastle Royal Infirmary

Gateshead

SunderlandMDT

Lessons from my tryst (KD)Will it all come out in the wash?

• 100% concordance

• Confident use of tools

• Confident of low power

dx

• 5/103 (4.8%) rescans

• More time than

analogue (subjective)

The live experience

4 5

28

46

12

10

1

22

19

1

22

3 19

1

2

Total

Axillary Nodes

Bladder biopsy

Breast biopsy

Breast resection

Breast Sentinel LN

Cervical biopsy

Cervical loop

Gallbladder

GI biopsy

GI polyp

Liver biopsy

Tissue Type

Count of Episode Number

186 cases,

(24 off site/digital home reporting)

Rescans

Time and analogue

31.72%

31.72%

36.02%

0.54%

Total

less

longer

same

(blank)

Time to assess case cf glass

Count of Episode Number

72.28%

27.72%

Total

no

yes

Glass Required For Si...

Count of Episode Number

Time cf glass

Glass needed

Pass the glass

Diagnostic Discordance (6)1.35% minor 0.69 % major (2%)

Glass

Required For

Sign Out

(Y/N) If Yes state reason

Diagnostic

Concordance

(Y/N) If No state reason

no no

Underscoring of mitosis in

scans

yes lack of confidence no Difficult for VIN 1,2 at margins

yes lack of confidence no Missed small foci of invasion

yes lack of confidence no hazy scan

yes lack of confidence no mucosal prolapse in C

yes difficult case no

Partial atrophy mimicking

cancer

CONFIDENCE TREND

Summary (289 cases)

• Quality benefits- Breast, cervix- accuracy• NHSBCSP and CRC- quality neutral• Steep learning curve/mental barrier• Work flow, remote site reporting, virtual academy

of specialists• Much slower for single slide, few fragments, low

complexity cases (skin, GI, endometrium)• CAUTION- Subtle foci of malignancy in a large

volume- TURP, re resection of bladder tumours, post NAC breast/colon (ROI tool?)

Results

No of Cases Tissues

N Tees 983 (5 Consultants) Breast, GI, Gynae, Head and Neck, Respiratory, Skin and Urology

CDDFT 320 (2 Consultants) Breast (2), GI(49), Gyn(22) H&N(4), Skin (56),Uro(3), Other(9)

Gateshead 90 (1 Consultant) Breast, GI, Gyn, H&N, Skin, Uro, Soft Tissue

Northumbria 32 (1 Consultant) Breast, GI, Gynae, Head and Neck, Respiratory, Skin and Urology

Newcastle 12 (1 Consultant) Soft tissue, Lymphoma

S Tees 17 (3 Consultants) Prostate cores, kidney endometrial, Lletz biopsy cervix, pleural, renal, GI

Key Issues

• Time available (for Pathologists to review cases)

• Trust to Trust IT Firewalls

• IM&T Resource/ Involvement

• IM&T Storage Strategy

• Technology currently does not accommodate megablocks, fluorescence, polarisation, gynae-cytology

• RC Path guidance – (in the making)

• ‘Challenging’ for larger cases

• Affordability – pump priming – invest to save

Key Benefits

• Faster for an estimated 80% of general work• Measurements considerably quicker and reproducible• Algorithms save considerable time – reproducible – lab based • Can provide better outcomes for patients – grading/ staging• With LIMS I/F would significantly reduce ‘wrong slide’ risks• Better workload management – urgent cases – referrals to

specialists – workload balancing – pull vs push• Excellent for ‘sharing’ expertise/ knowledge/ opinions - annotations• Saves a lot of technical time – tissue exchange/ slide retrieval/ MDT

prep/ archiving• Improved access to images via web – mortuaries/ MDT rooms/

Home• Brilliant for training & education• LEAN

Objectives

Outcome Views :-• Does it improve the quality of diagnosis/ provide better

outcomes ? – yes – better staging• Does it speed up diagnosis ? – too short/ small a study

to be sure but overall feeling was ‘yes’ • What are the constraints ? – funding/ implementation

resource/ transition• What would be the impact from an IM&T perspective ?

– all ‘doable’ if they’re on-board from the start• Is it ‘affordable’ ? – only with pump-prime funding –

however the ROI should be worthwhile

Aims/ Outcomes

• Scanner quality/ reliability – No issues/ Reliable• Case management software – Easy to Use / Improvement • Quality of the images in comparison to traditional

microscopy – As good/ Acceptable• Pathologist Workstations – Easy to Use/ Learn• Speed – Improves with experience/ quicker once learnt• Imaging tools – annotation/ measurements - brilliant• Trust to Trust image sharing – worked without delays/

diagnostic quality• Access from home – worked without delays/ diagnostic

quality• Algorithmic Analysis – excellent/ reproducible

What Next ?

• Business Case– Regional ‘Academy of Pathologists’ approach

preferred

– Recent ‘sign-up’ in principle by 7 CEOs encouraging

– However still requires pump-prime funding they don’t have

– Needs greater support from NHS Digital, National Cancer Capacity Funding to get what is seen as a new, unproven technology onto a ROI testbed.

Thanks

• To All Consultants and Lab staff who took part

• Staff from GE Omnyx for their generous support of the project

• To Trust IM&T depts. for supporting the interconnectivity infrastructure

• Any questions ??– kaushik.dasgupta@nth.nhs.uk

– david.bottoms@nhs.net

– d.bottoms@djbhealthcare.com