Artificial Intelligence Technology to Improve Prostate Cancer Diagnosis
SBRI Outpatient Transformation Challenge
In 2020/2021, the way outpatient services being delivered in Wales was undergoing a period of transformation to ensure that it delivers the care for patients in a Prudent and Value-Based way for a sustainable health-care model for the future. This along with COVID led to a significant backlog in review of referrals in secondary care and as capacity is not going to match demand, innovative ways of managing patients differently to reduce unnecessary referrals were eagerly sought.
In April 2021, we launched an Outpatient Transformation challenge with businesses being able to apply for a share of up to £800,000.00 for a two Phase feasibility and development contract. The challenge asked for innovations with assistance with one of three themes and the use of Artificial intelligence to ‘review images/samples’ was one of them.
Within Pathology, Pathologists review 4 million glass slides holding prostate tissue samples from 100,000 men. This creates some issues for the NHS, firstly there is a shortage of Pathologists and secondly it is expected that the number of men biopsied will have doubled by 2030. The overall demand for Pathology services is growing, both in number and in complexity of tests that are performed.
Data shows that due to the COVID 19 pandemic this had led to diagnostic services having a backlog of patients with suspected cancer. This backlog will only put further strain on an already over-stretched pathology workforce as services brace themselves for an expected influx of tests.
Early detection saves lives; it was an amazing deployment journey to show how AI expedites and improves quality of prostatic pathology diagnoses in Wales. This technology not only improved our cancer pick up rate but also helped us to refine various prognostic indicators in more objective manner. The computation pathology in prostate lead the way to for development and deployment of technology in other specialities; currently work being carried out on breast, colorectal, lymphoma, upper GI and PDL1 lung in national level in various stages
Consultant Pathologist And Clinical Director, BCUHB.
This project was a collaboration with Welsh Government, SBRI Centre of Excellence and NHS Wales Pathology. The Histology Team within Betsi Cadwaladr University Health Board took the lead with colleagues from, Swansea Bay and Aneurin Bevan Health Boards.
There was a real enthusiasm and passion from the clinical teams within this project to make it a success, whilst working very busy day jobs. Weekly project teams were set up and attendance was excellent throughout.
The Call to Industry
The call to industry was released on 19th April 2021, with one month for applicants to apply, on closing there were 49 applications received.
Of these 49 applications, four were chosen to develop their solutions for the Outpatient Transformation Challenge. One of the successful four was a medical technology company Ibex Medical Analytics, which was chosen to test and develop the Galen Prostate technology, which is designed to aid clinicians in making prostate cancer diagnosis.
This funding enabled the team to collaborate with Ibex to identify a robust, validated, clinically assured Artificial Intelligence (AI) technology solution. This solution would process a high-volume of pathology slides, analyse them using high-power computing and machine learning technologies to support pathologists to ‘triage’ the prostate cases.
As a result, Betsi Cadwaladr UHB, Swansea Bay UHB and Aneurin Bevan UHB, began their own testing and development and jointly became the first to use the application to assist in diagnosing patients.
Weekly Project Team meetings were held with the supplier to ensure a two way iterative process throughout the 18 months of the project.
Clinical, Ethical assurance and bias was led by Pathology clinical colleagues and was discussed and considered at each of the meetings and documented by the SBRI Centre of Excellence.
It was important to document the benefits of the solution, baseline activities were undertaken and worked through during the project lifecycle.
There were many benefits to running this challenge learning how the use of AI technology and innovation can help to improve clinical practices.
The main areas where the successful Ibex Galen Prostate solution help improve patient care were:
- Improved diagnostic quality by reducing discrepancies and increasing accuracy in prostatic biopsies that may have been missed by Pathologists.
- Improved Pathologist productivity by reducing the time spent per prostate biopsy case, reducing the time required for Multi-Disciplinary Teams review and able to manage the increasing workloads.
- Improves the turnaround time of cancer diagnosis by enabling triage and prioritization of prostatic biopsies based on AI pre-screening and prostate biopsies requiring immunohistochemistry to confirm diagnosis, by pre-ordering the tests based on initial AI pre-screen.
- Increasing Pathologist confidence in diagnosis, leading to a reduction in the usage of Immunohistochemistry and therefore its associated costs.
- Improves overall Pathologist work environment and job satisfaction, leading to increase staff retention and ability to recruit.
The most important lesson learned for this project was the commitment of the project team and their excitement and enthusiasm for the technology. There was strength in numbers with clinical colleagues in attendance at Project meetings from across Wales. There was a combined ‘All Wales’ approach in development and testing.
During this project, Pathology teams within the Health Boards have allowed themselves to analyse their processes and discovered new ways to measure performance, accuracy and control. This has given them the confidence to use Artificial Intelligence.
Funding for Scale and Spread activities enabled a larger pilot/trial of the solution across NHS Wales with the inclusion now of Cardiff and Vale, Cwm Taf and Hywel Dda University Health Boards.
This technology requires a strong appraisal to consider its fuller clinical effectiveness, specifically, how does it change patient management, outcomes and cost-effectiveness. This is now ongoing with all the Health Boards over a 12 month period with assistance from colleagues within Health Technology Wales.
The Scale and Spread activities over the next 12 months are going to assist the Pathology teams to build up ‘real world evidence’ to feed into a technical appraisal to inform any procurement decision. An appraisal is also being undertaken by colleagues in Health Technology Wales to consider fuller clinical effectiveness (specifically, how does it actually change patient management and outcomes) and cost-effectiveness. This will also feed in to learning lessons of the use of AI in Pathology services across Wales.
Scale and Spread activities across Wales will enable pathology to learn more about AI technology. It may not be the IBEX solution that is procured, as any procurement will need to be undertaken via procurement rules and regulations but the process and use in the interim will enable continued lessons to be learned.
There are opportunities for advancements in use of AI technology across Wales that could help inform other cancer diagnosis and we are excited to undertake further work in this area.
The Pathology Service was a perfect example of what likeminded individuals with a passion for innovation can achieve when they work together for an overall aim and purpose.
If you wish to discuss further or learn more about running your own SBRI please contact the SBRI Team on SBRI.COE@wales.nhs.uk