AI tool developed to streamline pathology workflow

A team of researchers, academics and pathologists from the University of Oxford and Oxford University Hospitals NHS Foundation Trust (OUH) has developed an artificial intelligence tool to help streamline pathology workflows – which could speed up diagnoses of prostate cancer.

The multidisciplinary team from Oxford has created an algorithm that automates requests for “additional investigation of diagnostically uncertain prostate biopsies”, according to OUH.

As per the trust, over 50,000 cases of prostate cancer are diagnosed every year in the UK. However, the diagnostic process requires a biopsy to be analysed by pathology services and it’s estimated that over 60,000 of these are performed per year.

There is a high workload associated with these biopsies, and the need for pathologists to review and refer samples where the Hematoxylin & Eosin (H&E) staining process does not provide enough ‘evidence’ for a confident diagnosis –  which is considered to be between 25 and 50 per cent of cases – for additional immunohistochemistry (IHC), is described as a significant “bottleneck” in the workflow.

To address this challenge, thereby potentially helping save staff time and quickening diagnosis, pathologists from OUH and the Nuffield Department of Surgical Sciences partnered with biomedical image analysts from the University’s Institute of Biomedical Engineering, Big Data Institute and Ludwig Oxford Institute for Cancer Research to come up with a solution.

This involved using annotated prostate biopsies to train an AI tool to “detect tissue regions with ambiguous morphology” and identify those which would need IHC. By automating the request for further investigation, a pathologist would only need to review after all the processes had been carried out – potentially saving around 11 minutes of pathologist time per case.

Professor Clare Verrill of OUH said: “The NHS spends £27 million on locum and private services to make up for the shortfall in pathology service provision. By using this AI tool to triage prostate biopsies for IHC, pathologists would spend less time reviewing these cases, which would not only lead to financial savings, but it would also accelerate prostate cancer diagnoses to inform patients and treating clinicians earlier.”

The work, supported by the PathLAKE Centre of Excellence for digital pathology and artificial intelligence, is now set to be further developed and validated by using pathology data from different locations.