How Lucida Medical’s AI-powered software spotted man’s prostate cancer that a radiologist couldn’t detect
Lucida Medical’s Pi software, the result of a five-year collaboration with Hampshire Hospitals NHS Foundation Trust, has helped to diagnose a high-grade cancer that a radiologist hadn’t been able to see.
The Cambridge-based company is empowering radiologists to achieve better results with its software, which is now in use in NHS and European hospitals. Pi AI is the first commercial product of its kind with performance equivalent to that of expert radiologists to detect prostate cancer, the most common cancer in men.
A Lucida Medical spokesperson said of the ground-breaking moment in prostate cancer detection: “The reporting radiologist initially thought that this man’s prostate was normal. But Pi highlighted this lesion in the apex, the lowest part of the prostate that can be easily overlooked, prompting the radiologist to report this. Biopsy showed Gleason 9 high-grade prostate cancer. For this patient, early detection should enable their high-grade tumour to be treated effectively, before it spreads further.”
The results of Lucida’s collaboration with Hampshire Hospitals NHS Foundation Trust have been published in European Radiology. The study was part of PAIR-1, (Prostate AI Research – 1), a collaborative research study, approved by the NHS Health Research Authority, between eight NHS Trusts and Lucida Medical. The study partners gathered historical data from over 2,000 patients and used this to develop, train and validate Pi, a software platform which uses AI algorithms to analyse magnetic resonance imaging (MRI) scans to help distinguish clinically significant prostate cancer.
An MRI scan is a key step to diagnose prostate cancer. The MRI is used to help identify patients at low risk who can avoid a painful, invasive biopsy, and to locate possible lesions so that higher-risk patients can have a targeted biopsy to maximise the chance of finding cancers that need treatment.
Dr Aarti Shah, consultant radiologist at Hampshire Hospitals NHS Foundation Trust, was chief investigator on the study. She highlighted that “analysing MRI scans is a time-consuming task for expert radiologists, and there are too few of us in the UK and many other countries. Pi offers exciting potential as an aid to help reporting radiologists in triaging workloads as well as producing visual reports to aid contouring of lesions for biopsy”.
Mark Hinton, CTO at Lucida Medical, said: “Pi is medical device software that is CE-approved for use in clinics. We developed Pi to automate key steps like outlining lesions and calculating risk scores, to assist radiologists to make these challenging decisions.”
Dr Francesco Giganti, associate professor of radiology at University College London, presented the results of the PAIR-1 study last week at the European Congress of Radiology (ECR) in Vienna. He noted that “this research found that Pi is non-inferior to multidisciplinary team-supported radiologists across a validation set of sequential cases from six NHS hospitals with a wide range of MRI scanner types. This is the first time that a commercial AI for prostate MRI has been tested on diverse, real-world data”.
Dr Rix told the Cambridge Independent: “Lucida Medical has been working for over five years to tackle key challenges with how we diagnose prostate cancer, the most common cancer in men.
“While MRI is now a key test for the disease, we have a growing shortfall in expert radiologists needed to make the diagnosis, and there is wide variability in quality and waiting times. The result is that far too many men are only diagnosed when their cancer is too late to treat.
“We solve this by providing AI-based software, Pi, that automatically analyses prostate MRI scans. Pi uniquely operates at the level of experts, enabling doctors to complete the diagnosis faster and more accurately, to cut waiting times and ensure patients get the treatment they need.”
