Computer Vision News - December 2022

48 Medical Imaging Challenge “ This time around, we wanted to partner withdifferent centers andexpertsworldwide to get a unified perspective on how we should tackle this challenge, ” Anindo tells us. “ The study design was reviewed by 16 multidisciplinary experts from the fields of radiology, urology, and AI related to prostate . We wanted to make a bigger impact with the Grand Challenge and take it that one step further. ” Annually, one million men are diagnosed, and 300,000 die from clinically significant prostate cancer. When someone has a prostate MRI, it first identifies if there is an aggressive lesion. The AI models are expected to give a case-level diagnosis, with each patient receiving a percentage likelihood of cancer being present. “ There are many research topics within this field that specifically investigate the added The number of prostateMRI acquisitions to be collected is expected to rise by quite a large margin in the coming years, meaning there will be more and more cases to be evaluated. Expert radiologists perform this evaluation, but there is a global shortage of these experts. Prostate cancer detection in MRI with AI aims to alleviate this workflow and make these radiologists quicker. Then if there are sufficient radiologists, its goal is to make it more accurate . The PI-CAI Challenge targets clinically significant prostate cancer detection and diagnosis in MRI. One other challenge in this space, in 2016, was the PROSTATEx Challenge, also from Radboudumc. The PI- CAI team wanted to scale up with more data, updated evaluation strategies, and a multi-center, multi-vendor dataset and cohort . Anindo Saha and Joeran Bosma are PhD candidates with the Diagnostic Image Analysis Group at Radboud University Medical Centre, working on prostate cancer detection in MRI under the supervision of Henkjan Huisman. Jasper Twilt is a PhD candidate at Radboudumc, with a background in technical medicine, under the supervision of Henkjan andMD-PhDs Jurgen Fütterer andMaarten de Rooij. Anindo, Joeran, and Jasper are co-organizers of the PI-CAI (Prostate Imaging: Cancer AI) Grand Challenge, and they speak to us as it enters its testing phase. PI-CAI (PROSTATE IMAGING: CANCER AI) GRAND CHALLENGE