Sperling Medical Group

reading & research

Artificial Intelligence in Medicine: AI Finds Incidental Prostate Cancer in Routine CT Scans

Imagine this scenario. Your doctor orders a pelvic CT scan to check if the pain you’re experiencing is due to a kidney stone or other blockage. After the scan, he says, “I’m happy to tell you there’s no evidence of stones and your kidney looks fine. Oh, and incidentally, you have prostate cancer.”

Of course, no professional would sound so casual about prostate cancer. However, the term incidentally is clinically correct when a medical scan or test for one condition reveals another disease—often, cancer—that was not suspected to begin with. In the scenario described above, your medical record would accurately state that you have incidental prostate cancer (PCa).

Of course, learning you have cancer when your doctor wasn’t even looking for it is not good news. On the other hand, incidental cancer detection can be a lifesaver. For example, kidney cancer is increasingly detected as the incidental result of abdominal ultrasound or CT scans long before cancer symptoms appear. A 2012 Icelandic study compared the records of all living Icelandic patients diagnosed with kidney cancer from 1971 to 2005. The survival factors of those whose tumors were discovered incidentally were compared with those who were diagnosed as the result of symptoms. The study showed that survival had improved “… mainly related to a steep increase in incidental detection.”[i]

Artificial intelligence detects incidental prostate cancer

A research team from Australia has now developed an Artificial Intelligence (AI) tool that save the lives and lifestyles of many men whose PCa would otherwise lie undiscovered in their bodies until it was too late for Active Surveillance or a focal treatment like Focal Laser Ablation.

The 2021 study by Korevaar, et al. was published in the British journal Nature’s “Scientific Reports.”[ii] The authors note that CT scanning is “generally considered insufficient to diagnose PCa due to its inferior soft tissue characterization…” To overcome this, they trained a deep learning model using an image dataset composed of 139 clinically significant PCa patients and 432 controls. This included what to look for, and where. The computer can detect tiny disease characteristics that the human eye can’t see. As study author Dr. Ruwan Tennakoon says, “It’s like training a sniffer dog – we can teach the AI to see things that we can’t with our own eyes, in the same way a dog can smell things human noses can’t.”[iii]

Furthermore, the deep learning branch of AI technology applied by the team means that the computer can build on its own experience in order to generalize what it learns to various scans (different angles, body positions, equipment, etc.). It thus becomes self-learning in much the same way as a human brain.

The authors have made a significant contribution toward incidental detection of PCa. As fellow author Dr. Mark Page points out, “For example, emergency patients who have CT scans could be simultaneously screened for prostate cancer… If we can detect it earlier and refer them to specialist care faster, this could make a significant difference to their prognosis.”

Our team at Sperling Prostate Center is proud to have been early adopters of AI tools. Using such forward-looking technology advances, we are able to provide PCa detection and diagnosis with superior accuracy and efficiency. Thanks to earlier identification of each patient’s tumor, for those who are qualified we offer excellence in both Active Surveillance monitoring and Focal Laser Ablation.

NOTE: This content is solely for purposes of information and does not substitute for diagnostic or medical advice. Talk to your doctor if you are experiencing pelvic pain, or have any other health concerns or questions of a personal medical nature.

[i] Palsdottir HB, Hardarson S, Petursdottir V, Jonsson A et al. Incidental detection of renal cell carcinoma is an independent prognostic marker: results of a long-term, whole population study. J Urol. 2012 Jan;187(1):48-53.
[ii] Korevaar S, Tennakoon R, Page M, Brotchie P et al. Incidental detection of prostate cancer with computed tomography scans. Nature (UK) Scientific Reports. 2021;11(1):1-10.
[iii] “New AI tech for early detection of prostate cancer.” RMIT University, July 8, 2021. https://www.rmit.edu.au/news/media-releases-and-expert-comments/2021/jul/ai-prostate-cancer