Detecting Prostate Cancer Using AI

Written by Hannah Segal and Edited by Kevin Liu

Would you trust a robot to diagnose cancer? Perhaps in the future you will.

In the United States, prostate cancer is the second most common cancer among men. Cancer cells are found in the small male glands that produce seminal fluid to nourish and transport sperms; usually the cancer grows slowly and is confined to the prostate glands [1]. Most prostate cancers are detected through screening tests, and the diagnosis is confirmed after the patient undergoes a prostate biopsy, where some of the issue is removed to be analyzed, by a resident doctor [2]. However, a recent published study successfully developed an AI-based algorithm that was reported to accurately detect, grade, and evaluate clinically relevant findings in the diagnosis of prostate cancer [2].

Artificial intelligence, commonly referred to as AI, involves a digital computer or a computerized robot that can perform tasks commonly associated with humans [3]. A study published in the Lancet Digital Health by researchers at the University of Pittsburgh Medical Center has used and identified an AI program that demonstrates the highest accuracy known-to-date on recognizing and characterizing prostate cancer [2].

Over a million parts of stained tissue slides that were taken from patient biopsies were used to train the AI to differentiate between healthy and abnormal tissue. “Humans are good at recognizing anomalies, but they have their own biases or past experiences,” said senior author Rajiv Dhir, M.D., M.B.A., chief pathologist and vice chair of pathology at UPMC Shadyside Hospital [4]. Rajiv has noted how the AI system “taught itself” to detect prostate after analyzing data from over 1,200 patients. The AI program examines biopsies in a similar manner to that of a pathologist and evaluates them according to an expert-referenced standard [5] [6].

“During testing, the AI demonstrated 98% sensitivity and 97% specificity at detecting prostate cancer—significantly higher than previously reported for algorithms working from tissue slides.” – ITNonline [1].

All-in-all, this new AI technology brings in a new element to standardizing care. Enforcing AI algorithms does not necessarily mean that machines are superior to humans, but rather, it helps make an assessment to verify doctors’ findings, and is accessible for those that are less experienced in diagnosing prostate cancers. For less experienced pathologists, this algorithm could be used to catch cases that might otherwise have been missed. This program may also reduce the workload of pathologists since the pathologist would only need to confirm or reject the opinion of the AI system. While the currently produced results are promising, these algorithms continue to be revised with the aim to detect different types of cancer in the future [4].

References:

1. “Prostate Cancer – Symptoms & causes”. Mayo Clinic, Diseases and Conditions., 2020, Accessed 17 Feb. 2021. https://www.mayoclinic.org/diseases-conditions/prostate-cancer/symptoms-causes/syc- 20353087.
2. Pantanowitz, L., Quiroga-Garza, M.G., Bien, L., Heled, R., Laifenfeld, D., Linhart, C., Sandbank, J., Shach, A.A., Shalev, V., Vecsler, M., Michelow, P., Hazelhurst, S., Dhir, R. (2020). An artificial intelligence algorithm for prostate cancer diagnosis in whole slide images of core needle biopsies: a blinded clinical validation and deployment study. The Lancet Digital Health, 2: E407-E416.
3. Copeland, B. Artificial intelligence. Encyclopedia Britannica, 2020. Accessed 24 February 2021. https://www.britannica.com/technology/artificial-intelligence.
4. Artificial Intelligence Identifies Prostate Cancer with Near-perfect Accuracy”. ITNonline, Prostate Cancer, 2020, Accessed 17 Feb. 2021. https://www.itnonline.com/content/artificial-intelligence-identifies-prostate-cancer-near-p erfect-accuracy#:~:text=During%20testing%2C%20the%20AI%20demonstrated,algorith ms%20working%20from%20tissue%20slides.
5. Ström, P., Kartasalo, K., Olsson, H., Solorzano, L., Delahunt, B., Berney, D. M., Bostwick, D. G., Evans, A. J., Grignon, D. J., Humphrey, P. A., Iczkowski, K. A., Kench, J. G., Kristiansen, G., van der Kwast, T. H., Leite, K., McKenney, J. K., Oxley, J., Pan, C. C., Samaratunga, H., Srigley, J. R., Takahashi, H., Tszuki, T., Varman, M., Zhou, M., Lindberg, J., Lindskog, C., Ruusuvuori, P., Wählby, C., Grönberg, H., Rantalainen, M., Egevad, L., Eklund, M. (2020). Artificial intelligence for diagnosis and grading of prostate cancer in biopsies: a population-based, diagnostic study. The Lancet. Oncology, 2: 222–232.
6. Yadav, K.K. (2018). How AI Is Optimizing the Detection and Management of Prostate Cancer. IEEE pulse, 9: 19.