An AI healthcare firm asserts that its software can detect the extent of prostate cancer more accurately than medical professionals.
Avenda Health conducted a study involving ten doctors who evaluated 50 different prostate cancer cases each. The company’s Unfold AI software achieved an accuracy rate of 84.7% in detecting cancer, while the physicians’ manual evaluations ranged from 67.2% to 75.9%.
This research, carried out in collaboration with UCLA Health and published in the Journal of Urology, indicated that when AI assisted in cancer contouring, assessments of tumor size became 45 times more precise and consistent compared to evaluations without AI assistance.
According to Shyam Natarajan, an assistant adjunct professor of urology, surgery, and bioengineering at UCLA and the study’s senior author, the integration of AI made doctors’ assessments more accurate and increased their agreement levels when making evaluations.
Dr. Wayne Brisbane, an assistant professor of urology at the David Geffen School of Medicine at UCLA, noted that while doctors typically rely on MRIs to gauge tumor sizes, some tumors are not visible through MRI technology, highlighting the areas where AI can provide significant assistance.
Brisbane added that the implementation of AI in cancer treatment could enhance the effectiveness of personalized care, resulting in treatment plans that are better customized to meet patients’ individual requirements, ultimately improving success rates against the disease. He emphasized that AI has the capacity to exceed human capabilities.
Avenda Health’s CEO, Dr. Shyam Natarajan, expressed that seeing such innovations validated through research and acknowledged by the American Medical Association is empowering for physicians.
According to the American Cancer Society, approximately 1 in 8 men will receive a prostate cancer diagnosis in their lifetime, and 1 in 44 will succumb to the disease. It is projected that there will be 299,010 new prostate cancer cases in the US this year, with 35,250 related fatalities.