24/08/2026
AI in Cancer Care: Beyond the Technology
Artificial intelligence is no longer something that belongs to the future of healthcare.
The U.S. FDA has now authorized more than 1,500 AI-enabled medical devices. Many are used in radiology and imaging, but AI is also beginning to do something even more consequential: help estimate what may happen to a patient's cancer in the future.
For men with prostate cancer, that is already becoming a reality.
One FDA-authorized technology, ArteraAI Prostate, analyzes digitized prostate biopsy tissue and provides estimates of a man's risk of developing distant metastases and dying from prostate cancer over the following 10 years.
That raises an important question:
If AI can predict cancer risk, should patients trust the prediction?
The answer requires more than knowing that an AI system has received FDA authorization.
Patients—and their physicians—need to ask:
What exactly is the AI predicting, and how reliable is it?
An AI-generated risk estimate is not a crystal ball. It represents a probability based on the data used to develop and validate the system.
Was the AI tested in people like me?
This is particularly important when we consider health disparities. Men of African descent have historically been underrepresented in many prostate cancer clinical trials and research studies, despite carrying a substantially higher burden of prostate cancer.
If populations are inadequately represented when an AI system is developed and validated, can we assume its predictions will be equally reliable for everyone?
We shouldn't.
And there is another equity question:
Who has access to the technology and care needed to act on its results?
A sophisticated AI prediction has limited value if a patient cannot access the pathology expertise, diagnostic testing, specialists or treatments needed to turn that information into better care.
The Prostate Net's position
We don't believe patients should be either AI evangelists or AI skeptics.
We should be asking better questions:
What does the AI actually do?
What evidence supports it?
Was my population adequately represented?
How accurate is it for people like me?
Will it change my treatment or outcomes?
And can I actually access it?
AI has enormous potential to advance precision medicine.
But the goal should never be AI for AI's sake.
The goal is better information, better decisions—and ultimately better outcomes for every patient, regardless of race, geography or ability to pay.
We inform to fight!
