Artificial intelligence is steadily finding its way into routine hospital care, including prostate cancer diagnostics. In Norway, researchers at the Norwegian University of Science and Technology (NTNU) have developed an AI-based tool that can analyze prostate MRI scans and help clinicians determine whether a biopsy is necessary, and where it should be taken. Early clinical tests suggest the technology could significantly streamline diagnostic workflows, provided it is used with clear clinical oversight.
The AI tool, known as PROVIZ, was developed within a research project led by Tone Frost Bathen, professor at NTNU. According to Bathen, AI can play a valuable role in handling straightforward diagnostic cases. “AI tools can take over the detection of simple and clear-cut cases, allowing doctors to spend their time on more complex ones,” she said. At St. Olavs Hospital, where the tool has been tested on patients, results indicate that PROVIZ can help radiologists assess MRI scans more quickly and accurately.
“AI can enable radiologists to determine more quickly and more accurately whether a patient needs a biopsy, and where in the prostate it should be taken from,” Bathen explained. This is particularly relevant as the volume of prostate MRI scans continues to grow.
Rising demand for prostate diagnostics
Prostate cancer is the most common cancer among men in Western countries and is closely linked to aging. Studies show that prostate cancer can be detected in 10% of men aged 50, 50% of men aged 60, and around 70% of men over 80. “Prostate cancer is something most men die with, not from,” noted Simon A. Berger, a PhD research fellow at NTNU.
Widespread use of the PSA blood test has led to a sharp increase in detected cases, with approximately 5,000 new diagnoses each year in Norway. As more men undergo PSA testing, the demand for follow-up imaging and biopsies has also risen. MRI has become the standard next step, but interpreting these images requires time and highly specialized expertise, resources that are increasingly under pressure.
Trust determines acceptance
While the technical performance of AI tools like PROVIZ is promising, their success in clinical practice depends heavily on patient trust. In a recent qualitative study published in Qualitative Health Research, Berger interviewed 18 men diagnosed with prostate cancer using PROVIZ-supported imaging. The findings show that patients are willing to accept AI support only when an experienced physician confirms and explains the results.
“Trust in doctors and health professionals is key for artificial intelligence to gain a place in the diagnosis of prostate cancer. Technology alone is not enough. Human contact and professional assessment remain indispensable,” Berger said.
Patients were generally more comfortable with AI in lower-risk scenarios, such as detecting bone fractures. In higher-risk contexts like cancer, they placed the greatest trust in specialist physicians who could interpret and validate AI findings.
Doctors as guarantors of safety
Berger identified three layers of trust influencing patient acceptance: foundational trust in the healthcare system, interpersonal trust in health professionals, and conditional trust in AI itself. Even when patients recognized AI’s potential, they consistently wanted a human assessment alongside it. Particularly due to concerns about accountability and whether AI can capture the full clinical picture.
“The relationship between patient and doctor is still key,” Berger concluded. “For AI to be accepted in clinical practice, health professionals must be active communicators and guarantors of safety.” As NTNU moves toward patenting and commercializing PROVIZ, the research underscores a central lesson for digital health innovation: AI can enhance efficiency and accuracy, but its real value in oncology lies in supporting, rather than replacing, clinical expertise.
AI-driven prostate care
The Mayo Clinic recently developed the PSA Control Tower, an AI-supported monitoring solution designed to improve follow-up care after prostate cancer treatment. Traditional follow-up relies on fixed schedules and manual review of PSA test results, which can delay action when early signs of recurrence emerge. The PSA Control Tower continuously analyses PSA trends alongside broader clinical data to help clinicians identify which patients may need timely attention.
Built on the Mayo Clinic Platform, the system uses de-identified data from lab results, imaging, pathology and clinical notes to train predictive models that improve over time. Insights are presented through intuitive dashboards, supporting clinical judgment. According to Mayo Clinic, the approach enables continuous, guideline-aligned monitoring, earlier intervention and more personalised follow-up, while reducing workload pressure and supporting scalable precision oncology care.
