Length: 5 min read
Published: April 27th 2026
Medically Reviewed: By Dr Aqua Asif

At a Glance

AI is being used most often to help radiologists interpret prostate MRI – not to replace them. In well-designed UK and international studies, AI assistance can modestly improve detection of clinically significant prostate cancer, with the biggest gains for less experienced readers. AI cannot diagnose prostate cancer on its own: definitive diagnosis still relies on biopsy, and decisions about treatment still require a multidisciplinary specialist team.

Key Takeaways

  • Most common use – software that supports radiologists in reading prostate MRI (lesion detection, scoring, reporting)
  • UK Pathway – NICE NG131 recommends MRI as the first-line investigation for suspected localized prostate cancer.
  • Evidence today – PI-CAI (Lancet Oncology, 2024) and the JAMA Network Open 2025 Reader Study both show AI assistance improves detection, especially for non-expert readers.
  • Regulation matters – MHRA regulates many of these tools as medical devices with defined evidence and monitoring requirements.
  • AI does not replace clinicians – Biopsy confirmation, MDT review, and shared decision making remain the backbone of UK care.

What AI means in prostate cancer care

In healthcare, AI usually means software trained on large datasets to recognise patterns and generate outputs such as risk scores, lesion maps, or suggested measurements. In prostate cancer, AI is most commonly applied to imaging, but it is also being explored in pathology and decision-support models.

In the UK, many AI systems used for diagnosis, monitoring, or treatment planning are regulated as medical devices (or in vitro diagnostic devices). Regulation matters because it drives requirements around intended use, evidence, safety monitoring, transparency, and how updates and retraining are managed.

A practical way to think about AI in prostate cancer care is this: it can help clinicians do the technical parts faster and more consistently, but it does not replace judgement, accountability, or shared decision-making.

How AI may help with prostate MRI reporting and diagnosis

Prostate MRI has become a core part of modern diagnosis pathways. NICE recommends MRI as the first-line investigation for people with suspected clinically localised prostate cancer, reported using a 5-point Likert scale. The EAU also recommends performing MRI before prostate biopsy in men with suspected organ-confined disease.

There is room for variation

Internationally, many radiologists use the PI-RADS scoring system to standardise prostate MRI reporting and communication. The American College of Radiology describes PI-RADS as a tool intended to support high-quality prostate MRI and improve diagnosis of clinically significant prostate cancer, with the broader goal of reducing unnecessary biopsies and treatment. Even with structured reporting systems, MRI interpretation can still vary between readers and centres. This is one of the reasons AI is being developed rapidly in this space.

What AI tools can do in prostate MRI

Depending on the product and intended use, prostate MRI AI tools may:

  • Highlight areas that look suspicious for clinically significant cancer
  • Support prostate segmentation and automated measurements
  • Assist with reporting consistency by providing structured prompts or risk scoring alongside a human reader’s assessment

The evidence so far

High-quality international studies show that AI can perform well in controlled, study settings. In the PI-CAI study, an AI system was, on average, superior to radiologists using PI-RADS v2.1 for detecting clinically significant prostate cancer on MRI, and the authors emphasised that prospective studies are still needed. In a large reader study, AI assistance was associated with significant improvements in diagnostic performance, with greater benefit seen in non-expert readers. A systematic review found promising results across studies, but highlighted major limitations, including heterogeneity and a lack of external validation in many models, concluding that prospective trials and better standardisation are needed before broad routine adoption.

Important to note

It may be tempting to assume that ‘AI = faster diagnosis’. Sometimes this may be true, but it depends on the particular bottleneck. If the delay is scanner availability, staffing, reporting capacity, or biopsy scheduling, AI may only help part of the chain.

In England, the NHS has announced a pilot designed to test whether AI-supported MRI interpretation can speed up the diagnostic pathway for suspected prostate cancer with radiologist review still built into the process. This is promising, but it is still being evaluated. Large prospective studies are also underway to clarify how AI should be used safely at scale.

How AI relates to focal therapy

Focal therapy aims to treat only the part of the prostate affected by cancer, rather than treating the whole gland. In principle, better localisation and targeting should help clinicians minimise damage to surrounding healthy structures, with the aim of preserving urinary continence and erectile function.

Potential roles include assisting with:

  • Lesion localisation on MRI (supporting where to target)
  • Aligning MRI with real-time ultrasound during fusion procedures
  • Documentation and follow-up planning

It is important to note that outcomes depend on tumour biology, patient selection, imaging quality, treatment modality and operator experience, not on ‘AI’ as a magic ingredient.

Get Expert Advice & The Latest Research

Subscribe to our newsletter to receive the latest updates, expert insights, and breakthrough research on prostate cancer-delivered straight to your inbox.

    The human expert and multidisciplinary decision-making

    AI is only useful if it used well. ‘Well’ means within a safe clinical framework. That framework usually includes an experienced clinician interpreting AI outputs in context (PSA trend, MRI and biopsy findings, symptoms, overall health, life expectancy, patient preferences); multidisciplinary review for complex decisions; appropriate governance, consent and outcomes monitoring; and compliance with medical device regulation. The MHRA’s guidance makes clear that many software and AI products are regulated as medical devices, with expectations around evidence, transparency and post-market monitoring.

    From a patient’s perspective – the key reassurance is simple. Even when AI tools are used, decisions should be made by people, and you should be told what is driving those decisions so you can give fully informed consent.

    What this means for patients at The Focal Therapy Clinic

    If you are exploring treatment options, the most important step is a clear, evidence-based diagnosis and an honest discussion about what the realistic goals are for your specific cancer.

    AI-supported tools may help make MRI assessment and reporting more consistent, but they are only one part of the decision process. In the UK diagnostic pathway, MRI and biopsy findings are typically considered alongside PSA and other clinical factors, and treatment is not suitable for every patient.

    At a practical level, you may find it helpful to review:

    This article is general information and is not a substitute for personalised medical advice. If you have symptoms, a raised PSA, or a new diagnosis, discuss your results and options with a specialist team.

    FAQs

    Q: How is AI used in prostate cancer diagnosis?
    A: Most commonly, AI is used to support prostate MRI interpretation, for example by highlighting suspicious regions and providing risk scores that clinicians can consider alongside assessments. High-quality studies show AI assistance can improve detection performance, particularly for less experienced readers, but it is not a stand-alone diagnosis. AI is also being explored in pathology (analysis of biopsy slides) and risk prediction, but these applications still require careful validation and governance before widespread routine use.
    Q: Does AI replace doctors in prostate cancer care?
    A: Prostate cancer diagnosis and treatment planning require clinical judgement, shared decision-making, and if required, biopsy confirmation. AI tools are designed to support clinicians, and in NHS pilots AI outputs are still reviewed by radiologists and specialist teams.
    Q: Can AI help decide if focal therapy is suitable?
    A: AI may support imaging interpretation and lesion localisation, which can be relevant when considering focal approaches. However, suitability for focal therapy depends on tumour characteristics, whether disease is multisite, patient preference, and careful follow-up planning.

    References

    European Association of Urology. (2026). EAU Guidelines on Prostate Cancer. Available at: https://uroweb.org/guidelines/prostate-cancer

    National Institute for Health and Care Excellence. (2019). Prostate cancer: diagnosis and management (NG131, last updated 2021). Available at: https://www.nice.org.uk/guidance/ng131

    National Institute for Health and Care Excellence. (2023). Focal therapy using high-intensity focused ultrasound for localised prostate cancer (HTG667). Available at: https://www.nice.org.uk/guidance/htg667

    National Institute for Health and Care Excellence. (2023). Irreversible electroporation for treating prostate cancer (HTG688). Available at: https://www.nice.org.uk/guidance/htg688

    NHS. (2025). Tests and next steps for prostate cancer. Available at: https://www.nhs.uk/conditions/prostate-cancer/tests-and-next-steps/

    NHS England. (2025). NHS artificial intelligence (AI) trial to diagnose prostate cancer up to a month faster. Available at: https://www.england.nhs.uk/2025/10/nhs-artificial-intelligence-ai-trial-to-diagnose-prostate-cancer-up-to-a-month-faster/

    Medicines and Healthcare products Regulatory Agency. (2025). Software and artificial intelligence (AI) as a medical device. Available at: https://www.gov.uk/government/publications/software-and-artificial-intelligence-ai-as-a-medical-device/software-and-artificial-intelligence-ai-as-a-medical-device

    Saha, A., et al. (2024). Artificial intelligence and radiologists in prostate cancer detection on MRI (PI-CAI): an international, paired, non-inferiority, confirmatory study. The Lancet Oncology. Available at: https://doi.org/10.1016/s1470-2045(24)00220-1

    Twilt, J.J., et al. (2025). AI-Assisted vs Unassisted Identification of Prostate Cancer in Magnetic Resonance Images. JAMA Network Open. Available at: https://doi.org/10.1001/jamanetworkopen.2025.15672

    Molière, S., et al. (2025). A systematic review of the diagnostic accuracy of deep learning models for the automatic detection, localization, and characterization of clinically significant prostate cancer on magnetic resonance imaging. European Urology Oncology. Available at: https://doi.org/10.1016/j.euo.2024.11.001

    Wu, H., et al. (2025). Automated MRI system for clinically significant prostate cancer detection: development, validation and real-world implementation. Nature Communications. Available at: https://www.nature.com/articles/s41467-025-66593-z

    Prostate Cancer UK. (2025). Can this cutting-edge AI tool change the way we diagnose and treat prostate cancer? Available at: https://prostatecanceruk.org/about-us/news-and-views/2025/11/ai-diagnosis-and-treatment-research

    About UsBook Consultation