NHS hospitals to test an AI biopsy tool for prostate cancer decision-making
An NHS-backed study will evaluate an AI tool that reads prostate biopsy slides and helps clinicians decide who needs treatment, what kind of treatment, and who can be safely monitored. The £1.9m, three-year Vanguard Path study, funded by Prostate Cancer UK and led by the University of Oxford, will analyse more than 4,000 biopsy samples using the ArteraAI Prostate Biopsy Assay.
The goal is simple: reduce under-treatment and over-treatment by giving teams a clearer, evidence-based signal at the point of decision.
What the tool does
The assay analyses digitised biopsy images and generates a personalised risk score. In clinical trials, it identified which men with high-risk prostate cancer were likely to benefit from adding abiraterone. US studies suggest it may also indicate which men with less aggressive disease benefit from hormone therapy alongside radiotherapy, and who can be monitored without immediate intervention.
The tool is already in broad use in the US. This trial tests whether those results hold in the NHS and whether the readout meaningfully shapes clinical decisions.
Study design and sites
Phase 1 will retrospectively test UK biopsy samples from men with known outcomes to validate predictive performance. Phase 2 will run prospectively at three sites as part of routine care: North Bristol NHS Trust, Oxford University Hospitals NHS Foundation Trust, and NHS Greater Glasgow and Clyde.
Clinicians will make standard treatment decisions, receive the AI readout in parallel, and record whether it would have changed their recommendation. The study will also track any impact on time from diagnosis to treatment choice.
Why this matters for clinical teams
Better selection for intensification (for example, abiraterone or adding hormones to radiotherapy) and clearer confidence about who can stay on active surveillance reduces harm and waste. If validated, one readout could support decisions across the aggressiveness spectrum-useful for MDTs that need consistent, reproducible inputs under time pressure.
Practical considerations for services
- Digital pathology readiness: Confirm whole-slide imaging capacity, file formats, and LIS/VNA integration for rapid turnaround and audit trails.
- MDT workflow: Pre-specify how the risk score is interpreted, thresholds for treatment escalation/de-escalation, and who signs off when scores conflict with clinical judgement.
- Patient consent and communication: Prepare simple explanations of what the assay does and does not do; support shared decision-making.
- Data governance: Validate image quality standards, ensure UK data coverage, and check performance across ethnic and socioeconomic groups.
- Metrics to track: Concordance between initial and final decisions, time-to-decision, treatment changes, adverse events, QoL, and downstream resource use.
Evidence and limits
Signals for treatment benefit are promising, including abiraterone in high-risk disease and hormone therapy decisions alongside radiotherapy. But as Prof Gerhardt Attard noted, approaches that work in trials can face real-world hurdles-workflows, data quality, and population differences can shift performance.
Adoption will depend on three things: scientific validity in UK cohorts, measurable impact on decisions and outcomes, and cost-effectiveness. That last point will be critical for commissioning.
Timeline and what to prepare now
The study is expected to run for three years. If results are positive, expect follow-on work on commissioning and implementation. Services can prepare by auditing current decision timelines, overtreatment/undertreatment patterns, and digital pathology throughput to establish a baseline.
- Map biopsy digitisation capacity and turnaround times.
- Define roles for reviewing AI outputs (pathology vs oncology) and escalation rules.
- Draft SOPs covering when to follow or override the assay readout.
- Engage ICS leads and finance early to scope a cost-effectiveness case if the data land well.
Key organisations
- Funded by Prostate Cancer UK.
- Led by researchers at the University of Oxford.
- Trial sites: North Bristol NHS Trust, Oxford University Hospitals NHS Foundation Trust, and NHS Greater Glasgow and Clyde.
Policy context
Senior figures have welcomed the trial as part of a broader push to modernise NHS cancer services. The study will test whether AI can help diagnose earlier, treat more effectively, and improve patient experience without adding friction to already stretched pathways.
Useful resources
Optional: building AI skills in your team
If your service is planning AI pilots or needs a baseline of digital literacy across roles, see curated options by role at Complete AI Training.
Your membership also unlocks:
