LONDON, UK, 25 FEBRUARY 2026 – Concr, the techbio company using Bayesian AI for precision oncology, announces the completion of the VISION clinical study at The Royal Marsden NHS Foundation Trust, Europe's largest cancer centre. The 24-month retrospective observational study (NCT06409221) was conducted as part of the AI-VISION collaboration, supported by Innovate UK and run in partnership with The Institute of Cancer Research and Durham University. The trial marks a pivotal milestone in the clinical validation of FarrSight®-Twin, Concr’s proprietary Bayesian Digital Twin model designed to improve treatment stratification for patients with Triple Negative Breast Cancer (TNBC).
Why TNBC?
TNBC remains one of the most aggressive forms of breast cancer, characterised by a rapid growth rate and a high risk of developing metastases. It accounts for a significant proportion of breast cancer deaths: approximately 40% of women with stage 2 or 3 TNBC will succumb to the disease, with disproportionately high incidence observed in black women.
Clinicians currently lack reliable biomarkers to predict which patients will achieve a pathological complete response (pCR) to standard treatments. Because 50-70% of patients do not achieve a pCR, they face a higher risk of relapse and lower survival. Identifying these patients before treatment is a critical clinical priority, as it creates the opportunity to modify strategies early and avoid treatment-related toxicities without clinical benefit.
Harnessing Bayesian AI
The study used Concr’s FarrSight®-Twin model to create digital twins of individual patients by integrating their clinical information, existing biomarkers, and molecular data from routine diagnostic biopsies. Interim results showed:
The model reliably identified patients unlikely to benefit from standard neoadjuvant chemotherapy, achieving a 74% overall Negative Predictive Value (NPV) from only 66 patients.
For the anthracycline + taxane regimen, the NPV reached 91%, indicating that the technology is remarkably dependable at flagging patients who will not respond to the standard-of-care combination.
These results were achieved using routinely collected FFPE samples and limited clinical data rather than requiring new diagnostics, indicating that the technology is highly scalable for integration into standard clinical pathways.
The interim results are available in full here, and the poster presented at the recent San Antonio Breast Cancer Symposium can be viewed here.
Regulatory and Policy Alignment
The study’s methodology and “Bayesian-first” approach align with the latest regulatory and policy developments, including the FDA’s January 2026 draft guidance on Bayesian methods and the EMA’s January 2026 concept paper on the same topic. In the UK, the study exemplifies the MHRA's 2026 clinical trials reform, which commits to incorporating in silico tools for predicting treatment response, and supports the priorities outlined in the National Cancer Plan for England, published this month, which aims to revolutionise cancer care through AI-driven tools like digital twins.
A New Dawn for Precision Oncology
Dr Uzma Asghar, the Chief Scientific Officer at Concr, and the Project Lead of AI-VISION, commented: "The retrospective VISION trial marks the 'end of the beginning' for next-gen technologies like digital twins in oncology. Academia, industry, regulators, and healthcare providers are all now recognising the potential of these innovations, and VISION is crucial in pioneering this shift.
The early results show promising signals, validating FarrSight®-Twin’s ability to predict treatment response for individual patients using routinely collected FFPE samples. This provides a compelling case for the full promise of model-informed precision medicine. Moving past 'trial and error', and toward truly personalised, data-driven care for those who need it most.”
Looking Ahead
Scientific close-out for VISION will continue through Q1 2026, with the full results expected to be published later in the year.
The proprietary models behind these outcomes are already enabling drug developers to de-risk critical decisions through in silico simulations, helping design more ethical clinical trials and build better, smarter therapies. Concr invites research hospitals and healthcare providers to get in touch to explore partnership and joint funding opportunities to further validate these tools in the clinic.
About Concr
Concr is a London-based TechBio company that uses astrophysics-derived Bayesian models to predict which cancer treatments will work for individual patients. Its FarrSight® platform creates digital twins, simulations of a patient’s molecular biology, to match them with effective therapies and help drug developers design better clinical trials. Founded in 2018 by Matthew Foster, Dr Matthew Griffiths, and Dr Uzma Asghar, CEO Dr Irina Babina leads the team of eight. Concr has raised ~£4M from investors including Cambridge Enterprise, R42 Group, Oncology Ventures, and Debiopharm Innovation Fund, in addition to two Innovate UK-funded clinical studies. Concr works with partners including the NHS, Roche, and the Institute of Cancer Research.
For Media Enquiries
Yeona Choi
yeona@naracommunications.com
Back
LinkTree