Panakeia publishes landmark clinical validation showing AI can deliver rapid molecular profiling for colorectal cancer
Panakeia publishes landmark clinical validation showing AI can deliver rapid molecular profiling for colorectal cancer
- Largest real-world, blinded, multi-site validation of AI molecular profiling from routine diagnostic images
- Software identifies clinically actionable cancer biomarkers in minutes rather than weeks
- Study led by University of Leeds and Leeds Teaching Hospitals NHS Trust, with validation across three independent healthcare institutions
- Findings published in npj Digital Medicine, support faster, more accessible precision medicine for colorectal cancer patients
Cambridge, UK, 29th January — Panakeia, the world's first in silico multi-omics company, today announced the publication of a landmark clinical validation study in npj Digital Medicine demonstrating that its PANProfiler Colorectal (MSI/MMR) software can accurately deliver molecular profiling for colorectal cancer directly from routine brightfield images already generated in hospital care [1].
Led by the University of Leeds and Leeds Teaching Hospitals NHS Trust, the study represents the largest real-world, blinded, multi-site validation to date of AI-enabled molecular profiling from routine diagnostic images. The findings show that PANProfiler Colorectal (MSI/MMR) can identify key molecular biomarkers that inform critical treatment decision-making in minutes rather than weeks, without requiring additional laboratory testing or changes to existing hospital workflows.
Robust, real-world clinical validation across healthcare systems
The blinded, multi-centred evaluation included 3,576 images of routine tissue samples from 1,243 colorectal cancer patients across three independent healthcare institutions in the United Kingdom: University of Leeds / Leeds Teaching Hospitals NHS Trust, Queen's University Belfast / Belfast Health and Social Care Trust, and the University of St Andrews.
Across this diverse, real-world dataset, PANProfiler Colorectal (MSI/MMR) successfully returned results on 86.55% of cases (approximately 86 out of every 100 patients), the highest test replacement rate achieved to date among AI solutions for molecular profiling from tissue images, marking a significant scientific breakthrough. The software achieved an overall agreement of 93.83% with standard laboratory testing for microsatellite instability (MSI) and mismatch repair deficiency (MMR)—molecular biomarkers that directly inform treatment decision-making for colorectal cancer patients. It correctly identified 92.54% of cases where the biomarker was present (MSI-H/dMMR) and 94.02% of cases where it was absent (MSI-stable).
The study was conducted under rigorous blinded conditions across sites with differing patient populations, diagnostic protocols, tissue preparation methods, and scanner types, reflecting the variability encountered in routine clinical practice. This comprehensive validation demonstrates that PANProfiler Colorectal (MSI/MMR) delivers consistent, reliable performance across heterogeneous healthcare environments, an essential requirement for real-world clinical deployment.
Addressing a global gap in access to molecular profiling
Laboratory-based molecular profiling (wet-lab testing) is standard of care for patients diagnosed with colorectal cancer, as mismatch repair status (MSI/MMR) directly informs treatment decision-making, including confirmation, immunotherapy eligibility, inherited cancer risk assessment, and prognosis [2,3,4].
However, significant gaps persist between guideline recommendations and real-world practice. National audits and patient surveys show that a substantial proportion of eligible patients do not receive molecular biomarker testing, or experience delays of several weeks due to laboratory capacity constraints and fragmented workflows [5,6]. During these delays, clinical decisions may be made with incomplete molecular information.
PANProfiler Colorectal (MSI/MMR) is designed to address these challenges by delivering mismatch repair deficiency status directly from routine brightfield images already generated during standard diagnostic processes. By providing results in minutes rather than weeks, the software supports faster, more informed clinical decision-making, enables earlier triage for confirmatory testing where required, and helps healthcare teams act on molecular insights earlier in the care pathway, without adding burden to laboratory infrastructure.
Best-in-class performance for AI-enabled molecular profiling
Based on publicly available performance data from other AI algorithms, PANProfiler Colorectal (MSI/MMR) represents best-in-class performance for AI-enabled molecular profiling in colorectal cancer. The software achieves the highest test replacement rate demonstrated to date while maintaining high agreement with standard testing methods. This combination of broad applicability and accuracy is essential for supporting clinical decision-making at scale in routine care settings.
Importantly, PANProfiler Colorectal (MSI/MMR) integrates into existing hospital workflows by leveraging data already generated during standard of care processes. This combination of accuracy, speed, and workflow compatibility differentiates the approach from traditional molecular testing strategies that rely on additional assays, specialist laboratories, and extended turnaround times.
Dr Nic Orsi, Associate Professor in Histopathology and Industrial Innovation at St James's University Hospital and Leeds Teaching Hospitals NHS Trust, said: "This study was designed to rigorously and independently assess whether AI-based MSI/MMR testing can perform reliably across real-world clinical data from multiple centres. By analysing thousands of cases under blinded conditions, we were able to demonstrate that PANProfiler Colorectal (MSI/MMR) delivers consistent performance comparable to standard laboratory testing, using only routine histology images.
"These findings highlight the potential for AI tools like this to support more timely and accessible molecular testing, which is increasingly critical for guiding treatment decisions in colorectal cancer."
Platform-level validation and extensibility across cancer types
The colorectal cancer validation builds on prior platform-level evidence for Panakeia's PANProfiler technology. In 2024, Panakeia published a study in Nature Communications Medicine, recognised as one of the journal's top 25 papers of the year, demonstrating the platform's ability to identify thousands of biomarkers across more than 30 cancer indications from routine diagnostic images [7].
Independent validation of PANProfiler Breast, also conducted in collaboration with the University of Leeds and published in Clinical Breast Cancer, further supports the generalisability of the approach across tumour types and clinical contexts [8]. These studies provide a growing body of peer-reviewed evidence supporting PANProfiler as a scalable foundation for molecular profiling across oncology.
Shaping the future of safe, accessible AI in healthcare
Building on this expanding evidence base, Panakeia is actively engaged with regulators and ecosystem leaders shaping how AI-enabled medical technologies are evaluated and deployed in healthcare. PANProfiler Colorectal (MSI/MMR) has been selected for the UK Medicines and Healthcare products Regulatory Agency's (MHRA) AI Airlock programme, which supports real-world evaluation of AI medical devices [9]. Panakeia has also participated in the Friends of Cancer Research Digital PATH Project, a multi-stakeholder initiative examining agreement and reliability across AI-based digital pathology platforms [10].
These engagements reflect growing confidence from regulators and influential ecosystem organisations in PANProfiler's scientific rigour and clinical relevance, positioning Panakeia as a key player in helping to define evidence standards for safe, ethical AI that can support clinical decision-making and expand access to precision diagnostics for patients [9,10].
Pahini Pandya, Founder and Chief Executive Officer of Panakeia, commented: "This npj Digital Medicine publication validates our vision to make precision medicine accessible to patients and doctors globally. Our biology-first approach has enabled us to build software that delivers robust molecular insights from data already generated in care. This rigorous multi-site validation, combined with PANProfiler Colorectal (MSI/MMR) being UKCA-marked and deployed in hospitals, establishes it as a clinically validated solution that supports faster, more informed clinical decision-making while expanding access to precision diagnostics."
Looking ahead: Pioneering the future of precision medicine
This landmark study demonstrates what's possible when AI meets real-world clinical care. Panakeia has achieved real-world, multi-site validation for AI-enabled molecular profiling in colorectal cancer. PANProfiler Colorectal (MSI/MMR) is now UKCA-marked and ready to scale across multiple sites in the UK, with international deployment plans advancing rapidly. This is precision medicine moving from aspiration to reality across entire healthcare systems.
PANProfiler addresses critical issues affecting health systems globally, including diagnostic delays that compromise cancer pathway performance. It provides comprehensive molecular insights to support same-day clinical decision-making. Starting with deployment within the UK's National Health Service (NHS), Panakeia is building a scalable solution for healthcare systems worldwide. The NHS and UK government's commitment to become a global leader in AI-enabled healthcare makes it the ideal environment to validate technology that can get patients the right treatment sooner, improve outcomes, and ultimately save lives.
- Ends -
About PANProfiler Colorectal (MSI/MMR)
PANProfiler Colorectal (MSI/MMR) is an AI software that determines mismatch repair deficiency status (MSI/MMR) directly from H&E-stained images of colorectal cancer tissue samples, entirely digitally. The software delivers results in minutes, enabling same-day clinical decision-making without additional laboratory testing. PANProfiler Colorectal (MSI/MMR) is UKCA-marked and deployed in clinical sites across the UK.
About the PANProfiler Platform
The PANProfiler platform is a software platform powered by Panakeia's proprietary AI technology that determines multi-omic biomarker status across DNA, RNA, proteins, and metabolites directly from routine brightfield images of cells and tissues. Built using a biology-first approach with AI trained on petabytes of population-scale data, the platform can profile thousands of biomarkers in more than 30 cancer types. PANProfiler supports drug discovery, clinical development, and patient care—making comprehensive molecular profiling accessible, affordable, and rapid.
About Panakeia
Panakeia is the world's first in silico multi-omics company, building AI solutions to uncover complete biology at scale and accelerate precision medicine research, drug discovery, development and clinical decision-making.
Panakeia's PANProfiler platform uses a biology-first AI approach to determine biomarker status across DNA, RNA, proteins, and metabolites directly from routinely used brightfield images of cells and tissues. The PANProfiler platform has been independently validated in peer-reviewed publications, including Nature Communications Medicine, where it was recognised as one of the journal's top 25 papers of 2024. Two UKCA-marked clinical products derived from the platform are deployed in hospitals, enabling faster and more accessible precision diagnostics.
To learn more about Panakeia's platform visit: https://www.panakeia.ai/
For more information please contact:
Stella Lempidaki, PhD
Head of Communications
contact@panakeia.ai
References
[1] H&E-based MSI/MMR testing with AI in colorectal cancer: a multi-centred blinded evaluation, npj Digital Medicine, 2025, https://doi.org/10.1038/s41746-025-02218-5
[2] Molecular testing strategies for Lynch syndrome in people with colorectal cancer, NICE Diagnostics Guidance [DG27], 2017, https://www.nice.org.uk/guidance/dg27
[3] Mismatch repair and microsatellite instability testing for immune checkpoint inhibitor therapy: ASCO Endorsement of College of American Pathologists Guideline, Journal of Clinical Oncology, 2023, https://doi.org/10.1200/JCO.22.02462
[4] Immunotherapy for dMMR/MSI-H metastatic colorectal cancer, NICE Technology Appraisal Guidance [TA716], 2024, https://www.nice.org.uk/guidance/ta716
[5] National Bowel Cancer Audit: State of the Nation Report, National Cancer Audit Collaborating Centre, Royal College of Surgeons of England, 2024, https://www.nboca.org.uk/reports/annual-report-2024/
[6] Access to treatment-guiding biomarkers in the UK: surveys of cancer patients and healthcare professionals, Future Oncology, 2025, https://doi.org/10.1080/14796694.2025.2535098
[7] A systematic pan-cancer study on deep learning-based prediction of multi-omic biomarkers from routine pathology images, Nature Communications Medicine, 2024, https://doi.org/10.1038/s43856-024-00476-8 (Recognised as one of the journal's top 25 papers of 2024)
[8] Deep learning-based biomarker assessment in breast cancer from routine histology, Clinical Breast Cancer, 2025, https://www.clinical-breast-cancer.com/article/S1526-8209(25)00168-5/fulltext
[9] MHRA AI Airlock Programme – Phase 2 cohort, UK Medicines and Healthcare products Regulatory Agency (MHRA), https://www.gov.uk/government/publications/ai-airlock-phase-2-cohort
[10] Digital PATH Project: Agreement Across AI-based Digital and Computational Pathology Platforms, Modern Pathology, 2025, https://doi.org/10.1016/j.modpat.2025.100944
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
