Picture a surgeon in an operating theatre, performing a delicate procedure to remove a large rectal polyp. The question at the heart of the operation is simple yet deeply consequential: is this tissue cancerous? An accurate assessment means that the patient receives the treatment they need. Getting it wrong could mean sending a patient home with undetected cancer or committing them to unnecessarily invasive surgery.
Such decisions are part of the daily reality colorectal cancer surgeons face. Colorectal cancer is the third most common cancer globally and the second leading cause of cancer deaths, claiming nearly one million lives each year. Current methods for characterising these polyps before excision, including biopsies, can be unreliable for up to half of such lesions. Yet pre-operative characterisation shapes the choice of endoscopic technique, determines the dissection plane, and influences the patient’s overall management.
Through the collaborative efforts of researchers in the EU-funded CLASSICA project, led by Professor of Surgery Ronan Cahill, explainable AI is being used to help detect cancer in large rectal polyps during surgery. The project is coordinated in Ireland by University College Dublin, with Dublin-based Pintail providing project management and communication support.
From research prototype to operating theatre reality
CLASSICA represents a transformation in how surgeons make critical decisions during cancer operations. The project is developing breakthrough AI technology and deploying it as a practical tool in operating theatres across six European countries.
Tumour tissue forms blood vessels differently from healthy tissue – a process known as abnormal angiogenesis, which is a hallmark of cancer. In CLASSICA, researchers use a special fluorescent dye called indocyanine green and near-infrared light (NIR) to reveal these differences during surgery. NIR is a safe type of light just beyond the visible spectrum that can penetrate several millimetres into tissue. When the dye is illuminated, cancerous and healthy tissues emit light in distinct ways that change over time as blood flows through them.
CLASSICA’s AI system, developed by technology partner Arctur, analyses these subtle patterns frame by frame and tracks how blood flows through each part of the tissue. Within moments, surgeons receive classification information that helps guide their next move: does this polyp contain cancer, and if so, where exactly is it located?
Building confidence through clinical validation
CLASSICA encompasses a rigorous multi-country clinical trial involving around 500 patients across leading cancer surgery centres: Mater Misericordiae University Hospital and Beaumont Hospital in Ireland; Ziekenhuis Oost-Limburg in Belgium; Krankenhaus der Barmherzigen Brüder Graz in Austria; Università degli Studi di Torino in Italy; Aretaieion University Hospital in Athens, Greece; and Amsterdam UMC in the Netherlands. This real-world validation goes beyond testing whether AI can correctly identify cancer, though accuracy is critical. It also examines how the system performs across different countries, hospitals and patient populations.
'It's incredible to see this project move from theory to actual deployment. It's inspiring to see the partners bring their expertise together so impactfully to make for better cancer surgery and patient care’, explains project coordinator Cahill from University College Dublin and Mater Hospital. 'Personalised surgical stratification sounds great but is now becoming a reality through this broad consortium of many talents and great expertise.’
Surgical practice has historically varied between hospitals, shaped by differences in equipment, techniques and local experience. Today, modern training and evidence-based guidelines are key to achieving consistency and safety across surgical practice. In CLASSICA, IRCAD – the world’s leading surgical training institution – plays a central role in developing specialised training that helps surgeons learn to use CLASSICA tools effectively. The European Association for Endoscopic Surgery Guidelines Committee will provide guidance on using CLASSICA in the operating theatre.
Evolving legal and ethical standards for surgical AI
AI is transforming surgery, but innovation must move hand in hand with ethics and the law. Within CLASSICA, the University of Copenhagen’s CeBIL group and the University of Illinois College of Law lead pioneering work on the legal, ethical and liability challenges of AI-assisted surgery.
Working across Europe and the United States, the teams have analysed how regulations – from the EU Medical Device Regulation and the GDPR to the forthcoming AI Act – apply to high-risk surgical AI systems. CLASSICA’s legal experts are helping shape the global conversation on the responsible use of surgical AI.
Discover more at https://classicaproject.eu/