REAL-TIME CLASSIFICATION OF TUMOUR AND NON-TUMOUR COLORECTAL TISSUE USING DIFFUSE REFLECTANCE SPECTROSCOPY TO AID RESECTION MARGIN ASSESSMENT
Scarlet Nazarian*, Ioannis Gkouzionis, Ara Darzi, Nisha Patel, Daniel Elson, Christopher J. Peters
Imperial College London, London, London, United Kingdom
Background
Colorectal cancer is the 4th most commonly diagnosed malignancy and the 3rd leading cause of mortality worldwide. A positive resection margin following surgery for colon cancer is linked with higher rates of recurrence and worse survival. The aim of this study was to use a developed diffuse reflectance spectroscopy (DRS) probe and tracking system to distinguish cancer and non-cancer colorectal tissue live on-screen intraoperatively in order to aid margin assessment.
Methods
Patients undergoing elective colorectal cancer resection surgery at a tertiary hospital in London were prospectively recruited between April 2021 and July 2022. A hand-held DRS probe was used on the surface of freshly resected ex-vivo colorectal tissue. Spectral data was acquired for normal and cancerous tissue. Binary classification was achieved using supervised machine learning classifiers, which were evaluated in terms of sensitivity, specificity, accuracy and the area under the curve.
Results
A total of 2702 mean spectra were obtained for normal and cancerous colorectal tissue. The Light Gradient Boosting Model was the best performing machine learning algorithm for differentiating normal and cancerous colorectal tissue, with an overall diagnostic accuracy of 90.7% and area under the curve of 96.7%. Live on-screen classification of tissue type was achieved using a graduated colourmap.
Conclusion
Real-time classification of tissue type was achieved using a DRS system, with high diagnostic accuracy, allowing differentiation of cancerous and normal colorectal tissue. This is a promising step towards an in-vivo classification system that is able to aid surgeons with accurate resection margin assessment for colorectal cancer intra-operatively.
Back to 2023 Abstracts