Back to 2025 Abstracts
EX-VIVO LYMPH NODE ASSESSMENT DURING ESOPHAGEAL AND GASTRIC CANCER RESECTION USING A NOVEL HAND-HELD DIFFUSE REFLECTANCE SPECTROSCOPY PROBE
Naim Slim
*, Maxime Giot, Ria Ranjitkar, Ioannis Gkouzionis, Jo Lloyd, Priscilla Anketell, Robert Goldin, Ara Darzi, Daniel Elson, Christopher Peters
Imperial College London, London, United Kingdom
Introduction. Lymphadenectomy during esophageal and gastric cancer resection is an essential step for complete oncological resection and accurate disease staging, but the extent of lymphadenectomy required to achieve this remains unclear. Current image-guided methods with near-infrared fluorescence and indocyanine green improve overall nodal yield but discriminate poorly between benign and metastatic nodes. We investigated whether diffuse reflectance spectroscopy (DRS), a non-invasive points-based optical technique, can differentiate between benign and malignant nodal tissue with a view to providing real-time intraoperative feedback to the operating surgeon.
Methods. This was a prospective ex-vivo validation study. Patients undergoing esophageal or gastric cancer resection between March 2024 and November 2024 were prospectively included into the study. Lymph nodes were systematically sampled on freshly resected specimens with a handheld DRS probe and spectral data collected. Following correlation with histopathological ground truth, binary classification was performed using supervised machine learning classifiers, with values derived for sensitivity, specificity, and overall accuracy.
Results. Twenty-four patients, twelve of whom underwent gastrectomy and twelve esophagectomy with an average age of 63.6 were recruited to the study. A total of 7,660 mean spectra for 75 benign and 21 metastatic nodes were collected. The eXtreme Gradient Boosting (XGB) machine learning classifier achieved the best performance, with a mean overall diagnostic accuracy of 96.0%, sensitivity of 90.3% and specificity of 98.1%, for the differentiation between benign and metastatic lymph nodes.
Conclusion. Our findings demonstrate the feasibility for DRS technology for differentiation between benign and metastatic nodes in the context of esophagogastric cancer resection, and the potential for adoption of DRS in the intraoperative setting for real-time lymph node characterization.
(a) Experimental setup showing use of DRS probe in the interrogation of lymph nodes in sequential esophageal and gastric lymph node stations; (b) Indicative diffuse reflectance spectra of benign and metastatic nodes; (c) Receiver Operator Characteristic (ROC) curve illustrating performance of binary classifiers.
Back to 2025 Abstracts
