Computer-Assisted and Unassisted Optical Diagnosis of Colorectal Polyps: Results from a Prospective Pragmatic Implementation Study

Background and Aims Computer-aided diagnosis (CADx) may support optical diagnosis of diminutive (≤5 mm) colorectal polyps, replacing histopathology. We aimed to describe CADx usage and diagnostic performance under real-life, discretionary adoption. Patients and Methods We conducted a prospective implementation study enrolling patients undergoing colonoscopies at our tertiary centre (2021-2025). Patients <40y, with inflammatory bowel disease, polyposis syndromes or no diminutive polyp were excluded. For each polyp, endoscopists could choose whether to use CADx for optical diagnosis. CADx-assisted and unassisted endoscopist optical diagnosis and CADx output were documented. Main outcomes were CADx usage and sensitivity for adenomas of CADx-assisted and CADx-unassisted optical diagnosis, using histopathology as ground truth. Results 868 patients (66.1y, 47% female) with 1660 diminutive polyps (62% adenomas) were included. Endoscopists recorded CADx-assisted and CADx-unassisted optical diagnoses for 657 and 864 polyps, respectively. Sensitivity was 85.1% (78.3-90.1) and 85.1% (81.2-88.4), specificity 55.2% (45.8-64.3) and 61.1% (54.5-67.3), PPV 76.2% (69.1-82.1) and 79.3% (75.5-82.9), NPV 60.8% (52.6-68.5) and 67.1% (60.7-73.0), and accuracy 62.7% (56.5-68.4) and 66.4% (63.0-69.7), respectively. In a simulated resect-and-discard strategy, CADx assistance marginally increased the proportion of polyps for which histopathology could be avoided (84.9% vs 81.1%). For high-confidence rectosigmoid diagnoses, NPV for adenomas was 95.5% (86.9-98.5) for CADx-assisted and 93.8% (85.9-97.4) for CADx-unassisted predictions. Conclusions In a real-world descriptive implementation study, CADx-assisted and CADx-unassisted optical diagnosis showed good diagnostic performance. Assessment of the relationship between CADx use and optical diagnosis was limited by unavoidable selection bias related to discretionary use and variable endoscopist behavior despite adjusted analyses.