Background
Previous studies recommend a resect and discard strategy for small, non‐cancerous colorectal polyps without histological examination, serving to reduce procedural time and screening costs. As surveillance protocols depend on the type of identified polyps, such an approach necessitates adequate endoscopic identification skills.
Methods
We assessed the applicability of this approach through a retrospective study in a UK secondary‐care center. We reviewed the accuracy of identification of colorectal polyps by comparing the endoscopic Kudo polyp evaluation from the procedure reports to the subsequent histological outcome.
Results
In 3,060 colonoscopies, 2,487 polyps were detected and 1,041 were included in the study and reviewed against their histology. Polyps were correctly classified as adenomas and hyperplastic in 84.1% and 66.4% respectively. With regards to adenomas, the likelihood of correct endoscopic diagnosis of Kudo IIIL (92.47%) and IV (93.94%) was significantly higher compared to the Kudo IIIS polyps (74.57%), P < 0.001. Correct identification rate was more likely for expert endoscopists (P = 0.04), whereas there was no statistical difference between different specialty groups (P = 0.59), endoscopists’ grade (P = 0.93), buscopan use (P = 0.83) and quality of bowel preparation (P = 0.54). Regarding hyperplastic polyps, the only factor associated with higher likelihood of correct classification was the timing of the procedure (P = 0.003). Fifteen per cent of the misclassified hyperplastic polyps were sessile serrated lesions.
Conclusion
Discard strategy could be feasible for sub‐centimetre polyps of Kudo type IIIL and IV, for endoscopists achieving a recommended ≥ 90% agreement with the histopathology findings. We suggest that all Kudo II and IIIS polyps are sent for histological assessment.