Accuracy of Size Measurement of Large (≥ 2 cm) Colorectal Laterally Spreading Tumours: A Prospective Video-Based...

Size estimation of large (≥ 20 mm) colorectal laterally spreading tumors (LSTs) is important in endoscopic decision making and determining risk of incomplete resection and adverse events. Data is lacking on endoscopists' accuracy in estimating LST size.

We conducted a prospective multicenter video-based study. 46 LSTs were video recorded endoscopically, resected en-bloc and pinned for high-precision measurement prior to formalin fixation. An international group of 24 endoscopists were presented with LST videos and provided an LST size estimate within 10 mm categories (20-29 mm, 30-39 mm, etc.). Correct measurements were defined as an assigned 10 mm category containing the true LST size. Primary outcome was accuracy of endoscopists' LST size estimates, defined as the proportion of correct measurements.

1104 measurements were performed with an overall accuracy of 27.2% (95% CI 22.1-32.9). 55.6% (95% CI 46.5-64.4) were underestimated and 17.2% (95% CI 12.3-23.6) were overestimated. Accuracy for LSTs < 40 mm, ≥ 40 mm, ≥ 50 mm, and ≥ 60 mm was, respectively, 42.2, 19.2, 13.1, and 7.7%. The proportion and magnitude of size underestimation increased with polyp size. Granular mixed LSTs were the most underestimated in size. There were no clinically significant differences in accuracy when stratified by endoscopist sex, ESD volume, and location.

Visual size estimation of large colorectal LSTs is highly inaccurate, with considerable underestimation. This may result in misjudging the projected resection time, incomplete resection risk, and adverse event risk. Tools to assist endoscopists in measuring large LSTs are needed.