Assessing Endoscopist Accuracy of Measuring Colon Polyp Size

Abstract

Purpose: Colonoscopy is the gold standard for colorectal cancer screening and prevention. When a polyp is identified prior to removal, the endoscopist will estimate the polyp size. Inaccurately sized adenomas could lead to incorrect surveillance recommendations. The primary aim of our study was to assess the accuracy between actual polyp size versus estimated polyp size of colon polyps by endoscopists. Methods: A prospective survey using polyp images ranging from 4-12 mm in size. A total of 18 images were used and half had forceps opened to help better estimate polyp size. Participants evaluated each polyp image, estimated the size, ranked their confidence in estimating polyp size and frequency of forceps use in measuring polyp size. The polyp size estimations were compared to the true polyp size for each image. Results: 130 physicians (128 GI/2 CRS) participated, 86% male with a mean age of 47 years (range 28-70 yrs). Nine polyp images without forceps had a mean absolute error between the estimate and actual polyp size of 3.65 mm. Average mean absolute error between the polyp size estimate and the actual polyp size for polyp images with forceps was 1.91 mm. The average relative difference between estimated and actual polyp size in the non-forceps group was 0.50 mm and 0.28 mm in the forceps group. Forceps improved accuracy by an average of 1.74 mm (p<0.0001). Using relative difference values, use of forceps improved accuracy of polyp size estimation by 22.2% (p<0.0001). Given an absolute 20% margin of error from the true polyp size, 20-47% subjects accurately estimated the size of the nine polyps without forceps as a reference while this range increased to 26-72% for the same nine polyps with forceps present in the images. Difficulty in assessing polyp size using forceps was noted for small polyps. Confidence of estimating polyp size was low, 18% of subjects had high confidence. Confidence increased to 49% when forceps were used as a reference. There was no direct link seen in increasing confidence and increased accuracy. Each point increase in confidence level, the accuracy in polyp size estimation increased by 0.2-3 mm (p=0.347). Each point increase in confidence level, the accuracy in polyp size with forceps was reduced by 0.19 mm (p=0.426). Subjects used forceps to estimate polyp size 40.6% of the time, only 22% used forceps >75% of the time. Subjects measured the polyp size externally after polypectomy 26.8% of cases. Conclusion: Accurately estimating polyp size during colonoscopy is a challenge for the endoscopist. Similar to other areas of endoscopy, variation amongst endoscopists exists. Our study shows the use of forceps as a measuring tool can improve polyp size accuracy and has another role beside tissue acquisition.