Correlation between preoperative endoscopic and intraoperative findings in localizing colorectal lesions.

Background: Colonoscopy findings compared with findings at time of surgery have a discrepancy rate of 3-21%. The objective of our study was to investigate this discrepancy and provide potential resolutions.

Methods: In this retrospective study, we identified 400 patients who underwent colonoscopy followed by colon resection at our community hospitals in 1999-2006. Discrepancies between colonoscopy and intraoperative findings were noted. Each discrepancy was classified as major if the surgical procedure had to be altered, the lesion was missed, an unnecessary segment was removed, or the incision was extended. A discrepancy was classified as minor if there was no alteration in planned surgery.

Results: Of the 400 cases, 160 (40%) were located in the right colon, 13 (3%) were in the transverse colon, 185 (46%) were in the left colon, and 42 (11%) were in the rectum. A total of 48 (12%) discrepancies between colonoscopy and intraoperative findings were identified: 26 (54%) were major and 22 (46%) were minor. Thirteen (27%) were in the proximal colon (3 major and 10 minor discrepancies), 3 (6.3%) were in the transverse colon (all major), 22 (46%) were in the distal colon (17 major and 5 minor), and 10 (21%) were in the rectum (3 major, 7 minor). Major discrepancies were significantly higher in the left colon (17 of the 185 left-sided lesions; 9.1%) than in the right colon (3/160; 1.9%; P = 0.045).

Conclusions: In our study, colonoscopy has an error rate of 12% when used to localize tumors; more than half of these patients require significant unanticipated changes in their surgery. The discrepancies are significantly higher in left side of colon.