Surgeon Consistency and Accuracy of Endoscopic Image Analysis in the Prediction of Complete Pathological Response of Rectal Cancer to Non Operative Management
Serena Murphy, Peter Callas, Peter Cataldo
UVMMC, Burlington, Vermont, United States
Objective: Neoadjuvant therapy frequently downstages rectal cancer, making the future resection smaller and the postoperative risks lower. New evidence suggests that surgical treatment mandated at initial biopsy and before CRT may not be necessary. This is non operative management (NOM) after CRT, or a watch-and-wait method. Patients are monitored serially with endoscopy and surveillance MRI/CT to ensure no regrowth of the original tumor. But reliable methods to identify rectal cancer patients with complete responses to neoadjuvant treatments are needed to safely apply and expand NOM. Of the available tumor response assessment tools, endoscopic examination has often been considered the most robust method. However, the evaluation of endoscopic imaging is difference among different colorectal surgeons. In an effort to standardize endoscopic image gradation, there first needs to be an agreement of what surgeons observe in endoscopic images among patients treated with a "watch-and-wait"? strategy. Using a survey based system, multiple surgeons will be queried to their assessment of endoscopic images for multiple patients' treatment progression. Objective 1: Is there agreement between all colorectal surgeons surveyed on which patients have "complete clinical response"??
Design: A survey was designed with 22 patient scenarios with each patient scenario with three endoscopic pictures demonstrating their tumor from their colonoscopy/flexible sigmoidoscopy at 3 points in their treatment journey: 1) tumor at diagnosis 2) tumor at middle of therapy 3) tumor at completion chemoradiation. At the final picture, the surgeon was asked to categorize the patient response as 1) Complete clinical response, 2) Near complete clinical response or 3) Incomplete or no clinical response.This survey was sent to Colorectal Board Certified surgeons. The degree of agreement between surgeons was then measured with a Kappa-coefficient statistical analysis. Degree of certainty was measured with Spearman coefficient.
Setting: Colorectal Surgeons in community and academic centers
Patients (or other participants): Colorectal surgeons.
Interventions: None
Main Outcome Measures: Objective 1: Is there agreement between all colorectal surgeons surveyed on which patients have "complete clinical response"??
Results: A weighted kappa was used for interrater reliability; it was 0.76 (95% CI 0.64-0.86). The degree of correlation ranged from 49-10% on the 22 patient scenarios. Spearman's correlation coefficient was calculated to be 0.80 (p<0.0001) to correlate percent who gave modal rating response with mean certainty of that response.
Conclusions:Correlation among colorectal surgeons regarding the status of a tumor response to neoadjuvant chemotherapy ranged from 49-100%, with overall Spearman correlation coefficient of 0.8 (p<0.0001). This study overall reflects a reasonable degree of correlation among colorectal surgeons.
Back to 2022 Abstracts