TAilored SToma policY after TME for rectal cancer: The TASTY approach.

Aim: Diverting stomas are routinely used in restorative surgery following total mesorectal exicision (TME) for rectal cancer to mitigate the clinical risks of anastomotic leakage (AL). However, routine diverting stomas are associated with their own complication profile and may not be required in all patients. A tailored approach based on personalized risk of AL and selective use of diverting stoma may be more appropriate. The aim of the TAilored SToma policY (TASTY) project was to design and pilot a standardized, tailored approach to diverting stoma in low rectal cancer.

Method: A mixed-methods approach was employed. Phase I externally validated the anastomotic failure observed risk score (AFORS). We compared the observed rate of AL in our cohort to the theoretical, predicted risk of the AFORS score. To identify the subset of patients who would benefit from early closure of the diverting stoma using C-reactive protein (CRP) we calculated the Youden index. Phase II designed the TASTY approach based on the results of Phase I. This was evaluated within a second prospective cohort study in patients undergoing TME for rectal cancer between April 2018 and April 2020.

Results: A total of 80 patients undergoing TME surgery for rectal cancer between 2016 and 2018 participated in the external validation of the AFORS score. The overall observed AL rate in this cohort of patients was 17.5% (n = 14). There was a positive correlation between the predicted and observed rates of AL using the AFORS score. Using ROC curves, we calculated a CRP cutoff value of 115 mg/L on postoperative day 2 for AL with a sensitivity of 86% and a negative predictive value of 96%. The TASTY approach was designed to allocate patients with a low risk AFORS score to primary anastomosis with no diverting stoma and high risk AFORS score patients to a diverting stoma, with early closure at 8-14 days, if CRP values and postoperative CT were satisfactory. The TASTY approach was piloted in 122 patients, 48 (39%) were identified as low risk (AFORS score 0-1) and 74 (61%) were considered as high risk (AFORS score 2-6). The AL rate was 10% in the low-risk cohort of patient compared to 23% in the high-risk cohort of patients, p = 0.078 The grade of Clavien-Dindo morbidity was equivalent. The incidence of major LARS was lowest in the no stoma cohort at 3 months (p = 0.014).

Conclusion: This study demonstrates the feasibility and safety of employing a selective approach to diverting stoma in patients with a low anastomosis following TME surgery for rectal cancer.