Preoperative predictive risk to cancer quality in robotic rectal cancer surgery

Background: Circumferential resection margin (CRM) involvement is widely considered the strongest predictor of local recurrence after TME. This study aimed to determine preoperative factors associated with a higher risk of pathological CRM involvement in robotic rectal cancer surgery. Methods: This was a retrospective review of a prospectively maintained database of consecutive adult patients who underwent elective, curative robotic low anterior or abdominoperineal resection with curative intent for primary rectal adenocarcinoma in a tertiary referral cancer center from March 2012 to September 2019. Pretreatment magnetic resonance imaging (MRI) reports were reviewed for all the patients. Risk factors for pathological CRM involvement were investigated using Firth's logistic regression and a predictive model based on preoperative radiological features was formulated. Results: A total of 305 patients were included, and 14 (4.6%) had CRM involvement. Multivariable logistic regression found both T3 >5 mm (OR 6.12, CI 1.35–36.44) and threatened or involved mesorectal fascia (OR 4.54, CI 1.33–17.55) on baseline MRI to be preoperative predictors of pathologic CRM positivity, while anterior location (OR 3.44, CI 0.72–33.13) was significant only on univariate analysis. The predictive model showed good discrimination (area under the receiver-operating characteristic curve >0.80) and predicted a 32% risk of positive CRM if all risk factors were present. Conclusion: Patients with pre-operatively assessed threatened radiological margin, T3 tumors with greater than 5 mm extension and anterior location are at risk for a positive CRM. The predictive model can preoperatively estimate the CRM positivity risk for each patient, allowing surgeons to tailor management to improve oncological outcomes.