Individualizing surgical treatment based on tumour response following neoadjuvant therapy in T4 primary rectal cancer.

BACKGROUND Rectal cancer involving at least one adjacent organ (mrT4b) requires multi-visceral resection to achieve clear resection margin (R0). Performing pelvic compartment preservation according to the tumour response has not been considered. This study assesses the impact of changing the surgical strategy according to tumour response in rectal cancer mrT4b. METHODS Patients with non-metastatic T4b rectal cancer at two tertiary referral centres between 2008 and 2013 were grouped as "Responders" ypT0-3abNx versus "Non-responders" ypT3cd-4Nx and divided into three surgical procedures: total mesorectal excision (TME), extended-TME (eTME) and beyond-TME (b-TME). End-points were circumferential resection margin, postoperative morbidity, definitive stoma formation, 3-years local recurrence (3y-LR) and 3-years disease-free survival (3y-DFS) according to both tumours' response and surgical procedures. RESULTS Among 883 patients with rectal cancer, 101 were included. Responders had a higher rate of induction chemotherapy (59.7% vs. 38.2%; p = 0.04). Morbidity and definitive stoma formation were significantly higher in Non-responders. R0 was not impacted by either the tumour response or the surgical procedures. The 3y-LR was lower in Responders (14%) compared to Non Responders (32%) (HR 1.6; 95% CI: 1.02-2.59; p = 0.041), and was two-fold higher in e-TME compared to b-TME in Non-responders, whereas no difference was found in Responders. The 3y-DFS was higher in Responders irrespective to the surgery (71% vs. 47%; p = 0.07). CONCLUSION In Responders, TME or e-TME are technically and oncollogically feasible and should be considered in preferrence to b-TME. In Non-responders, allowing for high rates of morbidity and local recurrence in patients with e-TME, b-TME procedures should be preferred.