Induction CAPOX as part of total neoadjuvant therapy for high-risk locally advanced rectal cancer: Long-term real-world outcomes from two centers.

Abstract

e15650 Background: Recent randomised trials support total neoadjuvant therapy (TNT) for high-risk locally advanced rectal cancer (LARC); however, the optimal chemotherapy regimen and sequencing strategy remain uncertain. Induction chemotherapy administered prior to chemoradiotherapy (CRT) may improve systemic disease control, treatment compliance, dose intensification and tumour response. We report long-term outcomes from two centres using induction capecitabine-oxaliplatin (CAPOX) prior to CRT. Methods: Consecutive patients with MRI-defined high-risk non-metastatic LARC and threatened circumferential resection margin (CRM) treated at two oncology centres (first Oncology appointment Centre1:2004–2016; Centre2:2015–2018) were retrospectively reviewed via an agreed protocol. Patients were intended to receive four cycles of induction CAPOX followed by long-course CRT and total mesorectal excision or watch-and-wait if radiology and biopsy proven complete response. Patients were included irrespective of surgical approach. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan-Meier methods. Cox regression evaluated prognostic factors. Results: One hundred thirty-six patients were included (Centre1 n = 75; Centre2 n = 61), including ten patients with T4 disease considered potentially unresectable at baseline. Median follow-up was 72 and 104 months, respectively. Chemotherapy completion was 89% and CRT completion 98.5%. Five-year PFS and OS were 59.2% and 68.2% in Centre1 and 78% and 83% in Centre2. This likely reflects higher baseline risk at Centre1, with more T4 (38.7% vs 24%) and node-positive disease (94.7% vs 87%). Pathological complete response or sustained clinical complete response (watch-and-wait) occurred in 12% and 33%, respectively. Tumour and nodal downstaging occurred in 46.7% and 70.7%. On univariate analysis, R1 resection (HR9.9,95%CI1.12–87.5,p < 0.0001) and extramural venous invasion (HR4.59,95%CI1.5–14.04,p = 0.0006) were associated with inferior PFS. Conversion from node-positive to node-negative status was associated with improved PFS (HR3.25,95%CI1.36–7.76,p = 0.0014). On multivariable analysis adjusting for age, post-treatment nodal positivity remained independently associated with inferior PFS (HR6.02,95%CI1.09–33.39,p = 0.04). Conclusions: In this real-world cohort with long-term follow-up, induction CAPOX as part of TNT was associated with survival and response outcomes comparable to contemporary TNT studies. Post-treatment nodal response appears to be a dominant prognostic biomarker and may inform future risk-adapted sequencing, escalation and de-escalation strategies. This approach may be particularly relevant for patients unsuitable for intensive triplet chemotherapy regimens.