Prognosis of curative surgery versus radical radiotherapy after neoadjuvant chemotherapy with ICIs for esophageal squamous cell carcinoma: a real world study

The individualized local treatment selection after neoadjuvant chemoimmunotherapy remains an unresolved clinical issue and requires validated risk-stratification tools, such as the prognostic comparison between curative resection and radical radiotherapy as subsequent local treatments for locally advanced esophageal squamous cell carcinoma (ESCC) remains unreported after neoadjuvant paclitaxel plus carboplatin combined with immune checkpoint inhibitors (ICIs). A retrospective cohort study was conducted on locally advanced ESCC patients treated at the People’s Hospital of Guangxi Zhuang Autonomous Region from June 2020 to December 2025. The patients were divided into surgery group and radical radiotherapy group after neoadjuvant paclitaxel plus carboplatin combined with ICIs. Recurrence-free survival (RFS), overall survival (OS), and adverse events were compared among patients who received curative resection and radiotherapy. The risk factors of poor prognosis were analyzed. A total of 77 patients were enrolled, including 26 in the curative resection group and 51 in the radical radiotherapy group, with a median follow-up of 33.9 months (IQR 28.1–39.7 months). 47 patients experienced disease recurrence (42.3% in the surgery group vs. 70.6% in the radiotherapy group). Correspondingly, the 1-, 2-, and 3-year RFS rates were 73.1%, 65.4%, 57.7% and 49.0%, 35.3%, 29.4%, with median RFS and OS of 27.4 months vs. 11.9 months (HR = 2.07, 95%CI 1.15–3.72, P = 0.030) and not reached vs. 29.4 months (HR = 3.69, 95%CI 1.72–7.92, P = 0.012). These difference remained significant after propensity score matching as RFS (HR 3.32; 95%CI 1.41–7.80; P = 0.004) and OS (HR 4.96; 95%CI 1.38–17.86; P = 0.007). The most common complication was postoperative pneumonia in the surgery group and radiation esophagitis in the radiotherapy group. AEs were mainly grade 1 or 2 in both groups, with no treatment-related deaths reported. Curative resection was associated with better RFS and OS in this selected real-world cohort, while causal superiority over radical radiotherapy requires prospective validation.