Abstract
BACKGROUND: T4 esophageal squamous cell carcinoma (ESCC) is associated with dismal prognosis, and the optimal treatment remains uncertain. This study aimed to identify the clinical and pathological characteristics of T4 ESCC patients in China and compare the efficacy, adverse events, and treatment failure patterns of two treatment strategies: neoadjuvant chemotherapy followed by surgery and adjuvant chemoradiotherapy (nCT+S) or neoadjuvant chemotherapy followed by definitive chemoradiotherapy (nCT+dCRT). METHODS: We retrospectively analyzed the clinical records of 1242 cT4N0-3M0 ESCC patients from January 2015 to December 2020. After performing propensity score matching (PSM), a total of 195 patients were included in the analysis, comprising 73 patients in the nCT+S group and 122 in the nCT+dCRT group. To address potential selection bias related to resectability, we further performed a subgroup analysis limited to MDT-confirmed resectable T4 ESCC patients after nCT.Endpoints included overall survival (OS), progression-free survival (PFS), and related clinical outcomes. RESULTS: The median follow-up time was 77.0 months (95% CI: 68.1-85.9). For the entire post-propensity score matching (PSM) cohort, the 1-, 3-, and 5-year overall survival (OS) rates were 65.6%, 27.6%, and 17.4%, respectively, with a median OS of 18.0 months (95% CI:14.9-21.1); the corresponding 1-, 3-, and 5-year progression-free survival (PFS) rates were 49.7%, 22.6%, and 14.5%, with a median PFS of 12.0 months (95% CI:8.8-15.2).Multivariate Cox regression analysis confirmed that treatment modality (nCT+S vs nCT+dCRT) and N stage were independent prognostic factors for both OS (HR = 1.867, 95% CI:1.341-2.598;HR=1.595, 95% CI:1.286-1.977, both P<0.001) and PFS (HR = 1.576, 95%CI:1.146-2.167; HR = 1.741, 95%CI:1.286-1.977, both P<0.01).After Bonferroni correction for multiple testing (corrected α=0.0025), subgroup analysis demonstrated that patients with ECOG performance status 0-1, cT4a stage, cN0 stage, and mid-esophageal tumors derived significant OS benefits from nCT+S (all P<0.0025);for PFS, male patients, those aged >64 years, with tumor length >5.0 cm, cT4a stage, cN0 stage, and mid-esophageal tumors obtained significant benefits from nCT+S (all P<0.0025).In the subgroup restricted to MDT-assessed resectable T4 lesions after nCT, nCT+S still provided significantly superior OS and PFS versus nCT+dCRT (both P<0.0025 after Bonferroni correction).Notably, the incidence of severe treatment-related complications was significantly lower in the nCT+S group (4.1%, 3/73) than in the nCT+dCRT group (18.9%, 23/122; χ²=8.591, P = 0.003).Additionally, the rate of isolated local or regional lymph node recurrence was significantly higher in the nCT+dCRT group (χ²=7.348, P = 0.007), though the overall treatment failure rate showed no statistical difference between the two groups. CONCLUSION: In nCT-responsive, MDT-evaluated resectable or downstaged resectable cT4N0-3M0 ESCC, nCT+S achieves longer OS and PFS than nCT+dCRT, with more survival benefits in ECOG 0-1, cT4a, cN0 or mid-esophageal tumor subgroups; it also lowers severe complications and isolated local/regional recurrence risk.These results were validated in a resectable-only subgroup, minimizing selection bias related to tumor resectability.Large multicenter randomized trials are required to validate these findings and define optimal therapy for T4 ESCC due to single-center retrospective limitations.