Purpose Multiple prospective trials have validated the efficacy and safety of neoadjuvant immunochemotherapy (nICT) in resectable esophageal squamous cell carcinoma (ESCC), yet patients with stable disease (SD) or progressive disease (PD) following nICT have limited benefit and are at risk for suboptimal surgical and pathological outcomes. The secondary induction radiotherapy prior surgery may represent a potential therapeutic strategy in this setting.Methods We conducted a retrospective study involving 23 patients with locally advanced, resectable ESCC who demonstrated poor response to nICT (19 with SD and 4 with PD). All patients subsequently received secondary induction radiotherapy via intensity-modulated techniques or tomotherapy (median dose, 41.4 Gy). The primary endpoint was the rate of R0 resection. The secondary endpoints included pathological complete response (pCR), major pathological response (MPR), the objective response rate (ORR), the disease control rate (DCR), tumor downstaging, and treatment-related adverse events (TREAs).Results Among the 23 patients (median age, 63 years; 82.6% male), 30.4%, 60.9%, and 8.7% had clinical stages II, III, and IVa disease, respectively. Following secondary induction radiotherapy, 4 patients (17.4%) achieved a complete response (CR), and 8 (34.8%) achieved a partial response, yielding an ORR of 52.2% and a DCR of 100%. All patients underwent resection, with an R0 resection rate of 100%. MPR was achieved in 43.5% of patients, and pCR was achieved in 21.7%. Clinical downstaging occurred in 60.9% of patients, and pathological downstaging occurred in 69.6%. The median interval from the end of radiotherapy to surgery was 48 days. Grade 3-4 surgical complications occurred in 8.7% of the patients. Radiation-related grade 3-4 adverse events included myelosuppression in 13.0% and gastrointestinal symptoms in 8.7% of the patients.Conclusions In patients with resectable ESCC and a poor response to nICT, secondary induction radiotherapy was associated with encouraging surgical and pathological outcomes and an acceptable safety profile. Prospective studies are warranted to confirm these findings.