ObjectiveTo investigate whether radiation-induced lymphopenia (RIL) affects survival and identify the predictors of RIL in postoperative esophageal cancer.Materials and methodsPost hoc analysis was conducted on data from 116 patients with esophageal cancer from a randomized controlled trial comparing adjuvant therapy with surgery alone. Doses of 54 Gy in 27 fractions was delivered in the postoperative radiotherapy (PORT) group and 50.4 Gy in 28 fractions combined with chemotherapy was delivered in postoperative concurrent chemoradiotherapy (POCRT) group. Blood counts were obtained before, during, and at first follow-up after treatment. Lymphopenia was graded per version 4.03 of the Common Terminology Criteria for Adverse Events. Disease-free survival (DFS) and overall survival (OS) were analyzed using the Kaplan-Meier method, and compared between groups using the log-rank test. Receiver operating characteristic curves identified thresholds for preventing grade 4 (G4) lymphopenia.ResultsMedian follow-up duration was 56.0 months. During treatment, 16 patients (13.8%) had G4 lymphopenia. All cases of G4 lymphopenia occurred in group PORT (30.2% vs 0.0%, p<0.001). Baseline absolute lymphocyte count was comparable between G1-3 and G4 patients (2.0 ± 0.8 *109/L vs 1.7 ± 0.5 *109/L; p=0.101). The 3-year DFS was significantly lower in group G4 lymphopenia than that in group G1-3 (31.3% vs 57.6%, p=0.036). The 3-year OS was comparable between both groups (50.0% vs 66.5%, p=0.095). Logistic regression analysis revealed that exposed more thoracic marrow (TM V20 ≥75%; TVB V20 ≥71%), heart (V15 ≥40%) and PTV (volume ≥507 ml) were associated with G4 lymphopenia (p<0.05).ConclusionsG4 RIL had poor disease-free survival, which may be related to more dose exposure of thoracic marrow and heart due to larger PTV. Reasonably reducing the radiation field combined with concurrent chemotherapy, or radiation dose constraints for these normal tissues may be sufficient to decrease the incidence of G4 lymphopenia, but further prospective trials are needed to verify the results.Clinical Trial Registrationclinicaltrials.gov, identifier NCT02279134