Objective:To compare the genetic spectrums of esophageal squamous cell carcinoma (ESCC) patients with different prognosis after postoperative radiotherapy and to screen the genetic variants associated with radiotherapy resistance.Methods:A total of 32 ESCC patients who received radical surgery and postoperative adjuvant radiotherapy in Affiliated Hospital of Nantong University from January 2015 to December 2019 were selected as the study objects. According to whether there was any recurrence in the radiation field within 1 year, they were divided into a recurrence group (radiotherapy resistance group, n=16) and a stable group (radiotherapy sensitive group, n=16). Genomic DNA was extracted from patients and high-throughput sequencing was performed using whole exome sequencing (WES) technology. Biological information analysis software Trimmomatic, BWA and Picard were used to process the data and the alignment files were obtained by GATK comparison, then Vardict software was used to screen out various genetic variants from the sequencing data. The disease free survival (DFS) and overall survival (OS) were estimated by Kaplan-Meier method. Cox proportional hazard regression model was used to analyze the independent risk factors of DFS and OS of ESCC patients. Results:After quality control of the sample data, 26 patients were finally included in this study for follow-up analysis, 13 in each of the recurrence and stable groups. The median tumor mutation burden of non-silent tumors in the whole group was 0.95 mutations/Mb. The substitution types of mutant bases were mainly C>T conversion, followed by C>G transmutation. The genetic variants with the highest frequency were single nucleotide polymorphism (SNP) (75.1%), deletion mutation (13.7%) and insertion mutation (10.5%). The number of tumor-specific mutations in the recurrence group was slightly higher than that in the stable group (median mutation number was 36 and 34, respectively), and the top ten gene profiles of mutation frequency were significantly different between the two groups. In the recurrence group, 392 unique mutated genes were detected, and the top five were MUC19, NPIPA5, EPPK1, FLG and FOXG1. In the stable group, 192 unique mutation genes were detected, and the top five were TCHH, WNK1, AIM1L, COL6A5 and DPCR1. The median DFS and OS were 15.0 months (95% CI: 10.1 months-not reached) and 26.2 months (95% CI: 19.8 months-not reached) in the recurrence group respectively, and no recurrence or metastasis occurred in the stable group. Univariate analysis showed that GRIK2 ( χ2=6.81, P=0.009), MUC4 ( χ2=4.25, P=0.039), MUC5B ( χ2=4.03, P=0.045), PRRG1 ( χ2=5.15, P=0.023) gene mutations, 3p deletion ( χ2=4.16, P=0.041) and 14q deletion ( χ2=7.09, P=0.008) were correlated with DFS. FLG ( χ2=6.41, P=0.011), NPIPA5 ( χ2=4.57, P=0.033), PKD1L2 ( χ2=6.41, P=0.011), FOXG1 ( χ2=4.57, P=0.033) gene mutations, 3p deletion ( χ2=3.88, P=0.049), 14q deletion ( χ2=5.66, P=0.017) and 18p deletion ( χ2=3.85, P=0.050) were associated with OS. Multivariate analysis showed that 14q deletion ( HR=3.65, 95% CI: 1.18-11.32, P=0.025) was an independent risk factor for DFS of ESCC patients with postoperative adjuvant radiotherapy, and FLG ( HR=8.94, 95% CI: 1.52-52.74, P=0.016), NPIPA5 ( HR=6.36, 95% CI: 1.23-33.03, P=0.028) gene mutation and 14q deletion ( HR=3.82, 95% CI: 1.18-12.31, P=0.025) were independent risk factors for OS of ESCC patients with postoperative adjuvant radiotherapy. Conclusion:The WES results suggest that the types and rates of gene mutations of the ESCC patients with postoperative adjuvant radiotherapy in the recurrence and stable groups are basically the same, but the mutation spectrum of the two groups is significantly different. FLG, NPIPA5 gene mutations and 14q deletion can be used as molecular markers to predict the prognosis of ESCC patients treated with postoperative adjuvant radiotherapy.