Clinical-radiomics nomogram for the risk prediction of esophageal fistula in patients with esophageal squamous cell carcinoma treated with intensity-modulated radiation therapy or volumetric-modulated arc therapy

• Published in: Journal of thoracic disease Vol. 16; no. 3; pp. 2032 - 2048
• Main Authors: Li, Zhaohui, Gong, Jie, Shi, Liu, Li, Jie, Yang, Zhi, Chai, Guangjin, Lv, Bo, Xiang, Geng, Wang, Bin, Carr, Shamus R, Fiorelli, Alfonso, Shi, Mei, Zhao, Yilin, Zhao, Lina
• Format: Journal Article
• Language: English
• Published: China 29.03.2024
• Subjects: radiotherapy; esophageal fistula (EF); clinical-radiomics nomogram; risk factor; Esophageal squamous cell carcinoma (ESCC)
• ISSN: 2072-1439; 2077-6624
• DOI: 10.21037/jtd-24-191

Esophageal fistula (EF) is a serious adverse event as a result of radiotherapy in patients with esophageal cancer (EC). We aimed to identify the predictive factors and establish a prediction model of EF in patients with esophageal squamous cell carcinoma (ESCC) who underwent intensity-modulated radiation therapy (IMRT) or volumetric-modulated arc therapy (VMAT). Patients with ESCC treated with IMRT or VMAT from January 2013 to December 2020 at Xijing Hospital were retrospectively analyzed. Ultimately, 43 patients with EF and 129 patients without EF were included in the analysis and propensity-score matched in a 1:3 ratio. The clinical characteristics and radiomics features were extracted. Univariate and multivariate stepwise logistic regression analyses were used to determine the risk factors associated with EF. The median follow-up time was 24.0 months (range, 1.3-104.9 months), and the median overall survival (OS) was 13.1 months in patients with EF. A total of 1,158 radiomics features were extracted, and eight radiomics features were selected for inclusion into a model for predicting EF, with an area under the receiver operating characteristic curve (AUC) value of 0.794. Multivariate analysis showed that tumor length, tumor volume, T stage, lymphocyte rate (LR), and grade IV esophagus stenosis were related to EF, and the AUC value of clinical model for predicting EF was 0.849. The clinical-radiomics model had the best performance in predicting EF with an AUC value of 0.896. The clinical-radiomics nomogram can predict the risk of EF in ESCC patients and is helpful for the individualized treatment of EC.