Purpose To construct machine learning models for predicting progression free survival (PFS) and overall survival (OS) with esophageal squamous cell carcinoma (ESCC) patients. Methods 204 ESCC patients were randomly divided into training cohort (n = 143) and test cohort (n = 61) according to the ratio of 7:3. Two radiomics models were constructed by radiomics features, which were selected by LASSO Cox model to predict PFS and OS, respectively. Clinical features were selected by univariate and multivariate Cox proportional hazards model (p < 0.05). Combined radiomics and clinical model was developed by selected clinical and radiomics features. The receiver operating characteristic curve, Kaplan Meier curve and nomogram were used to display the capability of constructed models. Results There were 944 radiomics features extracted based on volume of interest in CT images. There were six radiomics features and seven clinical features for PFS prediction and three radiomics features and three clinical features for OS prediction; The radiomics models showed general performance in training cohort and test cohort for prediction for prediction PFS (AUC, 0.664, 0.676. C-index, 0.65, 0.64) and OS (AUC, 0.634, 0.646.C-index, 0.64, 0.65). The combined models displayed high performance in training cohort and test cohort for prediction PFS (AUC, 0.856, 0.833. C-index, 0.81, 0.79) and OS (AUC, 0.742, 0.768. C-index, 0.72, 0.71). Conclusion We developed combined radiomics and clinical machine learning models with better performance than radiomics or clinical alone, which were used to accurate predict 3 years PFS and OS of non-surgical ESCC patients. The prediction results could provide a reference for clinical decision.

ObjectiveThis study aimed to compare the efficacy and safety of induction chemotherapy followed by concurrent chemoradiotherapy (I-CCRT), induction chemotherapy followed by concurrent chemoradiotherapy and consolidation chemotherapy (I-CCRT-C), and concurrent chemoradiotherapy followed by consolidation chemotherapy (CCRT-C) for locally advanced esophageal squamous cell carcinoma (ESSC).Patients and MethodsPatients with locally advanced ESCC who underwent definitive chemoradiotherapy with cisplatin plus fluorouracil or docetaxel from February 2012 to December 2018 were retrospectively reviewed. Kaplan–Meier curve was used to estimate survival. Efficacy was assessed using RECIST, version 1.0. Prognosis factors were identified with Cox regression analysis.ResultsPatients were treated with CCRT-C (n = 59), I-CCRT (n = 20), and I-CCRT-C (n = 48). The median follow-up duration was 73.9 months for the entire cohort. The ORR of the CCRT-C, I-CCRT, and I-CCRT-C groups was 89.8%, 70.0%, and 77.1%, respectively (p = 0.078). The median PFS in the CCRT-C, I-CCRT, and I-CCRT-C groups was 32.5, 16.1, and 27.1 months, respectively (p = 0.464). The median OS of the CCRT-C, I-CCRT, and I-CCRT-C groups was 45.9, 35.5, and 54.0 months, respectively (p = 0.788). Cox regression analysis indicated that I-CCRT-C and I-CCRT did not significantly prolong PFS and OS compared with CCRT-C (p &gt; 0.05). Neutropenia grade ≥3 in CCRT-C, I-CCRT, and I-CCRT-C groups was 47.5%, 15%, and 33.3% of patients, respectively (p = 0.027).ConclusionsI-CCRT and I-CCRT-C using cisplatin plus fluorouracil or docetaxel regimen are not superior to CCRT-C in survival but seem to have less severe neutropenia than CCRT-C. Further randomized controlled studies are warranted.

Purpose: To construct machine learning models for predicting progression free survival (PFS) and overall survival (OS) with esophageal squamous cell carcinoma (ESCC) patients. Methods: 204 ESCC patients were randomly divided into training cohort (n=143) and validation cohort (n=61) according to the ratio of 7:3. Two radiomics models were constructed by features which were selected by LASSO Cox model to predict PFS and OS, respectively. Clinical features were selected by univariate and multivariate Cox proportional hazards model (p<0.05). Combined radiomics and clinical model was developed by selected clinical and radiomics features. The receiver operating characteristic (ROC) curve, Kaplan Meier (KM) curve and nomogram were used to display the capability of constructed models. Results: There were 944 radiomics features extracted based on region of interest (ROI) in CT images. There were six radiomics features and seven clinical features for PFS prediction and three radiomics features and three clinical features for OS prediction; The radiomics models showed general performance in training cohort and validation cohort for prediction for prediction PFS (AUC, 0.664, 0.676. C-index, 0.65, 0.64) and OS (AUC, 0.634, 0.646.C-index, 0.64, 0.65). The combined models displayed high performance in training cohort and validation cohort for prediction PFS (AUC, 0.856, 0.833. C-index, 0.81, 0.79) and OS (AUC, 0.742, 0.768. C-index, 0.72, 0.71) Conclusion: We developed combined radiomics and clinical machine learning models with better performance than radiomics or clinical alone, which were used to accurate predict 3 years PFS and OS of non-surgical ESCC patients. The prediction results could provide a reference for clinical decision.