3480 - A Multi-Dimensional Deep Ensemble Learning Model Predicts Pathological Response and Treatment Outcomes to Neoadjuvant Chemoradiotherapy in Esophageal Squamous Cell Carcinoma Using Pretreatment CT Images: A Multicenter, Retrospective Cohort Study
Presenter(s)
Y. Liu1, Y. Su2, P. Jun3, W. Zhang4, F. Zhao5, Y. Li6, X. Y. Song7, Z. Ma1, W. Zhang1, J. Ji1, Y. Chen1, Y. Men8, F. Ye9, K. Men1, J. Qin10, W. Liu1, X. Wang1, N. Bi1, L. Xue11, W. Yu6, Q. Wang12, Z. Men2, and Z. Hui8; 1Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 2School of Biomedical Engineering, Wenzhou Medical University, Wenzhou, China, 3Department of Thoracic Surgery, Sichuan Clinical Research Center for Cancer, Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, University of Electronic Science and Technology of China, Chengdu, China, 4Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin's Clinical Research Center for Cancer, Tianjin, China, 5Department of Radiation Oncology, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin's Clinical Research Center for Cancer, Tianjin, China, 6Department of Radiation Oncology, Shanghai Chest Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China, 7Department of Radiation Oncology, Shanghai Chest Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China, 8Department of VIP Medical Services, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 9Department of Diagnostic Radiology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 10Department of Thoracic Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 11Department of Pathology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 12Department of Radiation Oncology, Sichuan Cancer Hospital and Institution, Sichuan Cancer Center, School of Medicine, University of Electronic Science and Technology of China, Radiation Oncology Key Laboratory of Sichuan Province, Chengdu, China
Purpose/Objective(s): Neoadjuvant chemoradiotherapy (nCRT) followed by esophagectomy is the standard treatment for esophageal squamous cell carcinoma (ESCC). However, accurate prediction of pathological complete response (pCR) and treatment outcome remains a major clinical challenge. This study aimed to develop and validate a multidimensional deep ensemble learning model (DELRN) using pretreatment CT imaging to predict pCR and stratify prognostic risk for ESCC patients undergoing nCRT.
Materials/Methods: In this multicenter, retrospective cohort study, 485 ESCC patients were enrolled from four hospitals in China between May 2009 to August 2023, December 2017 to September 2021, May 2014 to September 2019 and March 2013 to July 2019. Patients were divided into a discovery cohort (n=194), an internal cohort (n=49) and three external validation cohorts (n=242). A multi-dimensional feature-driven deep ensemble learning model (DELRN) incorporating both radiomics and 3D convolutional neural networks was developed to predict pCR and treatment outcomes for nCRT based on pretreatment CT images. The performance of the model was evaluated by discrimination, calibration and clinical utility. Kaplan-Meier survival analysis was performed to assess the overall survival (OS) and disease-free survival (DFS) of ESCC patients at two follow-up centers.
Results: DELRN demonstrated robust predictive performance for pCR in the discovery, internal, and external validation cohorts, achieving area under the curve (AUC) values of 0.943 (95% CI: 0.912-0.973), 0.796 (95% CI: 0.661-0.930), 0.767 (95% CI: 0.646-0.887), 0.829 (95% CI: 0.715-0.942), and 0.782 (95% CI: 0.664-0.900), respectively, outperforming both single-domain radiomics and deep learning models. In addition, DELRN demonstrated robust prognostic value, effectively classifying patients into high-risk and low-risk groups for OS (log-rank P = 0.018 and 0.0053) and DFS (log-rank P = 0.00042 and 0.035). Multivariate analysis confirmed that DELRN was an independent prognostic factor for both OS and DFS.
Conclusion: The DELRN model demonstrated promising clinical potential as an effective, non-invasive tool for predicting nCRT response and treatment outcome in ESCC patients, enabling personalized treatment strategies and improving clinical decision-making with future prospective multicenter validation.