Personalized neoadjuvant treatment regimen selection in locally advanced rectal cancer based on regimen-specific response modeling.
Authors
Affiliations (18)
Affiliations (18)
- School of Life Science and Technology, Xidian University & Engineering Research Center of Molecular and Neuro Imaging, Ministry of Education, Xi'an, Shaanxi, China.
- Xi'an Key Laboratory of Intelligent Sensing and Regulation of trans-Scale Life Information & International Joint Research Center for Advanced Medical Imaging and Intelligent Diagnosis and Treatment, School of Life Science and Technology, Xidian University, Xi'an, Shaanxi, China.
- Division of Abdominal Tumor Multimodality Treatment, Cancer Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China.
- Department of Radiation Oncology, Cancer Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China.
- CAS Key Laboratory of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences, Beijing, China.
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China.
- Department of Colorectal Surgery, State Key Laboratory of Oncology in South China, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangzhou, Guangdong, China.
- Department of Radiation Oncology, The Sixth Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China.
- Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Diseases, The Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, China.
- Department of Colorectal Surgery, The Third Affiliated Hospital of Kunming Medical University, Yunnan Cancer Hospital, Peking University Cancer Hospital Yunnan, Kunming, Yunnan, China.
- Colorectal Cancer Center, Department of General Surgery, West China Hospital, Sichuan University, Chengdu, Sichuan, China.
- School of Life Science and Technology, Xidian University & Engineering Research Center of Molecular and Neuro Imaging, Ministry of Education, Xi'an, Shaanxi, China. [email protected].
- CAS Key Laboratory of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences, Beijing, China. [email protected].
- School of Engineering Medicine & School of Biological Science and Medical Engineering, Beihang University, Beijing, China. [email protected].
- CAS Key Laboratory of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences, Beijing, China. [email protected].
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China. [email protected].
- Division of Abdominal Tumor Multimodality Treatment, Cancer Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China. [email protected].
- Department of Radiation Oncology, Cancer Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China. [email protected].
Abstract
Neoadjuvant therapy is standard for locally advanced rectal cancer (LARC), yet regimen selection remains population-based, risking over- or undertreatment. We developed and validated a deep learning framework that provides a generalizable paradigm for data-driven treatment regimen selection by estimating patient-specific probabilities of pathological complete response (pCR) across multiple therapeutic options. In a multicenter cohort, a hard-gated mixture-of-experts model integrating pretreatment multiparametric MRI and clinical variables generated regimen-specific pCR probabilities to support clinician-led treatment decision-making. The model achieved strong predictive performance, with AUCs of 0.827 and 0.790 in the validation and prospective test cohorts. In the combined validation and test cohorts, 53.16% of patients were recommended treatment escalation, with an observed pCR rate of 11.11% and a model-estimated pCR probability of 30.95% under the model-supported regimen. Meanwhile, 5.91% of patients were identified for de-intensification while maintaining a high estimated likelihood of response. This framework provides probabilistic support for multidisciplinary optimization of neoadjuvant treatment intensity in LARC. The prospective cohort was registered in the Chinese Clinical Trial Registry (ChiCTR2400085797; June 18, 2024).