S720
Clinical – Lower GI
ESTRO 2026
and incorporated into the ML models, including age, BMI, SMI, SATI, SAT_HU, VATI, and SM_HU. The SVM model achieved an internal validation AUC of 0.736±0.068 and an external validation AUC of 0.719±0.028. Logistic regression yielded an internal validation AUC of 0.742±0.059 and an external validation AUC of 0.638 0.023. The RF model showed an internal validation AUC of 0.673±0.071 and an external validation AUC of 0.631±0.051. Among the three models, the SVM achieved the most balanced and robust performance across internal and external validations. Conclusion: SMI was significantly associated with severe hematologic toxicity during iTNT. Predictive models incorporating body composition parameters, particularly the SVM model, showed favorable discriminative and may provide guidance for decision- making. References: [1] Xia F, Wang Y, Wang H, et al. Randomized Phase II Trial of Immunotherapy-Based Total Neoadjuvant Therapy for Proficient Mismatch Repair or Microsatellite Stable Locally Advanced Rectal Cancer (TORCH). J CLIN ONCOL. 2024; 42(28):3308-3318. doi:10.1200/JCO.23.022611.[2] Yang W, Zhang Z, Zhou M, et al. Radiomics of skeletal muscle helps to predict gastrointestinal toxicity in locally advanced rectal cancer patients receiving neoadjuvant chemoradiotherapy. CLIN TRANSL RAD ONCO. 2024; 44100703. doi:10.1016/j.ctro.2023.100703 Keywords: Body composition, toxicities, iTNT Prospective assessment of Quality of Life in Locally Advanced Rectal Cancer following Neo Adjuvant Chemo Radiotherapy at a tertiary hospital in India. Pritanjali Singh, Jahnavi Kodela Radiation Oncology, All India Institute of Medical Sciences Patna, Patna, India Purpose/Objective: Neoadjuvant chemoradiotherapy followed by surgery is the standard of care for LARC patients. Although effective, radiotherapy (RT) can impact the patient’s physical, emotional, familial and social aspects of QoL. The effectiveness of cancer treatment has always been evaluated based on survival rates, treatment-related toxicity, and physician evaluated functional outcomes.However,there is knowledge gap in understanding of how patients perceive the adverse effects of RT(1,2) especially in lower- middle- income countries like India. Hence, we conducted this study to assess QoL in rectal cancer patients undergoing NACTRT utilizing validated patient- reported outcome Digital Poster 3217
Digital Poster 3191
Body composition helps to predict hematologic toxicities in rectal cancer patients receiving immunotherapy-based total neoadjuvant therapy Wang Yang, Ruone Xu, Xiaoming Sun, Hui Zhang, Menglong Zhou, Yajie Chen, Yaqi Wang, Fan Xia, Zhen Zhang Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Shanghai, China Purpose/Objective: Immunotherapy-based total neoadjuvant therapy (iTNT) represents a promising strategy for enhance complete response rates and facilitating organ preservation in patients with locally advanced rectal cancer (LARC). However, this regimen is associated with a considerable incidence of severe hematologic toxicities, yet reliable predictive biomarkers are still lacking. Body composition can serve as a surrogate for nutritional status and was associated with toxicity of chemoradiotherapy (CRT). Nevertheless, its relationship with toxicities of immunotherapy-based regimens remains unclear. This study aims to investigate the relationship between body composition and hematologic toxicity associated with iTNT, and to develop a predictive model to predict hematologic toxicities. Material/Methods: A cohort of 204 LARC patients treated with iTNT were included for model development, with internal training and validation performed via five-fold cross-validation. Additionally, 43 patients with metastatic rectal cancer treated with same regimen were included as the external validation cohort. Body composition parameters including BMI, SMI, skeletal muscle attenuation (SM_HU), subcutaneous adipose tissue index (SATI), subcutaneous adipose tissue attenuation (SAT_HU), visceral adipose tissue index (VATI), and visceral adipose tissue attenuation (VAT_HU), were obtained from pretreatment CT images acquired at the L3. The least absolute shrinkage and selection operator (LASSO) regression was applied to the discovery dataset to select variables. Three common machine learning (ML) algorithms including logistic regression, random forest (RF), and support vector machine (SVM), were initially developed, and the best- performing model was selected as the final predictive model. Results: 33.8% of patiens experienced grade 3-4 hematologic toxicities. The sex-specific cut-offs for L3 skeletal muscle index associated with severe hematologic toxicities were 50.0 cm ² /m ² for men and 39.6 cm ² /m ² for women. Patients with low SMI demonstrated a significantly higher incidence of severe hematologic toxicities (P<0.001). Seven features were identified
Made with FlippingBook - Share PDF online