2963 - Exploring the Potential Causal Relationship between Body Composition and Prognosis in Breast Cancer: A Counterfactual Retrospective Cohort Study
Presenter(s)
Y. H. Lai1,2, C. C. Su2, H. H. W. Chen1, P. F. Su3, M. R. Shen4, and Y. S. Tsai2,5; 1Department of Oncology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan, 2Clinical Innovation and Research Center, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan, 3Department of Statistics, National Cheng Kung University, Tainan, Taiwan, 4Department of Obstetrics and Gynecology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan, 5Department of Medical Imaging, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
Purpose/Objective(s): Body composition is modifiable and may influence breast cancer prognosis. This study aims to investigate the possible causal relationship between body composition and survival outcomes, offering insights into potential therapeutic targets.
Materials/Methods: This retrospective study included female breast cancer patients treated between 2010 and 2019 who underwent post-diagnosis computed tomography (CT) scans with unenhanced axial images at the L3 vertebral level for body composition analysis. CT images were analyzed using a proprietary artificial intelligence (AI) model to quantify muscle and fat distribution. Myopenia, myosteatosis, and obesity were evaluated using the skeletal muscle index (SMI), skeletal muscle density (SMD), and visceral adipose tissue (VAT) area, respectively. Muscle quality was assessed by SMD, normal-attenuation muscle area (NAMA), skewness, and kurtosis. Covariates included age, body mass index (BMI), comorbidities, tumor characteristics, and treatments. Disease-free survival (DFS) and overall survival (OS) were the primary outcomes. Survival area plots visualized adjusted survival probabilities. Generalized propensity score models with inverse probability of treatment weighting (IPTW) addressed confounders to establish potential causal relationships, and Cox additive hazard models with IPTW estimated adjusted hazard ratios (aHRs) for DFS and OS.
Results: Among 388 patients (mean age: 56.8 years; BMI: 24.4 kg/m²), 110 (28.4%) experienced relapse, and 91 (23.5%) died (mean DFS: 6.4 years; OS: 7.2 years). Lower SMI, SMD, NAMA, kurtosis, and higher skewness were significantly associated with worse DFS and OS. VAT significantly impacted DFS but not OS. IPTW-adjusted Cox additive hazard models indicated that lower values of SMD (<20 HU), NAMA (<30 cm²), kurtosis (<1.0), and higher skewness (>-1.0) may causally increase relapse and mortality risks (aHRs significantly > 1).
Conclusion: Poor muscle quality, characterized by low SMD, low NAMA, low kurtosis, and high skewness, may contribute to worse DFS and OS in breast cancer patients. CT-derived body composition indices may provide valuable prognostic information, highlighting muscle quality as a potential therapeutic target for improving breast cancer outcomes.