Associations of Obesity and Inherited Genetic Variation with Aromatase Inhibitor Induced Musculoskeletal Syndrome in the All of Us Cohort

BACKGROUND: Aromatase Inhibitor-Induced Musculoskeletal Syndrome (AIMSS) causes treatment non-adherence in up to 50% of breast cancer patients. Obesity is a known risk factor for AIMSS, but obesity indices are not well-studied in AIMSS. Further, the investigation of genetic factors in AIMSS has been limited. METHODS: We identified 1,695 women with breast cancer who received AIT in All of Us. Three obesity measures were calculated using height, waist, and/or hip circumference. Genetic analyses were performed using a subset of participants (n=1,202) with available whole-genome-sequencing and joint pain data, which was randomly partitioned into training (n=841) and validation (n=361) sets. A genome-wide association study was performed on the training set and LASSO regression models were fit to identify predictors of AIMSS. ROC and risk stratification analyses were performed on the validation set to assess predictive performance. RESULTS: New-onset joint pain was reported in 32.2% of participants. Body roundness index, waist-to-hip ratio and BMI were not statistically significantly independently associated with joint pain. GWAS revealed several suggestively significant SNPs (p<1.0x10-6). In ROC analyses, the hybrid model including body roundness index and SNPs (AUC=0.73) and the genetic-only model (AUC=0.72) showed good discriminating capacity compared to the clinical data only (AUC=0.58) model. Risk stratification yielded a negative predictive value of 90.1%. CONCLUSIONS: Predictive modeling in this sample suggests that AIMSS risk may be driven predominantly by polygenic architecture. IMPACT: Larger GWAS and studies in independent populations are needed to further investigate these findings, including potential for clinical translation.