Establishment of a risk prediction model for Aromatase Inhibitor-Induced Arthralgia in breast cancer patients: a retrospective cohort study in high-altitude regions

Abstract

Abstract Background: Aromatase inhibitor-induced arthralgia (AIA) is a common and debilitating adverse event in breast cancer patients receiving aromatase inhibitors (AI) treatment. One of the reasons why breast cancer patients discontinue AI treatment prematurely. However, the risk factors and predictive models specific to high-altitude regions are lacking. This study aimed to develop a predictive model for AIA in breast cancer patients in high-altitude regions. Methods: A retrospective cohort study was conducted in a high-altitude region at Qinghai University Affiliated Hospital from June 2021 to October 2023. This study involved a total of 315 breast cancer patients undergoing AI treatment. Participants were randomly assigned to either a training set (n=220) or a validation set (n=95) in a 7:3 ratio. Variable selection was carried out using the Least Absolute Shrinkage and Selection Operator (LASSO) regression, coupled with 7-fold cross-validation. A multivariate logistic regression analysis was performed on the training set to identify independent risk factors for AIA, leading to the establishment of a nomogram based on these risk factors. The model's performance was assessed using calibration plots, Receiver Operating Characteristic (ROC) curves, and Decision Curve Analysis (DCA). Results: Out of the 14 variables analyzed, five predictors were selected for the development of the predictive model. These included prior chemotherapy, years since the last menstrual period (LMP), menopause mode, stage, and psychological factors. The incidence rate of AIA in the cohort was 58.41%. The multivariate logistic regression analysis identified several significant independent predictors for AIA in high-altitude regions. These included previous use of taxane chemotherapy (Odds Ratio [OR] = 10.174, 95% Confidence Interval [CI] = 2.008-62.69, P=0.008), LMP (OR = 0.175, 95% CI = 0.052-0.494, P=0.002), drug-induced menopause (OR = 3.834, 95% CI = 1.109-14.13, P=0.036), stage (OR = 10.423, 95% CI = 4.114-32.15, P < 0.001), and psychological factors (OR = 25.108, 95% CI = 8.430-87.95, P<0.001). The developed nomogram exhibited a strong predictive capacity, with an area under the Receiver Operating Characteristic (ROC) curve value of 0.971. The calibration curve demonstrated a high degree of consistency between predicted probabilities and observed values. Decision Curve Analysis (DCA) underscored the clinical utility of the nomogram.