Background
Traditional anthropometric measures, including body mass index (BMI), waist circumference (WC), and waist-to-hip ratio (WHR), represent strong predictors of breast cancer (BC) risk. However, WC and WHR present some limitations due to strong correlation with BMI. A Body Shape Index (ABSI) has been proposed as a novel metric to capture central obesity independently of BMI.
Objectives
We aimed to investigate the prognostic relevance of body shape in early-stage BC survivors from Southern Italy by integrating traditional anthropometric measures with ABSI to define distinct body shape phenotypes using an unsupervised machine-learning approach. Methods
Data from 955 early-stage BC survivors were analyzed, including BMI, WC, WHR, and ABSI. Body shape phenotypes derived from clustering methods following principal component analysis (PCA) to combine general and central obesity metrics. Associations of derived phenotypes with all-cause and BC-specific mortality were evaluated using hazard ratios (HRs) including terms for age, center, cancer stage, therapy and surrogate molecular subtypes.
Results
Four distinct body-shape clusters were identified. The “Young & fit” cluster was characterized by the most favorable anthropometric profile, including lower BMI, adiposity indices (BMI 22 [21–24] kg/m²; WHR 0.81 [0.76–0.85]) and a younger age, whereas the “General obesity” (apple) cluster presented the most unfavorable profile, with higher BMI values (35 [32–38] kg/m²). Despite similar BMI, intermediate clusters, namely “Young & overweight” (peer) and “Short and overweight” (little-squared) differed in body fat distribution and age, with the “Short & overweight” cluster showing greater central adiposity indices (WHR 0.93 [0.90–0.96]) and older age. The “General obesity” (apple) cluster showed a more than twofold increased all-cause mortality risk (HR = 2.02; 95% CI 1.23–3.29; p = 0.005). Inclusion of metabolic factors (blood pressure, HDL, triglycerides, glycemia and the Metabolic syndrome) strengthened the risk gradient, indicating a synergistic effect of adiposity and metabolic dysfunctions on BC prognosis. In line with these findings, metabolic syndrome was independently associated with an increased risk of all-cause mortality (HR = 1.92; 95% CI 1.32–2.80; p < 0.001). Conversely, when both body shape–age clusters and metabolic syndrome were included in the model, body shape showed no significant association with BC mortality, although the “General obesity” cluster showed a non-significant trend (HR = 1.64; 0.92-2.91; p = 0.091). Metabolic syndrome remained a strong independent predictor, more than doubling the risk of BC death (HR = 2.07; 95% CI 1.29–3.32; p = 0.002).
Conclusions
Our findings highlight the importance of assessing body composition and metabolic health beyond BMI in BC survivors. Metabolic syndrome strongly mediates mortality risk, suggesting that targeting dyslipidemia and hyperglicemia may be more effective than weight-focused strategies alone.

Visual representation of body-shape phenotypes. (Body shape phenotypes were visually represented using the Meshcapade editor (https://me.meshcapade.com/editor))