In individuals with normal body weight, altered fat distribution contributes to metabolic dysfunction and hepatic steatosis. We hypothesized that the weight-adjusted waist index (WWI), an indicator of abdominal fat accumulation, is associated with lean MASLD.

We analyzed lean adults from 2017–March 2020 pre-pandemic (N = 824). MASLD was defined using controlled attenuation parameter (CAP) from vibration-controlled transient elastography together with metabolic criteria. Four machine learning models (XGB, GBM, LASSO, SVM) were trained to prioritize candidate predictors from anthropometric and metabolic indices. The top consensus predictor was then evaluated in an independent cohort (NHANES August 2021–August 2023, N = 782) using survey-weighted logistic regression, subgroup analyses, and smooth curve fitting to assess potential non-linearity.

Across models, WWI ranked as the most important feature for discriminating lean MASLD and exceeded established indices such as the lipid accumulation product (LAP) and cardiometabolic index (CMI) in global importance metrics. The XGB model achieved the highest discrimination (AUC = 0.756). In the validation cohort, MASLD prevalence increased across WWI quartiles (7.5% to 44.0%). As a continuous variable, WWI was associated with MASLD after full adjustment (OR = 3.08, 95% CI: 1.53–6.21). However, quartile-based associations were attenuated and not statistically significant in the fully adjusted model. A modest trend across quartiles remained. Subgroup analyses demonstrated generally consistent associations across demographic and clinical strata. The dose–response analysis suggested a non-linear relationship, with MASLD risk increasing more steeply beyond a WWI of approximately 11 cm/√kg.

Among lean adults, WWI is independently associated with MASLD and may capture central-adiposity–related risk not reflected by BMI alone. WWI could serve as a simple complementary anthropometric marker for risk stratification in normal-weight populations, pending prospective validation.