Longitudinal Opportunistic T12 CT-BMD Assessment from Serial Health-Checkup Chest CT in Midlife Women.

This single-center retrospective study evaluated longitudinal artificial intelligence (AI)-derived T12 bone mineral density (BMD) trajectories from serial noncontrast chest computed tomography (CT) in women aged 45-65 years undergoing health checkups. Automated T11 and T12 BMD quantification was performed from serial CT examinations; no study-specific dual-energy X-ray absorptiometry or phantom-calibrated quantitative computed tomography reference standard was available. Low baseline bone status was defined as mean T11-T12 AI-derived CT-BMD ≤ 128 mg/cm3 and was considered an exploratory CT-derived classification rather than a clinical diagnosis of osteoporosis or osteopenia. Absolute and relative T12 BMD changes were assessed using linear mixed-effects models, with T11-only and protocol-consistent sensitivity analyses. Rapid relative loss (≥ 3%/year) and prediction modeling were exploratory. The cohort included 1,034 women and 4,185 examinations over 3.22 ± 0.30 years. The low-status group showed smaller absolute BMD loss (10.82 vs 12.52 mg/cm3; P = 0.022) and a less negative T12 BMD slope (- 3.08 vs - 3.91 mg/cm3/year; P = 0.011). By contrast, the greater relative decline observed in unadjusted analysis was not sustained after covariate adjustment or T11-only sensitivity analysis, suggesting that relative-change estimates were sensitive to baseline dependence, baseline-status definition, and covariate adjustment. Adding T11 BMD to clinical variables provided no clear incremental discrimination for rapid loss (AUC, 0.708 vs 0.716; P = 0.128). Serial chest CT enabled construction of individual AI-derived BMD trajectories, showing that low baseline bone status was not a consistent marker of faster subsequent BMD loss. Reference-standard comparison, repeatability assessment, and external validation are required to determine the clinical utility of longitudinal CT-derived BMD trajectories.