Latent profile analysis of multidimensional muscle profiles versus traditional sarcopenia metrics for capturing sagittal malalignment in adult spinal deformity: a retrospective study.

STUDY DESIGN: Retrospective cohort study. PURPOSE: Using latent profile analysis (LPA) to multidimensionally profile trunk muscles, we aimed to identify muscle phenotypes, compare LPA-based classification with sarcopenia definitions for explaining sagittal alignment, and examine malalignment patterns across LPA-derived subtypes. OVERVIEW OF LITERATURE: Paraspinal sarcopenia is closely associated with sagittal malalignment in adult spinal deformity (ASD), but traditional single-muscle or single-metric definitions fail to capture heterogeneous muscle degeneration. METHODS: This retrospective cohort study included 145 patients with ASD undergoing long-segment fusion (≥5 levels; upper instrumented vertebra T1-L1) with ≥1-year follow-up. We measured muscle quantity using the psoas vertebral body index (PVBI) and paralumbar vertebral body index (PLVBI), and muscle quality using fatty infiltration (FI) of the psoas and paralumbar (multifidus and erector spinae) muscles. LPA was applied to PVBI, PLVBI, and psoas and paralumbar FI to identify clusters. For comparison, sarcopenia was defined by each metric. We compared LPA with single-metric definitions for predicting sagittal measures and examined clinical, radiographic (pre/ post), and surgical outcomes across LPA subtypes. RESULTS: LPA explained sagittal alignment better than any single measure and revealed four phenotypes: Mild-Deg (high quantity, low FI; n=49), Mod-Deg (moderate; n=39), Seve-Deg (low quantity, high FI; n=34), and Hete-Deg (low quantity, high paralumbar FI but low psoas FI; n=23). Preoperative deformity progressively worsened from Mild-Deg to Seve-Deg, while Hete-Deg also showed pronounced malalignment. Postoperatively, most metrics were similar across groups. At last follow-up, however, Mod-Deg/Seve-Deg had greater pelvic incidence (PI)-lumbar lordosis (LL), and LL in Seve-Deg trended toward progressive worsening from postoperative to follow-up (PI-LL and T4 pelvic angle). CONCLUSIONS: Multidimensional LPA muscle phenotyping outperformed single-metric sarcopenia definitions in explaining sagittal alignment. It also identified a heterogeneous Hete-Deg subtype suggestive of functional compensation despite severe preoperative deformity, whereas Seve-Deg reflected irreversible muscle loss with progressive imbalance.