327 Unsupervised Machine Learning Reveals Clinically Relevant Body Composition Phenotypes in Adult Spinal Deformity Patients

View at publisher Summarise with VIRAT
Authors: Tej Deepak Azad, Carlos Andres Aude, Terry Nguyen, Jaemin Kim, Vedat Deviren, Christopher Pearson Ames
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

INTRODUCTION: Body composition is increasingly recognized as a determinant of surgical outcomes. Traditional metrics, like body mass index (BMI), fail to capture the interplay of sarcopenia, adiposity, and fluid imbalance—factors relevant in adult spinal deformity (ASD) surgery. We leveraged the inBody 970 bioimpedance platform to define and evaluate physiologically grounded body composition phenotypes. METHODS: We prospectively collected preoperative in Body 970 data from 178 ASD patients, focusing on six validated parameters: phase angle, body cell mass (BCM), ECW/TBW ratio, % skeletal muscle/weight, visceral fat area, and % body fat. Three composite indices were engineered to capture sarcopenic obesity and fluid imbalance. Robust PCA and PAM clustering yielded two distinct phenotypes: Lean-Muscular (n = 91) and Fluid-Adipose (n = 87). Outcomes included medical complications, neurological deficits, infections, and reoperations; a composite endpoint (any complication) was the primary outcome. RESULTS: Patients with the Fluid-Adipose phenotype had significantly elevated risk of complications: adjusted risk ratio 2.38 (95% CI: 1.39–4.06, p = 0.001). Logistic regression yielded an OR of 6.32 (p = 0.004). The phenotype remained independently predictive after adjusting for BMI, age, sex, surgical complexity, and revision status. Notably, the phenotype explained risk beyond BMI (likelihood ratio test p = 0.03; NRI = 0.448). Physiologically, the Fluid-Adipose group exhibited lower phase angle, reduced BCM, higher ECW/TBW, and increased visceral fat—features associated with sarcopenic obesity and impaired healing. CONCLUSIONS: BIA-derived body composition phenotypes offer robust, BMI-independent prognostic value in ASD surgery. These phenotypes capture subclinical physiologic vulnerabilities—such as sarcopenic obesity and fluid imbalance—that may guide risk stratification and prehabilitation. Prospective validation and integration into perioperative protocols could personalize care and improve outcomes in ASD surgery.

Fields of Study & Subjects

Scoliosis diagnosis and treatment Body Composition Measurement Techniques Total Knee Arthroplasty Outcomes

Publication details

Journal / source
Neurosurgery
Volume
72
Issue
Supplement_1
Pages
85–85
Publication date
26 March 2026
Language
EN