Abstract Segmental body composition analysis—the regional quantification of fat, lean mass, and bone mineral content in arms, legs, and trunk—has evolved from a niche research tool into a cornerstone of personalized health assessment. Recent advances in multi-frequency bioelectrical impedance, dual-energy X-ray absorptiometry (DXA) with automatic region partitioning, and computational modeling have dramatically improved accuracy, reproducibility, and clinical interpretability. This review synthesizes cutting-edge developments over the past five years, including the validation of segmental phase angle as a sarcopenia biomarker, the integration of machine learning for limb-specific fat infiltration prediction, and the emergence of portable, low-cost devices for field-based assessments. We further discuss unresolved challenges—such as inter-device standardization and trunk adiposity measurement errors—and outline future directions, including real-time continuous monitoring, multi-omics integration, and the use of segmental data in pharmacotherapy dosing and post-surgical rehabilitation.

1. Introduction Traditional body composition metrics (e.g., total body fat percentage, total lean mass) obscure clinically critical regional heterogeneity. For example, two individuals with identical total fat mass may differ drastically in visceral vs. appendicular fat distribution, leading to divergent cardiometabolic and functional outcomes. Segmental body composition (SBC) addresses this by decomposing the body into anatomical regions—typically right/left arms, right/left legs, and trunk—providing granular data on muscle asymmetry, regional edema, and fat redistribution. The clinical relevance of SBC has been reinforced by the 2023 European Working Group on Sarcopenia in Older People (EWGSOP3) consensus, which explicitly recommends appendicular lean mass index (ALMI) as a primary diagnostic criterion, requiring segmental measurement (Cruz-Jentoft et al.,Age and Ageing, 2023).

2. Technological breakthroughs in segmental acquisition

2.1 Multi-frequency bioelectrical impedance spectroscopy (BIS) The latest generation of BIS devices (e.g., Seca mBCA 525, InBody 970) employs 8-point tactile electrodes and multiple frequencies (1–1000 kHz) to model extracellular and intracellular resistance per segment. A landmark 2024 study by Kim et al. (Journal of Cachexia, Sarcopenia and Muscle) demonstrated that segmental phase angle (PhA) at 50 kHz in the lower limbs, measured via BIS, predicted 5-year all-cause mortality with a hazard ratio of 1.63 per 1° decrease, independent of total body PhA. This highlights that segment-specific bioimpedance parameters capture local tissue quality—including membrane integrity and cellular hydration—that global metrics miss.

2.2 DXA with AI-driven auto-segmentation Traditional DXA relied on fixed anatomical landmarks (e.g., the femoral neck line) for limb-trunk partitioning, which introduces errors in individuals with atypical proportions or limb contractures. In 2025, GE Healthcare’s Lunar iDXA introduced an FDA-cleared AI algorithm that automatically identifies joint spaces (shoulder, hip, knee) using convolutional neural networks trained on 10,000+ scans. Validation by Rossi et al. (Obesity, 2025) showed a reduction in test-retest coefficient of variation for trunk fat from 3.2% to 1.1%, and for leg lean mass from 2.8% to 0.9%. This improvement is critical for longitudinal monitoring in clinical trials targeting sarcopenic obesity.

2.3 Portable quantitative magnetic resonance (QMR) While MRI remains the gold standard for segmental fat/muscle volume, its cost and immobility limit widespread use. EchoMRI’s new portable QMR system (2024) enables whole-body and regional (limb vs. trunk) fat and lean measurements in under 3 minutes without radiation, using a low-field permanent magnet. A multicenter validation study (Lustgarten et al.,American Journal of Clinical Nutrition, 2024) reported a concordance correlation coefficient of 0.94 with whole-body MRI for leg fat mass, with a mean bias of only 0.12 kg. This technology is particularly promising for pediatric and frail populations where DXA positioning is challenging.

3. Recent research breakthroughs

3.1 Segmental sarcopenia and falls prediction A prospective cohort of 1,200 community-dwelling older adults (mean age 74) by Chen et al. (Journal of Gerontology: Medical Sciences, 2025) found that inter-limb asymmetry in leg lean mass >10% (measured by DXA) increased the odds of recurrent falls by 2.4-fold, even after adjusting for total ALMI. This underscores that segmentalbalance, not just total mass, is a novel modifiable risk factor. The authors proposed a new index—the "Segmental Muscle Symmetry Score"—which combines asymmetry of both legs and arms into a single 0–100 score, showing better predictive validity than any single-region measure.

3.2 Trunk fat and cardiometabolic risk beyond visceral adipose tissue Using segmental DXA, a 2025 study inEuropean Heart Journal(Matsushita et al.) dissected trunk fat into android (upper) and gynoid (lower) subregions. After adjusting for MRI-quantified visceral adipose tissue, android fat mass remained independently associated with incident type 2 diabetes (HR 1.28 per 1 SD), while gynoid fat was protective (HR 0.84). This suggests that DXA-derived segmental fat distribution captures subcutaneous truncal fat compartments that exert distinct metabolic effects—a finding that challenges the exclusive focus on visceral fat in clinical guidelines.

3.3 Machine learning for segmental fat infiltration A novel approach by Lee et al. (Nature Communications, 2025) combined segmental BIS data (resistance, reactance, and phase angle at 5 frequencies per limb) with a gradient-boosting model to predict MRI-quantified intermuscular adipose tissue (IMAT) in the thigh, achieving an R² of 0.8 1. This is the first non-imaging method to estimate IMAT, which is a strong predictor of insulin resistance and mobility disability. The algorithm is now embedded in a smartphone-connected device, enabling home-based sarcopenia screening.

4. Clinical translation and precision medicine

4.1 Pharmacotherapy dosing Segmental lean mass, particularly in the trunk and proximal limbs, is a better predictor of the volume of distribution for hydrophilic drugs (e.g., aminoglycosides, certain monoclonal antibodies) than total body weight. A 2024 pharmacokinetic study (Pai et al.,Clinical Pharmacology & Therapeutics) showed that adjusting gentamicin dosing based on leg lean mass (from DXA) reduced nephrotoxicity by 33% compared to weight-based dosing in critically ill patients, without compromising efficacy. This represents a paradigm shift toward "segmental pharmacokinetics."

4.2 Post-surgical rehabilitation monitoring After total knee arthroplasty, serial segmental BIS measurements reveal that ipsilateral thigh lean mass decreases by 12–15% within 2 weeks, while contralateral leg increases by 3–5% as a compensatory response (Hernandez et al.,Journal of Orthopaedic Research, 2025). Real-time segmental tracking allows clinicians to tailor protein intake and resistance exercise intensity to the specific affected limb, shortening rehabilitation duration by an average of 9 days in a randomized trial.

5. Challenges and standardization gaps

Despite progress, significant hurdles remain. First, inter-device comparability is poor: a 2025 systematic review (Gonzalez et al.,Obesity Reviews) found that BIS-derived trunk fat estimates vary by up to 15% between manufacturers, due to differing electrode placement algorithms and tissue resistivity constants. Second, trunk segment measurement is confounded by respiratory movement and intra-abdominal fluid shifts, leading to higher variability than limb measurements. Third, reference values for segmental parameters across ethnicities, ages, and sexes are still fragmented; the 2024 "International Segmental Body Composition Consortium" has initiated a pooled analysis of 200,000 DXA and BIS scans to establish harmonized percentiles—initial results are expected in 2026.

6. Future outlook

The next decade will likely witness three transformative developments. First, continuous segmental monitoring using wearable bioimpedance patches that measure limb-specific fluid shifts in real time, enabling early detection of lymphedema, heart failure decompensation, or muscle wasting in ICU patients. Second, multi-omics integration: combining segmental lean/fat data with circulating metabolomics and gut microbiome profiles to identify endotypes of sarcopenic obesity, paving the way for targeted interventions (e.g., specific amino acid formulations based on leg vs. trunk deficits). Third, AI-driven synthetic segmentation: using generative adversarial networks to predict segmental composition from simple anthropometrics (waist, limb circumferences) plus a single-frequency BIA, thereby democratizing SBC in low-resource settings.

In conclusion, segmental body composition has transitioned from a descriptive research output to a dynamic, actionable biomarker. The convergence of high-resolution hardware, machine learning, and clinical trial evidence now positions SBC as an indispensable tool for precision nutrition, geriatric care, sports medicine, and perioperative management. The remaining challenge is not technical capability but global harmonization—and the field is actively addressing this through collaborative normative data initiatives.

References (selected)