# Extremity Ultrasound vs. Computed Tomography at the Third Lumbar Vertebra Level for Assessing the Subcutaneous Adipose Tissue-to-Muscle Ratio

## TL;DR

This study shows that ultrasound measurements at the upper arm can predict a body fat-to-muscle ratio typically measured by CT scans, offering a non-invasive alternative.

## Contribution

The study identifies specific ultrasound sites, particularly the upper arm, that effectively predict the CT-derived SAT-to-muscle ratio.

## Key findings

- The upper-arm anterolateral site most strongly predicted the CT SAT-to-muscle ratio.
- A LASSO model using ultrasound and clinical data achieved a cross-validated R2 of 0.63.
- Bedside ultrasound can provide a within-patient comparison of body compartments non-invasively.

## Abstract

Background/Objectives: A ratio of subcutaneous adipose tissue to muscle mass may be more informative than defining low subcutaneous adipose tissue and muscle mass separately. The objective of this study was to determine which ultrasound measurement points in the upper and lower extremities predict the subcutaneous adipose tissue (SAT)-to-muscle ratio as measured by gold-standard computed tomography (CT) at the third lumbar vertebra (L3) level. Methods: Two hundred hospitalised patients (41% female; median (Q1–Q3) age: 61.3 (51.0–70.1) years) who underwent an abdominal CT scan for any clinical reason within 48 h prior to extremity ultrasound were included in this prospective observational study conducted from 2017 to 2019. Ultrasound measurements of subcutaneous adipose tissue and muscle thickness were obtained at three measuring points on the thigh and two on the upper arm. On the CT scan at the L3 level, subcutaneous (SAT) and visceral adipose tissue and skeletal muscle area were measured. A linear LASSO (Least Absolute Shrinkage and Selection Operator) model was used to identify which ultrasound sites best predicted the CT L3 SAT-to-muscle ratio. Results: Height, weight, sex, SAT-to-muscle ratio at four ultrasound measuring points and abdominal circumference predicted the CT SAT-to-muscle ratio in the LASSO model (R2 = 0.70; cross-validated R2 = 0.63; p values are not reported in LASSO regression and R2 is used instead). The upper-arm anterolateral ultrasound site most strongly influenced the CT SAT-to-muscle ratio (estimate × standard deviation of predictor: 0.24). Conclusions: The CT SAT-to-muscle ratio at the L3 level can be predicted non-invasively using bedside ultrasound, particularly at the anterolateral measuring point of the upper arm. Bedside ultrasound assessment of the ratio of subcutaneous adipose tissue to muscle on the anterolateral upper arm provides a within-patient comparison of body compartments.