LAUSR

Background The anthropometric index is not accurate but shows a great advantage in accessibility. Simple body composition formulas should be investigated before proceeding with the universal nutrition screening. Materials and Methods Clinical data of patients with a malignant tumor of the digestive system were collected. SliceOmatic 5.0 software (TOMOVISION, Canada) was used to analyze abdominal CT images and taken as references. A linear regression analysis was adopted to establish the formula for calculating skeletal muscle index (SMI) and visceral fat area (VFA). In addition, the relweights function was adopted to measure the contribution of each variable. Results In total, 344 patients were divided into the training set and 134 patients into the validation set. The selected formulas were SMI.pre = 0.540 × weight (kg) – 0.559 × height (cm) – 13.877 × sex (male = 1, female = 2) + 123.583, and VFA.pre = 5.146 × weight (kg) – 2.666 × height (cm) + 1.436 × age (year) + 134.096, of which the adjusted R2 were 0.597 and 0.581, respectively. The “weight” explained more than 80% of R2 in the prediction of VFA. In addition, “sex” occupied approximately 40% of R2 in the prediction of SMI. The paired t-test showed no significant difference between the real measured indices and the predicting ones (p = 0.123 for SMI and p = 0.299 for VFA). The logistic regression analysis exhibited similar diagnostic efficacy of the real measured parameters and formulas. Conclusion The SMI and VFA formulas were developed through basic indices, such as weight, height, sex, and age. According to the contribution of each variable, weight should always be focused on preserving appropriate muscle and adipose tissue.