Deep-learning model for predicting physical fitness in possible sarcopenia: analysis of the Korean physical fitness award from 2010 to 2023

Abstract

IntroductionPhysical fitness is regarded as a significant indicator of sarcopenia. This study aimed to develop and evaluate a deep-learning model for predicting the decline in physical fitness due to sarcopenia in individuals with potential sarcopenia.MethodsThis study used the 2010–2023 Korean National Physical Fitness Award data. The data comprised exercise- and health-related measurements in Koreans aged >65 years and included body composition and physical fitness variables. Appendicular muscle mass (ASM) was calculated as ASM/height2 to define normal and possible sarcopenia. The deep-learning model was created with EarlyStopping and ModelCheckpoint to prevent overfitting and was evaluated using stratified k-fold cross-validation (k = 5). The model was trained and tested using training data and validation data from each fold. The model’s performance was assessed using a confusion matrix, receiver operating characteristic curve, and area under the curve. The average performance metrics obtained from each cross-validation were determined. For the analysis of feature importance, SHAP, permutation feature importance, and LIME were employed as model-agnostic explanation methods.ResultsThe deep-learning model proved effective in distinguishing from sarcopenia, with an accuracy of 87.55%, precision of 85.57%, recall of 90.34%, and F1 score of 87.89%. Waist circumference (WC, cm), absolute grip strength (kg), and body fat (BF, %) had an influence on the model output. SHAP, LIME, and permutation feature importance analyses revealed that WC and absolute grip strength were the most important variables. WC, figure-of-8 walk, BF, timed up-and-go, and sit-and-reach emerged as key factors for predicting possible sarcopenia.ConclusionThe deep-learning model showed high accuracy and recall with respect to possible sarcopenia prediction. Considering the need for the development of a more detailed and accurate sarcopenia prediction model, the study findings hold promise for enhancing sarcopenia prediction using deep learning.

Neural network model for prediction of possible sarcopenic obesity using Korean national fitness award data (2010–2023)

“…Possible SO prevalence was higher among women than among men, suggesting that sarcopenia and SO in women were associated with an increased all-cause mortality risk independent of obesity, although this association was not significant in men 50 . Lower BF levels were reported in women with sarcopenia in a previous study, and our finding of a high BF level was related to possible SO 32 , 51 . The probability of possible SO had the highest feature importance factor of BF because obese adipose tissue inflammation dominates skeletal muscle inflammation, leading to SO, which is exacerbated by the age-related decline in muscle mass and function 52 .…”

“…The possibility of low grip strength was associated with elevated levels of pro-inflammatory cytokines, which decreased muscle strength and developed SO 55 . Using the same dataset as the Korean National Physical Fitness Award between 2010 and 2023, the sequential deep learning model showed that absolute grip strength was the most important factor for predicting possible sarcopenia in older adults 32 . The results of the deep learning model had high accuracy (AUCRC = 0.95) and recall with possible sarcopenia, and our results also showed that the possible SO prediction model had a high accuracy of performance (AUCRC = 0.97).…”

“…The dependent variable was binary (i.e., normal vs possible). Trained physical fitness instructors followed all physical measures 32 . The sit-and-stand-up test quantified the participant's ability to transition from a seated to a standing position by recording the total number of complete cycles achieved within a 30-s interval.…”

“…After training, the model with the optimized performance on the validation set was loaded with the TensorFlow and Keras model functions. These functions are stored in an optimized neural network model 32 , 44 .…”

“…EarlyStopping monitored the validation loss, terminated training (patience = 20), and set weights of the model to those of the optimized model. Mod-elCheckpoint monitors validation loss and saves the model with the lowest validation loss during training 32,43 .…”