Identify the Characteristics of Metabolic Syndrome and Non-obese Phenotype: Data Visualization and a Machine Learning Approach

Abstract

Introduction: A third of the world's population is classified as having Metabolic Syndrome (MetS). Traditional diagnostic criteria for MetS are based on three or more of five components. However, the outcomes of patients with different combinations of specific metabolic components are undefined. It is challenging to be discovered and introduce treatment in advance for intervention, since the related research is still insufficient.Methods: This retrospective cohort study attempted to establish a method of visualizing metabolic components by using unsupervised machine learning and treemap technology to discover the relations between predicting factors and different metabolic components. Several supervised machine-learning models were used to explore significant predictors of MetS and to construct a powerful prediction model for preventive medicine.Results: The random forest had the best performance with accuracy and c-statistic of 0.947 and 0.921, respectively, and found that body mass index, glycated hemoglobin, and controlled attenuation parameter (CAP) score were the optimal primary predictors of MetS. In treemap, high triglyceride level plus high fasting blood glucose or large waist circumference group had higher CAP scores (>260) than other groups. Moreover, 32.2% of patients with high CAP scores during 3 years of follow-up had metabolic diseases are observed. This reveals that the CAP score may be used for detecting MetS, especially for the non-obese MetS phenotype.Conclusions: Machine learning and data visualization can illustrate the complicated relationships between metabolic components and potential risk factors for MetS.

“Big Data” Approaches for Prevention of the Metabolic Syndrome

“…However, it has been reported that carriers of the PNPLA3 Ile148Met allele have an increased risk of the disease but do not typically display features of metabolic syndrome. In a later study, the authors further identified patients with different combinations of specific metabolic traits and especially potential patients in the non-obese population ( Yu et al, 2021 ), and with random forest algorithm, further confirmed the capability of the three parameters, namely, CAP score, HbA1c, and body mass index (study focusing on non-obese patients), in predicting MetS in a 3-year follow-up. This is an example of the versatility of data research in shifting the traditional diagnostic concept of diseases, especially with continuous variables such as CAP score to replace an arbitrary diagnosis of NAFLD, reflecting a progressive development of MetS.…”

“…For the same purpose, different models could exhibit different efficacy, such as how in supervised machine learning, the sensitivity and specificity vary among different models and detailed studies are needed to designate the most suitable model. For MetS, various algorithms have been tested, and many highlighted the “random forest” as the most appropriate model ( Xia et al, 2021 ; Yu et al, 2021 ), and studies have also used deep learning tools to analyze data from medical images to further contribute to the prediction of MetS occurrence and outcomes, contributing to secondary and tertiary prevention strategies ( Lin et al, 2021a ; Pickhardt et al, 2021 ).…”

“…Indeed, many studies of machine learning in metabolic disorders, including MetS, type II diabetes, and hypertension, focus on mining data from extensive medical documentation such as electronic records and medical images from daily practices and large-scale trials ( Lin et al, 2021b ; Yu et al, 2021 ). Common data types would include anthropometric data, laboratory tests covering different systems, and medical imaging of other modalities such as x-ray based, ultrasound-based, and even MRI-based techniques.…”

Association between liver stiffness measurement by transient elastography and chronic kidney disease

“…[ 12 ] The value of the liver TE controlled attenuation parameter is also revealed the feasibility of predicting possible CKD-related diseases. [ 13 ]…”

“…[ 27 , 28 ] Moreover, novel machine learning applications also reveal that LSM and CAP are potential factors to evaluate metabolic syndrome. [ 13 , 14 , 29 ] Therefore, the 2 parameters are appropriate screening tools for the determination of patient liver conditions in clinical contexts.…”