Multi-task deep learning identifies four MASLD subtypes for precision cardiovascular– liver–kidney–metabolic management

Metabolic-associated steatotic liver disease (MASLD) is a clinically heterogeneous condition with highly variable outcomes affecting more than 30% individuals globally. The disease is conventionally staged by histological progression, ranging from simple steatosis to metabolic dysfunction-associated steatohepatitis (MASH) and ultimately fibrosis or cirrhosis. Beyond liver-related outcomes, MASLD significantly elevates the risk of extrahepatic complications, including cardiovascular disease (CVD), type 2 diabetes mellitus (T2DM), and chronic kidney disease (CKD). Currently, personalized management strategies are lacking, underscoring an urgent need for a prognostic stratification system that integrates both hepatic and extrahepatic risks to guide clinical decision-making. A groundbreaking study led by Professor Yan Bi from Drum Tower Hospital, Medical School of Nanjing University, has developed a novel algorithm enabling precise MASLD subtyping for individualized intervention. This study was published online on January 28, 2026, in the Chinese Medical Journal.

The study analyzed 1,111 individuals who underwent liver biopsy and developed a multi-task deep LASSO algorithm for feature selection. This model identified six core clinical indicators: age, BMI, HbA1c, TyG, TC/HDL, and GGT/PLT. Cluster analysis using these variables initially established four stable MASLD subtypes. To evaluate the generalizability of this classification, the cluster analysis was replicated in two large, independent cohorts: a health check-up cohort of 6,172 adults (MASLD prevalence: 43.9%; mean follow-up: 27.6 months) and the NHANES-III cohort comprising 7,406 participants (MASLD prevalence: 37.3%; mean follow-up: 280.2 months). The four-cluster structure remained consistent across both validation cohorts.

Cluster 1 low CVD risk subgroup (41%):
highest percentages of body fat
lowest levels of visceral fat

Cluster 2 high fibrosis risk subgroup (26%):
significant lipid profile disorders
substantial liver damage

Cluster 3 high cardiovascular–kidney risk subgroup (19%):
lowest muscle mass
obvious chronic systemic inflammation

Cluster 4 high cardiovascular–liver–kidney risk subgroup (14%):
severe insulin resistance
poor glucose control (>98% with diabetes)
substantial liver damage
high visceral adiposity
highest frequencies of PNPLA3 risk alleles (>70%)

To further explore the influence of genetic variants on fibrosis, we conducted an analysis examining the association between SNP genotypes and phenotypes in a subset of individuals. Cluster 4 (high cardiovascular–liver–kidney risk subgroup) exhibited the highest frequencies of risk alleles in PNPLA3, TM6SF2, and MBOAT7, followed by Cluster 2 (high fibrosis risk subgroup). PNPLA3 rs738409 C > G variant carriers showed a 3.2-fold increase in significant fibrosis among those with the PNPLA3 CG genotype and a 2.7-fold increase among those with the PNPLA3 GG genotype.

This classification facilitates the precise integration of MASLD risk stratification and management within the cardiovascular–liver–kidney–metabolic framework. Professor Bi emphasized: Our subtyping enables targeted interventions, for example, prioritizing fibrosis screening for Cluster 2 while implementing aggressive cardiorenal protection for Cluster 3–4. The algorithm-based stratification system represents a paradigm shift toward precision hepatology.

Reference
DOI: https://doi.org/10.1097/CM9.0000000000003984

Yan Bi from Nangjing Drum Tower Hospital
Vice President and Academic Leader of Endocrinology Department at Drum Tower Hospital affiliated to Nanjing University Medical School.
The member of Chinese Diabetes Society and the Candidate Chairman of Jiangsu Association of Diabetes.
Hosting National Natural Science Foundation Key Program and Original Exploration Program.
Focused on the clinical and basic research of diabetes and obesity, published 106 SCI papers, including Cell Metabolism, Diabetes Care, Nature Communications, Journal of Hepatology and other journals as the first or corresponding author.

Tianwei Gu from Nangjing Drum Tower Hospital
Associate Chief Physician, Associate Professor
Focus on diabetes and metabolic dysfunction-associated steatotic liver disease.
Published 16 SCI papers as first or corresponding author.
Hosting 3 projects from National Natural Science Foundation.

Yinghuan Shi from Nanjing University
Professor in the Department of Computer Science and Technology at Nanjing University
Focus on machine learning, pattern recognition, computer vision, and medical image analysis.
Hosting National Natural Science Foundation of China Excellent Young Scholars Fund, National Key Research and Development Program of China, and Science and Technology Innovation 2030 Major Project.