Features and interaction between features of liver disease is of great significance for the classification of liver disease. Based on least absolute shrinkage and selection operator (LASSO) and interaction LASSO, the generalized interaction LASSO model is proposed in this paper for liver disease classification and compared with other methods. Firstly, the generalized interaction logistic classification model was constructed and the LASSO penalty constraints were added to the interactive model parameters. Then the model parameters were solved by an efficient alternating directions method of multipliers (ADMM) algorithm. The solutions of model parameters were sparse. Finally, the test samples were fed to the model and the classification results were obtained by the largest statistical probability. The experimental results of liver disorder dataset and India liver dataset obtained by the proposed methods showed that the coefficients of interaction features of the model were not zero, indicating that interaction features were contributive to classification. The accuracy of the generalized interaction LASSO method is better than that of the interaction LASSO method, and it is also better than that of traditional pattern recognition methods. The generalized interaction LASSO method can also be popularized to other disease classification areas.
Objective To identify potential clinical phenotypes in immunoglobulin G4-related disease (IgG4-RD) and to characterize differences in immune and metabolic profiles across distinct phenotypic subgroups. Methods A total of 125 patients diagnosed with IgG4-RD at West China Hospital, Sichuan University between January 2020 and December 2024 were retrospectively selected, and data regarding 12 organs (including the prostate) were collected. The prostate data were utilized solely for descriptive statistical purposes. Latent class analysis (LCA) was conducted using 11 organ variables, excluding the prostate, to perform a data-driven unsupervised classification of multiple organ involvement patterns. The optimal model was selected by comprehensively considering model fit indices, class size distribution, and clinical interpretability. Serum levels of immunoglobulin (Ig)G4, IgE, and uric acid were compared across different phenotypes. Results Among 125 patients with IgG4-RD, 89 were male (71.2%) and 36 were female (28.8%). Involvement of the pancreas was observed in 25 cases (20.0%), lymph nodes in 23 cases (18.4%), biliary system in 18 cases (14.4%), lungs in 16 cases (12.8%), lacrimal glands in 13 cases (10.4%), salivary glands and retroperitoneal fibrosis in 11 cases each (8.8%), eyes and pituitary glands in 7 cases each (5.6%), kidneys and aorta in 6 cases each (4.8%). LCA supported a three-class solution as optimal. Three distinct clinical phenotypes were identified: Class 1 (pulmonary-involvement type, n=26), characterized by frequent lung and lymph node involvement; Class 2 (mild systemic involvement type, n=89), showing limited multi-organ engagement; and Class 3 (glandular-lymphatic type, n=10), defined by predominant involvement of salivary glands, lacrimal glands, and lymph nodes, features resembling Mikulicz’s disease. Immunological and metabolic analysis revealed that the serum IgG4 level in Group 3 was higher than that in Group 1 and Group 2, but there was no statistically significant difference between the three groups (P>0.05). Spearman’s rank correlation analysis indicated a positive correlation between uric acid and IgG4 levels (r=0.2454, P=0.0067) as well as between IgE and IgG4 levels (r=0.4169, P<0.001). Conclusions Three distinct intrinsic clinical phenotypes of IgG4-RD were identified by LCA. Each phenotype exhibits characteristic patterns of organ involvement, along with specific immune and metabolic profiles. This suggests that interactions between metabolic and immune pathways may contribute to phenotypic differentiation and disease progression.