Predicting NAFLD severity: the diagnostic edge of ultrasound-guided attenuation parameter and key biomarkers

To investigate the diagnostic value of ultrasound-guided attenuation parameter (UGAP) combined with clinical characteristics of patients with non-alcoholic fatty liver disease (NAFLD) for the severity of their condition. Two hundred NAFLD patients admitted to our hospital were retrospectively selected as study subjects, with the selection period spanning from March 2022 to July 2024. According to the degree of hepatic steatosis, patients were divided into mild group (degeneration degree ≤ 33%) with 93 cases, the moderate group (degeneration degree 34–66%) with 68 cases, and the severe group (degeneration degree ≥ 67%) with 39 cases. Transient elastography examination was performed to record the patient's UGAP. A logistic regression model was used to analyze the factors affecting the severity of NAFLD patients' condition. LSM was incorporated as an alternative indicator for fibrosis stratification into the multivariate model, and subgroup analysis was conducted based on the risk of fibrosis. Pearson correlation test was used to analyze the correlation between UGAP and clinical indicators. The receiver operating characteristic curve (ROC) was used to analyze the predictive value of each indicator alone and in combination for severe NAFLD. Two independent Logistic regression models were constructed: Model 1 included clinical indicators (BMI, ALT, AIP, PAI-1), and Model 2 added UGAP to Model 1. The incremental diagnostic value of UGAP was assessed by comparing the area under the curve, net reclassification index (NRI), and integrated discrimination improvement (IDI) between the two models. BMI, ALT, AIP, PAI-1, and UGAP were all independent influencing factors for the severity of NAFLD ( P < 0.05). UGAP was an independent influencing factor in both the low fibrosis risk group (OR = 1.752, 95% CI 1.204–2.550) and the high fibrosis risk group (OR = 1.841, 95% CI 1.276–2.657) after stratification by fibrosis. Pearson correlation analysis confirmed that UGAP was positively associated with BMI ( r = 0.156, P = 0.028), ALT ( r = 0.236, P = 0.001), AIP ( r = 0.314, P < 0.001 ), and PAI-1 ( r = 0.577, P < 0.001 ). ROC curve analysis showed that the Model 1 (clinical indicators) had an AUC of 0.824, and the Model 2 (clinical indicators plus UGAP) had an AUC of 0.920 for predicting severe NAFLD, with ΔAUC = 0.096 ( P = 0.018), the NRI was 0.312 (95% CI 0.108–0.516, P = 0.003), and the IDI was 0.087 (95% CI 0.034–0.140, P = 0.001). BMI, ALT, AIP, PAI-1, and UGAP were all independent influencing factors on disease severity of NAFLD. Adding UGAP to clinical indicators significantly improved the predictive efficacy for severe NAFLD. Both NRI and IDI confirmed the incremental diagnostic value of UGAP.