Dairy products intake and prevalence, incidence, and recovery of non-alcoholic fatty liver disease in Chinese population.

BACKGROUND: Non-alcoholic fatty liver disease (NAFLD) is a growing public health concern. Modifiable factors such as diet and lifestyle are of research interest in preventing or reversing the disease. The relationship between dairy products and NAFLD remains unclear. METHODS: In this cohort study, 36,122 participants aged 20-74 were enrolled by multi-stage, stratified, randomized cluster sampling from 2016 to 2017. A total of 25,085 participants finished at least one follow-up visit from 2019 to 2023. Dairy intake was collected by food frequency questionnaire at baseline. NAFLD was defined as fatty liver diagnosed by ultrasonography with excessive alcohol drink excluded. Logistic regression and Cox proportional hazard models were used to analyze the association between dairy intake and NAFLD. RESULTS: A total of 34,040 participants were included in the baseline analysis. The prevalence of NAFLD was inversely associated with dairy intake (OR>7vs 0 servings/week = 0.91, 95% CI 0.84-0.98; ORper serving/day increase = 0.95, 95% CI 0.92-0.99). 20,460 participants entered the follow-up analysis. Among 12,204 without NAFLD at baseline, 4,470 developed NAFLD after a median time of 4.3 years. The incidence of NAFLD was inversely associated with dairy intake (HR>7 vs 0 servings/week = 0.89, 95% CI 0.81-0.98; HRper serving/day increase = 0.94, 95% CI 0.89-0.99). Among 8256 with NAFLD at baseline, 3,885 recovered after 4.2-year follow-up. Total dairy intake did not show significant associations with recovery of NAFLD, and the HRs (95% CI) were 0.96 (0.87-1.06) for > 7 servings/week and 0.98 (0.93-1.03) for per serving/day increase. CONCLUSION: Dairy product intake of more than one serving per day was associated with a lower prevalence and incidence of NAFLD in Chinese population. However, total dairy intake did not show significant association in NAFLD reversal.

An adaptive weighted ensemble learning network for diabetic retinopathy classification.

Diabetic retinopathy (DR) is one of the leading causes of blindness. However, because the data distribution of classes is not always balanced, it is challenging for automated early DR detection using deep learning techniques. In this paper, we propose an adaptive weighted ensemble learning method for DR detection based on optical coherence tomography (OCT) images. Specifically, we develop an ensemble learning model based on three advanced deep learning models for higher performance. To better utilize the cues implied in these base models, a novel decision fusion scheme is proposed based on the Bayesian theory in terms of the key evaluation indicators, to dynamically adjust the weighting distribution of base models to alleviate the negative effects potentially caused by the problem of unbalanced data size. Extensive experiments are performed on two public datasets to verify the effectiveness of the proposed method. A quadratic weighted kappa of 0.8487 and an accuracy of 0.9343 on the DRAC2022 dataset, and a quadratic weighted kappa of 0.9007 and an accuracy of 0.8956 on the APTOS2019 dataset are obtained, respectively. The results demonstrate that our method has the ability to enhance the ovearall performance of DR detection on OCT images.