bims-aukdir Biomed News
on Automated knowledge discovery in diabetes research
Issue of 2025–07–13
seven papers selected by
Mott Given
  1. Diabetic retinopathy detection using adaptive deep convolutional neural networks on fundus images.
  2. Applications of Artificial Intelligence and Machine Learning in Prediabetes: A Scoping Review.
  3. Evaluation of IDx-DR software for diabetic retinopathy screening in outpatient clinics: Efficacy, safety, and feasibility in a real-world setting.
  4. Optimized prediction of diabetes complications using ensemble learning with Bayesian optimization: a cost-efficient laboratory-based approach.
  5. A comparative approach of machine learning models to predict attrition in a diabetes management program.
  6. Enhancing diabetes risk prediction through focal active learning and machine learning models.
  7. The duplexity of insulin: The integrated bioinformatics analysis and machine learning identified key genes for type 2 diabetes.
  1. Sci Rep. 2025 Jul 09. 15(1): 24647
    Diabetic retinopathy detection using adaptive deep convolutional neural networks on fundus images.
    Rashid Abbasi, Farhan Amin, Amerah Alabrah, Gyu Sang Choi, Salabat Khan, Md Belal Bin Heyat, Muhammad Shahid Iqbal, Huiling Chen.
      Diabetic retinopathy (DR) is an age-related macular degeneration eye disease problem that causes pathological changes in the retinal neural and vascular system. Recently, fundus imaging is a popular technology and widely used for clinical diagnosis, diabetic retinopathy, etc. It is evident from the literature that image quality changes due to uneven illumination, pigmentation level effect, and camera sensitivity affect clinical performance, particularly in automated image analysis systems. In addition, low-quality retinal images make the subsequent precise segmentation a challenging task for the computer diagnosis of retinal images. Thus, in order to solve this issue, herein, we proposed an adaptive enhancement-based Deep Convolutional Neural Network (DCNN) model for diabetic retinopathy (DR). In our proposed model, we used an adaptive gamma enhancement matrix to optimize the color channels and contrast standardization used in images. The proposed model integrates quantile-based histogram equalization to expand the perceptibility of the fundus image. Our proposed model provides a remarkable improvement in fundus color images and can be used particularly for low-contrast quality images. We performed several experiments, and the efficiency is evaluated using a large public dataset named Messidor's. Our proposed model efficiently classifies a distinct group of retinal images. The average assessment score for the original and enhanced images is 0.1942 (standard deviation: 0.0799), Peak Signal-to-Noise Ratio (PSNR) 28.79, and Structural Similarity Index (SSIM) 0.71. The best classification accuracy is [Formula: see text], indicating that Convolutional Neural Networks (CNNs) and transfer learning are superior to traditional methods. The results show that the proposed model increases the contrast of a particular color image without altering its structural information.
    Keywords:  Artificial intelligence; Deep learning and Health; Machine learning; Medical imaging
    DOI:  https://doi.org/10.1038/s41598-025-09394-0
  2. J Diabetes Sci Technol. 2025 Jul 08. 19322968251351995
    Applications of Artificial Intelligence and Machine Learning in Prediabetes: A Scoping Review.
    Benjamin Lalani, Rohan Herur, Daniel Zade, Grace Collins, Devin M Dishong, Setu Mehta, Jalene Shim, Yllka Valdez, Nestoras Mathioudakis.
       INTRODUCTION: Prediabetes is a prevalent condition in which early detection and lifestyle interventions can prevent or delay progression to diabetes. Artificial intelligence (AI) and machine learning (ML) offer enhanced tools for diagnosis, risk stratification, and scalable delivery of lifestyle interventions. This review synthesizes current applications of AI/ML in patients with prediabetes.
    METHODS: We conducted a scoping review using PubMed, EMBASE, and Web of Science (through May 2025) to identify original studies applying AI/ML to prediabetes prediction or management. Population-level forecasting and models combining prediabetes with other conditions were excluded. Data were extracted via structured REDCap instruments and validated through secondary review. Descriptive statistics summarized findings.
    RESULTS: Of 2072 records screened, 149 studies met criteria: 118 prediction model studies, 20 intervention studies, and 11 miscellaneous. Machine learning models primarily targeted prediction of prediabetes, progression to diabetes, diabetic complications, and glucose metrics. Overall model performance was favorable (mean C-statistic 0.81), with random forests, neural networks, and support vector machines showing better performance. Only 20 studies reported external validation, few compared ML to standard risk tools, and data/code availability was limited. Six AI-based diabetes prevention programs showed positive clinical outcomes, though randomized controlled trial (RCT) evidence was limited. Three personalized nutrition interventions showed mixed efficacy.
    CONCLUSION: Most AI/ML research in prediabetes focused on predictive modeling, which shows promise but limited translation to real-world settings. Artificial intelligence-based interventions may scale behavioral change support but need further evaluation versus standard care. Future efforts should prioritize external validation, assess added value over standard tools, and address barriers to integration into care.
    Keywords:  artificial intelligence; diabetes; impaired glucose tolerance; lifestyle intervention; machine learning; prediabetes
    DOI:  https://doi.org/10.1177/19322968251351995
  3. J Diabetes Complications. 2025 Jul 02. pii: S1056-8727(25)00173-4. [Epub ahead of print]39(10): 109120
    Evaluation of IDx-DR software for diabetic retinopathy screening in outpatient clinics: Efficacy, safety, and feasibility in a real-world setting.
    Simone Leonora Huber, Verena Parzer, Bernhard Ludvik, Andreas Pollreisz, Nikolaus Mahnert, Johanna Maria Brix.
       AIMS: This study assessed the IDx-DR software's effectiveness as a diabetic retinopathy (DR) screening tool in a routine outpatient setting. It also evaluated the software's safety and feasibility.
    METHODS: RA prospectively planned analysis of patients with diabetes was conducted at the diabetes outpatient clinic of a specialized tertiary care center from March 2021 to October 2022. These patients underwent retinal imaging with IDx-DR during routine visits.
    RESULTS: The majority of the 996 included patients were female (53.1 %), and the median age was 61.1 years. Notably, 40.2 % had a BMI ≥ 30 kg/m2, and 35.8 % were active smokers. DR was detected by IDx-DR in n = 178 (26 %) of patients. 73.1 % of those patients were newly diagnosed with retinopathy without any history of retinopathy in the medical history (p < 0.001). DR patients were older (median 60.4 vs 57.5 years; p = 0.050), had higher HbA1c levels (7.0 % vs. 7.6 %; p < 0.001) and a higher frequency of ophthalmologic check-ups (p = 0.015). In multiple binary logistic regression, use of insulin (OR = 1.735 [1.096; 2.748], p = 0.19) and diabetes duration (medium vs. short: OR = 2.356 [1.36; 4.07], p = 0.002, long vs. short: OR = 1.776 [1.03; 3.06], p = 0.39) independently predicted DR, while age and sex were not significant predictors.
    CONCLUSIONS: This study supports integrating AI tools like IDx-DR in DR screening. It highlights IDx-DR's utility and efficacy in improving DR detection and patient care, suggesting its potential for broader, cost-effective screening in Austria and possibly elsewhere.
    Keywords:  AI-based diagnostics; Artificial intelligence; Automated screening; Diabetic eye disease; Diabetic retinopathy; IDx-DR software; Machine learning; Predictive analytics; Telemedicine; Vision loss prevention
    DOI:  https://doi.org/10.1016/j.jdiacomp.2025.109120
  4. Front Endocrinol (Lausanne). 2025 ;16 1593068
    Optimized prediction of diabetes complications using ensemble learning with Bayesian optimization: a cost-efficient laboratory-based approach.
    Dapeng Yan, Xiaohan Li, Yifan Wang, Zhikuang Cai.
       Background and objective: The increasing global prevalence of diabetes has led to a surge in complications, significantly burdening healthcare systems and affecting patient quality of life. Early prediction of these complications is critical for timely intervention, yet existing models often rely heavily on clinical indicators while underutilizing fundamental laboratory test parameters. This study aims to bridge this gap by leveraging the 12 most frequently tested laboratory indicators in diabetic patients to develop an optimized predictive model for diabetes complications.
    Methods: A comprehensive dataset was established through meticulous data collection from a high-volume tertiary hospital, followed by extensive data cleaning and classification. Various machine learning classifiers, including Random Forest, XGBoost, Support Vector Machine (SVM), and Multilayer Perceptron (MLP), were trained on this dataset to evaluate their predictive performance. We further introduced an ensemble learning model with Bayesian optimization to enhance accuracy and cost-efficiency. Additionally, feature importance analysis was conducted to refine the model by reducing testing costs while maintaining high predictive accuracy.
    Results: Our ensemble model with Bayesian optimization demonstrated superior performance, achieving over 90% accuracy in predicting various diabetic complications, with an outstanding 98.50% accuracy and 99.76% AUC for diabetic nephropathy. Feature correlation analysis enabled a refined model that not only improved predictive accuracy but also reduced overall medical costs by 2.5% through strategic feature elimination.
    Conclusions: This study makes three key contributions: (1) Development of a high-quality dataset based on fundamental laboratory indicators, (2) Creation of a highly accurate predictive model using ensemble learning and Bayesian optimization, particularly excelling in diabetic nephropathy prediction, and (3) Implementation of a cost-efficient diagnostic approach that reduces testing expenses without compromising accuracy. The proposed model provides a strong foundation for future research and practical clinical applications, demonstrating the potential of integrating machine learning with cost-conscious medical testing.
    Keywords:  Bayesian optimization; clinical laboratory indicators; cost-efficient diagnosis; diabetes complications; machine learning; predictive modeling
    DOI:  https://doi.org/10.3389/fendo.2025.1593068
  5. PLOS Digit Health. 2025 Jul;4(7): e0000930
    A comparative approach of machine learning models to predict attrition in a diabetes management program.
    Samantha Kanny, Grisha Post, Patricia Carbajales-Dale, William Cummings, Janet Evatt, Windsor Westbrook Sherrill.
      Approximately 11.6% of Americans have diabetes and South Carolina has one of the highest rates of adults with diabetes. Diabetes self-management programs have been observed to be effective in promoting weight loss and improving diabetes knowledge and self-care behaviors. The ability to keep vulnerable individuals in these programs is critical to helping the growing diabetic population. Utilizing machine learning is gaining popularity in healthcare settings. The objective of this study is to assess the effectiveness of several machine learning methods in predicting attrition from a diabetes self-management program, utilizing participant demographics and various evaluation measures. Data were collected from participants enrolled in Health Extension for Diabetes (HED). Descriptive statistics were used to examine HED participant demographics, while Mann-Whitney U tests and chi-square tests were used to examine relationships between demographics and pre-program evaluation measures. Through the various analyses, health-related measures - specifically the SF-12 quality of life scores, Distressed Communities Index (DCI) score, along with demographic factors (race, age, height, and educational attainment), and spatial variables (drive time to the nearest grocery store) emerged as influential predictors of attrition. However, the machine learning models showed poor overall performance, with AUC values ranging from 0.53 - 0.64 and F-1 scores between 0.19 - 0.36, indicating low predictive power. Among the models tested, XGBoost with downsampling yielded the highest AUC value (0.64) and a slightly higher F-1 score (0.36). To enhance model interpretability, SHAP (SHapley Additive exPlanations) was applied. While these models are not suitable for accurately predicting individual attrition risk in diabetes self-management programs, they identify potential factors influencing dropout rates. These findings underscore the difficulty for models to accurately predict health behavior outcomes, highlighting the need for future research to improve predictive modeling to better support patient engagement and retention.
    DOI:  https://doi.org/10.1371/journal.pdig.0000930
  6. PLoS One. 2025 ;20(7): e0327120
    Enhancing diabetes risk prediction through focal active learning and machine learning models.
    Wangyouchen Zhang, Zhenhua Xia, Guoqing Cai, Junhao Wang, Xutao Dong.
      To improve the effectiveness of diabetes risk prediction, this study proposes a novel method based on focal active learning strategies combined with machine learning models. Existing machine learning models often suffer from poor performance on imbalanced medical datasets, where minority class instances such as diabetic cases are underrepresented. Our proposed Focal Active Learning method selectively samples informative instances to mitigate this imbalance, leading to better prediction outcomes with fewer labeled samples. The method integrates SHAP (SHapley Additive Explanations) to quantify feature importance and applies attention mechanisms to dynamically adjust feature weights, enhancing model interpretability and performance in predicting diabetes risk. To address the issue of imbalanced classification in diabetes datasets, we employed a clustering-based method to identify representative data points (called foci), and iteratively constructed a smaller labeled dataset (sub-pool) around them using similarity-based sampling. This method aims to overcome common challenges, such as poor performance on minority classes and limited generalization, by enabling more efficient data utilization and reducing labeling costs. The experimental results demonstrated that our approach significantly improved the evaluation metrics for diabetes risk prediction, achieving an accuracy of 97.41% and a recall rate of 94.70%, clearly outperforming traditional models that typically achieve 95% accuracy and 92% recall. Additionally, the model's generalization ability was further validated on the public PIMA Indians Diabetes DataBase, outperforming traditional models in both accuracy and recall. This approach can enhance early diabetes screening in clinical settings, helping healthcare professionals reduce diagnostic errors and optimize resource allocation.
    DOI:  https://doi.org/10.1371/journal.pone.0327120
  7. Biochem Biophys Rep. 2025 Sep;43 102099
    The duplexity of insulin: The integrated bioinformatics analysis and machine learning identified key genes for type 2 diabetes.
    Nan Gao, Xiteng Chen, Jun Yang, Yuanfeng Jiang, Shaochong Bu, Xiaomei Bai, Zhenyu Kou, Chunjun Li, Fang Tian.
       Background: Insulin therapy is still the most important treatment for T2DM, but the discussion about whether insulin brings more benefits or harms to T2DM patients has not stopped. Therefore, we used high-throughput RNA sequencing to investigate the role of insulin in T2DM and its molecular changes.
    Method: We collected peripheral blood samples from 16 patients with T2DM, and performed RNA-seq on peripheral blood mononuclear cells. Bioinformatics analysis and machine learning were uesd to identify the key differential genes and transcription factor networks. In addition, we performed the flow cytometry and staining to observe ROS level and endothelial-monocyte adhesion in PBMCs of both groups.
    Results: A total of 529 differential genes were identified by bioinformatics analysis. 8 genes were identified as key genes, among which IL-6 had high importance in the random forest model. In transcription factor analysis, IL-6, RETN, CTSG and ELANE have abundant transcriptional regulatory relationships. Flow cytometry showed that ROS production, phagocytosis, leukocyte adhesion in insulin treatment group were lower than that in non-insulin treatment group.
    Conclusion: Insulin therapy is bidirectional, it can cause islet B cell damage and vascular complications, but also can reduce the level of inflammation and oxidative stress.
    Keywords:  Inflammation; Insulin; Machine learning; Oxidative stress; RNA-Seq; Type 2 diabetes mellitus
    DOI:  https://doi.org/10.1016/j.bbrep.2025.102099
This page is being maintained by Thomas Krichel. It was last updated on 2025‒07‒13 at 0:24.