Machine learning predictors of cardiovascular autonomic neuropathy in type 2 diabetes: insights from a prospectively collected single-center registry
8 affiliations
- 1Department of Medicine, Albert Szent-Györgyi Medical School, University of Szeged
- 2Department of Genetics, Faculty of Science and Informatics, University of Szeged
- 3Department of Pharmacology and Pharmacotherapy, Albert Szent-Györgyi Medical School, University of Szeged
- 4Center of Reproductive Medicine, Department of Obstetrics and Gynecology, Albert Szent-Györgyi Medical School, University of Szeged
- 5Institute for Translational Medicine, Medical School, University of Pécs
- 6Centre for Translational Medicine, Semmelweis University
- 7Institute of Pancreatic Diseases, Semmelweis University
- 8Translational Pancreatology Research Group, Interdisciplinary Centre of Excellence for Research Development and Innovation, University of Szeged
Abstract
Background: Cardiac autonomic neuropathy (CAN) is a severe complication of type 2 diabetes (T2DM) associated with increased cardiovascular morbidity and mortality. Machine learning (ML) can integrate multidimensional clinical data and identify features associated with CAN. We aimed to characterize predictors of CAN categories using ML.
Methods: Data were obtained from a prospective single-center observational registry of patients with T2DM undergoing standardized cardiovascular autonomic reflex tests (CARTs) based on Ewing’s protocol. Demographic, anthropometric, clinical, laboratory, and medication data were analyzed. Patients were classified as CAN0 (no or borderline CAN), CAN1 (early CAN), or CAN2 (definite or severe CAN). Three extreme gradient boosting models were evaluated: a CART-inclusive model, a CART-excluded model restricted to patients with T2DM, and a CART-excluded model supplemented with non-diabetic controls without abnormal CART results. Model performance was assessed using area under the receiver operating characteristic curve (AUC), and feature contributions were interpreted using SHapley Additive exPlanations values.
Results: Among 693 patients with T2DM, 19 (2.7%) had completely normal autonomic function, whereas 674 (97.3%) exhibited at least one borderline or abnormal CART result. Greater CAN severity was associated with older age, longer diabetes duration, higher systolic blood pressure, and poorer renal function. The CART-inclusive model achieved perfect discrimination in the best iteration (weighted AUC, 1.00; accuracy, 100%), with all five CARTs ranking among the leading predictors. After CART exclusion, performance declined markedly (weighted AUC, 0.64), with class-specific AUCs of 0.59 for CAN0, 0.59 for CAN1, and 0.68 for CAN2. Inclusion of non-diabetic controls improved the weighted AUC to 0.78, with the greatest improvement for CAN0 (AUC, 0.85), followed by CAN2 (0.78) and CAN1 (0.70). Age, diabetes duration, and renal function markers were recurrent non-CART predictors, while anthropometric, hematological, metabolic, and treatment-related features showed category- and model-specific contributions. Uniform Manifold Approximation and Projection demonstrated substantial overlap among CAN categories, particularly for CAN1.
Conclusions: CARTs remained essential for accurate CAN classification and staging. Routinely available clinical variables provided complementary phenotypic information but were insufficient to robustly distinguish CAN categories without CART data, particularly CAN0 and CAN1. ML may support multidimensional phenotyping and hypothesis generation but cannot replace standardized autonomic testing.
Related
- Assessment of Handgrip Strength and Cardiac Autonomic Neuropathy Using Cardiovascular Autonomic Reflex Tests and Heart Rate Variability in Adults With Type 2 Diabetes Mellitus: A Single-Center Comparative Cross-Sectional Study
- Socioeconomic deprivation and postoperative mortality: a retrospective analysis of prospectively collected data
- Prognostic significance of indocyanine green fluorescence patterns in resected hepatocellular carcinoma: a retrospective analysis of prospectively collected data.
- Delayed Diagnosis of Hereditary Transthyretin Cardiac Amyloidosis Presenting with Autonomic Neuropathy: A Case Report
- Associations Between Upright Cardiovascular-Autonomic Characteristics, Temporal Gait Measures, and Mediolateral Plantar Center-of-Force Measures in Adults with Developmental Disabilities: An Exploratory Cross-Sectional Study
BibTeX
@article{farkas2026machine,
title = {{Machine learning predictors of cardiovascular autonomic neuropathy in type 2 diabetes: insights from a prospectively collected single-center registry}},
author = {Katalin Farkas and Viktor Vedelek and Máté Vámos and Éva Csajbók and István Baczkó and János Zádori and Andrea Szentesi and Péter Hegyi and Tamás Várkonyi and Csaba Lengyel and Anna Vágvölgyi},
journal = {Frontiers in Endocrinology},
year = {2026},
month = oct,
doi = {10.3389/fendo.2026.1965013},
url = {https://www.frontiersin.org/articles/10.3389/fendo.2026.1965013/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Machine learning predictors of cardiovascular autonomic neuropathy in type 2 diabetes: insights from a prospectively collected single-center registry
AU - Katalin Farkas
AU - Viktor Vedelek
AU - Máté Vámos
AU - Éva Csajbók
AU - István Baczkó
AU - János Zádori
AU - Andrea Szentesi
AU - Péter Hegyi
AU - Tamás Várkonyi
AU - Csaba Lengyel
AU - Anna Vágvölgyi
PY - 2026
DA - 2026/10/15
JO - Frontiers in Endocrinology
DO - 10.3389/fendo.2026.1965013
UR - https://www.frontiersin.org/articles/10.3389/fendo.2026.1965013/full
AB - Background: Cardiac autonomic neuropathy (CAN) is a severe complication of type 2 diabetes (T2DM) associated with increased cardiovascular morbidity and mortality. Machine learning (ML) can integrate multidimensional clinical data and identify features associated with CAN. We aimed to characterize predictors of CAN categories using ML. Methods: Data were obtained from a prospective single-center observational registry of patients with T2DM undergoing standardized cardiovascular autonomic reflex tests (CARTs) based on Ewing’s protocol. Demographic, anthropometric, clinical, laboratory, and medication data were analyzed. Patients were classified as CAN0 (no or borderline CAN), CAN1 (early CAN), or CAN2 (definite or severe CAN). Three extreme gradient boosting models were evaluated: a CART-inclusive model, a CART-excluded model restricted to patients with T2DM, and a CART-excluded model supplemented with non-diabetic controls without abnormal CART results. Model performance was assessed using area under the receiver operating characteristic curve (AUC), and feature contributions were interpreted using SHapley Additive exPlanations values. Results: Among 693 patients with T2DM, 19 (2.7%) had completely normal autonomic function, whereas 674 (97.3%) exhibited at least one borderline or abnormal CART result. Greater CAN severity was associated with older age, longer diabetes duration, higher systolic blood pressure, and poorer renal function. The CART-inclusive model achieved perfect discrimination in the best iteration (weighted AUC, 1.00; accuracy, 100%), with all five CARTs ranking among the leading predictors. After CART exclusion, performance declined markedly (weighted AUC, 0.64), with class-specific AUCs of 0.59 for CAN0, 0.59 for CAN1, and 0.68 for CAN2. Inclusion of non-diabetic controls improved the weighted AUC to 0.78, with the greatest improvement for CAN0 (AUC, 0.85), followed by CAN2 (0.78) and CAN1 (0.70). Age, diabetes duration, and renal function markers were recurrent non-CART predictors, while anthropometric, hematological, metabolic, and treatment-related features showed category- and model-specific contributions. Uniform Manifold Approximation and Projection demonstrated substantial overlap among CAN categories, particularly for CAN1. Conclusions: CARTs remained essential for accurate CAN classification and staging. Routinely available clinical variables provided complementary phenotypic information but were insufficient to robustly distinguish CAN categories without CART data, particularly CAN0 and CAN1. ML may support multidimensional phenotyping and hypothesis generation but cannot replace standardized autonomic testing.
SN - 1664-2392
ER -