EUVIMEDEUVIMED entdecken
Open Access

Predicting the Timing of the Metabolic Inflection Point in Type 1 Diabetes Progression Using Machine Learning and Survival Analysis Models

Eslam Montaser iD, Jay M Sosenko iD, Heba M Ismail iD

Diabetes · 01.06.2026 · 75 (6) · 901-910

DOI: 10.2337/db25-0961 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Diabetes, online erschienen am 20.03.2026. Verfasst von Eslam Montaser und 2 weiteren Autor:innen. Beteiligte Einrichtung: Division of Endocrinology and Metabolism, Indiana University School of Medicine, Indianapolis, IN. Hauptthemen laut MEDLINE: Diabetes Mellitus, Type 1, Machine Learning. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 21 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NIDDK NIH HHS, National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

An inflection point (IP) marking accelerated β-cell decline occurs ∼1–2 years before type 1 diabetes diagnosis. Precisely determining this timing could optimize clinical intervention. We developed machine learning models to predict proximity to the inferred metabolic IP using oral glucose tolerance test (OGTT) data from islet autoantibody-positive individuals in the TrialNet Pathway to Prevention study. OGTTs were retrospectively labeled by estimating time to the IP using thresholds from 1.2 to 1.6 years. Models were trained on engineered glucose and C-peptide dynamics, slopes, and composite indices, with feature selection by recursive feature elimination (RFE). Classifiers included support vector machines (SVM), random forests, and gradient boosting; a Cox proportional hazards model was also applied to estimate visit-level time to diagnosis and derive time-to-IP predictions. External validation was performed in the independent Diabetes Prevention Trial–Type 1 cohort. The best-performing model—SVM with RFE at the 1.4-year threshold—achieved an area under the curve of 0.77 (95% CI 0.72–0.82). The Cox model provided complementary numeric estimates of time to diagnosis and time to IP, sensitive to the cohort-level lead time offset (Δ). These findings show that machine learning and survival analysis can support early metabolic shift detection, enabling timely risk stratification and personalized monitoring in at-risk individuals. Article Highlights We undertook this study to improve early identification of the metabolic inflection point (IP) preceding clinical type 1 diabetes in autoantibody-positive individuals. We aimed to develop and validate machine learning models using oral glucose tolerance test–derived dynamic features to detect proximity to the IP. A support vector machine trained on TrialNet Pathway to Prevention and tested on Diabetes Prevention Trial–Type 1 achieved an area under the curve of 0.77 at 1.4 years prior to diagnosis, with strong calibration and interpretability. Additionally, a Cox proportional hazards model provided numeric estimates of time to IP, offering complementary predictions. These results can support earlier intervention and timely monitoring through personalized oral glucose tolerance test–based risk stratification.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Diabetes Mellitus, Type 1Machine LearningHumansDisease ProgressionGlucose Tolerance TestFemaleSurvival AnalysisProportional Hazards ModelsRandom ForestMalePredictive Learning ModelsInsulin-Secreting CellsBlood GlucoseRetrospective StudiesPrediction AlgorithmsBoosting Machine Learning AlgorithmsSupport Vector MachineAutoantibodies

Substanzen

Blood GlucoseAutoantibodies

Autor:innen und Institutionen

  • Division of Endocrinology and Metabolism, Indiana University School of Medicine, Indianapolis, IN.
  • Division of Endocrinology, University of Miami, Miami, FL.
  • Department of Pediatrics, Indiana University School of Medicine, Indianapolis, IN.

ORCID-Profile: Eslam Montaser, Jay M Sosenko, Heba M Ismail

Förderung

  • NIDDK NIH HHS
  • National Institute of Allergy and Infectious Diseases
  • National Institutes of Health (NIH)
  • Breakthrough T1D
  • Eunice Kennedy Shriver National Institute of Child Health and Human Development
  • National Institute of Diabetes and Digestive and Kidney Diseases

Zitieren

APA

Montaser, E., Sosenko, J. M., & Ismail, H. M. (2026). Predicting the Timing of the Metabolic Inflection Point in Type 1 Diabetes Progression Using Machine Learning and Survival Analysis Models. Diabetes, 75(6), 901-910. https://doi.org/10.2337/db25-0961

Vancouver

Montaser E, Sosenko JM, Ismail HM. Predicting the Timing of the Metabolic Inflection Point in Type 1 Diabetes Progression Using Machine Learning and Survival Analysis Models. Diabetes. 2026;75(6):901-910. doi:10.2337/db25-0961

BibTeX
@article{Montaser2026,
  author = {Montaser, Eslam and Sosenko, Jay M and Ismail, Heba M},
  title = {Predicting the Timing of the Metabolic Inflection Point in Type 1 Diabetes Progression Using Machine Learning and Survival Analysis Models},
  journal = {Diabetes},
  year = {2026},
  volume = {75},
  number = {6},
  pages = {901--910},
  doi = {10.2337/db25-0961}
}
Kostenlos starten →