Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 10, 2025JMIR AgingOpen Access

AI model accurately predicts biological age using 27 routine clinical factors.

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Current biological age prediction models are limited by reliance on conventional statistical methods and constrained clinical information.

Population

Koreans undergoing health checkups at Seoul National University Hospital Gangnam Center and from the Korean Genome and Epidemiology Study

Comparison

Multiple machine learning algorithms incorporating 27 clinical factors

Design

Model development and validation study

Key result

An artificial intelligence-driven Gradient Boosting model accurately predicted biological age using 27 routine clinical factors, achieving a mean squared error of 4.219 and an R2 of 0.967.

Authors

CJChang-Uk JeongJLJacob S. LeibyDKDokyoon Kim

Discussion

Loading...

Member takes

Overview

High-accuracy AI biological age prediction from routine data is feasible; leaves open prospective outcome validation before clinical use.

Study Design

Type

Cohort (n=40,385)

Multicenter

Yes

Structured PICO

P
Population
Koreans who underwent health checkups at the Seoul National University Hospital Gangnam Center and from the Korean Genome and Epidemiology Study
I
Intervention
Artificial intelligence-driven biological age prediction model incorporating 27 clinical factors
O
Outcome
Model performance evaluated using adjusted R2 and mean squared error (MSE)surrogate

Main Result

Absolute Event Rate: 4.219% vs 8.244%

An AI-driven biological age prediction model using routine health checkup data demonstrated high accuracy and clinical relevance for personalized health monitoring.

Limitations

  • Predominantly Korean population limits generalizability to other ethnic groups.
  • Reliance on clinical data from health checkups may not capture genetic, epigenetic, or environmental influences.
  • Need for longitudinal studies to validate long-term predictive capabilities.

Cite This Study

Jeong et al. (2025) conducted a cohort in Healthy adults (n=40,385). Gradient Boosting biological age prediction model vs. Support Vector Machine (SVM) model was evaluated on Mean squared error (MSE) for biological age prediction. An artificial intelligence-driven Gradient Boosting model accurately predicted biological age using 27 routine clinical factors, achieving a mean squared error of 4.219 and an R2 of 0.967.

synapsesocial.com/papers/6a12f53e06ed52b5c2c0d46ahttps://doi.org/10.2196/64473
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Development of an AI-Driven Biological Age Prediction Model Using Comprehensive Health Check-Up Data (Preprint)2024
  2. 2Prediction of biological age using machine learning2025 · 1 citations
  3. 3Artificial intelligence approaches in biological age prediction: current status and challenges2026
  4. 4Blood biochemical and gut microbiotic neural network models forecasting human biological age2026
  5. 5A sex-adjusted 7-biomarker clinical aging clock for translational preventative medicine2025 · 2 citations