Fig. 6: Validation of disease risk patterns and progression trajectories of T2D in the Chinese tree models. | Nature Communications

Fig. 6: Validation of disease risk patterns and progression trajectories of T2D in the Chinese tree models.

From: Precision phenotyping of type 2 diabetes in chinese populations using a variational autoencoder-informed tree model

Fig. 6: Validation of disease risk patterns and progression trajectories of T2D in the Chinese tree models.

a Visualization of outcomes associated with T2D in the Chinese Tree based on the JADE validation dataset. Each point represents a hazard ratio (HR), and horizontal error bars indicate the corresponding 95% confidence intervals (CI). Predicted 10-year probabilities and HRs with 95% CI of myocardial infarction (MI, n = 28), heart failure (HF, n = 32), stroke (n = 50), cirrhosis (n = 4), chronic kidney disease (CKD, n = 250), and end-stage renal disease (ESRD, n = 38) following T2D diagnosis are shown for the JADE validation dataset. b Predicted probability and hazard ratios (HRs) with 95% CI of DR (87 events) in the DR cohort (n = 1006) on the Chinese Tree. Each point represents a hazard ratio (HR), and error bars denote the 95% confidence intervals. c Sensitivity analysis excluding patients with hepatitis at baseline in the CRDS cohort, FIB-4 for liver fibrosis and AST/ALT ratio marker for steatotic liver disease (n = 24,186). d, e Predicted probability of MASLD (6097 events) and cirrhosis (220 events) at 10 years from the diagnosis of T2D excluding patients with hepatitis (n = 24,186). f Sankey diagram showing the 5-year trajectory of changes classified into seven groups based on the Chinese tree coordinates. Seven groups were manually defined based on approximate regions in the tree coordinates to facilitate visualization of potential transitions.

Back to article page