A prognostic model based on metabolic indicators for recurrent acute pancreatitis in hypertriglyceridemia patients: a retrospective cohort study

Figures

• 

  Figure 1.. Flowchart. Retrospective selection process of patients.

• 

  Figure 2.. Nomogram for predicting recurrence of first-time HTG-AP in patients.

• 

  Figure 3.. Model-dependent K-M survival curve analysis.

• 

  Figure 4.. Case diagram under the overall indicator. Male patient, 47 years old, scored 255 points (90.7% diagnostic probability), indicating a high risk of recurrent hypertriglyceridemic acute pancreatitis (HTG-AP recurrence). Clinical presentation: Alcohol abuse, smoking, and visceral adiposity: 102.00 TYG: 11.51 On day 583 of follow-up, diagnosed with recurrent acute pancreatitis and admitted for treatment. This case demonstrates the model’s high-confidence capability in identifying recurrent AP cases.

• 

  Figure 5.. ROC curves of the HTG-AP recurrence prediction model and individual predictors over a 2-year period. The AUCs for the model, VATI, TyG index, smoking, and alcohol consumption were 0.927, 0.776, 0.843, 0.764, and 0.781, respectively.

• 

  Figure 6.. Model calibration validation. calibration curve of the AP recurrence model. The horizontal axis represents the AP recurrence probability calculated by the model, while the vertical axis displays the actual AP recurrence rate. The apparent calibration curve (dashed line) reflects model performance based on raw data, while the bias-corrected curve (solid line) shows model performance after optimistic adjustment via 1000-step guided resampling. The confidence interval for the apparent calibration curve is marked by light gray lines; perfect prediction should align perfectly with the 45-degree baseline (dashed line). (Schematic) (Insert).

• 

  Figure 7.. Model clinical utility assessment. Decision curve analysis (DCA) evaluates the clinical utility of the HTG-AP recurrence prediction model. The y-axis represents net benefit, quantifying the model’s clinical value, while the x-axis denotes threshold probabilities corresponding to clinical decision thresholds. The black horizontal line represents the assumption that no HTG-AP recurrence patients require treatment, whereas the solid gray line assumes all patients require treatment. The red curve reflects the performance of our HTG-AP recurrence model. When the threshold probability exceeds 0%, the model generates higher net benefit than either extreme strategy (black/gray lines), indicating its clinical utility in guiding interventions for patients at risk of HTG-AP recurrence.