Huasheng Lv,1,2 Fengyu Sun,1,2 Yuchen Zhang,1,2 Xinrong Zhou1,2 

1State Key Laboratory of Pathogenesis, Prevention and Treatment of High Incidence Diseases in Central Asia, Xinjiang Medical University, Urumqi, Xinjiang, People’s Republic of China; 2The First Affiliated Hospital of Xinjiang Medical University, Urumqi, Xinjiang, People’s Republic of China

Correspondence: Xinrong Zhou, Email Dochealcare@163.com

Objective: To develop and validate a machine learning (ML) model for predicting long-term depression risk in ACS patients following percutaneous coronary intervention (PCI).
Methods: This retrospective cohort study included 1951 ACS patients who underwent PCI in 2023. Feature selection was conducted using the Boruta algorithm, and restricted cubic spline (RCS) analysis was applied to assess non-linear associations. Six ML models were trained and tested using a 70:30 train-validation split. Model performance was evaluated using Area under the curve(AUC), sensitivity, specificity, F1-score, calibration curves, and decision curve analysis. SHapley Additive exPlanations (SHAP) were used to interpret feature contributions.
Results: Among the 1951 patients, 382 (19.6%) developed long-term depression. After feature selection via the Boruta algorithm, ten key predictors were identified, including NYHA classification, diabetes, thyroid-stimulating hormone (TSH), and left ventricular ejection fraction (LVEF). The LGBM and XGBoost models achieved the highest discrimination, with AUCs of 0.849 (training) and 0.652 (validation) for LGBM, and 0.814 (training) and 0.699 (validation) for XGBoost. Calibration curves showed good alignment between predicted and observed outcomes. SHAP analysis confirmed NYHA classification, TSH, and diabetes as the most influential features. Decision curve analysis demonstrated the clinical benefit of both models across a range of thresholds.
Conclusion: The models demonstrated potential for early risk stratification of post-PCI depression and may inform targeted clinical interventions.
Plain Language Summary: Depression is a common but often overlooked problem in people recovering from heart procedures. Patients with acute coronary syndrome (ACS) who undergo percutaneous coronary intervention (PCI)—a procedure to open blocked arteries—face a higher risk of long-term depression, which can worsen heart health and reduce quality of life.
In this study, we used machine learning (ML), a type of artificial intelligence, to analyze the medical data of 1951 patients who had PCI. Our goal was to predict who might develop depression in the year after their procedure. We tested six ML models and found that two of them—Light Gradient Boosting Machine (LGBM) and XGBoost—performed the best. These models used routine clinical information like heart function, thyroid levels, and diabetes status to make predictions.
The results showed that about 1 in 5 patients developed depression. The ML models were able to highlight high-risk individuals early, which could help doctors provide mental health support sooner.
This research shows how advanced data analysis can improve patient care after heart procedures. By identifying people at risk earlier, we can move toward more personalized and proactive treatment—not just for the heart, but for mental health too.

Keywords: acute coronary syndrome, percutaneous coronary intervention, depression, machine learning, predictive model