JMIR Publications

ABSTRACT

Background:

Given the challenges faced during percutaneous coronary intervention (PCI) for heavily calcified lesions, accurately predicting PCI success is crucial for enhancing patient outcomes and optimizing procedural strategies.

Objective:

This study aimed to utilize machine learning to identify coronary angiographic vascular characteristics and PCI procedures associated with the immediate procedural success rates of PCI in patients exhibiting moderate-to-severe coronary artery calcification (MSCAC).

Methods:

The development cohort consisted of patients who underwent PCI between January 2017 and December 2018 including MSCAC patients and no or mild CAC patients, and an external testing cohort included MSCAC patients from January 2013 to December 2013. Six machine learning models were developed and validated in the external testing cohort, with SMOTE used for dealing with imbalanced data.

Results:

3271 patients with MSCAC met the inclusion criteria, with an overall immediate PCI success rate of 92.4%. The XGBoost model emerged as the most predictive, achieving an AUC of 0.984 in the development set and an AUC of 0.972 in the external testing dataset. The key predictive factors identified included lesion length, minimum lumen diameter, TIMI, chronic total occlusion, reference vessel diameter, and diffuse lesion. The use of modified balloons for the preparation of calcified lesions had a specific impact on MSCAC patients for PCI success with a SHAP value of 0.16.

Conclusions:

This study successfully revealed the important PCI failure risk factors, such as lesion length and modified balloons, utilizing machine learning models to help clinicians manage PCI strategies in patients with complex coronary artery disease such as MSCAC.