Clinical Impact of Intravascular Ultrasound-Based Artificial Intelligence Technologies (INNOVATE-PCI)

Overview

This study is a prospective, multicenter study in the real practice to validate the diagnostic performances and clinical impact of coronary angiography \& intravascular ultrasound (IVUS)-based models developed by machine learning (ML).

The aim of the study is to evaluate the performances and prognostic impact of coronary angiography \& IVUS-based algorithms for decision making and stent optimization in a multicenter, prospective cohort. Between January 2020 and June 2027, a total of 3,000 patients who performed coronary angiography (± FFR) and have at least one coronary stenosis requiring PCI (as culprit) will be enrolled from 16 centers in South Korea. In addition, the deferred lesions with visual estimated diameter stenosis of \>30% will be evaluated as non-culprits. Brief study design is as depicted in the following figure. Supervised ML algorithms include: 1) angiography- and IVUS-based algorithms for predicting FFR, 2) IVUS-based algorithm for plaque characterization, 3) IVUS-based algorithm for predicting stent expansion, and 4) post-stenting IVUS-based algorithm for predicting stent failure. In the prospective cohort, the performance of each model will be assessed. This registry trial composed of the treated (culprit) and the deferred (nonculprit) coronary lesions has two primary objectives as follow; 1) Primary objectives in treated (culprit) lesions is to see the impact of the integrated ML model on the development of culprit-related 2-year target vessel failure (TVF). 2) Primary objectives in deferred (nonculprit) lesions is to see the impact of the integrated ML model on the development of nonculprit-related 2-year TVF.

Conditions

Interventions

Eligibility

Locations (16)

Outcomes

Central Contacts