Impact of Artificial Intelligence-Enhanced Optical Coherence Tomography Software on Percutaneous Coronary Intervention Decisions

Abstract

Background

Integration of intravascular imaging into percutaneous coronary intervention (PCI) workflow demands physician time and expertise. Artificial intelligence (AI)-enabled software that automates the identification of key intravascular imaging parameters has the potential to streamline physician workflow, increase accuracy, and reduce variability in PCI planning decisions. This study investigated if AI-enabled software, Ultreon (Abbott), compared with traditional software, AptiVue (Abbott), improved physician decision-making accuracy, variability, and efficiency in optical coherence tomography (OCT)-based PCI planning.

Methods

In this multireader, multicase study, 30 interventional cardiologists of varying OCT imaging experience evaluated 21 pre-PCI OCT pullbacks using both Ultreon and AptiVue platforms. Physician PCI planning decisions about lesion morphology, length, and diameter were compared to published best practices. Decision accuracy, variability, and time efficiency were assessed using statistical models.

Results

Physician OCT-based planning decisions were more accurate using Ultreon compared to AptiVue in the identification of calcium severity by 1.77 (95% CI, 1.27-2.50; P < .001), vessel preparation strategy by 2.00 (95% CI, 1.12-3.4; P = .018), and stent diameter by 2.83 (95% CI, 1.79-4.50; P < .001). Physicians exhibited less variability in assessments using Ultreon, especially for distal and proximal stent landing zone, and planned stent length (P < .0001). The efficiency of OCT assessments was improved with Ultreon, reducing the duration of OCT assessments by 0.5 minutes (P < .0001). The benefits were observed irrespective of the physician’s prior OCT experience.

Conclusions

Physician OCT-based PCI planning decisions were more accurate, less variable, and more efficient with AI-enhanced Ultreon software. This could potentially aid in the fuller adoption of intravascular imaging in PCI workflow.