Overview

Every year, over three million individuals globally receive stents to alleviate blocked blood vessels due to heart disease. However, tracking the healing process post-implantation poses significant challenges. Irregular tissue growth over the stent can lead to complications such as re-narrowing or complete blockage of the blood vessel. Currently, analyzing healing patterns through intravascular optical coherence tomography (OCT) images is a lengthy and impractical process for routine clinical use.

Development of DeepNeo

A research team from Helmholtz Munich and TUM University Hospital has introduced DeepNeo, an AI algorithm capable of automatically evaluating stent healing in OCT images. Key features include:

• Accurate differentiation of various healing patterns, matching the precision of clinical experts.
• Significantly reduced analysis time.
• Detailed measurements of tissue thickness and stent coverage, aiding patient management.

Research Findings

According to Valentin Koch, the study’s lead author, “DeepNeo allows for an automated, standardized, and highly accurate analysis of stent and vascular healing that was previously reliant on extensive manual effort.” He emphasized that the AI tool operates as effectively as a doctor but at a much faster pace.

Training and Testing

To develop DeepNeo, researchers utilized 1,148 OCT images from 92 patient scans, which were manually annotated to classify different tissue growth types. The algorithm was tested in an animal model, achieving an 87% accuracy in identifying unhealthy tissue compared to the current laboratory analysis standard. In human scans, DeepNeo also demonstrated high precision, closely aligning with expert evaluations.

Future Integration

Dr. Carsten Marr, Director at the Institute of AI for Health at Helmholtz Munich, stated, “DeepNeo illustrates how machine learning can assist clinicians in making quicker, more informed treatment decisions.” His colleague, Prof. Julia Schnabel, envisions DeepNeo as a component of an AI-driven healthcare system that could enhance clinical decision-making.

Support and Impact

The project has been awarded a Helmholtz Innovation Grant, and a patent application is underway. The team is collaborating with Ascenion, a technology transfer partner, to identify potential industry collaborators. According to cardiologists PD Dr. med. Philipp Nicol and Prof. Dr. med. Michael Joner, “DeepNeo standardizes OCT imaging assessments post-stent implantation, improving clinical decision-making. This advancement could lower healthcare costs and lead to more effective, personalized cardiovascular treatments.”