CT evaluation prior to transapical aortic valve replacement: semi-automatic versus manual image segmentation

Area of Science:

• Cardiovascular Imaging
• Radiology
• Medical Device Technology

Background:

• Transcatheter aortic valve replacement (TAVR) requires precise pre-procedural imaging analysis.
• ECG-triggered cardiovascular computed tomography (CT) is crucial for TAVR planning.
• Segmentation accuracy and speed impact procedural efficiency.

Purpose of the Study:

• To compare semi-automatic versus manual segmentation performance in ECG-triggered CT for TAVR.
• To evaluate the impact of observer experience (novice vs. expert) on segmentation speed and precision.
• To assess inter-software and inter-observer reliability.

Main Methods:

• Retrospective analysis of 30 TAVR patient CT datasets.
• Evaluation by two radiologists (1 and 5 years experience) using software with and without semi-automatic TAVR workflow.
• Assessment of time for data loading and segmentation steps.
• Inter-software and inter-observer reliability analysis.

Main Results:

• Semi-automatic segmentation improved evaluation speed by 10.3-12.3% for both novice and expert observers.
• Time savings were statistically significant (p < 0.001) for both groups.
• Excellent inter-software and inter-observer reliability was observed (mean ICC 0.95).
• Automatic aortic annulus detection had a 6.7% failure rate.

Conclusions:

• Semi-automatic segmentation offers a significant speed advantage for TAVR planning using cardiovascular CT.
• The method provides comparable precision to manual segmentation.
• Benefits are evident for both experienced and inexperienced radiologists, enhancing workflow efficiency.