Related Experiment Video
Updated: Jul 27, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Global left ventricular function in cardiac CT. Evaluation of an automated 3D region-growing segmentation algorithm
Georg Mühlenbruch1, Marco Das, Christian Hohl
1Department of Diagnostic Radiology, University Hospital RWTH Aachen, Pauwelsstrasse 30, 52057 Aachen, Germany. gmuehlen@rad.rwth-aachen.de
European Radiology
|December 24, 2005
Summary
A new semi-automated 3D region-growing algorithm accurately segments the left ventricle in cardiac CT scans. This method is faster and feasible for functional analysis, improving efficiency in clinical workflows.
Area of Science:
- Cardiovascular Imaging
- Medical Image Analysis
- Computational Anatomy
Background:
- Accurate functional analysis of the left ventricle is crucial for diagnosing and managing cardiovascular diseases.
- Multislice CT (MSCT) provides detailed anatomical information of the heart.
- Manual segmentation of the left ventricle can be time-consuming and subject to inter-observer variability.
Purpose of the Study:
- To evaluate a novel semi-automated 3D region-growing segmentation algorithm for left ventricular analysis using cardiac MSCT.
- To compare the accuracy and efficiency of the new algorithm against manual segmentation methods.
- To assess the feasibility of the algorithm for routine clinical application.
Main Methods:
- Twenty patients underwent contrast-enhanced cardiac MSCT.
- Semi-automated region-growing segmentation was applied to 1-mm axial slices.
- Left ventricular volumes (end-diastolic, end-systolic, ejection fraction, stroke volume) were compared to manual segmentation from short-axis slices.
- Post-processing time for both methods was recorded.
Main Results:
- The region-growing algorithm successfully segmented the left ventricle in 65% of patients.
- Volume measurements showed excellent correlation between the two methods (P
- The semi-automated method reduced post-processing time by 44.2% in successfully segmented cases.
- Higher signal-to-noise ratios were observed in patients with successful segmentation.
Conclusions:
- The semi-automated 3D region-growing algorithm is a technically feasible and accurate tool for left ventricular volume analysis.
- This method offers significant time-saving benefits compared to manual segmentation.
- The algorithm's effectiveness is dependent on the quality of the contrast-enhanced CT dataset.
