Ultrafast Computation of Left Ventricular Ejection Fraction by Using Temporal Intensity Variation in Cine Cardiac...

Area of Science:

• Cardiovascular Imaging
• Medical Physics
• Computational Biology

Background:

• Cardiac magnetic resonance (CMR) offers comprehensive cardiac evaluation but faces challenges due to extensive time and labor requirements for data acquisition and processing.
• These demands have limited widespread clinical adoption of CMR.
• Developing efficient image-processing algorithms is crucial for enhancing CMR accessibility.

Purpose of the Study:

• To develop and evaluate an ultrafast image-processing algorithm for computing left ventricular ejection fraction (LVEF) with minimal user interaction.
• To assess the algorithm's accuracy and performance against expert analysis using cine balanced steady-state free precession (bSSFP) short-axis images.
• To determine the algorithm's suitability for both pre- and post-contrast enhanced cardiac MRI.

Main Methods:

• An ultrafast algorithm utilizing temporal intensity variation in cine bSSFP short-axis images was developed to compute LVEF.
• The algorithm's performance was evaluated against an expert observer's analysis for left ventricular (LV) cavity segmentation in 65 participants.
• Bland-Altman analysis was employed to quantify differences in volumetric measurements and LVEF, with computation time recorded.

Main Results:

• The algorithm achieved accurate endocardial boundary delineation in 93% of slices (1,132 of 1,216).
• Bland-Altman analysis showed mean differences of <1.2 mL for end-diastolic volume, <5 mL for end-systolic volume, and <3% for LVEF compared to expert analysis.
• Total computation time for the LV stack was less than 2.5 seconds, enabling real-time processing on a commercial scanner console.

Conclusions:

• Temporal intensity variation-based ultrafast LVEF computation is clinically accurate across diverse LV shapes and wall motions.
• The algorithm is suitable for post-contrast cine SSFP imaging, offering a significant reduction in processing time.
• This automated, rapid LVEF calculation method enhances the clinical utility and accessibility of cardiac MRI.