Automatic Identification of the Cardiac Rest Period Using Template Updating for Magnetic Resonance Coronary Angiography

Neural network-based fully automated cardiac resting phase detection algorithm compared with manual detection in patients

Background

A cardiac resting phase is used when performing free-breathing cardiac magnetic resonance examinations.

Purpose

The purpose of this study was to test a cardiac resting phase detection system based on neural networks in clinical practice.

Material and Methods

Four chamber-view cine images were obtained from 32 patients and analyzed. The rest duration, start point, and end point were compared between that determined by the experts and general operators, and a similar comparison was done between that determined by the experts and neural networks: the normalized root-mean-square error (RMSE) was also calculated.

Results

Unlike manual detection, the neural network was able to determine the resting phase almost simultaneously as the image was obtained. The rest duration and start point were not significantly different between the neural network and expert (p = .30, .90, respectively), whereas the end point was significantly different between the two groups (p < .05). The start point was not significantly different between the general operator and expert (p = .09), whereas the rest duration and end point were significantly different between the two groups (p < .05). The normalized RMSEs of the rest duration, start point, and end point of the neural network were 0.88, 0.64, and 0.33 ms, respectively, which were lower than those of the general operator (normalized RMSE values were 0.98, 0.68, and 0.51 ms, respectively).

Conclusions

The neural network can determine the resting phase instantly with better accuracy than the manual detection of general operators.

Velocity encoded mitral valve inflow cine: A novel and more reproducible method to determine cardiac rest periods during coronary magnetic resonance angiography

A high temporal resolution, 4-chamber (4CH) cine is the standard method for determining cardiac rest periods during whole heart coronary magnetic resonance angiography (CMRA). We evaluated the image quality and reproducibility between the 4CH cine method and a novel approach using a velocity encoded mitral valve inflow cine (MVI). The goal of this study was to compare the quality of CMRAs utilizing MVI versus 4CH methods. Sharpness and vessel length for the LCA and RCA using each method were determined using Soap Bubble and two blinded observers independently assessed coronary image quality. Offline analysis on a separate, retrospective cohort (n = 25) was used to compare MVI and 4CH reproducibility. In the prospectively evaluated cohort there was no difference in overall vessel sharpness (4CH vs MVI mean ± SD) (31.0 ± 5.5% vs 30.5 ± 5.7%, p = .63), LCA vessel sharpness (30.0 ± 5.4% vs 31.1 ± 8.2%, p = .44), LCA length (4.7 ± 1.4 cm vs 4.6 ± 1.6 cm, p = .66), RCA vessel sharpness (32.1 ± 6.9% vs 31.1 ± 7.7%, p = .55), RCA length (5.51 ± 2.6 cm vs 5.95 ± 2.4 cm, p = .38), or image quality rating (2.66 vs 2.62, p = .80) between methods. In the retrospective cohort, the MVI method had 5.4% lower inter-observer variability (95% CI 3.7,7.2%, p < .0001) and 3.9% lower intra-observer variability (95% CI 2.4,5.4%, p < .0001) than the 4CH method.

MVI is a technically feasible and more reproducible method to determine cardiac rest periods compared to 4CH while preserving vessel sharpness, vessel length & image quality.

Automated determination of cardiac rest period on whole-heart coronary magnetic resonance angiography by extracting high-speed motion of coronary arteries

PURPOSE

The aim of the present study was to develop an automated system for determining the cardiac rest period during whole-heart coronary magnetic resonance angiography (CMRA) examination.

MATERIALS AND METHODS

Ten healthy male volunteers (25-51 years old, 50-77 beats/min heart rate) were enrolled in this prospective study. A motion area map was generated from a cine image set by extracting high-speed component of cardiac motion, and it was used to specify the rest period in the proposed CMRA. In conventional CMRA, the rest period was determined based on the visual inspection of cine images. Agreement of the start time, end time, and trigger time between the two methods was assessed by the Bland-Altman plot analysis. Two observers visually evaluated the quality of the curved planar reformation (CPR) image of the coronary arteries.

RESULTS

The proposed method significantly prolonged the start time (mean systematic difference 37.7 ms, P < 0.05) compared with the conventional method. Good agreement was observed for the end time (mean systematic difference 8.9 ms) and trigger time (mean systematic difference -28.8 ms) between the two methods. A significantly higher image quality (P < 0.05) was provided for the left circumflex artery in the proposed CMRA (mean grading score 3.88) than in conventional CMRA (mean grading score 3.68).

CONCLUSION

Our system enabled detection of the rest period automatically without operator intervention and demonstrated somewhat higher image quality compared with conventional CMRA. Its use may be useful to improve the imaging workflow for CMRA in clinical practice.