Visualized breath-hold training for reducing respiratory motion artifacts in liver dynamic contrast-enhanced magnetic resonance imaging

Abstract

Background: Liver dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) plays a critical role in detecting hepatic lesions but is highly susceptible to respiratory motion artifacts. Although advancements in rapid magnetic resonance (MR) acquisition techniques have been made, breath-hold acquisitions are predominantly used for abdominal DCE-MRI due to their superior image quality. Therefore, prescan breath-hold training is essential to improving patients' compliance and minimizing respiratory artifacts. Traditional oral training methods, which rely on operators monitoring chest/abdominal movement, are subjective and often insufficient for DCE-MRI requirements, leading to suboptimal image quality and potential risks from repeated contrast administration. To address these limitations, we developed a novel visual breath-hold training instrument and evaluated its efficacy in reducing respiratory artifacts.

Methods: This retrospective study included 174 patients undergoing liver DCE-MRI (pre-contrast, early arterial, late arterial, portal, and parenchymal phases). Patients were divided into two groups: the visualized breath-hold training (VBT) group (n=87) and the oral breath-hold training (OBT) group (n=87). In the VBT group, breath-hold training was performed with a self-designed visualized respiration training device. Operators objectively assessed training performance and provided individual guidance. In the OBT group, conventional training, with standard verbal instructions and tactile examination of the patients' chest and abdomen, was applied. For analyzing image quality, quantitative metrics [signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR)] and qualitative respiratory artifact scores were compared between groups with the Mann-Whitney U test. For the comparison of diaphragmatic stability across sequential breath-holds, slice pages of the top or bottom of liver were recorded and analyzed via the Mann-Whitney U test. A subgroup analysis was performed on non-contrast images from 18 patients, with respiratory artifacts being scored for paired intra-individual comparisons.

Results: The VBT group had a significantly higher SNR and CNR than did the OBT group in the pre-contrast (SNR: P=0.018; CNR: P=0.006), late arterial (SNR: P=0.016; CNR: P=0.029), portal (SNR: P=0.003; CNR: P=0.002), and parenchymal phases (SNR: P=0.044; CNR: P=0.010). Respiratory artifact scores were lower in the VBT group across all phases (pre-contrast: P=0.021; early arterial: P=0.002; late arterial: P=0.001; portal: P<0.001; parenchymal: P<0.001). Diaphragmatic consistency during sequential breath-holds was significantly higher in the VBT group between the early arterial and portal phases (P=0.046) and the late arterial and portal phases (P=0.011) as compared to the OBT group. Intra-individual subgroup analysis further confirmed the reduction of artifacts in the pre-contrast phase scans (P=0.010).

Conclusions: VBT effectively reduces respiratory motion artifacts and enhances image quality in liver DCE-MRI by improving patient compliance. This approach holds promise for optimizing clinical workflows and diagnostic accuracy, especially in populations with limited breath-hold capacity.

Keywords: Visualized breath-hold training (VBT); image quality; liver dynamic contrast-enhanced magnetic resonance imaging (liver DCE-MRI); respiratory artifacts.

Figure 1

The procedure of VBT and OBT. (A) The schematic diagram of VBT (the black band around the mid-abdomen was a pressure-sensitive band and the right panel was a monitor). (B) The flow chart of OBT and VBT. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; OBT, oral breath-hold training; VBT, visualized breath-hold training.

Figure 2

Common problems associated with poor breath-holds. (A) Failure to maintain breath-hold (black arrow: inhalation during breath-hold). (B) Delayed initiation of breath-holding (black arrow: continued inhalation after attempting to start the breath-hold). (C) Delayed initiation combined with slight inhalation during the breath-hold (black arrow: inhalation after initiation of breath-holding and slight breathing during breath-hold). (D) End-inspiration breath-hold (black arrow: inhalation before breath-holding).

Figure 3

The criteria of the respiratory artifacts scores. The representative examples (A-E) of respiratory artifacts scores [1–5] in the pre-contrast phase.

Figure 4

Pair-wise intra-individual comparison of respiratory artifacts in liver DCE-MRI. Two multiple-phases liver DCE-MRI scans (pre-contrast, early arterial phase, late arterial phase, portal phase, and parenchymal phase) at an interval of 3 months in a 59-year-old woman who underwent OBT [(A-E) scores of respiratory artifacts: 3, 2, 2, 2, and 2, respectively] and VBT [(F-J) scores of respiratory artifacts: 1, 1, 1, 1, and 1, respectively]. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; OBT, oral breath-hold training; VBT, visualized breath-hold training.

Figure 5

Quantitative analysis of image quality. (A) The SNR and (B) muscle CNR of the liver parenchyma on multiple-phase liver DCE-MRI for the OBT group and VBT group. *, P<0.05 (statistically significant difference). CNR, contrast-to-noise ratio; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; OBT, oral breath-hold training; SNR, signal-to-noise ratio; VBT, visualized breath-hold training.