Four dimensional magnetic resonance imaging optimization and implementation for magnetic resonance imaging simulation

Purpose: Precise radiation therapy for abdominal lesions is complicated by respiratory motion and suboptimal soft tissue contrast from 4-dimensional (4D) computed tomography, whereas 4D magnetic resonance imaging MRI (4DMRI) provides superior tissue discrimination. This work evaluates a novel 4DMRI algorithm for motion management in radiation therapy. Methods and materials: Respiratory-triggered, T2-weighted, single-shot 4DMRI was evaluated for an open 1.0T magnetic resonance simulation platform. An in-house programmable platform was devised that translated objects for a variety of breathing patterns. Coronal 4DMRIs were acquired to evaluate the impact of number of phases on excursion and scan time. The impact of breathing period and regularity on scan time was assessed. A novel clinical 4D prototype phantom was scanned to characterize excursion and absolute volume differences between phase acquisitions. Optimized parameters were applied to abdominal 4DMRIs of 5 volunteers and 2 abdominal cancer patients on an institutional review board-approved protocol. Duty cycle, scan time, and waveform analysis were evaluated. Maximum intensity projection datasets were analyzed. Results: Two- to 5-fold acquisition time increase was measured for 10-phase versus 2-phase phantom experiments. Regular breathing patterns yielded higher duty cycles than irregular (48.5% and 35.9%, respectively, P <.001), whereas faster breathing rates yielded shorter 4DMRI acquisition times. Volumes of a hypodense target were underestimated 4% to 5% for 2 and 4 phases compared with 10 phases. Better agreement was obtained for 6- and 8-phase acquisitions (~. 3% different from 10 phase). Internal target volume centroids on minimum and maximum images across all phases were <. 2 mm different across all 10 phases, although slight target excursion variations (up to 4 mm) were observed. In humans, a strong negative association between breathing rate and acquisition time (Pearson's r = - 0.68, P <.05) was observed. Eight-phase acquisition times ranged from 7 to 15 minutes, depending on the patient. Conclusion: 4DMRI has been optimized and implemented. Irregular breathing patterns and slow breathing rate adversely impacted 4DMRI efficiency; thus, interventions such as biofeedback may be desirable.