Impact of incorporating visual biofeedback in 4D MRI

Abstract

Precise radiation therapy (RT) for abdominal lesions is complicated by respiratory motion and suboptimal soft tissue contrast in 4D CT. 4D MRI offers improved contrast although long scan times and irregular breathing patterns can be limiting. To address this, visual biofeedback (VBF) was introduced into 4D MRI. Ten volunteers were consented to an IRB‐approved protocol. Prospective respiratory‐triggered, T2‐weighted, coronal 4D MRIs were acquired on an open 1.0T MR‐SIM. VBF was integrated using an MR‐compatible interactive breath‐hold control system. Subjects visually monitored their breathing patterns to stay within predetermined tolerances. 4D MRIs were acquired with and without VBF for 2‐ and 8‐phase acquisitions. Normalized respiratory waveforms were evaluated for scan time, duty cycle (programmed/acquisition time), breathing period, and breathing regularity (end‐inhale coefficient of variation, EI‐COV). Three reviewers performed image quality assessment to compare artifacts with and without VBF. Respiration‐induced liver motion was calculated via centroid difference analysis of end‐exhale (EE) and EI liver contours. Incorporating VBF reduced 2‐phase acquisition time (4.7±1.0 and 5.4±1.5trueprefixmin with and without VBF, respectively) while reducing EI‐COV by 43.8%±16.6%. For 8‐phase acquisitions, VBF reduced acquisition time by 1.9±1.6trueprefixmin and EI‐COVs by 38.8%±25.7% despite breathing rate remaining similar (11.1±3.8 breaths/min with vs. 10.5±2.9 without). Using VBF yielded higher duty cycles than unguided free breathing (34.4%±5.8% vs. 28.1%±6.6%, respectively). Image grading showed that out of 40 paired evaluations, 20 cases had equivalent and 17 had improved image quality scores with VBF, particularly for mid‐exhale and EI. Increased liver excursion was observed with VBF, where superior–inferior, anterior–posterior, and left–right EE‐EI displacements were 14.1±5.8, 4.9±2.1, and 1.5±1.0 mm, respectively, with VBF compared to 11.9±4.5, 3.7±2.1, and 1.2±1.4 mm without. Incorporating VBF into 4D MRI substantially reduced acquisition time, breathing irregularity, and image artifacts. However, differences in excursion were observed, thus implementation will be required throughout the RT workflow.PACS number(s): 87.55.‐x, 87.61.‐c, 87.19.xj

Image‐based retrospective 4D MRI in external beam radiotherapy: A comparative study with a digital phantom

“…To our knowledge, the few studies reporting a quantitative comparison between retrospective sorting methods do not include an analysis of different image‐based approaches. In one of these studies, visual biofeedback based on an MR‐compatible active control device was investigated and compared with a free‐breathing acquisition, with the aim to regularize respiration and reduce artifacts. Li et al directly compared an internal surrogate (1D navigator) with a concurrently acquired external surrogate (bellows respiratory signal).…”

MRI-guidance for motion management in external beam radiotherapy: current status and future challenges

“…In retrospective methods inherited from 4DCT, sorting of slices is usually based on an external surrogate (Hu et al 2013). Different strategies were investigated to improve the performance of the external surrogate, either making use of audio-visual biofeedback (To et al 2016b) or advanced sorting (Liu et al 2015, Liang et al 2016, Tryggestad et al 2013d, Du et al 2015. As previously mentioned however, the use of internal breathing surrogates directly extracted from the acquired 2D images has been shown to increase robustness in organ motion description with respect to external surrogates (Stemkens et al 2015, Liu et al 2016a, Li et al 2017.…”

“…To our knowledge, a comprehensive comparison of these approaches for the evaluation of the best solution is not available, thus limiting their application in a clinical setting. Visual biofeedback was compared against a free-breathing acquisition (To et al 2016b), whereas a direct comparison of an internal surrogate (1D navigator) with a concurrently acquired external surrogate was reported (Li et al 2016). Multi-slice 2D acquisition based on navigator approaches can substantially reduce image artefacts compared with some of the image-derived approaches (Paganelli et al 2018) and could describe intra-cycle variations more effectively (Von Siebenthal et al 2007).…”

A retrospective 4D‐MRI based on 2D diaphragm profiles for lung cancer patients

“…To et al 31 used an air-filled cushion; Yang et al 32 used an MR pulse sequence of 3D radial k-space sampling; and Han et al 33 used an MR pulse sequence of 3D spiral k-space sampling.…”

“…There have been various triggering methods; Hu et al . used a bellows belt; To et al . used an air‐filled cushion; Yang et al .…”