4D-MRI Images Research Articles

Ultra-fast multi-parametric 4D-MRI image reconstruction for real-time applications using a downsampling-invariant deformable registration (D2R) model

Background and purposeMotion estimation from severely downsampled 4D-MRI is essential for real-time imaging and tumor tracking. This simulation study developed a novel deep learning model for simultaneous MR image reconstruction and motion estimation, named the Downsampling-Invariant Deformable Registration (D2R) model. Materials and methodsForty-three patients undergoing radiotherapy for liver tumors were recruited for model training and internal validation. Five prospective patients from another center were recruited for external validation. Patients received 4D-MRI scans and 3D MRI scans. The 4D-MRI was retrospectively down-sampled to simulate real-time acquisition. Motion estimation was performed using the proposed D2R model. The accuracy and robustness of the proposed D2R model and baseline methods, including Demons, Elastix, the parametric total variation (pTV) algorithm, and VoxelMorph, were compared. High-quality (HQ) 4D-MR images were also constructed using the D2R model for real-time imaging feasibility verification. The image quality and motion accuracy of the constructed HQ 4D-MRI were evaluated. ResultsThe D2R model showed significantly superior and robust registration performance than all the baseline methods at downsampling factors up to 500. HQ T1-weighted and T2-weighted 4D-MR images were also successfully constructed with significantly improved image quality, sub-voxel level motion error, and real-time efficiency. External validation demonstrated the robustness and generalizability of the technique. ConclusionIn this study, we developed a novel D2R model for deformation estimation of downsampled 4D-MR images. HQ 4D-MR images were successfully constructed using the D2R model. This model may expand the clinical implementation of 4D-MRI for real-time motion management during liver cancer treatment.

Conditional-based Transformer network with learnable queries for 4D deformation forecasting and tracking.

Real-time motion management for image-guided radiation therapy interventions plays an important role for accurate dose delivery. Forecasting future 4D deformations from in-plane image acquisitions is fundamental for accurate dose delivery and tumor targeting. However, anticipating visual representations is challenging and is not exempt from hurdles such as the prediction from limited dynamics, and the high-dimensionality inherent to complex deformations. Also, existing 3D tracking approaches typically need both template and search volumes as inputs, which are not available during real-time treatments. In this work, we propose an attention-based temporal prediction network where features extracted from input images are treated as tokens for the predictive task. Moreover, we employ a set of learnable queries, conditioned on prior knowledge, to predict future latent representation of deformations. Specifically, the conditioning scheme is based on estimated time-wise prior distributions computed from future images available during the training stage. Finally, we propose a new framework to address the problem of temporal 3D local tracking using cine 2D images as inputs, by employing latent vectors as gating variables to refine the motion fields over the tracked region. The tracker module is anchored on a 4D motion model, which provides both the latent vectors and the volumetric motion estimates to be refined. Our approach avoids auto-regression and leverages spatial transformations to generate the forecasted images. The tracking module reduces the error by 63% compared to a conditional-based transformer 4D motion model, yielding a mean error of 1.5±1.1 mm. Furthermore, for the studied cohort of abdominal 4D MRI images, the proposed method is able to predict future deformations with a mean geometrical error of 1.2±0.7 mm.

A computational study of aortic reconstruction in single ventricle patients.

Patients with hypoplastic left heart syndrome (HLHS) are born with an underdeveloped left heart. They typically receive a sequence of surgeries that result in a single ventricle physiology called the Fontan circulation. While these patients usually survive into early adulthood, they are at risk for medical complications, partially due to their lower than normal cardiac output, which leads to insufficient cerebral and gut perfusion. While clinical imaging data can provide detailed insight into cardiovascular function within the imaged region, it is difficult to use these data for assessing deficiencies in the rest of the body and for deriving blood pressure dynamics. Data from patients used in this paper include three-dimensional, magnetic resonance angiograms (MRA), time-resolved phase contrast cardiac magnetic resonance images (4D-MRI) and sphygmomanometer blood pressure measurements. The 4D-MRI images provide detailed insight into velocity and flow in vessels within the imaged region, but they cannot predict flow in the rest of the body, nor do they provide values of blood pressure. To remedy these limitations, this study combines the MRA, 4D-MRI, and pressure data with 1D fluid dynamics models to predict hemodynamics in the major systemic arteries, including the cerebral and gut vasculature. A specific focus is placed on studying the impact of aortic reconstruction occurring during the first surgery that results in abnormal vessel morphology. To study these effects, we compare simulations for an HLHS patient with simulations for a matched control patient that has double outlet right ventricle (DORV) physiology with a native aorta. Our results show that the HLHS patient has hypertensive pressures in the brain as well as reduced flow to the gut. Wave intensity analysis suggests that the HLHS patient has irregular circulatory function during light upright exercise conditions and that predicted wall shear stresses are lower than normal, suggesting the HLHS patient may have hypertension.

A Model for Gastrointestinal Tract Motility in a 4D Imaging Phantom of Human Anatomy

<h3>Purpose/Objective(s)</h3> Gastrointestinal (GI) tract motility is one of the main sources for intra/inter-fraction variability and uncertainty in validating deformable image registration (DIR) algorithms. We present an implementation of four modes of GI motility in the 4D extended cardiac-torso (XCAT) digital phantom. <h3>Materials/Methods</h3> We consider motility modes that exhibit large amplitude changes in the diameter of the GI tract and may persist over timescales comparable to online adaptive planning and radiotherapy delivery. The following modes are modeled: peristalsis, rhythmic segmentation, high amplitude propagating contractions (HAPC), and tonic contractions. The models are intended for the organ of origin (i.e., peristalsis in esophagus and small intestine; HAPCs in colon) but can be applied to any user-specified GI tract organ. Peristalsis and rhythmic segmentations are modeled by traveling and standing sinusoidal waves. HAPCs and tonic contractions are modeled by traveling and stationary Gaussian waves. Spatial and temporal dispersion are implemented by exponential functions. The model is applied at the control points of the nonuniform rational B-spline (NURBS) surfaces defined in the reference XCAT phantom. The reference NURBS surfaces are dynamically updated at each timepoint based on the user-defined parameters shown in Table 1. GI motility is combined with the respiratory and cardiac motion available in XCAT. Cine MRI acquisitions from 50 fractions of 10 patients treated in a 1.5 T MR-Linac were analyzed to estimate default parameters for each model. <h3>Results</h3> We demonstrate the ability to generate realistic 4D MRI images that simulate four modes of GI motility combined with respiratory and cardiac motion. All modes of motility, except tonic contractions, were observed in the analysis of our cine MRI acquisitions. Peristalsis was the most common. Default parameters estimated from cine MRI data or from literature (if not observed in our measurements) are used as initial values for simulation experiments. <h3>Conclusion</h3> The digital phantom provides realistic models to aid in multimodality imaging research. The addition of GI motility will further contribute to the development, testing, and validation of DIR algorithms for MR-guided radiotherapy.

Improve correlation matrix of Discrete Fourier Transformation technique for finding the missing values of MRI images.

Missing values in the k-NN algorithm are a significant research concern, especially in low-grade tumours and CSF fluid, which are commonly identified in MRI scans. Missing values are usually ignored, but when data is mined, they can lead to bias and errors. In addition, the data is not missing at random. This study improves image accuracy, boosts the efficiency of missing k-NN hybrid values, and develops a research technique for detecting CSF fluid deposits in brain areas separated from non-tumor areas. We also offer a new method for detecting low-grade tumours or cerebrospinal fluid (CSF) formation in its early stages. In this study, we combine the hybrid K-Nearest Neighbor algorithm with the Discrete Fourier transform (DFT), as well as Time-Lagged analysis of four-dimensional (4D) MRI images. These dependencies exist in both space and time, but present techniques do not account for both sequential linkages and numerous types of missingness. To address this, we propose the DFLk-NN imputation method, which combines two imputation approaches based on a hybrid k-NN extension and the DFT to capture time-lag correlations both within and across variables. There are several types of missingness are enables the imputation of missing values across the variable even when all the data for a given time point is missing. The proposed method gives high accuracies of MRI datasets and retrieves the missing data in the images.

Multi-contrast four-dimensional magnetic resonance imaging (MC-4D-MRI): Development and initial evaluation in liver tumor patients.

To develop a novel multi-contrast four-dimensional magnetic resonance imaging (MC-4D-MRI) technique that expands single image contrast 4D-MRI to a spectrum of native and synthetic image contrasts and to evaluate its feasibility in liver tumor patients. The MC-4D-MRI technique integrates multi-parametric MRI fusion, 4D-MRI, and deformable image registration (DIR) techniques. The fusion technique consists of native MRI as input, image pre-processing, fusion algorithm, adaptation, and fused multi-contrast MRI as output. Four-dimensional deformation vector fields (4D-DVF) were generated from an original T2/T1-w 4D-MRI by deforming end-of-inhalation (EOI) to nine other phase volumes via DIR. The 4D-DVF were applied to multi-contrast MRI to generate a spectrum of 4D-MRI in different image contrasts. The MC-4D-MRI technique was evaluated in five liver tumor patients on tumor contrast-to-noise ratio (CNR), internal target volume (ITV) contouring consistency, diaphragm motion range, and tumor motion trajectory; and in digital anthropomorphic phantoms on 4D-DIR introduced errors in tumor motion range, centroid location, extent, and volume. MC-4D-MRI consisting of 4D-MRIs in native image contrasts (T1-w, T2-w, and T2/T1-w) and synthetic image contrasts, such as tumor-enhanced contrast (TEC) were generated in five liver tumor patients. Patient tumor CNR increased from 2.6±1.8 in the T2/T1-w MRI, to -4.4±2.4, 6.6±3.0, and 9.6±3.9 in the T1-w, T2-w, and TEC MRI, respectively. Patient ITV inter-observer mean Dice similarity coefficient (mDSC) increased from 0.65±0.10 in the original T2/T1-w 4D-MRI, to 0.76±0.14, 0.77±0.12, and 0.86±0.05 in the T1-w, T2-w, and TEC 4D-MRI, respectively. Patient diaphragm motion range absolute differences between the three new 4D-MRIs and original T2/T1-w 4D-MRI were 1.2±1.3, 0.3±0.7, and 0.5±0.5mm, respectively. Patient tumor displacement phase-averaged absolute differences between the three 4D-MRIs and the original 4D-MRI were 0.72±0.33, 0.62±0.54, and 0.74±0.43mm in the superior-inferior (SI) direction, and 0.59±0.36, 0.51±0.30, and 0.50±0.24mm in the anterior-posterior (AP) direction, respectively. In the digital phantoms, phase-averaged absolute tumor centroid shift caused by the 4D-DIR were at or below 0.5mm in SI, AP, and left-right (LR) directions. We developed an MC-4D-MRI technique capable of expanding single image contrast 4D-MRI along a new dimension of image contrast. Initial evaluations in liver tumor patients showed enhancements in image contrast variety, tumor contrast, and ITV contouring consistencies using MC-4D-MRI. The technique might offer new perspectives on the image contrast of MRI and 4D-MRI in MR-guided radiotherapy.

Feasibility of MR-guided radiotherapy using beam-eye-view 2D-cine with tumor-volume projection

Purpose. Current magnetic resonance imaging (MRI) guided radiotherapy (MRgRT) applies sagittal/coronal 2D-cine to monitor major tumor motions, however, the beam eye’s view (BEV) with volumetric tumor projection would be the best measure for radiation beam conformality, independent of tumor through-plane motion. The goal is to assess the feasibility, accuracy, and performance of the BEV approach. Methods. Beam-specific BEV 2D-cine with volume-projected tumor contours were simulated to establish a 2D/3D tumor match against a tumor-motion library based on multi-breath time-resolved (TR) 4DMRI images. Two BEV-library-matching methods were developed: (1) fast screening with tumor center-of-mass (∆COM), in-plane area ratio, and DICE similarity, and finalizing with the highest DICE score and (2) DICE screening for top-3 candidates and finalizing with rigid registration. A 4D-XCAT digital phantom and 8 lung-cancer patients were used for assessment. For each patient, 3 sets of 40 s TR-4DMRI were acquired at 2 Hz and 6 representative BEV were created with the isocenter set at tumor COM in mid-respiration. One TR-4DMRI set (40 × 2 = 80-images) was used to simulate BEV 2D-cine and the other two (160-images) were used to create a library. The matching result was validated against the ground truth within the test set. Using a leave-one-out strategy, the success rate, accuracy, and speed of tumor matching were assessed for volume-projected tumors over 11520 time-points (=8patients•3sets•80images•6BEVs). Results. Volume-projected tumor contour area on the 6 BEVs varies by 60% ± 8% and (in-plane/volume-projected) varies by 82% ± 9%. The changes with tumor shape, orientation, and through-plane motion. Method-1 produces 96% matching success (ΔCOM = 0.7 ± 0.2 mm, =1.01 ± 0.02, Dice=0.92 ± 0.02) with the computational time of 15 ± 1 ms/match, while method-2 produces 94% ± 1% success (ΔCOM = 0.2 ± 0.1 mm, =1.00 ± 0.01, Dice = 0.94 ± 0.02) with 223 ± 13 ms/match. Conclusion. This study has demonstrated the feasibility, accuracy, and benefits of BEV 2D-cine imaging with tumor-volume projection, allowing real-time tumor motion monitoring and beam conformality checking. Further clinical evaluation is necessary before MRgRT applications.

PO-1644: Inter-session variability of 4DMRI image quality after outlier rejection

PO-1644: Inter-session variability of 4DMRI image quality after outlier rejection

Abstract 427: A Patient-Specific 3D Bioprinted Platform for in vitro Disease Modeling and Treatment Planning in Pulmonary Vein Stenosis

Pulmonary vein stenosis (PVS) is an acute pediatric cardiovascular disease that is always lethal if not treated early. While current clinical interventions (stenting and angioplasties) have shown promising results in treating PVS, they require multiple re-interventions that can lead to re-stenosis and diminished long-term efficacy. Thus, there is an unmet need to develop functional in vitro models of PVS that can serve as a platform to study clinical interventions. Patient-inspired 3D bioprinted tissue models provide a unique model to recapitulate and analyze the complex tissue microenvironment impacted by PVS. Here, we developed perfusable in vitro models of healthy and stenotic pulmonary vein by 3D reconstruction and bioprinting inspired by patient CT data ( Figure 1 ). Models were seeded with human endothelial (ECs) and smooth muscle cells (SMCs) to form a bilayer structure and perfused using a bioreactor to study cell response to stenotic geometry, and to the stent-based treatment. Flow hemodynamics through printed veins were quantified via Computational Fluid Dynamics (CFD) modeling, 4D MRI and 3D Ultrasound Particle Imaging Velocimetry (echo PIV). Cell growth and endothelialization were analyzed. Our work demonstrates the feasibility of bioprinting various cardiovascular cells, to create perfusable, patient-specific vascular constructs that mimic complex in vivo geometries. Deeper understanding of EC-SMC crosstalk mechanisms in in vitro biomimetic models that incorporate tissue-like geometrical, chemical, and biomechanical ques could offer substantial insights for prevention and treatment of PVS, as well as other cardiovascular disease.

Motion robust 4D-MRI sorting based on anatomic feature matching: A digital phantom simulation study

PurposeMotion artifacts induced by breathing variations are common in 4D-MRI images. This study aims to reduce the motion artifacts by developing a novel, robust 4D-MRI sorting method based on anatomic feature matching and applicable in both cine and sequential acquisition. MethodThe proposed method uses the diaphragm as the anatomic feature to guide the sorting of 4D-MRI images. Initially, both abdominal 2D sagittal cine MRI images and axial MRI images were acquired. The sagittal cine MRI images were divided into 10 phases as ground truth. Next, the phase of each axial MRI image is determined by matching its diaphragm position in the intersection plane to the ground truth cine MRI. Then, those matched phases axial images were sorted into 10-phase bins which were identical to the ground truth cine images. Finally, 10-phase 4D-MRI were reconstructed from these sorted axial images. The accuracy of reconstructed 4D-MRI data was evaluated by comparing with the ground truth using the 4D eXtended Cardiac Torso (XCAT) digital phantom. The effects of breathing signal, including both regular (cosine function) and irregular (patient data) in both axial cine and sequential scanning modes, on reconstruction accuracy were investigated by calculating total relative error (TRE) of the 4D volumes, Volume-Percent-Difference (VPD) and Center-of-Mass-Shift (COMS) of the estimated tumor volume, compared with the ground truth XCAT images. ResultsIn both scanning modes, reconstructed 4D-MRI images matched well with ground truth with minimal motion artifacts. The averaged TRE of the 4D volume, VPD and COMS of the EOE phase in both scanning modes are 0.32%/1.20%/±0.05 ​mm for regular breathing, and 1.13%/4.26%/±0.21 ​mm for patient irregular breathing. ConclusionThe preliminary evaluation results illustrated the feasibility of the robust 4D-MRI sorting method based on anatomic feature matching. This method provides improved image quality with reduced motion artifacts for both cine and sequential scanning modes.

Probability-based 3D k-space sorting for motion robust 4D-MRI.

Current 4D-MRI techniques are prone to breathing-variation-induced motion artifacts. This study developed a novel method for motion-robust multi-cycle 4D-MRI using probability-based multi-cycle sorting to overcome this deficiency. The main cycles were first extracted from the breathing signal. 3D k-space data were then sorted using a result-driven method for each main cycle. The new method was tested on a 4D-extended cardiac-torso (XCAT) phantom with a patient and an artificially generated breathing curve. For comparison, the k-space data were sorted using conventional phase sorting to generate single-cycle 4D-MRI images. Signal-to-noise ratio (SNR) of tumor and liver, tumor volume consistency, and average intensity projection (AIP) accuracy were compared between the two methods. The original phantom images were used as references for the evaluation. The new method showed improved tumor-to-liver SNR and tumor volume consistency as compared to 3D k-space phase sorting in both the simulated artificial and real patient breathing signals. For the artificial breathing cycles, the average tumor-to-liver SNR and standard deviation (SD) of tumor volume were 2.53 and 3.80% for cycle 1, 2.24 and 6.16% for cycle 2 of probability-based sorting as compared to 1.47 and 21.83% obtained using the phase sorting method; for the patient breathing curve, values of 1.99 and 2.71%, 1.97 and 3.29%, 1.88 and 4.16% were observed for cycle 1, cycle 2 and cycle 3 of probability-based sorting, versus 1.44 and 7.20% for phase sorting method. Furthermore, the AIP accuracy was improved in the probability-based sorting approach when compared to phase sorting, with the average intensity difference per voxel reduced from 0.39 to 0.15 for the artificial curve, and from 0.46 to 0.21 for the patient curve. We demonstrated the feasibility of probability-based 3D k-space sorting for motion-robust multi-cycle 4D-MRI reconstruction with breathing variation induced motion artifact reduction compared with conventional 2D image sorting and 3D phase sorting methods. This new technique can potentially improve the accuracy of radiation treatment guidance for mobile targets.

A dual-purpose MRI acquisition to combine 4D-MRI and dynamic contrast-enhanced imaging for abdominal radiotherapy planning

For successful abdominal radiotherapy it is crucial to have a clear tumor definition and an accurate characterization of the motion. While dynamic contrast-enhanced (DCE) MRI aids tumor visualization, it is often hampered by motion artifacts. 4D-MRI characterizes this motion, but often lacks the contrast to clearly visualize the tumor. This dual requirement is challenging due to time constraints. Here, we propose combining both into a single acquisition by reconstructing the data in various ways in order to achieve both high spatio-temporal resolution DCE-MRI and accurate 4D-MRI motion estimates.A 5 min T1-weigthed DCE acquisition was collected in five renal-cell carcinoma patients and simulated in a digital phantom. Data were acquired continuously using a 3D golden angle radial stack-of-stars acquisition. This enabled three types of reconstruction; (1) a high spatio-temporal resolution DCE time series, (2) a 5D reconstruction and (3) a contrast-enhanced 4D-MRI for motion characterization. Motion extracted from the 4D- and 5D-MRI was compared with a separately acquired 4D-MRI and additional 2D cine MR imaging.Simulations on XCAT showed that 5D reconstructions severely underestimated motion (), whereas contrast-enhanced 4D-MRI only underestimated motion by . This was confirmed in the in vivo data where motion of the contrast-enhanced 4D-MRI was approximately smaller than the motion in the 2D cine MRI (5.8 mm versus 6.5 mm), but equal to a separately acquired 4D-MRI (5.8 mm versus 5.9 mm). 5D reconstructions underestimated the motion by more than , but minimized respiratory-induced blurring in the contrast enhanced images. DCE time-series demonstrated clear tumor visualization and contrast enhancement throughout the entire field-of-view.DCE- and 4D-MRI can be integrated into a single acquisition that enables different reconstructions with complementary information for abdominal radiotherapy planning and, in an MRI-guided treatment, precise motion information, input for motion models, and rapid feedback on the contrast enhancement.

A post-processing method based on interphase motion correction and averaging to improve image quality of 4D magnetic resonance imaging: a clinical feasibility study.

Nine patients (seven pancreas, one liver, and one lung) were recruited. 4D-MRI was performed using two prototype k-space sorted techniques, stack-of-stars (SOS) and koosh-ball (KB) acquisitions. Post-processing using MoCoAve was implemented for both methods. Image quality score, apparent SNR (aSNR), sharpness, motion trajectory and standard deviation (σ_GTV) of the gross tumor volumes were compared between original and MoCoAve image sets. All subjects successfully underwent 4D-MRI scans and MoCoAve was performed on all data sets. Significantly higher image quality scores (2.64 ± 0.39 vs 1.18 ± 0.34, p = 0.001) and aSNR (37.6 ± 15.3 vs 18.1 ± 5.7, p = 0.001) was observed in the MoCoAve images when compared to the original images. High correlation in tumor motion trajectories in the superoinferior direction (SI: 0.91 ± 0.08) and weaker in the anteroposterior (AP: 0.51 ± 0.44) and mediolateral (ML: 0.37 ± 0.23) directions, similar image sharpness (0.367 ± 0.068 vs 0.369 ± 0.072, p = 0.805), and minimal average absolute difference (0.47 ± 0.34 mm) of the motion trajectory profiles was found between the two image sets. The σ_GTV in pancreas patients was significantly (p = 0.039) lower in MoCoAve images (1.48 ± 1.35 cm3) than in the original images (2.17 ± 1.31 cm3). MoCoAve using interphase motion correction and averaging has shown promise as a post-processing method for improving k-space sorted (SOS and KB) 4D-MRI image quality in thoracic and abdominal cancer patients. The proposed method is an image based post-processing method that could be applied to many k-space sorted 4D-MRI methods for improved image quality and signal-to-noise ratio while preserving image sharpness and respiratory motion fidelity. It is a useful technique for the radiotherapy planning community who are interested in using 4D-MRI but aren't satisfied with their current MR image quality.

47 Automated respiratory cycle binning for liver 4D-MR imaging

Introduction Stereotactic body radiation therapy (SBRT) has been validated for liver lesions [1] . Because high dose per fraction is delivered on small volume, very accurate lesion segmentation is necessary. MRI is the most relevant imaging for liver lesion contouring but it is prone to motion artifacts if breathing is not managed. Evaluation of liver lesion motion induced by breathing is necessary in radiotherapy to guide patient treatment planning (respiratory gating). A 4D-MRI sequence was previously validated for liver [2] . In this study, an automated method was developed to sort images according to the breathing cycle. Methods The 4D-MRI acquisition is performed with an experimental sequence (bSSFP TrueFISP [3] ) on 1.5T Magnetom AeraTM (Siemens). This sequence allows very fast axial and sagittal 2D acquisition during free breathing. Axial slice positions (0.88 s/slice) are incremented to cover the liver volume. Sagittal slices (navigators) are acquired at a fixed position. The navigators are used to assign a phase for each axial slice which depend on an amplitude sequencing of the respiratory cycle (0% = inspiration, 16%, 33%, 50% = expiration, 66%, 83%). Lung volume is estimated on each navigator by automatic segmentation and allows to assign a phase for the coupled axial image in the breathing cycle. As the image number per respiratory cycle is low, the sequence is repeated 20 times in order to achieve satisfactory sampling. The sequence was evaluated on 4 volunteers with an audio coaching (5 and 6 s period) and 2 patients without audio coaching. Results Patient image sorting according to their position in the breathing cycle was realized by automated phase detection (Fig. 1) and images were imported into the treatment planning system Eclipse 13.7TM (Varian) in 6 phases. Audio coaching is not used here. It is required for patients with irregular breathing cycle. Download : Download high-res image (278KB) Download : Download full-size image From axial images, Eclipse reconstructs slices in sagittal and coronal plans, and allows to display a cine view with the 6 breathing phases in order to analyze the dynamic behavior of the liver. Conclusions 4D-MRI images from volunteers and patients were acquired with bSSFP TrueFISP sequence. After automated sorting with the method developed in this study, 6 breathing cycle phases were imported into Eclipse. Patient 4D-MRI can be registered with 4D-CT to increase the physician’s segmentation accuracy.

Novel 4D-MRI of tumor infiltrating vasculature: characterizing tumor and vessel volume motion for selective boost volume definition in pancreatic radiotherapy

BackgroundPancreatic ductal adenocarcinoma has dismal prognosis. Most patients receive radiation therapy (RT), which is complicated by respiration induced organ motion in upper abdomen. The purpose of this study is to report our early clinical experience in a novel self-gated k-space sorted four-dimensional magnetic resonance imaging (4D-MRI) with slab-selective (SS) excitation to highlight tumor infiltrating blood vessels for pancreatic RT.MethodsTen consecutive patients with borderline resectable or locally advanced pancreatic cancer were recruited to the study. Non-contrast 4D-MRI with and without slab-selective excitation and 4D-CT with delay contrast were performed on all patients. Vessel-tissue CNR were calculated for aorta and critical vessels (superior mesenteric artery or superior mesenteric vein) encompassed by tumor. Respiratory motion trajectories for tumor, as well as involved vessels were analyzed on SS-4D-MRI. Intra-class cross correlation (ICC) between tumor volume and involved vessels were calculated.ResultsAmong all 4D imaging modalities evaluated, SS-4D-MRI sampling trajectory results in images with highest vessel-tissue CNR comparing to non-slab-selective 4D-MRI and 4D-CT for all patients studied. Average (±standard deviation) CNR for involved vessels are 13.1 ± 8.4 and 3.2 ± 2.7 for SS-4D-MRI and 4D-CT, respectively. The ICC factors comparing tumor and involved vessels motion trajectories are 0.93 ± 0.10, 0.65 ± 0.31 and 0.77 ± 0.23 for superior-inferior, anterior-posterior and medial-lateral directions respectively.ConclusionsA novel 4D-MRI sequence based on 3D-radial sampling and slab-selective excitation has been assessed for pancreatic cancer patients. The non-contrast 4D-MRI images showed significantly better contrast to noise ratio for the vessels that limit tumor resectability compared to 4D-CT with delayed contrast. The sequence has great potential in accurately defining both the tumor and boost volume margins for pancreas RT with simultaneous integrated boost.

Evaluation of automatic contour propagation in T2-weighted 4DMRI for normal-tissue motion assessment using internal organ-at-risk volume (IRV).

PurposeThe purpose of this study was to evaluate the quality of automatically propagated contours of organs at risk (OARs) based on respiratory‐correlated navigator‐triggered four‐dimensional magnetic resonance imaging (RC‐4DMRI) for calculation of internal organ‐at‐risk volume (IRV) to account for intra‐fractional OAR motion.Methods and MaterialsT2‐weighted RC‐4DMRI images were of 10 volunteers acquired and reconstructed using an internal navigator‐echo surrogate and concurrent external bellows under an IRB‐approved protocol. Four major OARs (lungs, heart, liver, and stomach) were delineated in the 10‐phase 4DMRI. Two manual‐contour sets were delineated by two clinical personnel and two automatic‐contour sets were propagated using free‐form deformable image registration. The OAR volume variation within the 10‐phase cycle was assessed and the IRV was calculated as the union of all OAR contours. The OAR contour similarity between the navigator‐triggered and bellows‐rebinned 4DMRI was compared. A total of 2400 contours were compared to the most probable ground truth with a 95% confidence level (S95) in similarity, sensitivity, and specificity using the simultaneous truth and performance level estimation (STAPLE) algorithm.ResultsVisual inspection of automatically propagated contours finds that approximately 5–10% require manual correction. The similarity, sensitivity, and specificity between manual and automatic contours are indistinguishable (P > 0.05). The Jaccard similarity indexes are 0.92 ± 0.02 (lungs), 0.89 ± 0.03 (heart), 0.92 ± 0.02 (liver), and 0.83 ± 0.04 (stomach). Volume variations within the breathing cycle are small for the heart (2.6 ± 1.5%), liver (1.2 ± 0.6%), and stomach (2.6 ± 0.8%), whereas the IRV is much larger than the OAR volume by: 20.3 ± 8.6% (heart), 24.0 ± 8.6% (liver), and 47.6 ± 20.2% (stomach). The Jaccard index is higher in navigator‐triggered than bellows‐rebinned 4DMRI by 4% (P < 0.05), due to the higher image quality of navigator‐based 4DMRI.ConclusionAutomatic and manual OAR contours from Navigator‐triggered 4DMRI are not statistically distinguishable. The navigator‐triggered 4DMRI image provides higher contour quality than bellows‐rebinned 4DMRI. The IRVs are 20–50% larger than OAR volumes and should be considered in dose estimation.

Respiratory motion-resolved, self-gated 4D-MRI using Rotating Cartesian K-space (ROCK): Initial clinical experience on an MRI-guided radiotherapy system

PurposeTo optimize and evaluate the respiratory motion-resolved, self-gated 4D-MRI using Rotating Cartesian K-space (ROCK-4D-MRI) method in a 0.35 T MRI-guided radiotherapy (MRgRT) system. Methods and materialsThe study included seven patients with abdominal tumors treated on the MRgRT system. ROCK-4D-MRI and 2D-CINE, was performed immediately after one of the treatment fractions. Motion quantification based on 4D-MRI was compared with those based on 2D-CINE. The image quality of 4D-MRI was evaluated against 4D-CT. The gross tumor volumes (GTV) were defined based on individual respiratory phases of both 4D-MRI and 4D-CT and compared for their variability over the respiratory cycle. ResultThe motion measurements based on 4D-MRI matched well with 2D-CINE, with differences of 1.04 ± 0.52 mm in the superior–inferior and 0.54 ± 0.21 mm in the anterior–posterior directions. The image quality scores of 4D-MRI were significantly higher than 4D-CT, with better tumor contrast (3.29 ± 0.76 vs. 1.86 ± 0.90) and less motion artifacts (3.57 ± 0.53 vs. 2.29 ± 0.95). The GTVs were more consistent in 4D-MRI than in 4D-CT, with significantly smaller GTV variability (9.31 ± 4.58% vs. 34.27 ± 23.33%). ConclusionOur study demonstrated the clinical feasibility of using the ROCK-4D-MRI to acquire high quality, respiratory motion-resolved 4D-MRI in a low-field MRgRT system. The 4D-MRI image could provide accurate dynamic information for radiotherapy treatment planning.

An efficient model to guide prospective T2-weighted 4D magnetic resonance imaging acquisition.

To establish a mathematical model to guide prospective T2-weighted four-dimensional magnetic resonance imaging (4DMRI) acquisition and to propose an efficient solution to expedite prospective T2-weighted 4DMRI acquisition. Prospective T2-weighted 4DMRI acquisition was characterized by a mathematical model with 4DMRI acquisition time as the objective function and completeness of the image set, acquisition timing, image contrast, and image artifacts as constraints. Given the irregular nature of human respiration, an efficient solution based on the greedy strategy (ESGS) was proposed. The efficiency of the ESGS method was validated using healthy human subjects. Comparisons were made with the previous 4DMRI method incorporating the prefixed-order respiratory state splitting (PO-RSS) technique. 4DMRI image sets acquired using the ESGS and PO-RSS methods had similar image quality. The average time to acquire a 4DMRI image set covering 60 slices at 10 respiratory states was reduced by 30%, from 13.1min using the PO-RSS method to 9.0min using the ESGS method. It was demonstrated that high-quality T2-weighted 4DMRI could be obtained within a reasonable amount of time and all slices within each of the three-dimensional volumes were indeed acquired at the same respiratory state. The ESGS method substantially reduces the acquisition time for T2-weighted 4DMRI, making it ready for clinical evaluation to obtain abdominal tumor motion for radiotherapy treatment planning.

Surface as a motion surrogate for gated re-scanned pencil beam proton therapy

This simulation study investigated the dosimetric effectiveness and treatment efficiency of surface motion guided gating of pencil beam scanning (PBS) proton therapy for liver tumour treatments. Dedicated 4D dose calculations were performed for simulating gated treatments using 4DCT data for six patients derived from 4DMRI (4DCT(MRI)). Surface motion as a surrogate for tumour motion was extracted from the 4DMRI images and a linear internal–external correlation model applied to derive amplitude-based gating windows (GWs) of 10 and 5 mm. 4D treatments were simulated using gating and layered/volumetric rescanning (either alone or combined) and four assumed system latencies (50/100/200/500 ms) for the response time of the beam gating to the surrogate. Resulting 4D plans were compared using D5–D95 and V95 in the CTV as the primary metrics, as well as dose to the healthy liver and total treatment time. With no motion mitigation, interplay effects deteriorate the dose homogeneity by more than 30% with respect to the static reference plan, whereas with surface motion guided gating alone, this could be reduced to 12/20% and 5/10% (mean/max over all cases) for 10 mm and 5 mm GWs, respectively. Furthermore, by combining ×5 layered rescans with 5 mm GW, plan homogeneities to within 1/5% of the static references could be achieved. Dose inhomogeneities were however still pronounced for latencies ⩾200 ms but limited when ⩽100 ms. ITV volumes could be decreased by 19/25% when 10/5 mm GW was employed, leading to reductions in mean dose to the healthy liver tissue of 2.6/3.3%. Our results confirm the potential of combining gating and re-scanning (re-gating) for mitigating large tumor motions, and the potential of surface motion monitoring as a gating signal.

Investigation of undersampling and reconstruction algorithm dependence on respiratory correlated 4D-MRI for online MR-guided radiation therapy

The purpose of this work is to investigate the effects of undersampling and reconstruction algorithm on the total processing time and image quality of respiratory phase-resolved 4D MRI data. Specifically, the goal is to obtain quality 4D-MRI data with a combined acquisition and reconstruction time of five minutes or less, which we reasoned would be satisfactory for pre-treatment 4D-MRI in online MRI-gRT. A 3D stack-of-stars, self-navigated, 4D-MRI acquisition was used to scan three healthy volunteers at three image resolutions and two scan durations. The NUFFT, CG-SENSE, SPIRiT, and XD-GRASP reconstruction algorithms were used to reconstruct each dataset on a high performance reconstruction computer. The overall image quality, reconstruction time, artifact prevalence, and motion estimates were compared. The CG-SENSE and XD-GRASP reconstructions provided superior image quality over the other algorithms. The combination of a 3D SoS sequence and parallelized reconstruction algorithms using computing hardware more advanced than those typically seen on product MRI scanners, can result in acquisition and reconstruction of high quality respiratory correlated 4D-MRI images in less than five minutes.