MRI-guided Radiotherapy Research Articles

Edema Progression during MRI-Guided Glioblastoma Radiotherapy

There are limited studies evaluating brain changes on MRI during the standard six weeks of concurrent chemotherapy and radiotherapy (RT) for glioma. Existing studies suggest that MRI during chemoRT could be useful to determine response to treatment. At our institution, we have been treating select gliomas with one of the first commercially available 0.35T MRI-RT systems. The system includes one clinically useable pulse sequence called balanced steady state free precession (bSSFP), which produces a highly T2-weighted 3D image of the brain. The bSSFP MRIs are obtained daily for treatment setup and have demonstrated changes occurring in and around gliomas during patients’ chemoradiotherapy course. Our objective is to quantify the frequency, magnitude, and time course of these changes. Under a prospective registry protocol, we reviewed 14 patients at our institution who were treated for high grade glioma with chemoradiation, 60 Gy in 30 fractions plus temozolomide, on our institution’s MRI-RT system. We contoured the T2-weighted hyper-intense volumes from the MRI-RT set-up scans for all 30 fractions of treatment for each patient and then calculated the contoured MRI volumes. MRI-RT set-up scans during the last week of treatment were compared to the patients’ diagnostic MRIs obtained 3-4 weeks post-treatment completion to determine whether there were similar findings. 5 of 14 (36%) patients demonstrated greater than a 10% increase in T2-weighted MRI volume during the course of their treatment. Alternatively, 3 of 14 (21%) patients demonstrated greater than a 10% decrease in T2-weighted MRI volume during treatment. In 3 of the patients with volume growth, changes predominantly were observed during weeks 5 and 6 of treatment. The bSSFP images obtained during the last week of treatment demonstrated similar abnormalities compared to the diagnostic FLAIR MRIs obtained 3-4 weeks post-treatment. Patients who had volume expansion during treatment that resolved on serial follow-up MRIs were counted as pseudoprogressors. The time to true progression in the cohort was 19 months. Median follow-up was 14.7 months. 36% of high-grade glioma patients had growth of T2-weighted MRI volume during chemoradiotherapy, and these changes were similar to those observed on post-treatment diagnostic scans. Tumor volume expansion could be due to pseudoprogression, true progression, or changes in edema. Our future work aims to differentiate these possibilities with additional patients, follow-up, and MRI pulse sequences on the MRI-RT system. Groups using limited CTV margins for treatment planning should be aware that MRI volumes could significantly change during radiotherapy. Patients demonstrated T2-weighted MRI volume increase during weeks 5-6 of treatment, suggesting that studies using diagnostic MRI scans mid-treatment (week 3-4) may not fully capture the extent of progression.

Deep Learning Approach for Multi-Reference Tissue Normalization on T2-weighted MRI in Longitudinal Dataset from Prospective Radiotherapy Trial for Prostate Cancer

To remove the variations in T2-weighted (T2-w) intensities caused by different scanners and sequence parameters in longitudinal MRI series obtained before and after radiotherapy in a prospective trial for prostate cancer. Multi-reference tissue normalization utilizes the average T2-w signal in three anatomical structures. In our application, the reference structures were bladder, femur and Gluteus Maximus (GM). A spline model was fitted between the raw and reference values for the three structures for each patient. Using this model, the intensity values of each pixel on the image were normalized. A MASK R-CNN deep-learning network was trained to identify and automatically extract the intensity values in the reference structures. The normalization was applied to T2-w longitudinal data from patients, enrolled in a single institution Phase I trial of MRI-guided radiotherapy. All patients underwent an mpMRI before and at 3, 9, and 24 months post-RT. Prostate, peripheral zone (PZ) and transition zone (TZ) were contoured in MIM. GTV(s) were translated from the planning CT to all 4 mpMRI exams. The average T2-w signal in the GTVs, normal appearing (NA)-PZ and NA-TZ were evaluated before and after normalization. The three reference structures were contoured on T2-w MRI from 32 patients enrolled in unrelated study. MRIs from 28 patients were used for training (with 10% set aside for validation) and 4 patients for test. The network was evaluated using the Dice Similarity Coefficient (DSC) between the manual and automatic contours. DSC for bladder, femur, and GM for the training, validation and test datasets were: 0.93, 0.82, 0.88; 0.94, 0.85, 0.88; and 0.9, 0.77, 0.87; respectively. Twenty-five patients were enrolled on the trial. Two patients had only diagnostic mpMRI and two patients with hip implants were excluded from the study. A total of 71 mpMRI were analyzed: 13 patients had 3 longitudinal mpMRI studies and 8 patients – 4. Images were acquired on 3T Discovery MR750 magnet (n = 51) (GE, Waukesha, WI) and 3T MR Magnetom Trio (n = 6) and Skyra (n = 14) (Siemens, Erlagen, Germany). The T2-w signal of NA-PZ and NA-TZ in the normalized dataset exhibited gradual decrease post-RT with the reduction in 3 and 24 mo reaching statistical significance. There was no observable trend in NA-PZ and NA-TZ before normalization. Since GTV on T2 prior to RT is hypo-intense, we expected an increase in signal post-RT. However, in the longitudinal normalized dataset, GTVs signals remained flat, possibly due to the overall intensity decrease in the prostate. Multi-reference tissue normalization is a promising technique for harmonization of T2-w MRI series. This is particularly important in the settings of clinical trials where longitudinal data are collected over long periods during which MRI scanners undergo multiple upgrades. The deep learning approach makes this method feasible as removes the need for manual contouring of multiple structures.

The Effects of Androgen-Deprivation Therapy on MRI Delta Radiomics Features in a Prospective Radiotherapy Treatment Trial for Prostate Cancer

The aim of this study is to evaluate the effects of androgen-deprivation therapy (ADT) on the variation of multiparametric (mp)MRI-radiomics features, defined as "delta radiomics", in patients from a prospective radiotherapy trial in which 48% of men received short term androgen deprivation therapy (STADT). Twenty-five patients with favorable to high-risk prostate cancer were enrolled on a Phase I trial of MRI-guided radiotherapy treatment trial. Median follow up is 63 mo, range 17 to 90. mpMRI was carried out pre-RT, 3 and 9 months after completion of RT and at 2.0-2.5 years after completion of all treatment. A high dose (12-14 Gy) was given on day 1 to a portion near the center of the mpMRI delineated GTV(s). The mpMRI exam consisted of T2 weighted, T1 non-contrast, T1 dynamic contrast-enhanced (DCE)-MRI, and diffusion weighted imaging (DWI) with the generation of an Apparent Diffusion Coefficient (ADC) map. Prostate, peripheral zone (PZ) and transition zone (TZ) were contoured in MIM. GTV(s) were translated from the planning CT to all 4 mpMRI. Thirty variables (mean±SD) were extracted from GTV, normal appearing (NA)-PZ and NA-TZ on T2w and ADC, as well as the perfusion parameters Ktrans, Kep and ve on DCE-MRI. Baseline variables were compared between patients with and without ADT using Student t-test. Area-under-the-curve (s) (AUCs) for the variables over the baseline, 3 and 9 mo were compared between the +/-ADT groups using Mann-Whitney test. Two patients received only diagnostic mpMRI and were excluded from the study. From the reminding 23 patients, 14 had 3 longitudinal mpMRI studies and 9 had 4. The data was acquired on 3T Discovery MR750 (n = 45) (GE, Waukesha, WI), 3T MR Magnetom Trio (n = 7), Skyra (23) and 1.5T Symphony (n = 3) (Siemens, Erlagen, Germany). Twelve patients received STADT. There were no significant differences in the means of baseline radiomics variables, including the volumes of prostate, PZ and TZ between the (+/-ADT) groups. Similarly, there was no statistical difference between the AUC of the GTV on ADC and perfusion parameters. ADC values increased significantly on average 30%, 44% and 47% after treatment. ve also increased significantly on average 108%, 91% and 124% from baseline. Because of the different scanners, T2-weighted intensities were not compared. The study indicates that ADT does not affect the overall delta radiomics variables before and after radiotherapy. Prostate cancer has long natural history with relatively low number of failures. We demonstrate and +/-ADT patients can be analyzed together, allowing for increased number of events. The ultimate goal of the project is to identify delta radiomics signature that relates very early response to patient outcome such that therapeutic adjustments may be made to improve long term outcome. While 17% increase in ADC before and after ADT have been previously reported, the large magnitude of changes (∼40 %) here indicate that the radiation effects drive the observed variations.

Quantitative analysis of MRI-guided radiotherapy treatment process time for tumor real-time gating efficiency.

PurposeMagnetic Resonance‐guided radiotherapy (MRgRT) systems allow continuous monitoring of therapy volumes during treatment delivery and personalized respiratory gating approaches. Treatment length may therefore be significantly affected by patient’s compliance and breathing control. We quantitatively analyzed treatment process time efficiency (TE) using data obtained from real‐world patient treatment logs to optimize MRgRT delivery settings.MethodsData corresponding to the first 100 patients treated with a low T hybrid MRI‐Linac system, both in free breathing (FB) and in breath hold inspiration (BHI) were collected. TE has been computed as the percentage difference of the actual single fraction’s total treatment time and the predicted treatment process time, as computed by the TPS during plan optimization.Differences between the scheduled and actual treatment room occupancy time were also evaluated. Finally, possible correlations with planning, delivery and clinical parameters with TE were also investigated.ResultsNine hundred and nineteen treatment fractions were evaluated. TE difference between BHI and FB patients’ groups was statistically significant and the mean TE were 42.4%, and −0.5% respectively.No correlation was found with TE for BHI and FB groups. Planning, delivering and clinical parameters classified BHI and FB groups, but no correlation with TE was found.ConclusionThe use of BHI gating technique can increase the treatment process time significantly. BHI technique could be not always an adequate delivery technique to optimize the treatment process time. Further gating techniques should be considered to improve the use of MRgRT.

Quantitating Interfraction Target Dynamics During Concurrent Chemoradiation for Glioblastoma: A Prospective Serial Imaging Study

Magnetic resonance image (MRI) guided radiation therapy has the potential to improve outcomes for glioblastoma by adapting to tumor changes during radiation therapy. This study quantifies interfraction dynamics (tumor size, position, and geometry) based on sequential magnetic resonance imaging scans obtained during standard 6-week chemoradiation. Sixty-one patients were prospectively imaged with gadolinium-enhanced T1 (T1c) and T2/FLAIR axial sequences at planning (Fx0), fraction 10 (Fx10), fraction 20 (Fx20), and 1 month after the final fraction of chemoradiation therapy (P1M). Gross tumor volumes (GTVs) and clinical target volumes (CTVs) were contoured at all time points. Target dynamics were quantified by absolute volume (V), volume relative to Fx0 (Vrel), and the migration distance (dmigrate; the linear displacement of the GTV or CTV relative to Fx0). Temporal changes were assessed using a linear mixed-effects model. Median volumes at Fx0, Fx10, Fx20, and P1M for the GTV were 18.4 cm3 (range, 1.1-110.5 cm3), 14.7 cm3 (range, 0.9-115.1 cm3), 13.7 cm3 (range, 0.6-174.2 cm3), and 13.0 cm3 (range, 0.9-76.3 cm3), respectively, with corresponding median Vrel of 0.88 at Fx10, 0.77 at Fx20, and 0.71 at P1M relative to Fx0 (P < .001 for all). The GTV (CTV) migration distances were greater than 5 mm in 46% (54%) of patients at Fx10, 50% (58%) of patients at Fx20, and 52% (57%) of patients at P1M. Dynamic tumor morphologic changes were observed, with 40% of patients exhibiting a decreased GTV (Vrel ≤1) with a dmigrate >5 mm during chemoradiation therapy. Clinically meaningful tumor dynamics were observed during chemoradiation therapy for glioblastoma, supporting evaluation of daily MRI guided radiation therapy and treatment plan adaptation.

Enhanced MRI-guided radiotherapy with gadolinium-based nanoparticles: preclinical evaluation with an MRI-linac

BackgroundThe AGuIX® (NH TherAguix) nanoparticle has been developed to enhance radiotherapy treatment and provide strong MR contrast. These two properties have previously been investigated separately and progressed to clinical trial following a clinical workflow of separate MR imaging followed some time later by radiotherapy treatment. The recent development of MRI-linacs (combined Magnetic Resonance Imaging–linear accelerator systems enabling MRI-guided radiotherapy) opens up a new workflow where MR confirmation of nanoparticle uptake can be carried out at the time of treatment. A preclinical study was carried out to assess the suitability of a gadolinium-containing nanoparticle AGuIX® (NH TherAguix) for nano-enhanced image-guided radiotherapy on an MRI-linac.MethodsTreatments were carried out on F344 Fischer rats bearing a 9L glioma brain tumour. Animals received either: (A) no treatment; (B) injection of nanoparticles followed by MRI; (C) radiotherapy with MRI; or (D) injection of nanoparticles followed by radiotherapy with MRI. Pre-clinical irradiations were carried out on the 1.0 T, 6 MV in-line Australian MRI-linac. Imaging used a custom head coil specially designed to minimise interference from the radiotherapy beam. Anaesthetised rats were not restrained during treatment but were monitored with a cine-MRI sequence. Inductively Coupled Plasma Mass Spectrometry (ICP-MS) analysis was used to quantify residual gadolinium in the brain in normal and tumour tissue.ResultsA preclinical evaluation of nano-enhanced radiation treatment has been carried out on a 1.0 T MRI-linac, establishing a workflow on these novel systems. Extension of life when combining radiotherapy with nanoparticles was not statistically different from that for rats receiving radiotherapy only. However, there was no detrimental effect for animals receiving nanoparticles and radiation treatment in the magnetic field compared with control branches. Cine-MR imaging was sufficient to carry out monitoring of anaesthetised animals during treatment. AGuIX nanoparticles demonstrated good positive contrast on the MRI-linac system allowing confirmation of tumour extent and nanoparticle uptake at the time of treatment.ConclusionsNovel nano-enhanced radiotherapy with gadolinium-containing nanoparticles is ideally suited for implementation on an MRI-linac, allowing a workflow with time-of-treatment imaging. Live irradiations using this treatment workflow, carried out for the first time at the Australian MRI-linac, confirm the safety and feasibility of performing MRI-guided radiotherapy with AGuIX® nanoparticles. Follow-up studies are needed to demonstrate on an MRI-linac the radiation enhancement effects previously shown with conventional radiotherapy.

Dose response of three-dimensional silicone-based radiochromic dosimeters for photon irradiation in the presence of a magnetic field.

The magnetic field in magnetic resonance imaging guided radiotherapy (MRgRT) delivery systems influences charged-particle trajectories and hence the three-dimensional (3D) radiation dose distributions. This study investigated the dose-response as well as dose-rate and fractionation dependencies of silicone-based 3D radiochromic dosimeters for photon irradiation in a magnetic field using a 0.35 T MRgRT system. We found a linear dose response up to 22.6 Gy and no significant dose-rate dependency as a function of depth. A difference in optical response was observed for dosimeters irradiated in a single compared to multiple fractions. The dosimeter showed clinical potential for verification of MRgRT delivery.

Imaging and radiation isocentre determination for inline MR-guided radiotherapy systems – proof of principle using MR-phantom with embedded monolithic silicon detector

Multiple vendors are now offering real-time MRI-guided radiotherapy systems. Quality assurance of the imaging and radiation isocentre alignment is an essential component of the QA programme for any linear accelerator used for delivering image-guided radiotherapy. In this work, the authors describe the design and feasibility testing of a device that combines a high resolution monolithic silicon strip detector with an MRI visible phantom for characterisation of optical, MR imaging and radiation isocentre for inline MR-guided radiotherapy systems.

Feasibility of magnetic resonance guided radiotherapy for the treatment of bladder cancer

Whole bladder magnetic resonance image-guided radiotherapy using the 1.5 Telsa MR-linac is feasible. Full online adaptive planning workflow based on the anatomy seen at each fraction was performed. This was delivered within 45 min. Intra-fraction bladder filling did not compromise target coverage. Patients reported acceptable tolerance of treatment.

Edema Progression during MRI-Guided Glioblastoma Radiotherapy

There are limited studies evaluating brain changes on MRI during the six weeks of concurrent chemotherapy and radiotherapy (RT) for glioma. Existing studies suggest that MRI during chemoradiotherapy could be useful to determine early response to treatment. At our center, we have been treating selected gliomas with one of the first commercially available 0.35T MRI-RT systems. The system includes one clinically useable pulse sequence, balanced steady state free precession (bSSFP), which obtains a 1.5 mm isotropic highly T2-weighted 3D image volume through the brain with minimal distortions. The bSSFP MRIs are obtained daily for treatment setup, and we have seen changes occurring in and around gliomas during chemoradiotherapy. Our objective is to quantify the frequency, magnitude, and time course of these changes. Under a prospective registry protocol, we reviewed 14 patients at our institution who were treated for high grade glioma with chemoradiation 60 Gy in 30 fractions plus temozolomide on our MRI-RT system. We contoured the T2-weighted hyperintense volumes from the MRI-RT bSSFP images for all 30 treatment fractions for each patient and measured the volumes of those MRIs. The MRIs from fraction 30 were compared to the diagnostic MRIs obtained 3-4 weeks post-treatment to determine whether there were similar findings. 5 of 14 (36%) patients demonstrated greater than a 10% increase in tumor volume during the course of their treatment. In 3 of these patients with growth, changes predominantly were observed during week 5 and 6 of treatment. One patient’s non-enhancing unresected tumor decreased 24% in volume during the course of treatment. The bSSFP images obtained at the end of treatment demonstrated similar FLAIR abnormalities compared to diagnostic MRIs 3-4 weeks post treatment completion. Patients who had volume expansion during treatment that resolved on serial follow-up MRIs were counted as pseudoprogression, and excluding pseudoprogressors, the progression free survival in the cohort was 19 months. Median follow-up was 14.7 months. 36% of high grade glioma patients had growth of T2-weighted volume during chemoradiotherapy, and these changes were similar to those observed on post-treatment diagnostic scans. Tumor volume expansion could be due to pseudoprogression, true progression, or solely due to changes in edema. Our future work aims to differentiate these possibilities with additional patients and MRI pulse sequences. Changes were greatest during weeks 5-6 of the treatment course, suggesting that studies using diagnostic MRI scans during RT should be obtained late in the treatment course.

Problems and Promises of Introducing the Magnetic Resonance Imaging Linear Accelerator Into Routine Care: The Case of Prostate Cancer.

The new radiotherapy high field, 1.5 Tesla MRI-guided linear accelerator (MR-Linac) is being clinically introduced. Sensing and evaluating opportunities and barriers at an early stage will facilitate its eventual scale-up. This study investigates the opportunities and barriers to the implementation of MR-Linac into prostate cancer care based on 43 semi-structured interviews with Dutch oncology care professionals, hospital and division directors, patients, payers and industry. The analysis was guided by the Non-adoption, Abandonment, Scale-up, Spread, and Sustainability framework of new medical technologies and services. Opportunities included: the acquirement of (1) advanced MRI-guided radiotherapy technology with (2) the potential for improved patient outcomes and (3) economic benefits, as well as (4) professional development and (5) a higher hospital quality profile. Barriers included: (1) technical complexities, (2) substantial staffing and structural investments, (3) the current lack of empirical evidence of clinical benefits, (4) professional silos, and (5) the presence of patient referral patterns. While our study confirms the expected technical and clinical prospects from the literature, it also reveals economic, organizational, and socio-political challenges.

Treatment effect prediction for sarcoma patients treated with preoperative radiotherapy using radiomics features from longitudinal diffusion-weighted MRIs

The objective of this study was to explore radiomics features from longitudinal diffusion-weighted MRIs (DWIs) for pathologic treatment effect prediction in patients with localized soft tissue sarcoma (STS) undergoing hypofractionated preoperative radiotherapy (RT). Thirty patients with localized STS treated with preoperative hypofractionated RT were recruited to this longitudinal imaging study. DWIs were acquired at three time points using a 0.35 T MRI-guided radiotherapy system. Treatment effect score (TES) was obtained from the post-surgery pathology as a surrogate of treatment outcome. Patients were divided into two groups based on TES. Response prediction was first performed using a support vector machine (SVM) with only mean apparent diffusion coefficient (ADC) or delta ADC to serve as the benchmark. Radiomics features were then extracted from tumor ADC maps at each of the three time points. Logistic regression and SVM were constructed to predict the TES group using features selected by univariate analysis and sequential forward selection. Classification performance using SVM with features from different time points and with or without delta radiomics were evaluated. Prediction performance using only mean ADC or delta ADC was poor (area under the curve (AUC) < 0.7). For the radiomics study using features from all time points and corresponding delta radiomics, SVM significantly outperformed logistic regression (AUC of 0.91 ± 0.05 v.s. 0.85 ± 0.06). Prediction AUC values using single or multiple time points without delta radiomics were all below 0.74. Including delta radiomics of mid- or post-treatment relative to the baseline drastically boosted the prediction. In this work, an SVM model was built to predict the TES using radiomics features from longitudinal DWI. Based on this study, we found that use of mean ADC, delta ADC, or radiomics features alone was not sufficient for response prediction, and including delta radiomics features of mid- or post-treatment relative to the baseline can optimize the prediction of TES, a pathologic and clinical endpoint.

Toxicity reduction required for MRI-guided radiotherapy to be cost-effective in the treatment of localized prostate cancer.

To determine the toxicity reduction required to justify the added costs of MRI-guided radiotherapy (MR-IGRT) over CT-based image guided radiotherapy (CT-IGRT) for the treatment of localized prostate cancer. The costs of delivering prostate cancer radiotherapy with MR-IGRT and CT-IGRT in conventional 39 fractions and stereotactic body radiotherapy (SBRT) 5 fractions schedules were determined using literature values and cost accounting from two institutions. Gastrointestinal and genitourinary toxicity rates associated with CT-IGRT were summarized from 20 studies. Toxicity-related costs and utilities were obtained from literature values and cost databases. Markov modeling was used to determine the savings per patient for every 1% relative reduction in acute and chronic toxicities by MR-IGRT over 15 years. The costs and quality adjusted life years (QALYs) saved with toxicity reduction were juxtaposed with the cost increase of MR-IGRT to determine toxicity reduction thresholds for cost-effectiveness. One way sensitivity analyses were performed. Standard $100,000 and $50,000 per QALY ratios were used. The added cost of MR-IGRT was $1,459 per course of SBRT and $10,129 per course of conventionally fractionated radiotherapy. Relative toxicity reductions of 7 and 14% are required for SBRT to be cost-effective using $100,000 and $50,000 per QALY, respectively. Conventional radiotherapy requires relative toxicity reductions of 50 and 94% to be cost-effective. From a healthcare perspective, MR-IGRT can reasonably be expected to be cost-effective. Hypofractionated schedules, such a five fraction SBRT, are most likely to be cost-effective as they require only slight reductions in toxicity (7-14%). This is the first detailed economic assessment of MR-IGRT, and it suggests that MR-IGRT can be cost-effective for prostate cancer treatment through toxicity reduction alone.

Assessment of online adaptive MR-guided stereotactic body radiotherapy of liver cancers.

Online Adaptive Radiotherapy (ART) with daily MR-imaging has the potential to improve dosimetric accuracy by accounting for inter-fractional anatomical changes. This study provides an assessment for the feasibility and potential benefits of online adaptive MRI-Guided Stereotactic Body Radiotherapy (SBRT) for treatment of liver cancer. Ten patients with liver cancer treated with MR-Guided SBRT were included. Prescription doses ranged between 27 and 50Gy in 3-5 fx. All SBRT fractions employed daily MR-guided setup while utilizing cine-MR gating. Organs-at-risk (OARs) included duodenum, bowel, stomach, kidneys and spinal cord. Daily MRIs and contours were utilized to create each adapted plan. Adapted plans used the beam-parameters and optimization-objectives from the initial plan. Planning target volume (PTV) coverage and OAR constraints were used to compare non-adaptive and adaptive plans. PTV coverage for non-adapted treatment plans was below the prescribed coverage for 32/47 fractions (68%), with 11 fractions failing by more than 10%. All 47 adapted fractions met prescribed coverage. OAR constraint violations were also compared for several organs. The duodenum exceeded tolerance for 5/23 non-adapted and 0/23 for adapted fractions. The bowel exceeded tolerance for 5/34 non-adaptive and 1/34 adaptive fractions. The stomach exceeded tolerance for 4/19 non-adapted and 1/19 for adaptive fractions. Accumulated dose volume histograms were also generated for each patient. Online adaptive MR-Guided SBRT of liver cancer using daily re-optimization resulted in better target conformality, coverage and OAR sparing compared with non-adaptive SBRT. Daily adaptive planning may allow for PTV dose escalation without compromising OAR sparing.

Fast geometric distortion correction using a deep neural network: Implementation for the 1 Tesla MRI-Linac system.

Combining high-resolution magnetic resonance imaging (MRI) with a linear accelerator (Linac) as a single MRI-Linac system provides the capability to monitor intra-fractional motion and anatomical changes during radiotherapy, which facilitates more accurate delivery of radiation dose to the tumor and less exposure to healthy tissue. The gradient nonlinearity (GNL)-induced distortions in MRI, however, hinder the implementation of MRI-Linac system in image-guided radiotherapy where highly accurate geometry and anatomy of the target tumor is indispensable. To correct the geometric distortions in MR images, in particular, for the 1 Tesla (T) MRI-Linac system, a deep fully connected neural network was proposed to automatically learn the intricate relationship between the undistorted (theoretical) and distorted (real) space. A dataset, consisting of spatial samples acquired by phantom measurement that covers both inside and outside the working diameter of spherical volume (DSV), was utilized for training the neural network, which offers the ability to describe subtle deviations of the GNL field within the entire region of interest (ROI). The performance of the proposed method was evaluated on MR images of a three-dimensional (3D) phantom and the pelvic region of an adult volunteer scanned in the 1T MRI-Linac system. The experimental results showed that the severe geometric distortions within the entire ROI had been successfully corrected with an error less than the pixel size. Also, the presented network is highly efficient, which achieved significant improvement in terms of computational efficiency compared to existing methods. The feasibility of the presented deep neural network for characterizing the GNL field deviations in the 1T MRI-Linac system was demonstrated in this study, which shows promise in facilitating the MRI-Linac system to be routinely implemented in real-time MRI-guided radiotherapy.

Comparison of gross tumor volumes of pulmonary metastasis defined by CT and MRI in 0.345 T MRI-guided radiotherapy.

Objective:To assess the difference in gross tumor volumes (GTVs) defined by CT (GTV-CT) and by low magnetic field strength (0.345 T) MRI (GTV-MRI) in patients simulated for MRI-guided radiotherapy forlung metastasis.Methods:28 patients (148 lesions) who underwent CT and MRI simulation with the tri-60Co MRI-guided radiotherapy system (MRIdian, ViewRay) were included in this study. GTV-CT and GTV-MRI were compared using the paired t-test. The equivalence of variance between GTV-CT and GTV-MRI of small lesions (GTV-CT <1 ml) and large ones (GTV-CT >= 1 ml) was evaluated using F-test. The correlation between GTV-CT and GTV-MRI was evaluated by the correlation coefficient.Results:GTV-MRI was 120% larger than GTV-CT (p < 0.001) for small lesions, whereas GTV-MRI was 40% larger than GTV-CT (p < 0.001) for large lesions. In small lesions, the variation in GTV-MRI was significantly larger than that of GTV-CT (p < 0.001). There was no significant difference in the variation of GTV-MRI and GTV-CT in large lesions (p = 0.121). The correlation coefficient for small lesions was 0.93, whereas that for large lesions was 0.99, with large lesions having better correlation.Conclusions:GTV-MRI was larger than GTV-CT and the correlation between GTV-MRI and GTV-CT was better in large lesions. If the tumor volume is 1 ml or larger, the lesion can be accurately monitored even with a low magnetic field strength MRI.Advances in knowledge:This study is the first clinical report to evaluate the tolerability of MRI images in 0.345 T MRI-guided radiotherapy for lung metastasis. GTV contoured by MRI was larger than GTV by CT, and this tendency was more pronounced in small tumors of less than 1 ml.

Deep learning-based image reconstruction and motion estimation from undersampled radial k-space for real-time MRI-guided radiotherapy

To enable magnetic resonance imaging (MRI)-guided radiotherapy with real-time adaptation, motion must be quickly estimated with low latency. The motion estimate is used to adapt the radiation beam to the current anatomy, yielding a more conformal dose distribution. As the MR acquisition is the largest component of latency, deep learning (DL) may reduce the total latency by enabling much higher undersampling factors compared to conventional reconstruction and motion estimation methods. The benefit of DL on image reconstruction and motion estimation was investigated for obtaining accurate deformation vector fields (DVFs) with high temporal resolution and minimal latency.2D cine MRI acquired at 1.5 T from 135 abdominal cancer patients were retrospectively included in this study. Undersampled radial golden angle acquisitions were retrospectively simulated. DVFs were computed using different combinations of conventional- and DL-based methods for image reconstruction and motion estimation, allowing a comparison of four approaches to achieve real-time motion estimation. The four approaches were evaluated based on the end-point-error and root-mean-square error compared to a ground-truth optical flow estimate on fully-sampled images, the structural similarity (SSIM) after registration and time necessary to acquire k-space, reconstruct an image and estimate motion.The lowest DVF error and highest SSIM were obtained using conventional methods up to . For undersampling factors , the lowest DVF error and highest SSIM were obtained using conventional image reconstruction and DL-based motion estimation. We have found that, with this combination, accurate DVFs can be obtained up to with an average root-mean-square error up to 1 millimeter and an SSIM greater than 0.8 after registration, taking 60 milliseconds.High-quality 2D DVFs from highly undersampled k-space can be obtained with a high temporal resolution with conventional image reconstruction and a deep learning-based motion estimation approach for real-time adaptive MRI-guided radiotherapy.

Impact of a 1.5 T magnetic field on DNA damage in MRI-guided HDR brachytherapy

PurposeSome studies have suggested that the presence of a static magnetic field (SMF) during irradiation alters biological damage. Since MRI-guided radiotherapy is becoming increasingly common, we constructed a DNA-based detector to assess the effect of a 1.5 T SMF on DNA damage during high dose rate (HDR) brachytherapy irradiation. MethodsBlock phantoms containing a small cavity for the placement of plasmid DNA (pBR322) samples were 3-D printed with biocompatible tissue equivalent material. The phantom was CT scanned and an HDR brachytherapy treatment plan was designed to deliver 20 Gy and 30 Gy doses to the DNA samples in the presence and absence of a 1.5 T SMF. Relative yields of single- and double-strand breaks (SSBs and DSBs, respectively) were computed from gel electrophoresis images of the DNA band intensities and averaged over sample sizes ranging from 12 to 30. Radiation dose was also measured in the presence and absence of the 1.5 T SMF using GafChromic™ EBT3 film placed in the coronal, sagittal, and axial planes. ResultsThe average yield of DNA with SSBs and DSBs in the presence and absence of the SMF showed no statistically significant differences (all p ≥ 0.17). Differences in the net optical densities of the EBT3 films for each plane were within experimental uncertainty, suggesting no dose difference in the presence and absence of the SMF. ConclusionsHDR irradiation in the presence of the 1.5 T SMF did not alter dose deposition to the DNA cavity nor change SSB and DSB DNA damage.

Clinical workflow for treating patients with a metallic hip prosthesis using magnetic resonance imaging-guided radiotherapy.

Background & purposeMetallic prostheses distort the magnetic field during magnetic resonance imaging (MRI), leading to geometric distortions and signal loss. The purpose of this work was to develop a method to determine eligibility for MRI-guided radiotherapy (MRIgRT) on a per patient basis by estimating the magnitude of geometric distortions inside the clinical target volume (CTV).Materials & methodsThree patients with prostate cancer and hip prosthesis, treated using MRIgRT, were included. Eligibility for MRIgRT was based on computed tomography and associated CTV delineations, together with a field-distortion (B0) map and anatomical images acquired during MR simulation. To verify the method, B0 maps made during MR simulation and each MRIgRT treatment fraction were compared.ResultsEstimates made during MR simulation of the magnitude of distortions inside the CTV were 0.43 mm, 0.19 mm and 2.79 mm compared to the average over all treatment fractions of 1.40 mm, 0.32 mm and 1.81 mm, per patient respectively.ConclusionsB0 map acquisitions prior to treatment can be used to estimate the magnitude of distortions during MRIgRT to guide the decision on eligibility for MRIgRT of prostate cancer patients with metallic hip implants.

Label-driven magnetic resonance imaging (MRI)-transrectal ultrasound (TRUS) registration using weakly supervised learning for MRI-guided prostate radiotherapy

Registration and fusion of magnetic resonance imaging (MRI) and transrectal ultrasound (TRUS) of the prostate can provide guidance for prostate brachytherapy. However, accurate registration remains a challenging task due to the lack of ground truth regarding voxel-level spatial correspondence, limited field of view, low contrast-to-noise ratio, and signal-to-noise ratio in TRUS. In this study, we proposed a fully automated deep learning approach based on a weakly supervised method to address these issues. We employed deep learning techniques to combine image segmentation and registration, including affine and nonrigid registration, to perform an automated deformable MRI-TRUS registration. To start with, we trained two separate fully convolutional neural networks (CNNs) to perform a pixel-wise prediction for MRI and TRUS prostate segmentation. Then, to provide the initialization of the registration, a 2D CNN was used to register MRI-TRUS prostate images using an affine registration. After that, a 3D UNET-like network was applied for nonrigid registration. For both the affine and nonrigid registration, pairs of MRI-TRUS labels were concatenated and fed into the neural networks for training. Due to the unavailability of ground-truth voxel-level correspondences and the lack of accurate intensity-based image similarity measures, we propose to use prostate label-derived volume overlaps and surface agreements as an optimization objective function for weakly supervised network training. Specifically, we proposed a hybrid loss function that integrated a Dice loss, a surface-based loss, and a bending energy regularization loss for the nonrigid registration. The Dice and surface-based losses were used to encourage the alignment of the prostate label between the MRI and the TRUS. The bending energy regularization loss was used to achieve a smooth deformation field. Thirty-six sets of patient data were used to test our registration method. The image registration results showed that the deformed MR image aligned well with the TRUS image, as judged by corresponding cysts and calcifications in the prostate. The quantitative results showed that our method produced a mean target registration error (TRE) of 2.53 ± 1.39 mm and a mean Dice loss of 0.91 ± 0.02. The mean surface distance (MSD) and Hausdorff distance (HD) between the registered MR prostate shape and TRUS prostate shape were 0.88 and 4.41 mm, respectively. This work presents a deep learning-based, weakly supervised network for accurate MRI-TRUS image registration. Our proposed method has achieved promising registration performance in terms of Dice loss, TRE, MSD, and HD.