MRI-guided Radiotherapy Research Articles

A convolutional neural network for contouring metastatic lymph nodes on diffusion-weighted magnetic resonance images for assessment of radiotherapy response.

Retrieving quantitative parameters from magnetic resonance imaging (MRI), e.g. for early assessment of radiotherapy treatment response, necessitates contouring regions of interest, which is time-consuming and prone to errors. This becomes more pressing for daily imaging on MRI-guided radiotherapy systems. Therefore, we trained a deep convolutional neural network to automatically contour involved lymph nodes on diffusion-weighted (DW) MRI of head and neck cancer (HNC) patients receiving radiotherapy. DW-images from 48 HNC patients (18 induction-chemotherapy+chemoradiotherapy; 30 definitive chemoradiotherapy) with 68 involved lymph nodes were obtained on a diagnostic 1.5T MR-scanner prior to and 2-3 timepoints throughout treatment. A radiation oncologist delineated the lymph nodes on the b=50s/mm2 images. A 3D U-net was trained to contour involved lymph nodes. Its performance was evaluated in all 48 patients using 8-fold cross-validation and calculating the Dice similarity coefficient (DSC) and the absolute difference in median apparent diffusion coefficient (ΔADC) between the manual and generated contours. Additionally, the performance was evaluated in an independent dataset of three patients obtained on a 1.5T MR-Linac. In the definitive chemoradiotherapy patients (n=96 patients/lymphnodes/timepoints) the DSC was 0.87 (0.81-0.91) [median (1st-3rd quantiles)] and ΔADC was 1.9% (0.8-3.4%) and both remained stable throughout treatment. The network performed worse in the patients receiving induction-chemotherapy (n=65), with DSC=0.80 (0.71-0.87) and ΔADC=3.3% (1.6-8.0%). The network performed well on the MR-Linac data (n=8) with DSC=0.80 (0.75-0.82) and ΔADC=4.0% (0.6-9.1%). We established accurate automatic contouring of involved lymph nodes for HNC patients on diagnostic and MR-Linac DW-images.

The noise navigator for MRI-guided radiotherapy: an independent method to detect physiological motion

Motion is problematic during radiotherapy as it could lead to potential underdosage of the tumor, and/or overdosage in organs-at-risk. A solution is adaptive radiotherapy guided by magnetic resonance imaging (MRI). MRI allows for imaging of target volumes and organs-at-risk before and during treatment delivery with superb soft tissue contrast in any desired orientation, enabling motion management by means of (real-time) adaptive radiotherapy.The noise navigator, which is independent of the MR signal, could serve as a secondary motion detection method in synergy with MR imaging. The feasibility of respiratory motion detection by means of the noise navigator was demonstrated previously. Furthermore, from electromagnetic simulations we know that the noise navigator is sensitive to tissue displacement and thus could in principle be used for the detection of various types of motion.In this study we demonstrate the detection of various types of motion for three anatomical use cases of MRI-guided radiotherapy, i.e. torso (bulk movement and variable breathing), head-and-neck (swallowing) and cardiac. Furthermore, it is shown that the noise navigator can detect bulk movement, variable breathing and swallowing on a hybrid 1.5 T MRI-linac system. Cardiac activity detection through the noise navigator seems feasible in an MRI-guided radiotherapy setting, but needs further optimization. The noise navigator is a versatile and fast (millisecond temporal resolution) motion detection method independent of MR signal that could serve as an independent verification method to detect the occurrence of motion in synergy with real-time MRI-guided radiotherapy.

Gel dosimetry provides the optimal end-to-end quality assurance dosimetry for MR-linacs.

Major advances in external beam radiation therapy in the last decade have provided technologies for planning and delivering highly conformal radiation dose distributions made available by refined beam targeting methods, including intensity-modulated radiation therapy using tomotherapy, volumetric modulated arc therapy and many other auspicious tools for 3D dose painting. The emergence of magnetic resonance imaging (MRI) systems integrated with linear accelerators (linacs), commonly known as an MR-linacs, have realized MR-image guided radiation therapy (MRgRT) systems.

A motion prediction confidence estimation framework for prediction-based radiotherapy gating.

Motion prediction can compensate for latency in image-guided radiotherapy and has been an active area of research. However, motion predictions are subject to error and variations. We have developed and evaluated a novel motion prediction confidence estimation framework to improve the efficacy and robustness of prediction-based radiotherapy gating decision-making. The specific scenario of adaptive gating in magnetic resonance imaging (MRI)-guided radiotherapy is studied as an example, but the method generalizes to other modalities and motion management setups. The proposed prediction confidence estimator is based on a generic training/testing paradigm and consists of a weighted combination of three components: the prediction model's goodness of fit, variation in the prediction using a leave-one-out process and the velocity of the tracked target. Roughly, these terms quantify respectively the consistency between prediction and the training data, the robustness of model inference, and the stability due to target speed. The weight parameters and the action level in triggering beam-off decision are optimized. The method is assessed and validated in 8 healthy volunteer and 13 patient studies using a 0.35T MRI-guided radiotherapy system predicting 0.25-0.33s ahead. The effect of the action level on the predicted gating decision accuracy, beam-on positive predictive value (PPV) and median distance between the predicted and ground-truth target centroids were evaluated. Statistical significance was evaluated using a paired t-test. The tradeoff between these performance metrics and gating duty cycle was assessed. Use of the confidence estimator threshold increased gating accuracy by up to 2.42%, increased PPV by up to 3.00%, and reduced the median centroid distance up to 0.28mm. The confidence estimator threshold on average increased gating accuracy to 96.5% (P=2.08×10-4 ), increased PPV to 96.7% (P=1.46×10-5 ), reduced the median centroid distance to 0.54mm (P=1.71×10-5 ) at the cost of reducing the gating duty cycle by 14.3% to 48.5%. Hyperparameter tuning revealed that contrary to intuition, the velocity term offered only minimal performance improvement in some cases but also introduced potential stability issues. The combination of goodness of fit and leave-one-out prediction variation provided the most effective confidence estimator, yielding universally better performance in gating decisions. Confidence estimation utilizing prediction model fitness criterion and validation principles can complement prediction methods to guide MRI-guided radiotherapy gating. Results from both volunteer and patient studies showed improved gating quality.

Alanine dosimetry in strong magnetic fields: use as a transfer standard in MRI-guided radiotherapy

Reference dosimetry in the presence of a strong magnetic field is challenging. Ionisation chambers have shown to be strongly affected by magnetic fields. There is a need for robust and stable detectors in MRI-guided radiotherapy (MRIgRT). This study investigates the behaviour of the alanine dosimeter in magnetic fields and assesses its suitability to act as a reference detector in MRIgRT.Alanine pellets were loaded in a waterproof holder, placed in an electromagnet and irradiated by 60Co and 6 MV and 8 MV linac beams over a range of magnetic flux densities. Monte Carlo simulations were performed to calculate the absorbed dose, to water and to alanine, with and without magnetic fields. Combining measurements with simulations, the effect of magnetic fields on alanine response was quantified and a correction factor for the presence of magnetic fields on alanine was determined.This study finds that the response of alanine to ionising radiation is modified when the irradiation is in the presence of a magnetic field. The effect is energy independent and may increase the alanine/electron paramagnetic resonance (EPR) signal by 0.2% at 0.35 T and 0.7% at 1.5 T. In alanine dosimetry for MRIgRT, this effect, if left uncorrected, would lead to an overestimate of dose. Accordingly, a correction factor, , is defined. Values are obtained for this correction as a function of magnetic flux density, with a standard uncertainty which depends on the magnetic field and is 0.6% or less.The strong magnetic field has a measurable effect on alanine dosimetry. For alanine which is used to measure absorbed dose to water in a strong magnetic field, but which has been calibrated in the absence of a magnetic field, a small correction to the reported dose is required. With the inclusion of this correction, alanine/EPR is a suitable reference dosimeter for measurements in MRIgRT.

Impact of the cavity on sinus wall dose in magnetic resonance image-guided radiation therapy

PurposeThis study aims to investigate the impact of the cavity on the sinus wall dose by comparing dose distributions with and without the sinus under magnetic fields using Monte Carlo calculations. MethodsA water phantom containing a sinus cavity (Empty) was created, and dose distributions were calculated for 1, 2, and 4 irradiation fields with 6 MV photons. The sinus in the phantom was then filled with water (Full), and the dose distributions were calculated again. The sinus was set to cubes of 2 cm and 4 cm. The magnetic field was applied to the transverse and inline direction under the magnetic flux densities of 0 T, 0.35 T, 0.5 T, 1.0 T, and 1.5 T. The dose distributions were analyzed by the dose difference, dose volume histogram, and D2 with sinus wall thicknesses of 1 and 5 mm. ResultsD2 in the “Empty” sinus wall under transverse magnetic fields for the 1-field and 4-field cases was 51.9% higher and 3.7% lower than that in the “Full” sinus wall at 1.5 T, respectively. Meanwhile, D2 in the Empty sinus wall under inline magnetic fields for 1-field and 4-fields was 2.3% and 2.6% lower than that in the “Full” sinus at B = 0 T, respectively, whereas D2 was 0.9% and 0.7% larger at 1.0 T, respectively. ConclusionsThe impact of the cavity on the sinus wall dose depends on the magnetic flux density, direction of the magnetic field and irradiation beam, and number of irradiation fields.

Commissioning of a 1.5T Elekta Unity MR-linac: A single institution experience.

MR image‐guided radiotherapy has the potential to improve patient care, but integration of an MRI scanner with a linear accelerator adds complexity to the commissioning process. This work describes a single institution experience of commissioning an Elekta Unity MR‐linac, including mechanical testing, MRI scanner commissioning, and dosimetric validation. Mechanical testing included multileaf collimator (MLC) positional accuracy, measurement of radiation isocenter diameter, and MR‐to‐MV coincidence. Key MRI tests included magnetic field homogeneity, geometric accuracy, image quality, and the accuracy of navigator‐triggered imaging for motion management. Dosimetric validation consisted of comparison between measured and calculated PDDs and profiles, IMRT measurements, and end‐to‐end testing. Multileaf collimator positional accuracy was within 1.0 mm, the measured radiation isocenter walkout was 0.20 mm, and the coincidence between MR and MV isocenter was 1.06 mm, which is accounted for in the treatment planning system (TPS). For a 350‐mm‐diameter spherical volume, the peak‐to‐peak deviation of the magnetic field homogeneity was 4.44 ppm and the geometric distortion was 0.8 mm. All image quality metrics were within ACR recommendations. Navigator‐triggered images showed a maximum deviation of 0.42, 0.75, and 3.0 mm in the target centroid location compared to the stationary target for a 20 mm motion at 10, 15, and 20 breaths per minute, respectively. TPS‐calculated PDDs and profiles showed excellent agreement with measurement. The gamma passing rate for IMRT plans was 98.4 ± 1.1% (3%/ 2 mm) and end‐to‐end testing of adapted plans showed agreement within 0.4% between ion‐chamber measurement and TPS calculation. All credentialing criteria were satisfied in an independent end‐to‐end test using an IROC MRgRT phantom.

Magnetic resonance-guided radiotherapy feasibility in elderly cancer patients: proposal of the MASTER scoring system.

Elderly patients are often excluded from advanced treatments owing to clinical complexity or frailty. Magnetic resonance-guided radiotherapy (MRgRT) represents a new frontier of radiotherapy delivery that can play an important role in the management of these patients. To assess MRgRT feasibility in elderly patients, describe their compliance with this treatment, and provide a scoring system for elderly patient selection. Patients aged >75 years were enrolled. No restrictions on tumor site, staging, or treatment intent were applied. Patients underwent joint radiation oncology-geriatrics visits to assess the feasibility of MRgRT and to identify the most significant items (i.e. clinical variables) for the setup of a scoring system. The proposed scoring system was then internally validated on a prospectively enrolled cohort of elderly patients who were candidates for MRgRT. Thirty patients were enrolled between February and March 2018. Their mean age was 81.4 ± 3.4 years (range 75-88). Radiotherapy intent was curative in 26 patients; 14 patients were considered frail at screening tests before radiotherapy. Twelve items were identified as clinically significant for the setup of the MASTER score (MRI-Guided Radiotherapy Selection Elderly Score) score. Validation of the score showed 100% reliability, with no patient discharged after selection. MRgRT appears to be feasible in elderly patients and the MASTER score is proposed to support clinical decision-making in recommending elderly patients for this technology.

Impact of transverse magnetic fields on water equivalent ratios in carbon-ion radiotherapy

The correct estimation of water equivalent ratio (WER) and the associated errors is highly important for the delivery of carbon-ion (C-ion) beams. In the case of MRI-guided C-ion radiotherapy (MRgCT) scenario, as a therapeutic approach for future potential MR-guided particle therapy, C-ion beam will be deflected in the presence of the magnetic field. The curvature of the track causes a retraction of the range. This range variation might be different in various materials, which leads to change the WER values. To evaluate the WER variations versus magnetic field strength, in this research work, WER of some materials, including bladder, brain, prostate, muscle, bone, polymethylmethacrylate (PMMA), polyoxymethylene (POM), polyethyleneterephthalate (PET), titanium (Ti), gold (Au), platinum (Pt); potentially encountered in C-ion radiotherapy dosimetry were assessed. A mono-energetic C-ion beam, incident on a volume containing materials as mentioned above, was simulated using the FLUKA Monte Carlo code. To validate the simulated C-ion beams, the ion ranges in the water at the calculated energies (100–400 MeV/n) were compared with experimental and analytical data reported in the literature. Moreover, the calculated WER results were compared with available experimental data in the absence of the magnetic field. The WER values were calculated for the materials mentioned above in the presence of 0.35, 1.5, and 3 T magnetic fields and compared to the case without a magnetic field. Good agreement with the experimental data regarding range prediction in water was achieved. No change in the WER value was observed at 100 MeV/n and 0.35 T for all the studied materials and energies. In the case of bladder, brain, and prostate, and muscle materials, no change in WER values was observed by changing the magnetic field strength up to 3 T at the range of the studied energies. The maximum change in WER values by applying the magnetic fields is relevant to the Ti material (−0.4%), which occurs at 400 MeV/n energy and 3 T magnetic field. Considering the potential encounter of the studied materials in the clinical practice of MRgCT, in the form of phantoms, dosimeters, detectors, fiducial markers, radio-opaque clips, patient anatomies, etc., the results of this study highly contribute to assess the variation of WER values of the studied dosimetric materials and the changes in C-ion ranges within these materials when the transverse magnetic fields are applied to the materials subject to study.

How Technology Is Improving the Multidisciplinary Care of Sarcoma.

Sarcomas are rare tumors but comprise a wide histologic spectrum. Advances in technology have emerged to address the biologic complexity and challenging diagnosis and treatment of this disease. The diagnostic approach to sarcomas has historically been based on morphologic features, but technologic advances in immunohistochemistry and cytogenetic/molecular testing have transformed the interdisciplinary work-up of mesenchymal neoplasms in recent years. On the therapeutic side, technologic advances in the delivery of radiation have made it a linchpin in the treatment of localized and oligometastatic sarcoma. In this review, we discuss recent advances in the pathologic diagnosis of sarcomas and discuss select sarcoma types that illustrate how newly discovered diagnostic, prognostic, and predictive biomarkers have refined existing classification schemes and substantially shaped our diagnostic approach. Such examples include conventional and epithelioid malignant peripheral nerve sheath tumors (MPNSTs), emerging entities in the group of round cell sarcomas, and other mesenchymal neoplasms with distinct cytogenetic aberrations. Recent advances in radiation oncology, including intensity-modulated, stereotactic, MRI-guided, and proton radiotherapy (RT), will be reviewed in the context of neoadjuvant or adjuvant localized soft-tissue sarcoma and oligometastatic or oligoprogressive disease. Innovations in translational research are expected to be introduced into clinical practice over the next few years and will likely continue to affect the rapidly evolving field of sarcoma diagnostics and therapy.

Development and evaluation of machine learning models for voxel dose predictions in online adaptive magnetic resonance guided radiation therapy.

PurposeDaily online adaptive plan quality in magnetic resonance imaging guided radiation therapy (MRgRT) is difficult to assess in relation to the fully optimized, high quality plans traditionally established offline. Machine learning prediction models developed in this work are capable of predicting 3D dose distributions, enabling the evaluation of online adaptive plan quality to better inform adaptive decision‐making in MRgRT.MethodsArtificial neural networks predicted 3D dose distributions from input variables related to patient anatomy, geometry, and target/organ‐at‐risk relationships in over 300 treatment plans from 53 patients receiving adaptive, linac‐based MRgRT for abdominal cancers. The models do not include any beam related variables such as beam angles or fluence and were optimized to balance errors related to raw dose and specific plan quality metrics used to guide daily online adaptive decisions.ResultsAveraged over all plans, the dose prediction error and the absolute error were 0.1 ± 3.4 Gy (0.1 ± 6.2%) and 3.5 ± 2.4 Gy (6.4 ± 4.3%) respectively. Plan metric prediction errors were −0.1 ± 1.5%, −0.5 ± 2.1%, −0.9 ± 2.2 Gy, and 0.1 ± 2.7 Gy for V95, V100, D95, and Dmean respectively. Plan metric prediction absolute errors were 1.1 ± 1.1%, 1.5 ± 1.5%, 1.9 ± 1.4 Gy, and 2.2 ± 1.6 Gy. Approximately 10% (25) of the plans studied were clearly identified by the prediction models as inferior quality plans needing further optimization and refinement.ConclusionMachine learning prediction models for treatment plan 3D dose distributions in online adaptive MRgRT were developed and tested. Clinical integration of the models requires minimal effort, producing 3D dose predictions for a new patient’s plan using only target and OAR structures as inputs. These models can enable improved workflows for MRgRT through more informed plan optimization and plan quality assessment in real time.

Generative Adversarial Networks based synthetic-CT generation for patients with nasopharyngeal carcinoma

Objective To establish a correlation model between MRI and CT images to generate synthetic-CT (sCT) of head and neck cancer during MRI-guided radiotherapy by using generative adversarial networks (GAN). Methods Images and IMRT plans of 45 patients with nasopharyngeal carcinoma were collected before treatment. Firstly, the MRI (T1) and CT images were preprocessed, including rigid registration, clipping, background removal and data enhancement, etc. Secondly, the cases were trained by GAN, of which 30 cases were randomly selected and put into the network as training set images for modeling and learning, and the other 15 cases were used for testing. The image quality of predicted sCT and real CT were statistically compared, and the dose distribution recalculated upon predicted sCT was statistically compared with that of real planned dose distribution. Results The mean absolute error of the predicted sCT of the testing set was (79.15±11.37) HU, and the SSIM value was 0.83±0.03. The MAE values of dose distribution difference at different regional levels were less than 1% compared to the prescription dose. The gamma passing rate of the sCT dose distribution was higher than 92% and 98% under the 2mm/2% and 3mm/3% criteria. Conclusions We have successfully proposed and realized the generation of sCT for head and neck cancer using GAN, which lays a foundation for the implementation of MRI-guided radiotherapy. The comparison of image quality and dosimetry shows the feasibility and accuracy of this method. Key words: Nasopharyngeal neoplasm/magnetic resonance-image guided radiotherapy; Generative adversarial networks; Synthetic-CT generation

Impact of a parallel magnetic field on radiation dose beneath thin copper and aluminum foils

Purpose: The RF coils for magnetic resonance image guided radiotherapy (MRIgRT) may be constructed using thin and/or low-density conductors, along with thinner enclosure materials. This work measures the surface dose increases for lightweight conductors and enclosure materials in a magnetic field parallel to a 6 MV photon beam. Methods: Aluminum and copper foils (9–127 μm thick), as well as samples of polyimide (17 μm) and polyester (127 μm) films are positioned atop a polystyrene phantom. A parallel plate ion chamber embedded into the top of the phantom measures the surface dose in 6 MV photon beam. Measurements (% of dose at the depth of maximum dose) are performed with and without a parallel magnetic field (0.22T at magnet center). Results: In the presence of a magnetic field, the unobstructed surface dose is higher (31.9%Dmax versus 22.2%Dmax). The surface dose is found to increase linearly with thickness for thin (<25 μm) copper (0.339%Dmaxμm−1) and aluminum (0.116%Dmaxμm−1) foils. In the presence of a magnetic field the slope is lower (copper: 0.16%Dmaxμm−1, aluminum: 0.06%Dmaxμm−1). The effect of in-beam foils is reduced due to partial shielding of the surface from contaminant electrons. Copper causes a surface dose increase ≈3 times higher than aluminum of the same thickness, consistent with their relative electron density. Polyester film (127μm) increases the surface dose (to 35% Dmax with field) about as much as a gown (36% Dmax with field), while the increase with polyimide film (17μm) is less than 1% above the open field dose. Conclusions: Thin copper and aluminum conductors increase surface dose by an amount comparable to a hospital gown. Similarly, enclosure materials made of thin polyester or polyimide film increase surface dose by only a few %Dmax in excess of an unobstructed beam. Based on measurements in this study, in-beam, surface RF coils are feasible for MRIgRT systems.

Dose uncertainty and resolution of polymer gel dosimetry using an MRI guided radiation therapy system’s onboard 0.35 T scanner

Magnetic Resonance Imaging (MRI) scanners are widely used for 3D gel dosimeters readout. However, limited access to MRI scanners is a challenge in MRI-based gel dosimetry. Recent clinical implementation of MRI-guided radiation therapy machines provides potential opportunities for onboard gel dosimetry using its MRI subsystem. The objective of this study was to investigate the feasibility of gel dosimetry using ViewRay’s onboard 0.35 T MRI scanner. A BANG® polymer gel dosimeter was irradiated by three beams of 3 × 3 cm2 field size. The T2 relaxation rate (R2) of the irradiated gel was measured using a Philips 1.5 T Ingenia MRI and a ViewRay 0.35 T onboard MRI and spin-echo pulse sequences. The number of signal averages (NSA) was set to 16 for the ViewRay acquisitions and one for the Philips 1.5 T MRI to achieve similar signal-to-noise ratios. The in-plane spatial resolution was 1.5 × 1.5 mm2 and the slice thickness was 5 mm. The relative dose uncertainty was obtained using R2 versus dose curves to compare the performance of dosimetry using the two different MRIs and field strengths. The dose uncertainty decreased from 12% at 2 Gy to 3.5% at 7.5 Gy at 1.5 T. The dose uncertainty decreased from 13% at 2 Gy to 4% at 7.5 Gy with NSA = 16 and 3 × 3 mm2 pixel size, and from 10.5% at 2 Gy to 3.2% at 7.5 Gy with NSA = 16 and denoised R2 maps (1.5 × 1.5 mm2 pixel size) at 0.35 T. The mean of dose resolution was 0.4 Gy at 1.5 T while the mean of dose resolution was 0.8 Gy and 0.64 Gy at 0.35 T by downsampling and denoising the R2 map, respectively. Therefore, comparable dose uncertainty was achievable using the ViewRay’s onboard 0.35 T and Philips 1.5 T MRI scanners. 3D gel dosimetry using onboard low-field MRI scanner provides ViewRay users a 3D high resolution dosimetry option besides film and ionization chamber.

Abstract B69: Assessment of CCR5/maraviroc immunotherapy in combination with PD1 and MR-guided radiotherapy for treatment of pancreatic cancer

Abstract Background: Pancreatic ductal adenocarcinoma (PDAC) is the most prevalent form of pancreatic cancers with poor survival outcomes, due to late diagnosis, its propensity to metastasize, and lack of therapeutic efficacy under current chemotherapeutic regimens. Despite recent advancements in the field of immunotherapy, results from the clinic have demonstrated lack of efficacy when either radiotherapy (RT) or immunotherapy is used as a monotherapy in PDAC. However, recent publications revealed a synergistic effect on RT-induced immune modulation and reduced immune suppression when administered concurrently. Moreover, these reports demonstrate immune evasion in PDAC is dependent on the CCL5/CCR5 axis to recruit immunosuppressive T-regulatory cells (Tregs) into the tumor microenvironment. Therefore, targeting the migration of Tregs through modulation of CCL5/CCR5 can potentially inhibit tumor growth in pancreatic cancer. Aim: This preclinical study evaluates the impact of fractionated MR-image guided radiotherapy (MR-IGRT) in combination CCR5 inhibitor and simultaneous inhibition of the immune checkpoint axis PD1/PD-L1 in PDAC. Methods: For the purpose of this study we generated a syngeneic orthotopic pancreatic mouse model. Tumor cells derived from the Lox-Stop-Lox (LSL)-KrasG12D; LSL-Trp53R172H; Pdx1-cre (KPC) mouse model are injected in the tail of the pancreas. We are able to monitor tumor growth using a respiratory motion desensitized T2-weighted MRI imager, allowing the generation of high-resolution and high-contrast MRI data. Taking advantage of in-house developed technology, the MR-IGRT was delivered using the Small Animal Radiation Research Platform (SARRP) in combination with MRI imaging to deliver MR-guided fractionated radiotherapy. Concurrently with the radiotherapy, CCR5 inhibitor (maraviroc) and PD1 inhibitor were administrated at specific time points. To investigate the immunologic microenvironment, we developed a 17-color flow cytometry (FC) panel to immune-phenotype cytotoxic T, T regulatory, NK, NK/T and B cells, M-MDSC, PMN-MDSC, M1 and M2 macrophages in the peripheral blood and tumor infiltrate. Results/Conclusions: Using the established orthotopic mouse model and immune monitoring, we are investigating the therapeutic benefit of immunomodulation of the CCL5/CCR5 axis in combination with PD1/PDL1 and radiotherapy. Citation Format: Sophie Hughes, Simone Lanfredini, Asmita Thappa, Somnath Mukherjee, Eric O’Neill. Assessment of CCR5/maraviroc immunotherapy in combination with PD1 and MR-guided radiotherapy for treatment of pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference on Tumor Immunology and Immunotherapy; 2019 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Immunol Res 2020;8(3 Suppl):Abstract nr B69.

Development and Validation of a Clinically Relevant Workflow for MR-Guided Volumetric Arc Therapy in a Rabbit Model of Head and Neck Cancer.

There is increased interest in the use of magnetic resonance imaging (MRI) for guiding radiation therapy (RT) in the clinical setting. In this regard, preclinical studies can play an important role in understanding the added value of MRI in RT planning. In the present study, we developed and validated a clinically relevant integrated workflow for MRI-guided volumetric arc therapy (VMAT) in a VX2 rabbit neck tumor model of HNSCC. In addition to demonstrating safety and feasibility, we examined the therapeutic impact of MR-guided VMAT using a single high dose to obtain proof-of-concept and compared the response to conventional 2D-RT. Contrast-enhanced MRI (CE-MRI) provided excellent soft tissue contrast for accurate tumor segmentation for VMAT. Notably, MRI-guided RT enabled improved tumor targeting ability and minimal dose to organs at risk (OAR) compared to 2D-RT, which resulted in notable morbidity within a few weeks of RT. Our results highlight the value of integrating MRI into the workflow for VMAT for improved delineation of tumor anatomy and optimal treatment planning. The model combined with the multimodal imaging approach can serve as a valuable platform for the conduct of preclinical RT trials.

MRSIGMA: Magnetic Resonance SIGnature MAtching for real-time volumetric imaging.

To propose a real-time 3D MRI technique called MR SIGnature MAtching (MRSIGMA) for high-resolution volumetric imaging and motion tracking with very low imaging latency. MRSIGMA consists of two steps: (1) offline learning of a database of possible 3D motion states and corresponding motion signature ranges and (2) online matching of new motion signatures acquired in real time with prelearned motion states. Specifically, the offline learning step (non-real-time) reconstructs motion-resolved 4D images representing different motion states and assigns a unique motion range to each state. The online matching step (real-time) acquires motion signatures only and selects one of the prelearned 3D motion states for each newly acquired signature, which generates 3D images efficiently in real time. The MRSIGMA technique was evaluated on 15 golden-angle stack-of-stars liver data sets, and the performance of respiratory motion tracking with the online-generated real-time 3D MRI was compared with the corresponding 2D projections acquired in real time. The total latency of generating each 3D image during online matching was about 300ms, including acquisition of the motion signature data (~138ms) and corresponding matching process (~150ms). Linear correlation assessment suggested excellent correlation (R2 =0.948) between motion displacement measured from the online-generated real-time 3D images and the 2D real-time projections. This proof-of-concept study demonstrates the feasibility of MRSIGMA for high-resolution real-time volumetric imaging, which shifts the acquisition and reconstruction burden to an offline learning step and leaves fast online matching for online imaging with very low imaging latency. The MRSIGMA technique can potentially be used for real-time motion tracking in MRI-guided radiation therapy.

Feasibility of energy adaptive angular meshing for perpendicular and parallel magnetic fields in a grid based Boltzmann solver

Purpose: To develop the enabling algorithmic techniques which allow forward-peaked adaptive angular meshing to be compatible with angular advection of magnetic fields within a deterministic Grid Based Boltzmann Solver (GBBS) for MRI-guided radiotherapy, and establish appropriate energy adaptive meshing schemes which minimize total numerical degrees of freedom while preserving high dosimetric accuracy for parallel and perpendicular magnetic fields. Methods: A framework to independently adapt angular mesh resolution and basis function refinement of forward and backscattering hemispheres is developed, uniquely accommodating angular advection introduced by magnetic fields. Upwind stabilization techniques to accurately transfer fluence between hemispheres having different discretization are established. To facilitate oblique beam and magnetic field orientations, cardinal forward-peaked mesh orientations were devised to balance requirements for acyclic space-angle sweep ordering, while ensuring the beam predominantly overlaps the forward hemisphere. Energy-dependent fluence anisotropy is investigated, leading to adaptive angular meshing schemes for parallel and perpendicular magnetic fields. Calculated dose distributions were validated against GEANT4 Monte Carlo calculations on slab geometry and anthropomorphic phantoms. Results: Forward-peaked and isotropic energy adaptive angular meshing schemes were developed for parallel and perpendicular magnetic fields respectively, which reduce the number of elements solved by 52.8% and 47.7% respectively compared to static discretization using 32 quadratic elements while retaining over 97% of points passing the gamma 1%/1 mm criterion against Monte Carlo. Conclusions: Techniques to preserve angular upwind-stabilization between hemispheres of a forward-peaked mesh and establish an acyclic directed space-angle sweep graph enabled energy-adaptive meshing schemes to be developed while accurately solving for magnetic fields. This substantially reduced the numerical degrees of freedom while retaining excellent dosimetric agreement with Monte Carlo. These algorithmic underpinnings contribute towards a fast deterministic GBBS for MRI-guided radiotherapy.

Abstract P4-12-21: Implementing a real-time magnetic resonance imaging guided accelerated partial breast irradiation program

Abstract Motivation: Clinical evidence has shown that accelerated partial breast irradiation (APBI) for early stage breast cancer patients result in local recurrence control rates and cosmetic outcomes comparable to whole breast irradiation. Approaches to APBI include brachytherapy, intraoperative radiation therapy and external beam radiation therapy (EBRT). EBRT is the least invasive approach, most widely available, and does not require additional training. For EBRT APBI, based on NSABP B39/RTOG 0413 protocol, the clinical target volume (CTV) is generated by applying a 1.5 cm margin to the lumpectomy bed and the planning target volume (PTV) is generated by applying an additional 1 cm margin from the CTV to account for set up uncertainties.Real-time magnetic resonance imaging guided radiation therapy (MRgRT) allow for reduced treatment margins by decreasing treatment setup uncertainties. In this work, we share our experience in implementing a MRI guided accelerated partial breast irradiation (MRgAPBI) program into a new clinic with a linear accelerator capable of real-time MRI imaging. Methods: We have surveyed institutions in which rely on MRI guidance for APBI treatments through a comprehensive literature review and onsite visits. These visits included our radiation therapists, physicists, and radiation oncologists. There are two main components in implementing a new radiotherapy program: 1) patient simulation and immobilization techniques, and 2) treatment planning and delivery techniques. These techniques were studied and modified for our clinical workflow. Results: Literature review revealed a single institution report addressing their MRgAPBI experience. For immobilization, we adopted two techniques observed during two site visits. However, these techniques were later modified to allow for patient customization. For small breasted patients, a foam tube placed along the sternum is used and for large breasted patients, we employed custom made acrylic brackets placed superior to the shoulders which are then built into the patient's immobilization cradle for individualized set up. For patient comfort, the contralateral arm is placed down along the patient’s side, whereas the ipsilateral arm is moved up away from the radiation field. This also allows for lateral clearance for beam selection. For treatment planning, consensus on optimal margin for expansion from the lumpectomy bed is of utmost importance. A single institution MRgAPBI report described a 1 cm expansion from the lumpectomy bed to PTV while NSABP B39 indicated a 2.5 cm expansion. This reduction in margin was justified by the utilization of real-time MRgRT. We have come to consensus to use a 1.5 cm expansion around the lumpectomy cavity, which accounts for the CTV expansion per NSABP B39 and to forgo the PTV expansion given the minimal setup uncertainty with online MRgRT. Our initial experience showed significant changes in the lumpectomy cavity shape and volume when the gap between patient simulation and the start of treatment exceeded a week. We have thus reduced the time between simulation and treatment to 1 week to address this issue. Conclusion: Implementing a new MRgAPBI program requires a comprehensive evaluation of available resources and published reports, as well as, team visits to expert MRgAPBI centers. Both immobilization techniques and lumpectomy margin expansions for treatment planning should be well thought of prior to starting a similar program. In our experience, time between simulation and treatment of 1 week showed minimal changes in the shape and size of the planning volume. Additional verification will need to be conducted within a prospective setting and larger cohort. In order to setup a successful program, teamwork is essential between all groups including radiation therapists, dosimetrists, physicists, and radiation oncologists. Citation Format: Eenas A. Omari, Jennifer Gambla, Hahn P. Mai, Muaiad Kittaneh, Tarita O. Thomas, Raymond B. Wynn, Anil Sethi, William Small Jr, Tamer Refaat. Implementing a real-time magnetic resonance imaging guided accelerated partial breast irradiation program [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-12-21.

Auto-segmentation of pancreatic tumor in multi-parametric MRI using deep convolutional neural networks

PurposeThe recently introduced MR-Linac enables MRI-guided Online Adaptive Radiation Therapy (MRgOART) of pancreatic cancer, for which fast and accurate segmentation of the gross tumor volume (GTV) is essential. This work aims to develop an algorithm allowing automatic segmentation of the pancreatic GTV based on multi-parametric MRI using deep neural networks. MethodsWe employed a square-window based convolutional neural network (CNN) architecture with three convolutional layer blocks. The model was trained using about 245,000 normal and 230,000 tumor patches extracted from 37 DCE MRI sets acquired in 27 patients with data augmentation. These images were bias corrected, intensity standardized, and resampled to a fixed voxel size of 1 × 1 × 3 mm3. The trained model was tested on 19 DCE MRI sets from another 13 patients, and the model-generated GTVs were compared with the manually segmented GTVs by experienced radiologist and radiation oncologists based on Dice Similarity Coefficient (DSC), Hausdorff Distance (HD), and Mean Surface Distance (MSD). ResultsThe mean values and standard deviations of the performance metrics on the test set were DSC = 0.73 ± 0.09, HD = 8.11 ± 4.09 mm, and MSD = 1.82 ± 0.84 mm. The interobserver variations were estimated to be DSC = 0.71 ± 0.08, HD = 7.36 ± 2.72 mm, and MSD = 1.78 ± 0.66 mm, which had no significant difference with model performance at p values of 0.6, 0.52, and 0.88, respectively. ConclusionWe developed a CNN-based model for auto-segmentation of pancreatic GTV in multi-parametric MRI. Model performance was comparable to expert radiation oncologists. This model provides a framework to incorporate multimodality images and daily MRI for GTV auto-segmentation in MRgOART.