PURPOSE:We evaluated the feasibility of a magnetic resonance (MR)-only simulation, planning, and treatment (MROSPT) workflow for prostate cancer patients using synthetic computed tomography (sCT) generated from magnetic resonance imaging (MRI) data. By validating sCT-based dose calculations, we aimed to streamline radiotherapy workflows, eliminate the need for CT simulation, and enable reliable clinical implementation of MR-based radiotherapy for MR-linac (MRL). METHODS:We developed a comprehensive workflow encompassing the entire process from initial consultation to treatment delivery. After developing the workflow, a retrospective dosimetric validation study was performed on nine men with prostate cancer. They underwent CT and MRI simulations, and sCTs were generated from the MRI data. Contours and intensity-modulated radiation therapy treatment plans were created on the reference simulation CT (rCT) and transferred to sCTs for dose-calculation comparisons. Dosimetric accuracy was evaluated using gamma analysis (dose/distance; 2%/2mm). Bulk density sCTs (bCTs) were created by overriding organ density values with their mean (bulk) sCT-determined densities. bCT based on sCT allows treatment planning directly on MRI for MRL workflow efficiency. RESULTS:Minimal non-bone Hounsfield units (HU)-value differences between rCT and sCT (5.5 ± 2.9 HU for prostate) demonstrated the reliability of the sCT generation process. Dosimetric comparisons between treatment plans (rCT vs. sCT, rCT vs. bCT) showed agreement within ± 2% in gamma analysis, confirming robust accuracy. The gamma index pass rate for rCT versus sCT and rCT versus bCT were consistently > 95% using 2%/2 mm criteria. A dry run of the entire simulation-to-treatment workflow was successfully completed. CONCLUSION:The MROSPT workflow using sCT is clinically feasible and dosimetrically accurate for prostate cancer patients. Dose calculations based on sCT demonstrated high dosimetric agreement with simulation CT, with no statistically significant differences across all evaluated metrics. These findings support the adoption of sCT‑based planning for prostate cancer radiotherapy and suggest its potential applicability in other anatomical regions especially in the pelvis. Integration of robust quality‑assurance processes and treatment‑delivery flexibility will further enhance its clinical utility.

Background and purpose: Quantitative T1 mapping is a major building block in several multiparametric magnetic resonance imaging (MRI) protocols intended for adaptive radiation therapy. The implementation of these protocols is challenging in anatomical sites that experience large physiological motion. The purpose of this study was to implement and validate motion-resolved quantitative T1 mapping on a 1.5 T MRI linear accelerator (MR-Linac) combining non-Cartesian k-space sampling trajectories with compressed sensing (CS) reconstruction. Materials and methods: Four 3-dimensional non-Cartesian k-space trajectories were evaluated: kooshball and stack-of-stars sampling using half-and full-spoke coverage. A variable flip angle acquisition was performed using the spoiled gradient-echo sequence. Gradient delay timing was optimized to minimize trajectory-induced artifacts. Eight CS reconstruction strategies were tested using spatial/spatiotemporal regularization operators. Reconstructions were evaluated and sorted by spatial resolution, bias, and variability. Motion-resolved T1 mapping was validated using two standard phantoms, one healthy volunteer, and one kidney cancer patient using respiratory self-gating and phase-sorted reconstruction. Results: All non-Cartesian T1 maps demonstrated high repeatability and low longitudinal bias in phantom studies, with coefficients of variation below 3.3%. Spatiotemporal regularization preserved spatial resolution and quantitative accuracy at undersampling factors up to 20-fold. In human subjects, non-Cartesian T1 mapping provided improved accuracy and reduced variability in mobile abdominal tissues compared to Cartesian acquisitions. Conclusions: Quantitative T1 mapping using non-Cartesian trajectories and CS reconstruction is feasible on a 1.5 T MR-Linac. The proposed approach enables accurate motion-resolved quantitative imaging within clinically practical acquisition times, establishing a foundation for multiparametric MRI in adaptive radiotherapy.

BACKGROUND AND PURPOSE:MR-Linac (MRL) is well-suited for treatment of oligometastatic disease (OMD) in muscle due to its superior soft tissue contrast. The purpose of this study was to assess dosimetric differences and quantify margins when CT-guidance (CTgRT) is used for OMD treatment compared to MR-guidance (MRgRT) on the MRL. MATERIALS AND METHODS:Five patients with intramuscular oligometastasis were treated with the MRL using daily MRgRT. To simulate CT-based alignment, an in-house deep learning model was used to generate synthetic-CT images from MR images for each fraction. The synthetic-CTs were independently aligned to the corresponding simulation-CTs by seven physicists using rigid registration. The difference between CTgRT and MRgRT alignment was calculated to quantify the inter-fractional image-guidance uncertainty and estimate PTV margins. Dose from the clinical beam sets was recalculated on the synthetic-CTs for comparative analysis. RESULTS:The CTgRT alignment compared to MRgRT were -0.59 ± 3.49 mm, 2.04 ± 3.96 mm, and 1.23 ± 1.20 mm in left-right (LR), superior-inferior (SI), and anterior-posterior (AP) directions, respectively. Compared to clinical plans, the dosimetric parameters of recalculated synthetic-CT plans, V95 and D95, were significantly lower, while the homogeneity index was significantly higher (p < 0.001 for each metric). Compared to MRgRT, CTgRT required additional treatment margins of 10.0 mm in LR, 11.5 mm in SI, and 4.6 mm in AP directions. CONCLUSIONS:We quantified the dosimetric difference in treating intramuscular metastases using MRgRT versus CTgRT and showed that treatment margin can be reduced by at least 5 mm in each direction with MRgRT.

Conventional radiotherapy dose calculation algorithms are often computationally slow and non-differentiable, creating bottlenecks for online adaptive radiotherapy (ART) and limiting end-to-end automatic planning. Deep learning provides consistent inference performance and a differentiable framework essential for rapid optimization. In this study, we developed a generalized, site-independent deep learning dose engine using a beamlet-based input strategy. This establishes a computationally consistent and differentiable module that enables end-to-end training for autoplanning while maintaining accuracy across diverse geometries. A dataset of 3,600 plans from 120 patients across six anatomical sites was used to train two 3D convolutional neural networks, a standard U-Net and a Cascade U-Net, to predict 3D dose distributions from CT images and divergent MLC/jaw projections. Performance was validated via 3D gamma analysis on an independent cohort of 60 VMAT plans. The optimal model (U-Net with MAE loss) achieved a mean gamma passing rate of 98.9 ± 1.6% (3

Compression belts (CBs) are sometimes used to reduce respiratory motion during stereotactic body radiotherapy of abdominal cancers with magnetic resonance (MR)-guided online adaptive planning. This study evaluated the dosimetric effects of overriding the relative electron density (ED) value of CBs in creating synthetic computed tomography (CT) scans for MR-guided adaptive planning. We evaluated plans for 12 patients with abdominal cancer and identified that ED values between 0.2 and 0.3 achieved the best approximation of CB ED in dose calculation. Our study presented an approach to estimate appropriate ED overrides for CBs in MR-guided online adaptive planning

Background and purpose:Pancreatic gross tumor volume (GTV) delineation is challenging due to their variable morphology and uncertain ground truth. Previous deep learning-based auto-segmentation methods have struggled to handle tasks with uncertain ground truth and have not accommodated stylistic customizations. We aim to develop a human-in-the-loop pancreatic GTV segmentation tool using Tversky ensembles by leveraging uncertainty estimation techniques. Material and methods:In this study, we utilized a total of 282 patients from the pancreas task of the Medical Segmentation Decathlon. Thirty patients were randomly selected to form an independent test set, while the remaining 252 patients were divided into an 80-20 % training-validation split. We incorporated Tversky loss layer during training to train a five-member segmentation ensemble with varying contouring tendencies. The Tversky ensemble predicted probability maps by estimating pixel-level segmentation uncertainties. Probability thresholding was employed on the resulting probability maps to generate the final contours, from which eleven contours were extracted for quantitative evaluation against ground truths, with variations in the threshold values. Results:Our Tversky ensemble achieved DSC of 0.47, HD95 of 12.70 mm and MSD of 3.24 mm respectively using the optimal thresholding configuration. We outperformed the Swin-UNETR configuration that achieved the state-of-the-art result in the pancreas task of the medical segmentation decathlon. Conclusions:Our study demonstrated the effectiveness of employing an ensemble-based uncertainty estimation technique for pancreatic tumor segmentation. The approach provided clinicians with a consensus probability map that could be fine-tuned in line with their preferences, generating contours with greater confidence.

Background:Quantitative mapping of the longitudinal relaxation rate (R1=1/T1) is a major building block for several multiparametric MRI protocols intended for adaptive radiation therapy planning. The implementation of these protocols is challenging in anatomical sites that experience large physiological motion. Purpose:To implement and validate a motion-resolved quantitative T1 mapping method on a 1.5T MR-Linac that combines non-Cartesian k-space sampling trajectories with compressed sensing (CS) reconstruction techniques. Methods:Four 3D non-Cartesian k-space trajectories were evaluated: radial and stack-of-stars sampling using half- and full-spoke coverage. A variable flip angle acquisition was performed using the spoiled gradient-echo sequence, and T1 mapping was validated using two standard phantoms. Gradient delay timing was optimized empirically to minimize trajectory-induced artifacts. Eight compressed sensing reconstruction strategies were tested using spatial and spatiotemporal regularization operators. Reconstructions were evaluated across multiple implementation parameters and ranked based on spatial resolution, bias, and variability. In vivo studies included one healthy volunteer and one patient undergoing radiotherapy to a target in the kidney. Motion-resolved imaging was performed using respiratory self-gating and phase-sorted reconstruction. Results:All non-Cartesian trajectories demonstrated high repeatability and low longitudinal bias in phantom studies, with coefficients of variation below 3.3%. Radial half-spoke sampling achieved the shortest scan times and highest agreement with Cartesian benchmarks. Reconstruction methods incorporating spatiotemporal regularization maintained spatial resolution and quantitative accuracy across undersampling factors up to 20-fold. In human subjects, non-Cartesian T1 mapping provided improved accuracy and reduced variability in mobile abdominal tissues compared to Cartesian acquisitions, particularly in the kidney cortex and medulla, where motion artifacts led to overestimation and higher variance in the reference method. Conclusions:T1 mapping using non-Cartesian trajectories and compressed sensing reconstruction is feasible on a 1.5T MR-Linac. The proposed approach enables accurate, motion-resolved quantitative imaging within clinically practical acquisition times. These results support integration of quantitative T1 mapping into adaptive MR-guided radiotherapy workflows and establish a foundation for future development of multiparametric imaging and response-adaptive treatment strategies.

Purpose This study aims to evaluate organs-at-risk (OARs) segmentation variability across 8 commercial artificial intelligence (AI)-based segmentation software using independent multi-institutional data sets, and to provide recommendations for clinical practices using AI-segmentation. Methods and Materials A total of 160 planning computed tomography image sets from 4 anatomic sites: head and neck, thorax, abdomen, and pelvis were retrospectively pooled from 3 institutions. Contours for 31 OARs generated by the software were compared to clinical contours using multiple accuracy metrics, including: dice similarity coefficient (DSC), 95 percentile of Hausdorff distance, surface DSC, as well as relative added path length as an efficiency metric. A 2-factor analysis of variance was used to quantify variability in contouring accuracy across software platforms (intersoftware) and patients (interpatient). Pairwise comparisons were performed to categorize the software into different performance groups, and intersoftware variations were calculated as the average performance differences between the groups. Results Significant intersoftware and interpatient contouring accuracy variations (P < .05) were observed for most OARs. The largest intersoftware variations in DSC in each anatomic region were cervical esophagus (0.41), trachea (0.10), spinal cord (0.13), and prostate (0.17). Among the organs evaluated, 7 had mean DSC >0.9 (ie, heart, liver), 15 had DSC ranging from 0.7 to 0.89 (ie, parotid, esophagus). The remaining organs (ie, optic nerves, seminal vesicle) had DSC<0.7. Of the 31 organs, 16 (52%) had relative added path length less than 0.1. Conclusions Our results reveal significant intersoftware and interpatient variability in the performance of AI-segmentation software. These findings highlight the need of thorough software commissioning, testing, and quality assurance across disease sites, patient-specific anatomies, and image acquisition protocols.

Purpose: Accurate cardiac chamber segmentation is crucial for improving cardiac sparing in magnetic resonance (MR)-guided adaptive radiation therapy, especially in patients at risk for radiation-induced cardiotoxicity. Here, we developed and evaluated automatic segmentation models for cardiac chambers that use daily MR images acquired on a 1.5-T MR-Linac system. Methods and Materials: Twenty healthy volunteers underwent daily MR scanning on a 1.5-T MR-Linac, with 2 radial sequences: T2/T1 3DVaneXD balanced fast field echo with spectral attenuated inversion recovery (bFFE-SPAIR) and T1 3DVaneXD mDixon. Three flip angles were tested for each sequence to determine optimal image quality for chamber segmentation. Full-resolution 3D nnU-Net models were trained for the following: (1) bFFE-SPAIR (bFFE model); (2) T1 mDixon (mDixon model); and (3) both sequences (hybrid model). Models were evaluated based on Dice similarity coefficient (DSC) and mean surface distance against manual contours. Clinical acceptance of the automatic segmentation was assessed with a 5-point Likert scale. An in-silico planning study was performed to assess cardiac chamber sparing during plan adaptation. Results: The average contrast-to-noise ratios in bFFE-SPAIR were 8.7 (20°), 34.2 (50°), and 37.3 (80°); for T1 mDixon, these values were 3.6 (5°), 5.9 (10°), and 4.9 (20°). The bFFE model achieved the highest segmentation performance (average DSC 0.85 ± 0.05 and mean surface distance 2.2 ± 0.6 mm). The T1 mDixon sequence, despite lower contrast-to-noise ratios, provided similar segmentation accuracy (DSC 0.83 ± 0.06). A hybrid model combining both sequences showed no significant improvement over the bFFE model. Clinical evaluation indicated that 95% of the autosegmented contours from the bFFE model were acceptable for clinical use (score ≥4). Adaptive plan greatly reduced individual cardiac chamber dose while maintaining similar target coverage. Conclusions: This study demonstrated the feasibility of using bFFE-SPAIR and T1 mDixon sequences to accurately segment cardiac chambers on a 1.5-T MR-Linac. These models offer potential for improved cardiac sparing in MR-guided adaptive radiation therapy.

BACKGROUND/OBJECTIVES:Stereotactic body radiation therapy (SBRT) for skull base reirradiation is particularly challenging, as patients have already received substantial radiation doses to the region, and nearby normal organs may have approached their tolerance limit from prior treatments. In this study, we reviewed the characteristics and capabilities of four advanced external beam radiation delivery systems and four modern treatment planning systems and evaluated the treatment plan quality of each technique using skull base reirradiation patient cases. METHODS:SBRT plans were generated for sixteen skull base reirradiation patients using four modalities: the GK plan for the Elekta Leksell Gamma Knife Perfexion/ICON, the CyberKnife (CK) plan for the Accuray CyberKnife, the intensity-modulated proton therapy (IMPT) plan for the Hitachi ProBeat-FR proton therapy machine, and the volumetric-modulated arc therapy (VMAT) plan for the Varian TrueBeam STx. These plans were evaluated and compared using two novel gradient indices in addition to traditional dosimetry metrics for targets and organs at risk (OARs). The steepest border gradient quantified the percent prescription dose fall-off per millimeter at the boundary between the target and adjacent critical structures. This gradient index highlighted the system's ability to spare nearby critical OARs. The volume gradient assessed the extent of dose spread outside the target toward the patient's body. RESULTS:All plans achieved comparable target coverage and conformity, while IMPT and VMAT demonstrated significantly better uniformity. The GK plans exhibited the highest border gradient, up to 20.9%/mm, followed by small-spot-size IMPT plans and CK plans. Additionally, IMPT plans showed the benefit of reduced dose spread in low-dose regions and the lowest maximum and mean doses to the brainstem and carotid artery. CONCLUSIONS:The advanced external beam radiotherapy modalities evaluated in this study are well-suited for SBRT in skull base reirradiation, which demands precise targeting of tumors with highly conformal doses and steep dose gradients to protect nearby normal structures.

PURPOSE:Radiation-induced heart damage is a significant concern in the treatment of non-small cell lung cancer (NSCLC) that can have debilitating or life-threatening consequences. Current strategies focus on minimizing heart exposure, but individual susceptibility varies. Existing evidence also suggests that a uniform "one-size-fits-all" dosimetric constraint for the heart may not be optimal for all patients. METHODS AND MATERIALS:We developed a prospective study using Bayesian continuous learning and adaptation to develop a framework for personalized adaptive radiation treatment (PART) to reduce cardiovascular adverse events among patients with locally advanced NSCLC. The trial includes a Bayesian personalized risk prediction model to guide heart dose constraints; sequential learning to refine the model and the PART; continuous adaptation of the target risk level; and go/no-go monitoring of PART effectiveness in clinical implementation. Elevation of high-sensitivity cardiac troponin T (hs-cTnT) after radiation was used as a surrogate biomarker for grade ≥2 cardiovascular adverse events to allow real-time decision-making. RESULTS:As of July 31, 2025, 100 patients have been enrolled and completed radiation treatment. Standard radiation plans were implemented for cohort 1 (50 patients), and PART for cohort 2 (50 patients). The first model incorporated patient- and disease-related factors and mean heart dose (MHD) as risk factors. The average treated MHDs were 7.84 ± 6.30 Gy in cohort 1 and 6.36 ± 6.01 Gy in cohort 2 (P = .20). The incidence of hs-cTnT elevation was 20.5% in cohort 2 compared with 31.9% in cohort 1. Within cohort 2, patients who satisfied the PART dose constraint had a markedly lower incidence of hs-cTnT elevation (9.7%) compared with those who exceeded the PART dose constraint (46.2%, P = .012). CONCLUSIONS:Clinical implementation of PART model to guide treatment decision within a prospective trial is feasible. The recommended MHD constraints generated by the first version of PART appear reasonable and clinically relevant. PART was associated with lower incidence of hs-cTnT elevation.

This research used domestic satellite remote sensing images from 2020, 2021, and 2022 to conduct indoor interpretation of the mining geological environment in a certain city in Northeast China. The distribution location, mining types, land area, and types of dynamic changes in the mining geological environment in 2021 and 2022 were investigated, and the changes over the past two years were analyzed. This paper analyzes the homogeneity and difference of mining geological environment changes in the research area in recent years from multiple perspectives, which provides technical support for the future management of mining geological environment and related research in Northeast China., and at the same time fully proves the ability of domestic satellite remote sensing data to be applied in the field of mine monitoring.

Background: Stereotactic body radiation therapy (SBRT) has proven effective in controlling spinal lesions with minimal toxicity, primarily due to its ability to limit spinal cord dose. Recent advances in MR-linac (MRL) technology offer superior spinal cord visualization and real-time gating, which can facilitate dose escalation in spinal tumor treatment while maintaining safety. Purpose: This study aimed to optimize motion management for spine SBRT on an MRL by analyzing patient-specific motion dynamics and evaluating the most effective registration structures. We hypothesized that baseline shifts (BLS) would improve delivery efficiency while maintaining spinal cord dose constraints. The goal was to establish displacement thresholds and assess the role of baseline shift correction adaptative planning in improving treatment delivery efficiency. Methods: Twelve patients underwent two MRI sessions on the MRL. The optimal registration structure was identified, and intrafraction motion was assessed to calculate delivery efficiency. Baseline shift (BLS) simulations were applied for five cases that showed significant motion and suboptimal delivery efficiency, and the dosimetric impact of the BLS was evaluated. The simulated BLS-based plan adaptation was implemented via a segment aperture morphing adapt-to-position workflow. Results: The most stable registration structure was the spinal canal plus three adjacent vertebrae. Cine imaging revealed average intrafraction motion (95th to 5th percentiles) of 0.8 ± 0.5 mm in the right-left (RL) direction, 0.9 ± 0.6 mm in the anterior–posterior (AP) direction, and 0.7 ± 0.5 mm in the SI direction. Simulated BLS improved delivery efficiency to >80% in all but one case, with a ±1 mm displacement threshold tolerance. While target coverage remained consistent after BLS simulation, the spinal cord dose increased by 7–60%, exceeding the 14 Gy constraint in three of the five simulated cases. Conclusions: Cine imaging and BLS can enhance delivery efficiency in spine SBRT but may increase spinal cord dose. These findings underscore the need for careful patient selection, advanced motion management, and patient-specific BLS protocols.

Introduction To evaluate the feasibility of Monte Carlo (MC)-based patient-specific quality assurance (PSQA) for MR-guided online adaptive radiotherapy and to explore the potential to eliminate the post-delivery measurement-based PSQA. Material and methods A total of 113 cases from two institutions, treated on MR-Linac machines, were included in the study. A customized GPU-accelerated, Monte Carlo-based secondary dose verification software (ART2Dose) was developed and integrated into the QA workflow, accounting for a 1.5 Tesla magnetic field. PSQA included ArcCheck (AC) delivery QA and online MC calculation-based QA. Reference plans underwent offline validation with AC and MC, while adapt-to-shape (ATS) plans were processed through MC and post-delivery QA. Gamma pass rates (GPR) with 3 %/2mm criteria were compared statistically across methods. Radcalc was applied to compare point dose difference with MC. Results MC QA achieved GPRs of 97.5 % ± 2.0 % and 97.1 % ± 2.9 % for reference and ATS plans, comparable to AC QA (97.6 % ± 2.0 % and 96.9 % ± 3.0 %). Wilcoxon signed-rank test showed statistically significant differences between reference and ATS plan QA (p < 0.05), but a Pearson correlation coefficient of 0.76 confirmed a linear relationship for MC GPR. Lung cases exhibited lower GPRs with MC compared to AC QA. MC QA demonstrated supaireerior point dose agreement with TPS (1.7 % ± 1.2 %) compared to RadCalc (4.1 % ± 1.7 %). No significant differences were observed between institutions. Conclusion MC-based QA is a robust tool for adaptive QA workflows in 1.5-T MR-Linac systems. It enhances efficiency and potentially supports the elimination of post-delivery measurement-based QA for adaptive plans.