Publications

View details for DOI 10.1227/neuprac.0000000000000169

View details for Web of Science ID 001577502000001

Abstract

Occipital condyle metastasis (OCM) is a rare condition characterized by severe occipital pain and neurological symptoms due to lower cranial nerve (CN) deficits, stemming from its anatomic location. Despite the widespread use of stereotactic radiosurgery (SRS) for cranial metastases, its specific impact on OCM remains underexplored. This study evaluates the efficacy and safety of CyberKnife SRS in treating OCM, focusing on occipital pain, dysfunctions of lower CNs, and local tumor control.We retrospectively analyzed cases of OCM treated with SRS at our institute from 2012 to 2023, evaluating patient demographics, presenting symptoms, treatment parameters, and outcomes.Eighteen patients (10 females) with a mean age of 64 years (SD: 10.4) were treated. Common presentations included occipital pain (44.4%) and lower CN deficits (27.8%). The median target volume was 6.95 cc (IQR: 4.64-21.2). The mean single-fraction equivalent dose was 18.7 Gy10 (SD: 1.9). Ten tumors received 15-20 Gy in 1 fraction (50%), 2 tumors received 20-28 Gy in 2 fractions (10%), 4 tumors received 27 Gy in 3 fractions (20%), and 4 tumors received 30-40 Gy in 5 fractions (20%). Based on Kaplan-Meier estimate, SRS achieved 93.8% local tumor control rate over 3 years, with a median overall survival of 13 months (95% CI: 0-32.2). Among patients presenting with symptoms, 87.5% reported occipital pain relief (P = .04), and 80% observed improvements in CN function (P = .003). Four patients experienced local recurrence.CyberKnife SRS is a promising treatment of OCM, offering significant pain relief and improvement in neurological symptoms, along with favorable local control rates. This noninvasive therapy provides a valuable alternative to surgery, potentially enhancing the quality of life for patients with limited treatment options due to this challenging condition.

View details for DOI 10.1227/neuprac.0000000000000169

View details for PubMedID 41163734

View details for PubMedCentralID PMC12560716

Abstract

The GammaPod™ (GP) system, a contemporary platform dedicated to breast cancer (BC) radiotherapy, facilitates the delivery of accelerated partial breast irradiation (APBI) via the Co-60 prone-based stereotactic partial breast irradiation (CP-sPBI) technique. The precise CP-sPBI configuration permits reduced planning target volume (PTV) margins compared to other APBI techniques, creating an increased separation between PTV and organs at risk (OARs). This study explores the variability of heart-to-PTV distance and its effects on cardiac dosimetry.An APBI database of 102 consecutive patients treated with CP-sPBI between March 2019 and February 2023 was queried for retrospective analysis. Statistical analyses were performed to evaluate the mean and maximum (max) heart and left anterior descending artery (LAD) doses based on two parameters: 1) D-H, the minimum distance between the heart and the lumpectomy cavity PTV, and 2) D-LAD, the minimum distance between the LAD and the lumpectomy cavity PTV. The median values of D-H and D-LAD, measured on either axial or sagittal planes, were employed to categorize patients based on cardiac dose levels.The analysis revealed a statistically significant difference in the mean and max heart dose between patients with left-sided and right-sided breast cancer. Specifically, in left-sided breast cancer patients, median D-H and D-LAD cutoffs were identified as 2.67 and 3.22 cm, respectively. Patients with D-H less than 2.67 cm exhibited significantly higher mean (1.77 vs. 0.75 Gy; p < 0.01) and max heart doses (15.21 vs. 4.38 Gy; p < 0.01) compared to those with D-H greater than or equal to 2.67 cm. Similarly, lower D-LAD values (<3.22 cm) demonstrated a statistically significant association with increased arterial dose compared to higher D-LAD values (≥3.22 cm).Leveraging its sharp dose fall-off characteristic, the GP treatment delivery system facilitates the delivery of five-fraction APBI while maintaining acceptable cardiac dosimetry parameters. This is particularly advantageous for tumors situated further from the heart because heart doses dissipate with distance. The estimates of heart dose based on the distance to the heart and LAD from PTV have the potential to serve as a valuable tool for clinicians, aiding in more refined risk evaluation and patient selection for CP-sPBI.

View details for DOI 10.3389/fonc.2025.1458111

View details for PubMedID 40978062

View details for PubMedCentralID PMC12443537

Abstract

Spinal metastases (SM) with epidural spinal cord compression (ESCC) present a significant challenge because of the high risk of radiation-induced injury to critical structures such as the spinal cord and nerve roots. Traditional treatment approaches often avoid circumferential stereotactic body radiotherapy (SBRT) to reduce these risks. The efficacy and safety of donut-shaped circumferential SBRT, designed to target the spinal column while sparing the spinal cord, remains underexplored. The aim of this study was to evaluate the safety and efficacy of donut-shaped circumferential CyberKnife SBRT for SM, particularly in preventing radiation-induced myelopathy and achieving local tumor control (LTC).We retrospectively analyzed data from patients treated with donut-shaped circumferential SBRT between 2014 and 2023. Key parameters examined included patient demographics, ESCC grade (Bilsky), prior treatments, clinical symptoms, and treatment parameters. We focused on SBRT dosimetric data, radiation exposure to the spinal cord and cauda equina, adherence to dose-volume constraints, and post-SBRT outcomes, including myelopathy and LTC.Forty-eight lesions in 43 patients (median age: 65; range: 20-78) were reviewed. One patient required separation surgery for severe ESCC (Bilsky grade 3). The median clinical target volume was 63.77 cm3, and the median margin dose was 24 Gy. Over a median follow-up of 8 months, LTC was 91.1% at 6 months, 87.1% at 1 year, 82.8% at 3 years, and 62.1% at 5 years. The median overall survival was 17 months. Of the 21 lesions exceeding dose constraints, only one patient exhibited clinical myelopathy, which correlated with local tumor recurrence. No radiographic myelopathy or other radiation-induced complications were observed.Donut-shaped circumferential CyberKnife SBRT is a safe and effective treatment of SM, achieving high LTC with minimal radiation-induced complications, including myelopathy.

View details for DOI 10.1227/neu.0000000000003446

View details for PubMedID 40243341

Abstract

Radiotherapy requires precise, patient-specific treatment planning to achieve high-quality dose distributions that improve patient outcomes. Traditional manual planning is time-consuming and clinically impractical for performing necessary plan trade-off comparisons, including treatment modality selection, prescription dose settings, and organ at risk (OAR) constraints. A time-efficient dose prediction tool could accelerate the planning process by guiding clinical plan optimization and adjustments. While the deep convolutional neural networks (CNNs) are prominent in radiotherapy dose prediction tasks, most studies have attempted to customize network architectures for different diseases and treatment modalities.This study proposes a universal and efficient strategy, Seg2Dose, leveraging a state-of-the-art segmentation network for radiotherapy dose prediction without the need for model architecture modifications. We aim to provide a convenient off-the-shelf dose prediction tool that simplifies the dose prediction process, enhancing planning speed, and plan quality while minimizing the need for extensive coding and customization.The proposed Seg2Dose consists of three modules: the Adapter, the segmentation network, and the Smoother. Prior to model training, the Adapter processes dose distributions into dose level map with an adjustable interval, which serves as the ground truth of the segmentation network, and generates two input channels: weighted avoidance image and normalized prescribed dose image. The segmentation network predicts dose levels from input channels using the nnU-Net, which was trained, validated and tested on 304, 77, and 64 breast cancer GammaPod treatment plans from 90 patients. The Smoother converts the predicted dose levels into continuous dose distribution with a Gaussian filter. The performance of Seg2Dose models with two different dose level intervals, 2% (Seg2Dose 2%) and 5% (Seg2Dose 5%), was evaluated by the Dice similarity coefficients (DSCs), voxel-based mean absolute percent error (MAPE), dose-volume histogram (DVH) metrics, global 3%/2 mm and 3%/1 mm gamma passing rate (GPR), and a case study including normal and worst cases. Additionally, Seg2Dose was compared with an exciting cutting-edge Cascade 3D (C3D) dose prediction model, which was trained on continuous dose distributions, to investigate the impact of using dose level map.For dose level prediction, Seg2Dose achieved average DSCs of 0.94 and 0.93 for the 2% and 5% intervals, respectively. For dose distribution prediction, both Seg2Dose 2% and Seg2Dose 5% achieved MAPEs within 6% for targets and most OARs, with the exception of the skin, which had the highest MAPE at 8.58% for Seg2Dose 2% and 15.25% for Seg2Dose 5%. The DVH metrics showed consistent findings. The C3D model has a better performance in GPR than Seg2Dose models. However, the C3D model exhibited higher MAPEs in target areas with lower dose predictions. In the case study, Seg2Dose 2% and C3D predictions were more consistent with clinical plans, showing smaller dose differences compared to Seg2Dose 5%.Our study confirms the feasibility of leveraging the segmentation network for dose prediction and provides an efficient and off-the-shelf approach for dose prediction without requiring extensive coding efforts. This plug-in tool holds promise for quick dose planning, potentially aiding in the identification of optimal radiotherapy techniques and dosimetric tradeoffs prior to tedious treatment planning.

View details for DOI 10.1002/mp.17711

View details for PubMedID 40017352

Abstract

For patients with spine metastases, stereotactic radiosurgery (SRS) provides excellent local control and pain response. Despite increasing use of this treatment modality, there is no consensus on the optimal dose and fractionation of spine SRS for efficacy and toxicity. We have initiated a single-center phase III randomized trial that compares two dose regimens with similar biological equivalent dose (BED) to determine the isolated effect of SRS fractionation on local control.Patients with one to three cervical, thoracic, or lumbar spine metastases spanning no more than two contiguous vertebral levels in need of radiation will be eligible for enrollment. Patients will be assigned 1:1 to receive either 22 Gy in 1 fraction or 28 Gy in 2 fractions. Biased coin randomization will be used to randomly assign patients while balancing the following stratifying variables between the two treatment arms at baseline: gastrointestinal histology (yes/no), paraspinal tissue extension (yes/no), epidural compression (low-/high-grade), and number of sites treated (one to three). The primary endpoint is one-year local control, defined per Spine Response Assessment in Neuro-Oncology (SPINO) criteria. The secondary endpoints include patient-reported health-related quality of life (HRQOL), pain associated with the treated site, vertebral compression fracture (VCF), and two-year local control. Patients will be followed for these outcomes at one to two weeks, one month, three months, and six months after treatment, and every six months thereafter until 24 months after treatment. While on the study, patients will receive routine co-interventions as clinically indicated.The studies published thus far comparing the single- and multi-fraction SRS are lacking long-term local control outcomes and are limited by selection bias as well as single-fraction arms with higher BED, which is correlated with improved local control. Our study will isolate the effect of fractionation by comparing one-year local control in patients treated with single- and multi-fraction SRS with equivalent BED. We anticipate that the results of this, as well as secondary endpoints such as pain response, adverse effects, and quality of life will provide much-needed guidance regarding optimal dose and fractionation for both maximizing local control and minimizing toxicity.NCT#06173401. Approved by Stanford Scientific Review Committee (study ID: BRN0060) on 9/12/2023 and Stanford Institutional Review Board (study ID: IRB-72248) on 11/14/2023.

View details for DOI 10.1186/s12885-025-13655-6

View details for PubMedID 39984889

View details for PubMedCentralID PMC11846292

Abstract

Current standard of care treatment for patients with spine tumors includes multidisciplinary approaches, including the following: (1) surgical tumor debulking, epidural spinal cord decompression, and spine stabilization techniques; (2) systemic chemo/targeted therapies; (3) radiation therapy; and (4) surveillance imaging for local disease control and recurrence. Titanium pedicle screw and rod fixation have become commonplace in the spine surgeon's armamentarium for the stabilization of the spine following tumor resection and separation surgery. However, the high degree of imaging artifacts seen with titanium implants on postoperative CT and MRI scans can significantly hinder the accurate delineation of vertebral anatomy and adjacent neurovascular structures to allow for the safe and effective planning of downstream radiation therapies and detection of disease recurrence. Carbon fiber-reinforced polyetheretherketone (CFR-PEEK) spine implants have emerged as a promising alternative to titanium due to the lack of artifact signals on CT and MRI, allowing for more accurate and safe postoperative radiation planning. In this article, we review the tenants of the surgical and radiation management of spine tumors and discuss the safety, efficacy, and current limitations of CFR-PEEK spine implants in the multidisciplinary management of spine oncology patients.

View details for DOI 10.3390/brainsci15020199

View details for PubMedID 40002531

View details for PubMedCentralID PMC11852773

Abstract

Dosimetric commissioning and quality assurance (QA) for linear accelerators (LINACs) present a significant challenge for clinical physicists due to the high measurement workload and stringent precision standards. This challenge is exacerbated for radiosurgery LINACs because of increased measurement uncertainty and more demanding setup accuracy for small-field beams. Optimizing physicists' effort during beam measurements while ensuring the quality of the measured data is crucial for clinical efficiency and patient safety.To develop a radiosurgery LINAC beam model that embeds prior knowledge of beam data through implicit neural representation (NeRP) learning and to evaluate the model's effectiveness in guiding beam data sampling, predicting complete beam dataset from sparse samples, and verifying detector choice and setup during commissioning and QA.Beam data including lateral profile and tissue-phantom-ratio (TPR), collected from CyberKnife LINACs, were investigated. Multi-layer perceptron (MLP) neural networks were optimized to parameterize a continuous function of the beam data, implicitly defined by the mapping from measurement coordinates to measured dose values. Beam priors were embedded into network weights by first training the network to learn the NeRP of a vendor-provided reference dataset. The prior-embedded network was further fine-tuned with sparse clinical measurements and used to predict unacquired beam data. Prospective and retrospective evaluations of different beam data samples in finetuning the model were performed using the reference beam dataset and clinical testing datasets, respectively. Model prediction accuracy was evaluated over 10 clinical datasets collected from various LINACs with different manufacturing modes and collimation systems. Model sensitivity in detecting beam data acquisition errors including inaccurate detector positioning and inappropriate detector choice was evaluated using two additional datasets with intentionally introduced erroneous samples.Prospective and retrospective evaluations identified consistent beam data samples that are most effective in fine-tuning the model for complete beam data prediction. Despite of discrepancies between clinical beam and the reference beam, fine-tuning the model with sparse beam profile measured at a single depth or with beam TPR measured at a single collimator size predicted beam data that closely match ground truth water tank measurements. Across the 10 clinical beam datasets, the averaged mean absolute error (MAE) in percentage dose was lower than 0.5% and the averaged 1D Gamma passing rate (1%/0.5  mm for profile and 1%/1  mm for TPR) was higher than 99%. In contrast, the MAE and Gamma passing rates were above 1% and below 95% between the reference beam dataset and clinical beam datasets. Model sensitivity to beam data acquisition errors was demonstrated by significant model prediction changes when fine-tuned with erroneous versus correct beam data samples, as quantified by a Gamma passing rate as low as 18.16% between model predictions.A model for small-field radiosurgery beam was proposed that embeds prior knowledge of beam properties and predicts the entire beam data from sparse measurements. The model can serve as a valuable tool for clinical physicists to verify the accuracy of beam data acquisition and promises to improve commissioning and QA reliability and efficiency with substantially reduced number of beam measurements.

View details for DOI 10.1002/mp.17617

View details for PubMedID 39812551

Abstract

Accurate and automated early survival prediction is critical for patients with glioblastoma (GBM) as their poor prognosis requires timely treatment decision-making. To address this need, we developed a deep learning (DL)-based end-to-end workflow for GBM overall survival (OS) prediction using pre-resection basic structural multiparametric magnetic resonance images (Bas-mpMRI) with a multi-institutional public dataset and evaluated it with an independent dataset of patients on a prospective institutional clinical trial.The proposed end-to-end workflow includes a skull-stripping model, a GBM sub-region segmentation model and an ensemble learning-based OS prediction model. The segmentation model utilizes skull-stripped Bas-mpMRIs to segment three GBM sub-regions. The segmented GBM is fed into the contrastive learning-based OS prediction model to classify the patients into different survival groups. Our datasets include both a multi-institutional public dataset from Medical Image Computing and Computer Assisted Intervention (MICCAI) Brain Tumor Segmentation (BraTS) challenge 2020 with 235 patients, and an institutional dataset from a 5-fraction SRS clinical trial with 19 GBM patients. Each data entry consists of pre-operative Bas-mpMRIs, survival days and patient ages. Basic clinical characteristics are also available for SRS clinical trial data. The multi-institutional public dataset was used for workflow establishing (90% of data) and initial validation (10% of data). The validated workflow was then evaluated on the institutional clinical trial data.Our proposed OS prediction workflow achieved an area under the curve (AUC) of 0.86 on the public dataset and 0.72 on the institutional clinical trial dataset to classify patients into 2 OS classes as long-survivors (>12 months) and short-survivors (<12 months), despite the large variation in Bas-mpMRI protocols. In addition, as part of the intermediate results, the proposed workflow can also provide detailed GBM sub-regions auto-segmentation with a whole tumor Dice score of 0.91.Our study demonstrates the feasibility of employing this DL-based end-to-end workflow to predict the OS of patients with GBM using only the pre-resection Bas-mpMRIs. This DL-based workflow can be potentially applied to assist timely clinical decision-making.

View details for DOI 10.1016/j.compbiomed.2024.109436

View details for PubMedID 39637462

Abstract

Artificial intelligence (AI)-aided methods have made significant progress in the auto-delineation of normal tissues. However, these approaches struggle with the auto-contouring of radiotherapy target volume. Our goal is to model the delineation of target volume as a clinical decision-making problem, resolved by leveraging large language model-aided multimodal learning approaches.A vision-language model, termed Medformer, has been developed, employing the hierarchical vision transformer as its backbone, and incorporating large language models to extract text-rich features. The contextually embedded linguistic features are seamlessly integrated into visual features for language-aware visual encoding through the visual language attention module. Metrics, including Dice similarity coefficient (DSC), intersection over union (IOU), and 95th percentile Hausdorff distance (HD95), were used to quantitatively evaluate the performance of our model. The evaluation was conducted on an in-house prostate cancer dataset and a public oropharyngeal carcinoma (OPC) dataset, totaling 668 subjects.Our Medformer achieved a DSC of 0.81 ± 0.10 versus 0.72 ± 0.10, IOU of 0.73 ± 0.12 versus 0.65 ± 0.09, and HD95 of 9.86 ± 9.77 mm versus 19.13 ± 12.96 mm for delineation of gross tumor volume (GTV) on the prostate cancer dataset. Similarly, on the OPC dataset, it achieved a DSC of 0.77 ± 0.11 versus 0.72 ± 0.09, IOU of 0.70 ± 0.09 versus 0.65 ± 0.07, and HD95 of 7.52 ± 4.8 mm versus 13.63 ± 7.13 mm, representing significant improvements (p < 0.05). For delineating the clinical target volume (CTV), Medformer achieved a DSC of 0.91 ± 0.04, IOU of 0.85 ± 0.05, and HD95 of 2.98 ± 1.60 mm, comparable to other state-of-the-art algorithms.Auto-delineation of the treatment target based on multimodal learning outperforms conventional approaches that rely purely on visual features. Our method could be adopted into routine practice to rapidly contour CTV/GTV.

View details for DOI 10.1016/j.ijrobp.2024.07.2149

View details for PubMedID 39117164