Journal of Medical Physics最新文献

Background: Images from onboard electronic portal imaging devices (EPID) contain dose information that can be converted into dose maps. A cycle-consistent generative adversarial network (CycleGAN)-based model was developed for two-dimensional (2D) EPID dosimetry, and the dose characteristics of the model were evaluated carefully.

Materials and methods: All the measurements were done on a linac equipped with an EPID detector. This experiment involved: (1) assessing the dose characteristics of the EPID and (2) using a commercial treatment planning system to calculate dose distributions in a slab phantom, which were taken as the ground truth in CycleGAN-based models for converting the EPID images to 2D dose maps. There were about 780 beams delivered to EPID through a slab phantom. There were two normalization methods (NM): I: based on the highest possible value: 65,535 (16 bit); II: by its own maximum pixel value. To evaluate the model, gamma analyses between the ground truth and the output were performed with in-house software; and the dose linearity of the model was checked carefully. A comparative analysis was conducted to evaluate the outcomes stemming from two distinct NMs applied to the input data.

Results: The dose characteristics of the EPID demonstrated exceptional precision. Notably, the beam output factors exhibited considerable variations with the increasing thickness of the phantom. Specifically, when the phantom thickness surpassed 12 cm, the trend lines exhibited a pronounced linearity. Deep learning models efficiently transformed EPID images into planar dose maps, albeit exhibiting dose nonlinearity that could be mitigated by choose the suitable normalization medthods. The mean pass rates of gamma analyses (3 mm, 3%) of data normalized by the way I or II were 85.5%, 97.9%, respectively.

Conclusion: EPID is an excellent flat-panel detector that captures images rich in dose information, which can be effectively transformed into precise planar dose maps using CycleGAN-based models. The trained model could be used in the quality assurance of treatment plans.

In this study, we used Geant4 Monte Carlo simulations to explore how different two-layer combinations of moderator materials can improve the performance of accelerator-based boron neutron capture therapy (BNCT). We tested 16 pairings of aluminum oxide (Al₂O₃), titanium(III) fluoride, lithium bromide (LiBr), and lithium carbonate to see how each affected neutron and gamma radiation levels-both at the beam exit and within a simulated tumor. To evaluate their performance, we applied a weighted scoring system that considered both treatment effectiveness and patient safety. Our results showed that a dual-layer configuration of LiBr (Configuration N) delivered the highest thermal neutron dose to the tumor, making it the most effective for treatment. On the other hand, a double layer of Al₂O₃ (configuration P) excelled in minimizing harmful radiation outside the tumor area. Some setups, like LiBr + Al₂O₃ (configuration G), struck a good balance between efficacy and safety. These insights can help guide the development of more efficient and safer beam-shaping assemblies for clinical BNCT applications.

Purpose: This study presents a deep learning framework for automatic parotid segmentation using three-dimensional (3D) U-Net and attention-augmented 3D U-Net architectures trained with a novel combined loss function tailored for anatomical accuracy and class imbalance.

Materials and methods: A curated dataset of 379 noncontrast head-and-neck computed tomography scans with expert-verified contours was used. Two architectures a residual 3D U-Net and its attention-enhanced variant were implemented using TensorFlow. The networks were trained with both categorical cross-entropy and a proposed combined loss integrating modified Dice Score Coefficient (mDSC) and focal loss (FL) with weights 0.7 and 0.3. The models were evaluated using dice similarity coefficient (DSC), Intersection over Union (IoU), and categorical accuracy. A custom checkpointing strategy was designed to preserve model weights corresponding to both peak DSC and minimum validation loss. The code and pretrained models are hosted on a publicly available GitHub repository at: https://github.com/1aryantyagi/Segmentation-Paper.

Results: The 3D U-Net trained with the combined loss achieved a median Dice score of 0.8835 (left parotid) and 0.8709 (right), with mean IoU values of 0.7672 and 0.7358, indicating strong segmentation accuracy. The U-Net produced comparable results, supporting the combined loss's consistency. Bland-Altman analysis confirmed reduced variability and improved agreement.

Conclusion: The integration of mDSC and FL within a 3D U-Net architecture significantly improves segmentation performance, robustness, and spatial precision. These findings support the clinical feasibility of the proposed framework for automated, reproducible parotid delineation in radiotherapy planning.

Background and purpose: This study aimed to compare dose gradients and peripheral dose spillage across three advanced radiotherapy modalities: C-Arm Linac (TrueBeam STx), Ring Gantry (Radixact^®), and Robotic Arm Linac (CyberKnife^®). The focus was on evaluating their performance in delivering steep dose gradients and minimizing low-dose spillage, which is crucial for reducing toxicities and optimizing treatment outcomes.

Materials and methods: A retrospective analysis of 110 patients with multi-lesion brain tumors treated with stereotactic radiotherapy was conducted. Patients were grouped by target volume size, and treatment plans were created using the three modalities. Dosimetric parameters, including gradient index (GI5-GI50), dose conformity, and homogeneity, were analyzed following ICRU 91 guidelines. Phantom verification using the Rando Phantom and PTW SRS1600 detector ensured clinical reliability.

Results: The Robotic Arm Linac demonstrated the steepest dose gradients and lowest GI values at GI10 and GI5, highlighting superior precision and minimal low-dose spillage (P < 0.05). The C-Arm Linac and ring gantry showed comparable performance at higher GI values (GI20-GI50), with the ring gantry achieving broader dose coverage for larger targets. Phantom validation supported these findings, confirming modality-specific advantages.

Conclusion: Robotic Arm Linac is optimal for precise treatments with stringent organ-at-risk sparing, while C-Arm Linac and ring gantry are better suited for broader dose coverage in complex cases. These results provide guidance for modality selection based on tumor size and clinical needs, with potential for further optimization through artificial intelligence and advanced planning tools.

Background: Advanced radiotherapy techniques are widely used for the treatment of breast cancer. Breast cancer treatment often involves acute and late radiation effects on the skin. Quantification of surface dose is crucial due to its impact on skin toxicity, cosmetic outcomes, and local tumor control.

Aims: In this study, the feasibility of using unlaminated Gafchromic EBT3 film (Ashland Advanced Materials, Bridgewater, NJ, USA) for quantification of surface dose during breast radiotherapy was explored.

Materials and methods: The dosimetric properties of this film, such as dose response, dose rate dependence, orientation dependence, and Ultraviolet-spectrum analysis, were also studied. A CIRS thorax phantom (CIRS Inc. Norfolk, Virginia) with a removable breast attachment was used for surface dose measurement. Tangential beam three-dimensional-conformal radiotherapy (3D-CRT) and 7-field intensity-modulated radiotherapy (IMRT) treatment plans were generated on the computed tomography images of the CIRS thorax phantom. The unlaminated EBT3 films were positioned on the breast attachment, at medial and lateral positions, and the treatment was delivered using 6 megavoltage X-rays generated by a TrueBeam medical electron linear accelerator (Varian Medical Systems, Palo Alto, USA). The exposed films were scanned with an EPSON 10000XL flatbed scanner, and the percentage surface dose was determined by taking the ratio of the measured dose and prescribed dose.

Results: The average percentage surface doses for 3D-CRT and IMRT techniques were found to be 43.11% and 34.78%, respectively.

Conclusion: This study demonstrated the feasibility of using unlaminated EBT3 Gafchromic film for surface dose evaluation in breast radiotherapy, providing valuable insights to optimize treatment strategies and enhance patient outcomes.

Objective: The quality of positron emission tomography/computed tomography (PET/CT) and PET/magnetic resonance (MR) images is a critical factor affecting the diagnostic reliability of clinical studies. An image quality (IQ) phantom is used to evaluate PET IQ. This study aims to acquire PET IQ phantom images on PET/CT and PET/MR systems and investigate the effects of attenuation correction on IQ.

Materials and methods: The PET IQ phantom was filled with F-18 fluorodeoxyglucose to achieve a lesion/background ratio of 4/1. The phantom was imaged separately on two different PET/CT scanners and one PET/MR scanner. Region of interest (ROIs) was drawn on both raw and attenuation-corrected images. Contrast and contrast-to-noise ratios (CNRs) were calculated from ROI counts.

Results: For the smallest lesion with a diameter of 10 mm without attenuation correction, the contrast values in PET/CT and PET/MR images were found to be 1.32, 1.68, and 1.76, respectively. In attenuation-corrected images of the same lesion, the contrast values were 1.55, 1.89, and 2.22, respectively. The CNR values for the attenuation-corrected 10 cm lesion were 13.92, 12.05, and 13.36, respectively.

Conclusion: Attenuation correction significantly improved IQ, with attenuation-corrected images demonstrating higher contrast and CNR values.

Purpose: This study investigates the preclinical dosimetry of Lutetium-177 in targeted radionuclide therapy for cancer treatment, focusing on the radiopharmaceutical [¹⁷⁷Lu] Lu-DOTATATE.

Materials and methods: Subcutaneous HT-29 xenografts were established in nu/nu mice and administered [¹⁷⁷Lu] Lu-DOTATATE intravenously. Tumor uptake was quantified by radiometry, and absorbed doses were calculated using MIRD formalism and a linear-quadratic model to determine cell death probability.

Results: Using HT-29 tumor cell xenografts in nude mice, the research evaluated the absorbed doses in tumor cell nuclei, considering both self-radiation and cross-radiation from neighboring cells. A mathematical model of a spherical cell cluster (radius = 280 µm) was developed to simulate dose distribution, accounting for beta-particle range and linear energy transfer.

Conclusions: Results indicate that the absorbed dose varies with cell position within the cluster, ranging from 3.2 Gy at the surface to 8.4 Gy at the center. The study also calculates cell survival probabilities using a linear-quadratic model, revealing a relative biological effectiveness (RBE) of up to 2 for [¹⁷⁷Lu] Lu-DOTATATE compared to gamma irradiation. These findings underscore the therapeutic potential of [¹⁷⁷Lu] Lu-DOTATATE in inducing tumor regression.

Background: In July 2024, a major failure in our magnetic resonance-guided linear accelerator (MR-linac) led to a 2-week treatment interruption for four patients.

Objective: To evaluate the clinical outcomes and toxicities at the end of the treatment, and up to 6 months after dose compensation for treatment interruption.

Materials and methods: Following the guidelines of the Royal College of Radiologists, we hypo fractionated the treatments. The dose per fraction and number of fractions were recalculated using the linear-quadratic model.

Results: Despite the treatment delay, the adapted dose had minimal impact on the oncological outcomes in the follow-up up to 6 months posttreatment. All patients demonstrated adequate tolerance to toxicity with no disease progression. Two patients achieved complete responses, while two showed partial responses.

Conclusion: Hypofractionation proved to be a feasible strategy for managing treatment interruptions in MR-linac adaptive radiotherapy, ensuring treatment continuity without compromising safety and efficacy. The biological approach maintained tumor control while minimizing toxicity, demonstrating its potential for broader applications in similar clinical scenarios.

Purpose: The purpose of this study was to implement and evaluate an end-to-end (E2E) test using an anthropomorphic head-and-neck (H and N) phantom across different radiotherapy modality platforms for photon and proton.

Materials and methods: The CIRS SHANE anthropomorphic phantom was used for the E2E test across multiple modalities, including Varian Clinac iX, four Varian TrueBeam systems, two Varian Halcyon systems, one Varian ETHOS system, and one ProBeam system. The volumetric modulated arc therapy plans were independently optimized for each photon machine while the intensity modulated proton therapy was planned for proton. Dosimetric measurements were taken using a 0.125 cc semiflex ionization chamber at three planning target volumes (PTVs) and one spinal cord point.

Results: The Kruskal-Wallis test revealed no significant differences in dose/volume parameters across the different modalities (P = 0.976). However, the Varian TrueBeam system showed a higher percentage dose difference compared to the other modalities, with the largest deviation of -3.7%. The O-ring gantry systems, Varian Halcyon and Varian ETHOS, demonstrated smaller dose differences and deviations compared to the C-based gantry systems. The Varian ETHOS delivered the highest number of monitor units at 641.6, while the Clinac iX delivered the lowest at 407.0. The ProBeam system showed good results for PTV and spinal cord.

Conclusions: The implementation of the E2E test using an anthropomorphic H and N phantom has proven feasible across multiple radiotherapy modalities in both photon and proton machines. This approach offers a valuable tool for center with advanced radiotherapy technologies, particularly those without a national dosimetry audit program.