Medical physics最新文献

{"title":"OPT-former: A transformer-based approach for eye motion tracking in ocular proton therapy","authors":"Celeste Pinori,&nbsp;Gabriele Ariata,&nbsp;Maksym Hladchuk,&nbsp;Andrea Pella,&nbsp;Maria Rosaria Fiore,&nbsp;Giuseppe Magro,&nbsp;Chiara Paganelli,&nbsp;Agnieszka Chalaszczyk,&nbsp;Marco Rotondi,&nbsp;Mario Ciocca,&nbsp;Guido Baron","doi":"10.1002/mp.70526","DOIUrl":"https://doi.org/10.1002/mp.70526","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Ocular proton therapy (OPT) is the main conservative treatment option for ocular melanoma but requires precise gaze stability monitoring during irradiation. In current clinical workflows at the Centro Nazionale di Adroterapia Oncologica (CNAO) in Pavia, this task is performed manually by trained operators, which can introduce reaction delays and human error.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>To propose and assess the performance of <i>OPT-former</i>, a transformer-based method for real-time segmentation of the iris and pupil, designed for integration into automated safety interlock systems.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The model, inspired by the MedFormer architecture, performs direct three-class segmentation (background, iris, pupil) on 256 × 256 images acquired by CNAO's Eye Tracking System. A dataset of 1683 frames from 173 patients was split into independent training (1349 frames), validation (110 frames), and testing (224 frames) sets. Data augmentation was applied to increase variability, resulting in 3902 training images. Performance was evaluated using Intersection over Union (IoU), Dice coefficient, and Szymkiewicz–Simpson coefficient (SSC), assessing the <i>OPT-former</i> performance and comparing them with a previously proposed <i>3-step U-Net</i> and manual expert segmentations. Patient-specific fine-tuning was implemented using pre-treatment dry-run session videos.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The <i>OPT-former</i> achieved segmentation accuracy comparable to the <i>3-step U-Net</i>, but with substantially lower inference times (8.62 Hz on GPU vs. 3.26 Hz), enabling near frame-by-frame real-time analysis. Patient-specific fine-tuning significantly improved all metrics for both structures (<i>p</i> &lt; 0.05), with the largest gains for the iris. After fine-tuning, the model outperformed manual annotations from two experts under simulated clinical conditions.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The proposed <i>OPT-former</i> achieves high segmentation accuracy, with reduced pipeline complexity and inference time suitable for clinical application. These results endorse its role as a robust support tool in clinical practice, seamlessly integrating into established workflows without impacting the consolidated routine, and enabling automated gaze monitoring and real-time safety interlock activation in OPT.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148753882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Experimental quantification of frame-level active window duration and duty cycle in high frame rate x-ray detectors","authors":"Ruiran Lai,&nbsp;Linying Zhan,&nbsp;Bujar Mehmeti,&nbsp;Ran Zhang,&nbsp;Ke Li,&nbsp;Guang-Hong Chen","doi":"10.1002/mp.70627","DOIUrl":"10.1002/mp.70627","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Detector frame rate is commonly used as the primary metric to characterize the temporal behavior of x-ray detectors in medical imaging systems. However, frame rate alone specifies only the sampling rate of image frames and does not determine how x-ray interaction signals are temporally integrated to form the measured signal within each frame. In medical imaging applications, an important frame-level component of detector temporal performance is characterized by the actual x-ray active window and by the fraction of the frame time during which the detector is actively exposed to x-rays. These frame-level timing characteristics are crucial for assessing detector radiation dose efficiency, yet they are rarely measured experimentally.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;To develop and experimentally validate a method for quantifying the frame-level temporal behavior of high frame rate x-ray detectors beyond frame rate alone, by directly measuring the actual x-ray active window duration and the detector temporal duty cycle during active x-ray exposure.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;A rotating test object with sharp lead-foil edges was imaged using a high frame rate photon-counting detector over a wide range of nominal frame-rate settings up to 4500 fps. An independent laser tachometer was used to measure the angular velocity of the rotating object. Motion blur was analyzed using simple kinematic models that establish a direct relationship between the detector active window duration and the angular width of the blurred edge-spread function (ESF) in projection images. The active window duration was estimated from the measured angular blur width. The experimentally measured frame rate was determined independently from the angular displacement of recurring features between frames. The detector duty cycle was computed as the ratio of the measured active window duration to the measured frame time.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Results&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;The experimentally measured frame rate closely matched the nominal setting over the full tested range, with a linear relationship characterized by a slope of 1.002 and &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mrow&gt;\u0000 &lt;msup&gt;\u0000 &lt;mi&gt;R&lt;/mi&gt;\u0000 &lt;mn&gt;2&lt;/mn&gt;\u0000 &lt;/msup&gt;\u0000 &lt;mo&gt;=&lt;/mo&gt;\u0000 &lt;mn&gt;0.9999&lt;/mn&gt;\u0000 &lt;/mrow&gt;\u0000 &lt;annotation&gt;$R^{2}=0.99","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.70627","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148714820","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A single-view-based electroacoustic tomography imaging using deep learning for electroporation monitoring","authors":"Jie Zhang,&nbsp;Yifei Xu,&nbsp;Leshan Sun,&nbsp;Yankun Lang,&nbsp;Liangzhong Xiang,&nbsp;Lei Ren","doi":"10.1002/mp.70623","DOIUrl":"10.1002/mp.70623","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Electroacoustic tomography (EAT) is an emerging imaging technique with potential for guiding electroporation therapy. However, in clinical applications, EAT scan is typically limited to a single view acquisition, leading to severe artifacts in the reconstruction.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;This study aims to develop novel deep-learning models to substantially enhance the image quality of the single-view EAT reconstruction.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;We proposed a two-stage deep learning model (EXPP-Net) to address the limited view problem in EAT image reconstruction. The two-stage model comprises extrapolation and post-processing sub-models. The extrapolation sub-model receives a single-view sinogram &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt; acquired at any angle &lt;i&gt;α&lt;/i&gt; and a rotation angle &lt;i&gt;γ&lt;/i&gt;∈[−60°, −24°, 24°, 60°], and predicts four single-view sinograms at angles &lt;i&gt;α&lt;/i&gt;+&lt;i&gt;γ&lt;/i&gt;. The 5 sinograms (i.e., &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;-60°&lt;/sub&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;-24°&lt;/sub&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;+24°&lt;/sub&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;+60°&lt;/sub&gt;) are back-projected to reconstruct an initial sparse-view EAT image (&lt;b&gt;&lt;i&gt;m&lt;/i&gt;&lt;/b&gt;&lt;sub&gt;sv&lt;/sub&gt;). The post-processing sub-model enhances &lt;b&gt;&lt;i&gt;m&lt;/i&gt;&lt;/b&gt;&lt;sub&gt;sv&lt;/sub&gt;’s quality to a full-view image (&lt;b&gt;&lt;i&gt;m&lt;/i&gt;&lt;/b&gt;&lt;sub&gt;fv&lt;/sub&gt;). These sub-models were trained using experimental data. Using a linear-array probe, 54 full-view datasets (60 or 90 views each) were acquired by delivering electrical pulses via two tungsten electrodes rotated from −180° to 180° in a water tank. Each view &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;, together with {&lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;-60°&lt;/sub&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;-24°&lt;/sub&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;+24°&lt;/sub&gt;, &lt;b&gt;&lt;i&gt;u&lt;/i&gt;&lt;/b&gt;&lt;i&gt;&lt;sub&gt;α&lt;/sub&gt;&lt;/i&gt;&lt;sub&gt;+60°&lt;/sub&gt;} and &lt;b&gt;&lt;i&gt;m&lt;/i&gt;&lt;/b&gt;&lt;sub&gt;fv&lt;/sub&gt;, forms a view set for model training. Data splitting was performed at the dataset level, yielding 34/10/10 datasets for training/validation/test, corresponding to 2190/600/600 view sets. The model performance was evaluated using root mean square error (RMSE), structural similarity index measure (SSIM), and the iso-pressure line Dice similarity coefficients (DICE). Statistical analysis used a linear mixed-effects model with Benjamini–Hochberg FDR correction (&lt;i&gt;α &lt;/i&gt;= 0.05), and effect size was defined as standardized fixed-effect coefficients (&lt;i&gt;β&lt;/i&gt;/&lt;i&gt;σ&lt;/i&gt;). For comprehensive evaluation, the model was re-trained and evaluated on convex-array probe data, and further tes","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13455628/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148703175","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Magnetic resonance imaging safety-driven, heat-mitigating design for a magnetic resonance imaging-guided, magnetically actuated robotic catheter","authors":"E. Erdem Tuna,&nbsp;Gabriel J. Foss,&nbsp;Sanjana Kamath,&nbsp;Ridaa Z. Ali,&nbsp;Yen-Chun Chen,&nbsp;Ahmed B. Aydin,&nbsp;Reid Bolding,&nbsp;Mark Griswold,&nbsp;M. Cenk Cavusoglu","doi":"10.1002/mp.70630","DOIUrl":"10.1002/mp.70630","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Metallic components in magnetic resonance imaging (MRI)−guided robotic catheters, such as leadwires and microcoils, can interact with the scanner's radiofrequency (RF) fields, causing heating via the antenna effect and Joule heating, leading to localized temperature rises and an increased risk of thermal injury. Ensuring thermal safety will enable safe and reliable operation of magnetically actuated robotic catheters within MRI environments.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;This study aims to develop a novel MRI-compatible robotic catheter design that mitigates RF-induced and resistive heating, ensuring patient safety and proper functionality during MRI-guided interventions, particularly for cardiac ablation procedures. The catheter is directly actuated by magnetic torques generated on current-carrying microcoils at the tip by the MRI scanner's static magnetic field, allowing precise navigation.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;The catheter design incorporates systematically tuned microcoil parameters and distributed capacitors along the leadwires to suppress resonance and minimize RF energy absorption. A dithering technique applying actuation currents at a 50&lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mo&gt;%&lt;/mo&gt;\u0000 &lt;annotation&gt;$%$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt; duty cycle (10 Hz, 50 ms on/off) promotes forced convection to reduce resistive heating. A resistor-capacitor (RC) equivalent circuit model was used to characterize the catheter's thermal behavior. Thermal validation was conducted on a 3T MRI scanner using a saline-filled phantom, fiber-optic temperature probes, and a balanced steady-state free precession (bSSFP) imaging sequence with actuation currents up to 1 A. Hot spot detection sweeps along the leadwires and spatial heating assessments at multiple bore positions were also performed.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Results&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;The prototype exhibited a maximum temperature increase of 0.6&lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mrow&gt;\u0000 &lt;msup&gt;\u0000 &lt;mrow&gt;&lt;/mrow&gt;\u0000 &lt;mo&gt;∘&lt;/mo&gt;\u0000 &lt;/msup&gt;\u0000 &lt;mi&gt;C&lt;/mi&gt;\u0000 &lt;/mrow&gt;\u0000 &lt;annotation&gt;$^circ{rm C}$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt; during simultaneous imaging and 1 A dithered actuation, well within Food and Drug Administration (FDA) sa","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13455776/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148703240","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of an online neutron beam monitoring method for accelerator-based BNCT using an 8-MeV proton-beryllium source","authors":"Masashi Takada,&nbsp;Hiroaki Kumada,&nbsp;Masashi Oyauchi,&nbsp;Susumu Tanaka,&nbsp;Tomoya Nunomiya,&nbsp;Toshiyuki Ohba,&nbsp;Akihiko Masuda,&nbsp;Tetsuro Matsumoto","doi":"10.1002/mp.70629","DOIUrl":"10.1002/mp.70629","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Boron neutron capture therapy (BNCT) has emerged as a promising cell-selective clinical treatment for recurrent glioblastoma as well as head and neck cancers. In recent years, BNCT has been implemented in several hospitals using accelerator-based neutron sources. However, direct online measurement of the incident neutron beam intensity during clinical BNCT remains a major challenge. Neutron dosimetry is typically inferred from the accumulated electric charge of the incident proton beam delivered to the beryllium neutron-production target.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;To improve the reliability of neutron beam control, it is essential to implement an online neutron-monitoring system that complements proton beam current measurements. Such a system enables precise real-time regulation of the neutron dose delivered to patients. In this study, we developed a new online neutron-beam monitoring technique capable of measuring the neutron beam intensity in real time.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Reliable monitoring of neutron intensity at the patient position was achieved by detecting fast neutrons. The fast-neutron detector consisted of a silicon diode coupled with a high-density polyethylene neutron converter. The spatial distribution of neutrons was evaluated using Monte Carlo simulations to determine the optimal placement of the neutron beam monitor. The proposed method was validated through gold-activation analysis, cell irradiation experiments, and accelerator tuning.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Results&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Online neutron-beam monitoring was successfully achieved by detecting fast neutrons at the end of the vacuum beam duct, where neutrons propagate directly without traversing radiation shielding. The neutron counting rate increased immediately at the onset of proton-beam irradiation and decreased upon its termination, exhibiting periodic fluctuations throughout the irradiation period. These fluctuations did not significantly affect the integrated neutron counts, as they were averaged over a typical BNCT irradiation time of approximately 30 min. The fast-neutron detector produced a mean counting rate of 124.6 counts per mC of incident proton-beam charge, with a relative standard deviation of 1.2%, and yielded a neutron-calibration factor of &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mrow&gt;\u0000 &lt;mn&gt;4.9&lt;/mn&gt;\u0000 &lt;mo&gt;×&lt;/mo&gt;\u0000 &lt;msup&gt;\u0000 &lt;mn&gt;10&lt;/mn&gt;\u0000 &lt;mn&gt;6&lt;/mn&gt;\u0000 &lt;/msup&gt;\u0000 &lt;mspac","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148703209","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of a new set of Geant4-DNA Physics models for the simulation of electron track structures in liquid water up to 10 MeV","authors":"Sébastien Incerti,&nbsp;Ioanna Kyriakou,&nbsp;Hoang N Tran,&nbsp;Shogo Okada,&nbsp;Wook-Geun Shin,&nbsp;Dimitris Emfietzoglou","doi":"10.1002/mp.70618","DOIUrl":"10.1002/mp.70618","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Improved Physics models for the elastic and inelastic interactions of electrons in liquid water have recently been developed and implemented in the Geant4-DNA toolkit, allowing detailed electron transport simulations up to MeV energies.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>To present a preliminary performance evaluation of a new set of Geant4-DNA Physics models, which will become the default Physics list for liquid water in the upcoming Geant4 public releases.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A new set of Physics models (constructor), denoted as « Option4 X », is implemented in Geant4-DNA that combines a newly developed inelastic model with two alternative models for the simulation of elastic scattering, one based on the ELSEPA software and another one based on the analytic Uehara Screened Rutherford model. The inelastic model includes all the low-energy developments of the existing « Option4 » Physics constructor, upgrades its internal consistency (via sum rules), and extends its application from 10 keV to 10 MeV using the Relativistic Plane Wave Born Approximation (RPWBA).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Preliminary results for electron dose point kernels (DPKs), frequency-mean (yF) and dose-mean (yD) lineal energies in nanometric spheres, S-values in spherical cells, and the W-value of liquid water are simulated, and a comparison is presented between the new Physics constructor, « Option4 X », the current default Physics constructor, « Option2 », and the low-energy Physics constructor, « Option4 ». Significant improvement is offered by « Option4 X » in the simulation of DPK and the W-value, which have been two long-lasting performance limitations of the current default Physics constructor, while the choice of elastic model has a relatively small influence.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>«Option4 X» improves and extends the current Geant4-DNA transport capabilities up to 10 MeV and will become the default Geant4-DNA Physics constructor for liquid water in the upcoming public releases of Geant4. This development will allow Geant4-DNA users, along with users of Monte Carlo software that are built on Geant4 (like GATE and TOPAS-nBio), to perform more accurate electron track structure simulations over a wider energy range (10 eV ‒ 10 MeV) and cover radiotherapeutic applications of track structure simulations not reachable until now.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148703193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Generative AI, foundation models and large language models in radiation therapy physics: Clinical applications, challenges, and future directions","authors":"X. Sharon Qi,&nbsp;Yi Wang,&nbsp;Xiaofeng Yang,&nbsp;Lei Ren,&nbsp;Wei Liu,&nbsp;Stanley H. Benedict,&nbsp;Ying Xiao,&nbsp;Lei Xing,&nbsp;Issam M. El Naqa","doi":"10.1002/mp.70616","DOIUrl":"10.1002/mp.70616","url":null,"abstract":"<p>Generative AI (Gen AI), Foundation Models (FMs), and Large Language Models (LLMs) are powerful emerging technologies that demonstrate exceptional capabilities in processing vast amounts of unstructured and structured data, including text, voice, images, video and other formats, and adapting to a wide range of specific tasks. Their immense potential to drive meaningful improvements in treatment outcomes is increasingly evident. The advent of these technologies has marked a transformative era in healthcare, including the fields of radiation oncology and medical physics. Specifically, these powerful technologies offer unprecedented opportunities to analyze domain-specific data, process and synthesize medical images, automate routine tasks, support clinical decision-making, optimize and streamline clinical workflows, and enhance the quality of clinical trials. While these emerging technologies present new opportunities to revolutionize radiation therapy practice, their implementation also raises important educational, ethical, and regulatory considerations. This scoping review highlights benefits, promises, risks, and challenges, such as interpretability, data privacy, regulatory compliance, reproducibility, hallucination, and integration into existing clinical workflow. Finally, emerging opportunities are outlined to guide future research directions. This review paper provides a timely overview of Gen AI, FMs and LLMs, aiming to inform medical physicists, clinicians, and researchers of the evolving role of these disruptive technologies in shaping the future of radiation therapy.</p>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148687279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"HyperSight CBCT dose calculation accuracy for adaptive head and neck radiotherapy: Results from a prospective clinical trial","authors":"Abby Yashayaeva,&nbsp;R Lee MacDonald,&nbsp;Kenny Zhan,&nbsp;Jennifer DeGiobbi,&nbsp;Natasha McMaster,&nbsp;Dave McAloney,&nbsp;Lucy Ward,&nbsp;Cheryl Anderson,&nbsp;Marc LeBlanc,&nbsp;Lara Best,&nbsp;Murali Rajaraman,&nbsp;Derek Wilke,&nbsp;Amanda Cherpak","doi":"10.1002/mp.70632","DOIUrl":"10.1002/mp.70632","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Head and neck (H&amp;N) patients frequently undergo anatomical changes such as weight loss and tumour shrinkage during treatment, requiring repeat imaging and replanning to maintain accurate dose delivery. On-board cone-beam CT (CBCT) systems could support these workflows, however they have historically been limited by reduced Hounsfield unit accuracy, inferior image quality, and susceptibility to metal artifacts from dental hardware, all of which are critical for effective radiotherapy treatment planning and dose delivery. HyperSight CBCT offers improved image quality comparable to FBCT along with metal artifact reduction techniques, but its dosimetric performance in H&amp;N patients, particularly in the presence of dental metal implants, has not yet been evaluated.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;This work evaluates the dosimetric accuracy of the HyperSight CBCT system (Varian Medical Systems), including the impact of metal artifact reduction, to determine its reliability for H&amp;N treatment planning.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Images of 25 H&amp;N cancer patients were acquired on a conventional fan-beam CT (FBCT. GE Healthcare) and on two CBCT systems: an Ethos radiotherapy system equipped with HyperSight CBCT, which provides advanced iterative reconstruction (HS-Acuros) and metal artifact reduction reconstruction (HS-MAR) options, and a TrueBeam linear accelerator system with conventional CBCT, which uses filtered back projection reconstruction (Varian Medical Systems). For each patient, the CBCTs were rigidly registered to the FBCT, and the clinically accepted reference plans were forward calculated on all images. Gamma evaluations (3%/3 mm, 2%/2 mm, 1%/1 mm) for all CBCT modes were performed relative to FBCT on slices both unaffected and affected by metal artifacts, with the latter also analyzed as a function of the metal artifact index (AI). DVH metrics were assessed for the planning target volume (PTV), mandible, oral cavity, and parotid glands.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Results&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;For all gamma criteria, HS-Acuros images demonstrated significantly higher pass rates than both HS-MAR and TrueBeam CBCTs in the artifact-free slices (&lt;i&gt;p&lt;/i&gt; ≤ 0.033). In artifact-affected slices, HS-MAR reconstructions performed the best in slices affected by metal artifacts (&lt;i&gt;p&lt;/i&gt; ≤ 0.001), exhibiting the least reduction in gamma pass rates with increasing artifact index (AI). Conventional TrueBeam CBCT consistently yielded the lowest gamma pass rates in artifact-prone regi","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13447906/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148687252","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Estimating pediatric lung volumetric parameters via rapid, limited-slice, free-breathing thoracic dynamic MRI","authors":"You Hao,&nbsp;Yubing Tong,&nbsp;Caiyun Wu,&nbsp;Joseph M. McDonough,&nbsp;Samantha Gogel,&nbsp;David M. Biko,&nbsp;Patrick J. Cahill,&nbsp;Jason B. Anari,&nbsp;Drew A. Torigian,&nbsp;Jayaram K. Udupa","doi":"10.1002/mp.70628","DOIUrl":"10.1002/mp.70628","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Dynamic, separate lung volume assessment during breathing is critical for evaluating thoracic disorders like scoliosis, thoracic insufficiency syndrome (TIS), and pulmonary diseases, as well as for monitoring treatments. While traditional volumetric assessment relies on static 3D imaging during breath-holds, this is unfeasible for many pediatric patients or those with severe respiratory distress. Free-breathing 4D dynamic MRI (dMRI) is the ideal modality due to its lack of radiation, excellent soft-tissue contrast, and flexible imaging planes. However, its clinical use is hampered by long acquisition times, during which patients struggle to maintain stable, consistent breathing patterns or remain still. Consequently, methods to accelerate dMRI acquisition while maintaining volumetric accuracy are urgently needed to make free-breathing assessments clinically practical.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;We present an observational study involving free-breathing short-scan-time dynamic MRI (dMRI) method that can be routinely used for computing dynamic lung volumes accurately.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;(1) Full sampled free-breathing sagittal 2D dMRI scans are gathered from 45 normal children via bSSFP sequence. Sparse dMRI (s-dMRI) scans are simulated from these datasets by optimally subsampling in the spatio-temporal domains via a limited number of selected sagittal locations and time instances. (2) A 4D image is constructed from both scans. Lungs are segmented from 4D image, and their volumes from full and sparse dMRI scans are computed. (3) A regression model is developed to predict full-scan volumes from sparse-scan data on a training set. (4) The accuracy is analyzed on both synthesized sparse dMRI scans from a separate fully-sampled-scan test set and actual s-dMRI scans prospectively acquired from 10 normal children.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Results&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;With 5 slices per lung and 40 time points, the predicted volume showed a ∼2% deviation from the full-scan volume, with a total scan-time of ∼9 min (vs. 49.72 ± 5.01 min for the full scan with 15–22 slices per lung and 80 time points). When the spatial sampling was increased to full number of slices (15-22 per lung) but only 40 time points, these metrics become 0.4%, and 24.86 ± 2.5 min.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Conclusions&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;s-dMRI is a practical approach for computing dynamic lung volumes that can be used","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13447894/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148687291","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Monte Carlo simulations of \u0000 \u0000 \u0000 \u0000 \u0000 241\u0000 \u0000 Am\u0000 \u0000 ${}^{241}mathrm{Am}$\u0000 \u0000 \u0000 α\u0000 $alpha$\u0000 -particle energy distribution, LET distribution, and absorbed dose in a cell layer","authors":"Martyna Araszkiewicz,&nbsp;Katarzyna Tymińska,&nbsp;Urszula Kaźmierczak,&nbsp;Patrycja Chuchała,&nbsp;Jarosław Grzyb,&nbsp;Patrycja Kamińska,&nbsp;Roch Kwiatkowski,&nbsp;Marta Maleszewska,&nbsp;Aneta Malinowska,&nbsp;Kamila Maliszewska-Olejniczak,&nbsp;Bohdan Paterczyk,&nbsp;Agnieszka Korgul","doi":"10.1002/mp.70611","DOIUrl":"10.1002/mp.70611","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Background&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;Monte Carlo simulations constitute a well-established methodology for the quantitative characterization of &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;α&lt;/mi&gt;\u0000 &lt;annotation&gt;$alpha$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-particle transport and energy deposition. Owing to their high linear energy transfer (LET) and limited range in tissue, &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;α&lt;/mi&gt;\u0000 &lt;annotation&gt;$alpha$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-particles are increasingly regarded as a promising modality for targeted cancer therapy.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;The objective of this study was to perform a rigorous dosimetric characterization of an irradiation system utilizing a &lt;sup&gt;241&lt;/sup&gt;Am surface source, with emphasis on quantifying the &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;α&lt;/mi&gt;\u0000 &lt;annotation&gt;$alpha$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-particle energy spectrum, their energy deposition characteristics, and the resulting absorbed dose in a cell monolayer model, given its relevance to radiobiological applications and its potential translational implications for targeted &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;α&lt;/mi&gt;\u0000 &lt;annotation&gt;$alpha$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-emitter radionuclide therapy (TAT).&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Methods&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;A comprehensive geometric and material characterization of all irradiation system components was conducted. Cellular dimensions and culture medium thickness were quantified using confocal microscopy. A computational representation of the setup was implemented in the Monte Carlo N-Particle (MCNP) transport code to determine &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;α&lt;/mi&gt;\u0000 &lt;annotation&gt;$alpha$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-particle energy loss within relevant materials. Model predictions were validated against measurements obtained with a silicon semiconductor detector. Energy spectra of &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;α&lt;/mi&gt;\u0000 &lt;annotation&gt;$alpha$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-particles incident on the biological target were subsequently benchmarked against data acquired using CR-39 TASTRAK solid-state nuclear track detectors.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Results&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;The simulations revealed homogeneous ","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 8","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148687311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}