Medical physics最新文献

{"title":"2026 Joint AAPM COMP Meeting Abstracts","authors":"","doi":"10.1002/mp.70699","DOIUrl":"https://doi.org/10.1002/mp.70699","url":null,"abstract":"","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149146342","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":"Expansion and clinical validation of Markov model-based quality-adjusted life expectancy (QALE) simulations for head-and-neck cancer radiotherapy patients","authors":"Sebastian Tattenberg,&nbsp;Benoit Allignet,&nbsp;Xuanfeng Ding,&nbsp;Cornelia Hoehr,&nbsp;Peter Chen,&nbsp;Olivier Beuf","doi":"10.1002/mp.70696","DOIUrl":"10.1002/mp.70696","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Radiotherapy treatment plan comparisons often utilize structure doses, tumor control probabilities (TCPs), and normal tissue complication probabilities (NTCPs), but such metrics do not reflect the duration for which a patient suffers from a developed toxicity. Markov models allowing for the simulation of the patient's entire lifetime post-treatment and the determination of quality-adjusted life expectancy (QALE) have been developed but generally disregard important considerations such as additional treatments received and second primary malignancies (SPMs). Clinical validation also largely remains outstanding.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study aims to expand and clinically validate a Markov model focusing on head-and-neck cancer patients.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A Markov model considering tumor control, healthy tissue toxicities, distant metastases, radiation-associated secondary cancer, and death from various causes was expanded to further include additional treatments received, SPMs, and the patient's smoking history. For each of 148 head-and-neck cancer patients, 437,000 simulations of the patient's entire lifetime post-treatment were performed and compared to the patient's clinically-documented outcomes (wherever available).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>94.4% of simulations exhibited locoregional control, compared to 89.9% of patients in clinical reality. Simulated death with locoregional failure was 4.9%, compared to 5.4% in real life. Patients developed and died with distant metastases in 8.0% of simulations, while 10.8% of patients developed distant metastases in real life and 5.7% of patients passed away with distant disease. The simulated second primary cancer rate was 13.5%. Patients exhibited at least one healthy tissue toxicity in 43.8% of simulations, and simulated life expectancy and QALE were 76.5 ± 11.6 years and 74.3 ± 11.0 quality-adjusted life years (QALYs), respectively.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Agreement between simulation results and available clinically-documented outcomes was generally excellent, enabling the evaluation of prospective clinical outcomes and comparisons between different patient populations along with variable treatment combinations.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121082","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":"Radiogenomic inference of transcription factor regulatory architecture associated with a high-risk CT phenotype in Non-Small Cell Lung Cancer","authors":"Daisuke Kawahara,&nbsp;Misato Kishi,&nbsp;Yuzuha Kadooka,&nbsp;Yuji Murakami","doi":"10.1002/mp.70694","DOIUrl":"10.1002/mp.70694","url":null,"abstract":"&lt;div&gt;\u0000 \u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Purpose&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;To determine whether a CT radiomics-defined poor-prognosis phenotype in NSCLC is associated with reproducible transcriptomic programs and distinct transcription factor (TF) activity patterns.&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;This retrospective radiogenomic study used public datasets from The Cancer Imaging Archive. A prognostic radiomics model was developed from pretreatment CT in 402 patients from the NSCLC-Radiomics collection and evaluated in an independent cohort of 110 patients from the NSCLC Radiogenomics collection. Seven tumor-centered intra- and peritumoral regions were analyzed. Features were extracted from original, wavelet-, and Laplacian-of-Gaussian-filtered CT images. After preprocessing and training-cohort standardization, a LASSO-penalized Cox model was developed. The locked model retained eight features, all from the GTV+10-mm region, and was applied to the internal and external validation cohorts without refitting. The training-cohort median RadScore was used unchanged for Kaplan–Meier visualization. Matched transcriptomic data were analyzed using preranked gene set enrichment analysis (GSEA), gene set variation analysis (GSVA), target-gene-based TF activity inference, TF co-activity networks, exploratory regulator prioritization, unsupervised TF-state clustering, and LINCS-based perturbational analysis.&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 radiomics model achieved Harrell concordance indices of 0.785 (95% confidence interval [CI], 0.767–0.793), 0.819 (95% CI, 0.788–0.833), and 0.785 (95% CI, 0.752–0.802) in the training, internal validation, and external validation cohorts, respectively. The high-risk phenotype showed exploratory enrichment of proliferative and stress-adaptive programs, led by HALLMARK_E2F_TARGETS (normalized enrichment score [NES] = 1.696, false discovery rate [FDR] = 0.045), with additional enrichment of MYC targets and inflammatory response. HALLMARK_HYPOXIA showed a positive directional trend but did not meet the prespecified exploratory threshold. GSVA provided complementary sample-level evidence for MYC targets, hypoxia, E2F targets, DNA repair, and cell-cycle activity. Eleven of 13 evaluated TFs met the prespecified FDR threshold, including YAP1, MYC, HIF1A, E2F3, STAT3, KLF4, E2F1, and TP53. Integrated analysis prioritized HIF1A, STAT3, TP53, YAP1, and E2F1/E2F3 as candidate regulatory nodes. Unsupervised clustering of all 110 radiogenomic patients, performed independently of radiomics risk labels, identified three exploratory TF-defined states: Hypoxia_Metabolic (n = 69), Cell_Cycle_E2F (n = 31), and Hippo_Mechanosensing (n = 10). LINCS-based analysis generated regulatory-state-associated compound hypoth","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121155","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":"Expanding dose and dose-rate capabilities of a synchrotron-based pre-clinical proton FLASH irradiation platform","authors":"Zongsheng Hu,&nbsp;Yuting Li,&nbsp;Xiaochun Wang,&nbsp;Radhe Mohan,&nbsp;Uwe Titt","doi":"10.1002/mp.70695","DOIUrl":"10.1002/mp.70695","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;Proton FLASH radiotherapy has demonstrated promising normal tissue sparing for similar tumor control at ultra-high dose rates (&gt; 40 Gy/s), but systematic investigation of dose-rate-dependent biological mechanisms is limited. Most existing proton platforms operate within a narrow dose-rate range. Synchrotron-based systems are further constrained by limited beam current and cyclic (“spill-based”) delivery.&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 validate a synchrotron-based small-animal proton irradiation platform with expanded and controllable dose and dose-rate capabilities, enabling systematic studies across conventional, intermediate (meso)-dose-rate (μDR: 1–40 Gy/s) and FLASH regimes.&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 beamline was redesigned to incorporate three beam modulation modes (low, mid, and high dose) by modifying initial scattering conditions and re-optimizing and repositioning downstream components, including cone-shaped flattening filters, ridge filters, and range compensators. 14 discrete synchrotron beam extraction settings (denoted as “IDs”) with different beam fluxes were developed and combined with 3 beam modulation modes, resulting in 42 dose and dose-rate combinations. Ridge filters were fabricated using resin-based 3D printing with an iterative geometric correction workflow. Depth-dose distributions were measured using an Advanced Markus ionization chamber, and lateral dose profiles were verified using radiochromic film.&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 low-, mid-, and high-dose modes achieved maximum single-spill doses of 20.8 Gy (205 Gy/s), 30.2 Gy (297 Gy/s), and 53.8 Gy (531 Gy/s), respectively. Across all beam IDs and modulation modes, dose rates ranged from approximately 13 Gy/s to 2200 Gy/s, spanning μDR and FLASH regimes. Measured depth-dose profiles agreed with Monte Carlo simulations within ± 4%, and lateral dose flatness was within ± 3.7%.&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;This modular beamline redesign significantly expands the achievable dose and dose-rate combinations of a synchrotron-based proton irradiation platform without requiring accelerator modifications. The system enables 42 reproducible dose and dose-rate configurations with single-spill delivery, providing a flexible framework for systematic investigation of dose-rate-dependent radiobiological proton FLASH effects.&lt;/p&gt;\u0000 &lt;/secti","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13641327/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121095","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 quantum computing approach to beam angle optimization","authors":"Nimita Shinde,&nbsp;Ya-Nan Zhu,&nbsp;Haozheng Shen,&nbsp;Hao Gao","doi":"10.1002/mp.70688","DOIUrl":"10.1002/mp.70688","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Beam angle optimization (BAO) is a critical component of radiation therapy (RT) treatment planning, where small changes in beam configuration can significantly impact treatment quality, especially for proton RT. Mathematically, BAO is a mixed integer programming (MIP) problem, which is NP-hard due to its exponential growing search space. Traditional optimization techniques often struggle with computational efficiency, necessitating the development of novel approaches.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study introduces QC-BAO, a hybrid quantum-classical approach that leverages quantum inspired techniques to solve the MIP formulation of BAO.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The proposed approach, QC-BAO, models BAO as an MIP problem, incorporating binary variables for beam angle selection and continuous variables for optimizing spot intensities for proton therapy. The proposed approach employs a hybrid quantum-classical framework, utilizing quantum inspired techniques to solve the binary decision component while integrating classical optimization techniques, including iterative convex relaxation and the alternating direction method of multipliers.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Computational experiments were conducted on clinical test cases to evaluate QC-BAO's performance against institute-standard (IS) angles and a heuristic approaches, GS-BAO and AG-BAO. QC-BAO demonstrated consistently improved treatment plan quality over IS, GS-BAO and AG-BAO-selected angles. QC-BAO consistently increased the conformity index (CI) for target coverage while reducing mean and maximum doses to organs-at-risk (OAR). For instance, in the lung case, QC-BAO achieved a CI of 0.89, compared to 0.89 (IS), 0.76 (GS-BAO) and 0.86 (AG-BAO), while lowering the mean lung dose to 2.78 Gy from 3.36 Gy (IS), 4.80 Gy (GS-BAO) and 3.03 Gy (AG-BAO). Additionally, QC-BAO produced the lowest objective function values, confirming its superior optimization capability.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The findings highlight the potential of quantum-inspired algorithm to enhance the solution to BAO problem by demonstrated improvement in plan quality using the proposed method, QC-BAO. This study paves the way for future clinical implementation of quantum-accelerated optimization in RT.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121092","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":"Revisiting the humidity correction – A challenge to the current consensus","authors":"Yuichiro Morishita,&nbsp;Hideki Yamaoka","doi":"10.1002/mp.70697","DOIUrl":"10.1002/mp.70697","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;The humidity correction factor, &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;msub&gt;\u0000 &lt;mi&gt;k&lt;/mi&gt;\u0000 &lt;mi&gt;h&lt;/mi&gt;\u0000 &lt;/msub&gt;\u0000 &lt;annotation&gt;$k_mathrm{h}$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;, for ionization chambers in radiation dosimetry has relied on a consensus value established in ICRU Report 31 nearly half a century ago.&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 rigorously re-evaluate this long-standing consensus by employing modern instrumentation.&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;First, the equilibration time for the cavity air was examined by implanting a miniaturized humidity sensor directly into the cavity of a broken commercial ionization chamber to address uncertainties regarding stabilization in recent reports. Second, the &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;msub&gt;\u0000 &lt;mi&gt;k&lt;/mi&gt;\u0000 &lt;mi&gt;h&lt;/mi&gt;\u0000 &lt;/msub&gt;\u0000 &lt;annotation&gt;$k_mathrm{h}$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt; was measured using a vacuum-assisted gas exchange method, in which the air inside the cavity was exchanged actively. Finally, Monte Carlo simulations were performed to analyze energy deposition within the cavity, enabling the conversion of the measured &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;msub&gt;\u0000 &lt;mi&gt;k&lt;/mi&gt;\u0000 &lt;mi&gt;h&lt;/mi&gt;\u0000 &lt;/msub&gt;\u0000 &lt;annotation&gt;$k_mathrm{h}$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt; to the relative &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mi&gt;W&lt;/mi&gt;\u0000 &lt;annotation&gt;$W$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;-value.&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 direct monitoring of humidity inside the cavity revealed that the equilibration time scale extended to several days, which was much longer than the time scales assumed in the recent reports. The measured &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;msub&gt;\u0000 &lt;mi&gt;k&lt;/mi&gt;\u0000 &lt;mi&gt;h&lt;/mi&gt;\u0000 &lt;/msub&gt;\u0000 &lt;annotation&gt;$k_mathrm{h}$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt; for the Co-60 beam exhibited slight","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13641358/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121153","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":"Multimodal PET/MRI radiomics model for the classification of Alzheimer's disease and other dementia subtypes: A multicenter study","authors":"Bingjie Wang,&nbsp;Sheng Bi,&nbsp;Xiaoyin Xu,&nbsp;Hairong Zheng,&nbsp;Dong Liang,&nbsp;Greta S. P. Mok,&nbsp;Shaozhen Yan,&nbsp;Zhanli Hu","doi":"10.1002/mp.70690","DOIUrl":"10.1002/mp.70690","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;Accurate and early diagnosis of Alzheimer's disease (AD) remains a major clinical challenge. This study aims to explore the value of multimodal positron emission tomography (PET) and magnetic resonance imaging (MRI) radiomics to improve diagnostic accuracy.&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 evaluate hippocampus-based multimodal PET/MRI radiomics models for AD classification and exploratory five-class classification, and to compare early fusion (EF) with multi-criteria decision-making (MCDM)-based decision-level fusion.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Method&lt;/h3&gt;\u0000 \u0000 &lt;p&gt;A total of 758 patients who underwent both PET ([18F]FDG and Amyloid-β) and MRI (3DT1, T1WI, T2WI, T2-FLAIR) were retrospectively included as the primary dataset in this study. 945 radiomic features (RFs) per sequence were extracted from the bilateral hippocampus. Models were evaluated on the primary cohort using a five-fold stratified cross-validation at the subject level (80% training and 20% testing in each fold). Feature selection was performed independently for each sequence using the Least Absolute Shrinkage and Selection Operator (LASSO). For multimodal PET/MR models, features from all six sequences were concatenated, and feature selection was applied (EF). Six classifiers were used to construct classification models for two tasks, including (1) AD/non-AD classification, and (2) exploratory five-class classification among AD, FTD, VaD, MCI, and other dementia. In addition, an MCDM-based strategy was implemented to fuse multimodal RFs and multi-classifiers. Modality-matched external validation was performed in an independent cohort of 45 patients with 3DT1 and Aβ PET available. Decision curve analysis (DCA) was employed to assess the clinical usefulness of the models. SHapley Additive exPlanations (SHAP) was applied to explore the interpretability of the EF-based model.&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;Among the 758 patients in the primary dataset, AD, FTD, VaD, and MCI were present in 481 (63.5%), 34 (4.5%), 15 (2.0%), 136 (17.9%) cases, with 92 (12.1%) classified as the other dementia subtypes. The MCDM-based models achieved the highest area under the receiver operating characteristic curves (AUCs) for both tasks on the primary dataset (Task 1: AUC = 0.848; Task 2: AUC = 0.797), significantly higher than EF-based models in Task 2 (AUC = 0.782, &lt;i&gt;p&lt;/i&gt; &lt; 0.001) and showing comparable performance in Task 1 (AUC = 0.839, &lt;i&gt;p&lt;/i&gt; = 0.629). On the external dataset, EF and MCDM models achieved AUCs of 0.741 and 0.761, respectively, for Task 1. The MCDM model showed a higher AUC than EF.","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121211","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":"Deep learning-based CNN method for fiducial marker detection in kilovoltage X-ray images for liver tumor motion monitoring","authors":"Frances Kan,&nbsp;Freeman Jin,&nbsp;Adam Mylonas,&nbsp;Benjamin Zwan,&nbsp;Trang Nguyen,&nbsp;Trevor Moodie,&nbsp;Nicholas Hardcastle,&nbsp;Sau Fan Liu,&nbsp;Daniel Mason,&nbsp;Tim Wang,&nbsp;Yoo Young Lee,&nbsp;Paul Keall,&nbsp;Chandrima Sengupta","doi":"10.1002/mp.70691","DOIUrl":"10.1002/mp.70691","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;During radiation therapy, liver tumor motion can reduce dose delivery accuracy and increase irradiation of adjacent healthy tissues. Because liver tumors are difficult to visualize directly on kilovoltage (kV) X-ray images, fiducial markers are commonly implanted as surrogates for tumor position during treatment. Conventional marker-segmentation approaches such as template matching can lose accuracy when markers are obscured by bone, surgical clips, or stents. Although Convolutional Neural Network (CNN)-based approaches have shown strong performance for image analysis, their application has been constrained to more static organs such as the prostate or has used single-center and vendor-specific datasets, limiting assessment of model transportability across different institutions and anatomic sites.&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 aim of this study was to develop, integrate, and evaluate a CNN-based method for fiducial marker detection in kV X-ray images to support real-time image-guided radiation therapy (IGRT) during liver radiotherapy using a multi-institutional, multi-platform dataset spanning both large and fast free-breathing motion and slow and small breath-hold motion.&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 compact Convolutional Neural Network (CNN) was trained on 314,625 kilovoltage (kV) X-ray images encompassing 28 patients from the multi-institutional TROG 17.03 LARK clinical trial (NCT02984566). The model was validated using a hold-out set of 31,463 images (10%) from the same cohort of 12 patients and 55 treatment fractions, spanning three centers and three respiratory motion-management techniques. The CNN was tested on 4184 images from 16 patients and 55 fractions across three centers and three motion-management techniques. The ground truth for testing was manually segmented marker positions from every 10 degrees of gantry rotation for each fraction of each test patient. Based on AAPM guidelines (TG147 and TGB135.B), feasibility for clinical implementation was predefined as &gt; 95% of marker positions being within 2 mm of the ground truth position in each dimension, and processing time less than 150 ms per image using a simulated real-time Kilovoltage Intrafraction Monitoring (KIM) framework. Additional evaluation metrics included sensitivity, specificity, and the area under the precision-recall curve (AUC).&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 the unseen test patient data, the marker position was segmented by the CNN within 2 mm in 95.4% of frames on the X","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13635603/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121061","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":"Direct spatial convolution for pencil beam dose calculation and feasibility of dij-free IMRT optimization on GPU","authors":"Wen-Long Xia,&nbsp;Hong-Kai Wang,&nbsp;Bin Liang","doi":"10.1002/mp.70689","DOIUrl":"10.1002/mp.70689","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>In pencil beam dose calculation, the convolution is usually performed in the frequency domain via FFT. The dose is assembled into a sparse dose-influence matrix (d<i>ij</i>) for optimization. We propose to perform the convolution directly in the spatial domain with a deep-learning operator (Conv2d), and investigate the feasibility of a d<i>ij</i>-free optimization.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A 6 MV photon kernel (1 Gaussian penumbra and 3 singular value decomposition (SVD) scatter components, 0.5 mm resolution) was adopted. The Conv2d operator was validated against FFT on head-and-neck (HN), prostate and liver cases, with 7, 5, and 4 beams and corresponding 7503, 2649, and 845 bixels. The cost of building d<i>ij</i> was investigated to identify the bottleneck. A one-layer network optimization framework was implemented to investigate the feasibility of the d<i>ij</i>-free optimization (workflow A). Workflow A recomputes the dose with Conv2D at each iteration, with no pre-calculated d<i>ij</i>. The traditional d<i>ij</i>-based workflow (B) was implemented for comparison.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Across all 16 beams Conv2d reproduced FFT at machine precision (maximum absolute dose difference 1.35 × 10⁻¹³; Pearson correlation 1.0). The computational time of FFT versus Conv2d was 0.995s versus 0.502s (HN), 0.650s versus 0.199s (prostate), and 0.370s versus 0.082s (liver). D<i>ij</i> construction time was 132.9s, 76.4s, and 31.9s, with assembly accounting for ∼89%. Workflow A performs no build but a full forward and backward convolution (2.166s, 0.832s, 0.368s) per-iteration; workflow B performs the one-time d<i>ij</i> build (130.2s, 76.0s, 31.5s), then a sparse matrix-vector product (0.25s, 0.18s, 0.15s). The break-even points were 69, 116, and 145 iterations.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Pencil beam kernel convolution can be formulated as Conv2d spatial convolution with machine-precision equivalence to FFT and competitive speed. The spatial-domain formulation enables a differentiable, d<i>ij</i>-free optimization workflow that recomputes dose without storing a dose-influence matrix and that, for short optimization runs, is competitive with the traditional FFT-built sparse d<i>ij</i> pipeline.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121023","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":"End-to-end differentiable volumetric modulated arc therapy optimization via computation graph–driven automatic differentiation","authors":"Hongkai Wang,&nbsp;Wenzheng Sun,&nbsp;Wenlong Xia,&nbsp;Bin Liang","doi":"10.1002/mp.70685","DOIUrl":"10.1002/mp.70685","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;Conventional volumetric modulated arc therapy (VMAT) optimization relies on gradient-based methods in which analytical gradients must be explicitly derived for each objective function and machine constraint. This dedicated mathematical derivation imposes substantial barriers to extending the optimization framework with new clinical objectives or delivery constraints, and the resulting CPU-based implementations do not natively exploit modern GPU hardware.&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 an end-to-end differentiable VMAT optimization framework using computation graphs that enable automatic differentiation–based gradient computation for both fluence map optimization (FMO) and direct aperture optimization (DAO), while incorporating VMAT machine delivery constraints within a unified, GPU-accelerated pipeline.&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 two-stage pipeline—fluence map optimization (FMO) followed by DAO—was developed. FMO reformulates dose calculation as a differentiable computation graph, enabling gradient computation for arbitrary objective functions including non-smooth Heaviside step terms. For DAO, a novel analytically derived computation graph maps MLC leaf positions to delivered fluence through closed-form time-averaged bixel exposure integration with finite-width boundary corrections, ensuring exact differentiability. Machine delivery constraints—including leaf collision, speed limits, and dose rate bounds—are incorporated as differentiable penalty terms for joint end-to-end optimization. Twenty lung cancer cases from the GDP-HMM AAPM Challenge dataset (60 Gy/30 fractions) were optimized using objectives derived from clinical Eclipse reference plans and compared against MatRad and Eclipse.&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 proposed method achieved PTV D98 of 58.88 ± 0.33 Gy, comparable to Eclipse (59.07 ± 0.55 Gy, &lt;i&gt;p =&lt;/i&gt; 0.26) and significantly higher than MatRad (57.27 ± 0.48 Gy, &lt;i&gt;p &lt;&lt;/i&gt; 0.001). The Paddick conformity index was 0.81 ± 0.06, comparable to Eclipse (0.84 ± 0.22, &lt;i&gt;p =&lt;/i&gt; 0.48) and superior to MatRad (0.64 ± 0.11, &lt;i&gt;p &lt;&lt;/i&gt; 0.001). OAR sparing closely matched Eclipse: heart mean dose 10.74 ± 4.89 Gy vs. 11.55 ± 4.76 Gy; lung mean dose 15.33 ± 3.50 Gy vs. 16.70 ± 3.29 Gy; LAD mean dose 9.55 ± 4.97 Gy vs. 9.05 ± 3.94 Gy (&lt;i&gt;p =&lt;/i&gt; 0.18). MatRad achieved lower heart doses at the cost of degraded PTV coverage and conformity. Optimization time was significantly shorter (59.28 ± 43.95 s vs. 390.30 ± 184.90 s, &lt;i&gt;p &lt;&lt;/i&gt; 0.001), while delivery times were equivalent (&lt;i&gt;p =&lt;/i&gt; 0.92).&lt;/p&gt;\u0000 &lt;/se","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"53 10","pages":""},"PeriodicalIF":3.2,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149121075","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}