Medical physics最新文献

{"title":"4D Dynamic contrast-enhanced breast CT: Phantom-based reconstruction parameter optimization for iodine quantification","authors":"Mikhail Mikerov, Juan J. Pautasso, Liselot Goris, Koen Michielsen, Ioannis Sechopoulos","doi":"10.1002/mp.17658","DOIUrl":"10.1002/mp.17658","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Four-dimensional dynamic contrast-enhanced breast CT (4D DCE-bCT) offers promising high-resolution spatial and temporal imaging capabilities for the characterization and monitoring of breast tumors. However, the optimal combination of parameters for iodine quantification in image space remains to be determined.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study aims to optimize a dedicated bCT system to perform long dynamic contrast-enhanced scans with high spatio-temporal resolution while maintaining a reasonable radiation dose.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Our protocol includes the acquisition of a high-quality prior image that is reconstructed with a polychromatic iterative algorithm (IMPACT). The acquisition of the post-contrast sequence is continuous but sparse and these images are reconstructed using prior image constrained compressed sensing (PICCS). A four-step optimization process is performed using images of a physical phantom. First, the optimal tube current is selected by taking the noise level into account. Second, the optimal number of angles is selected based on the absence of streak artifacts. Third, the number of iterations in IMPACT is specified at the lowest value that achieves the highest spatial resolution. Finally, the number of iterations in PICCS is determined based on the quantitative accuracy of a range of iodine concentrations.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>When a high-quality prior image is available, the imaging of post-contrast images can be performed using just 40 projection angles with a tube current of 32 mA. The noise level in the post-contrast images is inherited from the prior image and no streak artifacts are visible. Mean difference between the linear attenuation coefficients of samples containing iodine reconstructed with IMPACT using all 360 projections and PICCS using 40 projections is 0.0004 <span></span><math>\u0000 <semantics>\u0000 <msup>\u0000 <mi>mm</mi>\u0000 <mrow>\u0000 <mo>−</mo>\u0000 <mn>1</mn>\u0000 </mrow>\u0000 </msup>\u0000 <annotation>$mathrm{mm}^{-1}$</annotation>\u0000 </semantics></math> at most. The spatial resolution of images reconstructed with PICCS is lower than the one of IMPACT images and is concentration dependent. The cut-off frequency at 10% modulation transfer function drops from 1.55 <span></span><math>\u0000 ","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 4","pages":"2212-2223"},"PeriodicalIF":3.2,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17658","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143124194","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":"Multi-scale dual attention embedded U-shaped network for accurate segmentation of coronary vessels in digital subtraction angiography","authors":"He Deng,&nbsp;Yuqing Li,&nbsp;Xu Liu,&nbsp;Kai Cheng,&nbsp;Tong Fang,&nbsp;Xiangde Min","doi":"10.1002/mp.17618","DOIUrl":"10.1002/mp.17618","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Most attention-based networks fall short in effectively integrating spatial and channel-wise information across different scales, which results in suboptimal performance for segmenting coronary vessels in x-ray digital subtraction angiography (DSA) images. This limitation becomes particularly evident when attempting to identify tiny sub-branches.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>To address this limitation, a multi-scale dual attention embedded network (named MDA-Net) is proposed to consolidate contextual spatial and channel information across contiguous levels and scales.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>MDA-Net employs five cascaded double-convolution blocks within its encoder to adeptly extract multi-scale features. It incorporates skip connections that facilitate the retention of low-level feature details throughout the decoding phase, thereby enhancing the reconstruction of detailed image information. Furthermore, MDA modules, which take in features from neighboring scales and hierarchical levels, are tasked with discerning subtle distinctions between foreground elements, such as coronary vessels of diverse morphologies and dimensions, and the complex background, which includes structures like catheters or other tissues with analogous intensities. To sharpen the segmentation accuracy, the network utilizes a composite loss function that integrates intersection over union (IoU) loss with binary cross-entropy loss, ensuring the precision of the segmentation outcomes and maintaining an equilibrium between positive and negative classifications.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Experimental results demonstrate that MDA-Net not only performs more robustly and effectively on DSA images under various image conditions, but also achieves significant advantages over state-of-the-art methods, achieving the optimal scores in terms of IoU, Dice, accuracy, and Hausdorff distance 95%.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>MDA-Net has high robustness for coronary vessels segmentation, providing an active strategy for early diagnosis of cardiovascular diseases. The code is publicly available at https://github.com/30410B/MDA-Net.git.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 5","pages":"3135-3150"},"PeriodicalIF":3.2,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143082926","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":"Optimization of sparse-view CT reconstruction based on convolutional neural network","authors":"Liangliang Lv,&nbsp;Chang Li,&nbsp;Wenjing Wei,&nbsp;Shuyi Sun,&nbsp;Xiaoxuan Ren,&nbsp;Xiaodong Pan,&nbsp;Gongping Li","doi":"10.1002/mp.17636","DOIUrl":"10.1002/mp.17636","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;Sparse-view CT shortens scan time and reduces radiation dose but results in severe streak artifacts due to insufficient sampling data. Deep learning methods can now suppress these artifacts and improve image quality in sparse-view CT 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;The quality of sparse-view CT reconstructed images can still be improved. Additionally, the interpretability of deep learning-based optimization methods for these reconstruction images is lacking, and the role of different network layers in artifact removal requires further study. Moreover, the optimization capability of these methods for reconstruction images from various sparse views needs enhancement. This study aims to improve the network's optimization ability for sparse-view reconstructed images, enhance interpretability, and boost generalization by establishing multiple network structures and datasets.&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;In this paper, we developed a sparse-view CT reconstruction images improvement network (SRII-Net) based on U-Net. We added a copy pathway in the network and designed a residual image output block to boost the network's performance. Multiple networks with different connectivity structures were established using SRII-Net to analyze the contribution of each layer to artifact removal, improving the network's interpretability. Additionally, we created multiple datasets with reconstructed images of various sampling views to train and test the proposed network, investigating how these datasets from different sampling views affect the network's generalization ability.&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 results show that the proposed method outperforms current networks, with significant improvements in metrics like PSNR and SSIM. Image optimization time is at the millisecond level. By comparing the performance of different network structures, we've identified the impact of various hierarchical structures. The image detail information learned by shallow layers and the high-level abstract feature information learned by deep layers play a crucial role in optimizing sparse-view CT reconstruction images. Training the network with multiple mixed datasets revealed that, under a certain amount of data, selecting the appropriate categories of sampling views and their corresponding samples can effectively enhance the network's optimization ability for reconstructing images with different sampling views.&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;The network in this pap","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 4","pages":"2089-2105"},"PeriodicalIF":3.2,"publicationDate":"2025-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143082881","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":"Cover","authors":"","doi":"10.1002/mp.17198","DOIUrl":"https://doi.org/10.1002/mp.17198","url":null,"abstract":"","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"C1"},"PeriodicalIF":3.2,"publicationDate":"2025-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17198","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143110443","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":"List of Advertisers","authors":"","doi":"10.1002/mp.17199","DOIUrl":"https://doi.org/10.1002/mp.17199","url":null,"abstract":"","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"1350"},"PeriodicalIF":3.2,"publicationDate":"2025-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17199","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143110425","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":"Automating the optimization of proton PBS treatment planning for head and neck cancers using policy gradient-based deep reinforcement learning","authors":"Qingqing Wang,&nbsp;Chang Chang","doi":"10.1002/mp.17654","DOIUrl":"10.1002/mp.17654","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 pencil beam scanning (PBS) treatment planning for head and neck (H&amp;N) cancers is a time-consuming and experience-demanding task where a large number of potentially conflicting planning objectives are involved. Deep reinforcement learning (DRL) has recently been introduced to the planning processes of intensity-modulated radiation therapy (IMRT) and brachytherapy for prostate, lung, and cervical cancers. However, existing DRL planning models are built upon the Q-learning framework and rely on weighted linear combinations of clinical metrics for reward calculation. These approaches suffer from poor scalability and flexibility, that is, they are only capable of adjusting a limited number of planning objectives in discrete action spaces and therefore fail to generalize to more complex planning problems.&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;Here we propose an automatic treatment planning model using the proximal policy optimization (PPO) algorithm in the policy gradient framework of DRL and a dose distribution-based reward function for proton PBS treatment planning of H&amp;N cancers.&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 planning process is formulated as an optimization problem. A set of empirical rules is used to create auxiliary planning structures from target volumes and organs-at-risk (OARs), along with their associated planning objectives. Special attention is given to overlapping structures with potentially conflicting objectives. These planning objectives are fed into an in-house optimization engine to generate the spot monitor unit (MU) values. A decision-making policy network trained using PPO is developed to iteratively adjust the involved planning objective parameters. The policy network predicts actions in a continuous action space and guides the treatment planning system to refine the PBS treatment plans using a novel dose distribution-based reward function. A total of 34 H&amp;N patients (30 for training and 4 for test) and 26 liver patients (20 for training, 6 for test) are included in this study to train and verify the effectiveness and generalizability of the proposed method.&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;Proton H&amp;N treatment plans generated by the model show improved OAR sparing with equal or superior target coverage when compared with human-generated plans. Moreover, additional experiments on liver cancer demonstrate that the proposed method can be successfully generalized to other treatment sites.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;h3&gt; Conclusions&lt;/h3&gt;\u0000 \u0000 &lt;","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 4","pages":"1997-2014"},"PeriodicalIF":3.2,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143069544","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":"Characterization of an MR-compatible motion platform for quality assurance of motion-compensated treatments on the 1.5 T MR-linac","authors":"Stijn Oolbekkink,&nbsp;Pim T. S. Borman,&nbsp;Jochem W. H. Wolthaus,&nbsp;Bram van Asselen,&nbsp;Astrid L. H. M. W. van Lier,&nbsp;Stephanie Dunn,&nbsp;Grant R. Koenig,&nbsp;Nick Hartman,&nbsp;Niusha Kheirkhah,&nbsp;Bas W. Raaymakers,&nbsp;Martin F. Fast","doi":"10.1002/mp.17632","DOIUrl":"10.1002/mp.17632","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Novel motion-compensated treatment techniques on the MR-linac can address adverse intra-fraction motion effects. These techniques involve beam gating or intra-fraction adaptations of the treatment plan based on real-time magnetic resonance imaging (MRI) performed during treatment. For quality assurance (QA) of these workflows, a multi-purpose motion platform is desirable. This platform should accommodate various phantoms, enabling multiple QA workflows.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study aims to evaluate the new IBA QUASAR Motion MR Platform for use in the 1.5 T MR-linac.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The motion platform was assessed for several magnetic resonance (MR) characteristics, including spurious noise generation and B0&amp;B1 homogeneity. In addition, the motion platform's motion accuracy and beam attenuation were assessed. An application was shown with a ScandiDos Delta4 Phantom+ MR demonstrating patient-specific plan QA of gated treatments using time-resolved dosimetry that includes motion based on a patient's respiratory motion trace.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>All MR characterization measurements were within the set tolerances for MRI QA. The motion platform motion accuracy showed excellent agreement with the reference, with a standard deviation of the amplitude of 0.01  mm (20 kg load) for the motor's self-estimated positions and 0.22 mm (no load) for the images acquired with the electronic portal imager. Beam attenuation was found to be 11.8%. The combination of the motion platform and Delta4 demonstrated motion-included dosimetry at high temporal and spatial resolutions. Motion influenced the measured dose in non-gated treatments by up to −20.1%, while gated deliveries showed differences of up to −1.7% for selected diodes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>The motion platform was found to be usable in a 1.5 T magnetic field, and for all MR characterization experiments, no influence from the motion platform was observed. This motion platform enables to perform motion-included QA, with a measurement device of choice.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 5","pages":"3391-3397"},"PeriodicalIF":3.2,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17632","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143069831","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":"TransAnaNet: Transformer-based anatomy change prediction network for head and neck cancer radiotherapy","authors":"Meixu Chen,&nbsp;Kai Wang,&nbsp;Michael Dohopolski,&nbsp;Howard Morgan,&nbsp;David Sher,&nbsp;Jing Wang","doi":"10.1002/mp.17655","DOIUrl":"10.1002/mp.17655","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;Adaptive radiotherapy (ART) can compensate for the dosimetric impact of anatomic change during radiotherapy of head–neck cancer (HNC) patients. However, implementing ART universally poses challenges in clinical workflow and resource allocation, given the variability in patient response and the constraints of available resources. Therefore, the prediction of anatomical change during radiotherapy for HNC patients is of importance to optimize patient clinical benefit and treatment resources. Current studies focus on developing binary ART eligibility classification models to identify patients who would experience significant anatomical change, but these models lack the ability to present the complex patterns and variations in anatomical changes over time. Vision Transformers (ViTs) represent a recent advancement in neural network architectures, utilizing self-attention mechanisms to process image data. Unlike traditional Convolutional Neural Networks (CNNs), ViTs can capture global contextual information more effectively, making them well-suited for image analysis and image generation tasks that involve complex patterns and structures, such as predicting anatomical changes in medical imaging.&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 purpose of this study is to assess the feasibility of using a ViT-based neural network to predict radiotherapy-induced anatomic change of HNC patients.&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 retrospectively included 121 HNC patients treated with definitive chemoradiotherapy (CRT) or radiation alone. We collected the planning computed tomography image (pCT), planned dose, cone beam computed tomography images (CBCTs) acquired at the initial treatment (CBCT01) and Fraction 21 (CBCT21), and primary tumor volume (GTVp) and involved nodal volume (GTVn) delineated on both pCT and CBCTs of each patient for model construction and evaluation. A UNet-style Swin-Transformer-based ViT network was designed to learn the spatial correspondence and contextual information from embedded image patches of CT, dose, CBCT01, GTVp, and GTVn. The deformation vector field between CBCT01 and CBCT21 was estimated by the model as the prediction of anatomic change, and deformed CBCT01 was used as the prediction of CBCT21. We also generated binary masks of GTVp, GTVn, and patient body for volumetric change evaluation. We used data from 101 patients for training and validation, and the remaining 20 patients for testing. Image and volumetric similarity metrics including mean square error (MSE), peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), Dice coefficient, and average surf","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 5","pages":"3015-3029"},"PeriodicalIF":3.2,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17655","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143070358","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":"Hybrid transformer-based model for mammogram classification by integrating prior and current images","authors":"Afsana Ahsan Jeny,&nbsp;Sahand Hamzehei,&nbsp;Annie Jin,&nbsp;Stephen Andrew Baker,&nbsp;Tucker Van Rathe,&nbsp;Jun Bai,&nbsp;Clifford Yang,&nbsp;Sheida Nabavi","doi":"10.1002/mp.17650","DOIUrl":"10.1002/mp.17650","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;Breast cancer screening via mammography plays a crucial role in early detection, significantly impacting women's health outcomes worldwide. However, the manual analysis of mammographic images is time-consuming and requires specialized expertise, presenting substantial challenges in medical practice.&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 address these challenges, we introduce a CNN-Transformer based model tailored for breast cancer classification through mammographic analysis. This model leverages both prior and current images to monitor temporal changes, aiming to enhance the efficiency and accuracy (ACC) of computer-aided diagnosis systems by mimicking the detailed examination process of radiologists.&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;In this study, our proposed model incorporates a novel integration of a position-wise feedforward network and multi-head self-attention, enabling it to detect abnormal or cancerous changes in mammograms over time. Additionally, the model employs positional encoding and channel attention methods to accurately highlight critical spatial features, thus precisely differentiating between normal and cancerous tissues. Our methodology utilizes focal loss (FL) to precisely address challenging instances that are difficult to classify, reducing false negatives and false positives to improve diagnostic ACC.&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;We compared our model with eight baseline models; specifically, we utilized only current images for the single model ResNet50 while employing both prior and current images for the remaining models in terms of accuracy (ACC), sensitivity (SEN), precision (PRE), specificity (SPE), F1 score, and area under the curve (AUC). The results demonstrate that the proposed model outperforms the baseline models, achieving an ACC of 90.80%, SEN of 90.80%, PRE of 90.80%, SPE of 90.88%, an F1 score of 90.95%, and an AUC of 92.58%. The codes and related information are available at https://github.com/NabaviLab/PCTM.&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;Our proposed CNN-Transformer model integrates both prior and current images, removes long-range dependencies, and enhances its capability for nuanced classification. The application of FL reduces false positive rate (FPR) and false negative rates (FNR), improving both SEN and SPE. Furthermore, the model achieves the lowest false discovery rate and FNR across various abnormalities, including masses, calcification, and architectural distortions (ADs). These low error rates highl","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 5","pages":"2999-3014"},"PeriodicalIF":3.2,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143070147","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":"Multilevel perception boundary-guided network for breast lesion segmentation in ultrasound images","authors":"Xing Yang,&nbsp;Jian Zhang,&nbsp;Yingfeng OU,&nbsp;Qijian Chen,&nbsp;Li Wang,&nbsp;Lihui Wang","doi":"10.1002/mp.17647","DOIUrl":"10.1002/mp.17647","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;Automatic segmentation of breast tumors from the ultrasound images is essential for the subsequent clinical diagnosis and treatment plan. Although the existing deep learning-based methods have achieved considerable progress in automatic segmentation of breast tumors, their performance on tumors with similar intensity to the normal tissues is still not satisfactory, especially for the tumor boundaries.&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 accurately segment the non-enhanced lesions with more accurate boundaries, a novel multilevel perception boundary-guided network (PBNet) is proposed to segment breast tumors from ultrasound images.&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;PBNet consists of a multilevel global perception module (MGPM) and a boundary guided module (BGM). MGPM models long-range spatial dependencies by fusing both intra- and inter-level semantic information to enhance tumor recognition. In BGM, the tumor boundaries are extracted from the high-level semantic maps using the dilation and erosion effects of max pooling; such boundaries are then used to guide the fusion of low- and high-level features. Additionally, a multi-level boundary-enhanced segmentation (BS) loss is introduced to improve boundary segmentation performance. To evaluate the effectiveness of the proposed method, we compared it with state-of-the-art methods on two datasets, one publicly available datasets BUSI containing 780 images and one in-house dataset containing 995 images. To verify the robustness of each method, a 5-fold cross-validation method was used to train and test the models. Dice score (Dice), Jaccard coefficients (Jac), Hausdorff Distance (HD), Sensitivity (Sen), and specificity(Spe) were used to evaluate the segmentation performance quantitatively. The Wilcoxon test and Benjamini-Hochberg false discovery rate (FDR) multi-comparison correction were then performed to assess whether the proposed method presents statistically significant performance (&lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mrow&gt;\u0000 &lt;mi&gt;p&lt;/mi&gt;\u0000 &lt;mo&gt;≤&lt;/mo&gt;\u0000 &lt;mn&gt;0.05&lt;/mn&gt;\u0000 &lt;/mrow&gt;\u0000 &lt;annotation&gt;$ple 0.05$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt;) difference comparing with existing methods. In addition, to comprehensively demonstrate the difference between different methods, the Cohen's d effect size and compound &lt;i&gt;p&lt;/i&gt;-value (c-Pvalue) obtained with Fisher's method were also calculated.&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;On the BUSI dataset, the mean Dice and Sen ","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 5","pages":"3117-3134"},"PeriodicalIF":3.2,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143070353","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}