Medical physics最新文献

{"title":"Design, manufacturing, and multi-modal imaging of stereolithography 3D printed flexible intracranial aneurysm phantoms","authors":"Alain Yalman,&nbsp;Arman Jafari,&nbsp;Étienne Léger,&nbsp;Michael-Anthony Mastroianni,&nbsp;Kowsar Teimouri,&nbsp;Houman Savoji,&nbsp;D. Louis Collins,&nbsp;Lyes Kadem,&nbsp;Yiming Xiao","doi":"10.1002/mp.17518","DOIUrl":"10.1002/mp.17518","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Physical vascular phantoms are instrumental in studying intracranial aneurysms and testing relevant imaging tools and training systems to provide improved clinical care. Current vascular phantom production methods have major limitations in capturing the biophysical and morphological characteristics of intracranial aneurysms with good fidelity and multi-modal imaging capacity. With stereolithography (SLA) 3D printing technology becoming more accessible, newer flexible and transparent printing materials with higher precision controls open the door for improving the efficiency and quality of producing anthropomorphic vascular phantoms but have rarely been explored for the application.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This technical note intends to report the feasibility of using SLA 3D printing technology to manufacture flexible intracranial aneurysm phantoms with similar scales to the real anatomy, as well as their capacity for multi-modal flow imaging and analysis, including ultrasound flow imaging, high-speed filming, and particle image velocimetry analysis.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We designed and 3D-printed two intracranial aneurysm phantoms with an SLA 3D printer using Formlabs Elastic 50A resin. By using a micropump to introduce cyclical flows in the phantoms, we first employed conventional Doppler and vector flow ultrasonography to observe and measure different fluidic properties. Then, a high-speed camera was used to record particles flowing within the phantom, and we further conducted a particle image velocimetry analysis, including the distribution of mean 2D velocity vectors, average velocity magnitudes, and the mean vorticity fields in the phantom for the high-speed imaging data.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>We successfully 3D-printed flexible intracranial aneurysm phantoms with similar dimensions to the real anatomy. Additionally, we validated the phantoms’ ability to allow visualization, measurement, and analysis of flow dynamics based on both real-time ultrasound and optical imaging.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Our proof-of-concept study illustrates that SLA 3D printing using commercial elastic resins can significantly contribute towards facilitating the fabrication of flexible intracranial aneurysms phantoms for training, research, and preoperative planning.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"742-749"},"PeriodicalIF":3.2,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17518","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142640302","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":"Open-chest cardiac ultrasound-mediated imaging with a vacuum coupler","authors":"Nilesh Mathuria,&nbsp;Krithik Vishwanath,&nbsp;Giorgio Brero,&nbsp;Blake C. Fallon,&nbsp;Antonio Martino,&nbsp;Richard C. Willson,&nbsp;Carly S. Filgueira,&nbsp;Richard R. Bouchard","doi":"10.1002/mp.17511","DOIUrl":"10.1002/mp.17511","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>A fundamental obstacle for the preclinical development of ultrasound-(US) mediated cardiac imaging remains cardiac motion, which limits interframe correlation during extended acquisition periods.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>To address this need, we present the design and implementation of a 3D-printed vacuum coupler that stabilizes a US transducer on the epicardial surface of the heart for feasibility assessment and development of advanced, cardiac, US-mediated imaging approaches.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The vacuum coupler was 3D printed with biocompatible resins and secured with a standard intraoperative suction aspirator. US-mediated imaging (i.e., B-mode and photoacoustic [PA] imaging) was performed in an open-chest porcine model with and without the vacuum coupler. Based on inter-frame displacement tracking and cross-correlation (CC) coefficients, changes in frame motion and stability were compared for each imaging mode/configuration through a prolonged (∼1 min) acquisition, while the impact on PA-based SO₂ accuracy was assessed.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>When compared to conventional handheld imaging, stand-off imaging, and coupler without suction, epicardial imaging with the vacuum coupler and suction applied led to a significantly reduced mean axial displacement of 0.15 mm versus 0.89, 0.49, &amp; 0.49 mm, respectively (<i>p</i>-values ≤ 8.65e-7). Comparing the coupler without suction to that with suction applied, physiologically unrealistic SO₂ estimates reduced from 1.72 to 0.81%, respectively, and lateral interframe displacement reduced from 4.58 to 2.01 mm, respectively (<i>p</i>-value = 5.07e-23). Overall, reduced cardiac tissue motion and increased interframe CC coefficient (baseline = 0.43 vs. coupler with suction = 0.80) allow for more accurate PA unmixing.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Epicardial US-mediated imaging with a vacuum coupler reduces cardiac motion artifact, providing a consistent sampling of an intended region of interest (ROI) over multiple cardiac cycles. This could help facilitate the development of advanced US-mediated imaging, which is often hindered by cardiac motion. Stable implementation of these imaging techniques could allow for intra-operative assessments of local cardiac perfusion as well as tissue characterization.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"880-888"},"PeriodicalIF":3.2,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17511","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142640264","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":"4DCT image artifact detection using deep learning","authors":"Joshua W. Carrizales,&nbsp;Mattison J. Flakus,&nbsp;Dallin Fairbourn,&nbsp;Wei Shao,&nbsp;Sarah E. Gerard,&nbsp;John E. Bayouth,&nbsp;Gary E. Christensen,&nbsp;Joseph M. Reinhardt","doi":"10.1002/mp.17513","DOIUrl":"10.1002/mp.17513","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Four-dimensional computed tomography (4DCT) is an es sential tool in radiation therapy. However, the 4D acquisition process may cause motion artifacts which can obscure anatomy and distort functional measurements from CT scans.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>We describe a deep learning algorithm to identify the location of artifacts within 4DCT images. Our method is flexible enough to handle different types of artifacts, including duplication, misalignment, truncation, and interpolation.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We trained and validated a U-net convolutional neural network artifact detection model on more than 23 000 coronal slices extracted from 98 4DCT scans. The receiver operating characteristic (ROC) curve and precision-recall curve were used to evaluate the model's performance at identifying artifacts compared to a manually identified ground truth. The model was adjusted so that the sensitivity in identifying artifacts was equivalent to that of a human observer, as measured by computing the average ratio of artifact volume to lung volume in a given scan.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The model achieved a sensitivity, specificity, and precision of 0.78, 0.99, and 0.58, respectively. The ROC area-under-the-curve (AUC) was 0.99 and the precision-recall AUC was 0.73. Our model sensitivity is 8% higher than previously reported state-of-the-art artifact detection methods.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The model developed in this study is versatile, designed to handle duplication, misalignment, truncation, and interpolation artifacts within a single image, unlike earlier models that were designed for a single artifact type.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"1096-1107"},"PeriodicalIF":3.2,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17513","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142634676","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":"Radiomics-guided generative adversarial network for automatic primary target volume segmentation for nasopharyngeal carcinoma using computed tomography images","authors":"Juebin Jin,&nbsp;Jicheng Zhang,&nbsp;Xianwen Yu,&nbsp;Ziqing Xiang,&nbsp;Xuanxuan Zhu,&nbsp;Mingrou Guo,&nbsp;Zeshuo Zhao,&nbsp;WenLong Li,&nbsp;Heng Li,&nbsp;Jiayi Xu,&nbsp;Xiance Jin","doi":"10.1002/mp.17493","DOIUrl":"10.1002/mp.17493","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Automatic primary gross tumor volume (GTVp) segmentation for nasopharyngeal carcinoma (NPC) is a quite challenging task because of the existence of similar visual characteristics between tumors and their surroundings, especially on computed tomography (CT) images with severe low contrast resolution. Therefore, most recently proposed methods based on radiomics or deep learning (DL) is difficult to achieve good results on CT datasets.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>A peritumoral radiomics-guided generative adversarial network (PRG-GAN) was proposed to address this challenge.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A total of 157 NPC patients with CT images was collected and divided into training, validation, and testing cohorts of 108, 9, and 30 patients, respectively. The proposed model was based on a standard GAN consisting of a generator network and a discriminator network. Morphological dilation on the initial segmentation results from GAN was first conducted to delineate annular peritumoral region, in which radiomics features were extracted as priori guide knowledge. Then, radiomics features were fused with semantic features by the discriminator's fully connected layer to achieve the voxel-level classification and segmentation. The dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), and average symmetric surface distance (ASSD) were used to evaluate the segmentation performance using a paired samples <i>t</i>-test with Bonferroni correction and Cohen's d (<i>d</i>) effect sizes. A two-sided <i>p</i>-value of less than 0.05 was considered statistically significant.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The model-generated predictions had a high overlap ratio with the ground truth. The average DSC, HD95, and ASSD were significantly improved from 0.80 ± 0.12, 4.65 ± 4.71  mm, and 1.35 ± 1.15  mm of GAN to 0.85 ± 0.18 (<i>p</i> = 0.001, <i>d</i> = 0.71), 4.15 ± 7.56 mm (<i>p</i> = 0.002, <i>d</i> = 0.67), and 1.11 ± 1.65 mm (<i>p</i> &lt; 0.001, <i>d</i> = 0.46) of PRG-GAN, respectively.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>Integrating radiomics features into GAN is promising to solve unclear border limitations and increase the delineation accuracy of GTVp for patients with NPC.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"1119-1132"},"PeriodicalIF":3.2,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142634705","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":"Radiomics analysis of dual-energy CT-derived iodine maps for differentiating malignant from benign thyroid nodules","authors":"Ni Liu,&nbsp;Zengfa Huang,&nbsp;Jun Chen,&nbsp;Yang Yang,&nbsp;Zuoqin Li,&nbsp;Yuanzhi Liu,&nbsp;Yuanliang Xie,&nbsp;Xiang Wang","doi":"10.1002/mp.17510","DOIUrl":"10.1002/mp.17510","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;Many thyroid nodules are detected incidentally with the widespread use of sensitive imaging techniques; however, only a fraction of these nodules are malignant, resulting in unnecessary medical expenditures and anxiety. The major challenge is to differentiate benign thyroid nodules from malignant ones. The application of dual-energy computed tomography (DECT) and radiomics provides a new diagnostic approach. Studies applying radiomics from primary tumours on iodine maps to differentiate malignant from benign thyroid nodules are still lacking.&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 determine the ability of an iodine map-based radiomic nomogram in the venous phase for differentiating malignant thyroid nodules from benign nodules.&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 total of 141 patients with thyroid nodules who underwent DECT were enrolled and randomly assigned to the training and test cohorts between January 2018 and January 2019. The radiomic score (Rad-score) was derived from nine quantitative features of the iodine maps. Stepwise logistic regression analysis was used to develop radiomic, clinical and combined models. Age, normalized iodine concentration (NIC), and cyst changes were used to construct the clinical model. Receiver operating characteristic (ROC) curve analysis, sensitivity and specificity were performed to analyse the ability of the models to predict malignant thyroid nodules. Calibration analysis was used to test the fitness of the models. Decision curve analysis (DCA) and nomogram construction 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;According to the clinical model, age (0.989 [0.984, 0.995]; &lt;i&gt;p &lt;/i&gt;&lt; 0.001), NIC (0.778 [0.640, 0.995]; &lt;i&gt;p &lt;/i&gt;= 0.01), and cyst changes (0.617 [0.507, 0.751]; &lt;i&gt;p &lt;/i&gt;&lt; 0.001) were independently associated with malignant thyroid nodules. According to the combined model, age (0.994 [0.989, 0.999]; &lt;i&gt;p &lt;/i&gt;= 0.01), NIC (0.797 [0.674, 0.941]; &lt;i&gt;p &lt;/i&gt;= 0.008), cyst changes (0.786 [0.653, 0.947]; &lt;i&gt;p &lt;/i&gt;= 0.01), and the rad-score (1.106 [1.070, 1.143]; &lt;i&gt;p &lt;/i&gt;&lt; 0.001) were independently associated with malignant thyroid nodules. The combined model achieved satisfactory discrimination in predicting malignant thyroid nodules and had greater predictive value in the training (AUC [areas under the curve], 0.96 vs. 0.87; &lt;i&gt;p &lt;/i&gt;= 0.01) and test (AUC, 0.90 vs. 0.79; &lt;i&gt;p &lt;/i&gt;= 0.04) cohorts than did the clinical model.&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 radiomics nomogram based on iodine ","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"826-836"},"PeriodicalIF":3.2,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142634701","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":"Imaging lung tumor motion using integrated-mode proton radiography—A phantom study towards tumor tracking in proton radiotherapy","authors":"Ryan Fullarton,&nbsp;Mikaël Simard,&nbsp;Lennart Volz,&nbsp;Allison Toltz,&nbsp;Savanna Chung,&nbsp;Christoph Schuy,&nbsp;Daniel G. Robertson,&nbsp;Gary Royle,&nbsp;Sam Beddar,&nbsp;Colin Baker,&nbsp;Christian Graeff,&nbsp;Charles-Antoine Collins-Fekete","doi":"10.1002/mp.17508","DOIUrl":"10.1002/mp.17508","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;Motion of lung tumors during radiotherapy leads to decreased accuracy of the delivered dose distribution. This is especially true for proton radiotherapy due to the finite range of the proton beam. Methods for mitigating motion rely on knowing the position of the tumor during treatment.&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;Proton radiography uses the treatment beam, at an energy high enough to traverse the patient, to produce a radiograph. This work shows the first results of using an integrated-mode proton radiography system to track the position of moving objects in an experimental phantom study; demonstrating the potential of using this method for measuring tumor 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;Proton radiographs of an anthropomorphic lung phantom, with a motor-driven tumor insert, were acquired approximately every 1 s, using tumor inserts of 10, 20, and 30 mm undergoing a known periodic motion. The proton radiography system used a monolithic scintillator block and digital cameras to capture the residual range of each pencil beam passing through the phantom. These ranges were then used to produce a water equivalent thickness map of the phantom. The centroid of the tumor insert in the radiographs was used to determine its position. This measured position was then compared to the known motion of the phantom to determine the accuracy.&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;Submillimeter accuracy on the measurement of the tumor insert was achieved when using a 30 mm tumor insert with a period of 24 s and was found to be improved for decreasing motion amplitudes with a mean absolute error (MAE) of 1.0, 0.9, and 0.7 mm for 20, 15, and 10 mm respectively. Using smaller tumor inserts reduced the accuracy with a MAE of 1.8 and 1.9 mm for a 20 and 10 mm insert respectively undergoing a periodic motion with an amplitude of 20 mm and a period of 24 s. Using a shorter period resulted in significant motion artifacts reducing the accuracy to a MAE of 2.2 mm for a 12 s period and 3.1 mm for a 6 s period for the 30 mm insert with an amplitude of 20 mm.&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 work demonstrates that the position of a lung tumor insert in a realistic anthropomorphic phantom can be measured with high accuracy using proton radiographs. Results show that the accuracy of the position measurement is the highest for slower tumor motions due to a reduction in moti","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"1146-1158"},"PeriodicalIF":3.2,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17508","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142634679","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 contextual learning for medical image segmentation via dual distillation","authors":"Ruize Cui,&nbsp;Lanqing Liu,&nbsp;Youyi Song,&nbsp;Ge Ren,&nbsp;Xiaowei Hu,&nbsp;Jing Qin","doi":"10.1002/mp.17506","DOIUrl":"10.1002/mp.17506","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Recently, many studies have explored fusing features extracted from Convolutional neural networks (CNNs) and transformers to integrate multi-scale representations for better performance in medical image segmentation tasks. Although these hybrid models have achieved better results than previous CNN-based and transformer-based methods, they suffer from high computation and space complexities.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>The purpose of this research is to address the prohibitive computation and space complexities of hybrid models, which limit their application in clinical practice where computational resources are usually constrained.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We propose a novel model equipped with a dual distillation scheme to sufficiently harness the complementary advantages of CNNs and transformers without compromising model efficiency. We further propose a multi-scale prior-knowledge distillation (MPD) module to effectively distill multi-scale knowledge from features extracted from transformers. In addition, to cooperate with the knowledge distillation scheme, we also propose an efficient and robust Selective Fusion module in the student network.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>We extensively evaluate the proposed model against fourteen different network frameworks on two representative datasets: SipakMed and ISIC 2017. In the SipakMed dataset, 3037 Pap smear images are used for training and 1012 for testing. In the ISIC 2017 dataset, 2000 dermoscopic images are used for training, 150 for validation, and 600 for testing. Experimental results demonstrate that our method not only surpasses existing methods by a considerable margin with respect to the evaluation metrics of mean Intersection over Union, mean Dice coefficient, mean average symmetric surface distance, but also requires fewer computational resources in terms of model parameters and floating-point operations per second.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Comprehensive comparisons in terms of segmentation accuracy and computational complexity unequivocally confirm that our method effectively and efficiently integrates the advantages of both CNNs and transformers, showing its suitability and significance for clinical applications.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"787-800"},"PeriodicalIF":3.2,"publicationDate":"2024-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142634685","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 apparent diffusion coefficient map generation from multi-parametric MR images for patients with diffuse gliomas","authors":"Zach Eidex,&nbsp;Mojtaba Safari,&nbsp;Jacob Wynne,&nbsp;Richard L. J. Qiu,&nbsp;Tonghe Wang,&nbsp;David Viar-Hernandez,&nbsp;Hui-Kuo Shu,&nbsp;Hui Mao,&nbsp;Xiaofeng Yang","doi":"10.1002/mp.17509","DOIUrl":"10.1002/mp.17509","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>Apparent diffusion coefficient (ADC) maps derived from diffusion weighted magnetic resonance imaging (DWI MRI) provides functional measurements about the water molecules in tissues. However, DWI is time consuming and very susceptible to image artifacts, leading to inaccurate ADC measurements. This study aims to develop a deep learning framework to synthesize ADC maps from multi-parametric MR images.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We proposed the multiparametric residual vision transformer model (MPR-ViT) that leverages the long-range context of vision transformer (ViT) layers along with the precision of convolutional operators. Residual blocks throughout the network significantly increasing the representational power of the model. The MPR-ViT model was applied to T1w and T2-fluid attenuated inversion recovery images of 501 glioma cases from a publicly available dataset including preprocessed ADC maps. Selected patients were divided into training (<i>N</i> = 400), validation (<i>N</i> = 50), and test (<i>N</i> = 51) sets, respectively. Using the preprocessed ADC maps as ground truth, model performance was evaluated and compared against the Vision Convolutional Transformer (VCT) and residual vision transformer (ResViT) models with the peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and mean squared error (MSE).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The results are as follows using T1w + T2-FLAIR MRI as inputs: MPR-ViT—PSNR: 31.0 ± 2.1, MSE: 0.009 ± 0.0005, SSIM: 0.950 ± 0.015. In addition, ablation studies showed the relative impact on performance of each input sequence. Both qualitative and quantitative results indicate that the proposed MR-ViT model performs favorably against the ground truth data.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>We show that high-quality ADC maps can be synthesized from structural MRI using a MPR-ViT model. Our predicted images show better conformality to the ground truth volume than ResViT and VCT predictions. These high-quality synthetic ADC maps would be particularly useful for disease diagnosis and intervention, especially when ADC maps have artifacts or are unavailable.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 2","pages":"847-855"},"PeriodicalIF":3.2,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142607447","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":"3D in-system calibration method for PET detectors","authors":"Yannick Kuhl,&nbsp;Florian Mueller,&nbsp;Julian Thull,&nbsp;Stephan Naunheim,&nbsp;David Schug,&nbsp;Volkmar Schulz","doi":"10.1002/mp.17475","DOIUrl":"10.1002/mp.17475","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;Light-sharing detector designs for positron emission tomography (PET) systems have sparked interest in the scientific community. Particularly, (semi-)monoliths show generally good performance characteristics regarding 2D positioning, energy-, and timing resolution, as well as readout area. This is combined with intrinsic depth-of-interaction (DOI) capability to ensure a homogeneous spatial resolution across the entire field of view (FoV). However, complex positioning calibration processes limit their use in PET systems, especially in large-scale clinical systems.&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 proposes a new 3D positioning in-system calibration method for fast and convenient (re-)calibration and quality control of assembled PET scanners. The method targets all kinds of PET detectors that achieve the best performance with individual calibration, including complex segmented detector designs. The in-system calibration method is evaluated and empirically compared to a state-of-the-art fan-beam calibration for a small-diameter proof of concept (PoC) scanner. A simulation study evaluates the method's applicability to different scanner geometries.&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 PoC scanner geometry of 120 mm inner diameter and 150 mm axial extent was set up consisting of five identical finely segmented slab detectors (one detector under test and four collimation detectors). A &lt;sup&gt;2&lt;/sup&gt;&lt;sup&gt;2&lt;/sup&gt;Na point source was moved in a circular path inside the FoV. Utilizing virtual collimation and by selecting gamma rays incident approximately perpendicular to the detector normal of the detector under test, training data was created for the training of a 2D positioning model with the machine-learning technique gradient tree boosting (GTB). Data with oblique ray angles was acquired in the same measurement for subsequent angular DOI calibration. For this, a 2D position estimate in the detector under test was calculated first. On this basis, the DOI label was calculated geometrically from the ray path within the detector to finally establish up to 3D training data.&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 a mean absolute error (MAE) of 0.8  and 1.19 mm full-width at half maximum (FWHM) along the planar-monolithic slab dimension, the in-system methods performed similarly within 1% to the fan-beam collimator results. The DOI performance was at ∼90% with 1.13 mm MAE and 2.47 mm FWHM to the fan-beam collimator. Analytical calculations suggest an improved performance for larger","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 1","pages":"232-245"},"PeriodicalIF":3.2,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11699997/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142592321","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 vessel bifurcation liver CT landmark pair dataset for evaluating deformable image registration algorithms","authors":"Zhendong Zhang,&nbsp;Edward Robert Criscuolo,&nbsp;Yao Hao,&nbsp;Trevor McKeown,&nbsp;Deshan Yang","doi":"10.1002/mp.17507","DOIUrl":"10.1002/mp.17507","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>Evaluating deformable image registration (DIR) algorithms is vital for enhancing algorithm performance and gaining clinical acceptance. However, there is a notable lack of dependable DIR benchmark datasets for assessing DIR performance except for lung images. To address this gap, we aim to introduce our comprehensive liver computed tomography (CT) DIR landmark dataset library. This dataset is designed for efficient and quantitative evaluation of various DIR methods for liver CTs, paving the way for more accurate and reliable image registration techniques.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Acquisition and validation methods</h3>\u0000 \u0000 <p>Forty CT liver image pairs were acquired from several publicly available image archives and authors’ institutions under institutional review board (IRB) approval. The images were processed with a semi-automatic procedure to generate landmark pairs: (1) for each case, liver vessels were automatically segmented on one image; (2) landmarks were automatically detected at vessel bifurcations; (3) corresponding landmarks in the second image were placed using two deformable image registration methods to avoid algorithm-specific biases; (4) a comprehensive validation process based on quantitative evaluation and manual assessment was applied to reject outliers and ensure the landmarks’ positional accuracy. This workflow resulted in an average of ∼56 landmark pairs per image pair, comprising a total of 2220 landmarks for 40 cases. The general landmarking accuracy of this procedure was evaluated using digital phantoms and manual landmark placement. The landmark pair target registration errors (TRE) on digital phantoms were 0.37 ± 0.26 and 0.55 ± 0.34 mm respectively for the two selected DIR algorithms used in our workflow, with 97% of landmark pairs having TREs below 1.5 mm. The distances from the calculated landmarks to the averaged manual placement were 1.27 ± 0.79 mm.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data format and usage notes</h3>\u0000 \u0000 <p>All data, including image files and landmark information, are publicly available at Zenodo (https://zenodo.org/records/13738577). Instructions for using our data can be found on our GitHub page at https://github.com/deshanyang/Liver-DIR-QA.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Potential applications</h3>\u0000 \u0000 <p>The landmark dataset generated in this work is the first collection of large-scale liver CT DIR landmarks prepared on real patient images. This dataset can provide researchers with a dense set of ground truth benchmarks for the quantitative evaluation of DIR algorithms within the liver.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 1","pages":"703-715"},"PeriodicalIF":3.2,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142592325","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}