Medical physics最新文献

{"title":"Unsupervised Bayesian generation of synthetic CT from CBCT using patient-specific score-based prior","authors":"Junbo Peng, Yuan Gao, Chih-Wei Chang, Richard Qiu, Tonghe Wang, Aparna Kesarwala, Kailin Yang, Jacob Scott, David Yu, Xiaofeng Yang","doi":"10.1002/mp.17572","DOIUrl":"10.1002/mp.17572","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Cone-beam computed tomography (CBCT) scans, performed fractionally (e.g., daily or weekly), are widely utilized for patient alignment in the image-guided radiotherapy (IGRT) process, thereby making it a potential imaging modality for the implementation of adaptive radiotherapy (ART) protocols. Nonetheless, significant artifacts and incorrect Hounsfield unit (HU) values hinder their application in quantitative tasks such as target and organ segmentations and dose calculation. Therefore, acquiring CT-quality images from the CBCT scans is essential to implement online ART in clinical settings.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This work aims to develop an unsupervised learning method using the patient-specific diffusion model for CBCT-based synthetic CT (sCT) generation to improve the image quality of CBCT.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The proposed method is in an unsupervised framework that utilizes a patient-specific score-based model as the image prior alongside a customized total variation (TV) regularization to enforce coherence across different transverse slices. The score-based model is unconditionally trained using the same patient's planning CT (pCT) images to characterize the manifold of CT-quality images and capture the unique anatomical information of the specific patient. The efficacy of the proposed method was assessed on images from anatomical sites including head and neck (H&N) cancer, pancreatic cancer, and lung cancer. The performance of the proposed CBCT correction method was evaluated using quantitative metrics, including mean absolute error (MAE), non-uniformity (NU), and structural similarity index measure (SSIM). Additionally, the proposed algorithm was benchmarked against other unsupervised learning-based CBCT correction algorithms.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The proposed method significantly reduced various kinds of CBCT artifacts in the studies of H&N, pancreatic, and lung cancer patients. In the lung stereotactic body radiation therapy (SBRT) patient study, the MAE, NU, and SSIM were improved from 47 HU, 45 HU, and 0.58 in the original CBCT images to 13 HU, 14 dB, and 0.67 in the generated sCT images. Compared to other unsupervised learning-based algorithms, the proposed method demonstrated superior performance in artifact reduction.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The proposed unsupervised method can generate sCT from CBCT with reduced artifacts and precise HU values, enabling CBCT-guided segmentation and replannin","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 4","pages":"2238-2246"},"PeriodicalIF":3.2,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142819719","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":"Estimating the CTDIvol with helical acquisitions: Results from a national generalizability study","authors":"Izabella L. Barreto, Dustin A. Gress, Stephanie M. Leon, Bryan C. Schwarz, Robert J. Kobistek, M. Mahesh, James A. Tomlinson, Chad M. Dillon","doi":"10.1002/mp.17543","DOIUrl":"10.1002/mp.17543","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>While many clinical computed tomography (CT) protocols use helical scanning, the traditional method for measuring the volume CT Dose Index (CTDI<sub>vol</sub>) requires modifying the helical protocol to perform a single axial rotation. This modification can present challenges and mismatched settings across various scanner models.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study investigates the generalizability of a helical methodology for estimating CTDI<sub>vol</sub> across a diverse range of participants, CT scanner models, and protocol parameters.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A web-based platform collected axial and helical CTDI<sub>vol</sub> measurements from 24 medical physicists who submitted 569 data sets obtained using four CT protocols on scanners from seven CT manufacturers. Various parameters were tested for tube voltage (70–140 kVp), rotation time (0.25–1.50 s), beam width (8–80 mm), and pitch (0.29–3.0) settings. Measurements from the two methodologies were assessed for reproducibility using three repeated exposures and then compared to each other and to the scanner-displayed CTDI<sub>vol</sub>. Agreement between the methodologies was assessed using Bland–Altman analysis, linear regression, paired <i>t</i>-tests, and a paired two one-sided tests (TOST) procedure with equivalence margins of 5% of the mean protocol CTDI<sub>vol</sub>. The impact of beam width and pitch on measurement accuracy was assessed using linear regression analysis and an independent <i>t</i>-test.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>This study demonstrated better measurement reproducibility with the helical method (<i>p </i>< 0.05) and excellent concordance between helical and axial measurements (CCC > 0.99), with an average difference of -0.61 mGy (limits of agreement: -4.54 and 3.32). The TOST analysis confirmed that the measurement methods were statistically equivalent within the defined equivalence margins. The number of measurements that differed from the displayed CTDI<sub>vol</sub> by more than ± 20% was 10 for the axial method and 22 for the helical method. We did not identify a linear correlation between measurement accuracy and beam width or pitch (<i>R</i><sup>2 </sup>< 0.06). However, differences between axial and helical methods were significant for protocols with beam widths up to 40 mm versus those greater than 40 mm, as well as for protocols with pitch factors up to 1.0 compared to those greater than 1.0 (<i>p </i>< 0.001).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 ","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1823-1832"},"PeriodicalIF":3.2,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142808977","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":"Predicting the high intensity focused ultrasound focus in vivo using acoustic radiation force imaging","authors":"Xinwang Shi,&nbsp;Fenglong Zhao,&nbsp;Lian Feng,&nbsp;Yijing Liu,&nbsp;Xiaowei Zhou","doi":"10.1002/mp.17564","DOIUrl":"10.1002/mp.17564","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>One big challenge in the noninvasive high-intensity focused ultrasound (HIFU) surgery is that the location and shape of its focus is unpredictable at the preoperative stage due to the complexity of sound wave propagation. The Acoustic Radiation Force Impulse (ARFI) imaging is a potential solution to this problem, but artifacts resulting from shear wave propagation remain to be solved.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>In this study, we proposed avoiding those artefacts by applying the ARFI technique at a high imaging frame rate within a very short time before the shear waves start to propagate.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Using single transmission with a convex imaging probe, two ultrafast imaging modalities (the diverging wave and the wide beam), were developed in the ARFI framework, and their reliabilities were validated on a nylon string phantom by the centroid tracking method borrowed from ultrasound localization microscopy (ULM). The proposed ARFI method was tested on a clinically equivalent HIFU system under different acoustic radiation intensities by in-vitro, ex-vivo and in-vivo experiments. In three experimental scenarios, we delivered short HIFU stimulation pulses at varying acoustic powers to induce tissue motion within the focal region. At each experimental site, both diverging wave and wide-beam imaging techniques were employed for motion estimation. Based on the focus prediction derived from the motion estimation, HIFU ablation treatment was performed. The treated samples were then incised to examine the damaged areas. Additionally, ultrasound B-mode images were acquired before and after the procedure and saved for analysis.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Quantitative analysis showed that the ARFI with wide beam imaging was able to predict the HIFU focus preoperatively, only with 1 to 3 mm of errors in focal central location, and less than 23% of percentage errors in focal area in most cases. However, the diverging wave imaging failed to predict the HIFU focus due to its low signal-to-noise ratio.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>In conclusion, the inherent shear wave artefacts in ARFI for predicting the HIFU focus can be successfully avoided by carefully designing the imaging strategy and its working sequence. This ARFI technique was validated through a series of experiments on a clinically equivalent HIFU system, which demonstrated its capability in assisting surgical planning.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1728-1745"},"PeriodicalIF":3.2,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142808994","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":"A Monte Carlo method to assess the spectral performance of photon counting detectors","authors":"Karl Stierstorfer,&nbsp;Martin Hupfer","doi":"10.1002/mp.17577","DOIUrl":"10.1002/mp.17577","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Assessing the performance of spectral detectors is an important but nontrivial problem. In the past few years, detective quantum efficiency-(DQE)-like quantities have been proposed that allow quantifying the spatial-spectral performance for certain tasks. In previous publications, we have presented and validated an approach to determine detector properties like the modulation transfer function (MTF), the noise power spectrum (NPS), and the DQE based on an end-to-end Monte Carlo model of the detection process. This approach has so far not been used to assess the task-dependent spatial-spectral performance of detectors.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>In this paper, we extend the Monte Carlo method to detectors with several spectral thresholds and show how it can be used to derive all relevant quantities for the assessment of the spectral performance of such detectors. We describe the method in detail and apply it to four interesting types of realistic detectors.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Method</h3>\u0000 \u0000 <p>The method is an extension of the Monte Carlo method presented previously. An end-to-end Monte Carlo simulation of the detection process directly provides the statistics necessary to obtain all relevant performance parameters, including task-based spectral DQEs. The method is applied to two direct converting photon counting detectors using CdTe and silicon: a CdTe-based photon counter with additional coincidence counters and an optical counting system using LaBr₃ as a scintillator.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The task-dependent DQEs show an advantage for CdTe, particularly for non-spectral tasks. Silicon has an advantage for material decomposition tasks at lower frequencies. Both hypothetical systems, the CdTe detector with coincidence counters and the scintillator-based detector, show the potential to outperform the two so-far-realized systems.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>The method presented is a direct method to obtain all relevant quantities (MTF, NPS, various spectral DQEs) from an end-to-end Monte Carlo simulation of the detector. It allows for assessing detector systems currently being used and potential novel detector systems.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1515-1525"},"PeriodicalIF":3.2,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142808975","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":"Automatic plan selection using deep network—A prostate study","authors":"Philippe Y. Chatigny,&nbsp;Cédric Bélanger,&nbsp;Éric Poulin,&nbsp;Luc Beaulieu","doi":"10.1002/mp.17550","DOIUrl":"10.1002/mp.17550","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;Recently, high-dose-rate (HDR) brachytherapy treatment plans generation was improved with the development of multicriteria optimization (MCO) algorithms that can generate thousands of pareto optimal plans within seconds. This brings a shift, from the objective of generating an acceptable plan to choosing the best plans out of thousands.&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;In order to choose the best plans, new criteria beyond usual dosimetrics volumes histogram (DVH) metrics are introduced and a deep learning (DL) framework is added as an automatic plan selection algorithm.&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 new criteria are visual-like criteria implemented for the bladder, rectum, and urethra. One criterion also takes into account the cold spot in the prostate. Those criteria, along with commonly used DVH criteria, are used to form classes on which to train the algorithm. The algorithm is trained with an input of two 3D images, dose and mask of the anatomy, in order to rank and automatically select a plan. The confidence in the output is used for ranking and the automatic plan selection. The algorithm is trained on 835 previously treated prostate cancer patients and evaluated on a separated 20 patients cohort previously evaluated by two experts (clinical medical physicists) in an inter-observer MCO study.&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 deep network takes 10 s to rank 2000 plans (vs. 5–10 min for experts to rank 4 preferred plans). A total of four different networks are trained which offer different trade-offs. The key trade-offs are the target coverage or the organs at risk (OAR) sparing. The algorithm with the best network achieves no statistical difference with the plans chosen by the two experts for 6 and 9 criteria, respectively, out of 13 criteria (paired &lt;i&gt;t&lt;/i&gt;-test with &lt;i&gt;p&lt;/i&gt; &lt;span&gt;&lt;/span&gt;&lt;math&gt;\u0000 &lt;semantics&gt;\u0000 &lt;mo&gt;&gt;&lt;/mo&gt;\u0000 &lt;annotation&gt;$&gt;$&lt;/annotation&gt;\u0000 &lt;/semantics&gt;&lt;/math&gt; 0.05) while the two experts have no statistical difference between them for 7 criteria.&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 developed approach is flexible since it allows the modification or addition of criteria to obtain different trade-offs in plan quality, per the institution standard. The approach is fast and robust while adding negligible time to MCO planning. These results ","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1717-1727"},"PeriodicalIF":3.2,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17550","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142831589","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":"Investigating LETd optimization strategies in carbon ion radiotherapy for pancreatic cancer: a dosimetric study using an anthropomorphic phantom","authors":"Filipa Baltazar,&nbsp;Friderike K. Longarino,&nbsp;Christina Stengl,&nbsp;Jakob Liermann,&nbsp;Stewart Mein,&nbsp;Jürgen Debus,&nbsp;Thomas Tessonnier,&nbsp;Andrea Mairani","doi":"10.1002/mp.17569","DOIUrl":"10.1002/mp.17569","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;Clinical carbon ion beams offer the potential to overcome hypoxia-induced radioresistance in pancreatic tumors, due to their high dose-averaged Linear Energy Transfer (LETd), as previous studies have linked a minimum LETd within the tumor to improved local control. Current clinical practices at the Heidelberg Ion-Beam Therapy Center (HIT), which use two posterior beams, do not fully exploit the LETd advantage of carbon ions, as the high LETd is primarily focused on the beams’ distal edges. Different LETd-boosting strategies, such as Spot-scanning Hadron Arc (SHArc), could enhance LETd distribution by concentrating high-LETd values in potential hypoxic tumor cores while sparing organs at risk.&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 investigate and verify different LETd-boosting strategies using an anthropomorphic pancreas phantom.&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;Various LETd-boosting strategies were investigated for a cylindrical and a pancreas-shaped target in an anthropomorphic pancreas phantom. Treatment plans were optimized using single field optimization (SFO) or multi field optimization (MFO), with objective functions based on either physical dose (Phys), relative biological effectiveness (RBE)-weighted dose, or a combination of RBE and LETd-based objectives (LETopt). The LETd-boosting planning strategies were optimized with the goal of increasing the minimum LETd in the tumor without compromising its homogeneous dose coverage. Beam configurations investigated included the two-beam in-house clinical standard (2-SFO&lt;sub&gt;Phys&lt;/sub&gt;, 2-SFO&lt;sub&gt;RBE&lt;/sub&gt; and 2-MFO&lt;sub&gt;RBE-LETopt&lt;/sub&gt;), a three-beam configuration (3-MFO&lt;sub&gt;RBE&lt;/sub&gt; and 3-MFO&lt;sub&gt;RBE-LETopt&lt;/sub&gt;) and SHArc (SHArc&lt;sub&gt;Phys&lt;/sub&gt;, SHArc&lt;sub&gt;RBE&lt;/sub&gt; and SHArc&lt;sub&gt;RBE-LETopt&lt;/sub&gt;) using step-and-shoot delivery. The different plans were verified using an anthropomorphic pancreas phantom at HIT and compared to treatment planning system (TPS) predictions.&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;All investigated LETd-boosting strategies altered the LETd distribution while meeting optimization goals and constraints, resulting in varying degrees of LETd enhancement. For the cylindrical volume, the SHArc plan resulted in the highest LETd concentration in the tumor core, with the minimum LETd in the GTV scaling up to 91 keV/µm. For the pancreas-shaped volume, however, the 3-MFO&lt;sub&gt;RBE-LETopt&lt;/sub&gt; achieved a higher minimum LETd in the GTV than SHArc&lt;sub&gt;RBE&lt;/sub&gt; (75.6 and 62.3 keV/µm, respectively). W","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1746-1757"},"PeriodicalIF":3.2,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17569","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142803786","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 neural network to create super-resolution MR from multiple 2D brain scans of pediatric patients","authors":"Jose Benitez-Aurioles,&nbsp;Eliana M. Vásquez Osorio,&nbsp;Marianne C. Aznar,&nbsp;Marcel Van Herk,&nbsp;Shermaine Pan,&nbsp;Peter Sitch,&nbsp;Anna France,&nbsp;Ed Smith,&nbsp;Angela Davey","doi":"10.1002/mp.17563","DOIUrl":"10.1002/mp.17563","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;High-resolution (HR) 3D MR images provide detailed soft-tissue information that is useful in assessing long-term side-effects after treatment in childhood cancer survivors, such as morphological changes in brain structures. However, these images require long acquisition times, so routinely acquired follow-up images after treatment often consist of 2D low-resolution (LR) images (with thick slices in multiple planes).&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;In this work, we present a super-resolution convolutional neural network, based on previous single-image MRI super-resolution work, that can reconstruct a HR image from 2D LR slices in multiple planes in order to facilitate the extraction of structural biomarkers from routine scans.&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 multilevel densely connected super-resolution convolutional neural network (mDCSRN) was adapted to take two perpendicular LR scans (e.g., coronal and axial) as tensors and reconstruct a 3D HR image. A training set of 90 HR T1 pediatric head scans from the Adolescent Brain Cognitive Development (ABCD) study was used, with 2D LR images simulated through a downsampling pipeline that introduces motion artifacts, blurring, and registration errors to make the LR scans more realistic to routinely acquired ones.&lt;/p&gt;\u0000 \u0000 &lt;p&gt;The outputs of the model were compared against simple interpolation in two steps. First, the quality of the reconstructed HR images was assessed using the peak signal-to-noise ratio and structural similarity index compared to baseline. Second, the precision of structure segmentation (using the autocontouring software Limbus AI) in the reconstructed versus the baseline HR images was assessed using mean distance-to-agreement (mDTA) and 95% Hausdorff distance. Three datasets were used: 10 new ABCD images (dataset 1), 18 images from the Children's Brain Tumor Network (CBTN) study (dataset 2) and 6 “real-world” follow-up images of a pediatric head and neck cancer patient (dataset 3).&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 mDCSRN outperformed simple interpolation in terms of visual quality. Similarly, structure segmentations were closer to baseline images after 3D reconstruction. The mDTA improved to, on average (95% confidence interval), 0.7 (0.4–1.0) and 0.8 (0.7–0.9) mm for datasets 1 and 3 respectively, from the interpolation performance of 6.5 (3.6–9.5) and 1.2 (1.0–1.3) mm.&lt;/p&gt;\u0000 &lt;/section&gt;\u0000 \u0000 &lt;section&gt;\u0000 \u0000 &lt;","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1693-1705"},"PeriodicalIF":3.2,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/mp.17563","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142831588","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 dual-decoder banded convolutional attention network for bone segmentation in ultrasound images","authors":"Chuanba Liu,&nbsp;Wenshuo Wang,&nbsp;Rui Sun,&nbsp;Teng Wang,&nbsp;Xiantao Shen,&nbsp;Tao Sun","doi":"10.1002/mp.17545","DOIUrl":"10.1002/mp.17545","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;Ultrasound (US) has great potential for application in computer-assisted orthopedic surgery (CAOS) due to its non-radiative, cost-effective, and portable traits. However, bone segmentation from low-quality US images has been challenging. Traditional segmentation methods cannot achieve satisfactory results due to their high customization and dependence on bone morphology. Existing deep learning-based methods make it difficult to ensure efficient and accurate segmentation due to the ignorance of prior knowledge of bone features during feature learning.&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 paper aims to systematically investigate feature extraction and segmentation methodologies of bone US images and then proposes an innovative convolutional neural network to address the need for precise and efficient bone structure extraction in CAOS.&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 paper has proposed a dual-decoder banded convolutional attention network (BCA-Net), which takes the raw US image as input and simplified U-Net as the baseline network. Multiscale banded convolution kernels are employed internally in the BCA-Net model, leveraging the prior knowledge that bone surfaces in US images are exhibited as bright bands of a few millimeters in width. Additionally, a shared encoder to extract input features and two independent decoders to generate outputs for the bone surface and bone shadow mask are integrated into the BCA-Net model, leveraging the prior knowledge that US bone surfaces manifest low-intensity hollow shadows below. Then, a new task consistency loss is introduced that can utilize inter-task dependency fully and enhance the performance of our model. In the network construction process, a dataset containing 1623 sets of US images was adopted, and a five-fold cross-validation strategy was divided into the training and validation sets for the model's training and validation. Many vital metrics were introduced to comprehensively evaluate the model performance, including overlap, edge distance, area under curve, and model efficiency. Finally, the model performance was subjected to a confidence interval, Tukey's honest significant difference, and Cohen's d statistics at a significance level (5%) to ensure the accuracy and reliability of the obtained findings.&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 experimental results show that the BCA-Net model performs well in the bone surface segmentation task. Its average Dice coefficient reaches 87.51%, 4.04% higher than U-Net's, proving its superior bone surface segmentation accuracy. Meanwhile, the average distance err","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1556-1572"},"PeriodicalIF":3.2,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142803707","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":"Learning soft tissue deformation from incremental simulations","authors":"Nathan Lampen,&nbsp;Daeseung Kim,&nbsp;Xuanang Xu,&nbsp;Xi Fang,&nbsp;Jungwook Lee,&nbsp;Tianshu Kuang,&nbsp;Hannah H. Deng,&nbsp;Michael A. K. Liebschner,&nbsp;Jaime Gateno,&nbsp;Pingkun Yan","doi":"10.1002/mp.17554","DOIUrl":"10.1002/mp.17554","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Surgical planning for orthognathic procedures demands swift and accurate biomechanical modeling of facial soft tissues. Efficient simulations are vital in the clinical pipeline, as surgeons may iterate through multiple plans. Biomechanical simulations typically use the finite element method (FEM). Prior works divide FEM simulations into increments to enhance convergence and accuracy. However, this practice elongates simulation time, thereby impeding clinical integration. To accelerate simulations, deep learning (DL) models have been explored. Yet, previous efforts either perform simulations in a single step or neglect the temporal aspects in incremental simulations.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study investigates the use of spatiotemporal incremental modeling for biomechanics simulations of facial soft tissue.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We implement the method using a graph neural network. Our method synergizes spatial features with temporal aggregation using DL networks trained on incremental FEM simulations from 17 subjects that underwent orthognathic surgery.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Our proposed spatiotemporal incremental method achieved a mean accuracy of 0.37 mm with a mean computation time of 1.52 s. In comparison, a spatial-only incremental method yielded a mean accuracy of 0.44 mm and a mean computation time of 1.60 s, while a spatial-only single-step method yielded a mean accuracy of 0.41 mm and a mean computation time of 0.05 s.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Statistical analysis demonstrated that the spatiotemporal incremental method reduced mean errors compared to the spatial-only incremental method, emphasizing the importance of incorporating temporal information in incremental simulations. Overall, we successfully implemented spatiotemporal incremental learning tailored to simulate soft tissue deformation while substantially reducing simulation time compared to FEM.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1914-1925"},"PeriodicalIF":3.2,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142789745","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":"Federated learning for enhanced dose–volume parameter prediction with decentralized data","authors":"Jiahan Zhang,&nbsp;Yang Lei,&nbsp;Junyi Xia,&nbsp;Ming Chao,&nbsp;Tian Liu","doi":"10.1002/mp.17566","DOIUrl":"10.1002/mp.17566","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>The widespread adoption of knowledge-based planning in radiation oncology clinics is hindered by the lack of data and the difficulty associated with sharing medical data.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Purpose</h3>\u0000 \u0000 <p>This study aims to assess the feasibility of mitigating this challenge through federated learning (FL): a centralized model trained with distributed datasets, while keeping data localized and private.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>This concept was tested using 273 prostate 45 Gy plans. The cases were split into a training set with 220 cases and a validation set with 53 cases. The training set was further separated into 10 subsets to simulate treatment plans from different clinics. A gradient-boosting model was used to predict bladder and rectum V_30Gy, V_35Gy, and V_40Gy. The Federated Averaging algorithm was employed to aggregate the individual model weights from distributed datasets. Grid search with five-fold in-training-set cross-validation was implemented to tune model hyperparameters. Additionally, we evaluated the robustness of the FL approach by varying the distribution of the training set data in several scenarios, including different number of sites and imbalanced data across sites.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The mean absolute error (MAE) for the FL model (4.7% ± 2.9%) is significantly lower than individual models trained separately (6.5% ± 4.9%, <i>p</i> &lt; 0.001) and similar to a traditional centralized model (4.4% ± 2.8%, <i>p</i> = 0.14). The federated model is robust to the number of subsets, showing MAE of 4.7% ± 3.2%, 4.8% ± 3.1%, 4.8% ± 2.9%, 4.5% ± 2.8%, 4.9% ± 3.3%, and 4.8% ± 3.1% for 5, 10, 15, 20, 25, and 30 subsets, respectively. For the two imbalanced datasets, the FL model achieves MAEs of 4.5% ± 2.9% and 5.6% ± 4.0%, non-inferior to the balanced data model. For all bladder and rectum metrics, the FL model significantly outperforms 36.7% of individual models.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>This study demonstrates the potential advantages of implementing a federated model over training individual models: the proposed FL approach achieves similar prediction accuracy as a conventional model without requiring centralized data storage. Even when local models struggle to produce accurate predictions due to data scarcity, the federated model consistently maintains high performance.</p>\u0000 </section>\u0000 </div>","PeriodicalId":18384,"journal":{"name":"Medical physics","volume":"52 3","pages":"1408-1415"},"PeriodicalIF":3.2,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142788248","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}