Search

Background: 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., Purpose: A peritumoral radiomics-guided generative adversarial network (PRG-GAN) was proposed to address this challenge., Methods: 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 t-test with Bonferroni correction and Cohen's d (d) effect sizes. A two-sided p-value of less than 0.05 was considered statistically significant., Results: 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 (p = 0.001, d = 0.71), 4.15 ± 7.56 mm (p = 0.002, d = 0.67), and 1.11 ± 1.65 mm (p < 0.001, d = 0.46) of PRG-GAN, respectively., Conclusion: Integrating radiomics features into GAN is promising to solve unclear border limitations and increase the delineation accuracy of GTVp for patients with NPC., (© 2024 American Association of Physicists in Medicine.)

Background: In MR-guided in-bore percutaneous needle interventions, typically 2D interactive real-time imaging is used for navigating the needle into the target. Misaligned 2D imaging planes can result in losing visibility of the needle in the 2D images, which impedes successful targeting. Necessary iterative manual slice adjustment can prolong interventional workflows. Therefore, rapid automatic alignment of the imaging planes with the needle would be preferable to improve such workflows., Purpose: To investigate rapid 3D localization of needles in MR-guided interventions via a convolutional neural network (CNN)-based localization algorithm using an undersampled white-marker contrast acquisition for the purpose of automatic imaging slice alignment., Methods: A radial 3D rf-spoiled gradient echo MR pulse sequence with white-marker encoding was implemented and a CNN-based localization algorithm was employed to extract position and orientation of an aspiration needle from the undersampled white-marker images. The CNN was trained using porcine tissue phantoms (257 needle trajectories, four-fold data augmentation, 90%/10% split into training and validation dataset). Achievable localization times and accuracy were evaluated retrospectively in an ex vivo study (109 needle trajectories) for a range of needle orientations between 78° and 90° relative to the B 0 field. A proof-of-concept in vivo experiment was performed in two porcine animal models and feasibility of automatic imaging slice alignment was evaluated retrospectively., Results: Ex vivo needle localization was achieved with a median localization accuracy of 1.9 mm (distance needle tip to detected needle axis) and a median angular deviation of 2.6° for needle orientations between 86° and 90° to the B 0 field from fully sampled WM images (resolution of (4 mm) 3 , 6434 acquired radial k-space spokes, acquisition time of 80.4 s) in a field-of-view of (256 mm) 3 . Localization accuracy decreased with increasing undersampling and needle trajectory increasingly aligned with B 0 . For needle orientations between 86° and 90° to the B 0 field, a highly accelerated acquisition of only 32 k-space spokes (acquisition time of 0.4 s) yielded a median localization accuracy of 3.1 mm and a median angular deviation of 4.7°. For needle orientations between 78° and 82° to the B 0 field, a median accuracy and angular deviation of 3.5 mm and 6.8° could still be achieved with 64 sampled spokes (acquisition time of 0.8 s). In vivo, a localization accuracy of 1.4 mm and angular deviation of 3.4° was achieved sampling 32 k-space spokes (acquisition time of 0.48 s) with the needle oriented at 87.7° to the B 0 field. For a needle oriented at 77.6° to the B 0 field, localization accuracy of 5.3 mm and angular deviation of 6.8° were still achieved sampling 128 k-space spokes (acquisition time of 1.92 s), allowing for retrospective slice alignment., Conclusion: The investigated approach enables passive biopsy needle localization in 3D. Acceleration of the localization to real-time applicability is feasible for needle orientations approximately perpendicular to B 0 . The method can potentially facilitate MR-guided needle interventions by enabling automatic imaging slice alignment with the needle., (© 2024 Siemens Healthineers AG, Cook Advanced Technologies and The Author(s). Medical Physics published by Wiley Periodicals LLC on behalf of American Association of Physicists in Medicine.)

Background: Accurate preoperative prediction of cervical lymph node metastasis (LNM) for papillary thyroid carcinoma (PTC) patients is essential for disease staging and individualized treatment planning, which can improve prognosis and facilitate better management., Purpose: To establish a fully automated deep learning-enabled model (FADLM) for automated tumor segmentation and cervical LNM prediction in PTC using ultrasound (US) video keyframes., Methods: The bicentral study retrospective enrolled 518 PTC patients, who were then randomly divided into the training (Hospital 1, n = 340), internal test (Hospital 1, n = 83), and external test cohorts (Hospital 2, n = 95). The FADLM integrated mask region-based convolutional neural network (Mask R-CNN) for automatic thyroid primary tumor segmentation and ResNet34 with Bayes strategy for cervical LNM diagnosis. A radiomics model (RM) using the same automated segmentation method, a traditional radiomics model (TRM) using manual segmentation, and a clinical-semantic model (CSM) were developed for comparison. The dice similarity coefficient (DSC) was used to evaluate segmentation performance. The prediction performance of the models was validated in terms of discrimination and clinical utility with the area under the receiver operator characteristic curve (AUC), heatmap analysis, and decision curve analysis (DCA). The comparison of the predictive performance among different models was conducted by DeLong test. The performances of two radiologists compared with FADLM and the diagnostic augmentation with FADLM's assistance were analyzed in terms of accuracy, sensitivity and specificity using McNemar's x 2 test. The p-value less than 0.05 was defined as a statistically significant difference. The Benjamini-Hochberg procedure was applied for multiple comparisons to deal with Type I error., Results: The FADLM yielded promising segmentation results in training (DSC: 0.88 ± 0.23), internal test (DSC: 0.88 ± 0.23), and external test cohorts (DSC: 0.85 ± 0.24). The AUCs of FADLM for cervical LNM prediction were 0.78 (95% CI: 0.73, 0.83), 0.83 (95% CI: 0.74, 0.92), and 0.83 (95% CI: 0.75, 0.92), respectively. It all significantly outperformed the RM (AUCs: 0.78 vs. 0.72; 0.83 vs. 0.65; 0.83 vs. 0.68, all adjusted p-values < 0.05) and CSM (AUCs: 0.78 vs. 0.71; 0.83 vs. 0.62; 0.83 vs. 0.68, all adjusted p-values < 0.05) across the three cohorts. The RM offered similar performance to that of TRM (AUCs: 0.61 vs. 0.63, adjusted p-value = 0.60) while significantly reducing the segmentation time (3.3 ± 3.8 vs. 14.1 ± 4.2 s, p-value < 0.001). Under the assistance of FADLM, the accuracies of junior and senior radiologists were improved by 18% and 15% (all adjusted p-values < 0.05) and the sensitivities by 25% and 21% (all adjusted p-values < 0.05) in the external test cohort., Conclusion: The FADLM with elaborately designed automated strategy using US video keyframes holds good potential to provide an efficient and consistent prediction of cervical LNM in PTC. The FADLM displays superior performance to RM, CSM, and radiologists with promising efficacy., (© 2024 American Association of Physicists in Medicine.)

Background: Computer-assisted planning methods have increasingly contributed to preoperative ablation planning; however, these methods cannot automatically obtain the final optimal solution within a short time and are rarely validated in practice, greatly limiting their clinical applicability., Purpose: We aimed to propose a full-automatic multi-stage ablation and needle trajectory planning method for CT-guided percutaneous liver ablation to attain the final optimal plans under multiple clinical constraints rapidly., Methods: Our proposed method integrates the ablation zone planning fulfilling complete tumor coverage and critical structure avoidance while reaching a trade-off between ablation number and healthy tissue damage, and needle trajectory planning under multiple clinical constraints. Our needle trajectory planning determines feasible skin entry regions based on hard constraints, where the multi-objective optimization (MOO) considering soft constraints is performed using the Pareto Optimality and Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) methods for the final optimal solution. The performance of our proposed method was evaluated on 30 tumors of various characteristics from 23 patients and clinically validated in five clinical cases., Results: Our proposed method achieved 99.8% treatment zone coverage and 40.5% ablation efficiency without involving critical structures, and completely satisfied multiple clinical constraints in all needle trajectory planning results. The average planning time was 23.6 s for tumors of different sizes. All the plans were considered clinically acceptable by the doctors' evaluation. Our method achieved complete tumor coverage without complications in clinical case validation., Conclusion: Our proposed planning method can generate a final optimal plan satisfying multiple clinical constraints within a short time, potentially facilitating preoperative planning for hepatic tumor ablation., (© 2024 The Author(s). Medical Physics published by Wiley Periodicals LLC on behalf of American Association of Physicists in Medicine.)

Background: Fluoroscopy guided interventions (FGIs) pose a risk of prolonged radiation exposure; personalized patient dosimetry is necessary to improve patient safety during these procedures. However, current FGIs systems do not capture the precise exposure regions of the patient, making it challenging to perform patient-procedure-specific dosimetry. Thus, there is a pressing need to develop approaches to extract and use this information to enable personalized radiation dosimetry for interventional procedures., Purpose: To propose a deep learning (DL) approach for the automatic localization of 3D anatomical landmarks on randomly collimated and magnified 2D head fluoroscopy images., Materials and Methods: The model was developed with datasets comprising 800 000 pseudo 2D synthetic images (mixture of vessel-enhanced and non-enhancement), each with 55 annotated anatomical landmarks (two are landmarks for eye lenses), generated from 135 retrospectively collected head computed tomography (CT) volumetric data. Before training, dynamic random cropping was performed to mimic the varied field-size collimation in FGI procedures. Gaussian-distributed additive noise was applied to each individual image to enhance the robustness of the DL model in handling image degradation that may occur during clinical image acquisition in a clinical environment. The model was trained with 629 370 synthetic images for approximately 275 000 iterations and evaluated against a synthetic image test set and a clinical fluoroscopy test set., Results: The model shows good performance in estimating in- and out-of-image landmark positions and shows feasibility to instantiate the skull shape. The model successfully detected 96.4% and 92.5% 2D and 3D landmarks, respectively, within a 10 mm error on synthetic test images. It demonstrated an average of 3.6 ± 2.3 mm mean radial error and successfully detected 96.8% 2D landmarks within 10 mm error on clinical fluoroscopy images., Conclusion: Our deep-learning model successfully localizes anatomical landmarks and estimates the gross shape of skull structures from collimated 2D projection views. This method may help identify the exposure region required for patient-specific organ dosimetry in FGIs procedures., (© 2024 American Association of Physicists in Medicine.)

Background: Automatic segmentation of vertebrae in spinal x-ray images is crucial for clinical diagnosis, case analysis, and surgical planning of spinal lesions., Purpose: However, due to the inherent characteristics of x-ray images, including low contrast, high noise, and uneven grey scale, it remains a critical and challenging problem in computer-aided spine image analysis and disease diagnosis applications., Methods: In this paper, a Multiscale Feature Enhancement Network (MFENet), is proposed for segmenting whole spinal x-ray images, to aid doctors in diagnosing spinal-related diseases. To enhance feature extraction, the network incorporates a Dual-branch Feature Extraction Module (DFEM) and a Semantic Aggregation Module (SAM). The DFEM has a parallel dual-branch structure. The upper branch utilizes multiscale convolutional kernels to extract features from images. Employing convolutional kernels of different sizes helps capture details and structural information at different scales. The lower branch incorporates attention mechanisms to further optimize feature representation. By modeling the feature maps spatially and across channels, the network becomes more focused on key feature regions and suppresses task-irrelevant information. The SAM leverages contextual semantic information to compensate for details lost during pooling and convolution operations. It integrates high-level feature information from different scales to reduce segmentation result discontinuity. In addition, a hybrid loss function is employed to enhance the network's feature extraction capability., Results: In this study, we conducted a multitude of experiments utilizing dataset provided by the Spine Surgery Department of Henan Provincial People's Hospital. The experimental results indicate that our proposed MFENet demonstrates superior segmentation performance in spinal segmentation on x-ray images compared to other advanced methods, achieving 92.61 ± 0.431 for MIoU, 92.42 ± 0.329 for DSC, and 99.51 ± 0.037 for Global&#95;accuracy., Conclusions: Our model is able to more effectively learn and extract global contextual semantic information, significantly improving spinal segmentation performance, further aiding doctors in analyzing patient conditions., (© 2024 American Association of Physicists in Medicine.)

Background: X-ray radiography is a widely used imaging technique worldwide, and its image quality directly affects diagnostic accuracy. Therefore, X-ray image quality control (QC) is essential. However, subjectively assessing image quality is inefficient and inconsistent, especially when large amounts of image data are being evaluated. Thus, subjective assessment cannot meet current QC needs., Purpose: To meet current QC needs and improve the efficiency of image quality assessment, a complete set of quality assessment criteria must be established and implemented using artificial intelligence (AI) technology. Therefore, we proposed a multi-criteria AI system for automatically assessing the image quality of knee radiographs., Methods: A knee radiograph QC knowledge graph containing 16 "acquisition technique" labels representing 16 image quality defects and five "clarity" labels representing five grades of clarity were developed. Ten radiographic technologists conducted three rounds of QC based on this graph. The single-person QC results were denoted as QC1 and QC2, and the multi-person QC results were denoted as QC3. Each technologist labeled each image only once. The ResNet model structure was then used to simultaneously perform classification (detection of image quality defects) and regression (output of a clarity score) tasks to construct an image QC system. The QC3 results, comprising 4324 anteroposterior and lateral knee radiographs, were used for model training (70% of the images), validation (10%), and testing (20%). The 865 test set data were used to evaluate the effectiveness of the AI model, and an AI QC result, QC4, was automatically generated by the model after training. Finally, using a double-blind method, the senior QC expert reviewed the final QC results of the test set with reference to the results QC3 and QC4 and used them as a reference standard to evaluate the performance of the model. The precision and mean absolute error (MAE) were used to evaluate the quality of all the labels in relation to the reference standard., Results: For the 16 "acquisition technique" features, QC4 exhibited the highest weighted average precision (98.42% ± 0.81%), followed by QC3 (91.39% ± 1.35%), QC2 (87.84% ± 1.68%), and QC1 (87.35% ± 1.71%). For the image clarity features, the MAEs between QC1, QC2, QC3, and QC4 and the reference standard were 0.508 ± 0.021, 0.475 ± 0.019, 0.237 ± 0.016, and 0.303 ± 0.018, respectively., Conclusions: The experimental results show that our automated quality assessment system performed well in classifying the acquisition technique used for knee radiographs. The image clarity quality evaluation accuracy of the model must be further improved but is generally close to that of radiographic technologists. Intelligent QC methods using knowledge graphs and convolutional neural networks have the potential for clinical applications., (© 2024 American Association of Physicists in Medicine.)

BackgroundIn cone-beam computed tomography (CBCT)-guided radiotherapy, off-by-one vertebral-body misalignments are rare but serious errors that lead to wrong-site treatments.PurposeAn automatic error detection algorithm was developed that uses a three-branch convolutional neural network error detection model (EDM) to detect off-by-one vertebral-body misalignments using planning computed tomography (CT) images and setup CBCT images.MethodsAlgorithm training and test data consisted of planning CTs and CBCTs from 480 patients undergoing radiotherapy treatment in the thoracic and abdominal regions at two radiotherapy clinics. The clinically applied registration was used to derive true-negative (no error) data. The setup and planning images were then misaligned by one vertebral-body in both the superior and inferior directions, simulating the most likely misalignment scenarios. For each of the aligned and misaligned 3D image pairs, 2D slice pairs were automatically extracted in each anatomical plane about a point within the vertebral column. The three slice pairs obtained were then inputted to the EDM that returned a probability of vertebral misalignment. One model (EDM1 ) was trained solely on data from institution 1. EDM1 was further trained using a lower learning rate on a dataset from institution 2 to produce a fine-tuned model, EDM2 . Another model, EDM3 , was trained from scratch using a training dataset composed of data from both institutions. These three models were validated on a randomly selected and unseen dataset composed of images from both institutions, for a total of 303 image pairs. The model performances were quantified using a receiver operating characteristic analysis. Due to the rarity of vertebral-body misalignments in the clinic, a minimum threshold value yielding a specificity of at least 99% was selected. Using this threshold, the sensitivity was calculated for each model, on each institution's test set separately.ResultsWhen applied to the combined test set, EDM1 , EDM2 , and EDM3 resulted in an area under curve of 99.5%, 99.4%, and 99.5%, respectively. EDM1 achieved a sensitivity of 96% and 88% on Institution 1 and Institution 2 test set, respectively. EDM2 obtained a sensitivity of 95% on each institution's test set. EDM3 achieved a sensitivity of 95% and 88% on Institution 1 and Institution 2 test set, respectively.ConclusionThe proposed algorithm demonstrated accuracy in identifying off-by-one vertebral-body misalignments in CBCT-guided radiotherapy that was sufficiently high to allow for practical implementation. It was found that fine-tuning the model on a multi-facility dataset can further enhance the generalizability of the algorithm.

Background and purpose: Magnetic resonance imaging (MRI) scanners are a major contributor to greenhouse gas emissions from the healthcare sector, and efforts to improve energy efficiency and reduce energy consumption rely on quantification of the characteristics of energy consumption. The purpose of this work was to develop a semi‐automatic analytical methodology for the characterization of the energy consumption of MRI systems using only the load duration curve (LDC). LDCs are a fundamental tool used across various fields to analyze and understand the behavior of loads over time. Methods: An electric current transformer sensor and data logger were installed on two 3T MRI scanners from two vendors, termed M1 (outpatient scanner) and M2 (inpatient/emergency scanner). Data was collected for 1 month (7/11/2023 to 8/11/2023). Active power was calculated, assuming a balanced three‐phase system, using the average current measured across all three phases, a 480 V reference voltage for both machines, and vendor‐provided power factors. An LDC was constructed for each system by sorting the active power values in descending order and computing the cumulative time (in units of percentage) for each data point. The first derivative of the LDC was then computed (LDC'), smoothed by convolution with a window function (sLDC'), and used to detect transitions between different system modes including (in descending power levels): scan, prepared‐to‐scan, idle, low‐power, and off. The final, segmented LDC was used to measure time (% total time), total energy (kWh), and mean power (kW) for each system mode on both scanners. The method was validated by comparing mean power values, computed using the segmented 1‐month LDC, for each nonproductive system mode (i.e., prepared‐to‐scan, idle, lower‐power, and off) against power levels measured after a deliberate system shutdown was performed for each scanner (1 day worth of data). Results: The validation revealed differences in mean power values <1.4% for all nonproductive modes and both scanners. In the scan system mode, the mean power values ranged from 29.8 to 37.2 kW and the total energy consumed for 1 month ranged from 11 106 to 14 466 kWh depending on the scanner. Over the course of 1 month, the portion of time the scanners were in nonproductive modes ranged from 76% to 80% across scanners and the nonproductive energy consumption ranged from 8010 to 6722 kWh depending on the scanner. The M1 (outpatient) scanner consumed 99.9 and 183.9 kWh/day in idle mode for weekdays and weekends, respectively, because the scanner spent 23% more time proportionally in idle mode on the weekends. Conclusions: A semi‐automatic method for quantifying energy consumption characteristics of MRI scanners was introduced and validated. This method is relatively simple to implement as it requires only power data from the scanners and avoids the technical challenges associated with extracting and processing scanner log files. The methodology enables quantitative evaluation of the power, time, and energy characteristics of MRI scanners in scan and nonproductive system modes, providing baseline data and the capability of identifying potential opportunities for enhancing the energy efficiency of MRI scanners. [ABSTRACT FROM AUTHOR]

Background: Electron therapy, vital for treating skin lesions and superficial tumors, demands precise electron cutouts to ensure accurate dose delivery. Traditional manual tracing methods introduce uncertainties and inefficiencies, necessitating innovative solutions for custom block creation., Purpose: To introduce and validate the Optical Automatic Contour Tracing (O-ACT) method, enhancing electron cutout generation's accuracy and efficiency through optical imaging and software automation., Methods: Utilizing a 3D-printed holder, a centrally-mounted charge-coupled device (CCD) camera on the electron applicator secured two distinct perspective images of the skin's contour at varying heights. Through image binarization and skeletonization, we identified the clinical target contour's region-of-interests (ROIs) in each setup image. Employing distances from each ROI's center, assessed at 5-degree intervals from both images, we reconstructed the target contour on the skin. The magnification factor, set at a 95 cm source-to-point distance, determined the final cutout shape. We crafted an in-house software in MATLAB for camera calibration and image processing and juxtaposed our results against the standard clinical cutout from the treatment planning system (TPS) using a correlation coefficient based on masked binary images' mutual information. Additionally, we performed dosimetric evaluations using abstract shapes to compare O-ACT with other methods., Results: Our methodology yielded cutout shapes exhibiting remarkable alignment with the TPS clinical cutout. O-ACT demonstrated superior precision in generating cutout shapes that closely align with the contours in TPS, improving upon other methods in terms of adaptability to patient body shapes and contour accuracy. Dosimetric evaluations showed minimal differences between methods, with O-ACT providing slightly more consistent results. Dose profile analyses in penumbra regions indicated O-ACT's improved accuracy compared to conventional methods., Conclusions: Pushing the boundaries of traditional practices, our O-ACT offers a more accurate, efficient, and reproducible method for custom electron cutout creation from clinical setup images. This innovation promises not only to streamline clinical workflow but also to potentially uplift clinical outcomes in radiation oncology by providing more accurate patient-specific treatment accessary., (© 2024 The Author(s). Medical Physics published by Wiley Periodicals LLC on behalf of American Association of Physicists in Medicine.)

Background: Head and neck (HN) gross tumor volume (GTV) auto-segmentation is challenging due to the morphological complexity and low image contrast of targets. Multi-modality images, including computed tomography (CT) and positron emission tomography (PET), are used in the routine clinic to assist radiation oncologists for accurate GTV delineation. However, the availability of PET imaging may not always be guaranteed., Purpose: To develop a deep learning segmentation framework for automated GTV delineation of HN cancers using a combination of PET/CT images, while addressing the challenge of missing PET data., Methods: Two datasets were included for this study: Dataset I: 524 (training) and 359 (testing) oropharyngeal cancer patients from different institutions with their PET/CT pairs provided by the HECKTOR Challenge; Dataset II: 90 HN patients(testing) from a local institution with their planning CT, PET/CT pairs. To handle potentially missing PET images, a model training strategy named the "Blank Channel" method was implemented. To simulate the absence of a PET image, a blank array with the same dimensions as the CT image was generated to meet the dual-channel input requirement of the deep learning model. During the model training process, the model was randomly presented with either a real PET/CT pair or a blank/CT pair. This allowed the model to learn the relationship between the CT image and the corresponding GTV delineation based on available modalities. As a result, our model had the ability to handle flexible inputs during prediction, making it suitable for cases where PET images are missing. To evaluate the performance of our proposed model, we trained it using training patients from Dataset I and tested it with Dataset II. We compared our model (Model 1) with two other models which were trained for specific modality segmentations: Model 2 trained with only CT images, and Model 3 trained with real PET/CT pairs. The performance of the models was evaluated using quantitative metrics, including Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff Distance (HD95). In addition, we evaluated our Model 1 and Model 3 using the 359 test cases in Dataset I., Results: Our proposed model(Model 1) achieved promising results for GTV auto-segmentation using PET/CT images, with the flexibility of missing PET images. Specifically, when assessed with only CT images in Dataset II, Model 1 achieved DSC of 0.56 ± 0.16, MSD of 3.4 ± 2.1 mm, and HD95 of 13.9 ± 7.6 mm. When the PET images were included, the performance of our model was improved to DSC of 0.62 ± 0.14, MSD of 2.8 ± 1.7 mm, and HD95 of 10.5 ± 6.5 mm. These results are comparable to those achieved by Model 2 and Model 3, illustrating Model 1's effectiveness in utilizing flexible input modalities. Further analysis using the test dataset from Dataset I showed that Model 1 achieved an average DSC of 0.77, surpassing the overall average DSC of 0.72 among all participants in the HECKTOR Challenge., Conclusions: We successfully refined a multi-modal segmentation tool for accurate GTV delineation for HN cancer. Our method addressed the issue of missing PET images by allowing flexible data input, thereby providing a practical solution for clinical settings where access to PET imaging may be limited., (© 2024 American Association of Physicists in Medicine.)

PurposesPreimplant diagnostic magnetic resonance imaging is the gold standard for image-guided tandem-and-ovoids (T&O) brachytherapy for cervical cancer. However, high dose rate brachytherapy planning is typically done on postimplant CT-based high-risk clinical target volume (HR-CTVCT ) because the transfer of preimplant Magnetic resonance (MR)-based HR-CTV (HR-CTVMR ) to the postimplant planning CT is difficult due to anatomical changes caused by applicator insertion, vaginal packing, and the filling status of the bladder and rectum. This study aims to train a dual-path convolutional neural network (CNN) for automatic segmentation of HR-CTVCT on postimplant planning CT with guidance from preimplant diagnostic MR.MethodsPreimplant T2-weighted MR and postimplant CT images for 65 (48 for training, eight for validation, and nine for testing) patients were retrospectively solicited from our institutional database. MR was aligned to the corresponding CT using rigid registration. HR-CTVCT and HR-CTVMR were manually contoured on CT and MR by an experienced radiation oncologist. All images were then resampled to a spatial resolution of 0.5 × 0.5 × 1.25 mm. A dual-path 3D asymmetric CNN architecture with two encoding paths was built to extract CT and MR image features. The MR was masked by HR-CTVMR contour while the entire CT volume was included. The network put an asymmetric weighting of 18:6 for CT: MR. Voxel-based dice similarity coefficient (DSCV ), sensitivity, precision, and 95% Hausdorff distance (95-HD) were used to evaluate model performance. Cross-validation was performed to assess model stability. The study cohort was divided into a small tumor group (40 cc) based on the HR-CTVCT for model evaluation. Single-path CNN models were trained with the same parameters as those in dual-path models.ResultsFor this patient cohort, the dual-path CNN model improved each of our objective findings, including DSCV , sensitivity, and precision, with an average improvement of 8%, 7%, and 12%, respectively. The 95-HD was improved by an average of 1.65 mm compared to the single-path model with only CT images as input. In addition, the area under the curve for different networks was 0.86 (dual-path with CT and MR) and 0.80 (single-path with CT), respectively. The dual-path CNN model with asymmetric weighting achieved the best performance with DSCV of 0.65 ± 0.03 (0.61-0.70), 0.79 ± 0.02 (0.74-0.85), and 0.75 ± 0.04 (0.68-0.79) for small, medium, and large group. 95-HD were 7.34 (5.35-10.45) mm, 5.48 (3.21-8.43) mm, and 6.21 (5.34-9.32) mm for the three size groups, respectively.ConclusionsAn asymmetric CNN model with two encoding paths from preimplant MR (masked by HR-CTVMR ) and postimplant CT images was successfully developed for automatic segmentation of HR-CTVCT for T&O brachytherapy patients.

Background Purpose Methods Results Conclusions The use of endovascular procedures is becoming increasingly popular across multiple clinical domains. These procedures are generally performed under image guidance using an interventional c‐arm x‐ray system. Radiation exposure to both patients and interventional staff due to use of fluoroscopy is a health and occupational concern, but modifications to the interventional workflow to address radiation may come at the cost of procedure time or quality.Interventional x‐ray collimation is a crucial task for improving image quality as well as reducing radiation exposure to both patients and operators who work in the x‐ray domain. However, collimation is heavily underutilized due to its cumbersome nature and the difficulty of manually manipulating multiple parameters during fast‐paced interventional procedures. Additionally, the widely varying collimation preferences of interventionalists across different procedure types, procedure phases, and anatomies makes the standardizing of collimation challenging for radiation technologist support staff.Automating collimation has the potential to bridge this gap, freeing up mental bandwidth for interventionalists and technologists and improving outcomes for patients. Here, we propose a tunable algorithm for automatic collimation based on a region‐of‐interest optimizer driven by a combination of image, system, device, and radiation based features and we illustrate its efficacy across varying personal preferences. Critically, we devise a method with a simple and easily understandable mapping between algorithm parameters and practical outcomes.We show a real‐time implementation of this algorithm using deep feature extraction by a convolutional neural network and evaluate its performance in a custom dataset of simulated fluoroscopy and recorded fluoroscopy from clinical radial access procedures. We evaluate the effects of a practically implemented mixed supervision training strategy on model performance and show potential for radiation reduction in simulation. An uncertainty analysis indicates that the algorithm is robust to noise and anatomical variation across our clinical dataset. Clinical acceptability and quality is evaluated through a reader study with expert neuro‐interventional radiologists, with participants indicating 100% clinical acceptability, high quality ratings, and improved radiation protection over their typical practice.The algorithm's modular design ensured that users' collimation requirements were met without disruption to the interventional workflow or procedure time, while exhibiting strong potential to reduce radiation risk to patients and operators. Evaluation in more varied clinical settings could support translation of this technology into the clinic. [ABSTRACT FROM AUTHOR]

Background Purpose Methods Results Conclusions Accurate and automatic segmentation of pericardial adipose tissue (PEAT) in cardiac magnetic resonance (MR) images is essential for the diagnosis and treatment of cardiovascular diseases. Precise segmentation is challenging due to high costs and the need for specialized knowledge, as a large amount of accurately annotated data is required, demanding significant time and medical resources.In order to reduce the burden of data annotation while maintaining the high accuracy of segmentation tasks, this paper introduces a semi‐supervised learning method to solve the limitations of current PEAT segmentation methods.In this paper, we propose a difference‐guided collaborative mean teacher (DCMT) semi‐supervised method, designed for the segmentation of PEAT from DCMT consists of two main components: a semi‐supervised framework with a difference fusion strategy and a backbone network MCM‐UNet using Mamba‐CNN mixture (MCM) blocks. The differential fusion strategy effectively utilizes the uncertain areas in unlabeled data, encouraging the model to reach a consensus in predictions across these difficult‐to‐segment yet information‐rich areas. In addition, considering the sparse and scattered distribution of PEAT in cardiac MR images, which makes it challenging to segment, we propose MCM‐UNet as the backbone network in our semi‐supervised framework. This not only enhances the processing ability of global information, but also accurately captures the detailed local features of the image, which greatly improves the accuracy of PEAT segmentation.Our experiments conducted on the MRPEAT dataset show that our DCMT method outperforms existing state‐of‐the‐art semi‐supervised methods in terms of segmentation accuracy. These findings underscore the effectiveness of our approach in handling the specific challenges associated with PEAT segmentation.The DCMT method significantly improves the accuracy of PEAT segmentation in cardiac MR images. By effectively utilizing uncertain areas in the data and enhancing feature capture with the MCM‐UNet, our approach demonstrates superior performance and offers a promising solution for semi‐supervised learning in medical image segmentation. This method can alleviate the extensive annotation requirements typically necessary for training accurate segmentation models in medical imaging. [ABSTRACT FROM AUTHOR]

Background: Adequate image enhancement of organs and blood vessels of interest is an important aspect of image quality in contrast-enhanced computed tomography (CT). There is a need for an objective method for evaluation of vessel contrast that can be automatically and systematically applied to large sets of CT exams., Purpose: The purpose of this work was to develop a method to automatically segment and measure attenuation Hounsfield Unit (HU) in the portal vein (PV) in contrast-enhanced abdomen CT examinations., Methods: Input CT images were processed by a vessel enhancing filter to determine candidate PV segmentations. Multiple machine learning (ML) classifiers were evaluated for classifying a segmentation as corresponding to the PV based on segmentation shape, location, and intensity features. A public data set of 82 contrast-enhanced abdomen CT examinations was used to train the method. An optimal ML classifier was selected by training and tuning on 66 out of the 82 exams (80% training split) in the public data set. The method was evaluated in terms of segmentation classification accuracy and PV attenuation measurement accuracy, compared to manually determined ground truth, on a test set of the remaining 16 exams (20% test split) held out from public data set. The method was further evaluated on a separate, independently collected test set of 21 examinations., Results: The best classifier was found to be a random forest, with a precision of 0.892 in the held-out test set to correctly identify the PV from among the input candidate segmentations. The mean absolute error of the measured PV attenuation relative to ground truth manual measurement was 13.4 HU. On the independent test set, the overall precision decreased to 0.684. However, the PV attenuation measurement remained relatively accurate with a mean absolute error of 15.2 HU., Conclusions: The method was shown to accurately measure PV attenuation over a large range of attenuation values, and was validated in an independently collected dataset. The method did not require time-consuming manual contouring to supervise training. The method may be applied to systematic quality control of contrast-enhanced CT examinations., (© 2024 American Association of Physicists in Medicine.)

Background Purpose Methods Results Conclusions Ablation zone segmentation in contrast‐enhanced computed tomography (CECT) images enables the quantitative assessment of treatment success in the ablation of liver lesions. However, fully automatic liver ablation zone segmentation in CT images still remains challenging, such as low accuracy and time‐consuming manual refinement of the incorrect regions.Therefore, in this study, we developed a semi‐automatic technique to address the remaining drawbacks and improve the accuracy of the liver ablation zone segmentation in the CT images.Our approach uses a combination of a CNN‐based automatic segmentation method and an interactive CNN‐based segmentation method. First, automatic segmentation is applied for coarse ablation zone segmentation in the whole CT image. Human experts then visually validate the segmentation results. If there are errors in the coarse segmentation, local corrections can be performed on each slice via an interactive CNN‐based segmentation method. The models were trained and the proposed method was evaluated using two internal datasets of post‐interventional CECT images (n1$n_{1}$ = 22, n2$n_{2}$ = 145; 62 patients in total) and then further tested using an external benchmark dataset (n3$n_{3}$ = 12; 10 patients).To evaluate the accuracy of the proposed approach, we used Dice similarity coefficient (DSC), average symmetric surface distance (ASSD), Hausdorff distance (HD), and volume difference (VD). The quantitative evaluation results show that the proposed approach obtained mean DSC, ASSD, HD, and VD scores of 94.0%, 0.4 mm, 8.4 mm, 0.02, respectively, on the internal dataset, and 87.8%, 0.9 mm, 9.5 mm, and −0.03, respectively, on the benchmark dataset. We also compared the performance of the proposed approach to that of five well‐known segmentation methods; the proposed semi‐automatic method achieved state‐of‐the‐art performance on ablation segmentation accuracy, and on average, 2 min are required to correct the segmentation. Furthermore, we found that the accuracy of the proposed method on the benchmark dataset is comparable to that of manual segmentation by human experts (p$p$ = 0.55, t$t$‐test).The proposed semi‐automatic CNN‐based segmentation method can be used to effectively segment the ablation zones, increasing the value of CECT for an assessment of treatment success. For reproducibility, the trained models, source code, and demonstration tool are publicly available at https://github.com/lqanh11/Interactive_AblationZone_Segmentation. [ABSTRACT FROM AUTHOR]

Background: Liver fibrosis poses a significant public health challenge given its elevated incidence and associated mortality rates. Diffusion-Weighted Imaging (DWI) serves as a non-invasive diagnostic tool for supporting the identification of liver fibrosis. Deep learning, as a computer-aided diagnostic technology, can assist in recognizing the stage of liver fibrosis by extracting abstract features from DWI images. However, gathering samples is often challenging, posing a common dilemma in previous research. Moreover, previous studies frequently overlooked the cross-comparison information and latent connections among different DWI parameters. Thus, it is becoming a challenge to identify effective DWI parameters and dig potential features from multiple categories in a dataset with limited samples., Purpose: A self-defined Multi-view Contrastive Learning Network is developed to automatically classify multi-parameter DWI images and explore synergies between different DWI parameters., Methods: A Dense-fusion Attention Contrastive Learning Network (DACLN) is designed and used to recognize DWI images. Concretely, a multi-view contrastive learning framework is constructed to train and extract features from raw multi-parameter DWI. Besides, a Dense-fusion module is designed to integrate feature and output predicted labels., Results: We evaluated the performance of the proposed model on a set of real clinical data and analyzed the interpretability by Grad-CAM and annotation analysis, achieving average scores of 0.8825, 0.8702, 0.8933, 0.8727, and 0.8779 for accuracy, precision, recall, specificity and F-1 score. Of note, the experimental results revealed that IVIM-f, CTRW-β, and MONO-ADC exhibited significant recognition ability and complementarity., Conclusion: Our method achieves competitive accuracy in liver fibrosis diagnosis using the limited multi-parameter DWI dataset and finds three types of DWI parameters with high sensitivity for diagnosing liver fibrosis, which suggests potential directions for future research., (© 2024 American Association of Physicists in Medicine.)

Background: Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is a standardized semi-quantitative method for early ischemic changes in acute ischemic stroke., Purpose: However, ASPECTS is still affected by expert experience and inconsistent results between readers in clinical. This study aims to propose an automatic ASPECTS scoring model based on diffusion-weighted imaging (DWI) mode to help clinicians make accurate treatment plans., Methods: Eighty-two patients with stroke were included in the study. First, we designed a new deep learning network for segmenting ASPECTS scoring brain regions. The network is improved based on U-net, which integrates multiple modules. Second, we proposed using hybrid classifiers to classify brain regions. For brain regions with larger areas, we used brain grayscale comparison algorithm to train machine learning classifiers, while using hybrid feature training for brain regions with smaller areas., Results: The average DICE coefficient of the segmented hindbrain area can reach 0.864. With the proposed hybrid classifier, our method performs significantly on both region-level ASPECTS and dichotomous ASPECTS. The sensitivity and accuracy on the test set are 95.51% and 93.43%, respectively. For dichotomous ASPECTS, the intraclass correlation coefficient (ICC) between our automated ASPECTS score and the expert reading was 0.87., Conclusions: This study proposed an automated model for ASPECTS scoring of patients with acute ischemic stroke based on DWI images. Experimental results show that the method of segmentation first and then classification is feasible. Our method has the potential to assist physicians in the Alberta Stroke Program with early CT scoring and clinical stroke diagnosis., (© 2024 American Association of Physicists in Medicine.)

Background: View planning for the acquisition of cardiac magnetic resonance (CMR) imaging remains a demanding task in clinical practice. Purpose: Existing approaches to its automation relied either on an additional volumetric image not typically acquired in clinic routine, or on laborious manual annotations of cardiac structural landmarks. This work presents a clinic‐compatible, annotation‐free system for automatic CMR view planning. Methods: The system mines the spatial relationship—more specifically, locates the intersecting lines—between the target planes and source views, and trains U‐Net‐based deep networks to regress heatmaps defined by distances from the intersecting lines. On the one hand, the intersection lines are the prescription lines prescribed by the technologists at the time of image acquisition using cardiac landmarks, and retrospectively identified from the spatial relationship. On the other hand, as the spatial relationship is self‐contained in properly stored data, for example, in the DICOM format, the need for additional manual annotation is eliminated. In addition, the interplay of the multiple target planes predicted in a source view is utilized in a stacked hourglass architecture consisting of repeated U‐Net‐style building blocks to gradually improve the regression. Then, a multiview planning strategy is proposed to aggregate information from the predicted heatmaps for all the source views of a target plane, for a globally optimal prescription, mimicking the similar strategy practiced by skilled human prescribers. For performance evaluation, the retrospectively identified planes prescribed by the technologists are used as the ground truth, and the plane angle differences and localization distances between the planes prescribed by our system and the ground truth are compared. Results: The retrospective experiments include 181 clinical CMR exams, which are randomly split into training, validation, and test sets in the ratio of 64:16:20. Our system yields the mean angular difference and point‐to‐plane distance of 5.68∘$^\circ$ and 3.12 mm, respectively, on the held‐out test set. It not only achieves superior accuracy to existing approaches including conventional atlas‐based and newer deep‐learning‐based in prescribing the four standard CMR planes but also demonstrates prescription of the first cardiac‐anatomy‐oriented plane(s) from the body‐oriented scout. Conclusions: The proposed system demonstrates accurate automatic CMR view plane prescription based on deep learning on properly archived data, without the need for further manual annotation. This work opens a new direction for automatic view planning of anatomy‐oriented medical imaging beyond CMR. [ABSTRACT FROM AUTHOR]

Background: Automatic solutions for generating radiotherapy treatment plans using deep learning (DL) have been investigated by mimicking the voxel's dose. However, plan optimization using voxel‐dose features has not been extensively studied. Purpose: This study aims to investigate the efficiency of a direct optimization strategy with finite elements (FEs) after DL dose prediction for automatic intensity‐modulated radiation therapy (IMRT) treatment planning. Methods: A double‐UNet DL model was adapted for 220 cervical cancer patients (200 for training and 20 for testing), who underwent IMRT between 2016 and 2020 at our clinic. The model inputs were computed tomography (CT) slices, organs at risk (OARs), and planning target volumes (PTVs), and the outputs were dose distributions of uniformly generated high‐dose region‐controlled plans. The FEs were discretized into equal intervals of the dose prediction value within the [OARs avoid PTV(O‐P)] and [body avoids OARs & PTV(B‐OP)] regions in the test cohort and used to define the objectives for IMRT plan optimization. The plans were optimized using a two‐step process. In the beginning, the plans of two extra cases with and without low‐dose region control were compared to pursue robust and optimal dose adjustment degree pattern of FEs. In the first step, the mean dose of O‐P FEs were constrained to differing degrees according to the pattern. The further the FEs from the PTV, the tighter the constraints. In the second step, the mean dose of O‐P FEs from first step were constrained again but weakly and the dose of the B‐OP FEs from dose prediction and PTV were tightly regulated. The dosimetric parameters of the OARs and PTV were evaluated and compared using an interstep approach. In another 10 cases, the plans optimized via the aforementioned steps (method 1) were compared with those directly generated by the double‐UNet dose prediction model trained by low and high region‐controlled plans (method 2). Results: The mean differences in dose metrics between the UNet‐predicted dose and the clinical plans were: 0.47 Gy for bladder D50%; 0.62 Gy for rectum D50%; 0% for small intestine V30Gy; 1% for small intestine V40Gy; 4% for left femoral head V30Gy; and 6% for right femoral head V30Gy. The reductions in mean dose (p < 0.001) after FE‐based optimization were: 4.0, 1.9, 2.8, 5.9, and 5.7 Gy for the bladder, rectum, small intestine, left femoral head, and right femoral head, respectively, with flat PTV homogeneity and conformity. Method 1 plans produced lower mean doses than those of method 2 for the bladder (0.7 Gy), rectum (1.0 Gy), and small intestine (0.6 Gy), while maintaining PTV homogeneity and conformity. Conclusion: FE‐based direct optimization produced lower OAR doses and adequate PTV doses after DL prediction. This solution offers rapid and automatic plan optimization without manual adjustment, particularly in low‐dose regions. [ABSTRACT FROM AUTHOR]

PurposeDomain knowledge (DK) acquired from prior studies is important for medical diagnosis. This paper leverages the population-level DK using an optimality design criterion to train a deep learning model in an end-to-end manner. In this study, the problem of interest is at the patient level to diagnose a subject with idiopathic pulmonary fibrosis (IPF) among subjects with interstitial lung disease (ILD) using a computed tomography (CT). IPF diagnosis is a complicated process with multidisciplinary discussion with experts and is subject to interobserver variability, even for experienced radiologists. To this end, we propose a new statistical method to construct a time/memory-efficient IPF diagnosis model using axial chest CT and DK, along with an optimality design criterion via a DK-enhanced loss function of deep learning.MethodsFour state-of-the-art two-dimensional convolutional neural network (2D-CNN) architectures (MobileNet, VGG16, ResNet-50, and DenseNet-121) and one baseline 2D-CNN are implemented to automatically diagnose IPF among ILD patients. Axial lung CT images are retrospectively acquired from 389 IPF patients and 700 non-IPF ILD patients in five multicenter clinical trials. To enrich the sample size and boost model performance, we sample 20 three-slice samples (triplets) from each CT scan, where these three slices are randomly selected from the top, middle, and bottom of both lungs respectively. Model performance is evaluated using a fivefold cross-validation, where each fold was stratified using a fixed proportion of IPF vs non-IPF.ResultsUsing DK-enhanced loss function increases the model performance of the baseline CNN model from 0.77 to 0.89 in terms of study-wise accuracy. Four other well-developed models reach satisfactory model performance with an overall accuracy >0.95 but the benefits brought on by the DK-enhanced loss function is not noticeable.ConclusionsWe believe this is the first attempt that (a) uses population-level DK with an optimal design criterion to train deep learning-based diagnostic models in an end-to-end manner and (b) focuses on patient-level IPF diagnosis. Further evaluation of using population-level DK on prospective studies is warranted and is underway.

Background: A clinical study regarding the potential of range verification in proton therapy (PT) by prompt gamma imaging (PGI) is carried out at our institution. Manual interpretation of the detected spot‐wise range shift information is time‐consuming, highly complex, and therefore not feasible in a broad routine application. Purpose: Here, we present an approach to automatically detect and classify treatment deviations in realistically simulated PGI data for head‐and‐neck cancer (HNC) treatments using convolutional neural networks (CNNs) and conventional machine learning (ML) approaches. Methods: For 12 HNC patients and 1 anthropomorphic head phantom (n = 13), pencil beam scanning (PBS) treatment plans were generated, and 1 field per plan was assumed to be monitored with a PGI slit camera system. In total, 386 scenarios resembling different relevant or non‐relevant treatment deviations were simulated on planning and control CTs and manually classified into 7 classes: non‐relevant changes (NR) and relevant changes (RE) triggering treatment intervention due to range prediction errors (±RP), setup errors in beam direction (±SE), anatomical changes (AC), or a combination of such errors (CB). PBS spots with reliable PGI information were considered with their nominal Bragg peak position for the generation of two 3D spatial maps of 16 × 16 × 16 voxels containing PGI‐determined range shift and proton number information. Three complexity levels of simulated PGI data were investigated: (I) optimal PGI data, (II) realistic PGI data with simulated Poisson noise based on the locally delivered proton number, and (III) realistic PGI data with an additional positioning uncertainty of the slit camera following an experimentally determined distribution. For each complexity level, 3D‐CNNs were trained on a data subset (n = 9) using patient‐wise leave‐one‐out cross‐validation and tested on an independent test cohort (n = 4). Both the binary task of detecting RE and the multi‐class task of classifying the underlying error source were investigated. Similarly, four different conventional ML classifiers (logistic regression, multilayer perceptron, random forest, and support vector machine) were trained using five previously established handcrafted features extracted from the PGI data and used for performance comparison. Results: On the test data, the CNN ensemble achieved a binary accuracy of 0.95, 0.96, and 0.93 and a multi‐class accuracy of 0.83, 0.81, and 0.76 for the complexity levels (I), (II), and (III), respectively. In the case of binary classification, the CNN ensemble detected treatment deviations in the most realistic scenario with a sensitivity of 0.95 and a specificity of 0.88. The best performing ML classifiers showed a similar test performance. Conclusions: This study demonstrates that CNNs can reliably detect relevant changes in realistically simulated PGI data and classify most of the underlying sources of treatment deviations. The CNNs extracted meaningful features from the PGI data with a performance comparable to ML classifiers trained on previously established handcrafted features. These results highlight the potential of a reliable, automatic interpretation of PGI data for treatment verification, which is highly desired for a broad clinical application and a prerequisite for the inclusion of PGI in an automated feedback loop for online adaptive PT. [ABSTRACT FROM AUTHOR]

Background: Seed implant brachytherapy (SIBT) is an effective treatment modality for head and neck (H&N) cancers; however, current clinical planning requires manual setting of needle paths and utilizes inaccurate dose calculation algorithms. Purpose: This study aims to develop an accurate and efficient deep convolutional neural network dose engine (DCNN‐DE) and an automatic SIBT planning method for H&N SIBT. Methods: A cohort of 25 H&N patients who received SIBT was utilized to develop and validate the methods. The DCNN‐DE was developed based on 3D‐unet model. It takes single seed dose distribution from a modified TG‐43 method, the CT image and a novel inter‐seed shadow map (ISSM) as inputs, and predicts the dose map of accuracy close to the one from Monte Carlo simulations (MCS). The ISSM was proposed to better handle inter‐seed attenuation. The accuracy and efficacy of the DCNN‐DE were validated by comparing with other methods taking MCS dose as reference. For SIBT planning, a novel strategy inspired by clinical practice was proposed to automatically generate parallel or non‐parallel potential needle paths that avoid puncturing bone and critical organs. A heuristic‐based optimization method was developed to optimize the seed positions to meet clinical prescription requirements. The proposed planning method was validated by re‐planning the 25 cases and comparing with clinical plans. Results: The absolute percentage error in the TG‐43 calculation for CTV V100 and D90 was reduced from 5.4% and 13.2% to 0.4% and 1.1% with DCNN‐DE, an accuracy improvement of 93% and 92%, respectively. The proposed planning method could automatically obtain a plan in 2.5 ± 1.5 min. The generated plans were judged clinically acceptable with dose distribution comparable with those of the clinical plans. Conclusions: The proposed method can generate clinically acceptable plans quickly with high accuracy in dose evaluation, and thus has a high potential for clinical use in SIBT. [ABSTRACT FROM AUTHOR]

Background: Cone-beam CT (CBCT) with non-circular scanning orbits can improve image quality for 3D intraoperative image guidance. However, geometric calibration of such scans can be challenging. Existing methods typically require a prior image, specialized phantoms, presumed repeatable orbits, or long computation time., Purpose: We propose a novel fully automatic online geometric calibration algorithm that does not require prior knowledge of fiducial configuration. The algorithm is fast, accurate, and can accommodate arbitrary scanning orbits and fiducial configurations., Methods: The algorithm uses an automatic initialization process to eliminate human intervention in fiducial localization and an iterative refinement process to ensure robustness and accuracy. We provide a detailed explanation and implementation of the proposed algorithm. Physical experiments on a lab test bench and a clinical robotic C-arm scanner were conducted to evaluate spatial resolution performance and robustness under realistic constraints., Results: Qualitative and quantitative results from the physical experiments demonstrate high accuracy, efficiency, and robustness of the proposed method. The spatial resolution performance matched that of our existing benchmark method, which used a 3D-2D registration-based geometric calibration algorithm., Conclusions: We have demonstrated an automatic online geometric calibration method that delivers high spatial resolution and robustness performance. This methodology enables arbitrary scan trajectories and should facilitate translation of such acquisition methods in a clinical setting., (© 2024 American Association of Physicists in Medicine.)

Background: Pancreatic cancer fine delineation in medical images by physicians is a major challenge due to the vast volume of medical images and the variability of patients., Purpose: A semi-automatic fine delineation scheme was designed to assist doctors in accurately and quickly delineating the cancer target region to improve the delineation accuracy of pancreatic cancer in computed tomography (CT) images and effectively reduce the workload of doctors., Methods: A target delineation scheme in image blocks was also designed to provide more information for the deep learning delineation model. The start and end slices of the image block were manually delineated by physicians, and the cancer in the middle slices were accurately segmented using a three-dimensional Res U-Net model. Specifically, the input of the network is the CT image of the image block and the delineation of the cancer in the start and end slices, while the output of the network is the cancer area in the middle slices of the image block. Meanwhile, the model performance of pancreatic cancer delineation and the workload of doctors in different image block sizes were studied., Results: We used 37 3D CT volumes for training, 11 volumes for validating and 11 volumes for testing. The influence of different image block sizes on doctors' workload was compared quantitatively. Experimental results showed that the physician's workload was minimal when the image block size was 5, and all cancer could be accurately delineated. The Dice similarity coefficient was 0.894 ± 0.029, the 95% Hausdorff distance was 3.465 ± 0.710 mm, the normalized surface Dice was 0.969 ± 0.019. By completing the accurate delineation of all the CT images, the speed of the new method is 2.16 times faster than that of manual sketching., Conclusion: Our proposed 3D semi-automatic delineative method based on the idea of block prediction could accurately delineate CT images of pancreatic cancer and effectively deal with the challenges of class imbalance, background distractions, and non-rigid geometrical features. This study had a significant advantage in reducing doctors' workload, and was expected to help doctors improve their work efficiency in clinical application., (© 2023 American Association of Physicists in Medicine.)

Background: The quality of segmentation of thyroid nodules in ultrasound images is a crucial factor in preventing the cancerization of thyroid nodules. However, the existing standards for the ultrasound imaging of cancerous nodules have limitations, and changes of the echo pattern of thyroid nodules pose challenges in accurately segmenting nodules, which can affect the diagnostic results of medical professionals., Purpose: The aim of this study is to address the challenges related to segmentation accuracy due to noise, low contrast, morphological scale variations, and blurred edges of thyroid nodules in ultrasound images and improve the accuracy of ultrasound-based thyroid nodule segmentation, thereby aiding the clinical diagnosis of thyroid nodules., Method: In this study, the dataset of thyroid ultrasound images was obtained from Hunan Provincial People's Hospital, consisting of a total of 3572 samples used for the training, validation, and testing of this model at a ratio of 8:1:1. A novel SK-Unet++ network was used to enhance the segmentation accuracy of thyroid nodules. SK-Unet++ is a novel deep learning architecture that adds the adaptive receptive fields based on the selective kernel (SK) attention mechanisms into the Unet++ network. The convolution blocks of the original UNet++ encoder part were replaced with finer SK convolution blocks in SK-Unet++. First, multiple skip connections were incorporated so that SK-Unet++ can make information from previous layers of the neural network to bypass certain layers and directly propagate to subsequent layers. The feature maps of the corresponding locations were fused on the channel, resulting in enhanced segmentation accuracy. Second, we added the adaptive receptive fields. The adaptive receptive fields were used to capture multiscale spatial features better by dynamically adjusting its receptive field. The assessment metrics contained dice similarity coefficient (Dsc), accuracy (Acc), precision (Pre), recall (Re), and Hausdorff distance, and all comparison experiments used the paired t-tests to assess whether statistically significant performance differences existed (p < 0.05). And to address the multi-comparison problem, we performed the false discovery rate (FDR) correction after the test., Results: The segmentation model had an Acc of 80.6%, Dsc of 84.7%, Pre of 77.5%, Re of 71.7%, and an average Hausdorff distance of 15.80 mm. Ablation experimental results demonstrated that each module in the network could contribute to the improved performance (p < 0.05) and determined the best combination of parameters. A comparison with other state-of-the-art methods showed that SK-Unet++ significantly outperformed them in terms of segmentation performance (p < 0.05), with a more accurate segmentation contour. Additionally, the adaptive weight changes of the SK module were monitored during the training process, and the resulting change curves demonstrated their convergence., Conclusion: Our proposed method demonstrates favorable performance in the segmentation of ultrasound images of thyroid nodules. Results confirmed that SK-Unet++ is a feasible and effective method for the automatic segmentation of thyroid nodules in ultrasound images. The high accuracy achieved by our method can facilitate efficient screening of patients with thyroid nodules, ultimately reducing the workload of clinicians and radiologists., (© 2023 American Association of Physicists in Medicine.)

Background: It is necessary to contour regions of interest (ROIs) for online magnetic resonance imaging (MRI)-guided adaptive radiotherapy (MRIgART). These updated contours are used for online replanning to obtain maximum dosimetric benefits. Contouring can be accomplished using deformable image registration (DIR) and deep learning (DL)-based autosegmentation methods. However, these methods may require considerable manual editing and thus prolong treatment time., Purpose: The present study aimed to improve autosegmentation performance by integrating patients' pretreatment information in a DL-based segmentation algorithm. It is expected to improve the efficiency of current MRIgART process., Methods: Forty patients with prostate cancer were enrolled retrospectively. The online adaptive MR images, patient-specific planning computed tomography (CT), and contours in CT were used for segmentation. The deformable registration of planning CT and MR images was performed first to obtain a deformable CT and corresponding contours. A novel DL network, which can integrate such patient-specific information (deformable CT and corresponding contours) into the segmentation task of MR images was designed. We performed a four-fold cross-validation for the DL models. The proposed method was compared with DIR and DL methods on segmentation of prostate cancer. The ROIs included the clinical target volume (CTV), bladder, rectum, left femur head, and right femur head. Dosimetric parameters of automatically generated ROIs were evaluated using a clinical treatment planning system., Results: The proposed method enhanced the segmentation accuracy of conventional procedures. Its mean value of the dice similarity coefficient (93.5%) over the five ROIs was higher than both DIR (87.5%) and DL (87.2%). The number of patients (n = 40) that required major editing using DIR, DL, and our method were 12, 18, and 7 (CTV); 17, 4, and 1 (bladder); 8, 11, and 5 (rectum); 2, 4, and 1 (left femur head); and 3, 7, and 1 (right femur head), respectively. The Spearman rank correlation coefficient of dosimetry parameters between the proposed method and ground truth was 0.972 ± 0.040, higher than that of DIR (0.897 ± 0.098) and DL (0.871 ± 0.134)., Conclusion: This study proposed a novel method that integrates patient-specific pretreatment information into DL-based segmentation algorithm. It outperformed baseline methods, thereby improving the efficiency and segmentation accuracy in adaptive radiotherapy., (© 2023 American Association of Physicists in Medicine.)

PurposeSegmentation of cardiac medical images, an important step in measuring cardiac function, is usually performed either manually or semiautomatically. Fully automatic segmentation of the left ventricle (LV), the right ventricle (RV) as well as the myocardium of three-dimensional (3D) magnetic resonance (MR) images throughout the entire cardiac cycle (four-dimensional, 4D), remains challenging. This study proposes a deformable-based segmentation methodology for efficiently segmenting 4D (3D + t) cardiac MR images.MethodsThe proposed methodology first used the Hough transform and the local Gaussian distribution method (LGD) to segment the LV endocardial contours from cardiac MR images. Following this, a novel level set-based shape prior method was applied to generate the LV epicardial contours and the RV boundary.ResultsThis automatic image segmentation approach has been applied to studies on 17 subjects. The results demonstrated that the proposed method was efficient compared to manual segmentation, achieving a segmentation accuracy with average Dice values of 88.62 ± 5.47%, 87.35 ± 7.26%, and 82.63 ± 6.22% for the LV endocardial, LV epicardial, and RV contours, respectively.ConclusionsWe have presented a method for accurate LV and RV segmentation. Compared to three existing methods, the proposed method can successfully segment the LV and yield the highest Dice value. This makes it an option for clinical assessment of the volume, size, and thickness of the ventricles.

Background: Medical image segmentation is one of the most key steps in computer-aided clinical diagnosis, geometric characterization, measurement, image registration, and so forth. Convolutional neural networks especially UNet and its variants have been successfully used in many medical image segmentation tasks. However, the results are limited by the deficiency in extracting high resolution edge information because of the design of the skip connections in UNet and the need for large available datasets., Purpose: In this paper, we proposed an edge-attending polar UNet (EPolar-UNet), which was trained on the polar coordinate system instead of classic Cartesian coordinate system with an edge-attending construction in skip connection path., Methods: EPolar-UNet extracted the location information from an eight-stacked hourglass network as the pole for polar transformation and extracted the boundary cues from an edge-attending UNet, which consisted of a deconvolution layer and a subtraction operation., Results: We evaluated the performance of EPolar-UNet across three imaging modalities for different segmentation tasks: CVC-ClinicDB dataset for polyp, ISIC-2018 dataset for skin lesion, and our private ultrasound dataset for liver tumor segmentation. Our proposed model outperformed state-of-the-art models on all three datasets and needed only 30%-60% of training data compared with the benchmark UNet model to achieve similar performances for medical image segmentation tasks., Conclusions: We proposed an end-to-end EPolar-UNet for automatic medical image segmentation and showed good performance on small datasets, which was critical in the field of medical image segmentation., (© 2024 American Association of Physicists in Medicine.)

Background: Despite extensive efforts to obtain accurate segmentation of magnetic resonance imaging (MRI) scans of a head, it remains challenging primarily due to variations in intensity distribution, which depend on the equipment and parameters used., Purpose: The goal of this study is to evaluate the effectiveness of an automatic segmentation method for head MRI scans using a multistep Dense U-Net (MDU-Net) architecture., Methods: The MDU-Net-based method comprises two steps. The first step is to segment the scalp, skull, and whole brain from head MRI scans using a convolutional neural network (CNN). In the first step, a hybrid network is used to combine 2.5D Dense U-Net and 3D Dense U-Net structure. This hybrid network acquires logits in three orthogonal planes (axial, coronal, and sagittal) using 2.5D Dense U-Nets and fuses them by averaging. The resultant fused probability map with head MRI scans then serves as the input to a 3D Dense U-Net. In this process, different ratios of active contour loss and focal loss are applied. The second step is to segment the cerebrospinal fluid (CSF), white matter, and gray matter from extracted brain MRI scans using CNNs. In the second step, the histogram of the extracted brain MRI scans is standardized and then a 2.5D Dense U-Net is used to further segment the brain's specific tissues using the focal loss. A dataset of 100 head MRI scans from an OASIS-3 dataset was used for training, internal validation, and testing, with ratios of 80%, 10%, and 10%, respectively. Using the proposed approach, we segmented the head MRI scans into five areas (scalp, skull, CSF, white matter, and gray matter) and evaluated the segmentation results using the Dice similarity coefficient (DSC) score, Hausdorff distance (HD), and the average symmetric surface distance (ASSD) as evaluation metrics. We compared these results with those obtained using the Res-U-Net, Dense U-Net, U-Net++, Swin-Unet, and H-Dense U-Net models., Results: The MDU-Net model showed DSC values of 0.933, 0.830, 0.833, 0.953, and 0.917 in the scalp, skull, CSF, white matter, and gray matter, respectively. The corresponding HD values were 2.37, 2.89, 2.13, 1.52, and 1.53 mm, respectively. The ASSD values were 0.50, 1.63, 1.28, 0.26, and 0.27 mm, respectively. Comparing these results with other models revealed that the MDU-Net model demonstrated the best performance in terms of the DSC values for the scalp, CSF, white matter, and gray matter. When compared with the H-Dense U-Net model, which showed the highest performance among the other models, the MDU-Net model showed substantial improvements in the HD view, particularly in the gray matter region, with a difference of approximately 9%. In addition, in terms of the ASSD, the MDU-Net model outperformed the H-Dense U-Net model, showing an approximately 7% improvements in the white matter and approximately 9% improvements in the gray matter., Conclusion: Compared with existing models in terms of DSC, HD, and ASSD, the proposed MDU-Net model demonstrated the best performance on average and showed its potential to enhance the accuracy of automatic segmentation for head MRI scans., (© 2023 American Association of Physicists in Medicine.)

Background: Functional assessment of right ventricle (RV) using gated myocardial perfusion single-photon emission computed tomography (MPS) heavily relies on the precise extraction of right ventricular contours., Purpose: In this paper, we present a new deep-learning-based model integrating both the spatial and temporal features in gated MPS images to perform the segmentation of the RV epicardium and endocardium., Methods: By integrating the spatial features from each cardiac frame of the gated MPS and the temporal features from the sequential cardiac frames of the gated MPS, we developed a Spatial-Temporal V-Net (ST-VNet) for automatic extraction of RV endocardial and epicardial contours. In the ST-VNet, a V-Net is employed to hierarchically extract spatial features, and convolutional long-term short-term memory (ConvLSTM) units are added to the skip-connection pathway to extract the temporal features. The input of the ST-VNet is ECG-gated sequential frames of the MPS images and the output is the probability map of the epicardial or endocardial masks. A Dice similarity coefficient (DSC) loss which penalizes the discrepancy between the model prediction and the manual annotation was adopted to optimize the segmentation model., Results: Our segmentation model was trained and validated on a retrospective dataset with 45 subjects, and the cardiac cycle of each subject was divided into eight gates. The proposed ST-VNet achieved a DSC of 0.8914 and 0.8157 for the RV epicardium and endocardium segmentation, respectively. The mean absolute error, the mean squared error, and the Pearson correlation coefficient of the RV ejection fraction (RVEF) between the manual annotation and the model prediction were 0.0609, 0.0830, and 0.6985., Conclusion: Our proposed ST-VNet is an effective model for RV segmentation. It has great promise for clinical use in RV functional assessment., (© 2023 American Association of Physicists in Medicine.)

Purpose: Deep learning-based networks have become increasingly popular in the field of medical image segmentation. The purpose of this research was to develop and optimize a new architecture for automatic segmentation of the prostate gland and normal organs in the pelvic, thoracic, and upper gastro-intestinal (GI) regions., Methods: We developed an architecture which combines a shifted-window (Swin) transformer with a convolutional U-Net. The network includes a parallel encoder, a cross-fusion block, and a CNN-based decoder to extract local and global information and merge related features on the same scale. A skip connection is applied between the cross-fusion block and decoder to integrate low-level semantic features. Attention gates (AGs) are integrated within the CNN to suppress features in image background regions. Our network is termed "SwinAttUNet." We optimized the architecture for automatic image segmentation. Training datasets consisted of planning-CT datasets from 300 prostate cancer patients from an institutional database and 100 CT datasets from a publicly available dataset (CT-ORG). Images were linearly interpolated and resampled to a spatial resolution of (1.0 × 1.0× 1.5) mm 3 . A volume patch (192 × 192 × 96) was used for training and inference, and the dataset was split into training (75%), validation (10%), and test (15%) cohorts. Data augmentation transforms were applied consisting of random flip, rotation, and intensity scaling. The loss function comprised Dice and cross-entropy equally weighted and summed. We evaluated Dice coefficients (DSC), 95th percentile Hausdorff Distances (HD95), and Average Surface Distances (ASD) between results of our network and ground truth data., Results: SwinAttUNet, DSC values were 86.54 ± 1.21, 94.15 ± 1.17, and 87.15 ± 1.68% and HD95 values were 5.06 ± 1.42, 3.16 ± 0.93, and 5.54 ± 1.63 mm for the prostate, bladder, and rectum, respectively. Respective ASD values were 1.45 ± 0.57, 0.82 ± 0.12, and 1.42 ± 0.38 mm. For the lung, liver, kidneys and pelvic bones, respective DSC values were: 97.90 ± 0.80, 96.16 ± 0.76, 93.74 ± 2.25, and 89.31 ± 3.87%. Respective HD95 values were: 5.13 ± 4.11, 2.73 ± 1.19, 2.29 ± 1.47, and 5.31 ± 1.25 mm. Respective ASD values were: 1.88 ± 1.45, 1.78 ± 1.21, 0.71 ± 0.43, and 1.21 ± 1.11 mm. Our network outperformed several existing deep learning approaches using only attention-based convolutional or Transformer-based feature strategies, as detailed in the results section., Conclusions: We have demonstrated that our new architecture combining Transformer- and convolution-based features is able to better learn the local and global context for automatic segmentation of multi-organ, CT-based anatomy., (© 2023 American Association of Physicists in Medicine.)

Background: The high morbidity and mortality of heart disease present a significant threat to human health. The development of methods for the quick and accurate diagnosis of heart diseases, enabling their effective treatment, has become a key issue of concern. Right ventricular (RV) segmentation from cine cardiac magnetic resonance (CMR) images plays a significant role in evaluating cardiac function for clinical diagnosis and prognosis. However, due to the complex structure of the RV, traditional segmentation methods are ineffective for RV segmentation., Purpose: In this paper, we propose a new deep atlas network to improve the learning efficiency and segmentation accuracy of a deep learning network by integrating multi-atlas., Methods: First, a dense multi-scale U-net (DMU-net) is presented to acquire transformation parameters from atlas images to target images. The transformation parameters map the atlas image labels to the target image labels. Second, using a spatial transformation layer, the atlas images are deformed based on these parameters. Finally, the network is optimized by backpropagation with two loss functions where the mean squared error function (MSE) is used to measure the similarity of the input images and transformed images. Further, the Dice metric (DM) is used to quantify the overlap between the predicted contours and the ground truth. In our experiments, 15 datasets are used in testing, and 20 cine CMR images are selected as atlas., Results: The mean values and standard deviations for the DM and Hausdorff distance are 0.871 and 4.67 mm, 0.104 and 2.528 mm, respectively. The correlation coefficients of endo-diastolic volume, endo-systolic volume, ejection fraction, and stroke volume are 0.984, 0.926, 0.980, and 0.991, respectively, and the mean differences between all of the mentioned parameters are 3.2, -1.7, 0.02, and 4.9, respectively. Most of these differences are within the allowable range of 95%, indicating that the results are acceptable and show good consistency. The segmentation results obtained in this method are compared with those obtained by other methods that provide satisfactory performance. The other methods provide better segmentation effects at the base, but either no segmentation or the wrong segmentation at the top, which demonstrate that the deep atlas network can improve top-area segmentation accuracy., Conclusion: Our results indicate that the proposed method can achieve better segmentation results than the previous methods, with both high relevance and consistency, and has the potential for clinical application., (© 2023 American Association of Physicists in Medicine.)

Background: The fusion of computed tomography (CT) and ultrasound (US) image can enhance lesion detection ability and improve the success rate of liver interventional radiology. The image-based fusion methods encounter the challenge of registration initialization due to the random scanning pose and limited field of view of US. Existing automatic methods those used vessel geometric information and intensity-based metric are sensitive to parameters and have low success rate. The learning-based methods require a large number of registered datasets for training., Purpose: The aim of this study is to provide a fully automatic and robust US-3D CT registration method without registered training data and user-specified parameters assisted by the revolutionary deep learning-based segmentation, which can further be used for preparing training samples for the study of learning-based methods., Methods: We propose a fully automatic CT-3D US registration method by two improved registration metrics. We propose to use 3D U-Net-based multi-organ segmentation of US and CT to assist the conventional registration. The rigid transform is searched in the space of any paired vessel bifurcation planes where the best transform is decided by a segmentation overlap metric, which is more related to the segmentation precision than Dice coefficient. In nonrigid registration phase, we propose a hybrid context and edge based image similarity metric with a simple mask that can remove most noisy US voxels to guide the B-spline transform registration. We evaluate our method on 42 paired CT-3D US datasets scanned with two different US devices from two hospitals. We compared our methods with other exsiting methods with both quantitative measures of target registration error (TRE) and the Jacobian determinent with paired t-test and qualitative registration imaging results., Results: The results show that our method achieves fully automatic rigid registration TRE of 4.895 mm, deformable registration TRE of 2.995 mm in average, which outperforms state-of-the-art automatic linear methods and nonlinear registration metrics with paired t-test's p value less than 0.05. The proposed overlap metric achieves better results than self similarity description (SSD), edge matching (EM), and block matching (BM) with p values of 1.624E-10, 4.235E-9, and 0.002, respectively. The proposed hybrid edge and context-based metric outperforms context-only, edge-only, and intensity statistics-only-based metrics with p values of 0.023, 3.81E-5, and 1.38E-15, respectively. The 3D US segmentation has achieved mean Dice similarity coefficient (DSC) of 0.799, 0.724, 0.788, and precision of 0.871, 0.769, 0.862 for gallbladder, vessel, and branch vessel, respectively., Conclusions: The deep learning-based US segmentation can achieve satisfied result to assist robust conventional rigid registration. The Dice similarity coefficient-based metrics, hybrid context, and edge image similarity metric contribute to robust and accurate registration., (© 2023 American Association of Physicists in Medicine.)

PurposeTo develop and evaluate a new automatic classification algorithm to identify voxels containing skin, vasculature, adipose, and fibroglandular tissue in dedicated breast CT images.MethodsThe proposed algorithm combines intensity- and region-based segmentation methods with energy minimizing splines and unsupervised data mining approaches for classifying and segmenting the different tissue types. Breast skin segmentation is achieved by a region-growing method which uses constraints from the previously extracted skin centerline to add robustness to the model and to reduce the false positive rate. An energy minimizing active contour model is then used to classify adipose tissue voxels by including gradient flow and region-based features. Finally, blood vessels are separated from fibroglandular tissue by a k-means clustering algorithm based on automatically extracted shape-based features. To evaluate the accuracy of the algorithm, two sets of 15 different patient breast CT scans, each acquired with different breast CT systems and acquisition settings were obtained. Three slices from each scan were manually segmented under the supervision of an experienced breast radiologist and considered the gold standard. Comparisons with manual segmentation were quantified using five similarity metrics: Dice similarity coefficient (DSC), sensitivity, conformity coefficient, and two Hausdorff distance measures. To evaluate the robustness to image noise, the segmentation was repeated after separately adding Gaussian noise with increasing standard deviation (in four steps, from 0.01 to 0.04) to an additional 15 slices from the first dataset. In addition, to evaluate vasculature classification, three different pre- and postcontrast injection patient breast CT images were classified and compared. Finally, DSC was also used for quantitative comparisons with previously proposed approaches for breast CT tissue classification using 10 images from the first dataset.ResultsThe algorithm showed a high accuracy in classifying the different tissue types for both breast CT systems, with an average DSC of 95% and 90% for the first and second image dataset, respectively. Furthermore, it demonstrated to be robust to image noise with a robustness to image noise of 85%, 83%, 79%, and 71% for the images corrupted with the four increasing noise levels. Previous methods for breast tissues classification resulted, for the tested dataset, in an average global DSC of 87%, while our approach resulted in a global average DSC of 94.5%.ConclusionsThe proposed algorithm resulted in accurate and robust breast tissue classification, with no prior training or threshold setting. Potential applications include breast density quantification and tissue pattern characterization (both biomarkers of cancer development), simulation-based radiation dose analysis, and patient data-based phantom design, which could be used for further breast imaging research.

Purpose: Automated treatment planning strategies are being widely implemented in clinical routines to reduce inter-planner variability, speed up the optimization process, and improve plan quality. This study aims to evaluate the feasibility and quality of intensity-modulated proton therapy (IMPT) plans generated with four different knowledge-based planning (KBP) pipelines fully integrated into a commercial treatment planning system (TPS)., Materials/methods: A data set containing 60 oropharyngeal cancer patients was split into 11 folds, each containing 47 patients for training, five patients for validation, and five patients for testing. A dose prediction model was trained on each of the folds, resulting in a total of 11 models. Three patients were left out in order to assess if the differences introduced between models were significant. From voxel-based dose predictions, we analyze the two steps that follow the dose prediction: post-processing of the predicted dose and dose mimicking (DM). We focused on the effect of post-processing (PP) or no post-processing (NPP) combined with two different DM algorithms for optimization: the one available in the commercial TPS RayStation (RSM) and a simpler isodose-based mimicking (IBM). Using 55 test patients (five test patients for each model), we evaluated the quality and robustness of the plans generated by the four proposed KBP pipelines (PP-RSM, PP-IBM, NPP-RSM, NPP-IBM). After robust evaluation, dose-volume histogram (DVH) metrics in nominal and worst-case scenarios were compared to those of the manually generated plans., Results: Nominal doses from the four KBP pipelines showed promising results achieving comparable target coverage and improved dose to organs at risk (OARs) compared to the manual plans. However, too optimistic post-processing applied to the dose prediction (i.e. important decrease of the dose to the organs) compromised the robustness of the plans. Even though RSM seemed to partially compensate for the lack of robustness in the PP plans, still 65% of the patients did not achieve the expected robustness levels. NPP-RSM plans seemed to achieve the best trade-off between robustness and OAR sparing., Discussion/conclusions: PP and DM strategies are crucial steps to generate acceptable robust and deliverable IMPT plans from ML-predicted doses. Before the clinical implementation of any KBP pipeline, the PP and DM parameters predefined by the commercial TPS need to be modified accordingly with a comprehensive feedback loop in which the robustness of the final dose calculations is evaluated. With the right choice of PP and DM parameters, KBP strategies have the potential to generate IMPT plans within clinically acceptable levels comparable to plans manually generated by dosimetrists., (© 2023 American Association of Physicists in Medicine.)

Objectives: Radiotherapy plays an important role in the treatment of nonsmall‐cell lung cancer (NSCLC). Accurate delineation of tumor is the key to successful radiotherapy. Compared with the commonly used manual delineation ways, which are time‐consuming and laborious, the automatic segmentation methods based on deep learning can greatly improve the treatment efficiency. Methods: In this paper, we introduce an automatic segmentation method by combining coarse and fine segmentations for NSCLC. Coarse segmentation network is the first level, identifing the rough region of the tumor. In this network, according to the tissue structure distribution of the thoracic cavity where tumor is located, we designed a competition method between tumors and organs at risk (OARs), which can increase the proportion of the identified tumor covering the ground truth and reduce false identification. Fine segmentation network is the second level, carrying out precise segmentation on the results of the coarse level. These two networks are independent of each other during training. When they are used, morphological processing of small scale corrosion and large scale expansion is used for the coarse segmentation results, and the outcomes are sent to the fine segmentation part as input, so as to achieve the complementary advantages of the two networks. Results: In the experiment, CT images of 200 patients with NSCLC are used to train the network, and CT images of 60 patients are used to test. Finally, our method produced the Dice similarity coefficient of 0.78 ± 0.10. Conclusions: The experimental results show that the proposed method can accurately segment the tumor with NSCLC, and can also provide support for clinical diagnosis and treatment. [ABSTRACT FROM AUTHOR]

PURPOSE:Intensity modulated radiation therapy (IMRT) is commonly employed for treating head and neck (H&N) cancer with uniform tumor dose and conformal critical organ sparing. Accurate delineation of organs-at-risk (OARs) on H&N CT images is thus essential to treatment quality. Manual contouring used in current clinical practice is tedious, time-consuming, and can produce inconsistent results. Existing automated segmentation methods are challenged by the substantial inter-patient anatomical variation and low CT soft tissue contrast. To overcome the challenges, we developed a novel automated H&N OARs segmentation method that combines a fully convolutional neural network (FCNN) with a shape representation model (SRM). METHODS:Based on manually segmented H&N CT, the SRM and FCNN were trained in two steps: (a) SRM learned the latent shape representation of H&N OARs from the training dataset; (b) the pre-trained SRM with fixed parameters were used to constrain the FCNN training. The combined segmentation network was then used to delineate nine OARs including the brainstem, optic chiasm, mandible, optical nerves, parotids, and submandibular glands on unseen H&N CT images. Twenty-two and 10 H&N CT scans provided by the Public Domain Database for Computational Anatomy (PDDCA) were utilized for training and validation, respectively. Dice similarity coefficient (DSC), positive predictive value (PPV), sensitivity (SEN), average surface distance (ASD), and 95% maximum surface distance (95%SD) were calculated to quantitatively evaluate the segmentation accuracy of the proposed method. The proposed method was compared with an active appearance model that won the 2015 MICCAI H&N Segmentation Grand Challenge based on the same dataset, an atlas method and a deep learning method based on different patient datasets. RESULTS:An average DSC = 0.870 (brainstem), DSC = 0.583 (optic chiasm), DSC = 0.937 (mandible), DSC = 0.653 (left optic nerve), DSC = 0.689 (right optic nerve), DSC = 0.835 (left parotid), DSC = 0.832 (right parotid), DSC = 0.755 (left submandibular), and DSC = 0.813 (right submandibular) were achieved. The segmentation results are consistently superior to the results of atlas and statistical shape based methods as well as a patch-wise convolutional neural network method. Once the networks are trained off-line, the average time to segment all 9 OARs for an unseen CT scan is 9.5 s. CONCLUSION:Experiments on clinical datasets of H&N patients demonstrated the effectiveness of the proposed deep neural network segmentation method for multi-organ segmentation on volumetric CT scans. The accuracy and robustness of the segmentation were further increased by incorporating shape priors using SMR. The proposed method showed competitive performance and took shorter time to segment multiple organs in comparison to state of the art methods.

PURPOSE: The vast majority of body CT exams are performed with automatic exposure control (AEC), which adapts the mean tube current to the patient size and modulates the tube current either angularly, longitudinally or both. However, most radiation dose estimation tools are based on fixed tube current scans. Accurate estimates of patient dose from AEC scans require knowledge of the tube current values, which is usually unavailable. The purpose of this work was to develop and validate methods to accurately estimate the tube current values prescribed by one manufacturers AEC system to enable accurate estimates of patient dose. METHODS: Methods were developed that took into account available patient attenuation information, user selected image quality reference parameters and x-ray system limits to estimate tube current values for patient scans. Methods consistent with AAPM Report 220 were developed that used patient attenuation data that were: (a) supplied by the manufacturer in the CT localizer radiograph and (b) based on a simulated CT localizer radiograph derived from image data. For comparison, actual tube current values were extracted from the projection data of each patient. Validation of each approach was based on data collected from 40 pediatric and adult patients who received clinically indicated chest (n = 20) and abdomen/pelvis (n = 20) scans on a 64 slice multidetector row CT (Sensation 64, Siemens Healthcare, Forchheim, Germany). For each patient dataset, the following were collected with Institutional Review Board (IRB) approval: (a) projection data containing actual tube current values at each projection view, (b) CT localizer radiograph (topogram) and (c) reconstructed image data. Tube current values were estimated based on the actual topogram (actual-topo) as well as the simulated topogram based on image data (sim-topo). Each of these was compared to the actual tube current values from the patient scan. In addition, to assess the accuracy of each method in estimating patient organ doses, Monte Carlo simulations were performed by creating voxelized models of each patient, identifying key organs and incorporating tube current values into the simulations to estimate dose to the lungs and breasts (females only) for chest scans and the liver, kidney, and spleen for abdomen/pelvis scans. Organ doses from simulations using the actual tube current values were compared to those using each of the estimated tube current values (actual-topo and sim-topo). RESULTS: When compared to the actual tube current values, the average error for tube current values estimated from the actual topogram (actual-topo) and simulated topogram (sim-topo) was 3.9% and 5.8% respectively. For Monte Carlo simulations of chest CT exams using the actual tube current values and estimated tube current values (based on the actual-topo and sim-topo methods), the average differences for lung and breast doses ranged from 3.4% to 6.6%. For abdomen/pelvis exams, the average differences for liver, kidney, and spleen doses ranged from 4.2% to 5.3%. CONCLUSIONS: Strong agreement between organ doses estimated using actual and estimated tube current values provides validation of both methods for estimating tube current values based on data provided in the topogram or simulated from image data.

Background: High dose rate (HDR) brachytherapy is commonly used to treat prostate cancer. Existing HDR planning systems solve the dwell time problem for predetermined catheters and a single energy source., Purpose: Additional degrees of freedom can be obtained by relaxing the catheters' pre-designation and introducing more source types, and may have a dosimetric benefit, particularly in improving conformality to spare the urethra. This study presents a novel analytical approach to solving the corresponding HDR planning problem., Methods: The catheter and dual-energy source selection problem was formulated as a constrained optimization problem with a non-convex group sparsity regularization. The optimization problem was solved using the fast-iterative shrinkage-thresholding algorithm (FISTA). Two isotopes were considered. The dose rates for the HDR 4140 Ytterbium (Yb-169) source and the Elekta Iridium (Ir-192) HDR Flexisource were modeled according to the TG-43U1 formalism and benchmarked accordingly. Twenty-two retrospective HDR prostate brachytherapy patients treated with Ir-192 were considered. An Ir-192 only (IRO), Yb-169 only (YBO), and dual-source (DS) plan with optimized catheter location was created for each patient with N catheters, where N is the number of catheters used in the clinically delivered plans. The DS plans jointly optimized Yb-169 and Ir-192 dwell times. All plans and the clinical plans were normalized to deliver a 15 Gy prescription (Rx) dose to 95% of the clinical treatment volume (CTV) and evaluated for the CTV D90%, V150%, and V200%, urethra D0.1cc and D1cc, bladder V75%, and rectum V75%. Dose-volume histograms (DVHs) were generated for each structure., Results: The DS plans ubiquitously selected Ir-192 as the only treatment source. IRO outperformed YBO in organ at risk (OARs) OAR sparing, reducing the urethra D0.1cc and D1cc by 0.98% ( p = 2.22 ∗ 10 - 9 $p\ = \ 2.22*{10^{ - 9}}$ ) and 1.09% ( p = 1.22 ∗ 10 - 10 $p\ = \ 1.22*{10^{ - 10}}$ ) of the Rx dose, respectively, and reducing the bladder and rectum V75% by 0.09 ( p = 0.0023 $p\ = \ 0.0023$ ) and 0.13 cubic centimeters (cc) ( p = 0.033 $p\ = \ 0.033$ ), respectively. The YBO plans delivered a more homogenous dose to the CTV, with a smaller V150% and V200% by 3.20 ( p = 4.67 ∗ 10 - 10 $p\ = \ 4.67*{10^{ - 10}}$ ) and 1.91 cc ( p = 5.79 ∗ 10 - 10 $p\ = \ 5.79*{10^{ - 10}}$ ), respectively, and a lower CTV D90% by 0.49% ( p = 0.0056 $p\ = \ 0.0056$ ) of the prescription dose. The IRO plans reduce the urethral D1cc by 2.82% ( p = 1.38 ∗ 10 - 4 $p\ = \ 1.38*{10^{ - 4}}$ ) of the Rx dose compared to the clinical plans, at the cost of increased bladder and rectal V75% by 0.57 ( p = 0.0022 $p\ = \ 0.0022$ ) and 0.21 cc ( p = 0.019 $p\ = \ 0.019$ ), respectively, and increased CTV V150% by a mean of 1.46 cc ( p = 0.010 $p\ = \ 0.010$ ) and CTV D90% by an average of 1.40% of the Rx dose ( p = 8.80 ∗ 10 - 8 $p\ = \ 8.80*{10^{ - 8}}$ ). While these differences are statistically significant, the clinical differences between the plans are minimal., Conclusions: The proposed analytical HDR planning algorithm integrates catheter and isotope selection with dwell time optimization for varying clinical goals, including urethra sparing. The planning method can guide HDR implants and identify promising isotopes for specific HDR clinical goals, such as target conformality or OAR sparing., (© 2023 American Association of Physicists in Medicine.)

Background: Pulmonary embolism (PE) is a common but fatal clinical condition and the gold standard of diagnosis is computed tomography pulmonary angiography (CTPA). Prompt diagnosis and rapid treatment can dramatically reduce mortality in patients. However, the diagnosis of PE is often delayed and missed., Methods: In this study, we identified a deep learning model Scaled-YOLOv4 that enables end-to-end automated detection of PE to help solve these problems. A total of 307 CTPA data (Tianjin 142 cases, Linyi 133 cases, and FUMPE 32 cases) were included in this study. The Tianjin dataset was divided 10 times in the ratio of training set: validation set: test set = 7:2:1 for model tuning, and both the Linyi and FUMPE datasets were used as independent external test sets to evaluate the generalization of the model., Results: Scaled-YOLOv4 was able to process one patient in average 3.55 s [95% CI: 3.51-3.59 s]. It also achieved an average precision (AP) of 83.04 [95% CI: 79.36-86.72] for PE detection on the Tianjin test set, and 75.86 [95% CI: 75.48-76.24] and 72.74 [95% CI: 72.10-73.38] on Linyi and FUMPE, respectively., Conclusions: This deep learning algorithm helps detect PE in real time, providing radiologists with aided diagnostic evidence without increasing their workload, and can effectively reduce the probability of delayed patient diagnosis., (© 2023 American Association of Physicists in Medicine.)

Purpose: The thrombus in the false lumen (FL) of aortic dissection (AD) patients is a meaningful indicator to determine aortic remodeling but difficult to measure in clinic. In this study, a novel segmentation strategy based on deep learning was proposed to automatically extract the thrombus in the FL in post-operative computed tomography angiography (CTA) images of AD patients, which provided an efficient and convenient segmentation method with high accuracy., Methods: A two-step segmentation strategy was proposed. Each step contained a convolutional neural network (CNN) to segment the aorta and the thrombus, respectively. In the first step, a CNN was used to obtain the binary segmentation mask of the whole aorta. In the second step, another CNN was introduced to segment the thrombus. The results of the first step were used as additional input to the second step to highlight the aorta in the complex background. Moreover, skip connection attention refinement (SAR) modules were designed and added in the second step to improve the segmentation accuracy of the thrombus details by efficiently using the low-level features., Results: The proposed method provided accurate thrombus segmentation results (0.903 ± 0.062 in dice score, 0.828 ± 0.092 in Jaccard index, and 2.209 ± 2.945 in 95% Hausdorff distance), which showed improvement compared to the methods without prior information (0.846 ± 0.085 in dice score) and the method without SAR (0.899 ± 0.060 in dice score). Moreover, the proposed method achieved 0.967 ± 0.029 and 0.948 ± 0.041 in dice score of true lumen (TL) and patent FL (PFL) segmentation, respectively, indicating the excellence of the proposed method in the segmentation task of the overall aorta., Conclusions: A novel CNN-based segmentation framework was proposed to automatically obtain thrombus segmentation for thrombosed AD in post-operative CTA images, which provided a useful tool for further application of thrombus-related indicators in clinical and research application., (© 2022 American Association of Physicists in Medicine.)

Purpose: Automatic liver segmentation from computed tomography (CT) images is an essential preprocessing step for computer-aided diagnosis of liver diseases. However, due to the large differences in liver shapes, low-contrast to adjacent tissues, and existence of tumors or other abnormalities, liver segmentation has been very challenging. This study presents an accurate and fast liver segmentation method based on a novel probabilistic active contour (PAC) model and its fast global minimization scheme (3D-FGMPAC), which is explainable as compared with deep learning methods., Methods: The proposed method first constructs a slice-indexed-histogram to localize the volume of interest (VOI) and estimate the probability that a voxel belongs to the liver according its intensity. The probabilistic image would be used to initialize the 3D PAC model. Secondly, a new contour indicator function, which is a component of the model, is produced by combining the gradient-based edge detection and Hessian-matrix-based surface detection. Then, a fast numerical scheme derived for the 3D PAC model is performed to evolve the initial probabilistic image into the global minimizer of the model, which is a smoothed probabilistic image showing a distinctly highlighted liver. Next, a simple region-growing strategy is applied to extract the whole liver mask from the smoothed probabilistic image. Finally, a B-spline surface is constructed to fit the patch of the rib cage to prevent possible leakage into adjacent intercostal tissues., Results: The proposed method is evaluated on two public datasets. The average Dice score, volume overlap error, volume difference, symmetric surface distance and volume processing time are 0.96, 7.35%, 0.02%, 1.17 mm and 19.8 s for the Sliver07 dataset, and 0.95, 8.89%, - 0.02 % $-0.02\%$ , 1.45 mm and 23.08 s for the 3Dircadb dataset, respectively., Conclusions: The proposed fully-automatic approach can effectively segment the liver from low-contrast and complex backgrounds. The quantitative and qualitative results demonstrate that the proposed segmentation method outperforms state-of-the-art traditional automatic liver segmentation algorithms and achieves very competitive performance compared with recent deep leaning-based methods., (© 2022 American Association of Physicists in Medicine.)

Purpose: Automatic detection and identification of setup devices, using a deep convolutional neural network (CNN) for real‐time multiclass object detection, has the potential to reduce errors in the treatment delivery process by avoiding documentation errors. Methods: A database of the setup device photos from the most recent 1200 patients treated at our institution was downloaded from the record and verify (R&V) system along with the corresponding setup notes. Images were manually labeled with bounding boxes of each device. A real‐time object detection CNN using the "you only look once" (YOLOv2) architecture was trained using transfer learning of a pretrained CNN (ResNet50). The CNN was trained to detect and identify 11 of the most common treatment accessories used at our institution. Results: Using transfer learning of a CNN for multiclass object detection, we are able to automatically detect and identify setup devices in photographs with an accuracy of 96%. Conclusions: Automation in radiation oncology has the potential to reduce risk. Automatic detection of setup devices is possible using a CNN and transfer learning. This work shows both the value of incident learning systems (ILS) in practice knowledge dissemination, and shows how automation of clinical processes and less reliance on manual documentation has the potential for risk reduction in radiation oncology treatments. [ABSTRACT FROM AUTHOR]

Background: EOSedge™* (EOS Imaging, Paris, France) is an X-ray imaging system using automatic exposure control (AEC) with tube current modulation, in order to optimize dose deposition in patients. Purpose: This study aims at characterizing EOSedge organ dose deposition in comparison to a digital radiography (DR) system and the previous EOS system (EOS-1st generation), in relation to their respective image quality levels. Method: Organ doses were measured in an anthropomorphic female adult phantom and a 5-year-old pediatric phantom using optically stimulated luminescence (OSL) dosimeters, which were carefully calibrated within the studied energy range. Organ doses were recorded on the EOSedge and the Fuji Visionary DRF (Fujifilm Medical Systems U.S.A., Inc, Lexington, MA). The resulting effective doses were compared to the EOS-1st-generation values present in the literature. Image quality assessment was carried out on end-user images. Quantitative image quality metrics were computed for all tested modalities on a quality assurance phantom. Qualitative assessment of EOSedge image quality was based on anthropomorphic phantom acquisitions against the EOS-1st-generation system, and on clinical images against the tested DR system. Results: For a full-spine exam, and on the female adult phantom (respectively, the pediatric phantom), an effective dose of 92 Sv (respectively, 32 Sv) was obtained on EOSedge, and 572 Sv (respectively, 179 Sv) on the DR system; these values were compared to effective dose values of 290 Sv (respectively, 200 Sv) from the literature on EOS-1st generation, leading to an effective dose reduction factor of 6 with respect to the DR system, and of 3-6 with respect to EOS-1st generation. EOSedge provides the best compromise between contrast-to-noise ratio (CNR) and dose, with more consistent CNR values than the other tested modalities, in a range of attenuation from 10 to 40 cm of poly(methyl methacrylate) (PMMA). Within this range, EOSedge is also comparable to DR for 10 and 20 cm of PMMA, and better than DR for 30 and 40 cm of PMMA, both in terms of spatial resolution and low-contrast detection. The anatomical landmarks of interest in the follow-up of spinal deformities can be detected in all tested modalities. Conclusion: Results showed that EOSedge provides significant dose reduction factors for full spine imaging in both adults and children compared to the other tested modalities, without compromising image quality. We believe that this work could help raise awareness on the capabilities of modern X-ray systems, when equipped with appropriate AEC strategies, to perform ultra-low-dose, long-axis images. [ABSTRACT FROM AUTHOR]

Purpose: Radiation damage on neurovascular bundles (NVBs) may be the cause of sexual dysfunction after radiotherapy for prostate cancer. However, it is challenging to delineate NVBs as organ-at-risks from planning CTs during radiotherapy. Recently, the integration of MR into radiotherapy made NVBs contour delineating possible. In this study, we aim to develop an MRI-based deep learning method for automatic NVB segmentation., Methods: The proposed method, named topological modulated network, consists of three subnetworks, that is, a focal modulation, a hierarchical block and a topological fully convolutional network (FCN). The focal modulation is used to derive the location and bounds of left and right NVBs', namely the candidate volume-of-interests (VOIs). The hierarchical block aims to highlight the NVB boundaries information on derived feature map. The topological FCN then segments the NVBs inside the VOIs by considering the topological consistency nature of the vascular delineating. Based on the location information of candidate VOIs, the segmentations of NVBs can then be brought back to the input MRI's coordinate system., Results: A five-fold cross-validation study was performed on 60 patient cases to evaluate the performance of the proposed method. The segmented results were compared with manual contours. The Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD 95 ) are (left NVB) 0.81 ± 0.10, 1.49 ± 0.88 mm, and (right NVB) 0.80 ± 0.15, 1.54 ± 1.22 mm, respectively., Conclusion: We proposed a novel deep learning-based segmentation method for NVBs on pelvic MR images. The good segmentation agreement of our method with the manually drawn ground truth contours supports the feasibility of the proposed method, which can be potentially used to spare NVBs during proton and photon radiotherapy and thereby improve the quality of life for prostate cancer patients., (© 2023 American Association of Physicists in Medicine.)