MRI Segmentation Research Articles

SDMT: Spatial Dependence Multi-Task Transformer Network for 3D Knee MRI Segmentation and Landmark Localization

Knee segmentation and landmark localization from 3D MRI are two significant tasks for diagnosis and treatment of knee diseases. With the development of deep learning, Convolutional Neural Network (CNN) based methods have become the mainstream. However, the existing CNN methods are mostly single-task methods. Due to the complex structure of bone, cartilage and ligament in the knee, it is challenging to complete the segmentation or landmark localization alone. And establishing independent models for all tasks will bring difficulties for surgeon's clinical using. In this paper, a Spatial Dependence Multi-task Transformer (SDMT) network is proposed for 3D knee MRI segmentation and landmark localization. We use a shared encoder for feature extraction, then SDMT utilizes the spatial dependence of segmentation results and landmark position to mutually promote the two tasks. Specifically, SDMT adds spatial encoding to the features, and a task hybrided multi-head attention mechanism is designed, in which the attention heads are divided into the inter-task attention head and the intra-task attention head. The two attention head deal with the spatial dependence between two tasks and correlation within the single task, respectively. Finally, we design a dynamic weight multi-task loss function to balance the training process of two task. The proposed method is validated on our 3D knee MRI multi-task datasets. Dice can reach 83.91% in the segmentation task, and MRE can reach 2.12 mm in the landmark localization task, it is competitive and superior over other state-of-the-art single-task methods.

Brain MRI segmentation of Zika-Exposed normocephalic infants shows smaller amygdala volumes.

Infants with congenital Zika syndrome (CZS) are known to exhibit characteristic brain abnormalities. However, the brain anatomy of Zika virus (ZIKV)-exposed infants, born to ZIKV-positive pregnant mothers, who have normal-appearing head characteristics at birth, has not been evaluated in detail. The aim of this prospective study is, therefore, to compare the cortical and subcortical brain structural volume measures of ZIKV-exposed normocephalic infants to age-matched healthy controls. We acquired T2-MRI of the whole brain of 18 ZIKV-exposed infants and 8 normal controls on a 3T MRI scanner. The MR images were auto-segmented into eight tissue types and anatomical regions including the white matter, cortical grey matter, deep nuclear grey matter, corticospinal fluid, amygdala, hippocampus, cerebellum, and brainstem. We determined the volumes of these regions and calculated the total intracranial volume (TICV) and head circumference (HC). We compared these measurements between the two groups, controlling for infant age at scan, by first comparing results for all subjects in each group and secondly performing a subgroup analysis for subjects below 8 weeks of postnatal age at scan. ZIKV-exposed infants demonstrated a significant decrease in amygdala volume compared to the control group in both the group and subgroup comparisons (p<0.05, corrected for multiple comparisons using FDR). No significant volume differences were observed in TICV, HC, or any specific brain tissue structures or regions. Study limitations include small sample size, which was due to abrupt cessation of extramural funding as the ZIKV epidemic waned. ZIKV-exposed infants exhibited smaller volumes in the amygdala, a brain region primarily involved in emotional and behavioral processing. This brain MRI finding may lead to poorer behavioral outcomes and warrants long-term monitoring of pediatric cases of infants with gestational exposure to Zika virus as well as other neurotropic viruses.

Automatic Localization of Key Structures for Subthalamic Nucleus–Deep Brain Stimulation Surgery via Prior-Enhanced Multi-Object Magnetic Resonance Imaging Segmentation

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is an established and effective neurosurgical treatment for relieving motor symptoms in Parkinson disease. The localization of key brain structures is critical to the success of DBS surgery. However, in clinical practice, this process is heavily dependent on the radiologist's experience. In this study, we propose an automatic localization method of key structures for STN-DBS surgery via prior-enhanced multi-object magnetic resonance imaging segmentation. We use the U-Net architecture for the multi-object segmentation, including STN, red nucleus, brain sulci, gyri, and ventricles. To address the challenge that only half of the brain sulci and gyri locate in the upper area, potentially causing interference in the lower area, we perform region of interest detection and ensemble joint processing to enhance the segmentation performance of brain sulci and gyri. We evaluate the segmentation accuracy by comparing our method with other state-of-the-art machine learning segmentation methods. The experimental results show that our approach outperforms state-of-the-art methods in terms of segmentation performance. Moreover, our method provides effective visualization of key brain structures from a clinical application perspective and can reduce the segmentation time compared with manual delineation. Our proposed method uses deep learning to achieve accurate segmentation of the key structures more quickly than and with comparable accuracy to human manual segmentation. Our method has the potential to improve the efficiency of surgical planning for STN-DBS.

A Novel Light U-Net Model for Left Ventricle Segmentation Using MRI

MRI segmentation and analysis are significant tasks in clinical cardiac computations. A cardiovascular MR scan with left ventricular segmentation seems necessary to diagnose and further treat the disease. The proposed method for left ventricle segmentation works as a combination of the intelligent histogram-based image enhancement technique with a Light U-Net model. This technique serves as the basis for choosing the low-contrast image subjected to the stretching technique and produces sharp object contours with good contrast settings for the segmentation process. After enhancement, the images are subjected to the encoder–decoder configuration of U-Net using a novel lightweight processing model. Encoder sampling is supported by a block of three parallel convolutional layers with supporting functions that improve the semantics for segmentation at various levels of resolutions and features. The proposed method finally increased segmentation efficiency, extracting the most relevant image resources from depth-to-depth convolutions, filtering them through each network block, and producing more precise resource maps. The dataset of MICCAI 2009 served as an assessment tool of the proposed methodology and provides a dice coefficient value of 97.7%, accuracy of 92%, and precision of 98.17%.

High-fidelity Database-free Deep Learning Reconstruction for Real-time Cine Cardiac MRI.

Real-time cine cardiac MRI provides an ECG-free free-breathing alternative to clinical gold-standard ECG-gated breath-hold segmented cine MRI for evaluation of heart function. Real-time cine MRI data acquisition during free breathing snapshot imaging enables imaging of patient cohorts that cannot be imaged with segmented or breath-hold acquisitions, but requires rapid imaging to achieve sufficient spatial-temporal resolutions. However, at high acceleration rates, conventional reconstruction techniques suffer from residual aliasing and temporal blurring, including advanced methods such as compressed sensing with radial trajectories. Recently, deep learning (DL) reconstruction has emerged as a powerful tool in MRI. However, its utility for free-breathing real-time cine MRI has been limited, as database-learning of spatio-temporal correlations with varying breathing and cardiac motion patterns across subjects has been challenging. Zero-shot self-supervised physics-guided deep learning (PG-DL) reconstruction has been proposed to overcome such challenges of database training by enabling subject-specific training. In this work, we adapt zero-shot PG-DL for real-time cine MRI with a spatio-temporal regularization. We compare our method to TGRAPPA, locally low-rank (LLR) regularized reconstruction and database-trained PG-DL reconstruction, both for retrospectively and prospectively accelerated datasets. Results on highly accelerated real-time Cartesian cine MRI show that the proposed method outperforms other reconstruction methods, both visibly in terms of noise and aliasing, and quantitatively.

Efficient Algorithm for Distinction Mild Cognitive Impairment from Alzheimer’s Disease Based on Specific View FCM White Matter Segmentation and Ensemble Learning

Purpose: Alzheimer's Disease (AD) is in the dementia group and is one of the most prevalent neurodegenerative disorders. Approximately 50 million people were affected in 2018, and that number is expected to triple by 2050. Several demographic properties, neuroimaging such as MRI, functional MRI (MRI), neuropsychiatric symptoms, and cognitive abilities are used to predict AD. Between existing characteristics, White Matter (WM) is a known marker for AD tracking, and WM segmentation in MRI based on clustering can be used to decrease the volume of data. Many algorithms have been developed to predict AD, but most concentrate on the distinction of AD from Cognitive Normal (CN), and fewer on the distinction of AD from Mild Cognitive Impairment (MCI), which has an important position in AD progression. In addition, there are not efficient algorithms with low computational costs and sufficient features in clinical use.&#x0D; In this study, we provided a new, simple, and efficient methodology for classifying patients into AD and MCI patients and evaluated the effect of the view dimension of Fuzzy C means (FCM) in prediction with ensemble classifiers. This work was based on the segmentation of WM and extracting two groups of features.&#x0D; Materials and Methods: We proposed our methodology in three steps; first, segmentation of WM from T1 MRI with FCM according to two specific viewpoints (3D and 2D). In the second step, two groups of features are extracted: approximate coefficients of Discrete Wavelet Transform (DWT) with three-level decomposition and statistical (mean, variance, skewness) features. In the final step, an ensemble classifier that is constructed with three simple classifiers, K-Nearest Neighbor (KNN), Decision Tree (DT), and Linear Discriminant Analysis (LDA), was used to distinguish MCI from AD.&#x0D; Results: The proposed method has been evaluated by using 1,280 slices (samples) from 64 patients with MCI (32) and AD (32) of the ADNI dataset. The best performance is for the 3D viewpoint, and the accuracy, precision, and f1-score achieved from the methodology are 94.22%, 94.45%, and 94.21%, respectively, by using a ten-fold Cross-Validation (CV) strategy.&#x0D; Conclusion: The experimental evaluation shows that WM segmentation increases the performance of the ensemble classifier; moreover, the 3D view FCM is better than the 2D view. According to the results, the proposed methodology has comparable performance for the detection of MCI from AD. The low computational cost algorithm and the three classifiers for generalization can be used in practical application by physicians in pre-clinical.

Combining neural networks and image synthesis to enable automatic thoracic cavity segmentation of hyperpolarized 129Xe MRI without proton scans

Rationale and objectivesQuantification of 129Xe MRI relies on accurate segmentation of the thoracic cavity, typically performed manually using a combination of 1H and 129Xe scans. This can be accelerated by using Convolutional Neural Networks (CNNs) that segment only the 129Xe scan. However, this task is complicated by peripheral ventilation defects, which requires training CNNs with large, diverse datasets. Here, we accelerate the creation of training data by synthesizing 129Xe images with a variety of defects. We use this to train a 3D model to provide thoracic cavity segmentation from 129Xe ventilation MRI alone. Materials and methodsTraining and testing data consisted of 22 and 33 3D 129Xe ventilation images. Training data were expanded to 484 using Template-based augmentation while an additional 298 images were synthesized using the Pix2Pix model. This data was used to train both a 2D U-net and 3D V-net-based segmentation model using a combination of Dice-Focal and Anatomical Constraint loss functions. Segmentation performance was compared using Dice coefficients calculated over the entire lung and within ventilation defects. ResultsPerformance of both U-net and 3D segmentation was improved by including synthetic training data. The 3D models performed significantly better than U-net, and the 3D model trained with synthetic 129Xe images exhibited the highest overall Dice score of 0.929. Moreover, addition of synthetic training data improved the Dice score in ventilation defect regions from 0.545 to 0.588 for U-net and 0.739 to 0.765 for the 3D model. ConclusionIt is feasible to obtain high-quality segmentations from 129Xe scan alone using 3D models trained with additional synthetic images.

Prediction of Age Older than 18 Years in Sub-adults by MRI Segmentation of 1st and 2nd Molars

PurposeTo investigate prediction of age older than 18 years in sub-adults using tooth tissue volumes from MRI segmentation of the entire 1st and 2nd molars, and to establish a model for combining information from two different molars.Materials and methodsWe acquired T2 weighted MRIs of 99 volunteers with a 1.5-T scanner. Segmentation was performed using SliceOmatic (Tomovision©). Linear regression was used to analyse the association between mathematical transformation outcomes of tissue volumes, age, and sex. Performance of different outcomes and tooth combinations were assessed based on the p-value of the age variable, common, or separate for each sex, depending on the selected model. The predictive probability of being older than 18 years was obtained by a Bayesian approach using information from the 1st and 2nd molars both separately and combined.Results1st molars from 87 participants, and 2nd molars from 93 participants were included. The age range was 14-24 years with a median age of 18 years. The transformation outcome (high signal soft tissue + low signal soft tissue)/total had the strongest statistical association with age for the lower right 1st (p= 7.1*10-4 for males) and 2nd molar (p=9.44×10-7 for males and p=7.4×10-10 for females). Combining the lower right 1st and 2nd molar in males did not increase the prediction performance compared to using the best tooth alone.ConclusionMRI segmentation of the lower right 1st and 2nd molar might prove useful in the prediction of age older than 18 years in sub-adults. We provided a statistical framework to combine the information from two molars.

Complementation-reinforced network for integrated reconstruction and segmentation of pulmonary gas MRI with high acceleration.

Hyperpolarized (HP) gas MRI enables the clear visualization of lung structure and function. Clinically relevant biomarkers, such as ventilated defect percentage (VDP) derived from this modality can quantify lung ventilation function. However, long imaging time leads to image quality degradation and causes discomfort to the patients. Although accelerating MRI by undersampling k-space data is available, accurate reconstruction and segmentation of lung images are quite challenging at high acceleration factors. To simultaneously improve the performance of reconstruction and segmentation of pulmonary gas MRI at high acceleration factors by effectively utilizing the complementary information in different tasks. A complementation-reinforced network is proposed, which takes the undersampled images as input and outputs both the reconstructed images and the segmentation results of lung ventilation defects. The proposed network comprises a reconstruction branch and a segmentation branch. To effectively exploit the complementary information, several strategies are designed in the proposed network. Firstly, both branches adopt the encoder-decoder architecture, and their encoders are designed to share convolutional weights for facilitating knowledge transfer. Secondly, a designed feature-selecting block discriminately feeds shared features into decoders of both branches, which can adaptively pick suitable features for each task. Thirdly, the segmentation branch incorporates the lung mask obtained from the reconstructed images to enhance the accuracy of the segmentation results. Lastly, the proposed network is optimized by a tailored loss function that efficiently combines and balances these two tasks, in order to achieve mutual benefits. Experimental results on the pulmonary HP 129 Xe MRI dataset (including 43 healthy subjects and 42 patients) show that the proposed network outperforms state-of-the-art methods at high acceleration factors (4, 5, and 6). The peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and Dice score of the proposed network are enhanced to 30.89, 0.875, and 0.892, respectively. Additionally, the VDP obtained from the proposed network has good correlations with that obtained from fully sampled images (r=0.984). At the highest acceleration factor of 6, the proposed network promotes PSNR, SSIM, and Dice score by 7.79%, 5.39%, and 9.52%, respectively, in comparison to the single-task models. The proposed method effectively enhances the reconstruction and segmentation performance at high acceleration factors up to 6. It facilitates fast and high-quality lung imaging and segmentation, and provides valuable support in the clinical diagnosis of lung diseases.

Deep Learning Segmentation of the Right Ventricle in Cardiac MRI: The M&Ms Challenge.

In recent years, several deep learning models have been proposed to accurately quantify and diagnose cardiac pathologies. These automated tools heavily rely on the accurate segmentation of cardiac structures in MRI images. However, segmentation of the right ventricle is challenging due to its highly complex shape and ill-defined borders. Hence, there is a need for new methods to handle such structure's geometrical and textural complexities, notably in the presence of pathologies such as Dilated Right Ventricle, Tricuspid Regurgitation, Arrhythmogenesis, Tetralogy of Fallot, and Inter-atrial Communication. The last MICCAI challenge on right ventricle segmentation was held in 2012 and included only 48 cases from a single clinical center. As part of the 12th Workshop on Statistical Atlases and Computational Models of the Heart (STACOM 2021), the M&Ms-2 challenge was organized to promote the interest of the research community around right ventricle segmentation in multi-disease, multi-view, and multi-center cardiac MRI. Three hundred sixty CMR cases, including short-axis and long-axis 4-chamber views, were collected from three Spanish hospitals using nine different scanners from three different vendors, and included a diverse set of right and left ventricle pathologies. The solutions provided by the participants show that nnU-Net achieved the best results overall. However, multi-view approaches were able to capture additional information, highlighting the need to integrate multiple cardiac diseases, views, scanners, and acquisition protocols to produce reliable automatic cardiac segmentation algorithms.

A New Framework of Swarm Learning Consolidating Knowledge From Multi-Center Non-IID Data for Medical Image Segmentation

Large training datasets are important for deep learning-based methods. For medical image segmentation, it could be however difficult to obtain large number of labeled training images solely from one center. Distributed learning, such as swarm learning, has the potential to use multi-center data without breaching data privacy. However, data distributions across centers can vary a lot due to the diverse imaging protocols and vendors (known as feature skew). Also, the regions of interest to be segmented could be different, leading to inhomogeneous label distributions (referred to as label skew). With such non-independently and identically distributed (Non-IID) data, the distributed learning could result in degraded models. In this work, we propose a novel swarm learning approach, which assembles local knowledge from each center while at the same time overcomes forgetting of global knowledge during local training. Specifically, the approach first leverages a label skew-awared loss to preserve the global label knowledge, and then aligns local feature distributions to consolidate global knowledge against local feature skew. We validated our method in three Non-IID scenarios using four public datasets, including the Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Segmentation (M&Ms) dataset, the Federated Tumor Segmentation (FeTS) dataset, the Multi-Modality Whole Heart Segmentation (MMWHS) dataset and the Multi-Site Prostate T2-weighted MRI segmentation (MSProsMRI) dataset. Results show that our method could achieve superior performance over existing methods. Code will be released via https://zmiclab.github.io/projects.html once the paper gets accepted.

Brain Tumor Diagnosis Using CNN and Image Processing

Abstract: Among the most tedious and time-consuming tasks in medical image processing is brain tumor identification and segmentation. Magnetic resonance imaging, or MRI, is a medical procedure that radiologists primarily utilize to examine the human body without having to do surgery. The substantial information that MRI provides on human soft tissue aids in the diagnosis of brain tumors. It is crucial to accurately segment MRI images in order to detect a brain tumour utilizing computeraided clinical techniques. Brain tumor identification facilitates determining the precise size and location of the tumor. Initially, the scanned image's quality is upgraded, and tumors are located in the image using morphological operators

Attention-guided residual W-Net for supervised cardiac magnetic resonance imaging segmentation

ObjectiveWith latest developments in deep learning approaches, automated, accurate, fast, and generalized segmentation model for left atrium, left ventricle, right ventricle, and myocardium from MR images is becoming increasingly desirable. In deep learning-based approaches, model generalizability is an essential concern. The strength of an approach that has proven competent for one dataset, and then executed for other without fine-tuning, started to decline. MethodsIn medical image segmentation, U-Net-based architecture with its skip-connection has attained the highest levels of success. This article, therefore presents a novel Attention-guided Residual W-Net (ARW-Net), a deep learning-based segmentation approach with residual links, attention gates at each feature dimensions, and upgraded deep supervision for cardiac MRI (CMRI) segmentation. As the basic building block, the W-Net, a unique methodology for medical image segmentation (centred on U-Net layout), is used. In a W-Net-based structure, ARW-Net suggests residual connections in encoder and attention gates at each feature dimension in decoder. Combination of Dice and Cross-entropy losses is used for model training. ResultsFour CMRI datasets, (i) ACDC 2017, (ii) 2018 ASC, (iii) M&Ms 2020, and (iv) LAScarQS 2022, which are publicly available, have been utilized to examine effectiveness of ARW-Net. It accomplished improved segmentation results and was among the top two for several metrics. ConclusionNumerous comparisons have demonstrated that ARW-Net is profoundly promising, with several segmentation findings delivering innovative state-of-the-art outcomes on four datasets. SignificanceThe results of experiments demonstrate that ARW-Net is able to effectively segment cardiac MR images, as demonstrated by comparisons with existing algorithms.

Placental MRI segmentation based on multi-receptive field and mixed attention separation mechanism

ObjectiveTo reduce the occurrence of massive bleeding during placental abruption in patients with placenta accrete, we established a medical imaging based on multi-receptive field and mixed attention separation mechanism (MRF-MAS) model to improve the accuracy of MRI placenta segmentation and provide a basis for subsequent placenta accreta. MethodsWe propose a placenta MRI segmentation technology using the MRF-MAS framework to develop a medical image diagnostic technique. The model first uses the multi-receptive field feature structure to obtain multi-level information, and improves the expression of features at differing scales. Note that the hybrid attention mechanism combines channel attention and spatial attention, separates the input feature sets and computes the attention separately, and finally reorganizes the feature maps. To show that the model can improve the accuracy of segmenting the placenta, we adopt mean Intersection over Union (IoU), Dice similarity coefficient (Dice) and area under the receiver operating characteristic curve (AUC) with U-Net, Mask RCNN, Deeplab v3 for comparison. ResultsThe four models achieved different outcomes based on our placenta dataset, with our model IoU and Dice up to 0.8169 and 0.8992, which are 5.51% and 3.03% higher than the average of the three comparison models. ConclusionThe model proposed by us is helpful to assist the imaging diagnosis and at the same time provides a quantitative reference for the precise treatment of placenta accreta, assists the Equationtion of the clinical operation plan of the physician, and promotes the precision medicine of placenta accreta.

A novel approach for brain tissue segmentation and classification in infants' MRI images based on seeded region growing, foster corner detection theory, and sparse autoencoder

Brain tissue segmentation and classification in infant MRI images play a crucial role in early diagnosis of neurological disorders. In this paper, we propose a novel approach based on Seeded Region Growing, Foster Corner Detection Theory, and Sparse Autoencoder. Proposed method incorporates advantages of these techniques to improve segmentation and classification accuracy of infant brain MRI. An unsupervised segmentation method proposed in this study has four stages. First stage involves pre-processing input image by removing noise using a non-local means method and rectifying intensity inhomogeneity field employing Contrast-limited adaptive histogram equalization. In second stage, modified and optimized approach of region growth is utilized, where Forstner Corner Identification principle is used to choose and position seeds to a producer. Third stage involves applying the region growth technique to image, where similarity criterion is optimized using Predator Prey algorithm. In fourth stage, tissues are classified using weighted KNN and Sparse Auto Encoder classifiers. Result shows high Dice factor of 93%, outperforming existing works. This method improves accuracy of brain tissue segmentation and classification, leading to better clinical decision-making. Applications of proposed method include early detection and diagnosis of neurological disorders, monitoring of disease progression, and assessment of treatment effectiveness.

Segmentation of 4D Flow MRI: Comparison between 3D Deep Learning and Velocity-Based Level Sets.

A thoracic aortic aneurysm is an abnormal dilatation of the aorta that can progress and lead to rupture. The decision to conduct surgery is made by considering the maximum diameter, but it is now well known that this metric alone is not completely reliable. The advent of 4D flow magnetic resonance imaging has allowed for the calculation of new biomarkers for the study of aortic diseases, such as wall shear stress. However, the calculation of these biomarkers requires the precise segmentation of the aorta during all phases of the cardiac cycle. The objective of this work was to compare two different methods for automatically segmenting the thoracic aorta in the systolic phase using 4D flow MRI. The first method is based on a level set framework and uses the velocity field in addition to 3D phase contrast magnetic resonance imaging. The second method is a U-Net-like approach that is only applied to magnitude images from 4D flow MRI. The used dataset was composed of 36 exams from different patients, with ground truth data for the systolic phase of the cardiac cycle. The comparison was performed based on selected metrics, such as the Dice similarity coefficient (DSC) and Hausdorf distance (HD), for the whole aorta and also three aortic regions. Wall shear stress was also assessed and the maximum wall shear stress values were used for comparison. The U-Net-based approach provided statistically better results for the 3D segmentation of the aorta, with a DSC of 0.92 ± 0.02 vs. 0.86 ± 0.5 and an HD of 21.49 ± 24.8 mm vs. 35.79 ± 31.33 mm for the whole aorta. The absolute difference between the wall shear stress and ground truth slightly favored the level set method, but not significantly (0.754 ± 1.07 Pa vs. 0.737 ± 0.79 Pa). The results showed that the deep learning-based method should be considered for the segmentation of all time steps in order to evaluate biomarkers based on 4D flow MRI.

Fully Automatic initialization and segmentation of left and right ventricles for large-scale cardiac MRI using a deeply supervised network and 3D-ASM

Background and objectiveThe sheer volume of data generated by population imaging studies is unparalleled by current capabilities to extract objective and quantitative cardiac phenotypes; subjective and time-consuming manual image analysis remains the gold standard. Automated image analytics to compute quantitative imaging biomarkers of cardiac function are desperately needed. Data volumes and their variability pose a challenge to most state-of-the-art methods for endo- and epicardial contours, which lack robustness when applied to very large datasets. Our aim is to develop an analysis pipeline for the automatic quantification of cardiac function from cine magnetic resonance imaging data. MethodThis work adopt 4,638 cardiac MRI cases coming from UK Biobank with ground truth available for left and RV contours. A hybrid and robust algorithm is proposed to improve the accuracy of automatic left and right ventricle segmentation by harnessing the localization accuracy of deep learning and the morphological accuracy of 3D-ASM (three-dimensional active shape models). The contributions of this paper are three-fold. First, a fully automatic method is proposed for left and right ventricle initialization and cardiac MRI segmentation by taking full advantage of spatiotemporal constraint. Second, a deeply supervised network is introduced to train and segment the heart. Third, the 3D-ASM image search procedure is improved by combining image intensity models with convolutional neural network (CNN) derived distance maps improving endo- and epicardial edge localization. ResultsThe proposed architecture outperformed the state of the art for cardiac MRI segmentation from UK Biobank. The statistics of RV landmarks detection errors for Triscuspid valve and RV apex are 4.17 mm and 5.58 mm separately. The overlap metric, mean contour distance, Hausdorff distance and cardiac functional parameters are calculated for the LV (Left Ventricle) and RV (Right Ventricle) contour segmentation. Bland–Altman analysis for clinical parameters shows that the results from our automated image analysis pipelines are in good agreement with results from expert manual analysis. ConclusionsOur hybrid scheme combines deep learning and statistical shape modeling for automatic segmentation of the LV/RV from cardiac MRI datasets is effective and robust and can compute cardiac functional indexes from population imaging.

Evaluating a Phase-Specific Approach to Aortic Flow: A 4D Flow MRI Study.

Aortic flow parameters can be quantified using 4D flow MRI. However, data are sparse on how different methods of analysis influence these parameters and how these parameters evolve during systole. To assess multiphase segmentations and multiphase quantification of flow-related parameters in aortic 4D flow MRI. Prospective. 40 healthy volunteers (50% male, 28.9 ± 5.0 years) and 10 patients with thoracic aortic aneurysm (80% male, 54 ± 8 years). 4D flow MRI with a velocity encoded turbo field echo sequence at 3 T. Phase-specific segmentations were obtained for the aortic root and the ascending aorta. The whole aorta was segmented in peak systole. In all aortic segments, time to peak (TTP; for flow velocity, vorticity, helicity, kinetic energy, and viscous energy loss) and peak and time-averaged values (for velocity and vorticity) were calculated. Static vs. phase-specific models were assessed using Bland-Altman plots. Other analyses were performed using phase-specific segmentations for aortic root and ascending aorta. TTP for all parameters was compared to TTP of flow rate using paired t-tests. Time-averaged and peak values were assessed using Pearson correlation coefficient. P < 0.05 was considered statistically significant. In the combined group, velocity in static vs. phase-specific segmentations differed by 0.8 cm/sec for the aortic root, and 0.1 cm/sec (P = 0.214) for the ascending aorta. Vorticity differed by 167 sec-1 mL-1 (P = 0.468) for the aortic root, and by 59 sec-1 mL-1 (P = 0.481) for the ascending aorta. Vorticity, helicity, and energy loss in the ascending aorta, aortic arch, and descending aorta peaked significantly later than flow rate. Time-averaged velocity and vorticity values correlated significantly in all segments. Static 4D flow MRI segmentation yields comparable results as multiphase segmentation for flow-related parameters, eliminating the need for time-consuming multiple segmentations. However, multiphase quantification is necessary for assessing peak values of aortic flow-related parameters. 2 TECHNICAL EFFICACY STAGE: 3.

Automated Neural Network-based Survival Prediction of Glioblastoma Patients Using Pre-operative MRI and Clinical Data

In this paper, we proposed a lightweight two-dimensional (2D) methodology to predict the survival time of Glioblastoma Multiforme patients. Firstly, we trained the 2D ResUNet-SEG (Residual UNet for Segmentation) model to perform semantic segmentation on brain tumour subregions. Then, we used the raw and segmented MRI volumes along with clinical data to train the 2D CNN-SP (Convolutional Neural Network for Survival Prediction) model to predict the survival time in days. The experiments showed that our proposed methodology achieved an accuracy of 0.517, Mean Square Error of 136,783.42, MSE, Median Square Error (medianSE) of 106,608.6, Standard Deviation Error (stdSE) of 139,210.8, and SpearmanR correlation score of 0.299 on the Multimodal Brain Tumour Segmentation (BraTS) 2020 validation set. The obtained results are competitive compared to the state-of-the-art automated techniques for survival prognosis of GBM patients validated on the same set of patients. Results proved that Deep Learning (DL) based feature learning is better than existing Machine Learning based techniques with handcrafted radiomics based feature extraction. It eliminates the need for feature selection as well. However, the results achieved are limited due to the unavailability of vast clinical data required to train Convolutional Neural Network (CNN) based deep architectures.

Automatic AI-based contouring of prostate MRI for online adaptive radiotherapy

Background and purposeMR-guided radiotherapy (MRgRT) online plan adaptation accounts for tumor volume changes, interfraction motion and thus allows daily sparing of relevant organs at risk. Due to the high interfraction variability of bladder and rectum, patients with tumors in the pelvic region may strongly benefit from adaptive MRgRT. Currently, fast automatic annotation of anatomical structures is not available within the online MRgRT workflow. Therefore, the aim of this study was to train and validate a fast, accurate deep learning model for automatic MRI segmentation at the MR-Linac for future implementation in a clinical MRgRT workflow. Materials and methodsFor a total of 47 patients, T2w MRI data were acquired on a 1.5 T MR-Linac (Unity, Elekta) on five different days. Prostate, seminal vesicles, rectum, anal canal, bladder, penile bulb, body and bony structures were manually annotated. These training data consisting of 232 data sets in total was used for the generation of a deep learning based autocontouring model and validated on 20 unseen T2w-MRIs. For quantitative evaluation the validation set was contoured by a radiation oncologist as gold standard contours (GSC) and compared in MATLAB to the automatic contours (AIC). For the evaluation, dice similarity coefficients (DSC), and 95% Hausdorff distances (95% HD), added path length (APL) and surface DSC (sDSC) were calculated in a caudal-cranial window of ± 4 cm with respect to the prostate ends. For qualitative evaluation, five radiation oncologists scored the AIC on the possible usage within an online adaptive workflow as follows: (1) no modifications needed, (2) minor adjustments needed, (3) major adjustments/ multiple minor adjustments needed, (4) not usable. ResultsThe quantitative evaluation revealed a maximum median 95% HD of 6.9 mm for the rectum and minimum median 95% HD of 2.7 mm for the bladder. Maximal and minimal median DSC were detected for bladder with 0.97 and for penile bulb with 0.73, respectively. Using a tolerance level of 3 mm, the highest and lowest sDSC were determined for rectum (0.94) and anal canal (0.68), respectively.Qualitative evaluation resulted in a mean score of 1.2 for AICs over all organs and patients across all expert ratings. For the different autocontoured structures, the highest mean score of 1.0 was observed for anal canal, sacrum, femur left and right, and pelvis left, whereas for prostate the lowest mean score of 2.0 was detected. In total, 80% of the contours were rated be clinically acceptable, 16% to require minor and 4% major adjustments for online adaptive MRgRT. ConclusionIn this study, an AI-based autocontouring was successfully trained for online adaptive MR-guided radiotherapy on the 1.5 T MR-Linac system. The developed model can automatically generate contours accepted by physicians (80%) or only with the need of minor corrections (16%) for the irradiation of primary prostate on the clinically employed sequences.