Keywords: Segmentation, Segmentation, Bone, Knee MRI, Pipeline, AI

Motivation: [18F]NaF PET imaging is a promising technique to study the role of bone metabolism in joint diseases such as osteoarthritis. To ease the corresponding burden of segmentations and data analysis, we developed an automated pipeline.

Goal(s): We developed a new automated pipeline for knee bone segmentations. We validated the creation of subchondral bone masks for applications to [18F]NaF PET imaging by measuring changes in PET measures.

Approach: We developed an automated segmentation pipeline for bone segmentations and validated the results for PET imaging. 

Results: DOSMA automated segmentations perform highly for bones and show applicability for quantitative musculoskeletal analysis.

Impact: Our automated bone segmentation and PET data analysis pipeline enables a streamlined way of automating PET-MRI processing, including registration, segmentation, quantitative mapping, and visualization of outcome measures.

Keywords: Segmentation, Heart

Motivation: Right ventricular(RV) segmentation is of great significance for the clinical diagnosis of heart diseases. However, due to the complex structure, RV segmentation is still challenging.

Goal(s): Fully automatic and accurate segmentation of the right ventricle.

Approach: A  new deep atlas network that combines atlas prior knowledge with Deformable Multi-scale Two-Stage U-net(DMTSU-net) is proposed to extract and fuse multi-scale RV features in Cine Cardiac Magnetic Resonance (CCMR) images.

Results: Compare with 8 classical methods, the segmentation results of DMTSU-net are mostly close to the gold standard and significantly correlate with it on all evaluation  indices in 15 testing datasets.

Impact: The proposed framework integrates prior information of atlases into a deep neural network to achieve accurate segmentation, which is promising for clinical heart disease diagnosis.

Keywords: Segmentation, Brain, High-Field MRI, Data Analsyis, Anaysis/Processing, Segementation, Multi-Contrast, Neuro-imaging, synthetic MPRAGE

Motivation: Brain segmentation and multiparameter mapping (MPM) are important for neurodegenerative disease characterization. Acquiring sub-millimeter images increases scan time and patient discomfort. At 7T, B1+ inhomogeneities challenge brain segmentation.

Goal(s): The quality of brain segmentations produced from FastSurferVINN was evaluated and compared between a 7T MPRAGE protocol and two synthetic MPRAGE approaches.

Approach: MPRAGE and MPM images were acquired on 16 subjects across three 7T sites using pTx pulses. MPRAGElike and synMPRAGE images were generated from MPM. All images were segmented with FastSurferVINN.

Results: FastSurferVINN seems to be a robust technique to segment sub-millimeter 7T images. MPRAGElike generated superior segmentations compared to synMPRAGE. 

Impact: Neuroscientists with Multi-Parameter Mapping sequences in their imaging protocol can approximate an MPRAGElike image (preferably over synthetic synMPRAGE). Acquiring an MPRAGE sequence solely for brain segmentation can be avoided resulting in a considerable amount of scan time saved.

Keywords: Segmentation, Brain, Magnetic Resonance Fingerprinting, Contrast-Invariant, Age-Agnostic

Motivation: Intensity-based brain segmentation methods face challenges with generalizability, as they are susceptible to site, age, and contrast variations.

Goal(s): Develop an efficient, unified framework to train segmentation networks that are insensitive to contrast variations.

Approach: MRF sequences encode image time series that include both common and uncommon contrasts. We develop MRF-Synth, a framework to generate contrasts from MRF quantitative maps for training and evaluating contrast-invariant networks.

Results: We show that a segmentation U-Net trained with MRF-Synth yields highly consistent results across contrasts, vendors, and ages (DICE > 0.86 in adults).

Impact: MRF-Synth represents an efficient, generalizable framework for developing and evaluating contrast-invariant segmentation networks. We demonstrate the utility of MRF-Synth in training a U-Net to segment healthy MR brain images into 18 anatomical regions regardless of contrast, scanner, vendor, or age.

Keywords: Segmentation, Brain

Motivation: The multi-modal images can provide complementary information to improve automatic MRI segmentation performance. However, most multi-modal methods require long training times due to the complex network structures and the large amounts of multi-modal images involved. Furthermore, obtaining numerous labeled data is time-consuming and laborious.

Goal(s): To achieve efficient and high-quality segmentation using only a few labeled data.

Approach: We propose an efficient training-free multi-modal fusion strategy based on superpixel method for semi-supervised brain tumor segmentation.

Results: The experiments on BraTS18 dataset show that our method results in superior overall performance, and can greatly reduce the time costs of doctors.

Impact: The strategy of using superpixel method to accelerate the network training process can assist in the timely diagnosis and treatment of diseases in the clinic, and provides a new idea to simplify multi-modal information fusion.

Keywords: MR-Guided Interventions, Diffusion/other diffusion imaging techniques

Motivation: Lack of direct visualization methods for thalamic subnuclei has resulted in variable patient outcomes and repeat surgeries in deep brain stimulation (DBS).

Goal(s): To generate images with sufficient intra-thalamic contrast to visualize subnuclei for surgical intervention.

Approach: We introduce DiMANI, an image obtained from combining diffusion-weighted volumes. We compared DiMANI to atlases as well as intra- and post-operative clinical DBS data.

Results: DiMANI showed strong correspondence to anatomical organization from atlases, was highly reproducible, and was observable at both 3T and 7T. Clinical data from six DBS patients corroborated DiMANI’s ability to identify the motor and sensory thalamus locations.

Impact: Visualization of thalamic subnuclei is now achievable using DiMANI, enabling direct targeting for DBS and MR-guided focused ultrasound procedures. This will provide immediate impact by enhancing clinical workflow efficiency, improving patient outcomes and advancing brain networks conception.

Keywords: Segmentation, Segmentation, Deep Learning, Digestive

Motivation: In abdominal MRI segmentation tasks, the need for high-quality support information for Segment Anything Model (SAM)-driven segmentation in limited data scenarios has motivated the search for an architecture with high performance and minimal support information requirements.

Goal(s): Our objective is to design a user-friendly architecture for segmentation, focusing on using only support information within the region of interest. We aim to verify its high-performance capabilities.

Approach: We developed Point-Guided 3D U-SAM, combining SAM and 3D U-Net with point-based support input. We compared its segmentation performance with existing methods.

Results: The model excelled in abdominal MRI segmentation across various contrast levels, ensuring high performance.

Impact: Point-Guided 3D U-SAM, which combines Segment Anything Model (SAM) and 3D U-Net with point-based inputs, would advance semi-automated organ segmentation, particularly where contrast is poor, such as MRCP, and in abdominal imaging, significantly reducing manual effort in clinical segmentation.

Keywords: Segmentation, Segmentation

Motivation:  Cine cardiac MRI provides a way to quantify additional cardiac indices beyond ejection fraction,  including ejection and filling rates, myocardial wall motion, and strain; segmentation of all temporal phases is required.

Goal(s): To develop an approach to segmenting images in all cardiac phases in cine MRI.

Approach: A U-net and a recurrent-neural-network were integrated to exploit the spatial-temporal information in cine time-series. 100 and 50 subjects labeled at the end-systole and end-diastole phases were used for network training and testing, respectively.

Results: The use of spatial-temporal information substantially improved the segmentation accuracy and the algorithm cardiac indices were strongly correlated with manual measurements.

Impact: The proposed method made effective use of the spatial-temporal information in a cine time-series and yielded highly accurate and precise segmentation and cardiac functional measurements, suggesting the utility of our approach for clinical cardiac patient care.

Keywords: Visualization, Visualization, Fairness; Bias

Motivation: A systematic analysis of the segmentation effectiveness for fairness helps enhance the effectiveness of artificial intelligence(AI) model, which has not been done before.

Goal(s): This study aims to compile statistics the relation between the segmentation effectiveness and aging, gender as well as anatomical regions.

Approach: The nnU-Net model is used for organ segmentation while the DICE was computed to evaluate the relation between the effectiveness with aging and gender and the heatmap was used to visualize the spatial error distribution regarding anatomical regions.

Results: The result demonstrates variations in nnU-Net's effectiveness within subgroups, highlighting the significance of attention mechanisms for segmentation model enhancement.

Impact: This study comprehensively evaluated the fairness and effectiveness of nnU-Net across multiple organs within the body. An analysis was conducted to investigate the relationship between segmentation errors and age, gender as well as anatomical regions for organ segmentation.

Keywords: Segmentation, Machine Learning/Artificial Intelligence

Motivation: Accurate segmentation of the hippocampus provides an important biomarker in neurodegenerative diseases, e.g., Alzheimer’s disease. However, currently available tools are not robust to disease-related atrophy.

Goal(s): We aim to demonstrate the accuracy of our InnerEye hippocampal segmentation tool on clinical data.

Approach: We fine-tuned our existing model on manually segmented data and externally validated the model on a clinical dataset of patients referred to a dementia clinic. We compare our model to three commonly used segmentation tools.

Results: Our model provides significant improvements over currently available tools when tested on an external, clinical dataset.

Impact: The hippocampal segmentation model presented in this work provides significant improvements over currently available tools in an external, clinical dataset. Segmentation performance was increased, while run-times were decreased. These results support the tool as a viable alternative in clinical settings.

Keywords: Microstructure, Microstructure

Motivation: Traditional image segmentation uses conventional MRI to classify tissue types based on image intensity. However, segmenting using microstructural data may provide more specific classification of tissue.

Goal(s): Cluster and segment brain tissue using quantitative MRI measures, to classify tissue based only on microstructural features without spatial input.

Approach: Measures denoting myelin content, anisotropy and tissue heterogeneity were clustered and used to label test datasets based on microstructural features alone, using an unsupervised approach.

Results: Segmentations were more informative than traditional segmentation, and consistent between healthy subjects. Differences between clusters reflect microstructural feature differences which would otherwise be invisible with conventional imaging.

Impact: The CAQE framework can be used for segmentation of tissue based on quantitative measures alone, providing better delineation of regions based on microstructural features. This allows for future comparisons between healthy and damaged tissue, to visualise and interpret pathological changes.

Keywords: Electromagnetic Tissue Properties, Machine Learning/Artificial Intelligence, Brachytherapy, Segmentation, Susceptibility

Motivation: Deep learning (DL) networks trained with synthetically generated data enable the visualization of I-125 brachytherapy seeds in prostate cancer patients in quantitative susceptibility mapping (QSM), possibly eliminating the need for a CT-scan in future.

Goal(s): The Goal was to automatically detect and segment I-125 seeds in-vivo by using a DL network directly (without QSM) on gradient-echo-sequence (GRE) data.

Approach: A U-Net was trained with synthetically generated multi-echo GRE magnitude and phase input data and corresponding target seed segmentations.

Results: The seed segmentations were of high visual quality and showed good agreement (85% detection rate) with corresponding CT-scans in five prostate cancer patients.

Impact: This work proposes a fast and completely automatic MRI-only based workflow for segmenting in-vivo brachytherapy seeds in prostate cancer patients. This approach has the potential to eliminate the need for a CT-scan, thereby reducing the use of ionizing radiation.

Keywords: Segmentation, Whole Body

Motivation: Numerous studies have made significant strides in the field of medical image segmentation. However, most studies have focused on specific localized regions rather than addressing the challenge of unified segmentation across the entire human body.

Goal(s): To improve the efficiency and accuracy of disease diagnosis and treatment, continuous advancements in multi-organ segmentation brings great advantages.

Approach:  In this paper, we present a prompt guided multi-organ segmentation model on total-body images, which can be adapted for CT, PET and MRI modalities. 

Results: Our extensive experiments demonstrate the superior performance of our model in accurately segmenting 21 organs. 

Impact: Our research leverages the power of prompts to tackle the challenge of multi-organ segmentation. It has potentially wide applications in the fields of CT, MRI and PET, enabling the simultaneous segmentation of multiple organs and images from diverse modalities.

Keywords: Segmentation, MR Fingerprinting

Motivation:  Segmentation of quantitative maps is required to compute anatomical region mean relaxation times. FreeSurfer is designed to automatically segment conventional weighted images. Voxels of CSF contaminate tissue segmentations in some healthy volunteers, skewing the mean T₁ or T₂.

Goal(s): To remove contaminant fluid from tissue regions of interest in T₁ and T₂ maps.

Approach:  Mean T₁ or T₂ of CSF in ventricles is used as a maximum threshold within brain tissue.

Results: Thresholding prior to 2D erosion of masks removes contaminant CSF and prevents erroneous variation between 10 healthy volunteers.

Impact: A simple threshold-based correction of FreeSurfer segmentation applied to quantitative T₁ and T₂ maps. The same threshold can be applied to all subjects in one step, eliminating the need for laborious manual adjustment.

Keywords: Cancer, Tumor, Nutritional Intervention, MR spectroscopy, Tumor segmentation

Motivation: The evaluation of MR spectroscopy imaging findings with multiparametric brain tumor segmentation might facilitate the understanding of altered tumor metabolism induced by intervention on an individual patient/tumor level.

Goal(s): In this study, we used this approach to elucidate the glioma metabolism under nutritional intervention.

Approach: The concentrations of ketone bodies in brain tumor after 72-hour-fasting were correlated with the volume of glioma sub-regions for 13 brain tumor patients.

Results: The outcome indicates that the accumulation of ketone bodies in solid tumors and necrotic areas after fasting might be a result of neovascularization and the blood-brain barrier compromise.

Impact: Here, we report on the validation of a dedicated, multi-voxel MRSI protocol with fully automated multiparametric segmentation of glioma sub-regions for monitoring fasting-induced changes. 

Keywords: Data Processing, Data Analysis, fiber quantification, tract segmentation, reliability

Motivation: The reliability of fiber quantification depends on the quality of the fiber tract reconstruction, which is influenced by the method used for fiber segmentation.

Goal(s): Our objective was to assess the reliability of different fiber segmentation methods and determine the most suitable strategy for the specific bundle.

Approach: We compared the distributions of intra-class coefficient (ICC) for different measurements and tracts across three widely used fiber segmentation methods. This analysis was conducted using the traveling subject dataset. The results are presented along with examples, followed by a discussion of the underlying reasons.

Results: We summarized the advantageous strategies for main tracts.

Impact: This strategy aims to enhance the stability of study results by facilitating the selection of more efficient segmentation methods for conducting fiber-specific quantification in the future.