Vessel Segmentation Research Articles

Retraction Note: An optimal approach to detect retinal diseases by performing segmentation of retinal blood vessels using image processing

Retraction Note: An optimal approach to detect retinal diseases by performing segmentation of retinal blood vessels using image processing

RetVes segmentation: A pseudo-labeling and feature knowledge distillation optimization technique for retinal vessel channel enhancement

Recent advancements in retinal vessel segmentation, which employ transformer-based and domain-adaptive approaches, show promise in addressing the complexity of ocular diseases such as diabetic retinopathy. However, current algorithms face challenges in effectively accommodating domain-specific variations and limitations of training datasets, which fail to represent real-world conditions comprehensively. Manual inspection by specialists remains time-consuming despite technological progress in medical imaging, underscoring the pressing need for automated and robust segmentation techniques. Additionally, these methods have deficiencies in handling unlabeled target sets, requiring extra preprocessing steps and manual intervention, which hinders their scalability and practical application in clinical settings. This research introduces a novel framework that employs semi-supervised domain adaptation and contrastive pre-training to address these limitations. The proposed model effectively learns from target data by implementing a novel pseudo-labeling approach and feature-based knowledge distillation within a temporal convolutional network (TCN) and extracts robust, domain-independent features. This approach enhances cross-domain adaptation, significantly enhancing the model's versatility and performance in clinical settings. The semi-supervised domain adaptation component overcomes the challenges posed by domain shifts, while pseudo-labeling utilizes the data's inherent structure for enhanced learning, which is particularly beneficial when labeled data is scarce. Evaluated on the DRIVE and CHASE_DB1 datasets, which contain clinical fundus images, the proposed model achieves outstanding performance, with accuracy, sensitivity, specificity, and AUC values of 0.9792, 0.8640, 0.9901, and 0.9868 on DRIVE, and 0.9830, 0.9058, 0.9888, and 0.9950 on CHASE_DB1, respectively, outperforming current state-of-the-art vessel segmentation methods. The partitioning of datasets into training and testing sets ensures thorough validation, while extensive ablation studies with thorough sensitivity analysis of the model's parameters and different percentages of labeled data further validate its robustness.

101 5-HT4 receptor-mediated vasorelaxation occurs through a cyclic mononucleotide-dependent mechanism in the isolated bovine lateral saphenous vein

Abstract Serotonin (5-HT)-mediated vasorelaxation of the isolated bovine lateral saphenous vein occurs primarily through 5-HT4 receptor activation. Vasorelaxation responses mediated through 5-HT4 have only been documented to occur in cattle and sheep and thus, may represent a novel vasorelaxation mechanism that is specific to ruminants. In tissues from other species, the 5-HT4 receptor and its associated G protein couple with adenylate cyclase. However, the mechanisms contributing to 5-HT4-mediated vasorelaxation in vascular tissue remain undefined. To better understand the mechanisms of 5-HT4-mediated vasorelaxation, lateral saphenous veins from cattle (n = 4) were collected and assessed for vasoactivity in response to increasing concentrations of a selective 5-HT4 receptor agonist (BIMU-8) in the absence and presence of selective inhibitors of downstream proteins involved with cyclic mononucleotide-mediated signaling. Isolated vessel segments were suspended in chambers of a multi-myograph containing 5 mL of continuously oxygenated Krebs-Henseleit buffer and equilibrated to 1 g tension for 90 min. After the equilibration period, vessels were exposed to dimethyl sulfoxide (DMSO; vehicle control) or selective inhibitors of downstream protein targets dissolved in DMSO (1×10-5 M) for 5 min. Vessels were pre-contracted with 1×10-4 M phenylephrine for 10 min to achieve a sustained increase in tension. After the 10-min pre-contraction period, vessels were exposed to increasing concentrations (1×10-9 to 1×10-4) of BIMU-8 non-cumulatively. Vasoactive response data were normalized as a percentage of the maximum contractile response induced by the phenylephrine pre-contraction. In the absence of selective inhibitors (control), BIMU-8 induced (P < 0.01) dose-dependent relaxations (mean -log EC50 = 6.30 M) with a 112% increase in maximal vasorelaxation activity occurring at 1×10-4 M. In the presence of the selective inhibitors for adenylate cyclase (NKY80) or guanylate cyclase (NS2028), 66% and 86% of the vasorelaxation responses to BIMU-8 were attenuated (P < 0.01), respectively. In the presence of the selective inhibitors for protein kinase A [protein kinase A inhibitor fragment (6-22) amide] or protein kinase G (Rp-8-bromo-PET-Cyclic GMPS), 66% and 97% of the maximal vasorelaxation responses to BIMU-8 were attenuated (P < 0.01), respectively. In the presence of non-selective inhibitors for Ca2+ channels and K+ channels, vasorelaxation responses to BIMU-8 were attenuated (P < 0.01). In the presence of selective inhibitors for the protein phosphatase 1 and 2A subunits of myosin light chain (calyculin A) and heat shock protein synthesis (quercetin), vasorelaxation responses to BIMU-8 were attenuated (P < 0.01). The findings of the current study demonstrate that 5-HT4-receptor mediated vasorelaxation occurs through a cyclic mononucleotide-dependent mechanism involving cross-activation of protein kinase A and G signaling. Downstream proteins involved in the signaling cascade leading to vasorelaxation include Ca2+ channels, K+ channels, myosin light chain phosphatase, and heat shock proteins. More research is needed to fully describe the cellular and molecular mechanisms responsible for 5-HT4 receptor-mediated vasorelaxation.

Predictive value of follow-up infarct volume on functional outcomes in middle cerebral artery M2 segment vessel occlusion stroke treated with mechanical thrombectomy.

Medium vessel occlusion (MeVO) strokes, particularly affecting the M2 segment of the middle cerebral artery, represent a critical proportion of acute ischemic strokes, posing significant challenges in management and outcome prediction. The efficacy of mechanical thrombectomy (MT) in MeVO stroke may warrant reliable predictors of functional outcomes. This study aimed to investigate the prognostic value of follow-up infarct volume (FIV) for predicting 90-day functional outcomes in MeVO stroke patients undergoing MT. This multicenter, retrospective cohort study analyzed data from the Multicenter Analysis of primary Distal medium vessel occlusions: effect of Mechanical Thrombectomy (MAD-MT) registry, covering patients with acute ischemic stroke due to M2 segment occlusion treated with MT. We examined the relationship between 90-day functional outcomes, measured by the modified Rankin Scale (mRS), and follow-up infarct volume (FIV), assessed through CT or MRI within 12-36 h post-MT. Among 130 participants, specific FIV thresholds were identified with high specificity and sensitivity for predicting outcomes. A FIV ⩽5 ml was highly specific for predicting favorable and excellent outcomes. The optimal cut-off for both prognostications was identified at ⩽15 ml by the Youden Index, with significant reductions in the likelihood of favorable outcomes observed above a 40 ml threshold. Receiver Operator Curve (ROC) analyses confirmed FIV as a superior predictor of functional outcomes compared to traditional recanalization scores, such as final modified thrombolysis in cerebral infarction score (mTICI). Multivariable analysis further highlighted the inverse relationship between FIV and positive functional outcomes. FIV within 36 h post-MT serves as a potent predictor of 90-day functional outcomes in patients with M2 segment MeVO strokes. Establishing FIV thresholds may aid in the prognostication of stroke outcomes, suggesting a role for FIV in guiding post intervention treatment decisions and informing clinical practice. Future research should focus on validating these findings across diverse patient populations and exploring the integration of FIV measurements with other clinical and imaging markers to enhance outcome prediction accuracy.

VesselBoost: A Python Toolbox for Small Blood Vessel Segmentation in Human Magnetic Resonance Angiography Data

Magnetic resonance angiography (MRA) performed at ultra-high magnetic field provides a unique opportunity to study the arteries of the living human brain at the mesoscopic level. From this, we can gain new insights into the brain’s blood supply and vascular disease affecting small vessels. However, for quantitative characterization and precise representation of human angioarchitecture to, for example, inform blood-flow simulations, detailed segmentations of the smallest vessels are required. Given the success of deep learning-based methods in many segmentation tasks, we explore their application to high-resolution MRA data and address the difficulty of obtaining large data sets of correctly and comprehensively labelled data. We introduce VesselBoost, a vessel segmentation toolbox, which utilizes deep learning and imperfect training labels for accurate vasculature segmentation. To enhance the segmentation models’ robustness and accuracy, VesselBoost employs an innovative data augmentation technique, which captures the resemblance of vascular structures across scales by zooming in or out on input image patches—virtually creating diverse scale vascular data. This approach enables detailed vascular segmentation and ensures the model’s ability to generalize across various scales of vascular structures.

Edge morphology attention mechanism and optimal geometric matching connection model for vascular segmentation

Over the past decades, medical image segmentation methods have been intensively developed, but there are still a number of unsolved challenging problems in vascular image segmentation, including broken vessels, insufficient vessel branches, and missing small vessels. In order to optimize the topology and accuracy of segmented vessels, we propose a novel Edge Morphology Attention Network (EMA-Net) for the segmentation of vessel-like structures, and an Optimal Geometric Matching Connection (OGMC) model to connect the broken vessel fragments. The EMA-Net has an edge attention module that is able to improve segmented performances of edges and small branches by morphology operation extracting boundary voxels on multi-scales. The OGMC model uses the geometry concept of curve touching to filter out fragmented vessel endpoints, and then employs minimal surfaces to determine the optimal connection order between blood vessels. Finally, we adopt geodesics to repair missing vessels under a Riemannian metric. Our method achieves superior or competitive results compared to state-of-the-art methods on four datasets of 3D vascular segmentation tasks, both effectively reducing vessel broken and increasing vessel branch richness, yielding blood vessels with more precise topological structures.

Thin vessel segmentation in fundus images using attention UNet and modified Frangi filtering

Thin vessel segmentation is an active research problem, with an emphasis on finding a universal approach for different types of fundus datasets. Enhancement of the thin vessels is the first and foremost task for proper segmentation, which is proposed to be done with the total variation (TV) decomposition method with layer-selective enhancement and illumination correction. The vessel segmentation task is carried out on two fronts. For thin vessels, we propose the attention UNet backbone, and for thick vessels, the modified Frangi method is used. The method is trained and tested on HRF, CHASE_DB1, and DRIVE datasets with accuracy of 96.73%, 96.38%, and 94.97%, and sensitivity of 92.66%, 81.05%, and 87.56%, respectively. The method was cross-validated on the STARE dataset with an accuracy of 96.30% and a sensitivity of 93.69%. The sensitivity performance surpasses the state of the art.

Multi-GlaucNet: A multi-task model for optic disc segmentation, blood vessel segmentation and glaucoma detection

Glaucoma is a common and severe ocular disease that often leads to vision loss. The information on the optic disc (OD) and blood vessels in fundus images can significantly aid in glaucoma detection. In addition, the use of deep learning models for glaucoma detection is a highly effective approach. We propose a multi-task deep learning model called Multi-GlaucNet that can simultaneously segment the OD and blood vessels, thereby assisting doctors in diagnosing glaucoma. Multi-GlaucNet consists of three modules: the OD segmentation module, blood vessel segmentation module and glaucoma detection module. The OD segmentation module and blood vessel segmentation module both adopt an encoder–decoder structure to segment OD and blood vessels, respectively. The segmentation module is constructed using bottleneck layers in the encoding process and uses Pixel Shuffle and channel attention mechanisms in the decoding process. The detection module uses the ResNet50 network to perform glaucoma detection based on the features extracted from the segmentation modules. Multi-GlaucNet demonstrates outstanding performance in three areas. It achieves a Dice coefficient of 96.7% for OD segmentation on the ORIGA dataset, accuracy of 0.9798 and F1 score of 0.8562 for blood vessel segmentation on a mixed dataset. For glaucoma detection on the REFUGE dataset, it attains the highest accuracy of 0.967 and an area under the curve (AUC) of 0.950. These results validate the effectiveness of Multi-GlaucNet for glaucoma detection. The model’s ability to perform multiple tasks with high accuracy and efficiency demonstrates that it is a valuable aiding tool for glaucoma diagnosis.

MPCCN: A Symmetry-Based Multi-Scale Position-Aware Cyclic Convolutional Network for Retinal Vessel Segmentation

In medical image analysis, precise retinal vessel segmentation is crucial for diagnosing and managing ocular diseases as the retinal vascular network reflects numerous health indicators. Despite decades of development, challenges such as intricate textures, vascular ruptures, and undetected areas persist, particularly in accurately segmenting small vessels and addressing low contrast in imaging. This study introduces a novel segmentation approach called MPCCN that combines position-aware cyclic convolution (PCC) with multi-scale resolution input to tackle these challenges. By integrating standard convolution with PCC, MPCCN effectively captures both global and local features. A multi-scale input module enhances feature extraction, while a weighted-shared residual and guided attention module minimizes background noise and emphasizes vascular structures. Our approach achieves sensitivity values of 98.87%, 99.17%, and 98.88%; specificity values of 98.93%, 97.25%, and 99.20%; accuracy scores of 97.38%, 97.85%, and 97.75%; and AUC values of 98.90%, 99.15%, and 99.05% on the DRIVE, STARE, and CHASE_DB1 datasets, respectively. In addition, it records F1 scores of 90.93%, 91.00%, and 90.55%. Experimental results demonstrate that our method outperforms existing techniques, especially in detecting small vessels.

Image decomposition based segmentation of retinal vessels

Image decomposition based segmentation of retinal vessels

A quasilinear hyperbolic one-dimensional model of the lymph flow through a lymphangion with valve dynamics and a contractile wall

This paper presents a one-dimensional model that describes fluid flow in lymphangions, the segments of lymphatic vessels between valves, using quasilinear hyperbolic systems. The model incorporates a phenomenological pressure-cross-sectional area relationship based on existing literature. Numerical solutions of the differential equations align with known results, offering insights into lymphatic flow dynamics. This model enhances the understanding of lymph movement through the lymphatic system, driven by lymphangion contractions.

Structural and functional fetal cardiac imaging using low field (0.55 T) MRI.

This study aims to investigate the feasibility of using a commercially available clinical 0.55 T MRI scanner for comprehensive structural and functional fetal cardiac imaging. Balanced steady-state free precession (bSSFP) and phase contrast (PC) sequences were optimized by in utero studies consisting of 14 subjects for bSSFP optimization and 9 subjects for PC optimization. The signal-to-noise ratio (SNR) of the optimized sequences were investigated. Flow measurements were performed in three vessels, umbilical vein (UV), descending aorta (DAo), and superior vena cava (SVC) using the PC sequences and retrospective gating. The optimized bSSFP, PC and half-Fourier single shot turbo spin-echo (HASTE) sequences were acquired in a cohort of 21 late gestation-age fetuses (>36 weeks) to demonstrate the feasibility of a fetal cardiac exam at 0.55 T. The HASTE stacks were reconstructed to create an isotropic reconstruction of the fetal thorax, followed by automatic great vessel segmentations. The intra-abdominal UV blood flow measurements acquired with MRI were compared to ultrasound UV free-loop flow measurements. Using the parameters from 1.5 T as a starting point, the bSSFP sequences were optimized at 0.55 T, resulting in a 1.6-fold SNR increase and improved image contrast compared to starting parameters, as well as good visibility of most cardiac structures as rated by two experienced fetal cardiologists. The PC sequence resulted in increased SNR and reduced scan time, subsequent retrospective gating enabled successful blood flow measurements. The reconstructions and automatic great vessel segmentations showed good quality, with 18/21 segmentations requiring no or minor refinements. Blood flow measurements were within the expected range. A comparison of the UV measurements performed with ultrasound and MRI showed agreement between the two sets of measurements, with better correlation observed at lower flows. We demonstrated the feasibility of low-field (0.55 T) MRI for fetal cardiac imaging. The reduced SNR at low field strength can be effectively compensated for by strategically optimizing sequence parameters. Major fetal cardiac structures and vessels were consistently visualized, and flow measurements were successfully obtained. The late gestation study demonstrated the robustness and reproducibility at low field strength. MRI performed at 0.55 T is a viable option for fetal cardiac examination.

Vascular liver segmentation: a narrative review on methods and new insights brought by artificial intelligence.

Liver vessel segmentation from routinely performed medical imaging is a useful tool for diagnosis, treatment planning and delivery, and prognosis evaluation for many diseases, particularly liver cancer. A precise representation of liver anatomy is crucial to define the extent of the disease and, when suitable, the consequent resective or ablative procedure, in order to guarantee a radical treatment without sacrificing an excessive volume of healthy liver. Once mainly performed manually, with notable cost in terms of time and human energies, vessel segmentation is currently realized through the application of artificial intelligence (AI), which has gained increased interest and development of the field. Many different AI-driven models adopted for this aim have been described and can be grouped into different categories: thresholding methods, edge- and region-based methods, model-based methods, and machine learning models. The latter includes neural network and deep learning models that now represent the principal algorithms exploited for vessel segmentation. The present narrative review describes how liver vessel segmentation can be realized through AI models, with a summary of model results in terms of accuracy, and an overview on the future progress of this topic.

3D Vessel Segmentation With Limited Guidance of 2D Structure-Agnostic Vessel Annotations.

Delineating 3D blood vessels of various anatomical structures is essential for clinical diagnosis and treatment, however, is challenging due to complex structure variations and varied imaging conditions. Although recent supervised deep learning models have demonstrated their superior capacity in automatic 3D vessel segmentation, the reliance on expensive 3D manual annotations and limited capacity for annotation reuse among different vascular structures hinder their clinical applications. To avoid the repetitive and costly annotating process for each vascular structure and make full use of existing annotations, this paper proposes a novel 3D shape-guided local discrimination (3D-SLD) model for 3D vascular segmentation under limited guidance from public 2D vessel annotations. The primary hypothesis is that 3D vessels are composed of semantically similar voxels and often exhibit tree-shaped morphology. Accordingly, the 3D region discrimination loss is firstly proposed to learn the discriminative representation measuring voxel-wise similarities and cluster semantically consistent voxels to form the candidate 3D vascular segmentation in unlabeled images. Secondly, the shape distribution from existing 2D structure-agnostic vessel annotations is introduced to guide the 3D vessels with the tree-shaped morphology by the adversarial shape constraint loss. Thirdly, to enhance training stability and prediction credibility, the highlighting-reviewing-summarizing (HRS) mechanism is proposed. This mechanism involves summarizing historical models to maintain temporal consistency and identifying credible pseudo labels as reliable supervision signals. Only guided by public 2D coronary artery annotations, our method achieves results comparable to SOTA barely-supervised methods in 3D cerebrovascular segmentation, and the best DSC in 3D hepatic vessel segmentation, demonstrating the effectiveness of our method.

Deep Learning-Based Liver Vessel Segmentation

Abstract Liver vessel segmentation in computed tomography represents a highly challenging task due to the imbalanced distribution within the liver parenchyma, the small and branched vessels with decreased image contrast to surrounding tissue and in general, due to the scarcity of highresolution and -contrast images, which hampers the efficient training of deep learning-based approaches. This study applies two state-of-the-art networks, the fully convolutional nnUnet and the transformer-based VT-Unet to three publicly available datasets, 3DIRCADb, one task of the Medical Segmentation Decathlon (MSD) and the more recent LiVS dataset. The nnUnet achieved Dice scores of 0.761, 0.714, and 0.696 on the 3DIRCADb, LiVS, and MSD datasets, respectively. In contrast, the experiments with the VT-UNet resulted in Dice scores of 0.795, 0.713, and 0.610. These findings indicates good accordance of the performance of the nnUnet and the transformer-based VT-Unet, with differences regarding individual datasets. Both network variants show competitive performances regarding the current state-of-the-art, yet the need for large-scale and high-quality datasets becomes evident to further enhance the accuracy of liver vessel segmentation.

Comparison of Duplex Ultrasound and Digital Subtraction Angiography for Assessing Tibial Vessel Disease.

Background Duplex ultrasonography (DUS) is readily available and often used as the first diagnostic test for patients with peripheral artery diseases (PADs). PAD is a disease that affects the general population but more commonly affects diabetics.To date, the role of DUS in the assessment of tibial vessel disease is inconclusive at best. The goal of our study is to assess the validity of DUS in characterizing the presence and severity of tibial diseasesvia comparison with digital subtraction angiography (DSA) findings. Methods This is a single-center retrospective cohort study analyzing three arterial segments (anterior tibial, posterior tibial, and fibular arteries) in patients who received a duplex study followed by DSA within a 30-day period. All arterial segments were graded from normal (Grade 0) to occluded (Grade 4), based on duplex interpretationand directly compared to direct visualization findings from DSA. Using statistical methods, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of DUS were determined. Results A total of 171 tibial vessel segments from 57 enrolled subjects with critical limb ischemia symptoms were analyzed in this study. The agreement between both modalities was poor (Kappa=0.19, p < 0.05), with DUS demonstrating a significant underestimation of vessel pathologies. This is also reflected by the overall sub-optimal sensitivity (23%), specificity (84%), PPV (69%), and NPV (41%) in DUS when compared to DSA results as the gold standard. Conclusion Significant disagreements were noted in this study between DUS and DSA findings, primarilysignificant underestimation of tibial vessel disease by the DUS when compared with the DSA. Caution is advised in the clinical application of DUS in patients with chronic limb-threatening ischemia (CLTI) symptoms and multi-segment tibial vessels due to its demonstrated limitations in this study.

Retinal Vessel Segmentation by a Transformer-U-Net Hybrid Model With Dual-Path Decoder.

This paper introduces an effective and efficient framework for retinal vessel segmentation. First, we design a Transformer-CNN hybrid model in which a Transformer module is inserted inside the U-Net to capture long-range interactions. Second, we design a dual-path decoder in the U-Net framework, which contains two decoding paths for multi-task outputs. Specifically, we train the extra decoder to predict vessel skeletons as an auxiliary task which helps the model learn balanced features. The proposed framework, named as TSNet, not only achieves good performances in a fully supervised learning manner but also enables a rough skeleton annotation process. The annotators only need to roughly delineate vessel skeletons instead of giving precise pixel-wise vessel annotations. To learn with rough skeleton annotations plus a few precise vessel annotations, we propose a skeleton semi-supervised learning scheme. We adopt a mean teacher model to produce pseudo vessel annotations and conduct annotation correction for roughly labeled skeletons annotations. This learning scheme can achieve promising performance with fewer annotation efforts. We have evaluated TSNet through extensive experiments on five benchmarking datasets. Experimental results show that TSNet yields state-of-the-art performances on retinal vessel segmentation and provides an efficient training scheme in practice.

Noncontrast free-breathing ECG-gated 3D balanced steady-state free precession in congenital heart disease and aortopathy evaluation.

High-fidelity cardiac magnetic resonance (MR) imaging plays a pivotal role in the surveillance of congenital heart disease (CHD) and aortopathy. We aimed to evaluate the quality and accuracy of free breathing, ECG-gated noncontrast three-dimensional (3D) balanced steady-state free precession (bSSFP) in cases of CHDs and aortopathies using contrast-enhanced 3D bSSFP as a reference. We also used one of our routinely used non-ECG-gated 2D-single-shot (SSh) bSSFP sequence as an adjunct to noncontrast 3D bSSFP. Institutional review board approval was obtained to perform a systematic retrospective analysis of image quality and vascular measurements. Patients with CHD and aortopathy, who were undergoing clinically indicated contrast-enhanced 3D bSSFP, were prospectively identified to also undergo additional noncontrast 3D bSSFP and 2D SSh bSSFP imaging as part of a clinical quality improvement initiative aimed at reducing the use of contrast when feasible. Two readers, blinded to each other's evaluations, graded image quality on a 5-point Likert scale and performed vascular measurements in separate sessions for both 3D bSSFP images. They also reported the visibility of various mediastinal great vessels on 2D SSh bSSFP images. Raw agreement, the weighted kappa statistic, and intra-class correlation coefficients (ICCs) were computed to assess the consistency and agreement between the two readers. Comparative analysis of noncontrast and contrast-enhanced 3D bSSFP imaging was performed in adult and pediatric patients using a two-sided paired t-test and Bland-Altman analysis. A P-value < 0.05 was considered significant for all inference testing. A total of 29 patients (17 males, median age 20.3years, interquartile range (IQR) 12.5, age range 7-39years), including 11 pediatric patients under the age of 18years (6 males, median age 14.5years, IQR 4.0, age range 7-17years), underwent retrospective analysis. The overall image quality score for contrast-enhanced 3D bSSFP was significantly higher (P < 0.0001) than that of noncontrast 3D bSSFP for both all subjects (4.4 ± 0.2, range 4.0-4.9 vs 3.7 ± 0.4, range 3.1-4.7) and only pediatric subjects (4.3 ± 0.3, range 4.0-4.9 vs 3.6 ± 0.5, range 3.1-4.4). By combining noncontrast 3D bSSFP and 2D bSSFP, reader 1 and reader 2 rated 423 and 420 vessels diagnostic, respectively, in a total of 435 vessel segments. All landmarks showed similar mean vessel diameters without significant differences between noncontrast and contrast-enhanced 3D bSSFP MR angiography (r = 0.99, bias -0.31mm, 95% limits of agreement -2.04mm to 1.43mm). Although contrast-enhanced images had better overall image quality, an imaging protocol consisting of noncontrast 2D SSh bSSFP and 3D bSSFP whole-chest images provides diagnostically adequate image quality, and accurate vascular measurements, comparable to free-breathing contrast-enhanced 3D bSSFP in both children and adults with CHD and aortopathies.

LVI-PathNet: Segmentation-classification pipeline for detection of lymphovascular invasion in whole slide images of lung adenocarcinoma

Lymphovascular invasion (LVI) in lung cancer is a significant prognostic factor that influences treatment and outcomes, yet its reliable detection is challenging due to interobserver variability. This study aims to develop a deep learning model for LVI detection using whole slide images (WSIs) and evaluate its effectiveness within a pathologist's information system. Experienced pathologists annotated blood vessels and invading tumor cells in 162 WSIs of non-mucinous lung adenocarcinoma sourced from two external and one internal datasets. Two models were trained to segment vessels and identify images with LVI features. DeepLabV3+ model achieved an Intersection-over-Union of 0.8840 and an area under the receiver operating characteristic curve (AUC-ROC) of 0.9869 in vessel segmentation. For LVI classification, the ensemble model achieved a F1-score of 0.9683 and an AUC-ROC of 0.9987. The model demonstrated robustness and was unaffected by variations in staining and image quality. The pilot study showed that pathologists' evaluation time for LVI detecting decreased by an average of 16.95%, and by 21.5% in “hard cases”. The model facilitated consistent diagnostic assessments, suggesting potential for broader applications in detecting pathological changes in blood vessels and other lung pathologies.

Real-time placental vessel segmentation in fetoscopic laser surgery for Twin-to-Twin Transfusion Syndrome

Twin-to-Twin Transfusion Syndrome (TTTS) is a rare condition that affects about 15% of monochorionic pregnancies, in which identical twins share a single placenta. Fetoscopic laser photocoagulation (FLP) is the standard treatment for TTTS, which significantly improves the survival of fetuses. The aim of FLP is to identify abnormal connections between blood vessels and to laser ablate them in order to equalize blood supply to both fetuses. However, performing fetoscopic surgery is challenging due to limited visibility, a narrow field of view, and significant variability among patients and domains. In order to enhance the visualization of placental vessels during surgery, we propose TTTSNet, a network architecture designed for real-time and accurate placental vessel segmentation. Our network architecture incorporates a novel channel attention module and multi-scale feature fusion module to precisely segment tiny placental vessels. To address the challenges posed by FLP-specific fiberscope and amniotic sac-based artifacts, we employed novel data augmentation techniques. These techniques simulate various artifacts, including laser pointer, amniotic sac particles, and structural and optical fiber artifacts. By incorporating these simulated artifacts during training, our network architecture demonstrated robust generalizability. We trained TTTSNet on a publicly available dataset of 2060 video frames from 18 independent fetoscopic procedures and evaluated it on a multi-center external dataset of 24 in-vivo procedures with a total of 2348 video frames. Our method achieved significant performance improvements compared to state-of-the-art methods, with a mean Intersection over Union of 78.26% for all placental vessels and 73.35% for a subset of tiny placental vessels. Moreover, our method achieved 172 and 152 frames per second on an A100 GPU, and Clara AGX, respectively. This potentially opens the door to real-time application during surgical procedures. The code is publicly available at https://github.com/SanoScience/TTTSNet.