Vessel Enhancement Research Articles

Abstract TP89: Ldl Levels May Correlate With Intracranial Atherosclerotic Plaque Enhancement

Introduction: The enhancement of atherosclerotic plaques after the intravenous administration of gadolinium (Gd) has been correlated with the presence of culprit plaques in patients with acute strokes due to underlying intracranial atherosclerosis (ICAS). A new method to objectively quantify Gd-enhancement was used to monitor plaques over time after an acute stroke. Currently there is no imaging method to accurately monitor atherosclerotic activity. Methods: Patients with stroke or transient ischemic attack caused by ICAS were included. Every patient was started on high-dose statins after ictus. 7T MRI images were obtained at baseline and after at least 6 months. LDL levels were obtained at the time of the event and at follow-up. Culprit plaques were identified based on degree of stenosis, presence of positive remodeling and Gd-enhancement. Arterial 3D enhancement maps in the territory of the culprit plaque were generated. Orthogonal signal intensity (SI) probes were expanded into the plaque and parent vessel to generate 3D enhancement color maps. Arterial segment variance (ASV) (SD 2 Arterial segment Baseline T1+Gd - SD 2 Arterial segment Follow up T1+Gd ) was used to compare the plaque and parent vessel enhancement. Results: Five patients underwent baseline and follow-up imaging. The median time between scans was 13.3 (IQR=13.3) months. ASV was strongly correlated with changes in LDL (Rho=0.9, p=0.037). Additionally, Gd-enhancement and plaque burden correlated with levels of LDL (Figure): Gd uptake (Rho=0.7, p=0.22) and plaque burden (Rho=0.67, p=0.22). Conclusion: This pilot study demonstrates that high-resolution imaging can be used in determining the response to statins in patients with high atherosclerotic burden. Gd-enhancement and plaque burden correlated with serum values of LDL. 3D enhancement maps and histogram analysis of atherosclerotic plaques are promising tools for evaluating disease activity and response to medical therapy.

Automatic Hepatic Vessels Segmentation Using RORPO Vessel Enhancement Filter and 3D V-Net with Variant Dice Loss Function

The segmentation of hepatic vessels is crucial for liver surgical planning. It is also a challenging task because of its small diameter. Hepatic vessels are often captured in images of low contrast and resolution. Our research uses filter enhancement to improve their contrast, which helps with their detection and final segmentation. We have designed a specific fusion of the Ranking Orientation Responses of Path Operators (RORPO) enhancement filter with a raw image, and we have compared it with the fusion of different enhancement filters based on Hessian eigenvectors. Additionally, we have evaluated the 3D U-Net and 3D V-Net neural networks as segmentation architectures, and have selected 3D V-Net as a better segmentation architecture in combination with the vessel enhancement technique. Furthermore, to tackle the pixel imbalance between the liver (background) and vessels (foreground), we have examined several variants of the Dice Loss functions, and have selected the Weighted Dice Loss for its performance. We have used public 3D Image Reconstruction for Comparison of Algorithm Database (3D-IRCADb) dataset, in which we have manually improved upon the annotations of vessels, since the dataset has poor-quality annotations for certain patients. The experiments demonstrate that our method achieves a mean dice score of 76.2%, which outperforms other state-of-the-art techniques.

A Benchmark Framework for Multiregion Analysis of Vesselness Filters.

Vessel enhancement (aka vesselness) filters, are part of angiographic image processing for more than twenty years. Their popularity comes from their ability to enhance tubular structures while filtering out other structures, especially as a preliminary step of vessel segmentation. Choosing the right vesselness filter among the many available can be difficult, and their parametrization requires an accurate understanding of their underlying concepts and a genuine expertise. In particular, using default parameters is often not enough to reach satisfactory results on specific data. Currently, only few benchmarks are available to help the users choosing the best filter and its parameters for a given application. In this article, we present a generic framework to compare vesselness filters. We use this framework to compare seven gold standard filters. Our experiments are performed on three public datasets: the hepatic Ircad dataset (CT images), the Bullit dataset (brain MRA images) and the synthetic VascuSynth dataset. We analyse the results of these seven filters both quantitatively and qualitatively. In particular, we assess their performances in key areas: the organ of interest, the whole vascular network neighbourhood and the vessel neighbourhood split into several classes, based on their diameters. We also focus on the vessels bifurcations, which are often missed by vesselness filters. We provide the code of the benchmark, which includes up-to-date C++ implementations of the seven filters, as well as the experimental setup (parameter optimization, result analysis, etc.). An online demonstrator is also provided to help the community apply and visually compare these vesselness filters.

Automatic blood vessel detection using fractional Hessian matrices

The enhancement of blood vessels is a vital stage in imaging. The goal of this project is to improve the evaluation of the performance of a method for enhancing arteries in coronary angiograms, which use fractional derivatives. In this work an algorithm for automatic enhancement of vessels in coronary angiograms is evaluated, the method uses the Hessian matrix, the eigenvalues and the Grünwald-Letnikov fractional derivative with fractional order ω the in the interval (1,3). The probes of the performance of the method were made using a set of 20 coronary angiograms with its respective ground-truth image and the area under the ROC’s curve. The fractional orders are 2<ω and the second interval 2≥ω. The results show that the maximum values of area under the ROC’s curve are obtained when the derivative order is in the interval 2<ω<2.15.

Virtual non-contrast reconstructions improve differentiation between vascular enhancement and calcifications in stereotactic planning CT scans of cystic intracranial tumors

PurposeTo assess the diagnostic value of spectral-detector CT (SDCT) derived virtual non-contrast images (VNC) for differentiation between vascular enhancement and wall calcifications of cystic intracranial tumors in contrast-enhanced stereotactic planning examinations. Method48 patients with cystic intracranial tumors who underwent stereotactic SDCT examinations between 02/2017 and 02/2020 were retrospectively included. In each patient, two separate hyperattenuating structures along the cyst wall were defined as either enhancement or calcification, respectively, using reference MRI examinations. Quantitative analysis was performed ROI-based in conventional images (CI) and VNC. In the subjective analysis, two radiologists diagnosed the predefined peri-cystic structures in binary decisions as either enhancement or calcification using CI and the combination of CI and VNC, and rated diagnostic confidence, image noise and removal of iodine in VNC. Moreover, a potential diagnostic benefit of VNC was indicated. ResultsAttenuation in CI was higher as compared to VNC across all assessed ROI (all p < 0.01). In VNC, CNR between calcification and white matter was significantly higher as compared to CNR between vascular enhancement and white matter (2.6 vs 1.3, p < 0.01), while there was no significant difference in CI. In the qualitative assessment, diagnostic accuracy was significantly higher using both VNC and CI compared to using CI alone. Raters reported less image noise in VNC as compared to CI. An additional diagnostic benefit of VNC was indicated in 84.4 % of all cases. ConclusionsSDCT-derived VNC images facilitate differentiation between peri-cystic contrast enhancement in blood vessels and calcifications in stereotactic planning scans of cystic intracranial tumors.

DS6, Deformation-Aware Semi-Supervised Learning: Application to Small Vessel Segmentation with Noisy Training Data.

Blood vessels of the brain provide the human brain with the required nutrients and oxygen. As a vulnerable part of the cerebral blood supply, pathology of small vessels can cause serious problems such as Cerebral Small Vessel Diseases (CSVD). It has also been shown that CSVD is related to neurodegeneration, such as Alzheimer’s disease. With the advancement of 7 Tesla MRI systems, higher spatial image resolution can be achieved, enabling the depiction of very small vessels in the brain. Non-Deep Learning-based approaches for vessel segmentation, e.g., Frangi’s vessel enhancement with subsequent thresholding, are capable of segmenting medium to large vessels but often fail to segment small vessels. The sensitivity of these methods to small vessels can be increased by extensive parameter tuning or by manual corrections, albeit making them time-consuming, laborious, and not feasible for larger datasets. This paper proposes a deep learning architecture to automatically segment small vessels in 7 Tesla 3D Time-of-Flight (ToF) Magnetic Resonance Angiography (MRA) data. The algorithm was trained and evaluated on a small imperfect semi-automatically segmented dataset of only 11 subjects; using six for training, two for validation, and three for testing. The deep learning model based on U-Net Multi-Scale Supervision was trained using the training subset and was made equivariant to elastic deformations in a self-supervised manner using deformation-aware learning to improve the generalisation performance. The proposed technique was evaluated quantitatively and qualitatively against the test set and achieved a Dice score of 80.44 ± 0.83. Furthermore, the result of the proposed method was compared against a selected manually segmented region (62.07 resultant Dice) and has shown a considerable improvement (18.98%) with deformation-aware learning.

Imaging characteristics of ovarian sclerosing stromal tumor

ObjectivesThis study was designed to evaluate the specific imaging features of ovarian sclerosing stromal tumor (SST), improve its accuracy as well as the specificity of imaging diagnosing, and prevent overestimation of malignancy to reduce unnecessary surgical procedures.MethodsEight patients with magnetic resonance imaging (MRI) and six with computed tomography (CT) images were analyzed in this retrospective observational study. All the cases were confirmed by postoperative pathological examination as those of ovarian SST. Imaging and pathological features were also evaluated.ResultsAll the 14 masses displayed cystic and solid components with outer surface of tumors contained a capsular and complete smooth rim. Eight tumors of MRI exhibited “lake-island” sign on T2 weighted imaging (T2WI). Two of the 6 CT cases displayed a flaky calcification. One case showed as a multiloculated cystic mass with irregularly thickened septae and the tumor wall. The solid components in other 13 masses were comb- or wheel-like enhanced. After injection of contrast agent, the solid components in 8 cases (57.1%) appeared as early enhancement, whereas the other 6 cases (42.9%) appeared as progressive enhanced, and the cystic components of all the cases had no enhancement in the whole course. Vascular flow signals or/and marked enhancement of the blood vessels were found in 12 lesions (85.7%). Pathological examination demonstrated pseudolobule patterns, round to spindle shaped cells, collagenous areas, edematous hypocellular areas and prominent vasculatures.ConclusionsThe results demonstrated that MRI with “lake-island” signs on T2WI and MRI/CT dynamic enhancement could potentially play a critical role in facilitating appropriate diagnosis preoperatively.

Computation of contrast-enhanced perfusion using only two CT scan phases: a proof-of-concept study on abdominal organs

BackgroundComputed tomography perfusion imaging (CTPI) by repeated scanning has clinical relevance but implies relatively high radiation exposure. We present a method to measure perfusion from two CT scan phases only, considering tissue enhancement, feeding vessel (aortic) peak enhancement, and bolus shape.MethodsCTPI scans (each with 40 frames acquired every 1.5 s) of 11 patients with advanced hepatocellular carcinoma (HCC) enrolled between 2012 and 2016 were retrospectively analysed (aged 69 ± 9 years, 8/11 males). Perfusion was defined as the maximal slope of the time-enhancement curve divided by the peak enhancement of the feeding vessel (aorta). Perfusion was computed two times, first using the maximum slope derived from all data points and then using the peak tissue enhancement and the bolus shape obtained from the aortic curve.ResultsPerfusion values from the two methods were linearly related (r2 = 0.92, p < 0.001; Bland–Altman analysis bias -0.12). The mathematical model showed that the perfusion ratio of two ROIs with the same feeding vessel (aorta) corresponds to their peak enhancement ratio (r2 = 0.55, p < 0.001; Bland–Altman analysis bias -0.68). The relationship between perfusion and tissue enhancement is predicted to be linear in the clinical range of interest, being only function of perfusion, peak feeding vessel enhancement, and bolus shape.ConclusionsThis proof-of-concept study showed that perfusion values of HCC, kidney, and pancreas could be computed using enhancement measured only with two CT scan phases, if aortic peak enhancement and bolus shape are known.

3D Enhancement Color Maps in the Characterization of Intracranial Atherosclerotic Plaques.

High-resolution MR imaging allows the identification of culprit symptomatic plaques after the administration of gadolinium. Current high-resolution MR imaging methods are limited by 2D multiplanar views and manual sampling of ROIs. We analyzed a new 3D method to objectively quantify gadolinium plaque enhancement. Patients with stroke due to intracranial atherosclerotic disease underwent 7T high-resolution MR imaging. 3D segmentations of the plaque and its parent vessel were generated. Signal intensity probes were automatically extended from the lumen into the plaque and the vessel wall to generate 3D enhancement color maps. Plaque gadolinium (Gd) uptake was quantified from 3D color maps as gadolinium uptake = (µPlaque T1 + Gd -µPlaque T1/SDPlaque T1). Additional metrics of enhancement such as enhancement ratio, variance, and plaque-versus-parent vessel enhancement were also calculated. Conventional 2D measures of enhancement were collected for comparison. Thirty-six culprit and 44 nonculprit plaques from 36 patients were analyzed. Culprit plaques had higher gadolinium uptake than nonculprit plaques (P < .001). Gadolinium uptake was the most accurate metric for identifying culprit plaques (OR, 3.9; 95% CI 2.1-8.3). Gadolinium uptake was more sensitive (86% versus 70%) and specific (71% versus 68%) in identifying culprit plaques than conventional 2D measurements. A multivariate model, including gadolinium uptake and plaque burden, identified culprit plaques with an 83% sensitivity and 86% specificity. The new 3D color map method of plaque-enhancement analysis is more accurate for identifying culprit plaques than conventional 2D methods. This new method generates a new set of metrics that could potentially be used to assess disease progression.

Estimating the Neovascularity of Human Finger Tendon Through High-Frequency Ultrasound Micro-Doppler Imaging.

Neovascularization of injured tendons prolongs the proliferative phase of healing, but prolonged neovascularization may cause improper healing and pain. Currently, ultrasound Doppler imaging is used for measuring the neovascularization of injured tendons (e.g., Achilles tendon). However, the resolution of state-of-the-art clinical ultrasound machines is insufficient for visualizing the neovascularization in finger tendons. In this study, a high-frequency micro-Doppler imaging (HFμDI) based on 40-MHz ultrafast ultrasound imaging was proposed for visualizing the neovascularization in injured finger tendons during multiple rehabilitation phases. The vessel visibility was enhanced through a block-wise singular value decomposition filter and several curvilinear structure enhancement strategies, including the bowler-hat transform and Hessian-based vessel enhancement filtering. HFμDI was verified through small animal kidney and spleen imaging because the related vessel structure patterns of mice are well studied. Five patients with finger tendon injuries underwent HFμDI examination at various rehabilitation phases after surgery (weeks 11-56), and finger function evaluations were performed for comparisons. The results of small animal experiments revealed that the proposed HFμDI provides excellent microvasculature imaging performance; the contrast-to-noise ratio of HFμDI was approximately 15 dB higher than that of the conventional singular value decomposition filter, and the minimum detectable vessel size for mouse kidney was 35 μm without the use of contrast agent. In the human study, neovascularization was clearly observed in injured finger tendons during the early phase of healing (weeks 11-21), but it regressed from week 52 to 56. Finger rehabilitation appears to help reduce neovascularization; neovascular density decreased by approximately 1.8%-8.0% in participants after 4 weeks of rehabilitation. The experimental results verified the performance of HFμDI for microvasculature imaging and its potential for injured finger tendon evaluations.

An optimal statistical feature‐based transformation function for enhancement of retinal images using adaptive enhanced leader particle swarm optimization

AbstractRetinal images play a crucial role in the clinical diagnosis and detection of many eye diseases. However, the retinal image is degraded with noise and suffers from low contrast, which makes it difficult to interpret the image precisely and accurately. So retinal image enhancement becomes a vital step before any further processing. Here, a new approach is suggested for fundus image enhancement, where the image is processed in the spatial domain through an optimal statistical feature‐based transformation function (SFTF). The proposed SFTF has four tunable parameters. An optimal set of these parameters is obtained through a newly suggested adaptive enhanced leader particle swarm optimization technique. The suggested method is investigated with freely accessible DRIVE, STARE, and IDRiD databases. The efficiency of the suggested technique is assessed through different validation indices. The results reveal superior performance compared with other popular methods used for vessel enhancement.

MC-DMD: A data-driven method for blood vessel enhancement in retinal images using morphological closing and dynamic mode decomposition

Detection of blood vessels in retinal images is an important step in any computer-aided pathological system for the early screening and detection of eye-related ailments such as retinal detachment, diabetic retinopathy, and macular degeneration. Propriety of the vessel enhancement process, as a pre-processing step, has been well established in medical corpus, for enhanced accuracy of vessel segmentation. Algorithms for retinal image extraction are still not qualitatively sufficient to be appropriate for diagnostics. This paper presents an effective and robust approach for retinal vessel enhancement that overcome prevailing deficiencies, using Morphological Closing based Dynamic Mode Decomposition (MC-DMD). The proffered algorithm exploits the power of mathematical morphology for the generation of input channel to the Dynamic Mode Decompostion (DMD) system, decomposing the vessel and non-vessel features of retinal images. We confirmed the effectiveness of the proposed enhancement method on three publicly available retinal image datasets: DRIVE, STARE and HRF, across nine existing vessel enhancement methods in terms of Receiver Operating Characteristic (ROC) curve and Area Under the ROC Curve (AUC). In addition, segmentation on final vessel map is performed and validated by enhanced performance in comparison with eight conventional vessel segmentation algorithms in terms of Sensitivity (SEN), Specificity (SP), Accuracy (ACC).

Automated segmentation of blood vessels in retinal images based on entropy weighted thresholding

ABSTRACT Automated extraction of retinal vasculature is becoming a crucial task during the study of pathological disorders present in the retinal fundus image. In this work, we propose an unsupervised approach for automatically extracting blood vessels from fundus photographs. In this paper, Logarithmic transformation (Log) and contrast limited adaptive histogram equalisation (CLAHE) are being used to boost the contrast level of the fundus photographs. After that, Matched filtering technique is used for further vessel enhancement. Then, different thresholding techniques are evaluated and compared to find out the best thresholding scheme for this segmentation. We have found that Otsu with entropy weighting thresholding is the best threshold method for segmenting the blood vessels. In the post-processing step, the area-based thresholding process is considered, and it gives the final vessel segmented image. Using the following statistical measures, the final vessel segmented image is compared to the ground truth image, such as sensitivity (Sens), precision (Spec), and accuracy (Acc). This work is evaluated by freely accessible DRIVE (Digital Retinal Images for Vessel Extraction), STARE (Structured Analysis of the Retina), and CHASE_DB1 (Child Heart and Health Study in England) datasets. This work achieves an average accuracy of 0.9562, 0.946, and 0.946 on the DRIVE, STARE, and CHASE_DB1 datasets, respectively.

Brainstem Small Lymphocytic Lymphoma (SLL) Presenting as Serpiginous Small Vessel Enhancement and Peduncular Hallucinosis: A First Case Report (P15-9.005)

Brainstem Small Lymphocytic Lymphoma (SLL) Presenting as Serpiginous Small Vessel Enhancement and Peduncular Hallucinosis: A First Case Report (P15-9.005)

Retinal blood vessels segmentation using Wald PDF and MSMO operator

ABSTRACT Glaucoma, hypertension, obesity, diabetes, etc. are the widely spread severe diseases nowadays. Glaucoma, hypertension, and diabetes are the main cause of vision loss, whereas diabetes, obesity, and hypertension are the basis of several other highly dangerous diseases. Blood vessels of a retina contain information about these critical diseases. Health professionals use this information to detect and diagnose these diseases. Hence, it is essential to segment retinal blood vessels. Many researchers have proposed different matched filter methods to segment retinal blood vessels, but their matched filter kernels are not appropriate to vessel intensity profile due to which performances of these methods are relatively low. Also, the quality of retinal images directly affects the accuracy of segmentation. Therefore, retinal images must be of good quality. Existing methods for retinal image enhancement are less efficient. The proposed method introduces the Multiscale switching morphological operator (MSMO) with Wald probability distribution function-based new matched filter for better enhancement and segmentation of retinal vessels in this paper. The proposed method is tested on two known datasets: DRIVE and STARE. The average value of the precision, specificity, accuracy, MSE and MAD of the proposed method are 83.74%, 98.80%, 95.58%, 0.0442, 0.0366 for DRIVE and 71.24%, 97.67%, 94.89%, 0.0511, 0.015 for STARE dataset. Results demonstrate significant improvement in the performance of the proposed method over existing state-of-the-art methods. The causes of improved performance are enhanced retinal images due to using a better pre-processing method and close matching of Wald matched filter kernel with vessel intensity profile.

Ultrasound-guided needle positioning for nodal dynamic contrast-enhanced MR lymphangiography

The aim of the study was to assess injection needle positioning for contrast-enhanced MR-lymphangiography (MRL) by ultrasound-guided injection of saline-solution. 80 patients (33 male, mean age 43.1 years) were referred for MRL. The injection needle position was assessed by injection of saline-solution. Consecutive lymph node distension was observed on sonography followed by MRL. Transpedal MRL was performed when no inguinal lymph nodes could be identified. The inguinal lymph node detection rate was recorded. MR-lymphangiograms were assessed regarding primary (i.e. enhancement of draining lymph vessels) and secondary technical success (i.e. lymph vessel enhancement after repositioning of the needle). MRL was considered as clinically successful if enhancement of the central lymphatic system and/or a lymphatic pathologies were observed. For a total of 92 MRLs 177 groins were evaluated sonographically. In 171/177 groins (96.6%) lymph nodes were identified. After needle placement lymph node distension was observed in 171/171 cases (100%) on saline injection. MR-contrast injection demonstrated enhancement of draining lymph vessels in 163/171 cases (95.3%). In 6/171 cases (3.5%) in-bore needle retraction lead to lymphatic enhancement. In one patient [2/171 nodes (1.1%)] no lymphatic enhancement was seen despite repeated needle repositioning. Overall contrast application was technically successful in 169/171 cases (98.8%). In the 6 groins in which no nodes were identifiable, transpedal MRL was successful. So overall 91/92 MRLs (98.9%) were clinically successful. No complications were recorded. Confirmation of the needle position for nodal MRL by sonographically controlled saline injection is a reliable technique with a high success rate of MRL.

Biphasic split-bolus injection protocol for routine contrast-enhanced chest CT: comparison with conventional early-phase single bolus technique.

To present a routine contrast-enhanced chest CT protocol with a split-bolus injection technique achieving combined early- and delayed phase images with a single aquisition, and to compare this technique with a conventional early-phase single-bolus chest CT protocol we formerly used at our institution, in terms of attenuation of great thoracic vessels, pleura, included hepatic and portal venous enhancement, contrast-related artifacts, and image quality. A total of 202 patients, who underwent routine contrast-enhanced chest CT examination aquired with either conventional early-phase single-bolus technique (group A,n = 102) or biphasic split-bolus protocol (group B,n = 100), were retrospectively included. Attenuation measurements were made by two radiologists independently on mediastinal window settings using a circular ROI at the following sites: main pulmonary artery (PA) at its bifurcation level, thoracal aorta (TA) at the level of MPA bifurcation,portal vein (PV) at porta hepatis, left and right hepatic lobe, and if present, thickened pleura (>2 mm) at the level with the most intense enhancement. Respective normalized enhancement values were also calculated. Contrast-related artifacts were graded and qualitative evaluation of mediastinal lymph nodes was performed by both reviewers independently. Background noise was measured and contrast-to-noise ratios (CNRs) of the liver and TA were calculated. While enhancement of thoracic vessels and normalised MPA enhancement did not differ significantly between both groups (p > 0.05), enhancement and normalised enhancement of pleura, liver parenchyma and PV was significantly greater in group B (p < 0.001). Perivenous artifacts limiting evaluation were less frequent in group B than in A and mediastinal lymph nodes were judged to be evaluated worse in group A than in group B with an excellent agreement between both observers. No significant difference was detected in CNRTA (p = 0.633), whereas CNR liver was higher in group B (p < 0.001). Our split-bolus chest CT injection protocol enables simultaneous enhancement for both vascular structures and soft tissues, and thus, might raise diagnostic confidence without the need of multiple acquisitions. We think that this CT protocol might also be a promising alternative in lung cancer staging, where combined contrast-enhanced CT of the chest and abdomen is indicated. We therefore suggest to further evaluate its diagnostic utility in this setting, in particular in comparison with a late delayed chest-upper abdominal CT imaging protocol.

A Tensor-Based Catheter and Wire Detection and Tracking Framework and Its Clinical Applications.

Catheters and wires are used extensively in cardiac catheterization procedures. Detecting their positions in fluoroscopic X-ray images is important for several clinical applications such as motion compensation and co-registration between 2D and 3D imaging modalities. Detecting the complete length of a catheter or wire object as well as electrode positions on the catheter or wire is a challenging task. In this paper, an automatic detection framework for catheters and wires is developed. It is based on path reconstruction from image tensors, which are eigen direction vectors generated from a multiscale vessel enhancement filter. A catheter or a wire object is detected as the smooth path along those eigen direction vectors. Furthermore, a real-time tracking method based on a template generated from the detection method was developed. The proposed framework was tested on a total of 7,754 X-ray images. Detection errors for catheters and guidewires are 0.56 ± 0.28 mm and 0.68 ± 0.33 mm, respectively. The proposed framework was also tested and validated in two clinical applications. For motion compensation using catheter tracking, the 2D target registration errors (TRE) of 1.8 mm ± 0.9 mm was achieved. For co-registration between 2D X-ray images and 3D models from MRI images, a TRE of 2.3 ± 0.9 mm was achieved. A novel and fully automatic detection framework and its clinical applications are developed. The proposed framework can be applied to improve the accuracy of image-guidance systems for cardiac catheterization procedures.

High-strength deep learning image reconstruction in coronary CT angiography at 70-kVp tube voltage significantly improves image quality and reduces both radiation and contrast doses

To explore the use of 70-kVp tube voltage combined with high-strength deep learning image reconstruction (DLIR-H) in reducing radiation and contrast doses in coronary CT angiography (CCTA) in patients with body mass index (BMI) < 26kg/m2, in comparison with the conventional scan protocol using 120 kVp and adaptive statistical iterative reconstruction (ASIR-V). A total of 100 patients referred to CCTA were prospectively enrolled and randomly divided into two groups: low-dose group (n = 50) with 70 kVp, Smart mA for noise index (NI) of 36HU, contrast dose rate of 16mgI/kg/s, and DLIR-H, and conventional group (n = 50) with 120kV, Smart mA for NI of 25HU, contrast dose rate of 32mgI/kg/s, and 60%ASIR-V. Radiation and contrast dose, subjective image quality score, and objective image quality measurement (image noise, contrast-noise-ratio (CNR), and signal-noise-ratio (SNR) for vessel) were compared between the two groups. Low-dose group used significantly reduced contrast dose (23.82 ± 3.69mL, 50.6% reduction) and radiation dose (0.75 ± 0.14mSv, 54.5% reduction) compared to the conventional group (48.23 ± 6.38mL and 1.65 ± 0.66mSv, respectively) (all p < 0.001). Both groups had similar enhancement in vessels. However, the low-dose group had lower background noise (23.57 ± 4.74 HU vs. 35.04 ± 8.41 HU), higher CNR in RCA (48.63 ± 10.76 vs. 29.32 ± 5.52), LAD (47.33 ± 10.20 vs. 29.27 ± 5.12), and LCX (46.74 ± 9.76 vs. 28.58 ± 5.12) (all p < 0.001) compared to the conventional group. The use of 70-kVp tube voltage combined with DLIR-H for CCTA in normal size patients significantly reduces radiation dose and contrast dose while further improving image quality compared with the conventional 120-kVp tube voltage with 60%ASIR-V. •The combination of 70-kVp tube voltage and high-strength deep learning image reconstruction (DLIR-H) algorithm protocol reduces approximately 50% of radiation and contrast doses in coronary computed tomography angiography (CCTA) compared with the conventional scan protocol. •CCTA of normal size (BMI < 26kg/m2) patients acquired at sub-mSv radiation dose and 24mL contrast dose through the combination of 70-kVp tube voltage and DLIR-H algorithm achieves excellent diagnostic image quality with a good inter-rater agreement. •DLIR-H algorithm shows a higher capacity of significantly reducing image noise than adaptive statistical iterative reconstruction algorithm in CCTA examination.

Robust deep 3-D architectures based on vascular patterns for liver vessel segmentation

Deep Learning is providing new solutions for medical image segmentation problems and becomes a key point for future clinical application. The study of vascular structures is a challenging task due to the extremely small size of the vessel structure, low SNR, and varying contrast in medical image data.In this study, we present an end-to-end deep learning segmentation method relying on the integration of vessel enhancement filters inside a 3-D U-Net based architecture. In particular, the raw data used in the learning process -or used as input- is preprocessed using these filters. The use of vesselness filters can significantly improve the contrast in raw images; this step can also help to improve segmentation decision especially on bifurcations. 3-D U-Net, Dense U-Net and MultiRes U-Net are pitted against each other in the vessel segmentation task with the public IRCAD dataset. Considering the integration of vesselness filters, the model parameters were optimized in order to identify the optimal configuration for fully automatic segmentation of hepatic vessels. In addition the three architectures were tested on full 3-D images and slabs (stacks of 2-D slices). The results showed that the most accurate setup is the full 3-D process, which provides the highest Dice for most of the considered models. The 3-D Dense U-Net also gives the best results compared to other models with or without vessel enhancement filters.