Mesh neural networks for SE(3)-equivariant hemodynamics estimation on the artery wall.

Computational fluid dynamics (CFD) is a valuable asset for patient-specific cardiovascular-disease diagnosis and prognosis, but its high computational demands hamper its adoption in practice. Machine-learning methods that estimate blood flow in individual patients could accelerate or replace CFD simulation to overcome these limitations. In this work, we consider the estimation of vector-valued quantities on the wall of three-dimensional geometric artery models. We employ group-equivariant graph convolution in an end-to-end SE(3)-equivariant neural network that operates directly on triangular surface meshes and makes efficient use of training data. We run experiments on a large dataset of synthetic coronary arteries and find that our method estimates directional wall shear stress (WSS) with an approximation error of 7.6% and normalised mean absolute error (NMAE) of 0.4% while up to two orders of magnitude faster than CFD. Furthermore, we show that our method is powerful enough to accurately predict transient, vector-valued WSS over the cardiac cycle while conditioned on a range of different inflow boundary conditions. These results demonstrate the potential of our proposed method as a plugin replacement for CFD in the personalised prediction of hemodynamic vector and scalar fields.

Automated multiclass segmentation, quantification, and visualization of the diseased aorta on hybrid PET/CT-SEQUOIA.

BACKGROUND: Cardiovascular disease is the most common cause of death worldwide, including infection and inflammation related conditions. Multiple studies have demonstrated potential advantages of hybrid positron emission tomography combined with computed tomography (PET/CT) as an adjunct to current clinical inflammatory and infectious biochemical markers. To quantitatively analyze vascular diseases at PET/CT, robust segmentation of the aorta is necessary. However, manual segmentation is extremely time-consuming and labor-intensive. PURPOSE: To investigate the feasibility and accuracy of an automated tool to segment and quantify multiple parts of the diseased aorta on unenhanced low-dose computed tomography (LDCT) as an anatomical reference for PET-assessed vascular disease. METHODS: A software pipeline was developed including automated segmentation using a 3D U-Net, calcium scoring, PET uptake quantification, background measurement, radiomics feature extraction, and 2D surface visualization of vessel wall calcium and tracer uptake distribution. To train the 3D U-Net, 352 non-contrast LDCTs from (2-[18 F]FDG and Na[18 F]F) PET/CTs performed in patients with various vascular pathologies with manual segmentation of the ascending aorta, aortic arch, descending aorta, and abdominal aorta were used. The last 22 consecutive scans were used as a hold-out internal test set. The remaining dataset was randomly split into training (n = 264; 80%) and validation (n = 66; 20%) sets. Further evaluation was performed on an external test set of 49 PET/CTs. The dice similarity coefficient (DSC) and Hausdorff distance (HD) were used to assess segmentation performance. Automatically obtained calcium scores and uptake values were compared with manual scoring obtained using clinical softwares (syngo.via and Affinity Viewer) in six patient images. intraclass correlation coefficients (ICC) were calculated to validate calcium and uptake values. RESULTS: Fully automated segmentation of the aorta using a 3D U-Net was feasible in LDCT obtained from PET/CT scans. The external test set yielded a DSC of 0.867 ± 0.030 and HD of 1.0 [0.6-1.4] mm, similar to an open-source model with a DSC of 0.864 ± 0.023 and HD of 1.4 [1.0-1.8] mm. Quantification of calcium and uptake values were in excellent agreement with clinical software (ICC: 1.00 [1.00-1.00] and 0.99 [0.93-1.00] for calcium and uptake values, respectively). CONCLUSIONS: We present an automated pipeline to segment the ascending aorta, aortic arch, descending aorta, and abdominal aorta on LDCT from PET/CT and to accurately provide uptake values, calcium scores, background measurement, radiomics features, and a 2D visualization. We call this algorithm SEQUOIA (SEgmentation, QUantification, and visualizatiOn of the dIseased Aorta) and is available at https://github.com/UMCG-CVI/SEQUOIA. This model could augment the utility of aortic evaluation at PET/CT studies tremendously, irrespective of the tracer, and potentially provide fast and reliable quantification of cardiovascular diseases in clinical practice, both for primary diagnosis and disease monitoring.

Quantitative analysis of aortic Na[18F]F uptake in macrocalcifications and microcalcifications in PET/CT scans.

BACKGROUND: Currently, computed tomography (CT) is used for risk profiling of (asymptomatic) individuals by calculating coronary artery calcium scores. Although this score is a strong predictor of major adverse cardiovascular events, this method has limitations. Sodium [18F]fluoride (Na[18F]F) positron emission tomography (PET) has shown promise as an early marker for atherosclerotic progression. However, evidence on Na[18F]F as a marker for high-risk plaques is limited, particularly on its presentation in clinical PET/CT. Besides, the relationship between microcalcifications visualized by Na[18F]F PET and macrocalcifications detectable on CT is unknown. PURPOSE: To establish a match/mismatch score in the aorta between macrocalcified plaque content on CT and microcalcification Na[18F]F PET uptake. METHODS: Na[18F]F-PET/CT scans acquired in our centre in 2019-2020 were retrospectively collected. The aorta of each low-dose CT was manually segmented. Background measurements were placed in the superior vena cava. The vertebrae were automatically segmented using an open-source convolutional neural network, dilated with 10 mm, and subtracted from the aortic mask. Per patient, calcium and Na[18F]F-hotspot masks were retrieved using an in-house developed algorithm. Three match/mismatch analyses were performed: a population analysis, a per slice analysis, and an overlap score. To generate a population image of calcium and Na[18F]F hotspot distribution, all aortic masks were aligned. Then, a heatmap of calcium HU and Na[18F]F-uptake on the surface was obtained by outward projection of HU and uptake values from the centerline. In each slice of the aortic wall of each patient, the calcium mass score and target-to-bloodpool ratios (TBR) were calculated within the calcium masks, in the aortic wall except the calcium masks, and in the aortic wall in slices without calcium. For the overlap score, three volumes were identified in the calcium and Na[18F]F masks: volume of PET (PET+/CT-), volume of CT (PET-/CT+), and overlapping volumes (PET+/CT+). A Spearman's correlation analysis with Bonferroni correction was performed on the population image, assessing the correlation between all HU and Na[18F]F vertex values. In the per slice analysis, a paired Wilcoxon signed-rank test was used to compare TBR values within each slice, while an ANOVA with post-hoc Kruskal-Wallis test was employed to compare TBR values between slices. p-values < 0.05 were considered significant. RESULTS: In total, 186 Na[18F]F-PET/CT scans were included. A moderate positive exponential correlation was observed between total aortic calcium mass and total aortic TBR (r = 0.68, p < 0.001). A strong positive correlation (r = 0.77, p < 0.0001) was observed between CT values and Na[18F]F values on the population image. Significantly higher TBR values were found outside calcium masks than inside calcium masks (p < 0.0001). TBR values in slices where no calcium was present, were significantly lower compared with outside calcium and inside calcium (both p < 0.0001). On average, only 3.7% of the mask volumes were overlapping. CONCLUSIONS: Na[18F]F-uptake in the aorta behaves similarly to macrocalcification detectable on CT. Na[18F]F-uptake values are also moderately correlated to calcium mass scores (match). Higher uptake values were found just outside macrocalcification masks instead of inside the macrocalcification masks (mismatch). Also, only a small percentage of the Na[18F]F-uptake volumes overlapped with the calcium volumes (mismatch).

Generative modeling of living cells with SO(3)-equivariant implicit neural representations.

Data-driven cell tracking and segmentation methods in biomedical imaging require diverse and information-rich training data. In cases where the number of training samples is limited, synthetic computer-generated data sets can be used to improve these methods. This requires the synthesis of cell shapes as well as corresponding microscopy images using generative models. To synthesize realistic living cell shapes, the shape representation used by the generative model should be able to accurately represent fine details and changes in topology, which are common in cells. These requirements are not met by 3D voxel masks, which are restricted in resolution, and polygon meshes, which do not easily model processes like cell growth and mitosis. In this work, we propose to represent living cell shapes as level sets of signed distance functions (SDFs) which are estimated by neural networks. We optimize a fully-connected neural network to provide an implicit representation of the SDF value at any point in a 3D+time domain, conditioned on a learned latent code that is disentangled from the rotation of the cell shape. We demonstrate the effectiveness of this approach on cells that exhibit rapid deformations (Platynereis dumerilii), cells that grow and divide (C. elegans), and cells that have growing and branching filopodial protrusions (A549 human lung carcinoma cells). A quantitative evaluation using shape features and Dice similarity coefficients of real and synthetic cell shapes shows that our model can generate topologically plausible complex cell shapes in 3D+time with high similarity to real living cell shapes. Finally, we show how microscopy images of living cells that correspond to our generated cell shapes can be synthesized using an image-to-image model.

Recent Developments in Speeding up Prostate MRI.

The increasing incidence of prostate cancer cases worldwide has led to a tremendous demand for multiparametric MRI (mpMRI). In order to relieve the pressure on healthcare, reducing mpMRI scan time is necessary. This review focuses on recent techniques proposed for faster mpMRI acquisition, specifically shortening T2W and DWI sequences while adhering to the PI-RADS (Prostate Imaging Reporting and Data System) guidelines. Speeding up techniques in the reviewed studies rely on more efficient sampling of data, ranging from the acquisition of fewer averages or b-values to adjustment of the pulse sequence. Novel acquisition methods based on undersampling techniques are often followed by suitable reconstruction methods typically incorporating synthetic priori information. These reconstruction methods often use artificial intelligence for various tasks such as denoising, artifact correction, improvement of image quality, and in the case of DWI, for the generation of synthetic high b-value images or apparent diffusion coefficient maps. Reduction of mpMRI scan time is possible, but it is crucial to maintain diagnostic quality, confirmed through radiological evaluation, to integrate the proposed methods into the standard mpMRI protocol. Additionally, before clinical integration, prospective studies are recommended to validate undersampling techniques to avoid potentially inaccurate results demonstrated by retrospective analysis. This review provides an overview of recently proposed techniques, discussing their implementation, advantages, disadvantages, and diagnostic performance according to PI-RADS guidelines compared to conventional methods. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 3.

Development of a Statistical Shape Model and Assessment of Anatomical Shape Variations in the Hemipelvis.

Knowledge about anatomical shape variations in the pelvis is mandatory for selection, fitting, positioning, and fixation in pelvic surgery. The current knowledge on pelvic shape variation mostly relies on point-to-point measurements on 2D X-ray images and computed tomography (CT) slices. Three-dimensional region-specific assessments of pelvic morphology are scarce. Our aim was to develop a statistical shape model of the hemipelvis to assess anatomical shape variations in the hemipelvis. CT scans of 200 patients (100 male and 100 female) were used to obtain segmentations. An iterative closest point algorithm was performed to register these 3D segmentations, so a principal component analysis (PCA) could be performed, and a statistical shape model (SSM) of the hemipelvis was developed. The first 15 principal components (PCs) described 90% of the total shape variation, and the reconstruction ability of this SSM resulted in a root mean square error of 1.58 (95% CI: 1.53-1.63) mm. In summary, an SSM of the hemipelvis was developed, which describes the shape variations in a Caucasian population and is able to reconstruct an aberrant hemipelvis. Principal component analyses demonstrated that, in a general population, anatomical shape variations were mostly related to differences in the size of the pelvis (e.g., PC1 describes 68% of the total shape variation, which is attributed to size). Differences between the male and female pelvis were most pronounced in the iliac wing and pubic rami regions. These regions are often subject to injuries. Future clinical applications of our newly developed SSM may be relevant for SSM-based semi-automatic virtual reconstruction of a fractured hemipelvis as part of preoperative planning. Lastly, for companies, using our SSM might be interesting in order to assess which sizes of pelvic implants should be produced to provide proper-fitting implants for most of the population.

A Statistical Shape Model of Infrarenal Aortic Necks in Patients With and Without Late Type Ia Endoleak After Endovascular Aneurysm Repair.

PURPOSE: Hostile aortic neck characteristics, including short length, severe suprarenal and infrarenal angulation, conicity, and large diameter, have been associated with increased risk for type Ia endoleak (T1aEL) after endovascular aneurysm repair (EVAR). This study investigates the mid-term discriminative ability of a statistical shape model (SSM) of the infrarenal aortic neck morphology compared with or in combination with conventional measurements in patients who developed T1aEL post-EVAR. MATERIALS AND METHODS: The dataset composed of EVAR patients who developed a T1aEL during follow-up and a control group without T1aEL. Principal component (PC) analysis was performed using a parametrization to create an SSM. Three logistic regression models were created. To discriminate between patients with and without T1aEL, sensitivity, specificity, and the area under the receiver operating characteristic (ROC) curve (AUC) were calculated. RESULTS: In total, 126 patients (84% male) were included. Median follow-up time in T1aEl group and control group was 52 (31, 78.5) and 51 (40, 62.5) months, respectively. Median follow-up time was not statistically different between the groups (p=0.72). A statistically significant difference between the median PC scores of the T1aEL and control groups was found for the first, eighth, and ninth PC. Sensitivity, specificity, and AUC values for the SSM-based versus the conventional measurements-based logistic regression models were 79%, 70%, and 0.82 versus 74%, 73%, and 0.85, respectively. The model of the SSM and conventional measurements combined resulted in sensitivity, specificity, and AUC of 81%, 81%, and 0.92. CONCLUSION: An SSM of the infrarenal aortic neck determines its 3-dimensional geometry. The SSM is a potential valuable tool for risk stratification and T1aEL prediction in EVAR. The SSM complements the conventional measurements of the individual preoperative infrarenal aortic neck geometry by increasing the predictive value for late type Ia endoleak after standard EVAR. CLINICAL IMPACT: A statistical shape model (SSM) determines the 3-dimensional geometry of the infrarenal aortic neck. The SSM complements the conventional measurements of the individual pre-operative infrarenal aortic neck geometry by increasing the predictive value for late type Ia endoleaks post-EVAR. The SSM is a potential valuable tool for risk stratification and late T1aEL prediction in EVAR and it is a first step toward implementation of a treatment planning support tool in daily clinical practice.

Deep Learning-Based Intraoperative Stent Graft Segmentation on Completion Digital Subtraction Angiography During Endovascular Aneurysm Repair.

PURPOSE: Modern endovascular hybrid operating rooms generate large amounts of medical images during a procedure, which are currently mostly assessed by eye. In this paper, we present fully automatic segmentation of the stent graft on the completion digital subtraction angiography during endovascular aneurysm repair, utilizing a deep learning network. TECHNIQUE: Completion digital subtraction angiographies (cDSAs) of 47 patients treated for an infrarenal aortic aneurysm using EVAR were collected retrospectively. A two-dimensional convolutional neural network (CNN) with a U-Net architecture was trained for segmentation of the stent graft from the completion angiographies. The cross-validation resulted in an average Dice similarity score of 0.957 ± 0.041 and median of 0.968 (IQR: 0.950 - 0.976). The mean and median of the average surface distance are 1.266 ± 1.506 mm and 0.870 mm (IQR: 0.490 - 1.430), respectively. CONCLUSION: We developed a fully automatic stent graft segmentation method based on the completion digital subtraction angiography during EVAR, utilizing a deep learning network. This can provide the platform for the development of intraoperative analytical applications in the endovascular hybrid operating room such as stent graft deployment accuracy, endoleak visualization, and image fusion correction.

Super-Resolved Microbubble Localization in Single-Channel Ultrasound RF Signals Using Deep Learning.

Recently, super-resolution ultrasound imaging with ultrasound localization microscopy (ULM) has received much attention. However, ULM relies on low concentrations of microbubbles in the blood vessels, ultimately resulting in long acquisition times. Here, we present an alternative super-resolution approach, based on direct deconvolution of single-channel ultrasound radio-frequency (RF) signals with a one-dimensional dilated convolutional neural network (CNN). This work focuses on low-frequency ultrasound (1.7 MHz) for deep imaging (10 cm) of a dense cloud of monodisperse microbubbles (up to 1000 microbubbles in the measurement volume, corresponding to an average echo overlap of 94%). Data are generated with a simulator that uses a large range of acoustic pressures (5-250 kPa) and captures the full, nonlinear response of resonant, lipid-coated microbubbles. The network is trained with a novel dual-loss function, which features elements of both a classification loss and a regression loss and improves the detection-localization characteristics of the output. Whereas imposing a localization tolerance of 0 yields poor detection metrics, imposing a localization tolerance corresponding to 4% of the wavelength yields a precision and recall of both 0.90. Furthermore, the detection improves with increasing acoustic pressure and deteriorates with increasing microbubble density. The potential of the presented approach to super-resolution ultrasound imaging is demonstrated with a delay-and-sum reconstruction with deconvolved element data. The resulting image shows an order-of-magnitude gain in axial resolution compared to a delay-and-sum reconstruction with unprocessed element data.

A Statistical Shape Model of the Morphological Variation of the Infrarenal Abdominal Aortic Aneurysm Neck.

Hostile aortic neck characteristics, such as short length and large diameter, have been associated with type Ia endoleaks and reintervention after endovascular aneurysm repair (EVAR). However, such characteristics partially describe the complex aortic neck morphology. A more comprehensive quantitative description of 3D neck shape might lead to new insights into the relationship between aortic neck morphology and EVAR outcomes in individual patients. This study identifies the 3D morphological shape components that describe the infrarenal aortic neck through a statistical shape model (SSM). Pre-EVAR CT scans of 97 patients were used to develop the SSM. Parameterization of the morphology was based on the center lumen line reconstruction, a triangular surface mesh of the aortic lumen, 3D coordinates of the renal arteries, and the distal end of the aortic neck. A principal component analysis of the parametrization of the aortic neck coordinates was used as input for the SSM. The SSM consisted of 96 principal components (PCs) that each described a unique shape feature. The first five PCs represented 95% of the total morphological variation in the dataset. The SSM is an objective model that provides a quantitative description of the neck morphology of an individual patient.

Deep learning-based whole-heart segmentation in 4D contrast-enhanced cardiac CT.

Automatic cardiac chamber and left ventricular (LV) myocardium segmentation over the cardiac cycle significantly extends the utilization of contrast-enhanced cardiac CT, potentially enabling in-depth assessment of cardiac function. Therefore, we evaluate an automatic method for cardiac chamber and LV myocardium segmentation in 4D cardiac CT. In this study, 4D contrast-enhanced cardiac CT scans of 1509 patients selected for transcatheter aortic valve implantation with 21,605 3D images, were divided into development (N = 12) and test set (N = 1497). 3D convolutional neural networks were trained with end-systolic (ES) and end-diastolic (ED) images. Dice similarity coefficient (DSC) and average symmetric surface distance (ASSD) were computed for 3D segmentations at ES and ED in the development set via cross-validation, and for 2D segmentations in four cardiac phases for 81 test set patients. Segmentation quality in the full test set of 1497 patients was assessed visually on a three-point scale per structure based on estimated overlap with the ground truth. Automatic segmentation resulted in a mean DSC of 0.89 ± 0.10 and ASSD of 1.43 ± 1.45 mm in 12 patients in 3D, and a DSC of 0.89 ± 0.08 and ASSD of 1.86 ± 1.20 mm in 81 patients in 2D. The qualitative evaluation in the whole test set of 1497 patients showed that automatic segmentations were assigned grade 1 (clinically useful) in 98.5%, 92.2%, 83.1%, 96.3%, and 91.6% of cases for LV cavity and myocardium, right ventricle, left atrium, and right atrium. Our automatic method using convolutional neural networks performed clinically useful segmentation across the cardiac cycle in a large set of 4D cardiac CT images, potentially enabling in-depth assessment of cardiac function.

AI-Based Quantification of Planned Radiation Therapy Dose to Cardiac Structures and Coronary Arteries in Patients With Breast Cancer.

PURPOSE: The purpose of this work is to develop and evaluate an automatic deep learning method for segmentation of cardiac chambers and large arteries, and localization of the 3 main coronary arteries in radiation therapy planning on computed tomography (CT). In addition, a second purpose is to determine the planned radiation therapy dose to cardiac structures for breast cancer therapy. METHODS AND MATERIALS: Eighteen contrast-enhanced cardiac scans acquired with a dual-layer-detector CT scanner were included for method development. Manual reference annotations of cardiac chambers, large arteries, and coronary artery locations were made in the contrast scans and transferred to virtual noncontrast images, mimicking noncontrast-enhanced CT. In addition, 31 noncontrast-enhanced radiation therapy treatment planning CTs with corresponding dose-distribution maps of breast cancer cases were included for evaluation. For reference, cardiac chambers and large vessels were manually annotated in two 2-dimensional (2D) slices per scan (26 scans, totaling 52 slices) and in 3-dimensional (3D) scan volumes in 5 scans. Coronary artery locations were annotated on 3D imaging. The method uses an ensemble of convolutional neural networks with 2 output branches that perform 2 distinct tasks: (1) segmentation of the cardiac chambers and large arteries and (2) localization of coronary arteries. Training was performed using reference annotations and virtual noncontrast cardiac scans. Automatic segmentation of the cardiac chambers and large vessels and the coronary artery locations was evaluated in radiation therapy planning CT with Dice score (DSC) and average symmetrical surface distance (ASSD). The correlation between dosimetric parameters derived from the automatic and reference segmentations was evaluated with R2. RESULTS: For cardiac chambers and large arteries, median DSC was 0.76 to 0.88, and the median ASSD was 0.17 to 0.27 cm in 2D slice evaluation. 3D evaluation found a DSC of 0.87 to 0.93 and an ASSD of 0.07 to 0.10 cm. Median DSC of the coronary artery locations ranged from 0.80 to 0.91. R2 values of dosimetric parameters were 0.77 to 1.00 for the cardiac chambers and large vessels, and 0.76 to 0.95 for the coronary arteries. CONCLUSIONS: The developed and evaluated method can automatically obtain accurate estimates of planned radiation dose and dosimetric parameters for the cardiac chambers, large arteries, and coronary arteries.

Anatomy-aided deep learning for medical image segmentation: a review.

Deep learning (DL) has become widely used for medical image segmentation in recent years. However, despite these advances, there are still problems for which DL-based segmentation fails. Recently, some DL approaches had a breakthrough by using anatomical information which is the crucial cue for manual segmentation. In this paper, we provide a review of anatomy-aided DL for medical image segmentation which covers systematically summarized anatomical information categories and corresponding representation methods. We address known and potentially solvable challenges in anatomy-aided DL and present a categorized methodology overview on using anatomical information with DL from over 70 papers. Finally, we discuss the strengths and limitations of the current anatomy-aided DL approaches and suggest potential future work.

Generative Adversarial Networks: A Primer for Radiologists.

Artificial intelligence techniques involving the use of artificial neural networks-that is, deep learning techniques-are expected to have a major effect on radiology. Some of the most exciting applications of deep learning in radiology make use of generative adversarial networks (GANs). GANs consist of two artificial neural networks that are jointly optimized but with opposing goals. One neural network, the generator, aims to synthesize images that cannot be distinguished from real images. The second neural network, the discriminator, aims to distinguish these synthetic images from real images. These deep learning models allow, among other applications, the synthesis of new images, acceleration of image acquisitions, reduction of imaging artifacts, efficient and accurate conversion between medical images acquired with different modalities, and identification of abnormalities depicted on images. The authors provide an introduction to GANs and adversarial deep learning methods. In addition, the different ways in which GANs can be used for image synthesis and image-to-image translation tasks, as well as the principles underlying conditional GANs and cycle-consistent GANs, are described. Illustrated examples of GAN applications in radiologic image analysis for different imaging modalities and different tasks are provided. The clinical potential of GANs, future clinical GAN applications, and potential pitfalls and caveats that radiologists should be aware of also are discussed in this review. The online slide presentation from the RSNA Annual Meeting is available for this article. ©RSNA, 2021.

Vascular uptake on 18F-sodium fluoride positron emission tomography: precursor of vascular calcification?

BACKGROUND: Microcalcifications cannot be identified with the present resolution of CT; however, 18F-sodium fluoride (18F-NaF) positron emission tomography (PET) imaging has been proposed for non-invasive identification of microcalcification. The primary objective of this study was to assess whether 18F-NaF activity can assess the presence and predict the progression of CT detectable vascular calcification. METHODS AND RESULTS: The data of two longitudinal studies in which patients received a 18F-NaF PET-CT at baseline and after 6 months or 1-year follow-up were used. The target to background ratio (TBR) was measured on PET at baseline and CT calcification was quantified in the femoral arteries at baseline and follow-up. 128 patients were included. A higher TBR at baseline was associated with higher calcification mass at baseline and calcification progression (ß = 1.006 [1.005-1.007] and ß = 1.002 [1.002-1.003] in the studies with 6 months and 1-year follow-up, respectively). In areas without calcification at baseline and where calcification developed at follow-up, the TBR was .11-.13 (P < .001) higher compared to areas where no calcification developed. CONCLUSION: The activity of 18F-NaF is related to the amount of calcification and calcification progression. In areas where calcification formation occurred, the TBR was slightly but significantly higher.

Six months vitamin K treatment does not affect systemic arterial calcification or bone mineral density in diabetes mellitus 2.

PURPOSE: Vitamin K-dependent proteins are involved in (patho)physiological calcification of the vasculature and the bones. Type 2 diabetes mellitus (DM2) is associated with increased arterial calcification and increased fractures. This study investigates the effect of 6 months vitamin K2 supplementation on systemic arterial calcification and bone mineral density (BMD) in DM2 patients with a history of cardiovascular disease (CVD). METHODS: In this pre-specified, post hoc analysis of a double-blind, randomized, controlled clinical trial, patients with DM2 and CVD were randomized to a daily, oral dose of 360 µg vitamin K2 or placebo for 6 months. CT scans were made at baseline and follow-up. Arterial calcification mass was quantified in several large arterial beds and a total arterial calcification mass score was calculated. BMD was assessed in all non-fractured thoracic and lumbar vertebrae. RESULTS: 68 participants were randomized, 35 to vitamin K2 (33 completed follow-up) and 33 to placebo (27 completed follow-up). The vitamin K group had higher arterial calcification mass at baseline [median (IQR): 1694 (812-3584) vs 1182 (235-2445)] for the total arterial calcification mass). Six months vitamin K supplementation did not reduce arterial calcification progression (ß [95% CI]: - 0.02 [- 0.10; 0.06] for the total arterial calcification mass) or slow BMD decline (ß [95% CI]: - 2.06 [- 11.26; 7.30] Hounsfield units for all vertebrae) when compared to placebo. CONCLUSION: Six months vitamin K supplementation did not halt progression of arterial calcification or decline of BMD in patients with DM2 and CVD. Future clinical trials may want to pre-select patients with very low vitamin K status and longer follow-up time might be warranted. This trial was registered at clinicaltrials.gov as NCT02839044.

Determinants of 18F-NaF uptake in femoral arteries in patients with type 2 diabetes mellitus.

BACKGROUND: The goal of this study was to investigate the potential determinants of 18F-NaF uptake in femoral arteries as a marker of arterial calcification in patients with type 2 diabetes and a history of arterial disease. METHODS AND RESULTS: The study consisted of participants of a randomized controlled trial to investigate the effect of vitamin K2 (NCT02839044). In this prespecified analysis, subjects with type 2 diabetes and known arterial disease underwent full body 18F-NaF PET/CT. Target-to-background ratio (TBR) was calculated by dividing the mean SUVmax from both superficial femoral arteries by the SUVmean in the superior vena cava (SVC) and calcium mass was measured on CT. The association between 18F-NaF TBR and cardiovascular risk factors was investigated using uni- and multivariate linear regression corrected for age and sex. In total, 68 patients (mean age: 69 ± 8 years; male: 52) underwent 18F-NaF PET/CT. Higher CT calcium mass, total cholesterol, and HbA1c were associated with higher 18F-NaF TBR after adjusting. CONCLUSION: This study shows that several modifiable cardiovascular risk factors (total cholesterol, triglycerides, HbA1c) are associated with femoral 18F-NaF tracer uptake in patients with type 2 diabetes.

Deep Learning-Based Regression and Classification for Automatic Landmark Localization in Medical Images.

In this study, we propose a fast and accurate method to automatically localize anatomical landmarks in medical images. We employ a global-to-local localization approach using fully convolutional neural networks (FCNNs). First, a global FCNN localizes multiple landmarks through the analysis of image patches, performing regression and classification simultaneously. In regression, displacement vectors pointing from the center of image patches towards landmark locations are determined. In classification, presence of landmarks of interest in the patch is established. Global landmark locations are obtained by averaging the predicted displacement vectors, where the contribution of each displacement vector is weighted by the posterior classification probability of the patch that it is pointing from. Subsequently, for each landmark localized with global localization, local analysis is performed. Specialized FCNNs refine the global landmark locations by analyzing local sub-images in a similar manner, i.e. by performing regression and classification simultaneously and combining the results. Evaluation was performed through localization of 8 anatomical landmarks in CCTA scans, 2 landmarks in olfactory MR scans, and 19 landmarks in cephalometric X-rays. We demonstrate that the method performs similarly to a second observer and is able to localize landmarks in a diverse set of medical images, differing in image modality, image dimensionality, and anatomical coverage.

Deep learning from dual-energy information for whole-heart segmentation in dual-energy and single-energy non-contrast-enhanced cardiac CT.

PURPOSE: Deep learning-based whole-heart segmentation in coronary computed tomography angiography (CCTA) allows the extraction of quantitative imaging measures for cardiovascular risk prediction. Automatic extraction of these measures in patients undergoing only non-contrast-enhanced CT (NCCT) scanning would be valuable, but defining a manual reference standard that would allow training a deep learning-based method for whole-heart segmentation in NCCT is challenging, if not impossible. In this work, we leverage dual-energy information provided by a dual-layer detector CT scanner to obtain a reference standard in virtual non-contrast (VNC) CT images mimicking NCCT images, and train a three-dimensional (3D) convolutional neural network (CNN) for the segmentation of VNC as well as NCCT images. METHODS: Eighteen patients were scanned with and without contrast enhancement on a dual-layer detector CT scanner. Contrast-enhanced acquisitions were reconstructed into a CCTA and a perfectly aligned VNC image. In each CCTA image, manual reference segmentations of the left ventricular (LV) myocardium, LV cavity, right ventricle, left atrium, right atrium, ascending aorta, and pulmonary artery trunk were obtained and propagated to the corresponding VNC image. These VNC images and reference segmentations were used to train 3D CNNs in a sixfold cross-validation for automatic segmentation in either VNC images or NCCT images reconstructed from the non-contrast-enhanced acquisition. Automatic segmentation in VNC images was evaluated using the Dice similarity coefficient (DSC) and average symmetric surface distance (ASSD). Automatically determined volumes of the cardiac chambers and LV myocardium in NCCT were compared to reference volumes of the same patient in CCTA by Bland-Altman analysis. An additional independent multivendor multicenter set of single-energy NCCT images from 290 patients was used for qualitative analysis, in which two observers graded segmentations on a five-point scale. RESULTS: Automatic segmentations in VNC images showed good agreement with reference segmentations, with an average DSC of 0.897 ± 0.034 and an average ASSD of 1.42 ± 0.45 mm. Volume differences [95% confidence interval] between automatic NCCT and reference CCTA segmentations were -19 [-67; 30] mL for LV myocardium, -25 [-78; 29] mL for LV cavity, -29 [-73; 14] mL for right ventricle, -20 [-62; 21] mL for left atrium, and -19 [-73; 34] mL for right atrium, respectively. In 214 (74%) NCCT images from the independent multivendor multicenter set, both observers agreed that the automatic segmentation was mostly accurate (grade 3) or better. CONCLUSION: Our automatic method produced accurate whole-heart segmentations in NCCT images using a CNN trained with VNC images from a dual-layer detector CT scanner. This method might enable quantification of additional cardiac measures from NCCT images for improved cardiovascular risk prediction.