Automatic thoracic aorta calcium quantification using deep learning in non-contrast ECG-gated CT images

Thoracic aorta calcium (TAC) can be assessed from cardiac computed tomography (CT) studies to improve cardiovascular risk prediction. The aim of this study was to develop a fully automatic system to detect TAC and to evaluate its performance for classifying the patients into four TAC risk categories. The method started by segmenting the thoracic aorta, combining three UNets trained with axial, sagittal and coronal CT images. Afterwards, the surrounding lesion candidates were classified using three combined convolutional neural networks (CNNs) trained with orthogonal patches. Image datasets included 1190 non-enhanced ECG-gated cardiac CT studies from a cohort of cardiovascular patients (age 57 ± 9 years, 80% men, 65% TAC > 0). In the test set (N = 119), the combination of UNets was able to successfully segment the thoracic aorta with a mean volume difference of 0.3 ± 11.7 ml (<6%) and a median Dice coefficient of 0.947. The combined CNNs accurately classified the lesion candidates and 87% of the patients (N = 104) were accurately placed in their corresponding risk categories (Kappa = 0.826, ICC = 0.9915). TAC measurement can be estimated automatically from cardiac CT images using UNets to isolate the thoracic aorta and CNNs to classify calcified lesions.

Abdominal aortic calcium and geometry in patients with essential hypertension

PURPOSE

Abdominal aorta calcium (AAC) burden and dilatation are associated with an increased risk of mortality. The purpose of this study was to investigate determinants of AAC and abdominal aorta size in patients with essential hypertension.

MATERIALS AND METHODS

Patients with uncomplicated essential hypertension who had undergone non-enhanced abdominal CT to rule out secondary hypertension in addition to biological test were recruited between 2010 and 2018. A semi-automatic system was designed to estimate the aortic size (diameter, length, volume) and quantify the AAC from mesenteric artery to bifurcation using the Agatston score. Determinants of aortic size and those related to AAC were searched for using uni- and multivariables analyses.

RESULTS

Among 293 randomly selected patients with hypertension (age 52 ± 11 [SD] years) included, 23% had resistant hypertension. Mean abdominal aorta diameter was 20.1 ± 2.1 (SD) mm. Eight (3%) patients had abdominal aorta aneurysm ≥ 30 mm and 58 (20%) had dilated abdominal aorta ≥ 27 mm. Median AAC score was 38 and calcifications were detected in the infra- and supra-renal abdominal aortic portions in 59% and 26% of the patients, respectively. After adjustment for age, male sex and body surface area, abdominal aorta diameter was positively associated with diastolic blood pressure (P = 0.0019). Smoking was the single variable associated with calcified abdominal aorta (P < 0.001) after adjustment for cofactors. In patients with calcifications of abdominal aorta, the score increased with smoking history (P < 0.001), statins treatment (P < 0.01), greater number of anti-hypertensive drugs (P < 0.01), larger abdominal aorta (P < 0.05) and greater systolic blood pressure (P < 0.05). Patients with resistant hypertension had more AAC in the supra-renal abdominal aorta portion than those without resistant hypertension (P < 0.01).

CONCLUSION

In patients with essential hypertension, abdominal aorta dilation is related with diastolic blood pressure while AAC is associated with smoking history and resistant hypertension when located to the supra-renal abdominal aorta portion.

Echocardiographic measurements of the left heart chambers size in a large cohort of subjects: comparison of body surface area and height indexation to account for effects of obesity

BACKGROUND

The assessment of cardiac chamber size in the obese population is a challenging subject. Values usually indexed to body surface area (BSA) are smaller in obese subjects and prone to overcorrection. This study aimed to find reference thresholds to account for effects of obesity from a large cohort of patients and to evaluate the indexation by height as an alternative to BSA.

METHODS

The last 10-years records of our Echocardiography Unit were retrospectively analyzed and 14,007 subjects without known cardiac disease were included (45±15 years, 54% women, 20% obese). Measurements included left atrial diameter, area and volume (LAD, LAA and LAV), left ventricular (LV) end-diastolic (LVEDD) and end-systolic diameters (LVESD), aortic root diameter (ARD) and LV mass (LVM). Absolute, BSA and height indexed maximum thresholds (mean+1.96.SD) were calculated. An allometric indexation of the form variable/Height^β was tested. Correlation coefficients between indexed and absolute values were calculated to evaluate their proportional association (ideally r=1). Correlations between indexed values and body size represented residual associations to be minimized (ideally r=0).

RESULTS

The strongest association of echocardiographic measurements with body size was observed for BSA (r=0.36-0.63), whereas the isometric and allometric height models showed lower comparable values (r=0.28-0.48). Positive correlations with BMI were mostly observed for LA size (r≈0.36) and LVM (r≈0.36) measurements. Scaling exponents β for the allometric height indexation were 1.72 for LAV and 2.33 for LVM. Correlations between indexed and absolute values were higher for height than BSA (0.80-0.98 vs 0.44-0.92). Correlations between indexed values and height were closer to zero than for BSA, particularly using the allometric model. The overcorrection observed with increasing obesity class after BSA indexation was avoided after height indexation.

CONCLUSIONS

Unlike BSA, height indexing provided an adequate body size scaling of left heart chambers size avoiding overcorrection using allometric models in particular.

Thoracic Aorta Calcium Detection and Quantification Using Convolutional Neural Networks in a Large Cohort of Intermediate-Risk Patients

Arterial calcification is an independent predictor of cardiovascular disease (CVD) events whereas thoracic aorta calcium (TAC) detection might anticipate extracoronary outcomes. In this work, we trained six convolutional neural networks (CNNs) to detect aortic calcifications and to automate the TAC score assessment in intermediate CVD risk patients. Cardiac computed tomography images from 1415 patients were analyzed together with their aortic geometry previously assessed. Orthogonal patches centered in each aortic candidate lesion were reconstructed and a dataset with 19,790 images (61% positives) was built. Three single-input 2D CNNs were trained using axial, coronal and sagittal patches together with two multi-input 2.5D CNNs combining the orthogonal patches and identifying their best regional combination (BRC) in terms of lesion location. Aortic calcifications were concentrated in the descending (66%) and aortic arch (26%) portions. The BRC of axial patches to detect ascending or aortic arch lesions and sagittal images for the descending portion had the best performance: 0.954 F1-Score, 98.4% sensitivity, 87% of the subjects correctly classified in their TAC category and an average false positive TAC score per patient of 30. A CNN that combined axial and sagittal patches depending on the candidate aortic location ensured an accurate TAC score prediction.

4D flow MRI: impact of region of interest size, angulation and spatial resolution on aortic flow assessment

Objectives.In cardiovascular magnetic resonance, the 3D time-resolved phase-contrast technique, also known as 4D flow, is gaining increasing attention due to applications that exploit three-directional velocity encoding throughout the cardiac cycle. Blood flow volume assessment usually requires an expert to draw regions of interest (ROI) around the vessel cross section, whereas the errors involved in this estimation have not been thoroughly investigated. Our objective is to quantify the influence of ROI sizing, angulation and spatial resolution of the reconstructed plane employed in blood flow measurements using 4D flow.Approach.Three circular ROIs were drawn around the ascending, arch and descending aorta of healthy volunteers (n= 27) and patients with a dilated ascending aorta or bicuspid valve (n= 37). We applied systematic changes of ROI diameter (up to ±10%), tilt angle (up to ±25°) and spatial resolution (from 0.25 to 2 mm) of the reconstructed oblique planes, calculating the effects on net, forward and backward blood flow volumes.Main results.Patients had a larger ascending aorta than healthy volunteers with similar ages and male sex proportion (60 ± 15 y.o. vs 58 ± 16 y.o. and 84% vs 70%, respectively). Higher forward and backward flow volumes were observed in the ascending aorta and the aortic arch of the patients with respect to controls (p< 0.001), whereas net volumes were similar: 74.0 ± 20.8 ml versus 75.7 ± 21.8 ml (p= 0.37), respectively. The ascending aorta was the most sensitive to ROI modifications. Changes of ±10% in the ROI diameter and ±25° in tilt angles produced flow volume differences of up to 9 ml (10%) and 18 ml (15%) in controls and patients, respectively. Modifying the reconstructed planes spatial resolution produced flow volume changes below 2 ml.Significance.Since the setting of the ROI size and plane angle could produce errors that represent up to 20% of the forward and/or backward aortic flow volume, a good standardization for vessel segmentation and plane positioning is desirable.

Comprehensive assessment of local and regional aortic stiffness in patients with tricuspid or bicuspid aortic valve aortopathy using magnetic resonance imaging

BACKGROUND

We aimed to provide a comprehensive aortic stiffness description using magnetic resonance imaging (MRI) in patients with ascending thoracic aorta aneurysm and tricuspid (TAV-ATAA) or bicuspid (BAV) aortic valve.

METHODS

This case-control study included 18 TAV-ATAA and 19 BAV patients, with no aortic valve stenosis/severe regurgitation, who were 1:1 age-, gender- and central blood pressures (BP)-matched to healthy volunteers. Each underwent simultaneous aortic MRI and BP measurements. 3D anatomical MRI provided aortic diameters. Stiffness indices included: regional ascending (AA) and descending (DA) aorta pulse wave velocity (PWV) from 4D flow MRI; local AA and DA strain, distensibility and theoretical Bramwell-Hill (BH) model-based PWV, as well as regional arch PWV from 2D flow MRI.

RESULTS

Patient groups had significantly higher maximal AA diameter (median[interquartile range], TAV-ATAA: 47.5[42.0-51.3]mm, BAV: 45.0[41.0-47.0]mm) than their respective controls (29.1[26.8-31.8] and 28.1[26.0-32.0]mm, p < 0.0001), while BP were similar (p ≥ 0.25). Stiffness indices were significantly associated with age (ρ ≥ 0.33), mean BP (arch PWV: ρ = 0.25, p = 0.05; DA distensibility: ρ = -0.30, p = 0.02) or AA diameter (arch PWV: ρ = 0.28, p = 0.03; DA PWV: ρ = 0.32, p = 0.009). None of them, however, was significantly different between TAV-ATAA or BAV patients and their matched controls. Finally, while direct PWV measures were significantly correlated to BH-PWV estimates in controls (ρ ≥ 0.40), associations were non-significant in TAV-ATAA and BAV groups (p ≥ 0.18).

CONCLUSIONS

The overlap of MRI-derived aortic stiffness indices between patients with TAV or BAV aortopathy and matched controls highlights another heterogeneous feature of aortopathy, and suggests the urgent need for more sensitive indices which might help better discriminate such diseases.

Effects on Aortoiliac Fluid Dynamics After Endovascular Sealing of Abdominal Aneurysms

OBJECTIVES:

To evaluate the effects on aortoiliac fluid dynamics after the implantation of an endograft based on endovascular aneurysm sealing (EVAS) versus endovascular aneurysm repair (EVAR) strategy.

METHODS:

An adaptive geometrical deformable model was used for aortic lumen segmentation in 8 patients before and after the surgery. Abdominal aneurysms were treated with an endograft based on the EVAS system (Nellix, n = 4) and with a device based on an anatomical fixation technology (n = 4). Pressure, blood velocity, and wall shear stress (WSS) were estimated at different aortic regions using computational fluid dynamics methods. Physiologic inlet/outlet flow values at the abdominal aorta, the celiac trunk, and the mesenteric and the renal arteries were set. Pressure references were set at iliac arteries outlet.

RESULTS:

Maximum aneurysm sizes were similar for both groups in the preoperative scans. The lumen area was lower after EVAR ( P < .05) and EVAS ( P < .01) compared to preoperative aortic lumen sizes. Pressure increase was higher in the proximal abdominal aorta after EVAS compared to EVAR (2.3 ± 0.3 mm Hg vs 0.9 ± 0.3 mm Hg, P < .001). Peak blood velocities inside the endografts were 3-fold higher for EVAS compared to EVAR (54 ± 5 cm/s vs 17 ± 4 cm/s, P < .01). Velocities at the iliac arteries also remained higher for EVAS (38 ± 4 cm/s vs 24 ± 4 cm/s, P < .05). Peak WSS at the iliac arteries remained higher for EVAS compared to EVAR group ( P < .05).

CONCLUSION:

The significant modification of the aortic bifurcation anatomy after EVAS alters aortoiliac fluid dynamics, showing a pressure impact at the renal arteries level and an acceleration of the blood velocity at the iliac region with a concomitant increase in peak WSS.

Continuous assessment of carotid intima-media thickness applied to estimate a volumetric compliance using B-mode ultrasound sequences

Recent reports have shown that the carotid artery wall had significant movements not only in the radial but also in the longitudinal direction during the cardiac cycle. Accordingly, the idea that longitudinal elongations could be systematically neglected for compliance estimations became controversial. Assuming a dynamic change in vessel length, the standard measurement of cross-sectional compliance can be revised. In this work, we propose to estimate a volumetric compliance based on continuous measurements of carotid diameter and intima-media thickness (IMT) from B-mode ultrasound sequences. Assuming the principle of conservation of the mass of wall volume (compressibility equals zero), a temporal longitudinal elongation can be calculated to estimate a volumetric compliance. Moreover, elongations can also be estimated allowing small compressibility factors to model some wall leakage. The cross-sectional and the volumetric compliance were estimated in 45 healthy volunteers and 19 asymptomatic patients. The standard measurement underestimated the volumetric compliance by 25% for young volunteers (p < 0.01) and 17% for patients (p < 0.05). When compressibility factors different from zero were allowed, volunteers and patients reached values of 9% and 4%, respectively. We conclude that a simultaneous assessment of carotid diameter and IMT can be employed to estimate a volumetric compliance incorporating a longitudinal elongation. The cross-sectional compliance, that neglects the change in vessel length, underestimates the volumetric compliance.

Calcifications of the thoracic aorta on extended non-contrast-enhanced cardiac CT

Background

The presence of calcified atherosclerosis in different vascular beds has been associated with a higher risk of mortality. Thoracic aorta calcium (TAC) can be assessed from computed tomography (CT) scans, originally aimed at coronary artery calcium (CAC) assessment. CAC screening improves cardiovascular risk prediction, beyond standard risk assessment, whereas TAC performance remains controversial. However, the curvilinear portion of the thoracic aorta (TA), that includes the aortic arch, is systematically excluded from TAC analysis. We investigated the prevalence and spatial distribution of TAC all along the TA, to see how those segments that remain invisible in standard TA evaluation were affected.

Methods and Results

A total of 970 patients (77% men) underwent extended non-contrast cardiac CT scans including the aortic arch. An automated algorithm was designed to extract the vessel centerline and to estimate the vessel diameter in perpendicular planes. Then, calcifications were quantified using the Agatston score and associated with the corresponding thoracic aorta segment. The aortic arch and the proximal descending aorta, “invisible” in routine CAC screening, appeared as two vulnerable sites concentrating 60% of almost 11000 calcifications. The aortic arch was the most affected segment per cm length. Using the extended measurement method, TAC prevalence doubled from 31% to 64%, meaning that 52% of patients would escape detection with a standard scan. In a stratified analysis for CAC and/or TAC assessment, 111 subjects (46% women) were exclusively identified with the enlarged scan.

Conclusions

Calcium screening in the TA revealed that the aortic arch and the proximal descending aorta, hidden in standard TA evaluations, concentrated most of the calcifications. Middle-aged women were more prone to have calcifications in those hidden portions and became candidates for reclassification.

Effects of aging on thoracic aorta size and shape: a non-contrast CT study

Measures of atherosclerosis burden like coronary artery calcification are performed using non-contrast heart CT. However, additional information can be derived from these studies, looking beyond the coronary arteries without exposing the patients to further radiation. We present a semi-automated method to assess ascending, arch and descending aorta geometry from non-contrast CT datasets in 250 normotensive patients. We investigated the effect of aging on thoracic aorta morphometry. The algorithm identifies the aortic centerline coordinates following a toroidal path for the curvilinear portion and axial planes for descending aorta. Then it reconstructs oblique planes orthogonal to the centerline direction and a circle fitting process estimates the vessel cross-section. Finally, global thoracic aorta dimensions (diameter, volume and length) and shape (vessel curvature and tortuosity, aortic arch width and height) are calculated. From a multivariate analysis, adjusted for gender and body-size area, aortic volume and arch width were the descriptors that better represented the aortic size and shape alterations with aging. The thoracic aorta suffers an expanding and unfolding process with aging that deserves further attention to prevent aortic aneurisms.

Three-dimensional evaluation of thoracic aorta enlargement and unfolding in hypertensive men using non-contrast computed tomography

Aging produces a simultaneous thoracic aorta (TA) enlargement and unfolding. We sought to analyze the impact of hypertension on these geometric changes. Non-contrast computed tomography images were obtained from coronary artery calcium scans, including the entire aortic arch, in 200 normotensive and 200 hypertensive asymptomatic men. An automated algorithm reconstructed the vessel in three-dimensions, estimating orthogonal aortic sections along the whole TA pathway, and calculated several geometric descriptors to assess TA morphology. Hypertensive patients were older with respect to normotensive (P<0.001). Diameter and volume of TA ascending, arch and descending segments were higher in hypertensive patients with respect to normotensive (P<0.001) and differences persisted after adjustment for age. Hypertension produced an accelerated unfolding effect on TA shape. We found increments in aortic arch width (P<0.001), radius of curvature (P<0.001) and area under the arch curve (P<0.01) with a concomitant tortuosity decrease (P<0.05) and no significant change in aortic arch height. Overall, hypertension produced an equivalent effect of 2-7-years of aging. In multivariate analysis adjusted for age and hypertension treatment, diastolic pressure was more associated to TA size and shape changes than systolic pressure. These data suggest that hypertension accelerates TA enlargement and unfolding deformation with respect to the aging effect.

Estimation of coronary length-volume allometric relations of human arteries in vivo using CT

Assessing the individual geometry of the coronary arteries in a patient can help to explain diffuse artery disease. Some allometric functions, relating arterial length and volume, were verified in porcine arteries and human autopsies but not in vivo. In this work we use skeletonization methods on MSCT images to render the whole coronary tree in healthy and cardiovascular patients. Twenty patients with and without coronary artery disease were recruited. The coronary was segmented with minimum user intervention. Vessels were separated and measured. A 3D coronary map was individually calculated. The allometric length-volume function L=k(v)V(β) was evaluated in each patient and plotted in a Log-Log scale. The coefficient k(v) ranged 1.00 ± 0.35. Slopes ranged β = 0.69-0.88 and seemed to overlap in the scatter Log plot. The analysis of covariance verified this perception and concluded that lines were parallel. In other words, the allometric function stood for all patients. Values were not different from other studies in humans and pigs. The combination of multislice CT with morphological extraction algorithms was effective to extract allometric functions from coronary arteries in patients and can be easily applied in the clinic.

Angle estimation of human femora in a three-dimensional virtual environment

The estimation of human femur morphology and angulation provide useful information for assisted surgery, follow-up evaluation and prosthesis design, cerebral palsy management, congenital dislocation of the hip and fractures of the femur. Conventional methods that estimate femoral neck anteversion employ planar projections because accurate 3D estimations require complex reconstruction routines. In a recent work, we proposed a cylinder fitting method to estimate bifurcation angles in coronary arteries and we thought to test it in the estimation of femoral neck anteversion, valgus and shaft-neck angles. Femora from 10 patients were scanned using multisliced computed tomography. Virtual cylinders were fitted to 3 regions of the bone painted by the user to automatically estimate the femoral angles. Comparisons were made with a conventional manual method. Inter- and intra-reading measurements were evaluated for each method. We found femoral angles from both methods strongly correlated. Average anteversion, neck-shaft and valgus angles were 17.5°, 139.5°, 99.1°, respectively. The repeatability and reproducibility of the automated method showed a 5-fold reduction in inter- and intra-reading variability. Accordingly, the coefficients of variation for the manual method were below 25% whereas for the automated method were below 6%. The valgus angle assessment was globally the most accurate with differences below 1°. Maximum distances from true surface bone points and fitting cylinders attained 6 mm. The employment of virtual cylinders fitted to different regions of human femora consistently helped to assess true 3D angulations.