Deriving coronary artery calcium scores from CT coronary angiography: a proposed algorithm for evaluating stable chest pain

A fully automated deep learning approach for coronary artery segmentation and comprehensive characterization

Coronary computed tomography angiography (CCTA) allows detailed assessment of early markers associated with coronary artery disease (CAD), such as coronary artery calcium (CAC) and tortuosity (CorT). However, their analysis can be time-demanding and biased. We present a fully automated pipeline that performs (i) coronary artery segmentation and (ii) CAC and CorT objective analysis. Our method exploits supervised learning for the segmentation of the lumen, and then, CAC and CorT are automatically quantified. 281 manually annotated CCTA images were used to train a two-stage U-Net-based architecture. The first stage employed a 2.5D U-Net trained on axial, coronal, and sagittal slices for preliminary segmentation, while the second stage utilized a multichannel 3D U-Net for refinement. Then, a geometric post-processing was implemented: vessel centerlines were extracted, and tortuosity score was quantified as the count of branches with three or more bends with change in direction forming an angle >45°. CAC scoring relied on image attenuation. CAC was detected by setting a patient specific threshold, then a region growing algorithm was applied for refinement. The application of the complete pipeline required <5 min per patient. The model trained for coronary segmentation yielded a Dice score of 0.896 and a mean surface distance of 1.027 mm compared to the reference ground truth. Tracts that presented stenosis were correctly segmented. The vessel tortuosity significantly increased locally, moving from proximal, to distal regions (p < 0.001). Calcium volume score exhibited an opposite trend (p < 0.001), with larger plaques in the proximal regions. Volume score was lower in patients with a higher tortuosity score (p < 0.001). Our results suggest a linked negative correlation between tortuosity and calcific plaque formation. We implemented a fast and objective tool, suitable for population studies, that can help clinician in the quantification of CAC and various coronary morphological parameters, which is helpful for CAD risk assessment.

Automatic coronary plaque detection, classification, and stenosis grading using deep learning and radiomics on computed tomography angiography images: a multi-center multi-vendor study

OBJECTIVES

An automatic system utilizing both the advantages of the neural network and the radiomics was proposed for coronary plaque detection, classification, and stenosis grading.

METHODS

This study retrospectively included 505 patients with 127,763 computed tomography angiography (CTA) images from 5 medical center. A convolutional neural network (CNN) model was used to segment the coronary artery, detect the plaque candidate, and extract the image patch with high computation efficiency. The manually designed radiomics feature extractor was utilized to collect plaque patterns, followed by the different classifiers to perform the plaque classification and stenosis grading.

RESULTS

The CNN model achieved 100% of sensitivity and the highest positive predictive value (83.9%) than U-Net and baseline model in plaque candidate detection. Twenty-six representative radiomics features were selected to construct the classifiers. Among different models, the gradient-boosting decision tree (GBDT) achieved the best performance in plaque classification (accuracy: 87.0%, sensitivity: 83.2%, specificity: 91.4%) and stenosis grading (accuracy: 90.9%, sensitivity: 84.1%, specificity: 95.7%). GBDT also achieved the highest AUC of 0.873 in plaque classification and 0.910 in stenosis grading. The computation time of processing one patient was 56.2 ± 5.7 s which was significantly less than radiologist manual analysis (285.6 ± 134.5 s, p = 0.0001).

CONCLUSIONS

In this study, an automatic workflow was proposed to detect and analyze coronary plaques with high accuracy and efficiency, showing the potential in clinical application.

KEY POINTS

• The proposed automatic system integrated deep learning and radiomics to perform the coronary plaque analysis. • The proposed automatic system achieved high accuracy in both plaque classification and stenosis grading. • The proposed automatic system was five times more efficient than radiologist manual analysis.

Machine Learning and Coronary Artery Calcium Scoring

PURPOSE OF REVIEW

To summarize current artificial intelligence (AI)-based applications for coronary artery calcium scoring (CACS) and their potential clinical impact.

RECENT FINDINGS

Recent evolution of AI-based technologies in medical imaging has accelerated progress in CACS performed in diverse types of CT examinations, providing promising results for future clinical application in this field. CACS plays a key role in risk stratification of coronary artery disease (CAD) and patient management. Recent emergence of AI algorithms, particularly deep learning (DL)-based applications, have provided considerable progress in CACS. Many investigations have focused on the clinical role of DL models in CACS and showed excellent agreement between those algorithms and manual scoring, not only in dedicated coronary calcium CT but also in coronary CT angiography (CCTA), low-dose chest CT, and standard chest CT. Therefore, the potential of AI-based CACS may become more influential in the future.

Automated plaque classification using computed tomography angiography and Gabor transformations

Cardiovascular diseases are the primary cause of death globally. These are often associated with atherosclerosis. This inflammation process triggers important variations in the coronary arteries (CA) and can lead to coronary artery disease (CAD). The presence of CA calcification (CAC) has recently been shown to be a strong predictor of CAD. In this clinical setting, computed tomography angiography (CTA) has begun to play a crucial role as a non-intrusive imaging method to characterize and study CA plaques. Herein, we describe an automated algorithm to classify plaque as either normal, calcified, or non-calcified using 2646 CTA images acquired from 73 patients. The automated technique is based on various features that are extracted from the Gabor transform of the acquired CTA images. Specifically, seven features are extracted from the Gabor coefficients : energy, and Kapur, Max, Rényi, Shannon, Vajda, and Yager entropies. The features were then ordered based on the F-value and input to numerous classification methods to achieve the best classification accuracy with the least number of features. Moreover, two well-known feature reduction techniques were employed, and the features acquired were also ranked according to F-value and input to several classifiers. The best classification results were obtained using all computed features without the employment of feature reduction, using a probabilistic neural network. An accuracy, positive predictive value, sensitivity, and specificity of 89.09%, 91.70%, 91.83% and 83.70% was obtained, respectively. Based on these results, it is evident that the technique can be helpful in the automated classification of plaques present in CTA images, and may become an important tool to reduce procedural costs and patient radiation dose. This could also aid clinicians in plaque diagnostics.

To fly as a pilot after cardiac surgery

Aircrew are responsible for safe and reliable aircraft operations. Cardiovascular disease accounts for 50% of all pilot licences declined or withdrawn for medical reasons in Western Europe and is the most common cases of sudden incapacitation in flight. Aircrew retirement age is increasing (up to age 65) in a growing number of airlines and the burden of subclinical, but potentially significant, coronary atherosclerosis is unknown in qualified pilots above age 40. Safety considerations are paramount in aviation medicine, and the most dreaded cardiovascular complications are thromboembolic events and rhythm disturbances due to their potential for sudden incapacitation. In aviation, the current consensus risk threshold for an acceptable level of controlled risk of acute incapacitation is 1% (for dual pilot commercial operations), a percentage calculated using engineering principles to ensure the incidence of a fatal air accident is no greater than 1 per 107 h of flying. This is known as the '1% safety rule'. To fly as a pilot after cardiac surgery is possible; however, special attention to perioperative planning is mandatory. Choice of procedure is crucial for license renewal. Licensing restrictions are likely to apply and the postoperative follow-up requires a tight scheduling. The cardiac surgeon should always liaise and communicate with the pilot's aviation medicine examiner prior to and following cardiac surgery.

The HASTE Protocol: a standardised CT Coronary Angiography service operated from a District General Hospital

Introduction

CT coronary angiography (CTCA) has excellent sensitivity but lacks specificity when compared with invasive coronary angiography (ICA) particularly in patients with a high coronary calcium burden. CTCA has been shown in large trials to decrease the requirement for diagnostic ICA and provide diagnostic clarity. We describe the methodology used to provide a standardised CTCA service established in a District General Hospital, which may assist other hospitals aiming to develop a cardiac CT service.

Methods

Scan request forms, authorisation and patient instruction were recorded. Patient preparation prior to CTCA as well as exclusion and inclusion criteria were documented. Scans were interpreted using a multidisciplinary team (MDT) approach in order to organise follow-up, medication and further investigation.

Results

Over 6 months, 157 consecutive scans were performed. CTCA was completed in 88% (n=138/157) and considered of diagnostic quality in 82% (n=129/157). The median radiation dose was 3.42 mSv. Overall, 64% of patients had evidence of coronary calcium. Following MDT review, 72% (n=113/157) of patients were discharged without requiring invasive angiography. 15% (n=24/157) of patients went on to have invasive angiography showing non-obstructive disease and 13% (20/157) of patients underwent percutaneous coronary intervention (11%) or bypass surgery (1%).

Discussion

Appropriate referrals, patient preparation and scan quality remain significant factors in running a CTCA service. Despite this, the vast majority of patients can be discharged on the basis of the CTCA alone. An MDT approach is key to the delivery of a cardiac CT service.

A prospective national survey of coronary CT angiography radiation doses in the United Kingdom

BACKGROUND

Little real-world radiation dose data exist for the majority of cardiovascular CT. Some data have been published for coronary CT angiography (coronary CTA) specifically, but they invariably arise from high-volume centres with access to the most recent technology.

OBJECTIVE

The aim of this study was to document real-world radiation doses for coronary CTA in the United Kingdom, and to establish their relationship to clinical protocol selection, acquisition heart rate, and scanner technology.

METHODS

A dose survey questionnaire was distributed to members of the British Society of Cardiovascular Imaging and other UK cardiac CT units. All participating centres collected data for consecutive coronary CTA cases over one month. The survey captured information about the exam conducted, patient demographics, pre-scan details such as beta-blocker administration, acquisition heart rate and scan technique, and post-scan dose indicators - series volumetric CT dose index (CTDI_vol), series dose-length product (DLP), and exam DLP.

RESULTS

Fifty centres provided data on a total of 1341 coronary CTA exams. Twenty-nine centres (58%) performed at least 20 coronary CTA scans in the collection period. The median BMI, acquisition heart rate and exam DLP were 28 kg/m², 60 bpm and 209 mGycm respectively. The corresponding effective dose was estimated as 5.9 mSv using a conversion factor of 0.028 mSv/mGycm. There was no statistically significant difference in radiation dose between low and high-volume centres. Median exam DLP increased with the acquisition heart rate due to the selection of wider temporal windows. The highest exam DLPs were obtained on the older scanner technology.

CONCLUSION

This study provides baseline data for benchmarking practice, optimizing radiation dose and improving service quality locally.

Assessment of Coronary Artery Calcium on Low-Dose Coronary Computed Tomography Angiography With Iterative Reconstruction

OBJECTIVE

This study aims to evaluate whether coronary calcium scoring (CCS) is also feasible using low-radiation-dose coronary computed tomography angiography (CCTA) in combination with iterative reconstruction.

METHODS

Forty-three individuals without known coronary artery disease underwent both noncontrast CCS (±1 mSv) for reference Agatston scores and low-dose CCTA (±3 mSv). Raw CCTA data were reconstructed with filtered back projection (FBP), hybrid iterative reconstruction (HIR), and model-based iterative reconstruction (MIR). Calcification volumes were derived with thresholds of >351 and >600 Hounsfield units (HU) and converted to proxy Agatston scores with linear regression analysis.

RESULTS

Intraclass correlation coefficients for Agatston scores versus CCTA volumes with FBP and iterative reconstruction were excellent (ranges 0.94-0.99 and 0.96-0.99 for >351 HU and >600 HU thresholds, respectively). The >351 HU threshold resulted in higher CCTA volume scores ranging from 65.9 (15.1-347.0) for HIR to 94.8 (42.0-423.0) for MIR (P = 0.001 and <0.001, respectively). The >600 HU threshold scores ranged from 14.1 (0.0-159.3) for HIR to 28.6 (0.0-215.6) for MIR (P = 0.003 and 0.027, respectively). At >351 HU, reclassification occurred in 21 individuals (49%) for FBP and HIR and 25 individuals (58%) for MIR. Reclassifications decreased with >600 HU to 10 (HIR, 23%), 8 (FBP, 19%), and 4 (MIR, 9%).

CONCLUSIONS

The CCS is feasible using iteratively reconstructed low-dose CCTA with a calcium threshold of >600 HU. Using MIR, only 9% of individuals were reclassified.

Automatic coronary artery calcium scoring in cardiac CT angiography using paired convolutional neural networks

The amount of coronary artery calcification (CAC) is a strong and independent predictor of cardiovascular events. CAC is clinically quantified in cardiac calcium scoring CT (CSCT), but it has been shown that cardiac CT angiography (CCTA) may also be used for this purpose. We present a method for automatic CAC quantification in CCTA. This method uses supervised learning to directly identify and quantify CAC without a need for coronary artery extraction commonly used in existing methods. The study included cardiac CT exams of 250 patients for whom both a CCTA and a CSCT scan were available. To restrict the volume-of-interest for analysis, a bounding box around the heart is automatically determined. The bounding box detection algorithm employs a combination of three ConvNets, where each detects the heart in a different orthogonal plane (axial, sagittal, coronal). These ConvNets were trained using 50 cardiac CT exams. In the remaining 200 exams, a reference standard for CAC was defined in CSCT and CCTA. Out of these, 100 CCTA scans were used for training, and the remaining 100 for evaluation of a voxel classification method for CAC identification. The method uses ConvPairs, pairs of convolutional neural networks (ConvNets). The first ConvNet in a pair identifies voxels likely to be CAC, thereby discarding the majority of non-CAC-like voxels such as lung and fatty tissue. The identified CAC-like voxels are further classified by the second ConvNet in the pair, which distinguishes between CAC and CAC-like negatives. Given the different task of each ConvNet, they share their architecture, but not their weights. Input patches are either 2.5D or 3D. The ConvNets are purely convolutional, i.e. no pooling layers are present and fully connected layers are implemented as convolutions, thereby allowing efficient voxel classification. The performance of individual 2.5D and 3D ConvPairs with input sizes of 15 and 25 voxels, as well as the performance of ensembles of these ConvPairs, were evaluated by a comparison with reference annotations in CCTA and CSCT. In all cases, ensembles of ConvPairs outperformed their individual members. The best performing individual ConvPair detected 72% of lesions in the test set, with on average 0.85 false positive (FP) errors per scan. The best performing ensemble combined all ConvPairs and obtained a sensitivity of 71% at 0.48 FP errors per scan. For this ensemble, agreement with the reference mass score in CSCT was excellent (ICC 0.944 [0.918-0.962]). Aditionally, based on the Agatston score in CCTA, this ensemble assigned 83% of patients to the same cardiovascular risk category as reference CSCT. In conclusion, CAC can be accurately automatically identified and quantified in CCTA using the proposed pattern recognition method. This might obviate the need to acquire a dedicated CSCT scan for CAC scoring, which is regularly acquired prior to a CCTA, and thus reduce the CT radiation dose received by patients.

Coronary Artery Disease Risk Factors, Coronary Artery Calcification and Coronary Bypass Surgery

INTRODUCTION

Atherosclerosis is an intimal disease which affects large and medium size arteries including aorta and carotid, coronary, cerebral and radial arteries. Calcium accumulated in the coronary arterial plaques have substantial contribution to the plaque volume. The aim of our study is to investigate the relationship between coronary artery disease (CAD) risk factors and coronary arterial calcification, and to delineate the importance of CACS in coronary artery bypass surgery.

MATERIALS AND METHODS

The current study is retrospective and 410 patients admitted to our clinic with atypical chest pain and without known CAD were included. These individuals were evaluated by 16 slice electron beam computed tomography with suspicion of CAD and their calcium scores were calculated. Detailed demographic and medical history were obtained from all of the patients.

RESULTS

In our study, we employed five different analyses using different coronary arterial calcification score (CACS) thresold levels reported in previous studies. All of the analyses, performed according to the previously defined thresold levels, showed that risk factors had strong positive relationship with CACS as mentioned in previous studies.

CONCLUSION

Coronary arterial calcification is part of the athero-sclerotic process and although it can be detected in atherosclerotic vessel, it is absent in a normal vessel. It can be concluded that the clinical scores, even they are helpful, have some limitations in a significant part of the population for cardiovascular risk determination. It is important for an anastomosis region to be noncalcified in coronary bypass surgery. In a coronary artery, it will be helpness for showing of calcific field and anostomosis spot.

Preoperative coronary calcium score is predictive of early postoperative cardiovascular complications in liver transplant recipients

BACKGROUND

Coronary computed tomographic angiography (coronary CT) is a non-invasive test for diagnosis of cardiac function. Coronary calcium scores determined by coronary CT are associated with cardiovascular risk factors. However, no studies have investigated the association between coronary calcium scores and cardiovascular complications after liver transplantation (LT). We therefore evaluated the utility of preoperative coronary calcium scores for predicting early postoperative cardiovascular complications in LT recipients.

METHODS

Between 2010 and 2012, 443 LT recipients were analysed retrospectively. Preoperative cardiovascular assessments, including coronary CT, were performed. A coronary calcium score >400 was defined as a positive finding. Predictive factors of early postoperative cardiovascular complications were evaluated by univariate and multivariate analyses. Major cardiovascular complications occurring during a period of 1 month after LT were noted.

RESULTS

Of the 443 patients, 38 (8.6%) experienced one or more cardiovascular complications. Positive coronary CT findings were seen in 11 (2.5%) patients. In the multivariate analysis, a coronary calcium score >400 {odds ratio (OR)=4.62 [95% confidence interval (CI): 1.14-18.72], P=0.032} and female sex [OR=2.76 (1.37-5.57), P=0.005] were predictive of cardiovascular complications.

CONCLUSIONS

A preoperative coronary calcium score of >400 predicted cardiovascular complications occurring 1 month after LT, suggesting that preoperative evaluation of coronary calcium scores could help predict early postoperative cardiovascular complications in LT recipients.