AI Evaluation of Stenosis on Coronary CTA, Comparison With Quantitative Coronary Angiography and Fractional Flow Reserve: A CREDENCE Trial Substudy

Artificial Intelligence-Driven Assessment of Coronary Computed Tomography Angiography for Intermediate Stenosis: Comparison With Quantitative Coronary Angiography and Fractional Flow Reserve

We aimed to compare artificial intelligence (AI)-based coronary stenosis evaluation of coronary computed tomography angiography (CCTA) with its quantitative counterpart of invasive coronary angiography (ICA) and invasive fractional flow reserve (FFR). This single-center retrospective study included 195 symptomatic patients (mean age 61 ± 10 years, 149 men, 585 coronary arteries) with 215 intermediate coronary lesions, with quantitative coronary angiography (QCA) diameter stenosis ranging from 20% to 80%. An AI-driven research prototype (AI-CCTA) was used to quantify stenosis on CCTA images. The diagnostic accuracy of AI-CCTA was assessed on a per-vessel basis using ICA stenosis grading (with ≥50% stenosis) or invasive FFR (≤0.80) as reference standards. AI-driven diameter stenosis was correlated with the QCA results and expert manual measurements subsequently. The disease prevalence in the 585 coronary arteries, as determined by invasive angiography (≥50%), was 46.5%. AI-CCTA exhibited sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve (AUC) of 71.7%, 89.8%, 85.9%, 78.5%, and 0.81, respectively. The diagnostic performance of AI-CCTA was moderate for the 215 intermediate lesions assessed using QCA and FFR, with an AUC of 0.63 for QCA and FFR. AI-CCTA demonstrated a moderate correlation with QCA (r = 0.42, p <0.001) for measuring the degree of stenosis, which was notably better than the results from manual quantification versus QCA (r = 0.26, p = 0.001). In conclusion, AI-driven CCTA analysis exhibited promising results. AI-CCTA demonstrated a moderate relation with QCA in intermediate coronary stenosis lesions; however, its results surpassed those of manual evaluations.

Fractional Flow Reserve Relates Stronger to Coronary Plaque Burden Than Nonhyperemic Pressure Indexes

BACKGROUND

The relationship between fractional flow reserve (FFR), resting full-cycle ratio (RFR), instantaneous wave-free ratio (iFR), resting distal pressure/aortic pressure (Pd/Pa), and plaque burden as well as phenotype requires further elucidation.

METHODS AND RESULTS

In this single-center cohort study, patients with suspected coronary artery disease who underwent invasive coronary angiography, including routine hyperemic (FFR) and nonhyperemic invasive pressure (Pd/Pa and iFR or RFR) interrogation and computed coronary tomography angiography were prospectively enrolled. Computed coronary tomography angiography was used to assess percentage atheroma volume (PAV), positive remodeling, and low-attenuation plaque.

Among 241 patients with 556 vessels, FFR correlated stronger to PAV compared with Pd/Pa (r=-0.56; versus r=-0.43; P<0.01) and iFR/RFR (r=-0.47; P=0.04). Vessels with FFR and Pd/Pa discordancy showed higher PAV in case of abnormal FFR (34% versus 14%; P<0.01), whereas vessels with FFR and iFR/RFR discordancy showed similar PAV levels. FFR and iFR/RFR, but not Pd/Pa, were independently associated with the presence of low-attenuation plaque (β, -0.03, P<0.01; β, -0.03, P=0.01; and β, -0.02, P=0.10, respectively). None of the invasive pressure measurements was independently associated with positive remodeling. Pressure index discordancy was not associated with positive remodeling or low-attenuation plaque.

CONCLUSIONS

FFR correlated stronger to plaque burden, as defined by PAV, than nonhyperemic pressure indexes. For plaque phenotype, both FFR and iFR/RFR were independently associated with low-attenuation plaque, whereas none of the invasive pressure indexes was associated with positive remodeling.

Plasma proteomics improves prediction of coronary plaque progression

AIMS

Coronary computed tomography angiography (CCTA) offers detailed imaging of plaque burden and composition, with plaque progression being a key determinant of future cardiovascular events. As repeated CCTA scans are burdensome and costly, there is a need for non-invasive identification of plaque progression. This study evaluated whether combining proteomics with traditional risk factors can detect patients at risk for accelerated plaque progression.

METHODS AND RESULTS

This long-term follow-up study included 97 participants who underwent two CCTA scans and plasma proteomics analysis using the Olink platform. Accelerated plaque progression was defined as rates above the median for percent atheroma volume (PAV), percent non-calcified plaque volume (NCPV), and percent calcified plaque volume (CPV). High-risk plaque (HRP) was identified by positive remodelling or low-density plaque at baseline and/or follow-up. Significant proteins associated with PAV, NCPV, CPV, and HRP development were incorporated into predictive models. The mean baseline age was 58.0 ± 7.4 years, with 63 (65%) male, and a median follow-up of 8.5 ± 0.6 years. The area under the curve (AUC) for accelerated PAV progression increased from 0.830 with traditional risk factors and baseline plaque volume to 0.909 with the protein panel (P = 0.023). For NCPV progression, AUC improved from 0.685 to 0.825 (P = 0.008), while no improvement was observed for CPV progression. For HRP development, AUC increased from 0.791 to 0.860 with the protein panel (P = 0.036).

CONCLUSION

Integrating proteomics with traditional risk factors enhances the prediction of accelerated plaque progression and high-risk plaque development, potentially improving risk stratification and treatment decisions without the need for repeated CCTAs.

Microvascular resistance reserve in relation to total and vessel-specific atherosclerotic burden

AIMS

The relationship between coronary artery atherosclerosis and microvascular resistance remains unclear. This study aims to clarify the relationship between total atherosclerotic and vessel-specific atherosclerotic burden and microvascular resistance reserve (MRR).

METHODS AND RESULTS

In this post hoc analysis of the PACIFIC 1 trial, symptomatic patients without prior coronary artery disease (CAD) underwent [15O]H2O positron emission tomography, coronary computed tomography angiography (CCTA), and invasive fractional flow reserve (FFR). MRR was assessed across all three coronary branches, utilizing PET-derived coronary flow reserve and invasive FFR measurements. CCTA was used to assess patient and vessel-specific plaque volumes. Percentage atheroma volume (PAV) was defined as total plaque volume divided by vessel volume. The study included 142 patients (55% male, 57.5 ± 8.6 years) with 426 vessels with a mean MRR of 3.77 ± 1.64. While a significantly higher PAV was observed in the left anterior descending artery territory, MRR was similar across the three coronary branches. Generalized estimating equations without correction for cardiovascular risk factors identified that patient-specific PAV tertiles but not vessel-specific PAV tertiles were related to vessel-specific MRR. After correction for cardiovascular risk factors, compared with the first tertile of patient-specific PAV, the second tertile showed a vessel-specific MRR decrease of β = -0.362, P = 0.018, and the third tertile showed a decrease of β = -0.347, P = 0.024.

CONCLUSION

In patients without prior CAD, patient-specific plaque burden was negatively associated to vessel-specific MRR; however, vessel-specific plaque burden was not related to vessel-specific MRR. Our findings suggest that the relation between atherosclerotic burden and an impaired microcirculatory function is of systemic origin.

Automated classification of coronary LEsions fRom coronary computed Tomography angiography scans with an updated deep learning model: ALERT study

OBJECTIVES

The use of deep learning models for quantitative measurements on coronary computed tomography angiography (CCTA) may reduce inter-reader variability and increase efficiency in clinical reporting. This study aimed to investigate the diagnostic performance of a recently updated deep learning model (CorEx-2.0) for quantifying coronary stenosis, compared separately with two expert CCTA readers as references.

METHODS

This single-center retrospective study included 50 patients that underwent CCTA to rule out obstructive coronary artery disease between 2017-2022. Two expert CCTA readers and CorEx-2.0 independently assessed all 150 vessels using Coronary Artery Disease-Reporting and Data System (CAD-RADS). Inter-reader agreement analysis and diagnostic performance of CorEx-2.0, compared with each expert reader as references, were evaluated using percent agreement, Cohen's kappa for the binary CAD-RADS classification (CAD-RADS 0-3 versus 4-5) at patient level, and linearly weighted kappa for the 6-group CAD-RADS classification at vessel level.

RESULTS

Overall, 50 patients and 150 vessels were evaluated. Inter-reader agreement using the binary classification at patient level was 91.8% (45/49) with a Cohen's kappa of 0.80. For the 6-group classification at vessel level, inter-reader agreement was 67.6% (100/148) with a linearly weighted kappa of 0.77. CorEx-2.0 showed 100% sensitivity for detecting CAD-RADS ≥ 4 and kappa values of 0.86 versus both readers using the binary classification at patient level. For the 6-group classification at vessel level, CorEx-2.0 demonstrated weighted kappa values of 0.71 versus reader 1 and 0.73 versus reader 2.

CONCLUSION

CorEx-2.0 identified all patients with severe stenosis (CAD-RADS ≥ 4) compared with expert readers and approached expert reader performance at vessel level (weighted kappa > 0.70).

KEY POINTS

Question Can deep learning models improve objectivity in coronary stenosis grading and reporting as coronary CT angiography (CTA) workloads rise? Findings The deep learning model (CorEx-2.0) identified all patients with severe stenoses when compared with expert readers and approached expert reader performance at vessel level. Clinical relevance CorEx-2.0 is a reliable tool for identifying patients with severe stenoses (≥ 70%), underscoring the potential of using this deep learning model to prioritize coronary CTA reading by flagging patients at risk of severe obstructive coronary artery disease.

Computed Tomography Angiography in the Catheterization Laboratory: A Guide Towards Optimizing Coronary Interventions

Cardiac computed tomography (CT) has become an essential tool in the pre-procedural planning and optimization of coronary interventions. Its non-invasive nature allows for the detailed visualization of coronary anatomy, including plaque burden, vessel morphology, and the presence of stenosis, aiding in precise decision making for revascularization strategies. Clinicians can assess not only the extent of coronary artery disease but also the functional significance of lesions using techniques like fractional flow reserve (FFR-CT). By providing comprehensive insights into coronary structure and hemodynamics, cardiac CT helps guide personalized treatment plans, ensuring the more accurate selection of patients for percutaneous coronary interventions or coronary artery bypass grafting and potentially improving patient outcomes.

Artificial Intelligence in Imaging for Personalized Management of Coronary Artery Disease

The precision of imaging and the number of other risk-assessing and diagnostic methods are constantly growing, allowing for the uptake of additional strategies for individualized therapies. Personalized medicine has the potential to deliver more adequate treatment, resulting in better clinical outcomes, based on each patient's vulnerability or genetic makeup. In addition to increased efficiency, costs related to this type of procedure can be significantly lower. Useful assistance in designing individual therapies may be assured by the adoption of artificial intelligence (AI). Recent years have brought essential developments in deep and machine learning techniques. Advances in technologies such as convolutional neural networks (CNNs) have enabled automatic analyses of images, numerical data, and video data, providing high efficiency in the creation of prediction models. The number of AI applications in medicine is constantly growing, and the effectiveness of these techniques has been demonstrated in coronary computed tomography angiography (CCTA), optical coherence tomography (OCT), and many others. Moreover, AI models may be useful in direct therapy optimization for patients with coronary artery disease (CAD), who are burdened with high risk. The combination of well-trained AI with the design of individual treatment pathways can lead to improvements in health care. However, existing limitations, such as non-adapted guidelines or the lack of randomized clinical trials to evaluate AI's true accuracy, may contribute to delays in introducing automatic methods into practical use. This review critically appraises the developed tools that are potentially useful for clinicians in guiding personalized patient management, as well as current trials in this field.

Diagnostic accuracy in coronary CT angiography analysis: artificial intelligence versus human assessment

BACKGROUND

Visual assessment of coronary CT angiography (CCTA) is time-consuming, influenced by reader experience and prone to interobserver variability. This study evaluated a novel algorithm for coronary stenosis quantification (atherosclerosis imaging quantitative CT, AI-QCT).

METHODS

The study included 208 patients with suspected coronary artery disease (CAD) undergoing CCTA in Perfusion Imaging and CT Coronary Angiography With Invasive Coronary Angiography-1. AI-QCT and blinded readers assessed coronary artery stenosis following the Coronary Artery Disease Reporting and Data System consensus. Accuracy of AI-QCT was compared with a level 3 and two level 2 clinical readers against an invasive quantitative coronary angiography (QCA) reference standard (≥50% stenosis) in an area under the curve (AUC) analysis, evaluated per-patient and per-vessel and stratified by plaque volume.

RESULTS

Among 208 patients with a mean age of 58±9 years and 37% women, AI-QCT demonstrated superior concordance with QCA compared with clinical CCTA assessments. For the detection of obstructive stenosis (≥50%), AI-QCT achieved an AUC of 0.91 on a per-patient level, outperforming level 3 (AUC 0.77; p<0.002) and level 2 readers (AUC 0.79; p<0.001 and AUC 0.76; p<0.001). The advantage of AI-QCT was most prominent in those with above median plaque volume. At the per-vessel level, AI-QCT achieved an AUC of 0.86, similar to level 3 (AUC 0.82; p=0.098) stenosis, but superior to level 2 readers (both AUC 0.69; p<0.001).

CONCLUSIONS

AI-QCT demonstrated superior agreement with invasive QCA compared to clinical CCTA assessments, particularly compared to level 2 readers in those with extensive CAD. Integrating AI-QCT into routine clinical practice holds promise for improving the accuracy of stenosis quantification through CCTA.

Location-specific prognostic significance of plaque burden, stenosis, and plaque morphology in coronary artery disease

AIMS

To investigate the location-specific prognostic significance of plaque burden, diameter stenosis, and plaque morphology.

METHODS AND RESULTS

Patients without a documented cardiac history that underwent coronary computed tomography angiography (CCTA) for suspected coronary artery disease were included. Percentage atheroma volume (PAV), maximum diameter stenosis, and plaque morphology were assessed and classified into proximal, mid, or distal segments of the coronary tree. Major adverse cardiac events (MACE) were defined as death or non-fatal myocardial infarction. Among 2819 patients 267 events (9.5%) occurred during a median follow-up of 6.9 years. When adjusted for traditional risk factors and the presence of PAV in other locations, only proximal PAV was independently associated with MACE. However, PAV of the proximal segments was strongly correlated to PAV localized at the mid (R = 0.76) and distal segments (R = 0.74, P < 0.01 for both). When only adjusted for cardiovascular risk factors, the area under the curve (AUC) to predict MACE for proximal PAV was 0.73 (95% CI 0.69-0.76), which was similar compared with mid PAV (AUC 0.72, 95% CI 0.68-0.76) and distal PAV (AUC 0.72, 95% CI 0.68-0.76). Similar results were obtained using diameter stenosis instead of PAV. The presence of proximal low-attenuation plaque had borderline additional prognostic value.

CONCLUSION

Proximal PAV was the strongest predictor of MACE when adjusted for cardiovascular risk factors and plaque at other locations. However, when the presence of plaque was only adjusted for cardiovascular risk factors, proximal, mid, and distal plaque localization showed a similar predictive ability for MACE.

First comparison between artificial intelligence-guided coronary computed tomography angiography versus single-photon emission computed tomography testing for ischemia in clinical practice

BACKGROUND

Noninvasive cardiac testing with coronary computed tomography angiography (CCTA) and single-photon emission computed tomography (SPECT) are becoming alternatives to invasive angiography for the evaluation of obstructive coronary artery disease. We aimed to evaluate whether a novel artificial intelligence (AI)-assisted CCTA program is comparable to SPECT imaging for ischemic testing.

METHODS

CCTA images were analyzed using an artificial intelligence convolutional neural network machine-learning-based model, atherosclerosis imaging-quantitative computed tomography (AI-QCT)ISCHEMIA. A total of 183 patients (75 females and 108 males, with an average age of 60.8 years ± 12.3 years) were selected. All patients underwent AI-QCTISCHEMIA-augmented CCTA, with 60 undergoing concurrent SPECT and 16 having invasive coronary angiograms. Eight studies were excluded from analysis due to incomplete data or coronary anomalies.

RESULTS

A total of 175 patients (95%) had CCTA performed, deemed acceptable for AI-QCTISCHEMIA interpretation. Compared to invasive angiography, AI-QCTISCHEMIA-driven CCTA showed a sensitivity of 75% and specificity of 70% for predicting coronary ischemia, versus 70% and 53%, respectively for SPECT. The negative predictive value was high for female patients when using AI-QCTISCHEMIA compared to SPECT (91% vs. 68%, P = 0.042). Area under the receiver operating characteristic curves were similar between both modalities (0.81 for AI-CCTA, 0.75 for SPECT, P = 0.526). When comparing both modalities, the correlation coefficient was r = 0.71 (P < 0.04).

CONCLUSION

AI-powered CCTA is a viable alternative to SPECT for detecting myocardial ischemia in patients with low- to intermediate-risk coronary artery disease, with significant positive and negative correlation in results. For patients who underwent confirmatory invasive angiography, the results of AI-CCTA and SPECT imaging were comparable. Future research focusing on prospective studies involving larger and more diverse patient populations is warranted to further investigate the benefits offered by AI-driven CCTA.

Polygenic Risk Is Associated With Long-Term Coronary Plaque Progression and High-Risk Plaque

BACKGROUND

The longitudinal relation between coronary artery disease (CAD) polygenic risk score (PRS) and long-term plaque progression and high-risk plaque (HRP) features is unknown.

OBJECTIVES

The goal of this study was to investigate the impact of CAD PRS on long-term coronary plaque progression and HRP.

METHODS

Patients underwent CAD PRS measurement and prospective serial coronary computed tomography angiography (CTA) imaging. Coronary CTA scans were analyzed with a previously validated artificial intelligence-based algorithm (atherosclerosis imaging-quantitative computed tomography imaging). The relationship between CAD PRS and change in percent atheroma volume (PAV), percent noncalcified plaque progression, and HRP prevalence was investigated in linear mixed-effect models adjusted for baseline plaque volume and conventional risk factors.

RESULTS

A total of 288 subjects (mean age 58 ± 7 years; 60% male) were included in this study with a median scan interval of 10.2 years. At baseline, patients with a high CAD PRS had a more than 5-fold higher PAV than those with a low CAD PRS (10.4% vs 1.9%; P < 0.001). Per 10 years of follow-up, a 1 SD increase in CAD PRS was associated with a 0.69% increase in PAV progression in the multivariable adjusted model. CAD PRS provided additional discriminatory benefit for above-median noncalcified plaque progression during follow-up when added to a model with conventional risk factors (AUC: 0.73 vs 0.69; P = 0.039). Patients with high CAD PRS had an OR of 2.85 (95% CI: 1.14-7.14; P = 0.026) and 6.16 (95% CI: 2.55-14.91; P < 0.001) for having HRP at baseline and follow-up compared with those with low CAD PRS.

CONCLUSIONS

Polygenic risk is strongly associated with future long-term plaque progression and HRP in patients suspected of having CAD.

SenseCare: a research platform for medical image informatics and interactive 3D visualization

Introduction

Clinical research on smart health has an increasing demand for intelligent and clinic-oriented medical image computing algorithms and platforms that support various applications. However, existing research platforms for medical image informatics have limited support for Artificial Intelligence (AI) algorithms and clinical applications.

Methods

To this end, we have developed SenseCare research platform, which is designed to facilitate translational research on intelligent diagnosis and treatment planning in various clinical scenarios. It has several appealing functions and features such as advanced 3D visualization, concurrent and efficient web-based access, fast data synchronization and high data security, multi-center deployment, support for collaborative research, etc.

Results and discussion

SenseCare provides a range of AI toolkits for different tasks, including image segmentation, registration, lesion and landmark detection from various image modalities ranging from radiology to pathology. It also facilitates the data annotation and model training processes, which makes it easier for clinical researchers to develop and deploy customized AI models. In addition, it is clinic-oriented and supports various clinical applications such as diagnosis and surgical planning for lung cancer, liver tumor, coronary artery disease, etc. By simplifying AI-based medical image analysis, SenseCare has a potential to promote clinical research in a wide range of disease diagnosis and treatment applications.

Atherosclerosis quantification and cardiovascular risk: the ISCHEMIA trial

BACKGROUND AND AIMS

The aim of this study was to determine the prognostic value of coronary computed tomography angiography (CCTA)-derived atherosclerotic plaque analysis in ISCHEMIA.

METHODS

Atherosclerosis imaging quantitative computed tomography (AI-QCT) was performed on all available baseline CCTAs to quantify plaque volume, composition, and distribution. Multivariable Cox regression was used to examine the association between baseline risk factors (age, sex, smoking, diabetes, hypertension, ejection fraction, prior coronary disease, estimated glomerular filtration rate, and statin use), number of diseased vessels, atherosclerotic plaque characteristics determined by AI-QCT, and a composite primary outcome of cardiovascular death or myocardial infarction over a median follow-up of 3.3 (interquartile range 2.2-4.4) years. The predictive value of plaque quantification over risk factors was compared in an area under the curve (AUC) analysis.

RESULTS

Analysable CCTA data were available from 3711 participants (mean age 64 years, 21% female, 79% multivessel coronary artery disease). Amongst the AI-QCT variables, total plaque volume was most strongly associated with the primary outcome (adjusted hazard ratio 1.56, 95% confidence interval 1.25-1.97 per interquartile range increase [559 mm3]; P = .001). The addition of AI-QCT plaque quantification and characterization to baseline risk factors improved the model's predictive value for the primary outcome at 6 months (AUC 0.688 vs. 0.637; P = .006), at 2 years (AUC 0.660 vs. 0.617; P = .003), and at 4 years of follow-up (AUC 0.654 vs. 0.608; P = .002). The findings were similar for the other reported outcomes.

CONCLUSIONS

In ISCHEMIA, total plaque volume was associated with cardiovascular death or myocardial infarction. In this highly diseased, high-risk population, enhanced assessment of atherosclerotic burden using AI-QCT-derived measures of plaque volume and composition modestly improved event prediction.

Cardiovascular care with digital twin technology in the era of generative artificial intelligence

Digital twins, which are in silico replications of an individual and its environment, have advanced clinical decision-making and prognostication in cardiovascular medicine. The technology enables personalized simulations of clinical scenarios, prediction of disease risk, and strategies for clinical trial augmentation. Current applications of cardiovascular digital twins have integrated multi-modal data into mechanistic and statistical models to build physiologically accurate cardiac replicas to enhance disease phenotyping, enrich diagnostic workflows, and optimize procedural planning. Digital twin technology is rapidly evolving in the setting of newly available data modalities and advances in generative artificial intelligence, enabling dynamic and comprehensive simulations unique to an individual. These twins fuse physiologic, environmental, and healthcare data into machine learning and generative models to build real-time patient predictions that can model interactions with the clinical environment to accelerate personalized patient care. This review summarizes digital twins in cardiovascular medicine and their potential future applications by incorporating new personalized data modalities. It examines the technical advances in deep learning and generative artificial intelligence that broaden the scope and predictive power of digital twins. Finally, it highlights the individual and societal challenges as well as ethical considerations that are essential to realizing the future vision of incorporating cardiology digital twins into personalized cardiovascular care.

Accuracy of deep learning in the differential diagnosis of coronary artery stenosis: a systematic review and meta-analysis

BACKGROUND

In recent years, as deep learning has received widespread attention in the field of heart disease, some studies have explored the potential of deep learning based on coronary angiography (CAG) or coronary CT angiography (CCTA) images in detecting the extent of coronary artery stenosis. However, there is still a lack of a systematic understanding of its diagnostic accuracy, impeding the advancement of intelligent diagnosis of coronary artery stenosis. Therefore, we conducted this study to review the accuracy of image-based deep learning in detecting coronary artery stenosis.

METHODS

We retrieved PubMed, Cochrane, Embase, and Web of Science until April 11, 2023. The risk of bias in the included studies was appraised using the QUADAS-2 tool. We extracted the accuracy of deep learning in the test set and performed subgroup analyses by binary and multiclass classification scenarios. We performed a subgroup analysis based on different degrees of stenosis and applied a double arcsine transformation to process the data. The analysis was done by using R.

RESULTS

Our systematic review finally included 18 studies, involving 3568 patients and 13,362 images. In the included studies, deep learning models were constructed based on CAG and CCTA. In binary classification tasks, the accuracy for detecting > 25%, > 50% and > 70% degrees of stenosis at the vessel level were 0.81 (95% CI: 0.71-0.85), 0.73 (95% CI: 0.58-0.88) and 0.61 (95% CI: 0.56-0.65), respectively. In multiclass classification tasks, the accuracy for detecting 0-25%, 25-50%, 50-70%, and 70-100% degrees of stenosis at the vessel level were 0.78 (95% CI: 0.73-0.84), 0.86 (95% CI: 0.78-0.93), 0.83 (95% CI: 0.70-0.97), and 0.70 (95% CI: 0.42-0.98), respectively.

CONCLUSIONS

Our study shows that deep learning models based on CAG and CCTA appear to be relatively accurate in diagnosing different degrees of coronary artery stenosis. However, for various degrees of stenosis, their accuracy still needs to be further improved.

Diagnostic value of artificial intelligence-assisted CTA for the assessment of atherosclerosis plaque: a systematic review and meta-analysis

Background

Artificial intelligence (AI) has increasingly been applied to computed tomography angiography (CTA) images to aid in the assessment of atherosclerotic plaque. Our aim was to explore the diagnostic accuracy of AI-assisted CTA for plaque diagnosis and classification through a systematic review and meta-analysis.

Methods

A systematic literature review was performed by searching PubMed, EMBASE, and the Cochrane Library according to PRISMA guidelines. Original studies evaluating the diagnostic accuracy of radiomics, machine-learning, or deep-learning techniques applied to CTA images for detecting stenosis, calcification, or plaque vulnerability were included. The quality and risk of bias of the included studies were evaluated using the QUADAS-2 tool. The meta-analysis was conducted using STATA software (version 17.0) to pool sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC) to determine the overall diagnostic performance.

Results

A total of 11 studies comprising 1,484 patients were included. There was low risk of bias and substantial heterogeneity. The overall pooled AUROC for atherosclerotic plaque assessment was 0.96 [95% confidence interval (CI) 0.94-0.97] across 21 trials. Of these, for ≥50% stenosis detection, the AUROC was 0.95 (95% CI 0.93-0.96) in five studies. For identifying ≥70% stenosis, the AUROC was 0.96 (95% CI 0.94-0.97) in six studies. For calcium detection, the AUROC was 0.92 (95% CI 0.90-0.94) in six studies.

Conclusion

Our meta-analysis demonstrates that AI-assisted CTA has high diagnostic accuracy for detecting stenosis and characterizing plaque composition, with optimal performance in detecting ≥70% stenosis.

Systematic Review Registration

https://www.crd.york.ac.uk/, PROSPERO, identifier (CRD42023431410).

Lipoprotein(a) and Long-Term Plaque Progression, Low-Density Plaque, and Pericoronary Inflammation

Importance

Lipoprotein(a) (Lp[a]) is a causal risk factor for cardiovascular disease; however, long-term effects on coronary atherosclerotic plaque phenotype, high-risk plaque formation, and pericoronary adipose tissue inflammation remain unknown.

Objective

To investigate the association of Lp(a) levels with long-term coronary artery plaque progression, high-risk plaque, and pericoronary adipose tissue inflammation.

Design, Setting, and Participants

This single-center prospective cohort study included 299 patients with suspected coronary artery disease (CAD) who underwent per-protocol repeated coronary computed tomography angiography (CCTA) imaging with an interscan interval of 10 years. Thirty-two patients were excluded because of coronary artery bypass grafting, resulting in a study population of 267 patients. Data for this study were collected from October 2008 to October 2022 and analyzed from March 2023 to March 2024.

Exposures

The median scan interval was 10.2 years. Lp(a) was measured at follow-up using an isoform-insensitive assay. CCTA scans were analyzed with a previously validated artificial intelligence-based algorithm (atherosclerosis imaging-quantitative computed tomography).

Main Outcome and Measures

The association between Lp(a) and change in percent plaque volumes was investigated in linear mixed-effects models adjusted for clinical risk factors. Secondary outcomes were presence of low-density plaque and presence of increased pericoronary adipose tissue attenuation at baseline and follow-up CCTA imaging.

Results

The 267 included patients had a mean age of 57.1 (SD, 7.3) years and 153 were male (57%). Patients with Lp(a) levels of 125 nmol/L or higher had twice as high percent atheroma volume (6.9% vs 3.0%; P = .01) compared with patients with Lp(a) levels less than 125 nmol/L. Adjusted for other risk factors, every doubling of Lp(a) resulted in an additional 0.32% (95% CI, 0.04-0.60) increment in percent atheroma volume during the 10 years of follow-up. Every doubling of Lp(a) resulted in an odds ratio of 1.23 (95% CI, 1.00-1.51) and 1.21 (95% CI, 1.01-1.45) for the presence of low-density plaque at baseline and follow-up, respectively. Patients with higher Lp(a) levels had increased pericoronary adipose tissue attenuation around both the right coronary artery and left anterior descending at baseline and follow-up.

Conclusions and Relevance

In this long-term prospective serial CCTA imaging study, higher Lp(a) levels were associated with increased progression of coronary plaque burden and increased presence of low-density noncalcified plaque and pericoronary adipose tissue inflammation. These data suggest an impact of elevated Lp(a) levels on coronary atherogenesis of high-risk, inflammatory, rupture-prone plaques over the long term.

Artificial Intelligence-Enabled Quantitative Coronary Plaque and Hemodynamic Analysis for Predicting Acute Coronary Syndrome

BACKGROUND

A lesion-level risk prediction for acute coronary syndrome (ACS) needs better characterization.

OBJECTIVES

This study sought to investigate the additive value of artificial intelligence-enabled quantitative coronary plaque and hemodynamic analysis (AI-QCPHA).

METHODS

Among ACS patients who underwent coronary computed tomography angiography (CTA) from 1 month to 3 years before the ACS event, culprit and nonculprit lesions on coronary CTA were adjudicated based on invasive coronary angiography. The primary endpoint was the predictability of the risk models for ACS culprit lesions. The reference model included the Coronary Artery Disease Reporting and Data System, a standardized classification for stenosis severity, and high-risk plaque, defined as lesions with ≥2 adverse plaque characteristics. The new prediction model was the reference model plus AI-QCPHA features, selected by hierarchical clustering and information gain in the derivation cohort. The model performance was assessed in the validation cohort.

RESULTS

Among 351 patients (age: 65.9 ± 11.7 years) with 2,088 nonculprit and 363 culprit lesions, the median interval from coronary CTA to ACS event was 375 days (Q1-Q3: 95-645 days), and 223 patients (63.5%) presented with myocardial infarction. In the derivation cohort (n = 243), the best AI-QCPHA features were fractional flow reserve across the lesion, plaque burden, total plaque volume, low-attenuation plaque volume, and averaged percent total myocardial blood flow. The addition of AI-QCPHA features showed higher predictability than the reference model in the validation cohort (n = 108) (AUC: 0.84 vs 0.78; P < 0.001). The additive value of AI-QCPHA features was consistent across different timepoints from coronary CTA.

CONCLUSIONS

AI-enabled plaque and hemodynamic quantification enhanced the predictability for ACS culprit lesions over the conventional coronary CTA analysis. (Exploring the Mechanism of Plaque Rupture in Acute Coronary Syndrome Using Coronary Computed Tomography Angiography and Computational Fluid Dynamics II [EMERALD-II]; NCT03591328).

Charting the Unseen: How Non-Invasive Imaging Could Redefine Cardiovascular Prevention

Cardiovascular diseases (CVDs) remain a major global health challenge, leading to significant morbidity and mortality while straining healthcare systems. Despite progress in medical treatments for CVDs, their increasing prevalence calls for a shift towards more effective prevention strategies. Traditional preventive approaches have centered around lifestyle changes, risk factors management, and medication. However, the integration of imaging methods offers a novel dimension in early disease detection, risk assessment, and ongoing monitoring of at-risk individuals. Imaging techniques such as supra-aortic trunks ultrasound, echocardiography, cardiac magnetic resonance, and coronary computed tomography angiography have broadened our understanding of the anatomical and functional aspects of cardiovascular health. These techniques enable personalized prevention strategies by providing detailed insights into the cardiac and vascular states, significantly enhancing our ability to combat the progression of CVDs. This review focuses on amalgamating current findings, technological innovations, and the impact of integrating advanced imaging modalities into cardiovascular risk prevention, aiming to offer a comprehensive perspective on their potential to transform preventive cardiology.

Revolutionizing Cardiac Imaging: A Scoping Review of Artificial Intelligence in Echocardiography, CTA, and Cardiac MRI

BACKGROUND AND INTRODUCTION

Cardiac imaging is crucial for diagnosing heart disorders. Methods like X-rays, ultrasounds, CT scans, and MRIs provide detailed anatomical and functional heart images. AI can enhance these imaging techniques with its advanced learning capabilities.

METHOD

In this scoping review, following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) Guidelines, we searched PubMed, Scopus, Web of Science, and Google Scholar using related keywords on 16 April 2024. From 3679 articles, we first screened titles and abstracts based on the initial inclusion criteria and then screened the full texts. The authors made the final selections collaboratively.

RESULT

The PRISMA chart shows that 3516 articles were initially selected for evaluation after removing duplicates. Upon reviewing titles, abstracts, and quality, 24 articles were deemed eligible for the review. The findings indicate that AI enhances image quality, speeds up imaging processes, and reduces radiation exposure with sensitivity and specificity comparable to or exceeding those of qualified radiologists or cardiologists. Further research is needed to assess AI's applicability in various types of cardiac imaging, especially in rural hospitals where access to medical doctors is limited.

CONCLUSIONS

AI improves image quality, reduces human errors and radiation exposure, and can predict cardiac events with acceptable sensitivity and specificity.

Development and Validation of a Quantitative Coronary CT Angiography Model for Diagnosis of Vessel-Specific Coronary Ischemia

BACKGROUND

Noninvasive stress testing is commonly used for detection of coronary ischemia but possesses variable accuracy and may result in excessive health care costs.

OBJECTIVES

This study aimed to derive and validate an artificial intelligence-guided quantitative coronary computed tomography angiography (AI-QCT) model for the diagnosis of coronary ischemia that integrates atherosclerosis and vascular morphology measures (AI-QCTISCHEMIA) and to evaluate its prognostic utility for major adverse cardiovascular events (MACE).

METHODS

A post hoc analysis of the CREDENCE (Computed Tomographic Evaluation of Atherosclerotic Determinants of Myocardial Ischemia) and PACIFIC-1 (Comparison of Coronary Computed Tomography Angiography, Single Photon Emission Computed Tomography [SPECT], Positron Emission Tomography [PET], and Hybrid Imaging for Diagnosis of Ischemic Heart Disease Determined by Fractional Flow Reserve) studies was performed. In both studies, symptomatic patients with suspected stable coronary artery disease had prospectively undergone coronary computed tomography angiography (CTA), myocardial perfusion imaging (MPI), SPECT, or PET, fractional flow reserve by CT (FFRCT), and invasive coronary angiography in conjunction with invasive FFR measurements. The AI-QCTISCHEMIA model was developed in the derivation cohort of the CREDENCE study, and its diagnostic performance for coronary ischemia (FFR ≤0.80) was evaluated in the CREDENCE validation cohort and PACIFIC-1. Its prognostic value was investigated in PACIFIC-1.

RESULTS

In CREDENCE validation (n = 305, age 64.4 ± 9.8 years, 210 [69%] male), the diagnostic performance by area under the receiver-operating characteristics curve (AUC) on per-patient level was 0.80 (95% CI: 0.75-0.85) for AI-QCTISCHEMIA, 0.69 (95% CI: 0.63-0.74; P < 0.001) for FFRCT, and 0.65 (95% CI: 0.59-0.71; P < 0.001) for MPI. In PACIFIC-1 (n = 208, age 58.1 ± 8.7 years, 132 [63%] male), the AUCs were 0.85 (95% CI: 0.79-0.91) for AI-QCTISCHEMIA, 0.78 (95% CI: 0.72-0.84; P = 0.037) for FFRCT, 0.89 (95% CI: 0.84-0.93; P = 0.262) for PET, and 0.72 (95% CI: 0.67-0.78; P < 0.001) for SPECT. Adjusted for clinical risk factors and coronary CTA-determined obstructive stenosis, a positive AI-QCTISCHEMIA test was associated with aHR: 7.6 (95% CI: 1.2-47.0; P = 0.030) for MACE.

CONCLUSIONS

This newly developed coronary CTA-based ischemia model using coronary atherosclerosis and vascular morphology characteristics accurately diagnoses coronary ischemia by invasive FFR and provides robust prognostic utility for MACE beyond presence of stenosis.

Prognostic value of a novel artificial intelligence-based coronary CTA-derived ischemia algorithm among patients with normal or abnormal myocardial perfusion

BACKGROUND

Among patients with obstructive coronary artery disease (CAD) on coronary computed tomography angiography (CTA), downstream positron emission tomography (PET) perfusion imaging can be performed to assess the presence of myocardial ischemia. A novel artificial-intelligence-guided quantitative computed tomography ischemia algorithm (AI-QCTischemia) aims to predict ischemia directly from coronary CTA images. We aimed to study the prognostic value of AI-QCTischemia among patients with obstructive CAD on coronary CTA and normal or abnormal downstream PET perfusion.

METHODS

AI-QCTischemia was calculated by blinded analysts among patients from the retrospective coronary CTA cohort at Turku University Hospital, Finland, with obstructive CAD on initial visual reading (diameter stenosis ≥50%) being referred for downstream 15O-H2O-PET adenosine stress perfusion imaging. All coronary arteries with their side branches were assessed by AI-QCTischemia. Absolute stress myocardial blood flow ≤2.3 ​ml/g/min in ≥2 adjacent segments was considered abnormal. The primary endpoint was death, myocardial infarction, or unstable angina pectoris. The median follow-up was 6.2 [IQR 4.4-8.3] years.

RESULTS

662 of 768 (86%) patients had conclusive AI-QCTischemia result. In patients with normal 15O-H2O-PET perfusion, an abnormal AI-QCTischemia result (n ​= ​147/331) vs. normal AI-QCTischemia result (n ​= ​184/331) was associated with a significantly higher crude and adjusted rates of the primary endpoint (adjusted HR 2.47, 95% CI 1.17-5.21, p ​= ​0.018). This did not pertain to patients with abnormal 15O-H2O-PET perfusion (abnormal AI-QCTischemia result (n ​= ​269/331) vs. normal AI-QCTischemia result (n ​= ​62/331); adjusted HR 1.09, 95% CI 0.58-2.02, p ​= ​0.794) (p-interaction ​= ​0.039).

CONCLUSION

Among patients with obstructive CAD on coronary CTA referred for downstream 15O-H2O-PET perfusion imaging, AI-QCTischemia showed incremental prognostic value among patients with preserved perfusion by 15O-H2O-PET imaging, but not among those with reduced perfusion.

Intra- and inter-reader reproducibility in quantitative coronary plaque analysis on coronary computed tomography angiography

PURPOSE

Coronary artery plaque burden, low attenuation non-calcified plaque (LAP), and pericoronary adipose tissue (PCAT) on coronary CT angiography (CCTA), have been linked to future cardiac events. The purpose of this study was to evaluate intra- and inter reader reproducibility in the quantification of coronary plaque burden and its characteristics using an artificial intelligence-enhanced semi-automated software.

MATERIALS AND METHODS

A total of 10 women and 6 men, aged 52 (IQR 49-58) underwent CCTA using a Siemens Somatom Force, Somatom Definition AS and Somatom Definition Flash scanners. Two expert readers utilized dedicated semi-automatic software (vascuCAP, Elucid Bioimaging, Wenham, MA) to assess calcified plaque, low attenuation plaque and PCAT. Readers were blinded to all clinical information and repeated their analysis at 6 weeks in random order to minimize recall bias. Data analysis was performed on the right and left coronary arteries. Intra- and inter-reader reproducibility was compared using Pearson correlation coefficient, while absolute values between analyses and readers were compared with paired non-parametric tests. This is a sub-study of the Specialized Center of Research Excellence (SCORE) clinical trial (5U54AG062334).

RESULTS

A total of 64 vessels from 16 patients were analyzed. Intra-reader Pearson correlation coefficients for calcified plaque volume, LAP volume and PCAT volumes were 0.96, 0.99 and 0.92 for reader 1 and 0.94, 0.94 and 0.95 for reader 2, respectively, (all p < 0.0001). Inter-reader Pearson correlation coefficients for calcified plaque volume, LAP and PCAT volumes were 0.92, 0.96 and 0.78, and 0.99, 0.99 and 0.93 on the second analyses, all had a p value <0.0001. There was no significant bias on the corresponding Bland-Altman analyses.

CONCLUSION

Volume measurement of coronary plaque burden and PCAT volume can be performed with high intra- and inter-reader agreement.

Application of Quantitative Assessment of Coronary Atherosclerosis by Coronary Computed Tomographic Angiography

Coronary computed tomography angiography (CCTA) has emerged as a pivotal tool for diagnosing and risk-stratifying patients with suspected coronary artery disease (CAD). Recent advancements in image analysis and artificial intelligence (AI) techniques have enabled the comprehensive quantitative analysis of coronary atherosclerosis. Fully quantitative assessments of coronary stenosis and lumen attenuation have improved the accuracy of assessing stenosis severity and predicting hemodynamically significant lesions. In addition to stenosis evaluation, quantitative plaque analysis plays a crucial role in predicting and monitoring CAD progression. Studies have demonstrated that the quantitative assessment of plaque subtypes based on CT attenuation provides a nuanced understanding of plaque characteristics and their association with cardiovascular events. Quantitative analysis of serial CCTA scans offers a unique perspective on the impact of medical therapies on plaque modification. However, challenges such as time-intensive analyses and variability in software platforms still need to be addressed for broader clinical implementation. The paradigm of CCTA has shifted towards comprehensive quantitative plaque analysis facilitated by technological advancements. As these methods continue to evolve, their integration into routine clinical practice has the potential to enhance risk assessment and guide individualized patient management. This article reviews the evolving landscape of quantitative plaque analysis in CCTA and explores its applications and limitations.

Assessment of atherosclerotic plaque burden: comparison of AI-QCT versus SIS, CAC, visual and CAD-RADS stenosis categories

This study assesses the agreement of Artificial Intelligence-Quantitative Computed Tomography (AI-QCT) with qualitative approaches to atherosclerotic disease burden codified in the multisociety 2022 CAD-RADS 2.0 Expert Consensus. 105 patients who underwent cardiac computed tomography angiography (CCTA) for chest pain were evaluated by a blinded core laboratory through FDA-cleared software (Cleerly, Denver, CO) that performs AI-QCT through artificial intelligence, analyzing factors such as % stenosis, plaque volume, and plaque composition. AI-QCT plaque volume was then staged by recently validated prognostic thresholds, and compared with CAD-RADS 2.0 clinical methods of plaque evaluation (segment involvement score (SIS), coronary artery calcium score (CACS), visual assessment, and CAD-RADS percent (%) stenosis) by expert consensus blinded to the AI-QCT core lab reads. Average age of subjects were 59 ± 11 years; 44% women, with 50% of patients at CAD-RADS 1-2 and 21% at CAD-RADS 3 and above by expert consensus. AI-QCT quantitative plaque burden staging had excellent agreement of 93% (k = 0.87 95% CI: 0.79-0.96) with SIS. There was moderate agreement between AI-QCT quantitative plaque volume and categories of visual assessment (64.4%; k = 0.488 [0.38-0.60]), and CACS (66.3%; k = 0.488 [0.36-0.61]). Agreement between AI-QCT plaque volume stage and CAD-RADS % stenosis category was also moderate. There was discordance at small plaque volumes. With ongoing validation, these results demonstrate a potential for AI-QCT as a rapid, reproducible approach to quantify total plaque burden.

Coronary computed tomography angiography for clinical practice

Coronary artery disease (CAD) is a common condition caused by the accumulation of atherosclerotic plaques. It can be classified into stable CAD or acute coronary syndrome. Coronary computed tomography angiography (CCTA) has a high negative predictive value and is used as the first examination for diagnosing stable CAD, particularly in patients at intermediate-to-high risk. CCTA is also adopted for diagnosing acute coronary syndrome, particularly in patients at low-to-intermediate risk. Myocardial ischemia does not always co-exist with coronary artery stenosis, and the positive predictive value of CCTA for myocardial ischemia is limited. However, CCTA has overcome this limitation with recent technological advancements such as CT perfusion and CT-fractional flow reserve. In addition, CCTA can be used to assess coronary artery plaques. Thus, the indications for CCTA have expanded, leading to an increased demand for radiologists. The CAD reporting and data system (CAD-RADS) 2.0 was recently proposed for standardizing CCTA reporting. This RADS evaluates and categorizes patients based on coronary artery stenosis and the overall amount of coronary artery plaque and links this to patient management. In this review, we aimed to review the major trials and guidelines for CCTA to understand its clinical role. Furthermore, we aimed to introduce the CAD-RADS 2.0 including the assessment of coronary artery stenosis, plaque, and other key findings, and highlight the steps for CCTA reporting. Finally, we aimed to present recent research trends including the perivascular fat attenuation index, artificial intelligence, and the advancements in CT technology.

Impact of atherosclerosis imaging-quantitative computed tomography on diagnostic certainty, downstream testing, coronary revascularization, and medical therapy: the CERTAIN study

AIMS

The incremental impact of atherosclerosis imaging-quantitative computed tomography (AI-QCT) on diagnostic certainty and downstream patient management is not yet known. The aim of this study was to compare the clinical utility of the routine implementation of AI-QCT versus conventional visual coronary CT angiography (CCTA) interpretation.

METHODS AND RESULTS

In this multi-centre cross-over study in 5 expert CCTA sites, 750 consecutive adult patients referred for CCTA were prospectively recruited. Blinded to the AI-QCT analysis, site physicians established patient diagnoses and plans for downstream non-invasive testing, coronary intervention, and medication management based on the conventional site assessment. Next, physicians were asked to repeat their assessments based upon AI-QCT results. The included patients had an age of 63.8 ± 12.2 years; 433 (57.7%) were male. Compared with the conventional site CCTA evaluation, AI-QCT analysis improved physician's confidence two- to five-fold at every step of the care pathway and was associated with change in diagnosis or management in the majority of patients (428; 57.1%; P < 0.001), including for measures such as Coronary Artery Disease-Reporting and Data System (CAD-RADS) (295; 39.3%; P < 0.001) and plaque burden (197; 26.3%; P < 0.001). After AI-QCT including ischaemia assessment, the need for downstream non-invasive and invasive testing was reduced by 37.1% (P < 0.001), compared with the conventional site CCTA evaluation. Incremental to the site CCTA evaluation alone, AI-QCT resulted in statin initiation/increase an aspirin initiation in an additional 28.1% (P < 0.001) and 23.0% (P < 0.001) of patients, respectively.

CONCLUSION

The use of AI-QCT improves diagnostic certainty and may result in reduced downstream need for non-invasive testing and increased rates of preventive medical therapy.

Revolutionizing Cardiology through Artificial Intelligence-Big Data from Proactive Prevention to Precise Diagnostics and Cutting-Edge Treatment-A Comprehensive Review of the Past 5 Years

BACKGROUND

Artificial intelligence (AI) can radically change almost every aspect of the human experience. In the medical field, there are numerous applications of AI and subsequently, in a relatively short time, significant progress has been made. Cardiology is not immune to this trend, this fact being supported by the exponential increase in the number of publications in which the algorithms play an important role in data analysis, pattern discovery, identification of anomalies, and therapeutic decision making. Furthermore, with technological development, there have appeared new models of machine learning (ML) and deep learning (DP) that are capable of exploring various applications of AI in cardiology, including areas such as prevention, cardiovascular imaging, electrophysiology, interventional cardiology, and many others. In this sense, the present article aims to provide a general vision of the current state of AI use in cardiology.

RESULTS

We identified and included a subset of 200 papers directly relevant to the current research covering a wide range of applications. Thus, this paper presents AI applications in cardiovascular imaging, arithmology, clinical or emergency cardiology, cardiovascular prevention, and interventional procedures in a summarized manner. Recent studies from the highly scientific literature demonstrate the feasibility and advantages of using AI in different branches of cardiology.

CONCLUSIONS

The integration of AI in cardiology offers promising perspectives for increasing accuracy by decreasing the error rate and increasing efficiency in cardiovascular practice. From predicting the risk of sudden death or the ability to respond to cardiac resynchronization therapy to the diagnosis of pulmonary embolism or the early detection of valvular diseases, AI algorithms have shown their potential to mitigate human error and provide feasible solutions. At the same time, limits imposed by the small samples studied are highlighted alongside the challenges presented by ethical implementation; these relate to legal implications regarding responsibility and decision making processes, ensuring patient confidentiality and data security. All these constitute future research directions that will allow the integration of AI in the progress of cardiology.

Beyond early LDL cholesterol lowering to prevent coronary atherosclerosis in familial hypercholesterolaemia

AIMS

Familial hypercholesterolaemia (FH) patients are subjected to a high lifetime exposure to low density lipoprotein cholesterol (LDL-C), despite use of lipid-lowering therapy (LLT). This study aimed to quantify the extent of subclinical atherosclerosis and to evaluate the association between lifetime cumulative LDL-C exposure and coronary atherosclerosis in young FH patients.

METHODS AND RESULTS

Familial hypercholesterolaemia patients, divided into a subgroup of early treated (LLT initiated <25 years) and late treated (LLT initiated ≥25 years) patients, and an age- and sex-matched unaffected control group, underwent coronary CT angiography (CCTA) with artificial intelligence-guided analysis. Ninety genetically diagnosed FH patients and 45 unaffected volunteers (mean age 41 ± 3 years, 51 (38%) female) were included. Familial hypercholesterolaemia patients had higher cumulative LDL-C exposure (181 ± 54 vs. 105 ± 33 mmol/L ∗ years) and higher prevalence of coronary plaque compared with controls (46 [51%] vs. 10 [22%], OR 3.66 [95%CI 1.62-8.27]). Every 75 mmol/L ∗ years cumulative exposure to LDL-C was associated with a doubling in per cent atheroma volume (total plaque volume divided by total vessel volume). Early treated patients had a modestly lower cumulative LDL-C exposure compared with late treated FH patients (167 ± 41 vs. 194 ± 61 mmol/L ∗ years; P = 0.045), without significant difference in coronary atherosclerosis. Familial hypercholesterolaemia patients with above-median cumulative LDL-C exposure had significantly higher plaque prevalence (OR 3.62 [95%CI 1.62-8.27]; P = 0.001), compared with patients with below-median exposure.

CONCLUSION

Lifetime exposure to LDL-C determines coronary plaque burden in FH, underlining the need of early as well as potent treatment initiation. Periodic CCTA may offer a unique opportunity to monitor coronary atherosclerosis and personalize treatment in FH.

The Role of Artificial Intelligence in Cardiac Imaging

Artificial intelligence (AI) is having a significant impact in medical imaging, advancing almost every aspect of the field, from image acquisition and postprocessing to automated image analysis with outreach toward supporting decision making. Noninvasive cardiac imaging is one of the main and most exciting fields for AI development. The aim of this review is to describe the main applications of AI in cardiac imaging, including CT and MR imaging, and provide an overview of recent advancements and available clinical applications that can improve clinical workflow, disease detection, and prognostication in cardiac disease.

Prognostic value of a novel artificial intelligence-based coronary computed tomography angiography-derived ischaemia algorithm for patients with suspected coronary artery disease

AIMS

Coronary computed tomography angiography (CTA) imaging is used to diagnose patients with suspected coronary artery disease (CAD). A novel artificial intelligence-guided quantitative computed tomography ischaemia algorithm (AI-QCTischaemia) aims to identify myocardial ischaemia directly from CTA images and may be helpful to improve risk stratification. The aims were to investigate (i) the prognostic value of AI-QCTischaemia amongst symptomatic patients with suspected CAD entering diagnostic imaging with coronary CTA and (ii) the prognostic value of AI-QCTischaemia separately amongst patients with no/non-obstructive CAD (≤50% visual diameter stenosis) and obstructive CAD (>50% visual diameter stenosis).

METHODS AND RESULTS

For this cohort study, AI-QCTischaemia was calculated by blinded analysts amongst patients with suspected CAD undergoing coronary CTA. The primary endpoint was the composite of death, myocardial infarction (MI), or unstable angina pectoris (uAP) (median follow-up 6.9 years). A total of 1880/2271 (83%) patients had conclusive AI-QCTischaemia result. Patients with an abnormal AI-QCTischaemia result (n = 509/1880) vs. patients with a normal AI-QCTischaemia result (n = 1371/1880) had significantly higher crude and adjusted rates of the primary endpoint [adjusted hazard ratio (HRadj) 1.96, 95% confidence interval (CI) 1.46-2.63, P < 0.001; covariates: age/sex/hypertension/diabetes/smoking/typical angina]. An abnormal AI-QCTischaemia result was associated with significantly higher crude and adjusted rates of the primary endpoint amongst patients with no/non-obstructive CAD (n = 1373/1847) (HRadj 1.81, 95% CI 1.09-3.00, P = 0.022), but not amongst those with obstructive CAD (n = 474/1847) (HRadj 1.26, 95% CI 0.75-2.12, P = 0.386) (P-interaction = 0.032).

CONCLUSION

Amongst patients with suspected CAD, an abnormal AI-QCTischaemia result was associated with a two-fold increased adjusted rate of long-term death, MI, or uAP. AI-QCTischaemia may be useful to improve risk stratification, especially amongst patients with no/non-obstructive CAD on coronary CTA.

Implications of Bias in Artificial Intelligence: Considerations for Cardiovascular Imaging

PURPOSE OF REVIEW

Bias in artificial intelligence (AI) models can result in unintended consequences. In cardiovascular imaging, biased AI models used in clinical practice can negatively affect patient outcomes. Biased AI models result from decisions made when training and evaluating a model. This paper is a comprehensive guide for AI development teams to understand assumptions in datasets and chosen metrics for outcome/ground truth, and how this translates to real-world performance for cardiovascular disease (CVD).

RECENT FINDINGS

CVDs are the number one cause of mortality worldwide; however, the prevalence, burden, and outcomes of CVD vary across gender and race. Several biomarkers are also shown to vary among different populations and ethnic/racial groups. Inequalities in clinical trial inclusion, clinical presentation, diagnosis, and treatment are preserved in health data that is ultimately used to train AI algorithms, leading to potential biases in model performance. Despite the notion that AI models themselves are biased, AI can also help to mitigate bias (e.g., bias auditing tools). In this review paper, we describe in detail implicit and explicit biases in the care of cardiovascular disease that may be present in existing datasets but are not obvious to model developers. We review disparities in CVD outcomes across different genders and race groups, differences in treatment of historically marginalized groups, and disparities in clinical trials for various cardiovascular diseases and outcomes. Thereafter, we summarize some CVD AI literature that shows bias in CVD AI as well as approaches that AI is being used to mitigate CVD bias.

AI-Guided Quantitative Plaque Staging Predicts Long-Term Cardiovascular Outcomes in Patients at Risk for Atherosclerotic CVD

BACKGROUND

The recent development of artificial intelligence-guided quantitative coronary computed tomography angiography analysis (AI-QCT) has enabled rapid analysis of atherosclerotic plaque burden and characteristics.

OBJECTIVES

This study set out to investigate the 10-year prognostic value of atherosclerotic burden derived from AI-QCT and to compare the spectrum of plaque to manually assessed coronary computed tomography angiography (CCTA), coronary artery calcium scoring (CACS), and clinical risk characteristics.

METHODS

This was a long-term follow-up study of 536 patients referred for suspected coronary artery disease. CCTA scans were analyzed with AI-QCT and plaque burden was classified with a plaque staging system (stage 0: 0% percentage atheroma volume [PAV]; stage 1: >0%-5% PAV; stage 2: >5%-15% PAV; stage 3: >15% PAV). The primary major adverse cardiac event (MACE) outcome was a composite of nonfatal myocardial infarction, nonfatal stroke, coronary revascularization, and all-cause mortality.

RESULTS

The mean age at baseline was 58.6 years and 297 patients (55%) were male. During a median follow-up of 10.3 years (IQR: 8.6-11.5 years), 114 patients (21%) experienced the primary outcome. Compared to stages 0 and 1, patients with stage 3 PAV and percentage of noncalcified plaque volume of >7.5% had a more than 3-fold (adjusted HR: 3.57; 95% CI 2.12-6.00; P < 0.001) and 4-fold (adjusted HR: 4.37; 95% CI: 2.51-7.62; P < 0.001) increased risk of MACE, respectively. Addition of AI-QCT improved a model with clinical risk factors and CACS at different time points during follow-up (10-year AUC: 0.82 [95% CI: 0.78-0.87] vs 0.73 [95% CI: 0.68-0.79]; P < 0.001; net reclassification improvement: 0.21 [95% CI: 0.09-0.38]). Furthermore, AI-QCT achieved an improved area under the curve compared to Coronary Artery Disease Reporting and Data System 2.0 (10-year AUC: 0.78; 95% CI: 0.73-0.83; P = 0.023) and manual QCT (10-year AUC: 0.78; 95% CI: 0.73-0.83; P = 0.040), although net reclassification improvement was modest (0.09 [95% CI: -0.02 to 0.29] and 0.04 [95% CI: -0.05 to 0.27], respectively).

CONCLUSIONS

Through 10-year follow-up, AI-QCT plaque staging showed important prognostic value for MACE and showed additional discriminatory value over clinical risk factors, CACS, and manual guideline-recommended CCTA assessment.

Value Creation Through Artificial Intelligence and Cardiovascular Imaging: A Scientific Statement From the American Heart Association

Multiple applications for machine learning and artificial intelligence (AI) in cardiovascular imaging are being proposed and developed. However, the processes involved in implementing AI in cardiovascular imaging are highly diverse, varying by imaging modality, patient subtype, features to be extracted and analyzed, and clinical application. This article establishes a framework that defines value from an organizational perspective, followed by value chain analysis to identify the activities in which AI might produce the greatest incremental value creation. The various perspectives that should be considered are highlighted, including clinicians, imagers, hospitals, patients, and payers. Integrating the perspectives of all health care stakeholders is critical for creating value and ensuring the successful deployment of AI tools in a real-world setting. Different AI tools are summarized, along with the unique aspects of AI applications to various cardiac imaging modalities, including cardiac computed tomography, magnetic resonance imaging, and positron emission tomography. AI is applicable and has the potential to add value to cardiovascular imaging at every step along the patient journey, from selecting the more appropriate test to optimizing image acquisition and analysis, interpreting the results for classification and diagnosis, and predicting the risk for major adverse cardiac events.

Distal-vessel fractional flow reserve by computed tomography to monitor epicardial coronary artery disease

AIMS

Coronary computed tomography angiography (CTA) and fractional flow reserve by computed tomography (FFR-CT) are increasingly utilized to characterize coronary artery disease (CAD). We evaluated the feasibility of distal-vessel FFR-CT as an integrated measure of epicardial CAD that can be followed serially, assessed the CTA parameters that correlate with distal-vessel FFR-CT, and determined the combination of clinical and CTA parameters that best predict distal-vessel FFR-CT and distal-vessel FFR-CT changes.

METHODS AND RESULTS

Patients (n = 71) who underwent serial CTA scans at ≥2 years interval (median = 5.2 years) over a 14-year period were included in this retrospective study. Coronary arteries were analysed blindly using artificial intelligence-enabled quantitative coronary CTA. Two investigators jointly determined the anatomic location and corresponding distal-vessel FFR-CT values at CT1 and CT2. A total of 45.3% had no significant change, 27.8% an improvement, and 26.9% a worsening in distal-vessel FFR-CT at CT2. Stepwise multiple logistic regression analysis identified a four-parameter model consisting of stenosis diameter ratio, lumen volume, low density plaque volume, and age, that best predicted distal-vessel FFR-CT ≤ 0.80 with an area under the curve (AUC) = 0.820 at CT1 and AUC = 0.799 at CT2. Improvement of distal-vessel FFR-CT was captured by a decrease in high-risk plaque and increases in lumen volume and remodelling index (AUC = 0.865), whereas increases in stenosis diameter ratio, medium density calcified plaque volume, and total cholesterol presaged worsening of distal-vessel FFR-CT (AUC = 0.707).

CONCLUSION

Distal-vessel FFR-CT permits the integrative assessment of epicardial atherosclerotic plaque burden in a vessel-specific manner and can be followed serially to determine changes in global CAD.

Artificial intelligence coronary computed tomography, coronary computed tomography angiography using fractional flow reserve, and physician visual interpretation in the per-vessel prediction of abnormal invasive adenosine fractional flow reserve

Aims

A comparison of diagnostic performance comparing AI-QCTISCHEMIA, coronary computed tomography angiography using fractional flow reserve (CT-FFR), and physician visual interpretation on the prediction of invasive adenosine FFR have not been evaluated. Furthermore, the coronary plaque characteristics impacting these tests have not been assessed.

Methods and results

In a single centre, 43-month retrospective review of 442 patients referred for coronary computed tomography angiography and CT-FFR, 44 patients with CT-FFR had 54 vessels assessed using intracoronary adenosine FFR within 60 days. A comparison of the diagnostic performance among these three techniques for the prediction of FFR ≤ 0.80 was reported. The mean age of the study population was 65 years, 76.9% were male, and the median coronary artery calcium was 654. When analysing the per-vessel ischaemia prediction, AI-QCTISCHEMIA had greater specificity, positive predictive value (PPV), diagnostic accuracy, and area under the curve (AUC) vs. CT-FFR and physician visual interpretation CAD-RADS. The AUC for AI-QCTISCHEMIA was 0.91 vs. 0.76 for CT-FFR and 0.62 for CAD-RADS ≥ 3. Plaque characteristics that were different in false positive vs. true positive cases for AI-QCTISCHEMIA were max stenosis diameter % (54% vs. 67%, P < 0.01); for CT-FFR were maximum stenosis diameter % (40% vs. 65%, P < 0.001), total non-calcified plaque (9% vs. 13%, P < 0.01); and for physician visual interpretation CAD-RADS ≥ 3 were total non-calcified plaque (8% vs. 12%, P < 0.01), lumen volume (681 vs. 510 mm3, P = 0.02), maximum stenosis diameter % (40% vs. 62%, P < 0.001), total plaque (19% vs. 33%, P = 0.002), and total calcified plaque (11% vs. 22%, P = 0.003).

Conclusion

Regarding per-vessel prediction of FFR ≤ 0.8, AI-QCTISCHEMIA revealed greater specificity, PPV, accuracy, and AUC vs. CT-FFR and physician visual interpretation CAD-RADS ≥ 3.

Use of coronary artery calcium score and coronary CT angiography to guide cardiovascular prevention and treatment

Currently, cardiovascular risk stratification to guide preventive therapy relies on clinical scores based on cardiovascular risk factors. However, the discriminative power of these scores is relatively modest. The use of coronary artery calcium score (CACS) and coronary CT angiography (CCTA) has surfaced as methods for enhancing the estimation of risk and potentially providing insights for personalized treatment in individual patients. CACS improves overall cardiovascular risk prediction and may be used to improve the yield of statin therapy in primary prevention, and possibly identify patients with a favorable risk/benefit relationship for antiplatelet therapies. CCTA holds promise to guide anti-atherosclerotic therapies and to monitor individual response to these treatments by assessing individual plaque features, quantifying total plaque volume and composition, and assessing peri-coronary adipose tissue. In this review, we aim to summarize current evidence regarding the use of CACS and CCTA for guiding lipid-lowering and antiplatelet therapy and discuss the possibility of using plaque burden and plaque phenotyping to monitor response to anti-atherosclerotic therapies.

Enhancing coronary artery plaque analysis via artificial intelligence-driven cardiovascular computed tomography

Coronary computed tomography angiography (CCTA) is a noninvasive imaging modality of cardiac structures and vasculature considered comparable to invasive coronary angiography for the evaluation of coronary artery disease (CAD) in several major cardiovascular guidelines. Conventional image acquisition, processing, and analysis of CCTA imaging have progressed significantly in the past decade through advances in technology, computation, and engineering. However, the advent of artificial intelligence (AI)-driven analysis of CCTA further drives past the limitations of conventional CCTA, allowing for greater achievements in speed, consistency, accuracy, and safety. AI-driven CCTA (AI-CCTA) has achieved a significant reduction in radiation exposure for patients, allowing for high-quality scans with sub-millisievert radiation doses. AI-CCTA has demonstrated comparable accuracy and consistency in manual coronary artery calcium scoring against expert human readers. An advantage over invasive coronary angiography, which provides luminal information only, CCTA allows for plaque characterization, providing detailed information on the quality of plaque and offering further prognosticative value for the management of CAD. Combined with AI, many recent studies demonstrate the efficacy, accuracy, efficiency, and precision of AI-driven analysis of CCTA imaging for the evaluation of CAD, including assessing degree stenosis, adverse plaque characteristics, and CT fractional flow reserve. The limitations of AI-CCTA include its early phase in investigation, the need for further improvements in AI modeling, possible medicolegal implications, and the need for further large-scale validation studies. Despite these limitations, AI-CCTA represents an important opportunity for improving cardiovascular care in an increasingly advanced and data-driven world of modern medicine.

Rationale and design of the CONFIRM2 (Quantitative COroNary CT Angiography Evaluation For Evaluation of Clinical Outcomes: An InteRnational, Multicenter Registry) study

BACKGROUND

In the last 15 years, large registries and several randomized clinical trials have demonstrated the diagnostic and prognostic value of coronary computed tomography angiography (CCTA). Advances in CT scanner technology and developments of analytic tools now enable accurate quantification of coronary artery disease (CAD), including total coronary plaque volume and low attenuation plaque volume. The primary aim of CONFIRM2, (Quantitative COroNary CT Angiography Evaluation For Evaluation of Clinical Outcomes: An InteRnational, Multicenter Registry) is to perform comprehensive quantification of CCTA findings, including coronary, non-coronary cardiac, non-cardiac vascular, non-cardiac findings, and relate them to clinical variables and cardiovascular clinical outcomes.

DESIGN

CONFIRM2 is a multicenter, international observational cohort study designed to evaluate multidimensional associations between quantitative phenotype of cardiovascular disease and future adverse clinical outcomes in subjects undergoing clinically indicated CCTA. The targeted population is heterogenous and includes patients undergoing CCTA for atherosclerotic evaluation, valvular heart disease, congenital heart disease or pre-procedural evaluation. Automated software will be utilized for quantification of coronary plaque, stenosis, vascular morphology and cardiac structures for rapid and reproducible tissue characterization. Up to 30,000 patients will be included from up to 50 international multi-continental clinical CCTA sites and followed for 3-4 years.

SUMMARY

CONFIRM2 is one of the largest CCTA studies to establish the clinical value of a multiparametric approach to quantify the phenotype of cardiovascular disease by CCTA using automated imaging solutions.

Determination of lipid-rich plaques by artificial intelligence-enabled quantitative computed tomography using near-infrared spectroscopy as reference

BACKGROUND AND AIMS

Artificial intelligence quantitative CT (AI-QCT) determines coronary plaque morphology with high efficiency and accuracy. Yet, its performance to quantify lipid-rich plaque remains unclear. This study investigated the performance of AI-QCT for the detection of low-density noncalcified plaque (LD-NCP) using near-infrared spectroscopy-intravascular ultrasound (NIRS-IVUS).

METHODS

The INVICTUS Registry is a multi-center registry enrolling patients undergoing clinically indicated coronary CT angiography and IVUS, NIRS-IVUS, or optical coherence tomography. We assessed the performance of various Hounsfield unit (HU) and volume thresholds of LD-NCP using maxLCBI4mm ≥ 400 as the reference standard and the correlation of the vessel area, lumen area, plaque burden, and lesion length between AI-QCT and IVUS.

RESULTS

This study included 133 atherosclerotic plaques from 47 patients who underwent coronary CT angiography and NIRS-IVUS The area under the curve of LD-NCP<30HU was 0.97 (95% confidence interval [CI]: 0.93-1.00] with an optimal volume threshold of 2.30 mm3. Accuracy, sensitivity, and specificity were 94% (95% CI: 88-96%], 93% (95% CI: 76-98%), and 94% (95% CI: 88-98%), respectively, using <30 HU and 2.3 mm3, versus 42%, 100%, and 27% using <30 HU and >0 mm3 volume of LD-NCP (p < 0.001 for accuracy and specificity). AI-QCT strongly correlated with IVUS measurements; vessel area (r2 = 0.87), lumen area (r2 = 0.87), plaque burden (r2 = 0.78) and lesion length (r2 = 0.88), respectively.

CONCLUSIONS

AI-QCT demonstrated excellent diagnostic performance in detecting significant LD-NCP using maxLCBI4mm ≥ 400 as the reference standard. Additionally, vessel area, lumen area, plaque burden, and lesion length derived from AI-QCT strongly correlated with respective IVUS measurements.

Artificial Intelligence-based Coronary Stenosis Quantification at Coronary CT Angiography versus Quantitative Coronary Angiography

Purpose To evaluate the performance of a new artificial intelligence (AI)-based tool by comparing the quantified stenosis severity at coronary CT angiography (CCTA) with a reference standard derived from invasive quantitative coronary angiography (QCA). Materials and Methods This secondary, post hoc analysis included 120 participants (mean age, 59.7 years ± 10.8 [SD]; 73 [60.8%] men, 47 [39.2%] women) from three large clinical trials (AFFECTS, P3, REFINE) who underwent CCTA and invasive coronary angiography with QCA. Quantitative analysis of coronary stenosis severity at CCTA was performed using an AI-based coronary stenosis quantification (AI-CSQ) software service. Blinded comparison between QCA and AI-CSQ was measured on a per-vessel and per-patient basis. Results The per-vessel AI-CSQ diagnostic sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were 80%, 88%, 86%, 65%, and 94%, respectively, for diameter stenosis (DS) 50% or greater; and 78%, 92%, 91%, 47%, and 98%, respectively, for DS 70% or greater. The areas under the receiver operating characteristic curve (AUCs) to predict DS of 50% or greater and 70% or greater on a per-vessel basis were 0.92 (95% CI: 0.88, 0.95; P < .001) and 0.93 (95% CI: 0.89, 0.97; P < .001), respectively. The AUCs to predict DS of 50% or greater and 70% or greater on a per-patient basis were 0.93 (95% CI: 0.88, 0.97; P < .001) and 0.88 (95% CI: 0.81, 0.94; P < .001), respectively. Conclusion AI-CSQ at CCTA demonstrated a high diagnostic performance compared with QCA both on a per-patient and per-vessel basis, with high sensitivity for stenosis detection. Keywords: CT Angiography, Cardiac, Coronary Arteries Supplemental material is available for this article. Published under a CC BY 4.0 license.