Abstract
Aims
Our aim was to evaluate the performance of machine learning (ML), integrating clinical parameters with coronary artery calcium (CAC), and automated epicardial adipose tissue (EAT) ...quantification, for the prediction of long-term risk of myocardial infarction (MI) and cardiac death in asymptomatic subjects.
Methods and results
Our study included 1912 asymptomatic subjects 1117 (58.4%) male, age: 55.8 ± 9.1 years from the prospective EISNER trial with long-term follow-up after CAC scoring. EAT volume and density were quantified using a fully automated deep learning method. ML extreme gradient boosting was trained using clinical co-variates, plasma lipid panel measurements, risk factors, CAC, aortic calcium, and automated EAT measures, and validated using repeated 10-fold cross validation. During mean follow-up of 14.5 ± 2 years, 76 events of MI and/or cardiac death occurred. ML obtained a significantly higher AUC than atherosclerotic cardiovascular disease (ASCVD) risk and CAC score for predicting events (ML: 0.82; ASCVD: 0.77; CAC: 0.77, P < 0.05 for all). Subjects with a higher ML score (by Youden’s index) had high hazard of suffering events (HR: 10.38, P < 0.001); the relationships persisted in multivariable analysis including ASCVD-risk and CAC measures (HR: 2.94, P = 0.005). Age, ASCVD-risk, and CAC were prognostically important for both genders. Systolic blood pressure was more important than cholesterol in women, and the opposite in men.
Conclusions
In this prospective study, machine learning used to integrate clinical and quantitative imaging-based variables significantly improves prediction of MI and cardiac death compared with standard clinical risk assessment. Following further validation, such a personalized paradigm could potentially be used to improve cardiovascular risk assessment.
Increased attenuation of pericoronary adipose tissue (PCAT) around the proximal right coronary artery (RCA) from coronary computed tomography angiography (CTA) has been shown to be associated with ...coronary inflammation and improved prediction of cardiac death over plaque features. Our aim was to investigate whether PCAT CT attenuation is related to progression of coronary plaque burden.
We analysed CTA studies of 111 stable patients (age 59.2 ± 9.8 years, 77% male) who underwent sequential CTA (3.4 ± 1.6 years between scans) with identical acquisition protocols. Total plaque (TP), calcified plaque (CP), non-calcified plaque (NCP), and low-density non-calcified plaque (LD-NCP) volumes and corresponding burden (plaque volume × 100%/vessel volume) were quantified using semi-automated software. PCAT CT attenuation (HU) was measured around the proximal RCA, the most standardized method for PCAT analysis. Patients with an increase in NCP burden (n = 51) showed an increase in PCAT attenuation, whereas patients with a decrease in NCP burden (n = 60) showed a decrease {4.4 95% confidence interval (CI) 2.6-6.2 vs. -2.78 (95% CI -4.6 to -1.0) HU, P < 0.0001}. Changes in PCAT attenuation correlated with changes in the burden of NCP (r = 0.55, P < 0.001) and LD-NCP (r = 0.24, P = 0.01); but not CP burden (P = 0.3). Increased baseline PCAT attenuation ≥-75 HU was independently associated with increase in NCP (odds ratio 3.07, 95% CI 1.4-7.0; P < 0.008) and TP burden on follow-up CTA.
PCAT attenuation measured from routine CTA is related to the progression of NCP and TP burden. This imaging biomarker may help to identify patients at increased risk of high-risk plaque progression and allow monitoring of beneficial changes from medical therapy.
Epicardial adipose tissue (EAT) is a visceral fat deposit related to coronary artery disease. Fully automated quantification of EAT volume in clinical routine could be a timesaving and reliable tool ...for cardiovascular risk assessment. We propose a new fully automated deep learning framework for EAT and thoracic adipose tissue (TAT) quantification from non-contrast coronary artery calcium computed tomography (CT) scans. The first multi-task convolutional neural network (ConvNet) is used to determine heart limits and perform segmentation of heart and adipose tissues. The second ConvNet, combined with a statistical shape model, allows for pericardium detection. EAT and TAT segmentations are then obtained from outputs of both ConvNets. We evaluate the performance of the method on CT data sets from 250 asymptomatic individuals. Strong agreement between automatic and expert manual quantification is obtained for both EAT and TAT with median Dice score coefficients of 0.823 (inter-quartile range (IQR): 0.779-0.860) and 0.905 (IQR: 0.862-0.928), respectively; with excellent correlations of 0.924 and 0.945 for EAT and TAT volumes. Computations are performed in <;26 s on a standard personal computer for one CT scan. Therefore, the proposed method represents a tool for rapid fully automated quantification of adipose tissue and may improve cardiovascular risk stratification in patients referred for routine CT calcium scans.
Combined analysis of SPECT myocardial perfusion imaging (MPI) performed with a solid-state camera on patients in 2 positions (semiupright, supine) is routinely used to mitigate attenuation artifacts. ...We evaluated the prediction of obstructive disease from combined analysis of semiupright and supine stress MPI by deep learning (DL) as compared with standard combined total perfusion deficit (TPD).
1,160 patients without known coronary artery disease (64% male) were studied. Patients underwent stress
Tc-sestamibi MPI with new-generation solid-state SPECT scanners in 4 different centers. All patients had on-site clinical reads and invasive coronary angiography correlations within 6 mo of MPI. Obstructive disease was defined as at least 70% narrowing of the 3 major coronary arteries and at least 50% for the left main coronary artery. Images were quantified at Cedars-Sinai. The left ventricular myocardium was segmented using standard clinical nuclear cardiology software. The contour placement was verified by an experienced technologist. Combined stress TPD was computed using sex- and camera-specific normal limits. DL was trained using polar distributions of normalized radiotracer counts, hypoperfusion defects, and hypoperfusion severities and was evaluated for prediction of obstructive disease in a novel leave-one-center-out cross-validation procedure equivalent to external validation. During the validation procedure, 4 DL models were trained using data from 3 centers and then evaluated on the 1 center left aside. Predictions for each center were merged to have an overall estimation of the multicenter performance.
718 (62%) patients and 1,272 of 3,480 (37%) arteries had obstructive disease. The area under the receiver operating characteristics curve for prediction of disease on a per-patient and per-vessel basis by DL was higher than for combined TPD (per-patient, 0.81 vs. 0.78; per-vessel, 0.77 vs. 0.73;
< 0.001). With the DL cutoff set to exhibit the same specificity as the standard cutoff for combined TPD, per-patient sensitivity improved from 61.8% (TPD) to 65.6% (DL) (
< 0.05), and per-vessel sensitivity improved from 54.6% (TPD) to 59.1% (DL) (
< 0.01). With the threshold matched to the specificity of a normal clinical read (56.3%), DL had a sensitivity of 84.8%, versus 82.6% for an on-site clinical read (
= 0.3).
DL improves automatic interpretation of MPI as compared with current quantitative methods.
Epicardial adipose tissue (EAT) volume (cm
) and attenuation (Hounsfield units) may predict major adverse cardiovascular events (MACE). We aimed to evaluate the prognostic value of fully automated ...deep learning-based EAT volume and attenuation measurements quantified from noncontrast cardiac computed tomography.
Our study included 2068 asymptomatic subjects (56±9 years, 59% male) from the EISNER trial (Early Identification of Subclinical Atherosclerosis by Noninvasive Imaging Research) with long-term follow-up after coronary artery calcium measurement. EAT volume and mean attenuation were quantified using automated deep learning software from noncontrast cardiac computed tomography. MACE was defined as myocardial infarction, late (>180 days) revascularization, and cardiac death. EAT measures were compared to coronary artery calcium score and atherosclerotic cardiovascular disease risk score for MACE prediction.
At 14±3 years, 223 subjects suffered MACE. Increased EAT volume and decreased EAT attenuation were both independently associated with MACE. Atherosclerotic cardiovascular disease risk score, coronary artery calcium, and EAT volume were associated with increased risk of MACE (hazard ratio 95%CI: 1.03 1.01-1.04; 1.25 1.19-1.30; and 1.35 1.07-1.68,
<0.01 for all) and EAT attenuation was inversely associated with MACE (hazard ratio, 0.83 95% CI, 0.72-0.96;
=0.01), with corresponding Harrell C statistic of 0.76. MACE risk progressively increased with EAT volume ≥113 cm
and coronary artery calcium ≥100 AU and was highest in subjects with both (
<0.02 for all). In 1317 subjects, EAT volume was correlated with inflammatory biomarkers C-reactive protein, myeloperoxidase, and adiponectin reduction; EAT attenuation was inversely related to these biomarkers.
Fully automated EAT volume and attenuation quantification by deep learning from noncontrast cardiac computed tomography can provide prognostic value for the asymptomatic patient, without additional imaging or physician interaction.
We investigated whether epicardial adipose tissue (EAT) volume and density are related to early atherosclerosis, plaque inflammation and major adverse cardiac events (MACE, cardiac death and ...myocardial infarction) in asymptomatic subjects.
EAT volume and density were quantified from non-contrast cardiac CT in 456 asymptomatic individuals (age 60.3 ± 8.3; 68% with CCS>0) from the prospective EISNER trial. EAT volume and density were examined in relation to coronary calcium score (CCS), inflammatory biomarkers and MACE.
EAT volume was higher and EAT density lower in subjects with coronary calcium compared to subjects without 89 vs 74 cm3, p < 0.001 -76.9 vs -75.7 HU,p = 0.024. EAT volume was lowest in individuals with no coronary calcium and was significant higher in subjects with early atherosclerosis (CCS 1-99) 74 vs 87 cm3,p = 0.016 and in subjects with more advanced atherosclerosis (CCS≥100) 89 cm3,p = 0.002). EAT volume was independently related to serum levels of PAI-1, and MCP-1 and inversely related to adiponectin and HDL-cholesterol (p < 0.05). EAT density was inversely related to PAI-1 and LDL-cholesterol and positively associated to adiponectin, sICAM-1 and HDL-cholesterol (p < 0.05). EAT density was more significantly associated with MACE (HR 0.8, 95%CI:0.7–0.98), p = 0.029 than EAT volume or CCS.
EAT volume was higher and density lower in subjects with coronary calcium compared to subjects with CCS = 0, with similar EAT volume in CCS<100 and CCS≥100. Lower EAT density and increased EAT volume were associated with coronary calcification, serum levels of plaque inflammatory markers and MACE, suggesting that dysfunctional EAT may be linked to early plaque formation and inflammation.
The study evaluated the automatic prediction of obstructive disease from myocardial perfusion imaging (MPI) by deep learning as compared with total perfusion deficit (TPD).
Deep convolutional neural ...networks trained with a large multicenter population may provide improved prediction of per-patient and per-vessel coronary artery disease from single-photon emission computed tomography MPI.
A total of 1,638 patients (67% men) without known coronary artery disease, undergoing stress
Tc-sestamibi or tetrofosmin MPI with new generation solid-state scanners in 9 different sites, with invasive coronary angiography performed within 6 months of MPI, were studied. Obstructive disease was defined as ≥70% narrowing of coronary arteries (≥50% for left main artery). Left ventricular myocardium was segmented using clinical nuclear cardiology software and verified by an expert reader. Stress TPD was computed using sex- and camera-specific normal limits. Deep learning was trained using raw and quantitative polar maps and evaluated for prediction of obstructive stenosis in a stratified 10-fold cross-validation procedure.
A total of 1,018 (62%) patients and 1,797 of 4,914 (37%) arteries had obstructive disease. Area under the receiver-operating characteristic curve for disease prediction by deep learning was higher than for TPD (per patient: 0.80 vs. 0.78; per vessel: 0.76 vs. 0.73: p < 0.01). With deep learning threshold set to the same specificity as TPD, per-patient sensitivity improved from 79.8% (TPD) to 82.3% (deep learning) (p < 0.05), and per-vessel sensitivity improved from 64.4% (TPD) to 69.8% (deep learning) (p < 0.01).
Deep learning has the potential to improve automatic interpretation of MPI as compared with current clinical methods.
Abstract
Aims
To optimize per-vessel prediction of early coronary revascularization (ECR) within 90 days after fast single-photon emission computed tomography (SPECT) myocardial perfusion imaging ...(MPI) using machine learning (ML) and introduce a method for a patient-specific explanation of ML results in a clinical setting.
Methods and results
A total of 1980 patients with suspected coronary artery disease (CAD) underwent stress/rest 99mTc-sestamibi/tetrofosmin MPI with new-generation SPECT scanners were included. All patients had invasive coronary angiography within 6 months after SPECT MPI. ML utilized 18 clinical, 9 stress test, and 28 imaging variables to predict per-vessel and per-patient ECR with 10-fold cross-validation. Area under the receiver operator characteristics curve (AUC) of ML was compared with standard quantitative analysis total perfusion deficit (TPD) and expert interpretation. ECR was performed in 958 patients (48%). Per-vessel, the AUC of ECR prediction by ML (AUC 0.79, 95% confidence interval (CI) 0.77, 0.80) was higher than by regional stress TPD (0.71, 0.70, 0.73), combined-view stress TPD (AUC 0.71, 95% CI 0.69, 0.72), or ischaemic TPD (AUC 0.72, 95% CI 0.71, 0.74), all P < 0.001. Per-patient, the AUC of ECR prediction by ML (AUC 0.81, 95% CI 0.79, 0.83) was higher than that of stress TPD, combined-view TPD, and ischaemic TPD, all P < 0.001. ML also outperformed nuclear cardiologists’ expert interpretation of MPI for the prediction of early revascularization performance. A method to explain ML prediction for an individual patient was also developed.
Conclusion
In patients with suspected CAD, the prediction of ECR by ML outperformed automatic MPI quantitation by TPDs (per-vessel and per-patient) or nuclear cardiologists’ expert interpretation (per-patient).
Abstract
Aims
Adverse plaque characteristics determined by coronary computed tomography angiography (CTA) have been associated with future cardiac events. Our aim was to investigate whether ...quantitative global per-patient plaque characteristics from coronary CTA can predict subsequent cardiac death during long-term follow-up.
Methods and results
Out of 2748 patients without prior history of coronary artery disease undergoing CTA with dual-source CT, 32 patients suffered cardiac death (mean follow-up of 5 ± 2 years). These patients were matched to 32 controls by age, gender, risk factors, and symptoms (total 64 patients, 59% male, age 69 ± 10 years). Coronary CTA data sets were analysed by semi-automated software to quantify plaque characteristics over the entire coronary tree, including total plaque volume, volumes of non-calcified plaque (NCP), low-density non-calcified plaque (LD-NCP, attenuation <30 Hounsfield units), calcified plaque (CP), and corresponding burden (plaque volume × 100%/vessel volume), as well as stenosis and contrast density difference (CDD, maximum percent difference in luminal attenuation/cross-sectional area compared to proximal cross-section). In patients who died from cardiac cause, NCP, LD-NCP, CP and total plaque volumes, quantitative stenosis, and CDD were significantly increased compared to controls (P < 0.025 for all). NCP > 146 mm³ hazards ratio (HR) 2.24; 1.09–4.58; P = 0.027, LD-NCP > 10.6 mm³ (HR 2.26; 1.11–4.63; P = 0.025), total plaque volume > 179 mm³ (HR 2.30; 1.12–4.71; P = 0.022), and CDD > 35% in any vessel (HR 2.85;1.4–5.9; P = 0.005) were associated with increased risk of future cardiac death, when adjusted for segment involvement score.
Conclusion
Among quantitative global plaque characteristics, total, non-calcified, and low-density plaque volumes as well as CDD predict cardiac death in long-term follow-up.
This study aimed to assess the association between increased lesion peri-coronary adipose tissue (PCAT) density and coronary 18F-sodium fluoride (18F-NaF) uptake on positron emission tomography (PET) ...in stable patients with high-risk coronary plaques (HRPs) shown on coronary computed tomography angiography (CTA).
Coronary 18F-NaF uptake reflects the rate of calcification of coronary atherosclerotic plaque. Increased PCAT density is associated with vascular inflammation. Currently, the relationship between increased PCAT density and 18F-NaF uptake in stable patients with HRPs on coronary CTA has not been characterized.
Patients who underwent coronary CTA were screened for HRP, which was defined by 3 concurrent plaque features: positive remodeling; low attenuation plaque (LAP) (<30 Hounsfield units HU) and spotty calcification; and obstructive coronary stenosis ≥50% (plaque volume >100 mm3). Patients with HRPs were recruited to undergo 18F-NaF PET/CT. In lesions with stenosis ≥25%, quantitative plaque analysis, mean PCAT density, maximal coronary motion−corrected 18F-NaF standard uptake values (SUVmax), and target-to-background ratios (TBR) were measured.
Forty-one patients (age 65 ± 6 years; 68% men) were recruited. Fifty-one lesions in 23 patients (56%) showed increased coronary 18F-NaF activity. Lesions with 18F-NaF uptake had higher surrounding PCAT density than those without 18F-NaF uptake (−73 HU; interquartile range −79 to −68 HU vs. −86 HU; interquartile range −94 to −80 HU; p < 0.001). 18F-NaF TBR and SUVmax were correlated with PCAT density (r = 0.63 and r = 0.68, respectively; all p < 0.001). On adjusted multiple regression analysis, increased lesion PCAT density and LAP volume were associated with 18F-NaF TBR (β = 0.25; 95% confidence interval: 0.17 to 0.34; p < 0.001 for PCAT, and β = 0.07; 95% confidence interval: 0.03 to 0.11; p = 0.002 for LAP).
In patients with HRP features on coronary CTA, increased density of PCAT was associated with focal 18F-NaF PET uptake. Simultaneous assessment of these imaging biomarkers by 18F-NaF PET and CTA might refine cardiovascular risk prediction in stable patients with HRP features.
Display omitted
PDF
