Yoav Arnson

Automatic Valve Plane Localization in Myocardial Perfusion SPECT/CT by Machine Learning: Anatomical and Clinical Validation

Precise definition of the mitral valve plane (VP) during segmentation of the left ventricle for SPECT myocardial perfusion imaging (MPI) quantification often requires manual adjustment, which affects the quantification of perfusion. We developed a machine learning approach using support vector machines (SVM) for automatic VP placement. Methods: A total of 392 consecutive patients undergoing 99mTc-tetrofosmin stress (5 min; mean ± SD, 350 ± 54 MBq) and rest (5 min; 1,024 ± 153 MBq) fast SPECT MPI attenuation corrected (AC) by CT and same-day coronary CT angiography were studied; included in the 392 patients were 48 patients who underwent invasive coronary angiography and had no known coronary artery disease. The left ventricle was segmented with standard clinical software (quantitative perfusion SPECT) by 2 experts, adjusting the VP if needed. Two-class SVM models were computed from the expert placements with 10-fold cross validation to separate the patients used for training and those used for validation. SVM probability estimates were used to compute the best VP position. Automatic VP localizations on AC and non-AC images were compared with expert placement on coronary CT angiography. Stress and rest total perfusion deficits and detection of per-vessel obstructive stenosis by invasive coronary angiography were also compared. Results: Bland–Altman 95% confidence intervals (CIs) for VP localization by SVM and experts for AC stress images (bias, 1; 95% CI, −5 to 7 mm) and AC rest images (bias, 1; 95% CI, −7 to 10 mm) were narrower than interexpert 95% CIs for AC stress images (bias, 0; 95% CI, −8 to 8 mm) and AC rest images (bias, 0; 95% CI, −10 to 10 mm) (P < 0.01). Bland–Altman 95% CIs for VP localization by SVM and experts for non-AC stress images (bias, 1; 95% CI, −4 to 6 mm) and non-AC rest images (bias, 2; 95% CI, −7 to 10 mm) were similar to interexpert 95% CIs for non-AC stress images (bias, 0; 95% CI, −6 to 5 mm) and non-AC rest images (bias, −1; 95% CI, −9 to 7 mm) (P was not significant [NS]). For regional detection of obstructive stenosis, ischemic total perfusion deficit areas under the receiver operating characteristic curve for the 2 experts (AUC, 0.79 [95% CI, 0.7–0.87]; AUC, 0.81 [95% CI, 0.73–0.89]) and the SVM (0.82 [0.74–0.9]) for AC data were the same (P = NS) and were higher than those for the unadjusted VP (0.63 [0.53–0.73]) (P < 0.01). Similarly, for non-AC data, areas under the receiver operating characteristic curve for the experts (AUC, 0.77 [95% CI, 0.69–0.89]; AUC, 0.8 [95% CI, 0.72–0.88]) and the SVM (0.79 [0.71–0.87]) were the same (P = NS) and were higher than those for the unadjusted VP (0.65 [0.56–0.75]) (P < 0.01). Conclusion: Machine learning with SVM allows automatic and accurate VP localization, decreasing user dependence in SPECT MPI quantification.

Molecular Imaging of Vulnerable Coronary Plaque: A Pathophysiologic Perspective

Atherothrombotic events in coronary arteries are most often due to rupture of unstable plaque resulting in myocardial infarction. Radiolabeled molecular imaging tracers directed toward cellular targets that are unique to unstable plaque can serve as a powerful tool for identifying high-risk patients and for assessing the potential of new therapeutic approaches. Two commonly available radiopharmaceuticals—18F-FDG and 18F-NaF—have been used in clinical research for imaging coronary artery plaque, and ongoing clinical studies are testing whether there is an association between 18F-NaF uptake and future atherothrombotic events. Other, less available, tracers that target macrophages, endothelial cells, and apoptotic cells have also been tested in small groups of patients. Adoption of molecular imaging of coronary plaque into clinical practice will depend on overcoming major hurdles, ultimately including evidence that the detection of unstable plaque can change patient management and improve outcomes.

Coronary Artery Calcium Scanning: The Agatston Score and Beyond

SEE PAGE 1407 C oronary artery calcium (CAC) has been promulgated as a specific marker of atherosclerosis since the 1940s. The extent of coronary artery calcification reflects the lifetime effect of all known and unknown factors that cause coronary artery disease. In 1991, Janowitz et al. (1) described a method for assessing the magnitude of CAC based on a single score that quantified the extent and density of computed tomography CAC, which became known as the Agatston score. Clinical risk increases with each increment of CAC score, as consistently demonstrated in numerous studies. Recently, studies have begun to address what additional information can be derived from CAC scanning. In a registry of 25,753 asymptomatic patients with CAC scans and followed for long-term mortality, Budoff et al. (2) and Tota-Maharaj et al. (3,4) demonstrated that the number of vessels with CAC score >100, location of CAC, and number of calcific lesions were all predictors of coronary events. In the MESA (Multi-Ethnic Study of Atherosclerosis) trial, the number of vessels manifesting CAC abnormality added to CAC in predicting subsequent coronary revascularization (5) and the magnitude of atherosclerosis as quantified by the calcium coverage score, representing percentage of the coronary tree with CAC, was highly predictive of coronary heart disease events (6). In other work, Criqui et al. (7) found that CAC density was inversely related to CVD events for a given CAC volume and was more predictive than the

HORMONE REPLACEMENT THERAPY IS ASSOCIATED WITH LESS CORONARY ATHEROSCLEROSIS AND LOWER MORTALITY

Background: There is a controversy regarding the role of hormone replacement therapy (HRT) as a cardio-protective agent in postmenopausal women. We aim to examine the effect of HRT treatment on coronary artery calcium (CAC) and on mortality in a large retrospective cohort of post-menopausal women

Assessment of Coronary Calcium Density on Noncontrast Computed Tomography

SEE PAGE 845 C oronary artery calcification (CAC) on electrocardiogram-gated noncontrast computed tomography (CT) provides a sensitive and specific marker of coronary atherosclerosis. The extent of CAC reflects the lifetime effect of known and unknown risk factors that cause coronary atherosclerosis in an individual patient. Nearly all outcome studies involving CAC scoring and virtually all clinicians who use CAC scanning in clinical practice rely on the Agatston score, first described by Agatston and Janowitz in 1991, and commonly referred to as the CAC score. This score has been consistently shown to provide strong predictive value for future cardiovascular events, providing incremental prognostic value over global risk scores (1). The Agatston score is a derived variable that increases based on the area (which is directly related to volume) and the density of calcified coronary plaques. It uses an arbitrary threshold of 130 Hounsfield units (HU) that seemed to best separate calcium from statistical noise on electron-beam CT imaging. The weighting for density that is applied to each coronary plaque identified on CAC scanning is based on the maximal Hounsfield units within each plaque, assessed by a categorical 4-point multiplication: 1 1⁄4 130 to 199; 2 1⁄4 200 to 299; 3 1⁄4 300 to 399; and 4 1⁄4

Comparison of the Coronary Artery Calcium Score and Number of Calcified Coronary Plaques for Predicting Patient Mortality Risk

400 HU. A prior study report has provided evidence that increased density might be associated with lower rather than higher risk of cardiac event (2).

Non-invasive fractional flow reserve in vessels without severe obstructive stenosis is associated with coronary plaque burden

Multiple coronary artery calcium (CAC) parameters have recently been proposed to improve risk prediction in patients with intermediate clinical risk based on CAC scoring, but outcome data that assess these variables are relatively sparse. We analyzed data from 11,633 consecutive asymptomatic patients undergoing CAC scanning that were followed for 8.8 ± 3.5 years for all-cause mortality (ACM). The patients who had coronary artery calcification were grouped by the number of calcified coronary plaques: 0, 1 to 5, 6 to 20, and >20 plaques. We examined the independent prognostic value of plaque number and its synergistic prognostic value when added to the CAC score. We observed a stepwise increase in ACM with increasing plaque number. In patients with a CAC score of 1 to 99, 6 plaques or more were associated with increased mortality. In patients with CAC scores of 100 to 399, there was a stepwise increase in ACM with increasing plaque number. For CAC >400, the risk of ACM was high regardless of plaque number. After risk adjustment, the number of plaques was a significant predictor of risk for ACM in the patients with an intermediate CAC score. In these patients, additional consideration of plaque number improved net reclassification improvement for predicting ACM by 29%. In conclusion, the number of calcified plaques adds to risk stratification beyond the CAC score in patients with intermediate CAC scores.

IMPROVEMENT IN LDL IS ASSOCIATED WITH DECREASE IN NON-CALCIFIED PLAQUE VOLUME ON CTA AS MEASURED BY AUTOMATED QUANTITATIVE SOFTWARE

AIMS Non-invasive fractional flow reserve derived from coronary CT angiography (FFRCT) has been shown to be predictive of lesion-specific ischemia as assessed by invasive fractional flow reserve (FFR). However, in practice, clinicians are often faced with an abnormal distal FFRCT in the absence of a discrete obstructive lesion. Using quantitative plaque analysis, we sought to determine the relationship between an abnormal whole vessel FFRCT (V-FFRCT) and quantitative measures of whole vessel atherosclerosis in coronary arteries without obstructive stenosis. METHODS FFRCT was calculated in 155 consecutive patients undergoing coronary CTA with ≥25% but less than 70% stenosis in at least one major epicardial vessel. Semi-automated software was used to quantify plaque volumes (total plaque [TP], calcified plaque [CP], non-calcified plaque [NCP], low-density non-calcified plaque [LD-NCP]), remodeling index [RI], maximal contrast density difference [CDD] and percent diameter stenosis [%DS]. Abnormal V-FFRCT was defined as a minimum value of ≤0.75 across the vessel (at the most distal region where FFRCT was computed). RESULTS Vessels with abnormal V-FFRCT had higher per-vessel TP (554 vs 331 mm3), CP (59 vs 25 mm3), NCP (429 vs 295 mm3), LD-NCP (65 vs 35 mm3) volume and maximum CDD (21 vs 14%) than those with normal V-FFRCT (median, p < 0.05 for all). Using a multivariate analysis to adjust for CDD and %DS, all measures of plaque volume were predictive of abnormal V-FFRCT (OR 2.09, 1.36, 1.95, 1.95 for TP, CP, NCP and LD-NCP volume, respectively; p < 0.05 for all). CONCLUSION Abnormal V-FFRCT in vessels without obstructive stenosis is associated with multiple markers of diffuse non-obstructive atherosclerosis, independent of stenosis severity. Whole vessel FFRCT may represent a novel measure of diffuse coronary plaque burden.

CORONARY ARTERY CALCIUM IN THE ELDERLY: PREVALENCE, SEVERITY, AND ASSOCIATION WITH ALL-CAUSE MORTALITY

Background: Plaque analysis of coronary CT angiography (CTA) can provide comprehensive quantitative assessment of coronary plaque characteristics. Low density lipoprotein cholesterol (LDL) lowering is associated with plaque regression by IVUS. We tested the hypothesis that improvement in LDL is

RELATIONSHIP BETWEEN WHOLE VESSEL QUANTITATIVE PLAQUE CHARACTERISTICS AND NONINVASIVE FRACTIONAL FLOW RESERVE FROM CORONARY COMPUTED TOMOGRAPHY ANGIOGRAPHY

Background: Coronary artery calcium (CAC) score is a marker of coronary atherosclerosis with increasing prevalence as a function of age; however, the prevalence and prognostic implications of CAC scanning in elderly patients are not well defined. Methods: We studied 1028 consecutive asymptomatic