Mortality impact of low CAC density predominantly occurs in early atherosclerosis: explainable ML in the CAC consortium.

BACKGROUND: Machine learning (ML) models of risk prediction with coronary artery calcium (CAC) and CAC characteristics exhibit high performance, but are not inherently interpretable. OBJECTIVES: To determine the direction and magnitude of impact of CAC characteristics on 10-year all-cause mortality (ACM) with explainable ML. METHODS: We analyzed asymptomatic subjects in the CAC consortium. We trained ML models on 80% and tested on 20% of the data with XGBoost, using clinical characteristics â€‹+ â€‹CAC (ML 1) and additional CAC characteristics of CAC density and number of calcified vessels (ML 2). We applied SHAP, an explainable ML tool, to explore the relationship of CAC and CAC characteristics with 10-year all-cause and CV mortality. RESULTS: 2376 deaths occurred among 63,215 patients [68% male, median age 54 (IQR 47-61), CAC 3 (IQR 0-94.3)]. ML2 was similar to ML1 to predict all-cause mortality (Area Under the Curve (AUC) 0.819 vs 0.821, p â€‹= â€‹0.23), but superior for CV mortality (0.847 vs 0.845, p â€‹= â€‹0.03). Low CAC density increased mortality impact, particularly ≤0.75. Very low CAC density ≤0.75 was present in only 4.3% of the patients with measurable density, and 75% occurred in CAC1-100. The number of diseased vessels did not increase mortality overall when simultaneously accounting for CAC and CAC density. CONCLUSION: CAC density contributes to mortality risk primarily when it is very low ≤0.75, which is primarily observed in CAC 1-100. CAC and CAC density are more important for mortality prediction than the number of diseased vessels, and improve prediction of CV but not all-cause mortality. Explainable ML techniques are useful to describe granular relationships in otherwise opaque prediction models.

A Boosted Ensemble Algorithm for Determination of Plaque Stability in High-Risk Patients on Coronary CTA.

OBJECTIVES: This study sought to identify culprit lesion (CL) precursors among acute coronary syndrome (ACS) patients based on qualitative and quantitative computed tomography-based plaque characteristics. BACKGROUND: Coronary computed tomography angiography (CTA) has been validated for patient-level prediction of ACS. However, the applicability of coronary CTA to CL assessment is not known. METHODS: Utilizing the ICONIC (Incident COroNary Syndromes Identified by Computed Tomography) study, a nested case-control study of 468 patients with baseline coronary CTA, the study included ACS patients with invasive coronary angiography-adjudicated CLs that could be aligned to CL precursors on baseline coronary CTA. Separate blinded core laboratories adjudicated CLs and performed atherosclerotic plaque evaluation. Thereafter, the study used a boosted ensemble algorithm (XGBoost) to develop a predictive model of CLs. Data were randomly split into a training set (80%) and a test set (20%). The area under the receiver-operating characteristic curve of this model was compared with that of diameter stenosis (model 1), high-risk plaque features (model 2), and lesion-level features of CL precursors from the ICONIC study (model 3). Thereafter, the machine learning (ML) model was applied to 234 non-ACS patients with 864 lesions to determine model performance for CL exclusion. RESULTS: CL precursors were identified by both coronary angiography and baseline coronary CTA in 124 of 234 (53.0%) patients, with a total of 582 lesions (containing 124 CLs) included in the analysis. The ML model demonstrated significantly higher area under the receiver-operating characteristic curve for discriminating CL precursors (0.774; 95% confidence interval [CI]: 0.758 to 0.790) compared with model 1 (0.599; 95% CI: 0.599 to 0.599; p < 0.01), model 2 (0.532; 95% CI: 0.501 to 0.563; p < 0.01), and model 3 (0.672; 95% CI: 0.662 to 0.682; p < 0.01). When applied to the non-ACS cohort, the ML model had a specificity of 89.3% for excluding CLs. CONCLUSIONS: In a high-risk cohort, a boosted ensemble algorithm can be used to predict CL from non-CL precursors on coronary CTA.

Extraction of radiographic findings from unstructured thoracoabdominal computed tomography reports using convolutional neural network based natural language processing.

BACKGROUND: Heart failure (HF) is a major cause of morbidity and mortality. However, much of the clinical data is unstructured in the form of radiology reports, while the process of data collection and curation is arduous and time-consuming. PURPOSE: We utilized a machine learning (ML)-based natural language processing (NLP) approach to extract clinical terms from unstructured radiology reports. Additionally, we investigate the prognostic value of the extracted data in predicting all-cause mortality (ACM) in HF patients. MATERIALS AND METHODS: This observational cohort study utilized 122,025 thoracoabdominal computed tomography (CT) reports from 11,808 HF patients obtained between 2008 and 2018. 1,560 CT reports were manually annotated for the presence or absence of 14 radiographic findings, in addition to age and gender. Thereafter, a Convolutional Neural Network (CNN) was trained, validated and tested to determine the presence or absence of these features. Further, the ability of CNN to predict ACM was evaluated using Cox regression analysis on the extracted features. RESULTS: 11,808 CT reports were analyzed from 11,808 patients (mean age 72.8 ± 14.8 years; 52.7% (6,217/11,808) male) from whom 3,107 died during the 10.6-year follow-up. The CNN demonstrated excellent accuracy for retrieval of the 14 radiographic findings with area-under-the-curve (AUC) ranging between 0.83-1.00 (F1 score 0.84-0.97). Cox model showed the time-dependent AUC for predicting ACM was 0.747 (95% confidence interval [CI] of 0.704-0.790) at 30 days. CONCLUSION: An ML-based NLP approach to unstructured CT reports demonstrates excellent accuracy for the extraction of predetermined radiographic findings, and provides prognostic value in HF patients.

Automatic segmentation of multiple cardiovascular structures from cardiac computed tomography angiography images using deep learning.

OBJECTIVES: To develop, demonstrate and evaluate an automated deep learning method for multiple cardiovascular structure segmentation. BACKGROUND: Segmentation of cardiovascular images is resource-intensive. We design an automated deep learning method for the segmentation of multiple structures from Coronary Computed Tomography Angiography (CCTA) images. METHODS: Images from a multicenter registry of patients that underwent clinically-indicated CCTA were used. The proximal ascending and descending aorta (PAA, DA), superior and inferior vena cavae (SVC, IVC), pulmonary artery (PA), coronary sinus (CS), right ventricular wall (RVW) and left atrial wall (LAW) were annotated as ground truth. The U-net-derived deep learning model was trained, validated and tested in a 70:20:10 split. RESULTS: The dataset comprised 206 patients, with 5.130 billion pixels. Mean age was 59.9 ± 9.4 yrs., and was 42.7% female. An overall median Dice score of 0.820 (0.782, 0.843) was achieved. Median Dice scores for PAA, DA, SVC, IVC, PA, CS, RVW and LAW were 0.969 (0.979, 0.988), 0.953 (0.955, 0.983), 0.937 (0.934, 0.965), 0.903 (0.897, 0.948), 0.775 (0.724, 0.925), 0.720 (0.642, 0.809), 0.685 (0.631, 0.761) and 0.625 (0.596, 0.749) respectively. Apart from the CS, there were no significant differences in performance between sexes or age groups. CONCLUSIONS: An automated deep learning model demonstrated segmentation of multiple cardiovascular structures from CCTA images with reasonable overall accuracy when evaluated on a pixel level.

Machine learning of clinical variables and coronary artery calcium scoring for the prediction of obstructive coronary artery disease on coronary computed tomography angiography: analysis from the CONFIRM registry.

AIMS: Symptom-based pretest probability scores that estimate the likelihood of obstructive coronary artery disease (CAD) in stable chest pain have moderate accuracy. We sought to develop a machine learning (ML) model, utilizing clinical factors and the coronary artery calcium score (CACS), to predict the presence of obstructive CAD on coronary computed tomography angiography (CCTA). METHODS AND RESULTS: The study screened 35 281 participants enrolled in the CONFIRM registry, who underwent ≥64 detector row CCTA evaluation because of either suspected or previously established CAD. A boosted ensemble algorithm (XGBoost) was used, with data split into a training set (80%) on which 10-fold cross-validation was done and a test set (20%). Performance was assessed of the (1) ML model (using 25 clinical and demographic features), (2) ML + CACS, (3) CAD consortium clinical score, (4) CAD consortium clinical score + CACS, and (5) updated Diamond-Forrester (UDF) score. The study population comprised of 13 054 patients, of whom 2380 (18.2%) had obstructive CAD (≥50% stenosis). Machine learning with CACS produced the best performance [area under the curve (AUC) of 0.881] compared with ML alone (AUC of 0.773), CAD consortium clinical score (AUC of 0.734), and with CACS (AUC of 0.866) and UDF (AUC of 0.682), P < 0.05 for all comparisons. CACS, age, and gender were the highest ranking features. CONCLUSION: A ML model incorporating clinical features in addition to CACS can accurately estimate the pretest likelihood of obstructive CAD on CCTA. In clinical practice, the utilization of such an approach could improve risk stratification and help guide downstream management.

Identification and Quantification of Cardiovascular Structures From CCTA: An End-to-End, Rapid, Pixel-Wise, Deep-Learning Method.

OBJECTIVES: This study designed and evaluated an end-to-end deep learning solution for cardiac segmentation and quantification. BACKGROUND: Segmentation of cardiac structures from coronary computed tomography angiography (CCTA) images is laborious. We designed an end-to-end deep-learning solution. METHODS: Scans were obtained from multicenter registries of 166 patients who underwent clinically indicated CCTA. Left ventricular volume (LVV) and right ventricular volume (RVV), left atrial volume (LAV) and right atrial volume (RAV), and left ventricular myocardial mass (LVM) were manually annotated as ground truth. A U-Net-inspired, deep-learning model was trained, validated, and tested in a 70:20:10 split. RESULTS: Mean age was 61.1 ± 8.4 years, and 49% were women. A combined overall median Dice score of 0.9246 (interquartile range: 0.8870 to 0.9475) was achieved. The median Dice scores for LVV, RVV, LAV, RAV, and LVM were 0.938 (interquartile range: 0.887 to 0.958), 0.927 (interquartile range: 0.916 to 0.946), 0.934 (interquartile range: 0.899 to 0.950), 0.915 (interquartile range: 0.890 to 0.920), and 0.920 (interquartile range: 0.811 to 0.944), respectively. Model prediction correlated and agreed well with manual annotation for LVV (r = 0.98), RVV (r = 0.97), LAV (r = 0.78), RAV (r = 0.97), and LVM (r = 0.94) (p < 0.05 for all). Mean difference and limits of agreement for LVV, RVV, LAV, RAV, and LVM were 1.20 ml (95% CI: -7.12 to 9.51), -0.78 ml (95% CI: -10.08 to 8.52), -3.75 ml (95% CI: -21.53 to 14.03), 0.97 ml (95% CI: -6.14 to 8.09), and 6.41 g (95% CI: -8.71 to 21.52), respectively. CONCLUSIONS: A deep-learning model rapidly segmented and quantified cardiac structures. This was done with high accuracy on a pixel level, with good agreement with manual annotation, facilitating its expansion into areas of research and clinical import.

Determinants of In-Hospital Mortality After Percutaneous Coronary Intervention: A Machine Learning Approach.

Background The ability to accurately predict the occurrence of in-hospital death after percutaneous coronary intervention is important for clinical decision-making. We sought to utilize the New York Percutaneous Coronary Intervention Reporting System in order to elucidate the determinants of in-hospital mortality in patients undergoing percutaneous coronary intervention across New York State. Methods and Results We examined 479 804 patients undergoing percutaneous coronary intervention between 2004 and 2012, utilizing traditional and advanced machine learning algorithms to determine the most significant predictors of in-hospital mortality. The entire data were randomly split into a training (80%) and a testing set (20%). Tuned hyperparameters were used to generate a trained model while the performance of the model was independently evaluated on the testing set after plotting a receiver-operator characteristic curve and using the output measure of the area under the curve ( AUC ) and the associated 95% CIs. Mean age was 65.2±11.9 years and 68.5% were women. There were 2549 in-hospital deaths within the patient population. A boosted ensemble algorithm (AdaBoost) had optimal discrimination with AUC of 0.927 (95% CI 0.923-0.929) compared with AUC of 0.913 for XGB oost (95% CI 0.906-0.919, P=0.02), AUC of 0.892 for Random Forest (95% CI 0.889-0.896, P<0.01), and AUC of 0.908 for logistic regression (95% CI 0.907-0.910, P<0.01). The 2 most significant predictors were age and ejection fraction. Conclusions A big data approach that utilizes advanced machine learning algorithms identifies new associations among risk factors and provides high accuracy for the prediction of in-hospital mortality in patients undergoing percutaneous coronary intervention.

Clinical applications of machine learning in cardiovascular disease and its relevance to cardiac imaging.

Artificial intelligence (AI) has transformed key aspects of human life. Machine learning (ML), which is a subset of AI wherein machines autonomously acquire information by extracting patterns from large databases, has been increasingly used within the medical community, and specifically within the domain of cardiovascular diseases. In this review, we present a brief overview of ML methodologies that are used for the construction of inferential and predictive data-driven models. We highlight several domains of ML application such as echocardiography, electrocardiography, and recently developed non-invasive imaging modalities such as coronary artery calcium scoring and coronary computed tomography angiography. We conclude by reviewing the limitations associated with contemporary application of ML algorithms within the cardiovascular disease field.

Machine learning in cardiac CT: Basic concepts and contemporary data.

Propelled by the synergy of the groundbreaking advancements in the ability to analyze high-dimensional datasets and the increasing availability of imaging and clinical data, machine learning (ML) is poised to transform the practice of cardiovascular medicine. Owing to the growing body of literature validating both the diagnostic performance as well as the prognostic implications of anatomic and physiologic findings, coronary computed tomography angiography (CCTA) is now a well-established non-invasive modality for the assessment of cardiovascular disease. ML has been increasingly utilized to optimize performance as well as extract data from CCTA as well as non-contrast enhanced cardiac CT scans. The purpose of this review is to describe the contemporary state of ML based algorithms applied to cardiac CT, as well as to provide clinicians with an understanding of its benefits and associated limitations.

Maximization of the usage of coronary CTA derived plaque information using a machine learning based algorithm to improve risk stratification; insights from the CONFIRM registry.

INTRODUCTION: Machine learning (ML) is a field in computer science that demonstrated to effectively integrate clinical and imaging data for the creation of prognostic scores. The current study investigated whether a ML score, incorporating only the 16 segment coronary tree information derived from coronary computed tomography angiography (CCTA), provides enhanced risk stratification compared with current CCTA based risk scores. METHODS: From the multi-center CONFIRM registry, patients were included with complete CCTA risk score information and ≥3 year follow-up for myocardial infarction and death (primary endpoint). Patients with prior coronary artery disease were excluded. Conventional CCTA risk scores (conventional CCTA approach, segment involvement score, duke prognostic index, segment stenosis score, and the Leaman risk score) and a score created using ML were compared for the area under the receiver operating characteristic curve (AUC). Only 16 segment based coronary stenosis (0%, 1-24%, 25-49%, 50-69%, 70-99% and 100%) and composition (calcified, mixed and non-calcified plaque) were provided to the ML model. A boosted ensemble algorithm (extreme gradient boosting; XGBoost) was used and the entire data was randomly split into a training set (80%) and testing set (20%). First, tuned hyperparameters were used to generate a trained model from the training data set (80% of data). Second, the performance of this trained model was independently tested on the unseen test set (20% of data). RESULTS: In total, 8844 patients (mean age 58.0 ±â€¯11.5 years, 57.7% male) were included. During a mean follow-up time of 4.6 ±â€¯1.5 years, 609 events occurred (6.9%). No CAD was observed in 48.7% (3.5% event), non-obstructive CAD in 31.8% (6.8% event), and obstructive CAD in 19.5% (15.6% event). Discrimination of events as expressed by AUC was significantly better for the ML based approach (0.771) vs the other scores (ranging from 0.685 to 0.701), P < 0.001. Net reclassification improvement analysis showed that the improved risk stratification was the result of down-classification of risk among patients that did not experience events (non-events). CONCLUSION: A risk score created by a ML based algorithm, that utilizes standard 16 coronary segment stenosis and composition information derived from detailed CCTA reading, has greater prognostic accuracy than current CCTA integrated risk scores. These findings indicate that a ML based algorithm can improve the integration of CCTA derived plaque information to improve risk stratification.