Artificial intelligence in medical imaging: A radiomic guide to precision phenotyping of cardiovascular disease

There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

Abstract

Rapid technological advances in non-invasive imaging, coupled with the availability of large data sets and the expansion of computational models and power, have revolutionized the role of imaging in medicine. Non-invasive imaging is the pillar of modern cardiovascular diagnostics, with modalities such as cardiac computed tomography (CT) now recognized as first-line options for cardiovascular risk stratification and the assessment of stable or even unstable patients. To date, cardiovascular imaging has lagged behind other fields, such as oncology, in the clinical translational of artificial intelligence (AI)-based approaches. We hereby review the current status of AI in non-invasive cardiovascular imaging, using cardiac CT as a running example of how novel machine learning (ML)-based radiomic approaches can improve clinical care. The integration of ML, deep learning, and radiomic methods has revealed direct links between tissue imaging phenotyping and tissue biology, with important clinical implications. More specifically, we discuss the current evidence, strengths, limitations, and future directions for AI in cardiac imaging and CT, as well as lessons that can be learned from other areas. Finally, we propose a scientific framework in order to ensure the clinical and scientific validity of future studies in this novel, yet highly promising field. Still in its infancy, AI-based cardiovascular imaging has a lot to offer to both the patients and their doctors as it catalyzes the transition towards a more precise phenotyping of cardiovascular disease.

Related collections

Most cited references 56

Record: found
Abstract: found
Article: not found

Coronary calcium as a predictor of coronary events in four racial or ethnic groups.

Robert Detrano, Alan Guerci, J Carr … (2008)

In white populations, computed tomographic measurements of coronary-artery calcium predict coronary heart disease independently of traditional coronary risk factors. However, it is not known whether coronary-artery calcium predicts coronary heart disease in other racial or ethnic groups. We collected data on risk factors and performed scanning for coronary calcium in a population-based sample of 6722 men and women, of whom 38.6% were white, 27.6% were black, 21.9% were Hispanic, and 11.9% were Chinese. The study subjects had no clinical cardiovascular disease at entry and were followed for a median of 3.8 years. There were 162 coronary events, of which 89 were major events (myocardial infarction or death from coronary heart disease). In comparison with participants with no coronary calcium, the adjusted risk of a coronary event was increased by a factor of 7.73 among participants with coronary calcium scores between 101 and 300 and by a factor of 9.67 among participants with scores above 300 (P<0.001 for both comparisons). Among the four racial and ethnic groups, a doubling of the calcium score increased the risk of a major coronary event by 15 to 35% and the risk of any coronary event by 18 to 39%. The areas under the receiver-operating-characteristic curves for the prediction of both major coronary events and any coronary event were higher when the calcium score was added to the standard risk factors. The coronary calcium score is a strong predictor of incident coronary heart disease and provides predictive information beyond that provided by standard risk factors in four major racial and ethnic groups in the United States. No major differences among racial and ethnic groups in the predictive value of calcium scores were detected. Copyright 2008 Massachusetts Medical Society.

0 comments Cited 654 times – based on 0 reviews      Review now

Bookmark

Record: found
Abstract: found
Article: not found

Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network

Awni Y. Hannun, Pranav Rajpurkar, Masoumeh Haghpanahi … (2019)

Computerized electrocardiogram (ECG) interpretation plays a critical role in the clinical ECG workflow1. Widely available digital ECG data and the algorithmic paradigm of deep learning2 present an opportunity to substantially improve the accuracy and scalability of automated ECG analysis. However, a comprehensive evaluation of an end-to-end deep learning approach for ECG analysis across a wide variety of diagnostic classes has not been previously reported. Here, we develop a deep neural network (DNN) to classify 12 rhythm classes using 91,232 single-lead ECGs from 53,549 patients who used a single-lead ambulatory ECG monitoring device. When validated against an independent test dataset annotated by a consensus committee of board-certified practicing cardiologists, the DNN achieved an average area under the receiver operating characteristic curve (ROC) of 0.97. The average F1 score, which is the harmonic mean of the positive predictive value and sensitivity, for the DNN (0.837) exceeded that of average cardiologists (0.780). With specificity fixed at the average specificity achieved by cardiologists, the sensitivity of the DNN exceeded the average cardiologist sensitivity for all rhythm classes. These findings demonstrate that an end-to-end deep learning approach can classify a broad range of distinct arrhythmias from single-lead ECGs with high diagnostic performance similar to that of cardiologists. If confirmed in clinical settings, this approach could reduce the rate of misdiagnosed computerized ECG interpretations and improve the efficiency of expert human ECG interpretation by accurately triaging or prioritizing the most urgent conditions.

0 comments Cited 619 times – based on 0 reviews      Review now

Bookmark

Record: found
Abstract: found
Article: found

Is Open Access

Non-invasive detection of coronary inflammation using computed tomography and prediction of residual cardiovascular risk (the CRISP CT study): a post-hoc analysis of prospective outcome data

Evangelos Oikonomou, Mohamed Marwan, Milind Desai … (2018)

Summary Background Coronary artery inflammation inhibits adipogenesis in adjacent perivascular fat. A novel imaging biomarker—the perivascular fat attenuation index (FAI)—captures coronary inflammation by mapping spatial changes of perivascular fat attenuation on coronary computed tomography angiography (CTA). However, the ability of the perivascular FAI to predict clinical outcomes is unknown. Methods In the Cardiovascular RISk Prediction using Computed Tomography (CRISP-CT) study, we did a post-hoc analysis of outcome data gathered prospectively from two independent cohorts of consecutive patients undergoing coronary CTA in Erlangen, Germany (derivation cohort) and Cleveland, OH, USA (validation cohort). Perivascular fat attenuation mapping was done around the three major coronary arteries—the proximal right coronary artery, the left anterior descending artery, and the left circumflex artery. We assessed the prognostic value of perivascular fat attenuation mapping for all-cause and cardiac mortality in Cox regression models, adjusted for age, sex, cardiovascular risk factors, tube voltage, modified Duke coronary artery disease index, and number of coronary CTA-derived high-risk plaque features. Findings Between 2005 and 2009, 1872 participants in the derivation cohort underwent coronary CTA (median age 62 years [range 17–89]). Between 2008 and 2016, 2040 patients in the validation cohort had coronary CTA (median age 53 years [range 19–87]). Median follow-up was 72 months (range 51–109) in the derivation cohort and 54 months (range 4–105) in the validation cohort. In both cohorts, high perivascular FAI values around the proximal right coronary artery and left anterior descending artery (but not around the left circumflex artery) were predictive of all-cause and cardiac mortality and correlated strongly with each other. Therefore, the perivascular FAI measured around the right coronary artery was used as a representative biomarker of global coronary inflammation (for prediction of cardiac mortality, hazard ratio [HR] 2·15, 95% CI 1·33–3·48; p=0·0017 in the derivation cohort, and 2·06, 1·50–2·83; p<0·0001 in the validation cohort). The optimum cutoff for the perivascular FAI, above which there is a steep increase in cardiac mortality, was ascertained as −70·1 Hounsfield units (HU) or higher in the derivation cohort (HR 9·04, 95% CI 3·35–24·40; p<0·0001 for cardiac mortality; 2·55, 1·65–3·92; p<0·0001 for all-cause mortality). This cutoff was confirmed in the validation cohort (HR 5·62, 95% CI 2·90–10·88; p<0·0001 for cardiac mortality; 3·69, 2·26–6·02; p<0·0001 for all-cause mortality). Perivascular FAI improved risk discrimination in both cohorts, leading to significant reclassification for all-cause and cardiac mortality. Interpretation The perivascular FAI enhances cardiac risk prediction and restratification over and above current state-of-the-art assessment in coronary CTA by providing a quantitative measure of coronary inflammation. High perivascular FAI values (cutoff ≥–70·1 HU) are an indicator of increased cardiac mortality and, therefore, could guide early targeted primary prevention and intensive secondary prevention in patients. Funding British Heart Foundation, and the National Institute of Health Research Oxford Biomedical Research Centre.