04 Feb AI Can Generate Predictive Coronary Calcium Score in Two Seconds
MedicalResearch.com Interview with:
Dr. Hugo Aerts, PhD
Dana-Farber Cancer Institute
Associate Professor, Brigham and Women’s Hospital
Harvard Medical School
Director, Program for Artificial Intelligence in Medicine
Brigham And Women’s Hospital
MedicalResearch.com: Deep convolutional neural networks to predict cardiovascular risk from computed tomography
Response: Cardiovascular disease is the most common preventable cause of death in Europe and the United States. Effective lifestyle and pharmacological prevention is available, but identifying those who would benefit most remains an ongoing challenge. Hence, efforts are needed to further improve cardiovascular risk prediction and stratification on an individual basis.
One of the strongest known predictors for adverse cardiovascular events is coronary artery calcification, which can be quantified on computed tomography (CT). The CT coronary calcium score is a measure of the burden of coronary atherosclerosis and is one of the most widely accepted measures of cardiovascular risk.
Recent strides in artificial intelligence, deep learning in particular, have shown its viability in several medical applications such as medical diagnostic and imaging, risk management, or virtual assistants. A major advantage is that deep learning can automate complex assessments that previously could only be done by radiologists, but now is feasible at scale with a higher speed and lower cost. This makes deep learning a promising technology for automating cardiovascular event prediction from imaging. However, before clinical introduction can be considered, generalizability of these systems needs to be demonstrated as they need to be able to predict cardiovascular events of asymptomatic and symptomatic individuals across multiple clinical scenarios, and work robustly on data from multiple institutions.
MedicalResearch.com: What are the main findings?
Response: We evaluated 20,084 individuals from distinct asymptomatic cohorts, including the Framingham Heart Study and the National Lung Screening Trial (NLST), as well as from stable and acute chest pain cohorts, including the Prospective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) and the Rule Out Myocardial Infarction using Computer Assisted Tomography (ROMICAT-II) trial.
Our results show that the automated score is a strong predictor of cardiovascular events, independent of risk factors (multivariable-adjusted hazard ratios up to 4.3), with high correlation with manual quantification, and robust test-retest reliability. These results demonstrate the clinical value of a deep learning system for the automated prediction of cardiovascular events.
MedicalResearch.com: Would the new program/software be compatible with existing CT technology?
Response: The presented deep learning system is able to fully automatically evaluate coronary calcium in, non-contrast enhanced, cardiac-ECG gated and non-gated chest CTs. It can be run on graphics processing unit (GPU) workstations and needs less than 2 seconds to process a scan.
MedicalResearch.com: What should readers take away from your report?
Response: In this study scientists and medical experts from the Artificial Intelligence in Medicine Program (AIM) group and Cardiovascular Imaging Research Center (CIRC) concentrated not only on developing a new deep learning system but also on showing its clinical applicability by using four large and diverse cohorts including CT scans from different institutes, recorded using different machines and different protocols.
MedicalResearch.com: What recommendations do you have for future research as a result of this work?
Response: In general we believe that the cooperation of experts from different fields and groups can benefit research especially for highly specialized applications. As for deep learning in general, we think it is important to evaluate its performance in sufficiently large test sets to show its real potential. Furthermore we shared the complete code of our deep learning system as well as the trained models, with the community, without restrictions, hoping that this system can be used in different studies to accelerate and improve cardiovascular risk prediction.
Any disclosures?
General: The authors thank the Framingham Heart Study, NCI, ACRIN, NLST, Prospective Multicenter Imaging Study for Evaluation of Chest Pain, and Rule Out Myocardial Infarction Using Computer Assisted Tomography II trial for access to trial data.
Funding: The authors acknowledge financial support from NIH (HA: NIH-USA U24CA194354, NIH-USA U01CA190234, NIH-USA U01CA209414, and NIH-USA R35CA22052; UH: NIH, 5R01-HL109711, NIH/NHLBI 5K24HL113128, NIH/NHLBI 5T32HL076136, NIH/NHLBI 5U01HL123339), the European Union – European Research Council (HA: 866504), as well as the German Research Foundation (DFG; TA: 1438/1-1 and WE: 6405/2-1), American Heart Association Institute for Precision Cardiovascular Medicine (MTL: 18UNPG34030172), Fulbright Visiting Researcher Grant (E0583118), Rosztoczy Foundation Grant. The Framingham Heart Study (FHS) acknowledges the support of contracts NO1-HC-25195, HHSN268201500001I and 75N92019D00031 from the National Heart, Lung and Blood Institute.
Competing interests:
The authors declare no competing interests to this work.
Citation:
Zeleznik, R., Foldyna, B., Eslami, P. et al. Deep convolutional neural networks to predict cardiovascular risk from computed tomography. Nat Commun 12, 715 (2021). https://doi.org/10.1038/s41467-021-20966-2
JOIN OUR EMAIL LIST
[mailpoet_form id="5"]We respect your privacy and will never share your details.
[last-modified]
The information on MedicalResearch.com is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition. Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.
Last Updated on February 4, 2021 by Marie Benz MD FAAD