Related Experiment Video
Updated: Jul 31, 2025

Author Spotlight: Advancing Cardiovascular Imaging - Introducing the Spatially Weighted Calcium Score for Early Disease Detection
Published on: September 22, 2023
Artificial intelligence using a deep learning versus expert computed tomography human reading in calcium score and
Jairo Aldana-Bitar1,2, Geoffrey W Cho3, Lauren Anderson2
1Division of Cardiology, The Lundquist Institute for Biomedical Innovation at Harbor-UCLA Medical Center, Los Angeles.
Artificial intelligence (AI) shows excellent correlation with human experts for coronary artery calcium (CAC) scoring. However, AI reclassification occurred in 14% of cases, particularly with minimal CAC scores, requiring algorithm optimization.
Area of Science:
- Cardiovascular imaging
- Artificial intelligence in medicine
- Radiology and medical imaging
Background:
- Artificial intelligence (AI) offers potential advancements in cardiac imaging processing and interpretation.
- Coronary artery calcium (CAC) scoring is a validated tool for cardiovascular risk stratification.
- Evaluating AI performance against expert human interpretation is crucial for clinical adoption.
Purpose of the Study:
- To assess the accuracy and correlation of an AI software (Coreline AVIEW) against expert human interpretation for CAC scoring.
- To evaluate the impact of applying the Coronary Artery Calcium Data and Reporting System (CAC-DRS) classification.
- To identify discrepancies and causes of reclassification between AI and human CAC interpretations.
Main Methods:
- 100 non-contrast CT calcium score images were analyzed using AI software and expert human readers.
- Pearson correlation coefficient was calculated to compare absolute CAC scores.
- CAC-DRS categories were applied, and reasons for reclassification were qualitatively assessed.
Main Results:
- A highly significant correlation (R=0.996) was found between AI and human CAC scores.
- Despite high correlation, 14% of patients experienced CAC-DRS category reclassification.
- Reclassification primarily occurred in the CAC-DRS 0-1 range, often due to AI under/overestimation of specific coronary arteries or ventricle densities.
Conclusions:
- AI demonstrates excellent correlation with human expert CAC scoring.
- AI performance requires optimization for accurately classifying minimal CAC volumes to improve clinical utility.
- Further algorithm refinement is needed to enhance AI's sensitivity and specificity for detecting low-volume calcium.
- AI may identify coronary calcium missed by human interpretation in rare instances.
Related Concept Videos
Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
Imaging Studies for Cardiovascular System V: CT
Coronary Artery Disease V: Interprofessional Care
Coronary Artery Disease I: Introduction
Acute Coronary Syndrome III: Diagnostic Studies
Atherosclerosis II: Clinical Manifestations and Diagnostic Tests

