Abstracts

RETURN TO ABSTRACT LISTING


2432. Improved Prognostic Value of Coronary CT Angiography-Derived Plaque Information and Clinical Parameter on MACE Using Machine Learning
Authors * Denotes Presenting Author
  1. Piotr Nikodem Rudzinski; Medical University of South Carolina; The Cardinal Stefan Wyszynski National Institute of Cardiology
  2. Christian Tesche; Munich University Clinic, Ludwig-Maximilians University
  3. Maximilian Bauer; Munich University Clinic, Ludwig-Maximilians University
  4. Verena Brandt *; Robert-Bosch-Krankenhaus
  5. Moritz Baquet; Munich University Clinic, Ludwig-Maximilians University
  6. U. Joseph Schoepf; Medical University of South Carolina
  7. Ulrich Ebersberger; Munich University Clinic, Ludwig-Maximilians University
Objective:
To evaluate the prognostic value of coronary CT angiography (cCTA)-derived plaque information and clinical parameter on adverse cardiac outcome using machine learning (ML).

Materials and Methods:
Datasets of 361 patients (61.9±10.3 years,65% male) with suspected coronary artery disease (CAD) who underwent cCTA were retrospectively analyzed. Major adverse cardiac events (MACE) more than 90 days after cCTA were recorded. Several cCTA-derived plaque measures and conventional CT risk scores together with cardiovascular risk factors were provided to a ML model to predict MACE. A boosted ensemble algorithm (RUSBoost) utilizing decision trees as weak learners with repeated nested cross-validation to train and validate the model was used. Performance of the ML model was calculated using the area under the receiver operating characteristic curve (AUC).

Results:
MACE was observed in 31 patients (8.6%) after a median follow-up of 5.4 years. Discriminatory power was significantly higher for the ML model (AUC 0.96[95%CI 0.93-0.98]) compared to conventional CT risk scores including Agatston calcium score (AUC 0.84 (95%CI 0.80-0.87)), segment involvement score (AUC 0.88(95%CI 0.84-0.91)), and segment stenosis score (AUC 0.89(95%CI 0.86-0.92),all p<0.05). Similar results were shown for plaque measures (AUCs 0.72-0.82, all p<0.05) and clinical parameters including the Framingham risk score (AUCs 0.71-0.76, all p<0.05). The ML model yielded significantly higher diagnostic performance when compared to logistic regression analysis (AUC 0.96vs.0.92, p=0.024).

Conclusion:
Integration of a ML model improves the prediction of MACE when compared to conventional CT risk scores, plaque measures and clinical information. ML algorithms may improve the integration of patient's information to improve risk stratification. ML based cCTA-derived plaque quantification and characterization may have utility in risk-stratifying the vulnerability of coronary lesions for the prediction of future major adverse cardiac events.