Machine Learning with 18F-Sodium Fluoride PET and Quantitative Plaque Analysis on CT Angiography for the Future Risk of Myocardial Infarction

• Published in: The Journal of nuclear medicine (1978) Vol. 63; no. 1; pp. 158 - 165
• Main Authors: Kwiecinski, Jacek, Tzolos, Evangelos, Meah, Mohammed N, Cadet, Sebastien, Adamson, Philip D, Grodecki, Kajetan, Joshi, Nikhil V, Moss, Alastair J, Williams, Michelle C, van Beek, Edwin J R, Berman, Daniel S, Newby, David E, Dey, Damini, Dweck, Marc R, Slomka, Piotr J
• Format: Journal Article
• Language: English
• Published: New York Society of Nuclear Medicine 01.01.2022
• Subjects: Angiography; Cardiovascular disease; Clinical Investigation; Computed tomography; Coronary artery; Coronary artery disease; Coronary vessels; Fluorides; Fluorine isotopes; Heart attacks; Heart diseases; Learning algorithms; Machine learning; Medical imaging; Myocardial infarction; Patients; Performance prediction; Positron emission tomography; Risk; Sodium fluoride; Statistical analysis; Vein & artery diseases
• ISSN: 0161-5505; 1535-5667; 1535-5667
• DOI: 10.2967/jnumed.121.262283

Coronary 18F-sodium fluoride (18F-NaF) PET and CT angiography–based quantitative plaque analysis have shown promise in refining risk stratification in patients with coronary artery disease. We combined both of these novel imaging approaches to develop an optimal machine-learning model for the future risk of myocardial infarction in patients with stable coronary disease. Methods: Patients with known coronary artery disease underwent coronary 18F-NaF PET and CT angiography on a hybrid PET/CT scanner. Machine-learning by extreme gradient boosting was trained using clinical data, CT quantitative plaque analysis, measures and 18F-NaF PET, and it was tested using repeated 10-fold hold-out testing. Results: Among 293 study participants (65 ± 9 y; 84% male), 22 subjects experienced a myocardial infarction over the 53 (40–59) months of follow-up. On univariable receiver-operator-curve analysis, only 18F-NaF coronary uptake emerged as a predictor of myocardial infarction (c-statistic 0.76, 95% CI 0.68–0.83). When incorporated into machine-learning models, clinical characteristics showed limited predictive performance (c-statistic 0.64, 95% CI 0.53–0.76) and were outperformed by a quantitative plaque analysis-based machine-learning model (c-statistic 0.72, 95% CI 0.60–0.84). After inclusion of all available data (clinical, quantitative plaque and 18F-NaF PET), we achieved a substantial improvement (P = 0.008 versus 18F-NaF PET alone) in the model performance (c-statistic 0.85, 95% CI 0.79–0.91). Conclusion: Both 18F-NaF uptake and quantitative plaque analysis measures are additive and strong predictors of outcome in patients with established coronary artery disease. Optimal risk stratification can be achieved by combining clinical data with these approaches in a machine-learning model.