Ramirez, Giselle

Abstract Background Quantitative myocardial blood flow (MBF) and myocardial flow reserve (MFR) provide incremental diagnostic and prognostic value in cardiac PET, but their widespread use is limited by the technical demands of dynamic imaging protocols. We evaluated the feasibility of using artificial intelligence (AI) to predict MBF and MFR from static and gated PET images, without the need for dynamic acquisition. Methods A machine learning (XGBoost) model was trained on 82Rb PET multi-center dataset using static perfusion imaging, injected dose, hemodynamic measures, clinical data and CT-derived features (including body composition) from the hybrid CT attenuation scan. Model performance was evaluated externally in an independent cohort. Results In total, 10,566 (derivation-cohort) and 7,842 (external-cohort) patients were included in this multi-center study. On the external-cohort, AI approach achieved an Area under the curve (AUC) of 0.92 (0.92–0.93) for abnormal stress MBF and 0.91 (0.90–0.92) for abnormal MFR; Intra-class correlation (ICC) 0.80 (0.78–0.82) and 0.78 (0.76–0.79), respectively. AI MFR closely mirrored the prognostic performance of measured MFR, showing nearly identical Kaplan–Meier risk stratification (both p<0.0001) and maintaining strong, and independently significant associations with all-cause mortality (HR 3.4 [2.8–4.2] vs. 4.6 [3.6–5.8]; both p<0.001), and demonstrated similar added value to perfusion for mortality prediction. Conclusion AI-predicted virtual stress MBF and MFR assessment using static and gated PET data is feasible and generalizable across cohorts. By removing the dependency on dynamic acquisitions, this approach has the potential to broaden the clinical adoption of flow quantification. STRUCTURED GRAPHICAL ABSTRACT PET: Positron Emission Tomography, CT: Computed Tomography, MFR: Myocardial Flow Reserve Key Question: Can machine learning models trained on dynamic PET datasets accurately predict regional stress myocardial blood flow (MBF) and myocardial flow reserve (MFR) from static image features, physiological parameters, and CT-based anatomical measures? Key Finding: Artificial intelligence can accurately estimate MBF and MFR from non-dynamic PET data, with strong agreement to reference standards. Take-home Message: By eliminating reliance on dynamic PET acquisitions, machine-learning has the potential to broaden clinical adoption of quantitative flow assessment.

BACKGROUND: Myocardial flow reserve (MFR), measured by positron emission tomography (PET) myocardial perfusion imaging, provides valuable information on epicardial coronary disease, diffuse atherosclerosis, and microvascular function. Despite its routine use, the prognostic efficacy of 13 N-ammonia PET MFR remains unconfirmed in larger multicenter cohorts of patients with suspected or known coronary artery disease. METHODS: We considered patients from 5 sites in the REFINE PET (Registry of Flow and Perfusion Imaging for Artificial Intelligence with PET) registry who underwent 13 N-ammonia PET myocardial perfusion imaging for coronary artery disease. Clinical and imaging data were collected at the time of myocardial perfusion imaging. MFR was quantified as the ratio of stress to rest myocardial blood flow, using QPET software (Cedars-Sinai Medical Center, Los Angeles, CA). The primary outcome was all-cause mortality. Survival analyses were performed using Kaplan-Meier and Cox regression models adjusted for clinical and imaging covariates. RESULTS: In total, 6277 patients were included (median age of 65 years, 56% male). Median follow-up time was 3.8 years. There were 1895 patients with MFR ≤2 and 4382 with MFR >2. Patients with MFR ≤2 had significantly higher mortality than those with MFR >2 (n=701 [37.0%] versus n=537 [12.3%], respectively; P <0.001). Annualized all-cause mortality rates by MFR and summed stress score ranged from 1.7 to 15.8. In multivariable analysis, MFR ≤2 was independently associated with increased all-cause mortality in the overall population (hazard ratio, 2.70 [95% CI, 2.41–3.03]; P <0.001), even among patients with no perfusion defects (hazard ratio, 2.36 [95% CI, 1.93–2.89]; P <0.001). Mortality risk decreased across increasing MFR deciles, ranging from hazard ratio, 2.73 (95% CI, 2.39–3.11) to hazard ratio, 0.35 (95% CI, 0.25–0.49). CONCLUSIONS: In this large multicenter cohort, MFR derived from 13 N-ammonia PET myocardial perfusion imaging is a strong, independent predictor of all-cause mortality, even in patients with normal perfusion. An MFR of ≤2 identifies elevated risk, while higher values are associated with improved survival. These findings support the routine integration of MFR to enhance risk stratification in patients with suspected or known coronary artery disease.