The Potential of Radiomic Analysis for Enhancing the Diagnostic Ability of PET and CMR in Cardiac Sarcoidosis

Abstract

Cardiac sarcoidosis (CS) is a granulomatous inflammatory disease whose aetiology is unknown, which features the existence of non-caseating granulomas. This thesis addresses the challenge of accurately diagnosing CS by enhancing the diagnostic capabilities of [18F]fluorodeoxyglucose positron emission tomography ([18F]FDG PET) and late gadolinium-enhanced cardiac magnetic resonance imaging (LGE-CMR). Independently, these modalities face limitations in isolating CS with high specificity and sensitivity. The thesis aimed to improve the diagnostic efficiency by integrating [18F]FDG PET and LGE-CMR through advanced radiomic feature analysis. Radiomic analysis was conducted across various scenarios, encompassing comparisons between positive and negative CS groups, distinguishing between active and inactive disease states, and differentiating CS patients from those experiencing myocardial inflammation due to another cause (post-COVID-19 patients).

The thesis concludes that radiomic analysis can enhance the objectivity and complementarity of PET and CMR in identifying cardiac sarcoidosis. While PET- based analyses demonstrate high performance, the project underscores the essential role of CMR-based analysis in mitigating challenges associated with PET image preparation variability.