Dynamic changes on serial PET predict short- and long-term outcomes in cardiac sarcoidosis

Background

Serial positron emission tomography (PET) is commonly used to monitor inflammatory activity in cardiac sarcoidosis (CS), yet follow-up strategies are not standardized, and the prognostic significance of early metabolic response remains uncertain.

Purpose

To determine whether qualitative and quantitative changes between baseline and first follow-up PET independently predict major clinical events in CS.

Methods

We analyzed 236 patients with CS from four tertiary centers undergoing serial PET imaging. The primary endpoint was time to first occurrence of death, sustained ventricular tachycardia, appropriate implantable cardioverter defibrillator shock, or heart transplantation. The index date was defined as the first clinical manifestation in symptomatic patients or the first diagnostic imaging (cardiac magnetic resonance or PET) in asymptomatic individuals. Median follow-up was 5.4 years [IQR 3.4-8.2 years]. Patients were stratified according to overall clinical impression at first follow-up PET (better/unchanged vs worse). Quantitative parameters, including myocardial and extracardiac standardized uptake values (SUV), were analyzed at baseline, follow-up, and as delta changes (Δ). To assess predictors of adverse events, multivariable Cox models were used, adjusted for age, sex, index year, Neutrophil-To-Lymphocyte Ratio (NLR), NT-proBNP, NYHA class, symptoms at onset, extracardiac PET uptake, cardiac magnetic left ventricular mass and ejection fraction, beta-blocker therapy, and steroid treatment. Predictors are reported as hazard ratios (HRs) and 95% confidence intervals (CIs).

Results

Composite events occurred in 103 patients (43.6%). Worsening PET overall impression was associated with significantly higher risk in univariable analysis (HR 4.12, 95% CI 2.36-7.16) and remained independently predictive after multivariable adjustment (HR 3.20, 95% CI 1.68-6.09) (Figure 1).

Baseline and follow-up absolute SUV values showed inconsistent associations. In contrast, dynamic quantitative changes demonstrated strong prognostic value: ΔSUVmax myocardium/blood pool (HR 0.25, 95% CI 0.16-0.38), ΔSUVmax myocardium (HR 0.46, 95% CI 0.34-0.62), and ΔSUVmax extracardiac (HR 0.53, 95% CI 0.37-0.76) were all associated with a lower risk of experiencing the composite outcome. These quantitative findings were directionally concordant with qualitative PET classification.

Conclusion

In this multicenter cohort, dynamic metabolic changes between baseline and first follow-up PET were strongly associated with short- and long-term outcomes in patients with CS. Early metabolic response provides clinically meaningful prognostic information and supports a structured approach to follow-up PET in CS.Figure 1