PT - JOURNAL ARTICLE AU - nie, binbin AU - Yun, Mingkai AU - Li, Xiang AU - Zhang, Xiaoli AU - Shan, Baoci TI - <strong>Heart failure related pattern of cerebral glucose metabolism for overall risk stratification in HF patients: A cross-sectional and longitudinal study</strong> DP - 2023 Jun 01 TA - Journal of Nuclear Medicine PG - P945--P945 VI - 64 IP - supplement 1 4099 - http://jnm.snmjournals.org/content/64/supplement_1/P945.short 4100 - http://jnm.snmjournals.org/content/64/supplement_1/P945.full SO - J Nucl Med2023 Jun 01; 64 AB - P945 Introduction: Heart failure (HF), a clinical syndrome characterized by complex cross-organs interactions, has been regarded as one of the leading cause of death in the modern world. Recently, it has been recognized from epidemiological studies that HF patients insult increase the risk of brain dysfunction. Remarkably, a significant association has been demonstrated between increased metabolic activities in the amygdala, a neural center involved in the processing of emotions, and deteriorated myocardial function. Thus, the brain's impairment potentially indicates further adverse mental effects of cardiovascular risk, which is associated with worse outcomes of HF. The Scaled Subprofile Modeling (SSM), a principal components analysis (PCA)-based spatial covariance method, could identify whole-brain characteristic abnormalities of functional organization associated a specific disease and simultaneously scoring disease-related severity of each individual. Therefore, this study aimed to construct a whole-brain heart failure related metabolic pattern (HFRP) using SSM for overall risk stratification of HF.Methods: A cohort of 81 HF patients (81 males; age, 59.20 ± 7.56 years, mean ± SD) and 29 age- and gender-matched normal controls (29 males; age: 51.3 ± 14.3years, mean ± SD) from Beijing Anzhen Hospital were enrolled and underwent FDG-PET imaging on a PET/CT system (Biograph mCT, Siemens Healthcare). The 18F-FDG-PET images were spatially normalized and smoothed in SPM12 (Welcome Department of Cognitive Neurology, London, UK). Then, 17 pairs of patient/control were randomly selected for HFRP construction (HFPC/NCPC), while the remained ones for validation (HFPV/NCPV). The HFRP was constructed in scanvp toolkit (version 6.1) based on SSM algorithm, and the maximum separation between the HFPC and NCPC was determined by two-sample Student’s t-test. For validation, the significance between HFPV and NCPV subgroups was also calculated and defined as a P &lt; 0.05. Finally, all the 81 HF patients were re-stratified into low- and high- HFRP expression subgroups by (1) one standard deviation (SD) above the mean cutoff, or (2) 90th percentile cutoff, respectively. Kaplan-Meier survival curves of both low- and high- HFRP expression subgroups were calculated, and statistically analyzed by the log-rank test. Statistical significance was defined as a P &lt; 0.05.Results: The HFRP was identified as the second group independent Subprofile (GIS2) accounting for 8.77% of subject × voxel variance (P = 0.002), and could identified HFPV from NCPV significantly (P &lt; 0.001). The HFRP itself features positive relative deviation in brainstem, cerebellum, thalamus, amygdala, pallidum, putamen, limbic lobe, parietal lobe and occipital lobe, as well as negative deviations in hippocampus, frontal lobe and temporal lobe. The survival rates of HF patients with high-HFRP expression were always significantly lower than the ones with low-HFRP expression, either stratified by one SD above the mean cutoff (P = 0.021) or by the 90th percentile cutoff (P = 0.002). Conclusions: In this study, a HFRP in association with cardiac events was constructed. The relatively high-HFRP expression patients were shown significantly higher overall risk than the low-HFRP expression ones. It may provide a new neuroimaging biomarker for risk stratifications of HF. Further multi-center validations are warranted.