Abstract Body (Do not enter title and authors here): Introduction
Fontan palliation surgery is the final step in a series of three operations to treat underdeveloped ventricles in patients with single ventricular heart disease. Fontan survivors exhibit right heart failure physiology with low cardiac output and high central venous pressures, with resulting brain morphometry changes. Iron deficiency is also a common comorbidity in these patients and could further impact brain development and cognition. Iron deposition in deep brain nuclei increases systematically with age in healthy children and serves as a marker of brain age. This study uses Quantitative Susceptibility Mapping (QSM) to probe for disturbed subcortical iron deposition in adolescent and young adult Fontan patients.

Hypothesis
The overall distribution of iron within the basal ganglia, hippocampus, and amygdala in Fontan patients would be increased compared to that of healthy control because of hypoxic brain damage and accelerated aging.

Methods
Participants: 26 Fontan patients of age=20.5±4.4 years, hemoglobin=15.4±2.3 g/dL, and hematocrit=42.8±5%. 35 healthy controls of age=22±2.5 years, hemoglobin =14.1±1.4 g/dL, and hematocrit =42.8±4%.
MRI data: 3D T1-weighted and Gradient echo QSM imaging acquired on a 3T Philips Achieva at Children’s Hospital Los Angeles.
MRI modeling: Brain segmentations based on the BCI-DNI anatomical atlas. Susceptibility maps calculated with SEPIA framework and then aligned to the BCI-DNI atlas.
Statistical analysis: Voxel-wise analysis of the individual regions controlling for age and sex. All statistical outcomes were corrected with 0.05 false discovery rate.

Results
Figure 1 shows statistically significant clusters with bilateral patterns of increased|decreased (white | yellow) iron distribution across all regions of interest. Figure 2 (left | right) illustrates a significant decrease in total brain blood flow|end-tidal CO₂ levels in Fontan patients. When adjusting for these two parameters, all the differences illustrated in Fig. 1 were accounted for. Mean cellular hemoglobin concentration (a marker of iron deficiency) was uncorrelated with brain iron distribution.

Conclusion
We demonstrate co-localized regions of increased and decreased iron in the same nuclei (redistribution) rather than the global changes anticipated with hypoxia or iron deficiency. This likely reflects regional changes in brain network activity and connectivity, which we will probe further with resting state and diffusion MRI.