Abstract
PURPOSE: Our primary purpose was to understand the correlation between pancreas T1-weighted signal intensity ratio (SIR) and T1 relaxation time in children. We also sought to characterize differences in T1 SIR between children without and with pancreatitis. METHODS: Retrospective study of patients < 18-years-old. SIR-pancreas:spleen (SIR-PS) and SIR-pancreas:paraspinal muscle (SIR-PM) were generated from T1-weighted gradient recalled echo images. Subdivided by field strength, T1 SIR was correlated (Spearman's) with T1 relaxation time. RESULTS: 220 participants were included, 144 imaged at 1.5T (mean: 11.4 ± 4.2 years) and 76 imaged at 3T (mean: 10.9 ± 4.5 years). At 1.5T, SIR-PS (rho=-0.62, 95% CI: -0.71 to -0.51, p < 0.0001) and SIR-PM (rho=-0.57, 95% CI: -0.67 to -0.45, p < 0.0001) moderately negatively correlated with T1 relaxation time. At 3T, correlations between T1 SIR and T1 relaxation time were moderate (rho=-0.40 to -0.43, p ≤ 0.0003). SIR-PS was significantly different between patient groups at 1.5T (p < 0.0001) with pairwise differences between: normal vs. acute on chronic pancreatitis (1.52 vs. 1.13; p < 0.0001). SIR-PM was also significantly different between groups at 1.5T (p < 0.0001) with differences between: normal vs. acute pancreatitis (1.65 vs. 1.40; p = 0.0006), normal vs. acute on chronic pancreatitis (1.65 vs. 1.18; p < 0.0001), and normal vs. chronic pancreatitis (1.65 vs. 1.52; p = 0.0066). A SIR-PS cut-off of ≤ 1.31 had 44% sensitivity and 95% specificity and SIR-PM cut-off of ≤ 1.53 had 69% sensitivity and 70% specificity for pancreatitis. At 3T, SIR-PS was significantly different between groups (p = 0.033) but without significant pairwise differences. CONCLUSION: At 1.5T pancreas T1 SIR moderately to strongly correlates with estimated T1 relaxation time and is significantly lower in children with pancreatitis.