Free-breathing 3D pancreatic water T1 mapping using radial sampling technique: comparison with conventional 2D technique.

PURPOSE: This study aimed to compare the clinical feasibility and measurement performance of free-breathing three-dimensional (3D) pancreatic water T1 mapping between a 3D radial 6-point DIXON Look-Locker (3D-LL) technique and a conventional breath-hold two-dimensional Look-Locker (2D-LL) technique.

METHODS

Six phantom syringes, comprising three organ-equivalent phantoms and three serial gadolinium dilution phantoms, were scanned using inversion recovery (IR), 2D-LL, and 3D-LL techniques. Pearson correlation coefficient and Bland-Altman analyses were used to assess measurement accuracy relative to IR and agreement between 2D-LL and 3D-LL, respectively. This prospective clinical study included 45 participants who underwent breath-hold 2D-LL T1 mapping and free-breathing 3D-LL pancreatic water T1 mapping. Two independent readers measured T1 values in the pancreatic head, body, and tail, and calculated whole-pancreas averages. Nonevaluable regions were recorded. T1 values were compared between the two techniques using the Wilcoxon signed-rank test.

RESULTS

Phantom experiments demonstrated very strong correlations with IR for both techniques (2D-LL, r = 0.9998; 3D-LL, r = 0.9993), and Bland-Altman analysis revealed a small positive bias of 3D-LL over 2D-LL (+15.1 ms). A clinical study revealed coverage- or motion-related regional failures that precluded T1 measurement in six participants (13%) with 2D-LL, but in none with 3D-LL. The median T1 values in the pancreatic head (P = 0.25), body (P = 0.21), tail (P = 0.25), and whole-pancreas average (P = 0.17) did not differ significantly between the two techniques.

CONCLUSION

The 3D-LL technique demonstrated accurate T1 quantification in phantoms, comparable to IR and 2D-LL. Clinical data indicated that the 3D-LL technique improves pancreatic measurability by avoiding coverage- and motion-related regions that are not amenable to T1 measurement.