Quantification and differentiation of composition of mixed pancreatic duct stones using single-source dual-energy CT: an ex vivo study.

PURPOSE: To evaluate the feasibility of using single-source dual-energy CT (SS DECT) to quantify and differentiate calcium carbonate (CA) and non-calcium carbonate (NCA) components of pancreatic duct stones (PDS) with mixed composition.

MATERIALS AND

METHODS

A total of 12 PDS harvested from general surgery department in our hospital were analyzed with micro-CT as a reference standard for CA and NCA composition. These stones were placed in a TOS water phantom of 35 cm diameter to simulate standard adult body size. High- and low-energy image sets were acquired from SS DECT scans with high/low tube potential pairs of 80 kVp/140 kVp. All the image sets were imported into an in-house software for further post-processing. CT number ratio (CTR), defined as the ratio of the CT number at 80 kVp to 140 kVp was calculated for each pixel of the images. Threshold was preset between 1.00 and 1.25 to classify CA and NCA components. Pixels in PDS with CTR higher than the threshold were classified as CA, and those with CTR lower than the threshold were classified as NCA. The percentages of CA and NCA for each stone were determined by calculating the number of CA and NCA pixels. Finally, the minimal, maximal and root-mean-square errors (RMSE) of composition measured by SS DECT under each threshold were calculated by referring to the composition data from micro-CT. The optimal threshold was determined with the minimal RMSE. A paired t test was used to compare the stone composition determined by DECT with micro-CT.

RESULTS

The optimal CTR threshold was 1.16, with RMSE of 6.0%. The minimum and maximum absolute errors were 0.22% and 11.35%, respectively. Paired t test showed no significant difference between DECT and micro-CT for characterizing CA and NCA composition (p = 0.414).

CONCLUSION

SS DECT is a potential approach for quantifying and differentiating CA and NCA components in PDS with mixed composition.