Reproducibility of VPCT parameters in the normal pancreas: comparison of two different kinetic calculation models

To assess the reproducibility of volume computed tomographic perfusion (VPCT) measurements in normal pancreatic tissue using two different kinetic perfusion calculation models at three different time points.

Institutional ethical board approval was obtained for retrospective analysis of pancreas perfusion data sets generated by our prospective study for liver response monitoring to local therapy in patients experiencing unresectable hepatocellular carcinoma, which was approved by the institutional review board. VPCT of the entire pancreas was performed in 41 patients (mean age, 64.8 years) using 26 consecutive volume measurements and intravenous injection of 50 mL of iodinated contrast at a flow rate of 5 mL/s. Blood volume(BV) and blood flow (BF) were calculated using two mathematical methods: maximum slope + Patlak analysis versus deconvolution method. Pancreas perfusion was calculated using two volume of interests. Median interval between the first and the second VPCT was 2 days and between the second and the third VPCT 82 days. Variability was assessed with within-patient coefficients of variation (CVs) and Bland-Altman analyses. Interobserver agreement for all perfusion parameters was calculated using intraclass correlation coefficients (ICCs).

BF and BV values varied widely by method of analysis as did within-patient CVs for BF and BV at the second versus the first VPCT by 22.4%/50.4% (method 1) and 24.6%/24.0% (method 2) measured in the pancreatic head and 18.4%/62.6% (method 1) and 23.8%/28.1% (method 2) measured in the pancreatic corpus and at the third versus the first VPCT by 21.7%/61.8% (method 1) and 25.7%/34.5% (method 2) measured also in the pancreatic head and 19.1%/66.1% (method 1) and 22.0%/31.8% (method 2) measured in the pancreatic corpus, respectively. Interobserver agreement measured with ICC shows fair-to-good reproducibility.

VPCT performed with the presented examinational protocol is reproducible and can be used for monitoring purposes. Best reproducibility was obtained with both methods for BF and with method 2 also for BV data for both follow-up studies.