Split-bolus pancreas CTA protocol for local staging of pancreatic cancer and detection and characterization of liver lesions.

PURPOSE: To validate the use of a split-bolus pancreas CTA protocol for local staging of pancreatic cancer and to evaluate its ability to detect and characterize liver lesions. METHODS: Consecutive patients with pancreatic cancer who underwent split-bolus pancreas CTA between 12/2015 and 12/2016 were included in this IRB-approved HIPAA-compliant retrospective study. Objective evaluation of the abdominal vessels, the pancreas, the liver, and lesions, if present, was performed with attenuation measurements and tumor conspicuity and contrast-to-noise ratio (CNR) calculations. An abdominal radiologist with 20 years of experience performed subjective evaluation of image quality and blindly detected and characterized liver lesions. Any inconclusive findings or grading scores were evaluated in consensus with another abdominal radiologist with 7 years of experience. Liver findings were validated using a composite reference standard to assess accuracy. RESULTS: There were 82 pancreatic cancer patients with a total of 91 liver findings. Tumor conspicuity and CNR were 60.8 ± 35.1 HU and 8.0 ± 5.8 for the pancreatic lesions and 58 ± 34.7 HU and 9.7 ± 6.3 for the liver lesions, respectively. The accuracy, sensitivity, and specificity of the split-bolus protocol for the hepatic findings were correspondingly 89/91 (97.8%, 95% CI 92.3-99.4), 58/60 (96.7%, 95% CI 88.6-99.1), and 33/33 (100%, 95% CI 89.6-100). The subjective image quality ratings were optimal in more than 89% of the cases for various structures, with no non-diagnostic ratings. CONCLUSION: Split-bolus pancreas CTA protocol allows for detection and staging of pancreatic cancer, both for the primary tumor and detection and characterization of liver lesions.