Analysis of free-form radiology dictations for completeness and clarity for pancreatic cancer staging

Purpose: To assess the completeness and clarity of current free-form radiology reports for pancreatic cancer staging by evaluating them against the elements of the RSNA CT oncology primary pancreas mass dictation template. Methods: This retrospective study was approved by our Institutional Review Board (IRB). 295 free-form computed tomography (CT) reports for baseline staging of pancreatic cancer (PC) generated between August 2008 and December 2010 were evaluated by one of two radiologists with expertise in pancreatic cancer imaging. Reports which indicated that metastatic disease was present were excluded. The completeness and clarity of the reports were analyzed against the elements of the RSNA CT pancreas mass dictation template. Fisher’s exact tests were used to analyze differences by year and type of radiologist. Results: Primary lesion location, size, and effect on bile duct (BD) were provided in 93.9% (277/295), 69.8% (206/295), and 67.5% (199/295) of reports, respectively. Standard terms to describe vascular involvement were used in 47.5% (140/295) of reports. In 20.3% (60/295), the resectability status could not be defined based on the report alone. In 36.9% (109/295) of reports, review of CT images was necessary to understand vascular involvement. Radiologists expert in pancreatic oncology had a higher proportion of reports using standardized terminology and reports in which vascular involvement was understood without revisiting the images. Conclusions: Free-form reports were more likely to use ambiguous terminology and/or require review of the actual images for understanding resectability status. The use of a standardized reporting template may improve the usefulness of pancreatic cancer staging reports.