Background: The optimal approach for detecting small pancreatic tumors is uncertain. We compared multidetector CT (MDCT) with follow-up endoscopic ultrasonography (EUS) without or with fine-needle aspiration (EUS-FNA) for diagnosing pancreatic cancer.
Methods: Patients with suspicion of pancreatic cancer who underwent dual-phase MDCT and follow-up EUS were retrospectively reviewed. This consisted of scoring MDCT scans independently by three radiologists on a 1-5 scale of certainty, determining whether a stent was present, scoring EUS reports regarding presence of a mass and analyzing EUS-FNA results.
Results: A total of 117 patients underwent MDCT and EUS. ROC values for MDCT were 0.85, 0.87, and 0.91. There was no significant difference in the accuracy of EUS and MDCT. Follow-up EUS (99%) was significantly more sensitive than MDCT (89% and 93%), as interpreted by two radiologists. Follow-up EUS was statistically significantly more sensitive than MDCT (96% vs. 70%) for one radiologist for tumors < 2 cm. Specificity of EUS was 50%, and sensitivity of EUS-FNA was 82%. Negative predictive value of EUS-FNA was significantly less in patients with (21%) than without (70%) biliary stents.
Conclusions: Follow-up EUS improves lesion detection over MDCT alone. Close follow-up/repeat biopsy should be considered if FNA is negative, but EUS is positive.