Aim: Compare the performances of EUS to helical CT in the diagnosis and staging of pancreatic adenocarcinoma.

Methods: Forty two consecutive patients (mean age 63 years; 25 men, 17 women) who had surgical exploration and histologically proved pancreatic cancer were retrospectively included. All our patients underwent with endoscopic ultrasonography (EUS) and helical computed tomography (helical CT). Data analysis compared helical CT, EUS with the surgical data with or without histological study in diagnosis, staging and resectability of pancreatic cancer. Surgical findings were used as gold standard.

Results: For positive diagnosis EUS was more sensitive 100% (CI:93-100) than helical CT 88% (CI:77-95). But helical CT was more specific 89% (CI:64-98) than EUS 83% (CI:58-96) for small tumors whose diameter is below 2.5 cm in witch EUS was more sensitive in their detection (100% versus 83%). In evaluating venous involvement EUS was more sensitive than helical CT (96% versus 50%; p<0.05), while CT was more specific (81% versus 75%; p<0.05). Regarding lymph nodes invasion, the two imaging technique had the same sensibility (56%) with better specificity for helical CT (83% versus 75%; p<0.05). The accuracy of EUS in identifying the T and N stages were 80% and 67% respectively, while helical CT have an accuracy of 50% and 71% respectively. EUS and helical CT correctly identified all resectable tumors while EUS was more accurate than helical CT in detecting non resectable tumors 94% versus 69%.

Conclusion: EUS remains superior to helical CT in positive diagnosis of pancreatic adenocarcinoma especially for small tumors and also for the diagnosis of venous invasion and in identifying non resectable tumors. The two techniques have the same accuracy in the detection of lymph node involvement.