Endoscopic ultrasound and Computed Tomography in the diagnosis, locoregional staging and assessment of vascular infiltration of pancreatic carcinoma

PURPOSE: The aim of this study was to evaluate the diagnostic accuracy of colour-Doppler Endoscopic Ultrasonography (EUS), in the detection, loco-regional staging and assessment of vascular infiltration in pancreatic carcinoma, and to compare the results with those obtained by Computed Tomography (CT). MATERIALS AND METHODS: A series of 57 patients with diagnosed or suspected pancreatic carcinoma was retrospectively analysed. All patients underwent EUS and thin-slice (< 5 mm) spiral dynamic CT. The final diagnosis (carcinoma in 37 patients and benign lesion in 20) was obtained by laparotomy in 21 patients, fine-needle aspiration cytology (FNAC) in 17, and follow-up in 19. RESULTS: The specificity and sensitivity for the diagnosis of malignancy were respectively 45% and 92% for EUS and 45% and 89% for CT, with an accuracy of 75% for EUS (p <0.05) and 74% for CT (p = 0.07). The specificity and sensitivity for the diagnosis of loco-regional nodal metastases were both 100% for EUS. The specificity and sensitivity for the diagnosis of vascular infiltration were 100% and 94% for EUS and 100% and 44% for CT, giving a diagnostic accuracy of 97% for EUS vs 74% for CT (p <0.001). CONCLUSIONS: EUS proved to be more sensitive and specific than CT in the loco-regional staging of pancreatic carcinoma. Its diagnostic accuracy is especially high in assessing vascular infiltration and loco-regional nodal metastases. CT still remains the examination of choice for staging pancreatic carcinoma and for assessing its resectability as it affords a panoramic view and ability to rule out distant metastases. Candidates to resection should all be examined by EUS, as, due to its high accuracy in loco-regional staging and assessing vascular infiltration, it might allow a large proportion of patients to be spared the operation.