Ductal adenocarcinoma is the most frequent pancreatic solid lesion and the most common tumor of the pancreas. Given its poor prognosis and the major therapeutic consequences, the discrimination between PA and other pancreatic solid lesions is mandatory. EUS is admitted as the most sensitive imaging procedure for the detection and characterization of pancreatic tumors [1-3]. Nevertheless it remains difficult to differentiate, on morphological features, PA from other solid masses. For 15 years, endoscopic ultrasound fine needle aspiration (EUS-FNA) has demonstrated its efficiency for tissue sampling and cyto-histologic diagnosis of PA. However, the negative predictive value (NPV) for the diagnosis of pancreatic adenocarcinoma (PA) remains low (30-70%) in the published prospective series [4]. So, in case of negative result, the choice between surgery and follow-up remains difficult. Additional criteria to get the decision are then warranted.

The assessment of pancreatic tumor enhancement using ultrasound contrast agents (UCAs) in real time with imaging specific methods seems useful to improve their characterization [4-8] either by contrast-enhanced EUS (CE-EUS) or, more recently, by contrast-enhanced harmonic EUS (CH-EUS).

The aims of this prospective multicenter study is:

1. to compare the NPV of contrast-enhanced endoscopic ultrasound (CH-EUS) and EUS-FNA for the diagnosis of PA;
2. to assess the intra- and inter-observer concordances of CH-EUS for the diagnosis of PA.

100 patients with a solid pancreatic mass of indeterminate origin must be prospectively included in 3 French centers Exclusion criteria: presence of a cystic component greater than 25 % of the total volume of the lesion, pregnancy, lactation, age <18 years, and usual contraindications to SonoVue® injection.

All EUS procedure will be performed by 5 experienced endosonographers as follows:

1. Conventional gray-scale B-mode and conventional power Doppler EUS to assess the EUS characteristics of the pancreatic lesion (localization, size, echogenicity, cystic component), the aspect of the surrounding parenchyma as well as the presence of proximal duct dilation, vessels infiltration, and collateral veins; tumor and nodes (uTN) staging. A systematic video of the 2 modes will be recorded.
2. CH-EUS will be performed to assess the microvascularization of the lesion and of the surrounding parenchyma: the echoendoscope will be placed in front of the pancreatic lesion and the contrast-specific mode will be engaged with simultaneous monitoring by fundamental B mode. A mecanical index (MI) of 0.4 will be chose based on previous studies [4-6]. An intravenous 4.8 ml SonoVue® bolus injection will be administrated through an antecubital vein, using a 20 Gauges catheter, followed by 20 ml saline flush. Examination of pancreatic lesion will be evaluated in real time and a video of each examination will be record and store. The examination lasted up to 3 minutes after SonoVue® injection to ensure full examination of the lesion in arterial (hyper echogenicity of the aorta, superior mesenteric, hepatic or splenic arteries) and venous phases (hyper echogenicity of the splenic-mesenteric-portal vessels). The pancreatic lesion enhancement pattern will be compare to the adjacent pancreatic parenchyma. We differentiated 3 patterns: hypo-, iso- or hyperenhancement. The operator classified the lesion as pancreatic adenocarcinoma (PA) or non PA. Based on previous pilot studies, lesion in hypoenhancement consider as PA while lesion in hyper or isoenhancement as non-PA [4-6]. In case of tumors with mixed pattern the lesion was considered as PA if a significant area (>10% of the surface) is in hypoenhancement.
3. EUS-FNA: a specimen will be obtain from all lesions using a 22 Gauge needle. Final diagnosis will be based on pathological findings obtained either surgically or by EUS-FNA. In the absence of histological evidence, follow-up (F-U) of patients for 12 months will be carried out. The diagnosis of PA ruled out if no sign of malignancy (disease regression or absence of disease progression) present at the end of F-U.
4. Images reviewing: at the end of the study an anonymous digital video recording of each procedure including B mode, power Doppler mode and CHE mode will be performed.

Statistical analysis. The McNemar test will be use to compare the CH-EUS performance for the diagnosis of PA to EUS-FNA and final diagnosis. Sensitivity (Se), specificity (Spe), predictive positive value (PPV), negative predictive value (NPV) and accuracy with 95% confidence intervals (95%CI) will be calculate. A p value of 0.05 considered statistically significant. Intra- and interobserver agreements of CH-EUS for the diagnosis of PA will assess using kappa statistics and associated 95% CI. Depending on Kappa values, agreement will considered as minor (0.01-0.20), fair (0.21-0.40), moderate (0.41-0.60), high (0.61-0.80), or almost perfect (0.81-1.00), beyond chance.

Inclusion Criteria:

• patients with a solid pancreatic mass of indeterminate origin

Exclusion Criteria:

• presence of a cystic component greater than 25 % of the total volume of the lesion, pregnancy, lactation, age <18 years, and usual contraindications to SonoVue® injection