RESUMEN
PURPOSE: We aimed to evaluate the agreement of single criteria and dedicated scores from transabdominal ultrasound of the pancreas (US) compared to standards by endoscopic ultrasound (EUS) and computed tomography (CT). MATERIALS AND METHODS: In this observational cohort study performed in a tertiary care center, US and EUS were performed in 110 patients referred for suspected CP. Based on the Mayo score, 52 patients were diagnosed with CP. The sonographic findings obtained by both methods were registered. The number of criteria was counted and scored according to the Rosemont score. RESULTS: Agreement between the number of detected US and EUS criteria was substantial (ICCâ=â0.74 [0.61-0.83]. Adding Rosemont weighting improved the agreement (ICCâ=â0.88 [0.81-0.92]). Regarding individual criteria, the agreement was substantial for the detection of calcifications (κâ=â0.86) and moderate for cysts and irregular or dilated pancreatic duct (κâ=â0.42-0.58). Agreement for the other criteria was poorer (κ≤â0.40). The diagnostic performance indices [95â% CI] of US for diagnosing CP (using Mayo score as reference standard) were for the unweighted score: Sensitivity: 0.65 [0.51-0.78], specificity: 0.97 [0.87-1.00]; and for Rosemont score: Sensitivity: 0.75 [0.61-0.86], specificity: 0.95 [0.83-0.99]. CONCLUSION: The agreement between US and EUS for the unweighted and weighted scores was substantial. For the features calcifications, cysts and main pancreatic duct (MPD) changes, agreement was moderate to substantial. For the other detected US criteria, the agreement with EUS was too poor to be clinically relevant.