Virtual Analysis of Pancreatic Cystic Lesion Fluid Content by Ultrasound Acoustic Radiation Force Impulse Quantification

Abstract

The purpose of this study was to prospectively evaluate the application of acoustic radiation force impulse (ARFI) imaging implemented with Virtual Touch tissue quantification (Siemens AG, Erlangen, Germany) in the study of pancreatic cystic lesions by using different analysis methods compared with the final diagnosis (pathologic or by magnetic resonance imaging and endoscopic sonographic findings). Thirty-eight patients with pancreatic cystic focal lesions (diameter >3 cm and located at a depth of 5.5 cm) were included in the study and underwent conventional sonography. For every patient, 5 measurements in the Virtual Touch tissue quantification region of interest were obtained. To distinguish mucinous (potentially malignant) from serous (mainly benign) cystic lesions, the result XXXX/0 was considered to mean simple liquids (comparable to water), and the accuracy of Virtual Touch tissue quantification in differentiating pancreatic cystic lesions was calculated. To consider a lesion as containing complex fluids (potentially mucinous), two different reading methods were applied: (1) at least 2 numerical values when obtaining 5 measurements; and (2) the prevalence of numerical values irrespective of the number of measurements. The sensitivity, specificity, positive and negative predictive values, and accuracy were calculated for the differential diagnosis between mucinous and nonmucinous cystic lesions. By the first reading method, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy in the group of cystic lesions were 68.8%, 77.3%, 68.8%, 77.3%, and 73.7%, respectively; by the second method, the values were 37.5%, 100%, 100%, 68.8%, and 73.3%. Acoustic radiation force impulse imaging with Virtual Touch tissue quantification can have a role in the noninvasive characterization of pancreatic cystic lesions during conventional sonographic examinations.