Résumé
PURPOSE: We aimed to evaluate objective ultrasonography (US) findings for diagnosis of liver cirrhosis and to correlate the utility between the use of the low frequency and high frequency probes for images analyzed on the picture archiving and communications system (PACS). MATERIALS AND METHODS: A total of 87 patients participated in the study; 19 patients that had a clinically proven normal liver and 68 patients that had biopsy-proven chronic liver disease were evaluated with the use of US for the status of the liver. Multiple variables such as a 'smooth surface,' 'irregular nodular surface,' 'homogeneous echotexture,' 'heterogeneous appearance mixed with hypoechoic and hyperechoic echotexture' and 'hypoechoic honeycomb like echotexture' were evaluated based on images obtained with the use of both low frequency and high frequency probes by two observers. The diagnoses obtained after US were correlated with the histological results using assess agreement as statistical method. The concordance rate was calculated to correlate the utility between the use of the low frequency and high frequency probes. RESULTS: An 'irregular nodular surface' showed high interobserver agreement for the diagnosis of liver cirrhosis on images obtained with the low frequency probe (kappa= 0.61). In addition, 'hypoechoic honeycomb' showed high interobserver agreement on images obtained with the high frequency probe (kappa= 0.60). The use of the low frequency probe was associated with more accuracy for the diagnosis of liver cirrhosis (73%). CONCLUSION: Objective US findings for the diagnosis of liver cirrhosis are an 'irregular nodular surface' on images obtained with use of a low frequency probe and a 'hypoechoic honeycomb' on images with use of a high frequency probe.