Detection of focal liver lesions: from the subjectivity of conventional ultrasound to the objectivity of volume ultrasound

Abstract

This study was undertaken to establish the reliability of automated volumetric liver scans in detecting focal liver lesions by evaluating the degree of agreement between conventional and volume ultrasound (US) examinations. Over a period of 3 months, we prospectively studied 100 consecutive patients (36 men and 64 women; age range 15-87 years; mean age 63 years) referred to our institute for US imaging of the liver. Volumetric acquisition of the liver was achieved with a 3D transducer (2.0-5.0 MHz) and a Logiq 9 US scanner. All patients underwent both 2- and 3D US studies performed by two expert radiologists. Volumetric acquisitions were subsequently reviewed by the second radiologist, who was blinded to the first radiologist's report. Three categories were established: 1=presence of focal liver lesions; 2=doubtful finding; 3=absence of focal liver lesions. Concordance between volume US and conventional US was calculated by using the k statistic. Out of 100 patients examined, 39 were found to be affected by focal liver lesions. All volume US examinations were technically adequate, allowing exploration of all hepatic sectors, except for five cases that were marred by major respiratory motion artefacts. Conventional and volume US identified the same number of focal liver lesions, with the exception of four cases of doubtful findings at volume US. Concordance between automated volume US and conventional US of the liver was high (k=0.92). The identification of focal liver lesions on automated volume US is possible, and the examination shows a high level of concordance with conventional US.