Three-dimensional ultrasound for volume measurement of liver tumors

In a clinical trial the accuracy of volumetry by use of three-dimensional u ltrasound (3-D-US) in comparison to three-dimensional computer tomography ( 3-D-CT) was evaluated. Overall, 42 patients with focal hepatic lesions were investigated with 2-D- and 3-D ultrasound. In 11 patients additional compu terized tomographic arterial portography using spiral technique was perform ed. The volumes of the lesions were calculated using the ellipsoid formula (for 2-D- and S-D-US) as well as with a planimetric reconstruction for 3-D- US and 3-D-CT. In addition the intra- and inter-investigator variability of 3-D ultrasound was determined. The volume of the investigated liver lesion s (planimetric reconstruction with 3-D ultrasound) ranged between 1.5 cm(3) and 1231 cm(3) with a mean volume of 155 cm(3) and a median volume of 65 c m(3). The deviation of 2-D-US vs 3-D-CT was - 62 % to + 68 %, in the case o f 3-D-US (ellipsoid) vs 3-D-CT it was - 28% to + 9%, and for 3-D-US (planim etry) vs 3-D-CT it was -21% to + 9 %. The concordance index kappa was 0.886 , showing very good agreement between the two investigators. The intra-inve stigator variability was 5%. Our data show that volume measurement by use o f 3-D-US is independent of the investigator. With regard to accuracy of vol ume measurements 3-D-US is comparable to 3-D-CT but more precise than 2-D-U S. These results indicate that 3-D-US may be applied in the follow-up of tu mor patients as an alternative diagnostic procedure to computer tomography. In addition 3-D-US might: be useful in planning liver resections by virtue of better evaluation of the volume of the liver tissue remaining after res ection and better visualization of the topography of liver tumors and major hepatic vessels.