Identification of liver disease by radionuclide tomographic imaging. Results of a prospective clinical trial and potential for further improvement

A prospective clinical trial has been undertaken to assess the contribution of a tomographic section view to conventional radionuclide imaging of the liver. The trial protocol is presented and some problems associated with analysis of the collected data by rating curves and ROC type curves are discussed. Results show that the tomographic section view improves the sensitivity of the examination, especially for focal liver disease where 91% successful detection of positives was achieved. The sensitivity for diffuse liver disease was appreciably lower. The incidence of false positive reports also increased when the tomographic section was included, such that the accuracy of the examination showed no overall improvement. Several developments either under way or under consideration may improve tomographic imaging and the manner in which images of the liver obtained by this method are interpreted. In due course better overall accuracy could result, so the present work provides an important quantitative base-line against which any such improvement can be monitored. (author)