Reliability of the hepatic perfusion index for the detection of liver metastases

The reliability of the technique for measuring the hepatic perfusion index (HPI) for the detection of liver metastases has been examined. Dynamic images from 173 patients with primary and recurrent colorectal cancer have been analysed by two operators using the peak values of both the right and left kidney time-activity curves. In 14.4% of analyses the right kidney HPI method failed, while only 5.7% of analyses failed using the left kidney method. The results showed (a) that disagreement between the operators HPI values was small but statistically significant (p less than 0.05) and (b) that disagreement between the two methods was highly significant (p less than 10(-7)), with the left kidney method resulting in a lower HPI value. An upper limit of normal for left kidney HPI of 0.37 was predicted from the data, assuming the corresponding right kidney value to be 0.40.