[Direct integral linear least square regression method for kinetic evaluation of hepatobiliary scintigraphy: clinical...

Kinetic evaluation of 170 hepatobiliary scintigrams was performed using direct integral linear least square regression method (DILS method). Liver time-activity curve (TAC) data were fitted to 2 types of 1 compartment model using heart TAC as input function. One model was 1 compartment model with 2 parameters (extraction rate (ku) and excretion rate (ke)), the other had another parameter for representing non-specific volume of distribution (Vn) in liver. Obtained parameter estimate was then examined to see how they correlate with three major visual findings of scintigram; low tracer uptake, intrahepatic cholestasis, extrahepatic cholestasis. In both models, ku showed significant difference between low uptake positive and negative groups and ke showed significant difference between extrahepatic cholestasis positive and negative groups. Only mean uptake time (Vn/ku) by 3 parameter model showed significant difference between intrahepatic cholestasis positive and negative groups, although the other 2 parameters (ku, ke) were insensitive to intrahepatic cholestasis in either model. Thus, the 3 parameter model was superior to the 2 parameter model because of its sensitivity to intrahepatic cholestasis and ability to differentiate extrahepatic and intrahepatic cholestasis as the difference of parameter values. Based on this 3 parameter model, parametric images were generated by calculating parameter value of each pixel using the DILS method. The DILS method could determine parameters by simple calculation, and hence was very powerful in processing parametric images, which processing otherwise required immense calculation.