Development and intra-institutional and inter-institutional validation of a comprehensive new hepatobiliary software: Part 1--Liver and gallbladder function

Abstract

Objective: To develop a software tool for quantification of liver and gallbladder function, and to assess the repeatability and reproducibility of measurements made with it.

Materials and methods: The software tool developed with the JAVA programming language uses the JAVA2 Standard Edition framework. After manual selection of the regions of interest on a 99mTc hepatic iminodiacetic acid study, the program calculates differential hepatic bile flow, basal duodeno-gastric bile reflux (B-DGBR), hepatic extraction fraction (HEF) of both the lobes with deconvolutional analysis and excretion half-time with nonlinear least squares fit. Gallbladder ejection fraction, ejection period (EP), ejection rate (ER), and postcholecystokinin (CCK) DGBR are calculated after stimulation with CCK-8. To assess intra-observer repeatability and intra-observer reproducibility, measurements from 10 normal participants were analyzed twice by three nuclear medicine technologists at the primary center. To assess inter-site reproducibility, measurements from a superset of 24 normal participants were also assessed once by three observers at the primary center and single observer at three other sites.

Results: For the 24 control participants, mean+/-SD of hepatic bile flow into gallbladder was 63.87+/-28.7%, HEF of the right lobe 100+/-0%, left lobe 99.43+2.63%, excretion half-time of the right lobe 21.50+6.98 min, left lobe 28.3+/-11.3 min. Basal DGBR was 1.2+/-1.0%. Gallbladder ejection fraction was 80+/-11%, EP 15.0+/-3.0 min, ER 5.8+/-1.6%/min, and DGBR-CCK 1.3+/-2.3%. Left and right lobe HEF was virtually identical across readers. All measures showed high repeatability except for gallbladder bile flow, basal DGBR, and EP, which exhibited marginal repeatability. Ejection fraction exhibited high reproducibility. There was high concordance among the three primary center observers except for basal DGBR, EP, and ER. Concordance between the primary site and one of the other sites was high, one was fair, and one was poor.

Conclusion: New United States Food and Drug Administration-approved personal computer-based Krishnamurthy Hepato-Biliary Software for quantification of the liver and gallbladder function shows promise for consistently repeatable and reproducible results both within and between institutions, and may help to promote universal standardization of data acquisition and analysis in nuclear hepatology.