Bernd Schnabl1, Salman Farshchi-Heydari2, Rohit Loomba1, Robert F Mattrey3, Carl K Hoh3, Claude B Sirlin3, David A Brenner1, Cynthia A Behling4, David R Vera5. 1. Department of Medicine, University of California, San Diego, La Jolla, CA. 2. UCSD In Vivo Cancer & Molecular Imaging Program, University of California, San Diego, La Jolla, CA. 3. UCSD In Vivo Cancer & Molecular Imaging Program, University of California, San Diego, La Jolla, CA; Department of Radiology, University of California, San Diego, La Jolla, CA; Moores UCSD Cancer Center, University of California, San Diego, La Jolla, CA. 4. Analytic Pathology Medical Group, Sharp Memorial Hospital, San Diego. 5. UCSD In Vivo Cancer & Molecular Imaging Program, University of California, San Diego, La Jolla, CA; Department of Radiology, University of California, San Diego, La Jolla, CA; Moores UCSD Cancer Center, University of California, San Diego, La Jolla, CA; Department of Surgery, University of California, San Diego, La Jolla, CA. Electronic address: dvera@ucsd.edu.
Abstract
OBJECTIVES: The aim of this study was to test the ability of hepatocyte-specific functional imaging to stage fibrosis in experimental rat models of liver fibrosis and progressive NASH. Using ROC analysis we tested the ability of a functional imaging metric to discriminate early (F1) from moderate (F2) fibrosis in the absence and presence of non-alcoholic steatohepatitis, which has not been achieved by any modality other than biopsy. METHODS: Galactosyl Human Serum Albumin (GSA) was radiolabeled with the positron-emitter, (68)Ga, and injected (i.v., 45-95 μCi, 1.5 pmol/g TBW) into 44 healthy, 19 DEN-, and 22 CDAA-treated male rats. Quantification of liver function was achieved by calculating T90, defined as the time for the liver to accumulate 90 percent of the [(68)Ga]GSA plateau value. All livers were excised immediately after imaging and prepared for a "blinded" histologic examination, which included fibrosis and fat content scores. Two sets of fibrosis scores were recorded for all of animals. The dominant fibrosis stage was recorded as the "Dominant Pattern" score and the "Maximum Pattern" score was assigned if a smaller distinct region with a higher fibrosis score was observed. RESULTS: Animals with Dominant Pattern F0-F1 liver fibrosis (D(-)=39) demonstrated significantly (P<0.0001) faster accumulation of [(68)Ga]GSA (2.40 ± 0.52 min) than those with moderate to advanced Dominant Pattern fibrosis F2 and F4 (D(+)=26) (3.48 ± 1.01 min). ROC analysis (F0-F1 vs F2-F4) produced an area under the binormal curve (AUC) of 0.867 ± 0.045. Twenty-seven of the 65 rats had small regions with higher fibrosis scores. Six of these Maximum Pattern scores reclassified the animals from D(-) to D(+). ROC analysis of F0-F1 versus F2-F4 rats without liver fat produced AUCs of 0.881 ± 0.053 for the Dominant Pattern Score and 0.944 ± 0.035 for the Maximum Pattern Score. CONCLUSIONS: PET Functional Imaging of [(68)Ga]GSA accurately discriminates early from moderate experimental fibrosis independent of steatosis grade. If validated in human studies, molecular imaging may emerge as a potential alternative to invasive liver biopsy.
OBJECTIVES: The aim of this study was to test the ability of hepatocyte-specific functional imaging to stage fibrosis in experimental rat models of liver fibrosis and progressive NASH. Using ROC analysis we tested the ability of a functional imaging metric to discriminate early (F1) from moderate (F2) fibrosis in the absence and presence of non-alcoholic steatohepatitis, which has not been achieved by any modality other than biopsy. METHODS: Galactosyl HumanSerum Albumin (GSA) was radiolabeled with the positron-emitter, (68)Ga, and injected (i.v., 45-95 μCi, 1.5 pmol/g TBW) into 44 healthy, 19 DEN-, and 22 CDAA-treated male rats. Quantification of liver function was achieved by calculating T90, defined as the time for the liver to accumulate 90 percent of the [(68)Ga]GSA plateau value. All livers were excised immediately after imaging and prepared for a "blinded" histologic examination, which included fibrosis and fat content scores. Two sets of fibrosis scores were recorded for all of animals. The dominant fibrosis stage was recorded as the "Dominant Pattern" score and the "Maximum Pattern" score was assigned if a smaller distinct region with a higher fibrosis score was observed. RESULTS: Animals with Dominant Pattern F0-F1 liver fibrosis (D(-)=39) demonstrated significantly (P<0.0001) faster accumulation of [(68)Ga]GSA (2.40 ± 0.52 min) than those with moderate to advanced Dominant Pattern fibrosis F2 and F4 (D(+)=26) (3.48 ± 1.01 min). ROC analysis (F0-F1 vs F2-F4) produced an area under the binormal curve (AUC) of 0.867 ± 0.045. Twenty-seven of the 65 rats had small regions with higher fibrosis scores. Six of these Maximum Pattern scores reclassified the animals from D(-) to D(+). ROC analysis of F0-F1 versus F2-F4 rats without liver fat produced AUCs of 0.881 ± 0.053 for the Dominant Pattern Score and 0.944 ± 0.035 for the Maximum Pattern Score. CONCLUSIONS: PET Functional Imaging of [(68)Ga]GSA accurately discriminates early from moderate experimental fibrosis independent of steatosis grade. If validated in human studies, molecular imaging may emerge as a potential alternative to invasive liver biopsy.