There is a need for further refinement of current histological systems for assessment of hepatic fibrosis in nonalcoholic fatty liver disease (NAFLD). This study evaluated hepatic fibrosis in NAFLD using dual-photon microscopy-based quantitation of fibrosis-related parameters (q-FPs). Fifty test cohort subjects and 42 validation cohort subjects with NAFLD and the full spectrum of fibrosis were studied. q-FPs were measured in specific predefined regions of interest (general, vessel, perisinusoid, and vascular septa). Seventy q-FPs had inter- and intraobserver concordance ≥0.8 and were related to the NASH Clinical Research Network fibrosis staging. Of these, 16 q-FPs with the strongest correlations (P < 0.001 for all) were entered in a principal component analysis model (odds ratio [OR] 7.8, P < 0.001), which separated any stage of fibrosis versus no fibrosis, and cirrhosis versus earlier stages with the areas under the receiver operating characteristic curves of 0.88 and 0.93 (P ≤ 0.01 for both), respectively. In an independent multivariable analysis, four q-FPs-the number of collagen strands (OR 8.5, P = 0.004), strand length (OR 12.0, P = 0.02), strand eccentricity (OR 8.3, P = 0.004), and strand solidity (OR 8.0, P = 0.003)-were independently associated with fibrosis stages and were used to model fibrosis along a continuous linear scale using desirability functions; this linear scale of fibrosis measurement was also related to fibrosis stage (P < 0.0001). The robustness of both the multivariable model and the linear scale of measurement was confirmed in the validation cohort.
Conclusion: The q-FP model provides an accurate reproducible method to evaluate fibrosis in NAFLD along a quantitative and continuous scale. (Hepatology 2017;65:1891-1903).
© 2017 by the American Association for the Study of Liver Diseases.