Pathologists currently use semi-quantitative image analysis based on scoring evaluation (NAS or SAF score) to assess NAFLD/NASH severity from human liver biopsies. This scoring evaluation is also widely used for pre-clinical assessment of the effectiveness of lead molecules in NAFLD/NASH rodent models. Biocellvia decided to perform a reader-independent NAFLD / NASH histological analysis.
Histological scoring is carried out on hematoxylin-eosin and Masson’s trichrome-stained slides in a masked manner by specialised pathologists. Individual features of NAFLD/NASH are scored according to the CNR pathology committee [Kleiner et al., 2005, 2019] and look at the steatosis, ballooning and lobular inflammation. A NAFLD Activity Score (NAS score) is defined corresponding to the sum of the severity scores for steatosis, lobular inflammation, and ballooning (range 0-8). In addition to the NAS scores the estimate of NAFLD/NASH is determined by the staging of fibrosis, scored from Masson’s trichrome-stained slides, according to the amount of collagen fibres in selected lobular regions (perisinusoidal and portal – range 0-4). The combination of scoring assessment of these various features of NAFLD/NASH provides a single semi-quantitative value that allows pathologists to establish an ultimate diagnosis.
Although scoring evaluation is widely used currently for both clinical and pre-clinical NAFLD/NASH evaluation, it is reader-dependent. Consequently, it is subject to significant intra- and inter-variabilities as reflected by the low statistical Kappa index of the whole features of NAFLD/NASH [Kleiner et al., 2019]. These variabilities can be significant barriers not only for the reliable clinical assessment of NAFLD and NASH diagnoses but also when determining the efficacy of molecules in pre-clinical development. In the pre-clinical development of lead molecules, the evaluation of their effectiveness by scoring, i.e. discontinuous values on a reduced scale, is not suitable for obtaining accurate and reliable measurements. This scoring may mask the true efficacy of lead molecules and ultimately lead to inconsistent results and misinformed decisions.
Biocellvia’s research team, composed of software engineers and scientists, has developed a new quantitative NAFLD / NASH histological analysis, which overcomes the limitations of scoring. Biocellvia’s analysis is based on computerised assisted discrimination of the main features of NAFLD/NASH disease, namely: steatosis, inflammation and fibrosis. Considering the significant discrepancies found currently in the literature for an accurate and reliable discrimination of ballooned hepatocytes, quantitative analysis of ballooning has been temporally discarded from our field of analysis and we are exploring this as a part of our research and development.
To eliminate any bias due to the reader-dependent evaluation of liver morphometric parameters, our quantitative analysis has been fully automated from the digitalisation of histology slides through to the tabulated data and report. Biocellvia’s algorithms have been developed in-house and designed for use with various rodent models as well as human biopsies.
Quantitative analysis of steatosis and fibrosis is carried out concomitantly on the same slide stained with picrosirius red and inflammation is analysed on serial slides immunostained with F4/80. Steatosis is defined by the area of lipid vesicles, and fibrosis by the area of collagen fibres specifically stained with picrosirius red. Inflammation is determined by the area of F4/80 immunolabelled crown-like structures formed around hepatocytes by active macrophages. The fully automated software provides simultaneous quantitative analysis of steatosis, fibrosis and inflammation with a very short turnaround time (<1h for 100 slides).
Biocellvia’s NAFLD/NASH analysis has been validated in two recent articles published in collaboration with Pr. Saurabh Chatterjee and Anna Mae Diehl (Duke University, USA) and Pr. Isabelle Leclercq (Leuven University, Belgium) [Albadrani et al., 2019; De Rudder et al., 2019]. These studies demonstrated that Biocellvia’s NAFLD/NASH quantitative analysis betters traditional scoring analysis in terms of accuracy, reliability, reproducibility and speed.
The quantified histology of NAFLD/NASH developed by Biocellvia opens a new and significant approach for the accurate and reliable evaluation of liver diseases. This level of precision cannot be achieved through the use of standard scoring analysis. Furthermore, Biocellvia’s NAFLD / NASH histological analysis is not limited to pre-clinical in-vivo investigations; it is used to analyse other pre-clinical models such as spheroids. The NAFLD/NASH analysis is also used to quantify human liver biopsies, providing a very rapid RUO insight into a compound’s efficacy. This robust, objective and fully automated analysis eliminates variability and provides researchers with better insights into their compounds’ efficacy.
Toward a reader-independent NAFLD-NASH histological analysis (download it as a PDF file)