The quantification of histological features of Idiopathic Pulmonary Fibrosis (IPF) from human lung biopsies or rodent models relies upon a scoring evaluation established over thirty years ago by Ashcroft and his collaborators (Ashcroft et al., 1988). Although such an evaluation is currently the gold standard, it is not without significant limitations linked to the reader-dependent evaluation and the discontinuous scoring measurement leading consequently to variability and semi-quantitative analysis, respectively.
IPF histological scoring is carried out on hematoxylin-eosin-stained lung sections in a blinded manner by specialized lung pathologists. As established by Ashcroft and his collaborators, the severity of IPF disease uses a scale of increasing scores ranging from 0 to 8. At the time, Ashcroft scoring provided « a detailed description of the changes in a lung, currently not possible with most existing methods» (Ashcroft et al., 1988). However, the variability of quantitative data due to the reader-dependent evaluation is a significant obstacle for both a reliable clinical assessment of IPF diagnosis but also for determining the efficacy of therapeutic molecules. In the preclinical development of lead molecules, the evaluation of their effectiveness by scoring, i.e. discontinuous values on a reduced scale, is not suitable for providing accurate and reliable measurements. Ultimately, Ashcroft scoring evaluation may mask significant efficacy of lead molecules giving rise to inconsistent results and misinformed decisions. If Ashcroft scoring was a breakthrough in 1988, in the following decades, new scoring analyses have emerged which have improved the accuracy of IPF assessment but have not yet eliminated two of the most significant limitations of scoring: reader-dependency and variability.
Biocellvia’s research team, composed of software engineers and scientists, has developed a new quantitative IPF histological analysis, which overcomes the limitations of the scoring evaluation method. Biocellvia’s IPF analysis leverages the computer-assisted discrimination of the two main features: the pulmonary foci and the collagen content. To eliminate any bias due to experimenter-dependent evaluation, Biocellvia’s IPF histological analysis is fully automated from the digitalization of histological lung sections to the tabulated data and their statistical analysis. Biocellvia’s algorithms have been developed in-house and designed for use with various rodent models as well as human biopsies.
Quantitative analysis of IPF is carried out on entire lung sections stained with H&E or picrosirius red. The quantification of pulmonary foci rests on the assessment of the density of pulmonary tissue (Biocellvia’s innovation) whereas that of collagen content on the area of stained collagen fibres. The fully automated software allows a simultaneous quantification of pulmonary foci and collagen content with a very short turnaround time (<1h for 100 slides). It is important to note that Biocellvia’s IPF analysis was the subject of an exhaustive study performed on bleomycin-treated mouse model for validation. This study gave rise to a publication realized in collaboration with Boehringer Ingelheim (Gilhodes et al., 2017). In the past two years, numerous preclinical studies were conducted using Biocellvia’s IPF analysis for determining the effectiveness of lead molecules or performing basic research. These preclinical studies carried out on various rodent models have contributed to establish the robustness of Biocellvia’s IPF analysis.
The quantitative analysis of IPF features developed by Biocellvia opens a new and robust methodological approach which outmatches conventional scoring analysis in terms of accuracy, reliability, rapidity and reproducibility. Furthermore, computerized quantitative analysis allows to conduct a fully automated analysis totally independent of the experimenter and therefore, without any intra- and inter variability, which cannot be achieved using scoring evaluation. Biocellvia’s IPF assay provides researchers with the reproducible quantification that is critical in determining the efficacy of their candidate molecules, giving them the insightful data that leads to better decisions.