Digital scanning and diagnostic scoring of kidney biopsies from children with steroid resistant nephritic syndrome
Background: Digital pathology is an attractive new tool for multicenter research. Practical improvements on conventional pathology include parallel evaluation by several experts in distant locations and reproducible identification of individual lesions. Also, digital image analysis allows more objective measures of cell staining and 3D reconstruction of glomeruli.
Design: We are establishing a digital pathology archive within WP2 of EURenOmics. Kidney biopsies from children with steroid resistant nephrotic syndrome (SRNS) enrolled in the PodoNet registry are collected. Analysis will be harmonized with digital pathology scoring developed by the NEPTUNE study (Nephrotic Syndrome Study Network) in the US.
Methods: The material is scanned at high resolution in Heidelberg at the Hamamatsu Tissue Image and AnalysisCenter using a Hamamatsu Nanozoomer. Digital images of existing electron microscopy and immunofluorescence scans and original pathology reports are also collected. After anonymization and manual quality controls (e.g. checking correct focus in all areas) data is stored centrally for remote review.
Digital images are first annotated, i.e. each glomerulus is given a number which can be tracked across different section levels. Histopathological scoring of each glomerulus in each section level will then be performed in an unbiased fashion by independent blinded pathologists. These scores will be correlated to clinical and genetic information, conventional histopathological diagnoses and molecular profiles obtained by multi-omics profiling.
Progress: So far 83 kidney biopsies have been collected and at least another 180 are expected. Test runs of different stain types have produced good quality images. Bulk scanning and annotations are commencing, so that pathology reviews will start in spring 2014.
Outlook: The new digital pathology archive for children with SRNS will enable standardized review and objective scoring of histopathological findings. We hope that this will improve correlations of histological findings with clinical outcomes and biomarkers. Setting up the relevant infrastructure will allow extension of the project to other registries within WP2.
