Histopathological evaluation of Onchocerca volvulus nodules by microscopy and by digital image analysis for the study of macrofilaricidal drug efficacy

BACKGROUND: Novel drugs or drug combinations that kill or permanently sterilize adult Onchocerca volvulus worms would be very helpful for treatment and elimination of onchocerciasis. In absence of a reliable biomarker for viable adult worms, histopathological assessment of worms within onchocercal nodules is a standard method to determine macrofilaricidal activity. The goal of the present study was to determine the agreement between two independent experts in the analysis of nodule sections and to assess the value of digital imaging as a means of standardizing the analysis. MATERIAL AND METHODS: Two expert microscopists independently assessed 605 nodules by direct microscopy. At least two sections with two different stains hematoxylin & eosin (H&E, APR immunostain) of paraffin-embedded, ethanol-fixed whole-nodule cross-sections were analyzed. After variables were identified prone to observer discrepancies, we performed a second study to compare consolidated results for 100 nodules obtained by the two readers by microscopy and by analysis of scanned, high resolution digital images (20x magnification). The last data set analyzed was a quality panel of 100 nodules that has been previously examined by microscopy, and included additional immunostains for Wolbachia endobacteria. These slides were digitalized, read by the two assessors and results were compared with original microscopy results. RESULTS: The degree of agreement between assessors varied for different parameters. Agreement for female worm counts in nodules was approximately 80%, while agreement regarding female worm viability was 98%. There were no major differences observed between results obtained by microscopy or digital images. Good agreement for important parameters was also observed for the nodules of the quality panel. CONCLUSION: Nodule analysis by experienced microscopists was reproducible with regard to important parameters such as identification of living female worms or detection of normal embryogenesis. Assessments varied more for other parameters, and we recommend continued use of two independent readers for detailed analyzes. Analysis of scanned images provided similar results to direct microscopy. This facilitates training and comparison of nodule findings by readers in different locations. Analysis of high quality digital images that can be viewed remotely should improve the quality and availability of nodule assessments that are primary endpoints for onchocerciasis clinical trials.