Performance evaluation of fungal DNA PCR amplification from formalin-fixed paraffin-embedded tissue for diagnosis; experience of a tertiary reference laboratory.

Background Diagnosis of invasive fungal infections from formalin-fixed paraffin-embedded (FFPE) tissues by PCR amplification is a developing technology. One of the difficulties of establishing a validated protocol for this testing is that the gold standard, culture, is much less sensitive than the test being validated. Objectives To validate FFPE PCR as a refence laboratory identification methodology in the absence of abundant gold standard specimens. Methods In this validation, PCR from FFPE tissue was compared to other diagnostic methods for genus/species identification. Four different groups of correlative data from FFPE tissues were used to validate this procedure. Thirteen specimens had culture or serology results and FFPE PCR results, 49 specimens had both immunohistochemistry (IHC) identification and FFPE PCR results, 118 specimens had histological evidence of fungal elements, 64 of which also had FFPE PCR results, and 36 fungal mock tissues or fungal negative tissues were used. Results The sensitivity determined from the tissues with positive fungal histopathology was 54%. The specificity of the cases for which there were both culture and FFPE PCR results was 100%. For the correlation with IHC, the specificity was 98%. For the mock tissues and fungal-negative tissues the calculated analytical sensitivity was 94%, specificity was 95% and accuracy was 94%. Conclusions By uniquely combining various data sources this study provides a comprehensive framework for how validation can be achieved in the absence of a gold standard and outlines the excellent performance of PCR from FFPE tissue, despite relatively the low sensitivity when compared to histopathology.