In resource-limited settings, where pneumocystosis in immunocompromised patients is infrequently observed, cost-efficient, reliable, and sensitive approaches for the diagnostic identification of Pneumocystis jirovecii in human tissue samples are desirable. Here, an in-house fluorescence in situ hybridization assay was comparatively evaluated against Grocott's staining as a reference standard with 30 paraffin-embedded tissue samples as well as against in-house real-time PCR with 30 respiratory secretions from immunocompromised patients with clinical suspicion of pneumocystosis. All pneumocystosis patients included in the study suffered from HIV/AIDS. Compared with Grocott's staining as the reference standard, sensitivity of the FISH assay was 100% (13/13), specificity was 41% (7/17), and the overall concordance was 66.7% with tissue samples. With respiratory specimens, sensitivity was 83.3% (10/12), specificity was 100% (18/18), and the overall concordance was 93.3% as compared with real-time PCR. It remained unresolved to which proportions sensitivity limitations of Grocott's staining or autofluorescence phenomena affecting the FISH assay accounted for the recorded reduced specificity with the tissue samples. The assessment confirmed Pneumocystis FISH in lung tissue as a highly sensitive screening approach; however, dissatisfying specificity in paraffin-embedded biopsies calls for confirmatory testing with other techniques in case of positive FISH screening results. In respiratory secretions, acceptable sensitivity and excellent specificity were demonstrated for the diagnostic application of the P. jirovecii-specific FISH assay.
Keywords: FISH; Grocott’s staining; Pneumocystis jirovecii; bronchoalveolar lavage; diagnosis; paraffin-embedded tissue; real-time PCR; sputum; test evaluation.