Automated in situ hybridization and immunohistochemistry for cytomegalovirus detection in paraffin-embedded tissue sections.

Cytomegalovirus (CMV) infections cause serious morbidity and mortality in immunocompromised individuals. CMV detection methods include serology, viral culture, polymerase chain reaction, and histologic examination [hematoxylin and eosin, immunohistochemistry (IHC), in situ hybridization (ISH)]. Until recently, ISH was performed manually. We compared automated ISH and automated IHC (Bond-max system, Leica Microsystems) in 72 cases (multiple organ systems) previously evaluated for CMV by conventional methods [hematoxylin and eosin, IHC (Dako Autostainer Plus), polymerase chain reaction, and/or viral culture]. By automated ISH, 27 cases were CMV-positive (25 positive, 2 equivocal on original diagnosis), 43 were negative (10 positive, 29 negative, 4 equivocal on original diagnosis), and 2 were equivocal (positive on original diagnosis). By automated IHC, 31 cases were CMV-positive (28 positive, 3 equivocal on original diagnosis), 39 were negative (7 positive, 29 negative, 3 equivocal on original diagnosis), and 2 were equivocal (positive on original diagnosis). Using original CMV diagnosis as gold standard, automated ISH and automated IHC had sensitivities of 67.6% and 75.7%, respectively (P value=0.25), and both had specificities of 100%. Combined automated ISH and IHC had sensitivity of 75.7% and specificity of 100%. Positive predictive values for automated ISH, automated IHC, and combination of ISH and IHC were 92.6%, 90.3%, and 90.3%, respectively. Negative predictive values were 67.4%, 74.4%, and 76.3%, respectively. We recommend either automated ISH or IHC for CMV detection in formalin-fixed, paraffin-embedded tissues. Although with moderate sensitivities, these methods have high specificities and positive predictive values, permit rapid turnover, require only set-up time, and result in strong stain intensity with minimal background.