Diagnostic accuracy of chip hybridization from formalin-fixed paraffin embedded bioptic specimens in a cohort with predominantly extrapulmonary mycobacterial disease.

INTRODUCTION

Extrapulmonary tuberculosis (TB) and non-tuberculous mycobacterial (NTM) disease remain a diagnostic challenge and conventional methods such as mycobacterial culture or molecular methods (e.g. GeneXpert) performed on formalin-fixed paraffin embedded (FFPE) specimens are often unsuccessful. In this study, we assessed the diagnostic accuracy of chip hybridization from FFPE specimens in a cohort with predominantly extrapulmonary mycobacterial disease.

METHODS

FFPE specimens from patients with presumed mycobacterial disease underwent chip hybridization with two different assays (MYCO Direct 1.7 and MYCO Chip Vision Array 2.0) in addition to conventional diagnostic methods for TB and NTM disease including mycobacterial culture, molecular methods and conventional histology. The different techniques were assessed with regards to sensitivity, specificity and species identification.

RESULTS

Overall, 184 samples were assessed by one of the two chip hybridization techniques (MYCO Direct 1.7 n=65 and MYCO Chip Vision Array n=119). Samples were mainly extrapulmonary (160/184, 87.0%). Of all samples, 32 (17.4%) turned positive in chip hybridization and 74/184 patients (40.2%) were diagnosed with clinically relevant mycobacterial disease (59 with TB and 15 with NTM disease). For the detection of Mycobacterium tuberculosis complex (MTBC), sensitivity was 38.9% when compared to mycobacterial culture from the same specimen. For the detection of NTM, sensitivity was 50% against NTM culture. Species identification was overall consistent with conventional methods but did not allow for species differentiation of members of the MTBC or Mycobacterium avium complex, respectively.

CONCLUSION

As extrapulmonary TB and NTM disease remain a diagnostic challenge, the investigated chip hybridization techniques might present a complementation of the diagnostic portfolio. However, with a low sensitivity to detect MTB from extrapulmonary FFPE specimens, they should not supplant conventional diagnostic methods but could be used to rule in TB when native specimens are not available.