Fungal Integrated Histomolecular Diagnosis Using Targeted Next-Generation Sequencing on Formalin-Fixed Paraffin- Embedded Tissues

Abstract

Histopathology is the gold standard for fungal infection (FI) diagnosis, but it does not provide a genus and/or species identification. The objective of the present study was to develop targeted next-generation sequencing (NGS) on formalinfixed tissue samples (FTs) to achieve a fungal integrated histomolecular diagnosis. Nucleic acid extraction was optimized on a first group of 30 FTs with Aspergillus fumigatus or Mucorales infection by macrodissecting the microscopically identified fungalrich area and comparing Qiagen and Promega extraction methods through DNA amplification by A. fumigatus and Mucorales primers. Targeted NGS was developed on a second group of 74 FTs using three primer pairs (ITS-3/ITS-4, MITS-2A/MITS-2B, and 28S-12-F/28S-13-R) and two databases (UNITE and RefSeq). A prior fungal identification of this group was established on fresh tissues. Targeted NGS and Sanger sequencing results on FTs were compared. To be valid, the molecular identifications had to be compatible with the histopathological analysis. In the first group, the Qiagen method yielded a better extraction efficiency than the Promega method (100% and 86.7% of positive PCRs, respectively). In the second group, targeted NGS allowed fungal identification in 82.4% (61/74) of FTs using all primer pairs, in 73% (54/74) using ITS-3/ITS-4, in 68.9% (51/74) using MITS-2A/MITS-2B, and in 23% (17/74) using 28S-12-F/28S-13-R. The sensitivity varied according to the database used (81% [60/74] using UNITE compared to 50% [37/74] using RefSeq [P = 0.000002]). The sensitivity of targeted NGS (82.4%) was higher than that of Sanger sequencing (45.9%; P < 0.00001). To conclude, fungal integrated histomolecular diagnosis using targeted NGS is suitable on FTs and improves fungal detection and identification.