By scrutiny in-field studies, several factors found to be affected the consequence of PCR assay in FFPE tissue samples. These factors are including DNA extraction method, inclusion of proper housekeeping gene, PCR method (i.e., panfungal, specific, nested, semi-nested, multiplex, and Real-Time), target gene(s) primers, amplicon length, thickness of FFPE cut, storage time of specimen, and contamination during sample preparation [24–26].
Here, we reached a sensitivity of 87.5% and a specificity of 89.1% for semi-nested PCR assay targeting ITS1-5.8s-ITS2 (ITS1-2) region of 110 FFPE-FESS samples. Bialek et al. indicated a 60.6% sensitivity and 100% specificity for nested PCR targeting 100-kDa-like-protein gene in 33 FFPE specimens, which were histopathologically proven for histoplasmosis [27]. Similarly, Willinger et al. indicated the 87% sensitivity for PCR of the 28S region [28]. This may result from short amplicon length (260 bp) and pertinent primer pairs (U1/U2) to amplification the 28S region. In another similar study, Rickerts et al. reported acceptable sensitively for molecular diagnosis of invasive aspergillosis and mucormycosis [29]. In both studies [22, 29], the DNA extraction method, gene targets, primer pairs, and amplicon length were similar. Lau et al. reported 82% positive results for panfungal PCR assay [30]. They also showed a sensitivity of 97% and 68% for fresh and FFPE samples, respectively which was lower than the results of current study. Cabaret et al. indicated 62.5% and 93.75 positive results of FFPE samples for conventional PCR and qPCR, respectively [31]. These differences may result from the differences in amplicon length (> 300 bp for conventional PCR vs. 150 bp for qPCR) and the type of PCR method. In a similar study, Hammond et al. reported 81.5% positive results by semi-nested PCR of 18S rDNA using ZM1/ZM2 and ZM1/ZM3 primer pairs targeting an amplicon < 200 bp [32]. Salehi et al. reported a 64% sensitivity for identification of fungi from FFPE by qPCR assay [33]. Drogari-Apiranthitou et al. [34] reached 45% positive PCR for Mucorales and 40% positive for Aspergillus spp. They reported 79.3% and 100% sensitivity and specificity for semi-nested PCR, respectively. The higher sensitivity of their method in comparison to us might be due to the more thickness of their tissue cuts (10 µm vs. 5 µm). In another study, Ganesan et al. reported the site of infection might affect the sensitivity of PCR test and reported that sensitivity rate increased from 63–83% for angioinvasion sites [35]. While Jung et al. indicated a 41.3% positive results for panfungal PCR, due to long amplicon length and may be inappropriate DNA extraction kit [36]. In our study we found a concordance rate of 76% for semi-nested PCR of the ITS region and histopathology assays which is similar to previous reports [30]. While others reported higher [33] or lower [37, 38] concordance rate between these two methods.
Due to the low amount of tissue in FFPE samples, the extracted DNA would be low. Hence, selection of proper extraction method is a key step in achieving a reliable amplification. Here, we used QIAamp DNA FFPE Tissue Kit (QIAGEN), as successfully used before [17, 28, 34, 37, 39–43]. In a comparative study between different commercial extraction kits, Muñoz-Cadavid et al. reported TaKaRa and QIAamp extraction kits yielded the best results for extraction of high quality DNA from the FFPE sample [44].
Similar to some previous studies [22, 32–34, 41, 44], we used human β-globin gene as internal control and the primers successfully amplified the expected region in all samples. While others used different genes such as IRBP [38], GAPDH [27], and ZP3 [45] as controls.
In this study we used a panfungal semi-nested PCR for amplification of the ITS1-5.8s-ITS2 followed by ITS1 regions to detect fungal element in FFPE tissue samples which is concordant with a previous report [24]. Others amplified different targets such as 28S rDNA [43], ITS2 [35], 18S rDNA [34, 37] and mitochondrial tRNA [34] for detection of fungal DNA in FFPE samples. Although, Cabaret et al. suggest targeting mitochondrial DNA is superior to ITS [31], they obtained lower positive rate by conventional PCR (10/16, 62.5%) in comparison to ours (56/64, 87.5%). In an another study, Jillwin et al. [37] targeted five different gene regions including universal ITS (ITS1-5.8s-ITS2), ITS1, ITS2, 18S rDNA, and D1/D2 of 28S rDNA and reported that ITS1 amplification leads to 61.9% positive results by PCR method.
The false-negative/positive consequences may result from the following reasons: First, false-positive histopathological results artifacts during staining. Second, the presence of conserved genes in multiple copies (rDNA) is a disadvantage in clinical specimens collected from nonsterile body sites, because nonpathogenic commensal fungi, environmental spores, or colonizing fungi can also cause considerable nonspecific amplification in samples consisting mainly of human cells with only a few fungal cells. Third, proteins-DNA cross-linking formation and fragmentation of DNA during the fixation process leads to a lack of intact DNA needed for amplification. More specifically, it is difficult to amplify the target gene when DNA is highly fragmented or cross-linked and the amplicon size is large. Fourth, amplification of a housekeeping gene (human β-globin) in tissue does not necessarily mean the presence of a sufficient amount of amplifiable fungal DNA.
It is recommended that a shorter amplicon length gives higher sensitivity and specificity rates for PCR results [26]. Also, it recommended that semi-nested PCR assays performed in a single tube, which showed less prone to contamination when compared with assays that carried out in two stages and in separate tubes. As Aspergillus species and Mucorales are most frequent causative agents of FRS, it can be a good idea to use species-specific primers and multiplex PCR to differentiate them as soon as possible in a single reaction.