PCR assays have not reached the same level of acceptance for the detection of human fungal pathogens as for other micro-organisms, mainly because the low number of micro-organisms challenges the detection limits of PCR. Therefore, whereas meta-analyses focusing on clinical validation suggest interest in adding PCR results to the diagnostic workup for invasive fungal disease (IFD) along with clinical evaluation, CT scans, classical mycology and antigen detection, no consensual PCR method has emerged. Compared with the end-point format of the 1990s, real-time quantitative PCR is a major breakthrough. This format prevents contamination with previously amplified products, provides the yield of amplification, allows for developing consensus procedures and should therefore be the only format used. An internal control is now mandatory to avoid false-negative results. Primer design strongly impacts on the objectives: pan-fungal primers can provide false-positive results due to environmental fungal DNA contamination; conversely, species-specific primers miss infections caused by untargeted fungi. Unresolved issues include the best specimens to be used; serum is currently preferred to blood because of the ease of the DNA extraction step. Work is in progress to establish standards at least for Aspergillus PCR, and the implementation of quality controls should help centres to improve assays. Eventually, the classical analysis of biomarker performance does not consider the evolving risk factors and changing treatments during IFD, which can lead to variable conclusions. New statistical methods such as event history analysis should be considered.