Evaluation of qPCR-based assays for leprosy diagnosis directly in clinical specimens

The increased reliability and efficiency of the quantitative polymerase chain reaction (qPCR) makes it a promising tool for performing large-scale screening for infectious disease among high-risk individuals. To date, no study has evaluated the specificity and sensitivity of different qPCR assays for leprosy diagnosis using a range of clinical samples that could bias molecular results such as difficult-to-diagnose cases. In this study, qPCR assays amplifying different M. leprae gene targets, sodA, 16S rRNA, RLEP and Ag 85B were compared for leprosy differential diagnosis. qPCR assays were performed on frozen skin biopsy samples from a total of 62 patients: 21 untreated multibacillary (MB), 26 untreated paucibacillary (PB) leprosy patients, as well as 10 patients suffering from other dermatological diseases and 5 healthy donors. To develop standardized protocols and to overcome the bias resulted from using chromosome count cutoffs arbitrarily defined for different assays, decision tree classifiers were used to estimate optimum cutoffs and to evaluate the assays. As a result, we found a decreasing sensitivity for Ag 85B (66.1%), 16S rRNA (62.9%), and sodA (59.7%) optimized assay classifiers, but with similar maximum specificity for leprosy diagnosis. Conversely, the RLEP assay showed to be the most sensitive (87.1%). Moreover, RLEP assay was positive for 3 samples of patients originally not diagnosed as having leprosy, but these patients developed leprosy 5-10 years after the collection of the biopsy. In addition, 4 other samples of patients clinically classified as non-leprosy presented detectable chromosome counts in their samples by the RLEP assay suggesting that those patients either had leprosy that was misdiagnosed or a subclinical state of leprosy. Overall, these results are encouraging and suggest that RLEP assay could be useful as a sensitive diagnostic test to detect M. leprae infection before major clinical manifestations.