Multicenter Studies to Evaluate the Performance of MALDI-TOF MS for Fungal Identification

传统的霉菌鉴定具有挑战性、繁琐和耗时，依赖于培养的非特定表型特征或昂贵和专门的DNA测序方法。MALDI-TOF MS使世界各地的临床微生物学实验室发生了革命性的变化，使细菌和酵母的快速鉴定成为可能。然而，由于开发有效的蛋白质提取方法的技术挑战和临床综合数据库的有限，这项技术还没有容易地应用于霉菌鉴定。 最近，我们开发和优化了一种简单的霉菌蛋白质提取程序，并构建了一个由来自90个属220个物种的410个单独分离物组成的高度严格的数据库，用于MALDI-TOF MS快速鉴定霉菌。NIH霉菌数据库是迄今为止开发的最全面的霉菌鉴定数据库，填补了临床微生物学领域的巨大诊断空白。NIH霉菌数据库现已与全球100家临床微生物学实验室共享，旨在通过提供更快的鉴定(在生长检测后30分钟内而不是数周内)来改进侵袭性真菌感染的诊断，从而指导最佳的抗真菌治疗。 在本财年，我们完成了一项十中心的国家对照研究，评估了两个真菌数据库(NIH霉菌数据库和布鲁克真菌数据库)针对80个临床霉菌分离株的临床性能和准确性。这项研究强调了重要的(以前未确定的)因素，如改善和实现MALDI-TOF MS在临床微生物实验室常规霉菌鉴定中的成功/统一实施所需的光谱采集设置。事实上，2019年8月，质谱仪制造商(Bruker Daltonics)发布了更新软件，对采集参数进行了具体更改，以提高模具识别能力。关于仪器间重复性差的数据得到了来自不同群体的最近两份出版物的支持。这一现象以前没有在大规模的细菌和酵母研究中得到证实，结果令人惊讶，值得进一步评估。 这项研究的结果已被接受发表在《前沿微生物学》上(印刷，2019年9月)。

Traditional identification of molds is challenging, cumbersome, and time-consuming, relying on non-specific phenotypic characteristics from culture or expensive and specialized DNA sequencing methods. MALDI-TOF MS has revolutionized clinical microbiology laboratories worldwide, enabling rapid identification of bacteria and yeast. This technology, however, has not been readily applied to mold identification due to technical challenges in developing an efficient protein extraction method and the limited availability of clinically comprehensive databases. Recently, we developed and optimized a simple protein extraction procedure for molds and constructed a highly stringent database comprising 410 individual isolates representing 90 genera and 220 species for the rapid identification of molds by MALDI-TOF MS. The NIH Mold Database is the most comprehensive database developed to date for mold identification and has filled a huge diagnostic gap in the field of clinical microbiology. The NIH Mold Database has now been shared with >100 clinical microbiology laboratories worldwide in an effort to improve the diagnosis of invasive fungal infections by providing faster identification (within 30 minutes of growth detection as opposed to weeks) that in turn can guide optimal antifungal therapy. During this fiscal year, we completed a ten-center national controlled study that evaluated the clinical performance and accuracy of two fungal databases (the NIH Mold Database and the Bruker Fungi Database) against 80 clinical mold isolates has come to a close. This study has highlighted important (previously undetermined) factors such as spectral acquisition settings required to improve and achieve successful/uniform implementation of MALDI-TOF MS for routine mold identification in clinical microbiology laboratories. In fact, in August 2019, the manufacturer of the mass spectrometer (Bruker Daltonics) released updated software with specific changes to acquisition parameters to improve mold identification. Data regarding poor inter-instrument reproducibility is supported by two recent publications from different groups. This phenomenon has not been demonstrated previously for large-scale bacterial and yeast studies, and results are surprising and warrant further evaluation. The results from this study have been accepted for publication in Frontiers Microbiology (in press, Sep 2019).