Quality of MALDI-TOF Mass Spectra in Routine Diagnostics: Results from an International External Quality Assessment including 36 Laboratories from 12 countries using 47 challenging bacterial strains.

Abstract

OBJECTIVE

MALDI-TOF MS is a widely used method for bacterial species identification. Incomplete databases and mass spectral quality (MSQ) still represent major challenges. Important proxies for MSQ are: number of detected marker masses, reproducibility, and measurement precision. We aimed to assess MSQs across diagnostic laboratories and the potential of simple workflow adaptations to improve it.

METHODS

For baseline MSQ assessment, 47 diverse bacterial strains which are challenging to identify by MALDI-TOF MS, were routinely measured in 36 laboratories from 12 countries, and well defined MSQ features were used. After an intervention consisting of detailed reported feedback and instructions on how to acquire MALDI-TOF mass spectra, measurements were repeated and MSQs were compared.

RESULTS

At baseline, we observed heterogeneous MSQ between the devices, considering the median number of marker masses detected (range = [5, 25]), reproducibility between technical replicates (range = [55%, 86%]), and measurement error (range = [147 parts per million (ppm), 588ppm]). As a general trend, the spectral quality was improved after the intervention for devices which yielded low MSQs in the baseline assessment: for 4/5 devices with a high measurement error, the measurement precision was improved (p-values<0.001, paired Wilcoxon test); for 6/10 devices, which detected a low number of marker masses, the number of detected marker masses increased (p-values<0.001, paired Wilcoxon test).

CONCLUSION

We have identified simple workflow adaptations, which, to some extent, improve MSQ of poorly performing devices and should be considered by laboratories yielding a low MSQ. Improving MALDI-TOF MSQ in routine diagnostics is essential for increasing the resolution of bacterial identification by MALDI-TOF MS, which is dependent on the reproducible detection of marker masses. The heterogeneity identified in this EQA requires further study.