From clinics to the environment: A systematic review of machine learning and MALDI-TOF MS in the identification of antimicrobial-resistant bacteria

Abstract

As a result of the increasing prevalence of Antibiotic-resistant bacteria (ARB) and antibiotic-resistant genes (ARGs) in both community and hospital settings, their identification by conventional approaches has been posing a significant issue for decades. A new approach, integrating matrix-assisted laser desorption/ionization time of-flight mass spectrometry (MALDI-TOF-MS) with machine learning (ML), has emerged as a valuable method for their identification. This review systematically evaluates the effectiveness of integrating MALDI-TOF-MS with ML for the identification of ARB. A comprehensive literature search was conducted using PubMed, Google Scholar, the Cochrane Library, Scopus, and Web of Science to identify original research articles focused on the application of MALDI-TOF-MS and ML in detecting ARB. Studies unrelated to bacteria or antibiotic resistance, as well as short communications, scientific reports, and case studies, were excluded. In accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, 401 potentially relevant articles were initially identified. Following the application of inclusion criteria and relevance assessment of titles and abstracts, 34 studies were selected for final analysis. The findings demonstrate that integrating MALDI-TOF-MS with ML models markedly improves the speed, accuracy, and reliability of ARB detection while offering valuable insights into the molecular mechanisms of resistance. Current evidence suggests that the integration of MALDI TOF MS with ML is an important approach for the identification of bacteria in clinics and environments, mainly ARB. Future research is needed to apply this approach to address the growing challenge of ARB both in clinical and environmental settings.