Bloodstream infection (BSI) is characterized by the presence of viable microorganisms in the bloodstream and may induce systemic immune responses. Early and appropriate antibiotic usage is crucial to effectively treating BSI. However, conventional culture-based microbiological diagnostics are timeconsuming and cannot provide timely bacterial identification for subsequent antimicrobial susceptibility test (AST) and clinical decision-making. To address this issue, modern microbiological diagnostics have been developed, such as surface-enhanced Raman scattering (SERS), which is a sensitive, label-free, and quick bacterial detection method measuring specific bacterial metabolites. In this study, we aim to integrate a new deep learning (DL) method, Vision Transformer (ViT), with bacterial SERS spectral analysis to build the SERS-DL model for rapid identification of Gram type, species, and resistant strains. To demonstrate the feasibility of our approach, we used 11,774 SERS spectra obtained directly from eight common bacterial species in clinical blood samples without artificial introduction as the training dataset for the SERS-DL model. Our results showed that ViT achieved excellent identification accuracy of 99.30% for Gram type and 97.56% for species. Moreover, we employed transfer learning by using the Gram-positive species identifier as a pre-trained model to perform the antibiotic-resistant strain task. The identification accuracy of methicillin-resistant and -susceptible Staphylococcus aureus (MRSA and MSSA) can reach 98.5% with only 200-dataset requirement. In summary, our SERS-DL model has great potential to provide a quick clinical reference to determine the bacterial Gram type, species, and even resistant strains, which can guide early antibiotic usage in BSI.

Effective management of sepsis requires timely administration of appropriate antibiotics; therefore, a reliable and rapid antimicrobial susceptibility test (AST) is crucial. To meet clinical needs, we developed a novel AST, referred to as SERS-AST, based on the surface-enhanced Raman Scattering (SERS) technology. In this study, we applied SERS-AST to eight most common pathogens causing bacteremia, including Staphylococcus aureus, Staphylococcus epidermidis, Enterococcus faecalis, E. faecium, Escherichia coli, Enterobacter cloacae, Klebsiella pneumoniae, and Acinetobacter baumannii. Seven different antibiotics were tested, including oxacillin, levofloxacin, vancomycin, ampicillin, cefotaxime, ceftazidime, levofloxacin, and imipenem. SERS-AST determines antibiotic susceptibility of bacteria directly from positive blood cultures based on variations in bacterial SERS signals derived from secreted purines and their derivatives. The whole process could be completed within 4 hours, and the agreement rates between SERS-AST and VITEK 2 results were 96% for Gram-positive bacteria and 97% for Gram-negative bacteria.

Background As effective management of sepsis requires timely administration of appropriate antibiotics, a reliable and rapid antimicrobial susceptibility testing (AST) is crucial. To meet clinical needs, we developed a novel AST, referred to as SERS-AST, based on the surface-enhanced Raman Scattering (SERS) technology. SERS-AST determines antibiotic susceptibility of bacteria based on variations in bacterial SERS signals derived from secreted purines and their derivatives. Methods SERS-AST was applied to blood culture samples of patients with bacteremia. Eight common causative organisms, including Staphylococcus aureus, Staphylococcus epidermidis, Enterococcus faecalis, Enterococcus faecium, Escherichia coli, Enterobacter cloacae, Klebsiella pneumoniae, and Acinetobacter baumannii and seven most prescribed antibiotics, including oxacillin, levofloxacin, vancomycin, ampicillin, cefotaxime, ceftazidime, levofloxacin, and imipenem were tested. Receiver operating characteristic analysis was performed to obtain the optimal cutoff SERS signal value to determine the susceptibility of tested bacteria. The results of SERS-AST were compared with those of VITEK 2. Results A total of 164 bacterial isolates from blood samples were examined, and the agreement rates between SERS-AST and VITEK 2 results were 96% for Gram-positive bacteria and 97% for Gram-negative bacteria. Conclusions By directly assaying positive blood cultures without additional subcultures, SERS-AST can be completed within 4 hours with high accuracy. It can be an alternative AST method to provide critical information to clinicians for timely administration of appropriate antibiotics to treat patients with blood stream infections.