A new study published by Research Square proposes a machine learning (ML) approach combined with UV absorbance spectroscopy to detect microbial contamination in cell therapy products (CTPs) more rapidly and accurately. Researchers developed a one-class support vector machine (SVM) model trained on sterile mesenchymal stromal cell (MSC) cultures, allowing them to detect bacterial contamination based on UV absorbance changes. This model achieved a high true positive accuracy of 92.7%, rapidly identifying bacterial contaminants such as Escherichia coli, Staphylococcus aureus, and Candida albicans.
Current methods for sterility testing in CTPs, like the USP <71> assay, often require up to 14 days to deliver results, during which they rely on turbidity and growth indicators, which can be slow and affected by non-microbial interferences. The ML-enhanced UV spectroscopy model, on the other hand, detects spectral changes from bacterial metabolic by-products like nicotinic acid. These metabolites shift the UV spectra, allowing the ML model to identify contamination in as little as 21 hours, thus offering a faster, label-free, and non-invasive testing solution.
The study also highlights the method’s versatility and sensitivity. It demonstrated a limit of detection of 10 colony-forming units per milliliter (CFU/mL) and worked across seven different bacterial strains and multiple MSC donors. Results indicated that while media components could vary between donors, the model’s performance remained stable, suggesting it could be adapted to handle variability across donor sources.
Overall, this research provides an innovative, faster approach to microbial detection in CTPs, which could significantly benefit manufacturing timelines and sterility assurance in the cell therapy industry. The ML-UV method not only addresses current sterility testing challenges but also holds promise for further applications in quality control for various biomedical products, thereby enhancing safety and efficiency in advanced cell-based therapies.
This article was created with the assistance of artificial intelligence and has been carefully reviewed and edited to ensure scientific accuracy and clarity. For more information on our AI usage policies, please visit our AI policy page.
Reference
Chelvam, S. P., Ng, A., Huang, J., Lee, E., Baranski, M., Yong, D., Williams, R. B. H., Springs, S. L., & Ram, R. J. (2024). Machine learning aided UV absorbance spectroscopy for microbial contamination in cell therapy products. Research Square. https://doi.org/10.21203/rs.3.rs-4880911/v1
