Machine learning aided UV Absorbance Spectroscopy for microbial contamination in cell therapy products

Abstract We demonstrate the use of machine-learning aided UV absorbance spectroscopy to detect contamination during cell therapy product (CTP) manufacturing. The proposed method leverages a one-class support vector machine model to analyse the absorbance spectra of cell cultures and predict if a sample is sterile or contaminated. This label-free and non-invasive method provides a rapid output (30 minutes) with minimal sample preparation and sample volume (<1mL). Spiking of 7 bacterial species into mesenchymal stromal cells supernatant aliquots from 6 commercial donors showed that contamination events could be detected at low inoculums with mean true positive and negative rates of 92.7% and 77.7% respectively. Accounting for donor variation improved specificity to 92.0%. In cells spiked with 10 CFUs/mL of E.coli, contamination was detected at the 21-hour timepoint, demonstrating comparable sensitivity to compendial USP <71> test (~24 hours). We hypothesize that spectral differences between nicotinic acid and nicotinamide in the UV region are the underlying mechanisms for contamination detection. This technique can be deployed as a preliminary test during different CTP manufacturing stages where cultures can be monitored real-time to detect early onset of microbial contamination, assuring safety of CTP prior to infusion into patients.