LAUSR

Abstract This aim of this study was to assess the ability of Raman spectroscopy to quantify antineoplastic drugs directly in the finished product in plastic bags using a handled Raman spectrometer. Gemcitabine diluted in 0.9% sodium chloride was analyzed at various concentrations ranging from 1 to 20 mg/mL directly through plastic bags using a handled 785 nm Raman spectrometer. In accordance with EMA guidelines, quantitative models were developed to predict gemcitabine concentration in bag using partial least squares (PLS) regression. In order to evaluate the transposability of the developed Raman method and the routine method (flow injection analysis with UV detection), independent samples were analyzed using both techniques. The impact of the plastic bag was also evaluated by analysis samples through two different bags. The best model was obtained after standard normal variates preprocessing (SNV) for 15 latent variables. This model presented an excellent correlation between predicted and theoretical concentration values (R² of 0.9938 from the calibration set), a low limit of quantification (LLOQ) of 3.68 mg/mL and acceptable repeatability and intermediate precision lower than the expected acceptance limit of 5% over the entire concentration range tested (except for the average concentration of 5.73 mg/mL). For the 48 preparations higher than the LLOQ, the Bland-Altman approach showed the interchangeability of the wo methods with a difference bias of 2%. Moreover, no significant difference of predicted concentrations between the two containers tested (p = 0.189) was observed. Despite some limitations for low concentrations, this study clearly shows promising results for real-time monitoring of gemcitabine infusion preparations without removing samples. The non-invasive nature of this method should ensure the correct dose before administration to patients and with heightened safety for operators.