Quantification of Tuberculosis targeting isoniazid in oral dosage tablets using low-cost and handheld spectroscopy

Description:
Isoniazid is an essential medicine distributed through global health supply chains as a treatment for patients with Tuberculosis. Handheld spectrometers offer potential for short-wave infrared (SWIR) data collection at a lower cost than benchtop spectrometers. Previous efforts have shown that these handheld spectrometers can effectively differentiate between brands of oral dosage isoniazid. Here, we assess the potential of a handheld spectrometer for quantifying the amount of active ingredient (isoniazid) in both in-house pressed solid dosage tablets as well as isoniazid tablets obtained from five manufacturers. Tablets pressed in house (n = 135) were formulated to contain between 0 – 100% isoniazid, while the five tablet brands (n = 84) contained 50 – 86% isoniazid. Tablets were scanned with the handheld spectrometer (900 – 1700 nm), and a benchtop spectrometer (350 – 2500 nm). For each spectrometer, a calibration sample set containing 150 tablets was constructed, while 65 tablets were reserved as an independent validation dataset. A partial least squares regression model (PLSR) was applied to the calibration datasets. The independent validation datasets resulted in an R2 of 0.998 and 0.996 and a root mean squared error of prediction (RMSEP) of 2.08 and 2.56 percentage of the isoniazid present, for the benchtop and handheld spectrometers, respectively. While the benchtop spectrometer resulted in a slightly lower RMSEP likely due to the expanded spectral range capturing more isoniazid spectral features than the handheld, the results for the handheld are promising. These results suggest the possibility for rapid and low-cost spectral analysis in resource limited settings.

Speaker: Matthew Eady - FHI 360
Matthew Eady is a Scientist with FHI 360’s Product Quality Compliance Department where he researches portable spectroscopy and other low-cost screening methods for quality compliance of essential medicines provided through global supply chains. Prior to joining FHI 360, he spent 8 years working with hyperspectral imaging at the U.S. Department of Agriculture. He has over 10 years of experience working with NIR and multivariate data analysis methods. His Ph.D. was obtained through the Food Science and Technology Department at the University of Georgia. He is also an Adjunct Professor in Agricultural Science at Murray State University.