Development of calibration-free/minimal calibration wavelength selection for iterative optimization technology...

Area of Science:

• Chemical Engineering
• Analytical Chemistry
• Pharmaceutical Manufacturing

Background:

• Process Analytical Technology (PAT) using Near-Infrared (NIR) spectroscopy is crucial for monitoring continuous manufacturing (CM) processes.
• Traditional multivariate models like Partial Least Squares (PLS) require extensive calibration, posing challenges for CM implementation.
• Calibration-free or minimal calibration methods are increasingly sought to streamline PAT deployment in pharmaceutical CM.

Purpose of the Study:

• To evaluate the efficacy of Iterative Optimization Technology (IOT) algorithms as a calibration-free PAT approach for CM.
• To compare the performance of IOT against traditional PLS models for monitoring active pharmaceutical ingredient (API) content and potency.
• To develop and implement a novel wavelength selection method to enhance IOT performance under dynamic CM conditions.

Main Methods:

• Implemented IOT algorithms for quantitative (API content) and qualitative (potency trends) monitoring in a CM system.
• Utilized pure component spectra as the sole requirement for IOT model calibration.
• Developed and applied a variable importance in projection (VIP) scores-based wavelength selection method to address non-steady state challenges.

Main Results:

• IOT algorithms successfully monitored API content and potency trends in the CM system.
• The novel wavelength selection method significantly improved IOT prediction performance during non-steady state conditions.
• IOT demonstrated comparable or superior performance to traditional PLS models with reduced calibration burden.

Conclusions:

• IOT algorithms offer a viable and efficient calibration-free/minimal calibration alternative for PAT in pharmaceutical CM.
• The developed wavelength selection strategy enhances the robustness of IOT for real-time process monitoring.
• IOT algorithms present significant value and potential for widespread adoption in advanced pharmaceutical manufacturing.