Making Spectroscopic PAT Practical with Lean Chemometrics

The Promise vs. Reality of Spectroscopic PAT

Spectroscopic Process Analytical Technology (PAT) has long been recognized as a powerful tool to enable faster development, better process control, and real-time insight across the product lifecycle. Its value is widely understood by process development, PAT, and scale-up teams.

Yet in practice, spectroscopic PAT is still not implemented as early or as broadly as it could be.

Spectral data is complex, requiring a chemometric model to extract actionable information. The calibration burden of chemometric models introduces significant time, material, and operational requirements that compete directly with development timelines and resources. Additionally, many process development teams may face a skills gap when it comes to chemometric modeling, making it even more challenging to efficiently build and maintain these models within existing project constraints.

As a result, spectroscopic PAT is often delayed, not because its value is uncertain, but because the effort required to build and maintain calibration models is difficult to justify early on.

This means that critical process insight comes too late, when decisions are already constrained, and changes become more costly and complex. It also limits the ability to generate early process understanding and apply it effectively during scale-up and transfer.

In this article, we discuss the main barriers to adoption and, more importantly, outline a more practical path forward with Lean chemometrics.

The Core Barrier: Calibration Burden

Calibration burden = the time, material, and financial resources required to build, validate, and maintain a chemometric model.

Time investment: Calibration development requires multiple steps, including calibration design selection, mixture sample preparation, physical sampling and transport, equipment cleaning, and reference method analysis. Each step adds time and operational complexity, requiring skilled resources.
Material usage: Calibration activities consume raw materials for mixture samples, consumables for cleaning, and materials used in reference analyses. These requirements increase development effort and waste generation.
Financial cost: Calibration work also drives operational cost, including FTE time for calibration development, cleaning equipment, facilities and utilities, sample containers and logistics, and waste disposal.

Some spectroscopic models require extensive calibration, while others can work with a much lower calibration burden. Comparing full calibration vs. calibration-free approaches shows clear differences in the number of mixture samples and materials required^1,2.

Reduced calibration requirements can help speed up adoption of spectroscopic PAT

Reference: Henson, S., Rish, A. J., Alom, M. A., Liu, Y., Drennen, J. K., & Anderson, C. A. (2024). Development of iterative optimization technology: Selecting pure component spectra using a small-scale feed frame simulator. International Journal of Pharmaceutics, 657, 124079.

Introducing Lean Chemometrics

Inspired by lean thinking, Lean Chemometrics has a simple goal: cut the excess, keep the value. It is defined as low calibration burden methods that “employ time-saving, material-sparing, and cost-cutting strategies for the purposes of reducing the overall calibration burden of developing, validating, and maintaining a fit-for-purpose analytical procedure”¹.

This means removing unnecessary data collection, using simpler, more interpretable models, prioritizing speed and robustness over complexity, and emphasizing fit-for-purpose chemometric modeling.

This approach makes methods affordable, understandable, and easy to implement for spectroscopic PAT, regardless of a practitioner’s individual expertise. Because in the real world, the “best” model isn’t the most sophisticated one; it’s the one that works when it counts.

“While lean chemometrics cannot always replace full calibration, it can complement traditional approaches by reducing effort where possible. It is especially valuable early in development, when API cost is high, and when API availability is limited”, explains Sam Henson, SME Lean Chemometrics.

Iterative Optimization Technology (IOT): A Promising Lean Chemometric Method

Iterative Optimization Technology (IOT) is a lean chemometric method for early multivariate spectral data analysis that uses only pure spectra of individual components, eliminating the need for conventional mixture-based calibration campaigns.

IOT is easier to deploy and maintain than traditional chemometric approaches and is ready to deliver real-world impact, particularly in an environment where pharma is under increasing pressure to reduce costs and accelerate sustainability goals.

Key benefits of IOT include:

Reduced calibration burden, with less material, less time, and lower calibration costs
Increased explainability, as predictions are directly based on actual mixture components and no spectral decomposition is required
Simplified model maintenance, with no need to rebuild full calibration sets and only pure component spectra updated as needed
Enables sustainable PAT, with no reference or mixture samples, resulting in fewer solvents, containers, and waste

Conclusion: A Practical Path Forward

Reducing calibration burden can improve the acceptance and adoption of spectroscopic PAT, helping to close the gap between its recognized value and its actual use across the product lifecycle. Lean chemometrics provides a practical approach to reducing the time, material, and cost constraints associated with calibration while maintaining fit-for-purpose analytical performance.

By lowering the effort required to implement and maintain chemometric models, these approaches make it more feasible to apply spectroscopic PAT earlier and more broadly in the development process.

This enables faster learning with less API consumed and less lab time spent on model calibration. As a result, process understanding can be obtained earlier in the product lifecycle, allowing better-informed decisions before constraints increase, and supporting more efficient scale-up and process development at commercial scale.

If you are interested in lean chemometrics or IOT and want to discuss with an expert, or if you would like to better understand how synTQ supports these approaches, please reach out to us.

Contact Us

References:

1. Rish, A.J., Henson, S., and Rehrauer O., “Lean Chemometrics in Spectroscopic Process Analytical Technology” Journal of Chemometrics 40, no. 2 (2026): e70105, https://doi.org/10.1002/cem.70105.