Building quality into drug development for safer medicines

 In this interview, News Medical-Life Sciences speaks with Anil Kane, pharmaceutical development expert at Thermo Fisher Scientific’s Pharma Services Group, about how quality is built into Drug Development from early design through manufacturing. He discusses phase-appropriate analytical strategies, predictive modeling, Quality by Design, and how artificial intelligence is transforming decision-making to bring safer, more effective medicines to patients faster. 

Can you please introduce yourself and your role at Thermo Fisher Scientific?

I began my career as a pharmacist, focusing on industrial pharmacy and research and development. Over the past 35 years, I have worked extensively in Drug Development, concentrating on the science and strategy required to move molecules from concept through clinical development and ultimately to commercialization.

At Thermo Fisher Scientific, within the Pharma Services Group, we partner with pharmaceutical and biotech companies to support end-to-end drug development. This includes early-stage development, clinical manufacturing, late-stage development, and commercialization support. Our role is to ensure that quality is built into every stage of the development process so that safe and effective medicines reach patients efficiently.

What does quality really mean in drug development?

Quality in drug development is not optional; it is fundamental. It is embedded at every step of the journey, which can take 5-12 years from concept to patient.

Quality begins with the strategy. The quality of the input data directly affects the quality of decision-making and ultimately the quality of the output. Drug development is driven by data. At each stage, scientists make go/no-go decisions based on robust datasets. If the data is incomplete or unreliable, the downstream consequences can be significant.

Therefore, quality is about generating reliable, scientifically sound data at every stage and interpreting it correctly to guide the right decisions.

How does quality differ between early-stage and late-stage development?

There is a clear distinction, particularly in terms of phase-appropriate quality.

In early development, analytical methods do not need to be fully validated to the same regulatory standard required for commercialization. However, they must still be robust and reliable enough to generate accurate data to support decision-making.

In later stages, particularly as commercialization approaches, analytical methods must meet regulatory expectations and validation requirements. At that point, consistency, reproducibility, and compliance become even more critical.

The principle remains the same: phase-appropriate quality ensures that decisions are supported by the right level of scientific rigor at the right time.

How important is early molecular characterization in drug development?

Understanding the molecule at an early stage is absolutely critical. Whether it is a small molecule from medicinal chemistry or a biologic developed through cell line engineering, complete characterization forms the foundation for development.

This includes selecting the correct polymorph, salt form, and assessing stability, absorption, distribution, and bioavailability. Analytical and bioanalytical techniques provide enabling datasets that inform stage-gate decisions, such as whether to proceed to toxicology studies or first-in-human trials.

Getting this right early is essential because there is little room for error or rework later. The cost and time pressures in drug development demand precision from the beginning.

How are predictive modeling and AI transforming drug development?

Predictive modeling tools have evolved significantly. What were once referred to as neural networks are now widely recognized as artificial intelligence and machine learning systems.

These tools rely on high-quality datasets to build learning algorithms capable of predicting solubility, absorption, bioavailability, excipient compatibility, and stability profiles. They help narrow down candidates from large pools of molecules to a smaller, more promising set.

By reducing trial-and-error, AI and machine learning improve efficiency, shorten timelines, and lower development costs. They enable faster and more informed decision-making, particularly in early-stage drug development.

How do you confirm predictions made by AI models?

Predictive modeling narrows the field, but experimental validation remains essential.

Advanced analytical techniques such as X-ray crystallography, powder X-ray diffraction, Raman spectroscopy, near-infrared spectroscopy, particle size analysis, and differential scanning calorimetry are used to confirm polymorph selection, distinguish crystalline from amorphous forms, and characterize structural properties.

These tools provide direct evidence that the molecule behaves as predicted. This integration of modeling and experimentation accelerates development while maintaining scientific rigor.

How does Quality by Design improve long-term development outcomes?

Quality by Design, or QbD, is an approach that integrates quality considerations throughout the entire development lifecycle.

This includes understanding critical quality attributes, defining acceptable parameter ranges, validating analytical methods, and designing robust, scalable manufacturing processes. Stability studies determine shelf life and ensure consistent performance throughout the product lifecycle.

By embedding QbD early, we minimize late-stage surprises, improve scalability, and ensure that the product consistently meets safety and efficacy standards.

Can early quality decisions prevent costly late-stage failures?

Absolutely. If early data are inaccurate or incomplete, problems may only surface later in development. For example, a product might prove unstable, non-scalable, or inadequately absorbed in the body. Discovering such issues late can lead to significant delays and financial losses.

From Molecule to Medicine: How Quality Is Built into Drug Development | Analyze That Ep. 7Play