Upfront

Imaging technologies such as computerized tomography (CT), which won Peter Mansfield and Paul Lauterbur the Nobel Prize for medicine last month, have proven instrumental in diagnosing many illnesses. As pharmaceutical companies refine process analytical technologies (PAT), similar imaging technologies may soon be used routinely to optimize production of the medicines used to treat those ills.

A number of R&D projects are now underway to evaluate medical imaging techniques for pharmaceutical manufacturing "patients." In the U.S., Pfizer Corp. (New York) and the U.S. Food and Drug Administration (FDA) have launched a cooperative research and development agreement (CRADA) that could take near-infrared imaging to the shop floor for formulations development, process scale up, monitoring and control, quality assurance and verification of packaged product.

Where NIR spectroscopy, a well established lab technique, measures a single spectrum of a sample presented to it, NIR imaging takes multiple spectra from discrete points within the sample, providing a spectroscopic map of the sample. "We intend to take imaging based on lab-based spectroscopic techniques and apply it on line, throughout products' lifecycles" says Norman Winskill, vice-president of global manufacturing services at Pfizer. Pfizer and FDA plan to have an experimental NIR unit on line at a pilot plant in New Jersey within the next six months, with commercial instrumentation available within two years.

The CRADA is part of FDA's push to commercialize PAT and update its current good manufacturing practices (cGMP). NIR, like CT scanning, has become established in medicine for studying the brain and other parts of the human body, and analyzing ailments such as tumors.

Pfizer first began to use NIR spectroscopy in its pharmaceutical manufacturing operations back in the 1980s, to monitor and troubleshoot specific reactions. By 1990, the company had formed a dedicated NIR group, and roughly five years ago, began to evaluate mid-IR, acoustic and Raman spectroscopy as well.

One important goal of the CRADA is to automate quality control so that testing occurs during processing, rather than as a separate adjunct to each process step. NIR imaging offers the benefit of being able to analyze all components within a given solid, including active ingredient, lubricant and disintegrant simultaneously. When solids are dissolved in solution to prepare samples for HPLC or other analysis, solid-solid interactions are destroyed or reduced. NIR allows these interactions, instead, to be analyzed, providing much more data on bioavailability, Winskill says. Sample-preparation and laboratory test time also are eliminated.

Last year, at its Brooklyn, N.Y. plant, Pfizer installed a battery-powered, radio-communicating NIR spectrometer to take continuous online readings from a blender. The unit, controlled by a PC in another room, featured a sample-reading head that automatically collected between 200 and 300 milligrams of material from the blender. Readings allowed Pfizer technicians to track not only one, but all ingredients in the blend in real time. The installation achieved in one minute what would otherwise have required taking eight samples and running HPLC analyses on each, to determine product uniformity, explained Steve Hammond, head of the NIR group within Pfizer's global manufacturing services, at FDA's PAT Initiative meeting last year.

The company is also working with Bruker Optics (Billerica, Mass.) and Schleuniger Pharmatron of Switzerland to develop a fully automated, integrated system that would scan tablets directly at the tablet press, testing 300 tablets at a time, rather than the 10 typically tested using conventional methods. In evaluations at a company plant in Australia, the technology was used to trace problems with a specific batch, and ultimately found that a transfer chute was causing segregation in the blend, Hammond explained at the meeting. Traditional testing would have taken much longer to trace the cause of the problem.

Pfizer has worked extensively with other vendors such as Spectral Dimensions (Olney, Md.), and, previously, with academic researchers at London's University College and Imperial College to develop its in-house NIR imaging technology.

Meanwhile, in the U.K., pharmaceutical major GlaxoSmithKline is looking closely at the CT imaging offshoot, electrical resistance tomography (ERT), as a process-development tool. The company discussed its work at the third World Congress on Industrial Process Tomography held in Banff, Alberta, Canada in late September 2003. Leading chemical processing companies, including Johnson Matthey Catalysts and Syngenta, also presented results of applying ERT at the conference.

Based on lab-scale studies, Glaxo concludes that ERT technology can provide the mixing time data needed to characterize reactors and other process vessels more effectively than existing method. The technique can also be used to validate computational fluid dynamic (CFD) models. The company has used ERT online to monitor ethyl acetate hydrolysis, a single-phase process, in a 3.5-l reactor that mimicked pilot-plant equipment and operating conditions as closely as possible. Glaxo is now evaluating ERT on multi-phase systems. Supplying the imaging equipment and know-how for this work is Industrial Tomography Systems (ITS), a company spun off by the University of Manchester's Institute of Science and Technology, U.K., specifically to develop the commercial potential of ERT.

Glaxo, working with Imperial College London, arranged platinum electrodes symmetrically in planes around the circumference of a model glass reactor. Glass was chosen for its transparency and because it is commonly used to line pilot- and full-scale pharmaceutical manufacturing vessels. Platinum was used because of its high chemical resistance and because its thermal expansion coefficient is close to that of glass.