(317f) Digital Design of a Lomustine Manufacturing Process Using Pharmapy

Model-based process design is being increasingly adopted by the pharmaceutical industry as it seeks to modernize process development methods and embrace Industry 4.0 technology and tools. This encompasses modeling, computational and numerical methods development, and experimental planning, leading to the construction of digital twins/virtual plants of the processes under study. In this way, it is possible to explore a variety of process conditions and configurations, aiming for higher productivity, lower processing times, lower waste formation and higher product (consumer) and process (operational) safety. Also, model-based process design supports industry and government decision-making to ensure a safe and affordable supply of critical medicines to the public.

Recently, lomustine, a high-value, low-volume cancer drug has seen a dramatic increase in price as a consequence of limited supply [1]. This unfortunate situation has spurred research into discovery of novel routes for its manufacture. One such effort is recent experimental work which identified a path for continuous lomustine synthesis through a series of telescoped reaction steps [2]. This innovation in lomustine synthesis have motivated the next steps towards optimal production of the drug substance lomustine, specifically i) the analysis of extended processes for the formation and separation of the crystalline active pharmaceutical ingredient (API), and ii) the evaluation and comparison of different process configurations, viz. end-to-end batch or continuous processes, and hybrid schemes involving batch, semi-batch and continuous operations. This extended analysis supports informed decision-making aimed at the successful commercial production of this important API.

The process alternatives considered in this work differ in their operating mode, i.e. batch/continuous/hybrid manufacturing, where unique benefits and concerns exist for each case. In the context of pharmaceutical manufacturing, where small product campaigns are common, the start-up of processes involving continuous operation can be a major issue, greatly impacting product yield and equipment size [3]. On the other hand, batch and semi-batch operation offer operational flexibility and are not limited by residence time; however, some advantages of continuous processing such as consistency of product critical quality attributes (CQAs), better controllability, and reduced equipment size [4] may be lost. For this reason, a framework that enables the systematic evaluation of process alternatives is necessary, to make valid comparisons based on technical and/or economic criteria.

In this work, we present the evaluation and comparison of processing alternatives for lomustine production via digital twin construction and analysis. This is made possible by the use of PharmaPy, a Python-based, object-oriented platform for the development of pharmaceutical processes [5]. PharmaPy offers tools for dynamic model construction and execution under alternative operating modes, and also permits the incorporation of experimental data for model calibration purposes. As a framework for pharmaceutical process development, PharmaPy is designed to facilitate generating digital twins via robust parameter estimation, in this case, using experimental data from ongoing experimental efforts on lomustine [6]. Then, different process alternatives are constructed via object creation and manipulation using Python syntax, as offered by the PharmaPy platform. Each processing alternative includes API synthesis, crystallization and vaporization (solvent switch), and a series of solid-liquid separation units, viz. filtration, deliquoring, washing and drying of the resulting solid cake [7]. Using the set of unit operations included in PharmaPy, a complete representation of the drug substance manufacturing process of lomustine can be conveniently constructed.

Comparison of the process alternatives is made on the basis of economic criteria (capital and operational expenses) for a desired product throughput, while meeting operational constraints (maximum concentrations in relation to solubility constraints, maximum operating temperature/pressure), and product CQAs, e.g. crystal size distribution of the solid cake, residual impurities/solvent(s) of the dried product. To analyze uncertainty in generated digital twin predictions, global sensitivity analysis using random sampling (viz. MonteCarlo and Sobol) is used to evaluate the impact of model parameters (kinetic constants, transport coefficients) and operating conditions (input flows, operating temperature/pressure) on the mentioned CQAs and on capital operational costs. Sensitivity analysis serves as a valuable input for further optimization efforts, identifying critical process parameters (CPPs) that have most impact on the outputs of interest, and also providing insights on of how parametric uncertainty propagates into the model predictions, which can support the determination of the operational design space [8], [9].

References

[1] J. Funaro, H. Friedman, and M. Weant, “A costly ‘rebranding’ of an old drug comes with a 700% price increase,” The Cancer Letter, 27-Sep-2017.

[2] Z. Jaman, T. J. P. Sobreira, A. Mufti, C. R. Ferreira, R. G. Cooks, and D. H. Thompson, “Rapid On-Demand Synthesis of Lomustine under Continuous Flow Conditions,” Org. Process Res. Dev., vol. 23, no. 3, pp. 334–341, 2019.

[3] D. Casas-Orozco et al., “Application of PharmaPy in the digital design of the manufacturing process of an active pharmaceutical ingredient,” in 31st European Symposium on Computer Aided Process Engineering, 2021.

[4] E. Içten et al., “A Virtual Plant for Integrated Continuous Manufacturing of a Carfilzomib Drug Substance Intermediate, Part 1: CDI-Promoted Amide Bond Formation,” Org. Process Res. Dev., 2020.

[5] D. Casas-Orozco et al., “PharmaPy: an object-oriented tool for the development of hybrid pharmaceutical flowsheets,” Comput. Chem. Eng., 2021 (submitted).

[6] J. Mackey et al., “End-to-End Reconfigurable Process Development of a Hybrid Manufacturing System for the Cancer Drug Lomustine,” in AIChE Annual Meeting, 2021.

[7] F. Destro et al., “Digital design of an intensified filtration-drying unit for pharmaceutical upstream manufacturing,” in AIChE Annual Meeting, 2020.

[8] D. Laky, S. Xu, J. S. Rodriguez, S. Vaidyaraman, S. G. Muñoz, and C. Laird, “An optimization-based framework to define the probabilistic design space of pharmaceutical processes with model uncertainty,” Processes, vol. 7, no. 2, 2019.

[9] L. X. Yu et al., “Understanding pharmaceutical quality by design,” AAPS J., vol. 16, no. 4, pp. 771–783, 2014.