(567c) Integrated System Modeling That Connects the Drug Product Manufacturing Process to Dissolution Testing

The quality-by-design paradigm in the pharmaceutical industry has been systematically incorporating science-based process understanding in all manufacturing stages. This increased interest in integrating mechanistic process models into the development workflow has led to improved product quality, reduced consumption of raw materials, reduced time to market, and reduced risk [1,2]. Adapting smart manufacturing also referred to as Industry 4.0 practices, provides further digitalization opportunities involving the application of model-based methodologies to the pharmaceutical industry [3].

Process system modeling or flowsheet modeling evaluates mechanistic models for unit operations when connected in series. Considering that a manufacturing process consists of multiple unit operations in series, a system modeling approach can aid design space exploration and optimized product quality testing.

In drug product manufacturing, the dry granulation process is commonly used for moisture-sensitive and highly cohesive powders. The manufacturing route consists of multiple unit operations: feeder, roller compactor, mill, and tablet press. The importance of a process system model in providing a fundamental understanding of the dry granulation process with limited experimental calibration has been demonstrated previously by Gavi, Emmanuela, and Gavin K. Reynolds [4].

In the presented work, the impact of material property in the dry granulation process system model is studied. The following improvements are made to the existing workflow [4].

1. A material sparing approach developed by Sousa, Ricardo, et al. [5] is used to model the roller compaction process. This is a simplified Johanson model where the ribbon strength is independent of the nip angle, thereby removing the need to measure the effective angle of internal friction and angle of wall friction of the feed material.
2. The impact of ribbon porosity on the screen mill model parameters is evaluated to predict the particle size distribution. The results show that the fines fraction in the bimodal fragment distribution function is highly sensitive to the ribbon porosity.
3. A bi-modal particle size distribution is also used to predict the tablet disintegration and dissolution process. The study compares the deviation of model prediction when considering a uniform, unimodal and bimodal particle generation during the tablet disintegration process.

A detailed analysis of multiple unit operations, such as roller compactor, screen mill, tablet press, and QC dissolution, provides the effect of each process parameter on final tablet product performance. The calibrated ‘digital twin’ is further utilized in decision-making, including evaluation of design and scale-up decisions, process optimization at the pilot-scale, sensitivity analysis, and technology transfer.

[1] Hakemeyer, Christian, et al. "Process characterization and design space definition." Biologicals 44.5 (2016): 306-318.

[2] Rogers, Amanda, and Marianthi Ierapetritou. "Challenges and opportunities in modeling pharmaceutical manufacturing processes." Computers & Chemical Engineering 81 (2015): 32-39.

[3] Litster, James, and Ian David L. Bogle. "Smart process manufacturing for formulated products." Engineering 5.6 (2019): 1003-1009.

[4] Gavi, Emmanuela, and Gavin K. Reynolds. "System model of a tablet manufacturing process." Computers & chemical engineering 71 (2014): 130-140.

[5] Sousa, Ricardo, et al. "Roller compaction scale-up made simple: an approximate analytical solution to johanson's rolling theory." Journal of Pharmaceutical Sciences 109.8 (2020): 2536-2543.

[6] Process Systems Enterprise, gPROMS, www.psenterprise.com/products/gproms, 1997-2020