(4ii) Sentinel Biomaterials for Precision Alloimmune Tolerance: Prenatal Complications and Transplantation

Driven by the emergence of artificial intelligence (AI) and the Industry 4.0 paradigm, an infinite number of modeling approaches are now available by utilizing both mechanistic and data-driven models. This opportunity has opened new avenues for process modeling in advanced manufacturing, such as pharmaceutical manufacturing. The pharmaceutical industry deals with a large variety of particles, including fine, needle-shaped, cohesive, and poorly soluble powders, and it must meet high-quality standards with limited experimental work. My PhD and postdoctoral research has focused on the modeling, simulation, and optimization of pharmaceutical manufacturing processes, serving as good case studies for integrating mechanistic and data-driven modeling.

Despite significant attention on combining these approaches, numerous challenges remain toward standardization and wider application. Firstly, the differences and appropriate applications of numerous types of combinations are not sufficiently understood. Hybrid modeling, which combines mechanistic and data-driven models to form a new model, is becoming a popular approach. This can be further divided into serial and parallel hybrid modeling. In addition to hybrid modeling, other options are available using both mechanistic and data-driven models, e.g., surrogate modeling and first-principle model identification through machine learning. Secondly, advanced analysis methods, e.g., sensitivity and uncertainty analysis, have yet to be fully developed.

Our group’s research goal is to develop an integration method for mechanistic and data-driven modeling for advanced manufacturing process design. The project consists of three work packages:

i) Process modeling of drug substance and drug product manufacturing

I have worked on mechanistic, hybrid, and data-driven modeling of drug product manufacturing at Ghent University [1, 2, P1]. Currently, I am involved in process modeling of drug substance manufacturing at Purdue University [P2]. After model development, calibration, and validation, the characteristics of each developed model are analyzed from diverse perspectives, such as accuracy, applicability, computational time, and necessary amount of data.

ii) Techno-economic and quality assessment method development

Based on the developed process models, techno-economic and quality assessments are performed for drug substance and product manufacturing. During my PhD at the University of Tokyo, I worked on techno-economic [3] and quality assessment [4] of tablet manufacturing processes. This work package aims to further advance assessment methods to predict net present value and critical quality attributes from material properties of raw materials, process settings, and equipment information. Finally, optimization algorithms are developed to identify necessary unit operations and their process settings once active pharmaceutical ingredients and the target quality profiles are fixed.

iii) Generalization of integrating mechanistic and data-driven models

Alongside the process modeling work in the first work package, possible combinations of mechanistic and data-driven models are comprehensively generated and reviewed. A flowchart suggesting appropriate modeling strategies will be developed based on model requirements, process complexity, and the quality of available data.

Long-term vision

From a long-term perspective, our research group aims to extend the applicability of process systems engineering (PSE) methodologies to various domains, including biologics, water treatment, energy resources, food production, and consumption behavior, thus showcasing the versatility of PSE on a global scale. Since PSE can incorporate new techniques, e.g., generative AI, our research group will continuously seek advanced methods from diverse research fields and integrate them as PSE tools.

[1] Kensaku Matsunami, Jonathan Meyer, Martin Rowland, Neil Dawson, Thomas De Beer, Daan Van Hauwermeiren. “T-shaped partial least squares for high-dosed new active pharmaceutical ingredients in continuous twin-screw wet granulation: Granule size prediction with limited material information,” Int. J. Pharm., 646, 123481 (2023)

[2] Kensaku Matsunami, Tuur Vandeputte, Ana Alejandra Barrera Jiménez, Michiel Peeters, Michael Ghijs, Daan Van Hauwermeiren, Fanny Stauffer, Eduardo dos Santos Schultz, Ingmar Nopens, Thomas De Beer. “Validation of model-based design of experiments for continuous wet granulation and drying,” Int. J. Pharm., 646, 123493 (2023)

[3] Kensaku Matsunami, Fabian Sternal, Keita Yaginuma, Shuichi Tanabe, Hiroshi Nakagawa, Hirokazu Sugiyama. “Superstructure-based process synthesis and economic assessment under uncertainty for solid drug product manufacturing,” BMC Chem. Eng., 2, 6 (2020)

[4] Kensaku Matsunami, Tomohiro Miura, Keita Yaginuma, Shuichi Tanabe, Sara Badr, Hirokazu Sugiyama. “Surrogate modeling of dissolution behavior toward efficient design of tablet manufacturing processes,” Comput. Chem. Eng., 171, 108141 (2023)

[P1] Kensaku Matsunami, Viktor Bultereys, Laure Descamps, and Ashish Kumar “In-depth understanding of the impact of material properties on the performance of jet milling of active pharmaceutical ingredients,” 2024 AIChE Annual Meeting

[P2] Kensaku Matsunami, Zoltan K. Nagy “Development of robust design space for integrated pharmaceutical processes through uncertainty analysis,” 2024 AIChE Annual Meeting

Teaching Interests

My primary education goal is to help students understand the value of chemical engineering from a holistic perspective and apply chemical engineering skills for their future careers. I believe that chemical engineering will play a critical role in improving industrial and societal systems toward sustainability. Therefore, I aim to educate students to utilize chemical engineering skills in academia, industry, and government globally.

I have extensive teaching experience as a teaching assistant and guest lecturer. During my PhD at the University of Tokyo, I taught basic chemical engineering, including transport phenomena, reaction engineering, and life-cycle assessment to first-year undergraduate students. I also mentored and supported undergraduate students in practical chemical engineering courses. At Ghent University, I lectured on process and equipment design for master’s students and organized an intensive course on flowsheet modeling for PhD students. These experiences have equipped me to teach chemical engineering courses tailored to the backgrounds and levels of students.

As a professor, I am interested in teaching both theoretical and practical courses in chemical engineering, particularly PSE. I would like to develop a course on process and equipment design using realistic data and scenarios, involving industrial partners. Educational exchanges with different universities would further enhance the quality of chemical engineering education worldwide. Furthermore, I will continuously update the courses, incorporating new techniques and topics, and ensure the courses are immediately relevant to societal needs.