(301f) New Methods for Design Space Identification and a Re-Interpretation of ICH Endorsed Guide for Q8Q9Q10 Implementation

Since its inception, the concept of the design space has always been a model-based one [1]. One cannot establish a bridge between two experiments without the assumption of continuity. The tools and methods used to support this continuity determine the type of experiments required to build the continuum upon which a design space is to be built. This continuity is pragmatically defined with a mathematical model. Establishing a design space strictly through empirical means implies a heavy burden down the road as it pertains the verification and lifecycle of the design space. This is mostly driven by the need to verify both, the structure of the continuum (equations) and its parameters. The use of a mathematical formulation that is supported by fundamental principles of physics, chemistry and biology permutes the verification of the equation structure with the verification of the assumptions that sustained the choice of mathematical model during development. If the assumptions still hold in the system, the model must hold too. Furthermore, the verification of a probability postulation remains subject to the same stochasticity and offers inconsequential evidence given the number of at-scale experimentation a pharmaceutical company is realistically willing to do. This talk will present a review of recent methods to derive candidate design spaces and furthermore refine them into multivariate operational ranges to be used in master batch records in commercial manufacture [2,3,4]. A discussion will follow as to the logical ways towards verification and lifecycle of the multiple components that assemble the design space [5].

[1] Garcia-Munoz, S., Luciani, C.V., Vaidyaraman, S. and Seibert, K.D., 2015. Definition of design spaces using mechanistic models and geometric projections of probability maps. Organic Process Research & Development, 19(8), pp.1012-1023.

[2] Laky, D., Xu, S., Rodriguez, J.S., Vaidyaraman, S., García Muñoz, S. and Laird, C., 2019. An optimization-based framework to define the probabilistic design space of pharmaceutical processes with model uncertainty. Processes, 7(2), p.96.

[3] Kucherenko, S., Giamalakis, D., Shah, N. and García-Muñoz, S., 2020. Computationally efficient identification of probabilistic design spaces through application of metamodeling and adaptive sampling. Computers & Chemical Engineering, 132, p.106608.

[4] Zhao, F., Grossmann, I.E., García‐Muñoz, S. and Stamatis, S.D., 2020. Flexibility index of black‐box models with parameter uncertainty through derivative‐free optimization. AIChE Journal, p.e17189.

[5] ICH-Endorsed Guide for ICH Q8/Q9/Q10 Implementation, 2011, International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use