A curve-based framework for dissolution profile modeling and formulation optimization.

We present a unified framework for dissolution profile modeling and formulation optimization to identify formulation settings that reproduce a target dissolution profile. In generic drug development, the reference formulation is often unavailable or proprietary, making the underlying release behavior difficult to verify in advance. The proposed framework integrates parametric dissolution models and functional data analysis using functional principal component analysis (FPCA) within a candidate-based formulation-space exploration workflow. Curve similarity is quantified using the integrated squared difference, with the regulatory similarity factor f2 used as an additional evaluation metric. The framework was evaluated using two extended-release dissolution datasets exhibiting different release behaviors. Results demonstrated that model performance depends on the relationship between the modeling approach and the underlying dissolution characteristics. The parametric approach achieved strong performance when the dissolution behavior was well represented by the assumed model structure, whereas FPCA provided a flexible alternative when the underlying release behavior was not adequately described by a predefined model. The proposed framework enables systematic identification of candidate formulations and provides a flexible strategy for dissolution-guided inverse formulation design under uncertain release behavior.