The proposed experiment consists of two steps: first, an extensive parametric study of human airways will be executed using advanced HPC to create a database of solutions via CFD. This extensive study will not use the conventional methodology of executing CT scans and cleaning them for CFD, but use a morphing tool based on Radial Basis Functions, to parameterize and adapt a single baseline geometry into new ones. Then, a vast amount of numerical simulations will be used to train the ROM and build the digital twin, allowing to predict flow and drug deposition on any lung geometry and visualize it without having to execute a CFD simulation. Such a tool will be able to provide patient-specific results in real-time and at affordable costs.