Digital Twins and Virtual Populations: Applications in Quantitative Systems Pharmacology

Quantitative Systems Pharmacology (QSP) models face challenges in estimating parameters from patient-specific data to obtain their digital counterparts. To this end, global and local fitting strategies involving optimization methods such as least squares and Bayesian inference are employed. In addition, identifiability and uncertainty quantification techniques can be instrumental in developing robust QSP models. Virtual populations generated by statistical approaches and Monte Carlo simulations can capture patient variability and incorporate genetic, demographic, and treatment factors. Digital patients (or twins) and virtual populations can help optimize dosing, inform clinical trials, and aid in understanding diseases. Here, we report two examples of their applications in studying neurofilament trafficking in spinal muscular atrophy patients and the rare Gaucher disease type 1, showing promising results. Overall, QSP combined with digital patients and virtual populations has the potential to push drug development toward personalized medicine.