Personalising docetaxel and G-CSF schedules in cancer patients by a clinically validated computational model

BACKGROUND: This study was aimed to develop a new method for personalising chemotherapeutic and granulocyte colony-stimulating factor (G-CSF) combined schedules, and use it for suggesting efficacious chemotherapy with reduced neutropenia. METHODS: Clinical data from 38 docetaxel (Doc)-treated metastatic breast cancer patients were employed for validating a new pharmacokinetic/pharmacodynamics model for Doc, combined with a mathematical model for granulopoiesis. An optimisation procedure was constructed and used for selecting improved treatment schedules. RESULTS: The combined model accurately predicted observed nadir timing (r=0.99), grade 3 or 4 neutropenia (86% success) and neutrophil counts over time in individual patients (r=0.63), and showed robustness to CYP3A-induced variability in Doc clearance. For average patients, the predicted optimal support for the standard chemotherapy regimen, Doc 100 μg m(−2) tri-weekly, is G-CSF, 300 μg, Q1D × 3, starting day 7 post-Doc. This regimen largely moderates chemotherapy-induced neutrophil nadir and neutropenia duration. The more intensive Doc dose, 150 mg m(−2), is optimally supported by the slightly less cost-effective G-CSF 300 μg, Q1D × 4, 5 days post-Doc. The latter regimen is optimal for borderline patients (2000 neutrophils per μl) under Doc, 100–150 mg m(−2) tri-weekly. CONCLUSIONS: The new computational method can serve for tailoring efficacious cytotoxic and supportive treatments, minimising side effects to individual patients. Prospective clinical validation is warranted.