Population PKPD of paclitaxel and carboplatin in ovarian cancer patients: A study by the EORTC-PAMM-NDDG

Introduction:  The paclitaxel/carboplatin doublet is standard treatment in advanced ovarian cancer and usually applied following cytoreductive surgery. Paclitaxel displays nonlinear PK, and the exposure-toxicity relationship has been described by threshold-models, whereas the time above a paclitaxel plasma concentration of 0.05–0.2 μmol/L (T>C0.05-T>C0.2) was correlated with hematological toxicity. A semimechanistic PKPD model has recently been described for paclitaxel-induced myelosuppression. The aim of this study was to investigate the population PKPD of paclitaxel and carboplatin in ovarian cancer patients. Methods:  139 ovarian cancer patients received paclitaxel 175 mg/m2 over three hours, followed by carboplatin AUC 5 mg*min/mL over 30 min, after cytoreductive surgery. Plasma concentration-time data of paclitaxel and carboplatin were measured, and the relationship between PK, hematological toxicity and drug response was evaluated using nonlinear mixed-effect modelling. Data were analysed according to a semiphysiological model, including proliferation –, transit – and circulation compartments for neutrophils and thrombocytes. Linear and sigmoidal maximum response models were tested for the influence of paclitaxel and carboplatin on the proliferation rate of neutrophils or thrombocytes. Toxicity and clinical outcome data were analysed by previously established threshold models. Results:  105 patients had complete PK and toxicity data. In 34 patients with measurable disease, objective response rate was 76%. Hematological toxicity was adequately described by an inhibitory linear drug effect model. Mean paclitaxel T>C0.1 was 16.4 hrs, mean T>C0.05 61.4 hrs. Patients with a complete response to paclitaxel/carboplatin had a higher T>C0.05 as compared to patients with progressive disease (91.9 vs 31.5 hrs, p=0.02). Similarly, T>C0.05 was higher in patients with at least stable as compared to progressive disease (78.3 vs 31.5 hrs, p=0.05). Patients with T>C0.05 > 61.4 hrs (mean value) had a longer time-to-disease-progression compared to patients with T>C0.05 C0.05 was also a good predictor for neutropenia, in that patients with absent or mild (WHO grade 1 or 2) neutropenia had a lower T>C0.05 as compared to patients with severe (WHO grade 3 or 4) neutropenia (50.3 vs 74.1 hrs, p=0.01). Carboplatin drug exposure was the best predictor for thrombocytopenia, in that patients with absent thrombopenia had lower values for Cmax and AUC as compared to patients experiencing mild or severe thrombopenia (82.6 vs 132 μmol/L for Cmax, 1.24 vs 1.79 mg*min/mL for AUC, p < 10−4 for both comparisons). Conclusions:  Paclitaxel T>C0.05 was useful to predict neutropenia and clinical outcome in ovarian cancer patients receiving paclitaxel/carboplatin, while carboplatin exposure was useful to predict thrombocytopenia. Prospective studies should assess the value of therapeutic drug monitoring in this patient group.