Purpose: Paclitaxel and carboplatin are frequently used in advanced ovarian cancer following cytoreductive surgery. Threshold models have been used to predict paclitaxel pharmacokinetic-pharmacodynamics, whereas the time above paclitaxel plasma concentration of 0.05 to 0.2 μmol/L (t C > 0.05-0.2) predicts neutropenia. The objective of this study was to build a population pharmacokinetic-pharmacodynamic model of paclitaxel/carboplatin in ovarian cancer patients. Experimental Design: One hundred thirty-nine ovarian cancer patients received paclitaxel (175 mg/m 2) over 3 h followed by carboplatin area under the concentration-time curve 5 mg/mL*min over 30 min. Plasma concentration-time data were measured, and data were processed using nonlinear mixed-effect modeling. Semiphysiologic models with linear or sigmoidal maximum response and threshold models were adapted to the data. Results: One hundred five patients had complete pharmacokinetic and toxicity data. In 34 patients with measurable disease, objective response rate was 76%. Neutrophil and thrombocyte counts were adequately described by an inhibitory linear response model. Paclitaxel t C > 0.05 was significantly higher in patients with a complete (91.8 h) or partial (76.3 h) response compared with patients with progressive disease (31.5 h; P = 0.02 and 0.05, respectively). Patients with paclitaxel tC > 0.05 > 61.4 h (mean value) had a longer time to disease progression compared with patients with paclitaxel tC > 0.05 < 61.4 h (89.0 versus 61.9 weeks; P = 0.05). Paclitaxel tC > 0.05 was a good predictor for severe neutropenia (P = 0.01), whereas carboplatin exposure (Cmax and area under the concentration-time curve) was the best predictor for thrombocytopenia (P < 10-4). Conclusions: In this group of patients, paclitaxel tC > 0.05 is a good predictive marker for severe neutropenia and clinical outcome, whereas carboplatin exposure is a good predictive marker for thrombocytopenia.