Thrombocytopenia, the dose-limiting toxicity of carboplatin, is manageable and predictable with the dosing equation: Dose (mg/m2) = (0.091) (creatinine clearance)/(body surface area) (desired percentage change in platelet count) + 86. We used this equation to dose patients receiving carboplatin (day 1) and etoposide (80 mg/m2 days 1 to 3). An initial cohort of 14 patients with non-small-cell lung cancer (NSCLC) were treated with 75% of the calculated dose of carboplatin (29 evaluable courses) as a precaution against added myelosuppression from combination chemotherapy. The observed reduction in platelets was essentially equal to the reduction in platelets predicted if patients had received carboplatin alone. Subsequently, a second cohort of 20 evaluable patients with NSCLC received the full calculated dose of carboplatin and etoposide (51 evaluable courses). There was a linear relationship between observed (o) and predicted (p) reductions in platelets. With full-dose carboplatin in combination with etoposide, there was a significantly greater carboplatin dosage administered, greater reduction in platelets, and lower platelet nadir. Among 34 evaluable patients (80 courses) there was one complete response (CR) and four partial responses (PRs) for an overall response rate of 15% (90% confidence +/- 9%). The median duration of survival for responders was 336+ days and for nonresponders was 204+ days. Therapy was well tolerated. This study, in addition, supports our concept of individualized dosing of carboplatin and the underlying pharmacokinetic/pharmacodynamic relationships and represents an interesting pharmacodynamic and quantitative approach to studying potential drug-drug interactions and defining appropriate dosages for combination chemotherapy.