Background Chemotherapy agents have dose-related effectiveness. The BTOG2 trial is a large phase III trial supported by the British Thoracic Oncology Group patients addressing this issue in which advanced non-small cell lung cancer were randomised to GC80 vs GC50 vs GCb6. Treatment delays and dose reductions due to toxicity mean that patients do not actually receive planned treatments and the BTOG2 trial provides an opportunity to investigate the delivered dose intensity (DDI) of these treatments in a large group of patients.

Methods Carboplatin dose was calculated using the Calvert equation, incorporating estimated GFR based on the Wright equation including creatinine kinase. Delivered dose intensity (DDI) for each patient was calculated as the mean of the per-cycle DDI which is the ratio of the delivered vs planned dose per day, calculated for platinum and gemcitabine separately.

Results Starting doses for cycle 1 were generally as per protocol. Doses of carboplatin are higher using estimated GFR from the novel Wright formula compared to standard Cockcroft–Gault approach. Dose reductions on cycles 2–4 were more apparent for GC80 compared to GC50 (56% vs 42% with =1 dose reduction) but dose delivered remained high with reductions to median of 77 mg/m² by cycle 4. Dose reduction rate was highest on GCb6 with 71% of patients experiencing =1 reduction, with median dose of AUC 4.5 at cycle 4. Gemcitabine dose reductions parallelled those seen with platins, occurring more frequently with GCb6. Overall DDI for platinum was high for all treatments but lowest for GCb6 (96% vs 99% vs 87%). Response rates were GC50 23%, GC80 33% and GCb6 28%. There was no evidence that dose reductions, treatment delays or DDI was associated with response thus the delivered dose of GC80 was sufficient to generate this 10% difference.

Conclusion Doses of cisplatin at 80 mg/m² and carboplatin at AUC6 based on the Wright formula in combination with gemcitabine are deliverable but individuals have higher chance of treatment delays and dose reductions with carboplatin. However the reduced DDI does not appear to have an effect on clinical outcomes.