e14074 Background: The glomerular filtration rate (GFR) is essential for carboplatin chemotherapy dosing, however, the best method to estimate GFR in patients with cancer is unknown. We identify the most accurate and least biased method. Methods: Data on age, sex, height, weight, serum creatinine, and results for GFR from 51Cr-EDTA excretion measurements (51Cr-EDTA GFR) were obtained from Caucasian patients aged 18 years or older with histologically confirmed cancer diagnoses at the University of Cambridge Hospital NHS Trust, UK. We developed a new multivariable linear model for GFR using statistical regression analysis. 51Cr-EDTA GFR was compared to the estimated GFR (eGFR) from seven published and our new model using an internal validation data set and root-mean-squared-error (RMSE) and median residuals. A comparison of carboplatin dosing accuracy based on an absolute percentage error more than 20% (APE > 20%) was undertaken. Results: Between August 2006 and January 2013 data from 2,471 patients were obtained. The new model improved the eGFR accuracy (RMSE 15.00ml/min (95% CI 14.12-16.00)) compared to all published models. Body surface area (BSA) adjusted CKD-EPI was the most accurate published models for eGFR (RMSE 16.30ml/min (95% CI 15.34-17.38)) for the internal validation set. Importantly, the new model reduced the fraction of patients with a carboplatin dose APE > 20% to 14.17% in contrast to 18.62% for BSA adjusted CKD-EPI and 25.51% for the Cockcroft-Gault model. The results were externally validated. Conclusions: In a large data set, from patients with cancer, a new model improves eGFR and carboplatin dose calculations, when compared to BSA adjusted CKD-EPI, the model we identified as the best published model for determination of eGFR in patients with cancer.