We aimed at developing a OS model incorporating TK metrics in platinum-resistant (PR)-ovarian cancer (OC) patients using data from the randomized, open label, phase 3 AURELIA trial designed to compare PFS in patients treated with chemotherapy alone (CT) or in combination with Bevacizumab (B).
Individual data from 361 patients randomly allocated to the B+CT or CT were available. Three types of CT were evenly distributed in both arms. Tumor size reported as sum of lesion diameter (SLD, RECIST 1.0) was collected at baseline and every 8 to 9 weeks until disease progression. Patients continued to be followed for OS even after treatment discontinuation. A non-linear mixed effect TK model accounting for the dynamics of tumor growth, drug effect and treatment resistance was used to fit the SLD. The final TK model contained a term of resistance specific to each CT type in both study arms while drug effect and tumor growth rate were common to all patients. TK metrics indicated that most of the patients experienced tumor shrinkage with a more profound shrinkage in the B+CT arm. From this model, two TK metrics were derived for each patient: early shrinkage at week 8 (ETS8) and predicted SLD at treatment onset (TS0). Then a Cox proportional-hazard survival model was developed. Covariates including ECOG and FIGO score at baseline, histological grade and subtype, time from 1st diagnosis to treatment onset, presence of ascites, CA-125 at baseline, TS0, and ETS8 were tested as prognostic factors.
In the bootstrap-based covariate analysis, two sets of factors were found to be influential on survival time: those reflecting the disease severity (ECOG, FIGO stage, presence of ascites) and those describing key features of the TK (ETS8 and TS0). Treatment group was not retained in the final model as its effect was masked by the influence of the other covariates.
For the 1st time, an OS model including individual TK metrics was developed for PR-OC patients. This analysis confirms previous findings indicating the low predictive value of CA-125 in this population. While treatment group and baseline CA-125 were not found to influence survival time, model derived TK metrics were predictive of OS, ETS8 notably.
Clinical trial identification
AURELIA study: NCT00976911 Completed July 2014
Legal entity responsible for the study
F. Hoffman La Roche Ltd
F. Hoffman La Roche Ltd
A. Sostelly: Employee of F-Hoffman La Roche (Basel, Switzerland). F. Jaminion, F.J. Mercier: Employee of F. Hoffman-La Roche Ltd.