How Disease Models can reduce late phase attrition rate to half by 2015
Only 5% of new molecules make it to the market in oncology, lowest compared to other therapeutic areas. Yet, cancer is one of the leading causes of deaths in US, and non-small cell lung cancer (NSCLC) being the top cause within cancer deaths. Given the urgent need for more effective NSCLC treatments and the low yield drug development, we elected to understand risk factors for death in patients with NSCLC and whether anticancer drug activity can be characterized more precisely early in clinical development based on predictive biomarker, such as tumor size. This knowledge might then aid drug developers to better screen drug molecules, design trials and select doses. Four registration trials for NSCLC provided nine different regimens that are either first-line or second-line treatments for locally advanced or metastatic NSCLC. Tumor size dynamic data were described with a disease model that incorporates both the tumor growth property and the regimen's anti-tumor activity. Patient survival times were described with a parametric survival model that includes various risk factors and tumor size change as predictors. Among 11 potential risk factors for survival, ECOG score and baseline tumor size were found to be significantly related to survival in almost all regimens. The disease model describes the longitudinal tumor size data fairly well, especially for early weeks after treatment initiation. The survival model based on one regimen predicted the survival outcomes for the other eight regimens reasonably well despite that these regimens have different mechanism of actions and were studied in different trials.
The drug effects on tumor size from early clinical trials in conjunction with the NSCLC model can be applied to screen compounds, simulate NSCLC clinical trials to optimize designs leading to more successful registration trials.