Exploration of the effect of treatment to progression with the VEGF signaling inhibitor cediranib (CED) plus chemotherapy (CT) on the results of the HORIZON (HZ) II trial in first-line metastatic colorectal cancer.

Abstract

e14052^ Background: The HZII trial (CED + FOLFOX [FF] or XELOX [X] vs placebo [PBO] + FF or X) met the co-primary endpoint of improved progression-free survival (PFS) with CED (hazard ratio [HR]=0.84), but not overall survival (HR=0.94). PFS results were similar to those of the N016966 trial (bevacizumab + FF4 or X vs PBO + FF4 or X; HR=0.83) in which treatment to progression was considered to affect outcome (Saltz et al 2008). The HZII protocol specified that patients (pts) should be treated to progression with CT + randomized treatment (RT), although individual components could be stopped for tolerability. This analysis assessed whether treatment to progression in HZII impacted PFS. Methods: Pts were randomized to CED or PBO, each with FF or X (per physician preference), and considered to be treated to progression if they received treatment to within 28 days of their disease progression date. Separate Cox proportional hazards models that included time-dependent covariate effects for the number of 5-FU/capecitabine cycles received were fitted for FF and X pts, and the treatment-effect PFS HRs (adjusted for amount of CT received to progression) assessed. Results: The proportion of pts treated to progression with RT was high in both arms (72% CED vs 81% PBO), but the proportion treated to progression with RT + fluoropyrimidine was lower in the CED vs PBO arm (42% vs 58%, respectively). Analysis by CT type showed more pts treated to progression with RT + oral capecitabine (48% CED vs 68% PBO) than RT + iv 5-FU (33% CED vs 45% PBO), but the difference between arms remained. The treatment-effect PFS HRs for pts receiving FF and X, adjusted for 5-FU/capecitabine cycles received, were 0.81 and 0.79, respectively. Conclusions: Treatment to progression with CT + CED may have a beneficial effect in colorectal cancer. Although the overall proportion of pts treated to progression with CED was high, adverse events probably affected the ability to maintain CT + CED to progression, which may have influenced overall outcome. As this analysis was not a randomized comparison, the magnitude of effect is difficult to quantify due to confounding factors.