Predictive value of chromosome 18q11.2-q12.1 loss for benefit from bevacizumab in metastatic colorectal cancer: A post hoc analysis of the randomized phase III-trial AGITG-MAX

Abstract

The VEGF-A monoclonal antibody bevacizumab is currently recommended for first-line treatment of all metastatic colorectal cancer (mCRC) patients. Cost-benefit ratio and side-effects however necessitate patient selection. A large retrospective yet nonrandomized study showed that patients with loss of chromosome 18q11.2-q12.1 in the tumor and treated with bevacizumab have 3 months improved median progression-free (PFS) and overall survival (OS) benefit compared to patients without this loss and/or treatment modality. Implementation for loss of chromosome 18q11.2-q12.1 as a marker in clinical practice mandates evidence in a randomized controlled trial for bevacizumab. Of the trials with randomization of chemotherapy vs chemotherapy with bevacizumab, the AGITG-MAX trial was the only one with tumor materials available. Chromosome 18q11.2-q12.1 copy number status was measured for 256 AGITG-MAX trial patients and correlated with PFS according to a predefined analysis plan with marker-treatment interaction as the primary end-point. Chromosome 18q11.2-q12.1 losses were detected in 71% of patients (181/256) characteristic for mCRC. Consistent with the nonrandomized study, significant PFS benefit of bevacizumab was observed in patients with chromosome 18q11.2-q12.1 loss (P = .009), and not in patients without 18q loss (P = .67). Although significance for marker-treatment interaction was not reached (P_interaction = .28), hazard ratio and 95% confidence interval of this randomized cohort (HR_interaction = 0.72; 95% CI = 0.39-1.32) shows striking overlap with the nonrandomized study cohorts (HR_interaction = 0.41; 95% CI = 0.32-0.8) supported by a nonsignificant Cochrane χ² test (P = .11) for heterogeneity. We conclude that post hoc analysis of the AGITG-MAX RCT provides supportive evidence for chromosome 18q11.2-q12.1 as a predictive marker for bevacizumab in mCRC patients.

Keywords: anti-VEGF monoclonal antibody; bevacizumab; chromosome 18q; metastatic colorectal cancer; predictive biomarker; randomized controlled trial.

FIGURE 1

Progression free survival characteristics stratified by bevacizumab treatment and chromosome 18q11.2‐q12.1 copy number status. (A) Kaplan‐Meier analysis for the 18q‐evaluable AGITG‐MAX RCT cohort (N = 256); (B) Kaplan‐Meier analysis for the combined nonrandomized study cohorts (N = 616). Patients treated with capecitabine monotherapy (C, blue lines), with capecitabine and bevacizumab with‐ or without mitomycin (B + M, red lines), patients with chromosome 18q11.2‐q12.1 loss tumors (dashed lines), with no‐loss tumors (solid lines). Number of patients at risk below the x‐axis. (C) Forest plot of hazard ratio (HRs, black squares) and 95% confidence intervals (CIs), horizontal lines bevacizumab vs no‐bevacizumab patients. First column; patient selection, randomized (MAX) and nonrandomized by 18q11.2‐q12.1 status (loss and no loss); second column number of patients (#patients). [Correction added on 26 July 2022, after first online publication: Figure 1B y‐axis label has been changed from OS to PFS.] [Color figure can be viewed at wileyonlinelibrary.com]