Explaining the unexplainable: discrepancies in results from the CALGB/SWOG 80405 and FIRE-3 studies

Abstract

We propose a working hypothesis that integrates data from the CALGB/SWOG 80405 and FIRE-3 studies to explain apparent discrepancies in their results. Both trials assessed the combination of either cetuximab or bevacizumab with a different chemotherapy backbone: irinotecan in all patients in the FIRE-3 study, or oxaliplatin in 75% of the patients in the CALGB/SWOG 80405 study. The hypothesis is divided into three parts. Firstly, in addition to the biology or microenvironment of the tumour and the selection of the biologically targeted agents common to both trials, chemotherapy itself is an important variable that determines treatment efficacy because of a complex interplay between the biological therapy, the chemotherapy, and the microenvironment. Secondly, the tumour microenvironment, as defined by the Consensus Molecular Subtypes (CMS) classification, determines the interaction of chemotherapeutic agents with biologically targeted agents such as bevacizumab and cetuximab. Whereas irinotecan synergises with cetuximab across all CMS subtypes, oxaliplatin might have variable effects, synergising with cetuximab in fibroblast-poor microenvironments, such as CMS2 and CMS3, but activating fibroblast-rich microenvironments, such as CMS1 and CMS4, to release cytokines that might antagonise some of the cetuximab effects. Thirdly, the previous assumptions integrate into a final concept, which is that overall survival is determined not only by the biological therapy or the first-line treatment, but specifically by the sequence of first-line and second-line regimens, and the degree of synergism between them. In a clinical setting, the optimal first-line combination of biological therapy and chemotherapy predetermines the crossover to a specific second-line treatment, which affects the overall survival of a patient with a specific tumour subtype. Our working hypothesis suggests that the CALGB/SWOG 80405 and FIRE-3 studies are complementary rather than discrepant, and it provides an explanation for their opposing interpretations. In conclusion, proper interpretation of the CALGB/SWOG 80405 and FIRE-3 results requires an in-depth examination of the complex interplay, not only between the targeted biological agents and chemotherapeutic drugs, but also between therapies and the tumour biology and microenvironment, for each line of treatment.