Right or Left Primary Site of Colorectal Cancer: Outcomes From the Molecular Analysis of the AGITG MAX Trial

Abstract

Background

For metastatic colorectal cancer, previous reports have described differences in biology and outcome, including response to biologic agents, based on whether the primary tumor is right- or left-sided. We explored the molecular markers from the AGITG MAX trial.

Patients and Methods

The AGITG MAX trial was a randomized study comparing capecitabine versus capecitabine + bevacizumab versus capecitabine + bevacizumab + mitomycin C as first-line therapy in advanced colorectal cancer. Patients were classified as having right-sided (caecum to transverse colon) or left-sided (descending colon to rectum) disease according to anatomic location. Baseline characteristics and previously described molecular profiles were compared by side of primary tumor. Survival outcomes were analyzed by the Kaplan-Meier approach and proportional hazards regression modeling.

Results

Among the 471 patients, the location of primary tumor was known in 440 patients (93%). Molecular profile was known in 298 patients (63%). Twenty-eight percent had right-sided primary tumors. Major differences between right and left are as follows: female 49% versus 33% (P < .01), BRAF mutant 16% versus 3.5% (P ≤ .001), and phosphatase and tensin homolog (PTEN) loss 27.6% versus 53% (P = .01). There were no differences in RAS mutation, PIK3CA mutation, or high versus low expression of assessed angiogenic markers. Right-sided primary lesion predicted a poor outcome for median overall survival: right-sided disease 13.2 months versus left-sided disease 20 months (P = .001; hazard ratio [HR] = 0.67; 95% confidence interval [CI], 0.53-0.85), but not for progression-free survival (HR 0.96; 95% CI, 0.78-1.20). The relative treatment effect did not differ significantly according to location of primary tumor: right primary tumor HR (bevacizumab containing arm vs. capecitabine monotherapy arm) was 0.82 (95% CI, 0.54-1.22), and left primary HR (bevacizumab containing arm vs. capecitabine monotherapy arm) was 0.51 (95% CI, 0.4-0.63) (interaction P = .10).

Conclusion

There are more negative prognostic factors in patients with right-sided primary tumors, in particular high BRAF mutations, and these patients have inferior overall survival compared to those with a left-sided primary tumor. There was no suggestion that side of primary site had any impact on bevacizumab effect on progression-free survival.

Introduction

Colorectal cancer (CRC) is a major health problem, with an estimated 1.2 million new cases and over 600,000 deaths occurring each year worldwide.¹ There are differences in reported population rates, with higher rates of CRC in New Zealand and Australia compared to other countries.² Hereditary familial conditions also define separate subgroups based on potentially differing biologic behavior and molecular profiles.

Over the last few years, considerable research analyzing the molecular differences within the colon itself has been a focus culminating in the 2012 report from the Cancer Genome Atlas project.³ Furthermore, there is renewed interest in clinical markers as predictors of cancer behavior. Side of primary site (right vs. left) for CRC is an example of this. The study of Bufill⁴ in 1990 was one of the first to propose differences in biology and outcome based on whether the primary lesion for metastatic CRC (mCRC) was right- or left-sided. There are a number of differences between the sides of the bowel—for example, embryologic beginnings, with the right bowel arising from the midgut and the left side from the hind gut. The vascular supply is therefore divided on this basis, and there are also capillary network⁵ and crypt length variations.⁶ Subsequent reports support the division of right- and left-sided CRCs.7, 8

As regards presentation and symptomatology in relation to the primary site, the differences between right- and left-sided primary colon cancers appear to be as follows: right-sided colon tumors are more likely to be seen in older female patients who usually present with subtle signs and symptoms such as microcytic anemia and weight loss. These are therefore generally diagnosed at a more advanced tumor stage. From the pathology perspective, right-sided colon tumors have a higher tumor grade and are more frequently associated with mucinous histology and exophytic growth patterns. By contrast, left-sided CRCs are seen more commonly in younger patients who present with obstructive symptoms or rectal bleeding at diagnosis, with often infiltrating lesions. Further support for molecular differences based on the primary site in the bowel comes from a recent analysis of KRAS, BRAF, and microsatellite instability (MSI).⁹ Thus, right-sided tumors are usually more likely to be associated with MSI, diploidy, epidermal growth factor receptor (EGFR) expression, BRAF mutation, and possibly KRAS mutation. In contrast, it has been postulated that left-sided CRC is associated with higher chromosomal instability, p53 mutation, COX2 expression, and aneuploidy.¹⁰ All these factors may contribute to the difference observed in patient prognosis, with increasing pooled data demonstrating a shorter overall survival (OS) for patients with right-sided colon tumors; a clear prognostic value of sidedness has therefore been proposed.

Differences in response to biologic agents have also been reported based on the side of the primary lesion. The effect of the primary tumor is more pronounced in patients treated with anti-EGFR therapy. In patients with left-sided tumors, treatment with anti-EGFR therapy in the first-line setting leads to a significantly longer median OS compared to patients treated with bevacizumab.¹¹ Left-sided primary tumors also tend to respond better to chemotherapy + bevacizumab compared to right-sided colon tumors.12, 13

The phase 3 MAX trial (capecitabine vs. capecitabine + bevacizumab [± mitomycin C]) has confirmed improved progression-free survival (PFS) with the addition of bevacizumab to capecitabine.¹⁴ The availability of a control arm without bevacizumab treatment makes this an ideal data set for examining predictive factors for bevacizumab. We have previously published molecular markers from the MAX trial: extended RAS, BRAF, phosphatase and tensin homolog (PTEN), phosphatidylinositol 3-kinase (PIK3CA), vascular endothelial growth factor (VEGF), interleukin (IL)-6 and IL-8, basic fibroblast growth factor (BFGF), and platelet-derived growth factor (PDGF-BB).15, 16, 17, 18, 19

We assessed the panel of markers based on right or left primary lesion site to assess if the primary site affects outcome with bevacizumab when combined with capecitabine, and to assess differences in biomarker patterns based on side.