Original Study
Impact of Primary Tumor Site on Bevacizumab Efficacy in Metastatic Colorectal Cancer

https://doi.org/10.1016/j.clcc.2016.02.007Get rights and content

Abstract

Background

With an ever-increasing focus on personalized medicine, all factors known to affect treatment response need to be considered when defining optimal therapy for individual patients. While the prognostic impact of primary tumor site on colorectal cancer (CRC) outcomes is established, emerging data suggest potential differences in response to biologic therapies. We studied the impact of tumor site on bevacizumab efficacy in patients with metastatic CRC.

Patients and Methods

We analyzed data of patients in an Australian prospective multicenter metastatic CRC (mCRC) registry who received first-line chemotherapy. Tumor site was defined as right colon, cecum to transverse; left colon, splenic flexure to rectosigmoid; and rectum. Kaplan-Meier and Cox models were used for survival analyses.

Results

Of 926 patients, 297 had right colon, 354 left colon, and 275 rectum primary disease. Median age was 68.6, 65.9, and 63.3 years, respectively (P = .001). Right colon disease was significantly associated with intraperitoneal spread (P < .0001), while left colon and rectum disease preferentially metastasized to the liver and lungs, respectively (P < .0001 in both settings). A total of 636 patients (68.7%) received bevacizumab. Progression-free survival was superior for bevacizumab-treated patients in all groups but appeared greatest in right colon disease (hazard ratio, 0.46; 95% confidence interval, 0.36-0.60; P ≤ .001). Overall survival was longest in patients with disease of the rectum, followed by left colon and right colon (median, 26.2, 23.6, and 18.2 months, respectively; P = .0004).

Conclusion

Tumor site appears to be prognostic in mCRC, with rectum and right colon disease associated with the best and worst outcomes, respectively. Patients who received bevacizumab in addition to chemotherapy had superior outcomes, with the effect appearing greatest in patients with right colon disease.

Introduction

Colorectal cancer (CRC) is a heterogeneous disease that evolves through varied genetic and epigenetic molecular pathways.1, 2 In 1990, Bufill proposed the concept of developmental and biologic differences in the proximal and distal colon resulting in differing susceptibility to neoplastic transformation, with proximal and distal colon cancers arising through different pathogenic mechanisms.3 Findings of varying molecular genetic alterations in proximal and distal cancers supported this concept.4 Since then, growing evidence regarding molecular changes at colonic subsites has supported as well as challenged the dichotomy of proximal (right-sided) colon versus distal (left-sided) colorectum divided at the splenic flexure.5, 6, 7, 8, 9, 10

Regardless of whether a strict right- versus left-sided dichotomy exists or molecular changes occur as a continuum, clear differences between cancers of the right colon and the left colorectum have been observed—eg, right-sided tumors are associated with CpG island methylator phenotype–high and microsatellite instability–high and with BRAF mutations, and left-sided tumors exhibit chromosomal instability.9, 11, 12 These factors may contribute to variations in patient outcomes, with consistent reports of the association of right-sided tumors and poorer prognosis.7, 10, 12, 13 Importantly, in the era of personalizing the management of metastatic CRC (mCRC), emerging data highlight the potential importance of primary tumor location (proximal vs. distal) in predicting response to targeted therapies. Indeed, 3 recent analyses reported that in patients with KRAS wild-type tumors, the efficacy of the anti–epidermal growth factor receptor (EGFR) monoclonal antibody cetuximab was confined to left-sided tumors (positioned at or distal to the splenic flexure).14, 15

For the anti-VEGF (vascular endothelial growth factor) agent bevacizumab, results to date regarding the impact of primary tumor site are conflicting. Boisen et al16 concluded the impact of bevacizumab may be site specific. In their analysis of an initial cohort of 213 patients who received capecitabine and oxaliplatin (CAPOX) alone, they found no impact of primary site on progression-free survival (PFS) or overall survival (OS). In a later cohort of 667 mCRC patients treated with CAPOX and bevacizumab, the authors noted superior survival outcomes in patients with sigmoid and rectal primary cancers versus those with colon cancers at any other location. In contrast, using the more traditional definition of proximal (right-sided) versus distal (left-sided) tumors (division at the splenic flexure), Loupakis et al, in an analysis of 3 independent studies where patients with mCRC received first-line chemotherapy with or without bevacizumab, found no association between tumor location and bevacizumab efficacy.8, 17

The present prospective observational study of consecutive mCRC patients treated in routine clinical practice sought to address the issue of potential interaction, if any, between primary tumor site and bevacizumab efficacy. We used data from TRACC (Treatment of Recurrent and Advanced Colorectal Cancer), a prospective multicenter Australian registry of consecutive patients with mCRC18 that utilizes a clinician-designed data set that includes key prognostic information such as Eastern Cooperative Oncology Group (ECOG) performance status (PS), comorbidities, and treatment intent. Using the traditional definition of left- versus right-sided mCRC (division at splenic flexure) with rectal tumors considered a separate entity, we assessed whether the location of the primary tumor was a predictor of PFS in patients receiving bevacizumab concurrently with first-line chemotherapy. Additionally, we conducted a similar analysis for a subgroup of tumors classified based on the alternative definition by Boisen et al16 of right and left colorectal tumors (right-sided, cecum to descending colon; left-sided, sigmoid and rectal tumors) to determine if the variation in definition could account for the differences reported.

Section snippets

Study Population and Description

The TRACC registry, a prospective multicenter registry enrolling consecutive patients with mCRC from centers across Australia, has been described previously.18 Key data points pertaining to patient and disease characteristics, treatment, and outcomes are collected in an electronic database at the point of care. For this analysis, all patients diagnosed between January 2009 and December 2014 who received first-line chemotherapy with palliative intent were included. Patients with multiple primary

Study Population

At the time of data analysis, 1604 patients with mCRC from 16 sites were registered within the TRACC database. Of these, 926 satisfied the inclusion criteria (297 had right-sided primary tumors, 354 had left-sided tumors, and 275 had rectal tumors). The study population's median age was 65.6 years (range, 26-92 years), and patients were followed for a median of 32.5 months.

Table 1 depicts baseline parameters. Patients with right-sided primary tumors were significantly older (median age, 68.6

Discussion

The findings of the present study concur with previous reports6, 7, 13 indicating that primary tumor location is a prognostic variable in mCRC, with right-sided tumors (proximal to the splenic flexure) associated with the least favorable survival outcomes and with the longest PFS and OS being seen in patients with rectal tumors. We also confirmed the role of baseline ECOG PS as an independent prognostic variable regardless of treatment administered, and that provision of bevacizumab alongside

Disclosure

The authors have stated that they have no conflict of interest.

Acknowledgments

Roche Products Pty Limited provided financial assistance for the development, installation, and maintenance of the TRACC registry. We thank all the participating centers of the TRACC registry who contributed to data collection and BioGrid Australia. We also thank Nalini Swaminathan for her writing assistance.

References (25)

  • S.D. Markowitz et al.

    Molecular origins of cancer: molecular basis of colorectal cancer

    N Engl J Med

    (2009)
  • J.A. Bufill

    Colorectal cancer: evidence for distinct genetic categories based on proximal or distal tumor location

    Ann Intern Med

    (1990)
  • Cited by (44)

    • Dichotomous colorectal cancer behaviour

      2023, Critical Reviews in Oncology/Hematology
    • Treatment sequencing of metastatic colorectal cancer based on primary tumor location

      2021, Seminars in Oncology
      Citation Excerpt :

      They considered the limited efficacy of bevacizumab in right-sided tumors may be attributable to a lower systemic inflammatory status and a higher expression of pro-angiogenic factors, both of which appear to characterize patients with right-sided tumors [18]. The Australian registry reported that PFS is superior for bevacizumab-treated patients in all groups but appeared greatest in right-sided colon disease [19]. However, results reported by Price et al did not suggest that the site of the primary tumor had any impact on PFS [20].

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

      2019, Clinical Colorectal Cancer
      Citation Excerpt :

      The authors suggested that the addition of bevacizumab to chemotherapy in first-line treatment may predominantly benefit patients with primary tumors originating in the rectum and sigmoid colon. Conflicting conclusions were reached by Wong et al,32 who analyzed the data of 926 patients from an Australian prospective multicenter mCRC registry treated with first-line chemotherapy ± bevacizumab. Patients who received anti-VEGF agent in addition to chemotherapy had superior outcomes, but the effect appeared greatest in patients with right-sided colon tumors.

    • Colonic Strictures

      2019, Clinical Gastrointestinal Endoscopy
    View all citing articles on Scopus

    The first 2 authors contributed equally to this article, and both should be considered first author.

    View full text