Statistics from Altmetric.com
Evidence accumulated over the past decade has clearly established excess body fatness as a risk factor for colorectal cancer. Specifically, using body mass index (BMI) as an approximation of body fatness, risk increases with increasing BMI in a sex- (men>women) and site- (colon>rectum) specific manner.1 A number of plausible biological mechanisms may underpin these associations,2 including increased insulin resistance, increased availability of insulin-like growth factor (IGF)-I (mitogenic, pro-apoptotic, pro-angiogenic, increases cell motility), and altered adipokine metabolism; for example, increased leptin (mitogenic, anti-apoptotic, pro-angiogenic) and/or decreased adiponectin (anti-angiogenic, anti-inflammatory). Population-based cohort data report that increased BMI is also associated with an increased risk of colon cancer-related mortality suggesting that excess body weight impacts unfavourably on outcome in patients with established disease.3 However, the above study design is limited by lack of details on staging and treatment (key determinants of outcome), and hence, it remains unclear whether the increased mortality observed among obese individuals represents increased incidence, delayed diagnosis, differential primary treatment, differential selection for adjuvant therapies, increased risk from treatment-related complications and/or a differential response to treatment, compared with leaner counterparts. Addressing one possible mechanism of adverse treatment outcome, secondary analyses from at least two randomised trials of adjuvant 5-fluorouracil (5-FU)-based chemotherapies in patients with colon cancer demonstrate increased recurrence rates and mortality in patients with increased BMI values.4 5
The study from Guiu and colleagues6 reported in this issue of Gut (see page 341) adds new and interesting dimensions to the story of obesity and treatment outcome in colorectal cancer. Recognising that obesity is associated with increased circulating levels of vascular endothelial growth factor (VEGF), a key regulator of tumour angiogenesis and the main target for bevacizumab antibody therapy, the authors tested the hypothesis that excess fatness adversely impacts upon the oncological endpoints of treatment response, time-to-progression (TTP) and overall survival (OS) in a retrospective series of patients of 80 patients undergoing first-line treatment with combined bevacizumab and conventional chemotherapy for metastatic colorectal cancer (MCC). The study related these outcomes to BMI, subcutaneous fat area (SFA), and visceral fat area (VFA), the latter two parameters determined from computed tomography (CT) scan, and in their multivariate analyses, showed that high VFA was associated with poorer response, shorter TTP and reduced OS. There were some ‘signals’ of adverse outcomes when BMI was the anthropometric measure of interest, but none for SFA. In a comparative analysis of 40 patients with MCC undergoing first-line treatment with conventional chemotherapy, there were no relationships between body fatness and outcomes.
The authors recognised there were potential limitations to this study, including small sample size, retrospective study design, clinical heterogeneity between the two treatment groups, and the risk of model overfitting. Despite these shortcomings, this study generates two interesting hypotheses: the first clinical; the second mechanistic. The role of bevacizumab-based therapy (in combination with conventional chemotherapy) has undoubtedly been established in the options for first-line treatment of MCC, but until now, there has been no predictive ‘biomarker’ to direct treatment algorithms. This contrasts with other mechanism-based therapies in MMC, for example K-RAS status and Cetuximab therapy, where measurement of the predictive biomarker clearly discriminates for subsequent treatment response. While the concept of body fatness as a ‘lifestyle’ biomarker may seem unconventional (compared with the traditional gene or protein biomarker), there is a precedent from other cancer types where lifestyle is predictive for treatment response; for example, smoking status and the use of anti-epidermal growth factor receptor (EGFR) therapy (erlotinib) in non-small cell lung cancer.7 It is also noteworthy that (not unexpectedly) the mean BMI at commencement in the present study was relatively low (23 kg/m2), but body weights greater than this value seem to matter; in other words, the adverse effect of excess body fatness on bevacizumab-based therapy in MCC may be relevant at modestly increased and not just extreme adiposity ranges.
The second series of hypotheses (or questions) are mechanistic. What is the origin of the elevated VEGF environment in obesity? The observation that the strongest relationships with adverse treatment response are with VFA points to this adipose source rather subcutaneous fat. But there seems to be a paradox, as increased VEGF expression and angiogenesis are features of adipose tissue development, yet in fully developed fatty tissues there is a reduction in VEGF mRNA expression and decreased capillary vessel density, a process termed rarefaction, which may in part be related to obesity-related tissue hypoxia.8 Moreover, adipose tissue endothelial cells may not be the only source of observed elevated VEGF levels, as it is conceivable that adipocytes secrete pro-angiogenic cytokines and stimulants capable for increasing VEGF production at distant sites. On the other hand, the key mechanisms favouring poor outcome in the presence of excess body fatness might be the development of anti-VEGF therapy resistance, either as an intrinsic non-responsiveness or an evasive resistance.9 In the latter, induced angiogenic factors circumvent or substitute traditional pathways and re-establish vascularisation. Thus, for example, overweight and obesity are also associated with increased levels of serum angiopoietin-2, angiogenin and endostatin.10
Going forward, there is a need to replicate these observations: secondary analyses in bevacizumab-based and other anti-angiogenesis trials in MCC relating anthropometric measurements to outcomes offer an obvious setting in which to test this hypothesis. In this context, there is a need to detail dose specifications; for example, full weight-based dosing versus dose limits in obese patients. Stratification by men and women would seem an obvious a priori sub-group analysis as men consistently report higher risks per unit change in BMI for incident risk and mortality, but interestingly, if central obesity is the anthropometric parameter of interest, associations for men and women seem similar. At a mechanistic level, serial measurements of angiogenic factors during therapy will be informative.
With the increasing availability of effective combinational therapies in patients with MCC, we are moving closer towards ‘personalised’ oncology in this setting. Some of the ‘signposts’ along the way may be cutting-edge expensive technologies, such as gene profiling, while others may come from readily measured and inexpensively determined patient characteristics, such as measures of body fatness, ultimately striving towards a longer and better quality of life for our patients with colorectal cancer.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.