Review of the Interaction Between Body Composition and Clinical Outcomes in Metastatic Renal Cell Cancer Treated With Targeted Therapies

Treatment of metastatic renal cell cancer (mRCC) currently focuses on inhibition of the vascular endothelial growth factor pathway and the mammalian target of rapamycin (mTOR) pathway. Obesity confers a higher risk of RCC. However, the influence of obesity on clinical outcomes in mRCC in the era of targeted therapy is less clear. This review focuses on the impact of body composition on targeted therapy outcomes in mRCC. The International Metastatic Renal Cell Carcinoma Database Consortium database has the largest series of patients evaluating the impact of body mass index (BMI) on outcomes in mRCC patients treated with targeted therapy. Overall survival was significantly improved in overweight patients (BMI ≥ 25 kg/m2), and this observation was externally validated in patients who participated in Pfizer trials. In contrast, sarcopenia is consistently associated with increased toxicity to inhibitors of angiogenesis and mTOR. Strengthening patients with mRCC and sarcopenia, through a structured exercise program and dietary intervention, may improve outcomes in mRCC treated with targeted therapies. At the same time, the paradox of obesity being a risk factor for RCC while offering a better overall survival in response to targeted therapy needs to be further evaluated.


Introduction
Treatment of metastatic renal cell cancer (mRCC) currently focuses on inhibition of the vascular endothelial growth factor (VEGF) pathway and the mammalian target of rapamycin (mTOR) pathway. Although predictive biomarkers for targeted therapy Obesity confers a higher risk of RCC (4)(5)(6)(7); however, the influence of obesity on clinical outcomes in mRCC in the era of targeted therapy is less clear. The World Health Organization utilizes body mass index (BMI; weight divided by height squared) to define the terms "overweight" (BMI 25.0-29.9 kg/m 2 ) and "obesity" (BMI ≥ 30 kg/m 2 ) (8). However, BMI does not accurately reflect body composition, the proportion of lean tissue to fat; nor does BMI account for sarcopenia, the loss of skeletal muscle tissue. Computed tomography (CT) is often used to assess response to therapy as part of routine care. Cross-sectional imaging can be utilized to quantify skeletal muscle density (SMD) and adipose tissue. Higher BMI may negatively influence outcomes through commonly associated comorbidities of diabetes and cardiovascular disease (9). It may alter drug concentrations and pharmacokinetics of targeted therapies that are dosed independent of weight. Obesity may activate oncogenic pathways and create an inflammatory state. This is postulated to occur via elevations in interleukins (IL-6, IL-1β, and IL-1 receptor antagonist), tumor necrosis factor, and Creactive protein (10). Furthermore, a proangiogenic state is created by the production of factors such as VEGF and leptin by adipose tissue (11). An obese body composition also can promote and activate the mTOR pathway through reactive oxygen species (12), as well as elevated levels of insulin and insulin-like growth factor (13).
Obesity is paradoxically associated with better prognosis, particularly in the setting of nephrectomized patients with RCC (14,15). Table 2 summarizes the findings of studies, which examine this relationship between survival rates in mRCC and body composition metrics. This review paper will focus on the impact of body composition on targeted therapy outcomes in mRCC.

Literature search strategy
A PubMed and Medline literature search was performed for the time period 1994 to 2015 with the following search terms: sarcopen*, BMI, body mass, cachexia, BSA, body surface area, body composition, renal cell ca*, RCC, kidney cancer, prognos*, outcome*, response, predict*, mTOR, everolimus, sirolimus, sunitinib, PD1, PDL1. Additionally, American Society of Clinical Oncology meeting proceedings were searched with the following search terms: BMI, body, BSA, renal cell. Articles in any language were included, and all levels of evidence were considered. The retrieved articles' relevant references were also reviewed for possible inclusion. Eleven articles (eight published, three abstracts) evaluated body composition as a prognostic factor for targeted therapy outcomes in mRCC.

Impact of BMI on outcomes
Body composition and its potential influence on targeted therapy outcomes were initially assessed in a retrospective study of 475 mRCC patients treated with antiangiogenic therapy. Choueiri et al. (16) identified that obesity (BMI > 30 kg/m 2 ) was independently associated with greater overall survival (OS) ( There was question of whether tolerability of therapy in higher BMI patients played a role in producing these findings. However, in the IMDC dataset, rates of early discontinuation due to adverse events did not differ between the two BMI groups, and therefore this was unlikely a cause of bias (17). Additionally, the toxicity patterns were similar in the high-and low-BMI groups in the external Pfizer validation set (18).
The biologic rationale for the association between BMI and outcomes is not clear. Fatty acid synthase (FASN) is a key enzyme involved in the production of fatty acids. FASN has emerged as a metabolic oncogene with an important role in tumor growth and survival (19). There was a trend to improved OS in the elevated BMI group (p = 0.07) in the Cancer Genome Atlas clear cell mRCC dataset (n = 61) (18). High BMI was associated with low FASN gene expression (p = 0.034), and FASN expression (using the median as a cutpoint) was inversely associated with OS (p = 0.002). FASN gene was evaluated using immunohistochemistry (IHC) in the IMDC biospecimen cohort (17).
Median OS was significantly improved in FASN IHC negative compared with positive patients (27.5 vs 14.5 months, p = 0.005). FASN metabolism may contribute to the development of mRCC and therefore may represent a novel therapeutic target. These results are hypothesis-generating.

Volumetric assessment of body composition
BMI is a relatively crude measurement of body composition. Other groups have evaluated more granular metrics of body composition such as volumetric assessment of fat and muscle ( Table 3). Software programs can be used to identify the visceral and subcutaneous visceral adipose compartments using Houndsfield units at specific landmarks on CT, followed by calculation of the cross-sectional area. In contrast to the IMDC data, BMI was not prognostic in other smaller retrospective studies that incorporate these volumetric assessments (20)(21)(22)

The impact of sarcopenia on adverse events from targeted therapy
Sarcopenia was evaluated in a subset of mRCC patients from the TARGET trial (sorafenib vs placebo after progression on standard therapy) (24). Sarcopenia was present in 72% of patients with a BMI < 25 and 34% of obese patients. Treatment with sorafenib was associated with a significant decrease in skeletal muscle in comparison with placebo (8.0% loss, p < 0.01). Preclinical models suggest that the skeletal muscle loss associated with sorafenib may be mediated by downstream effects of mTOR inhibition (25). Frequency of sorafenib-induced doselimiting toxicities was highest in sarcopenic patients whose BMI < 25 kg/m 2 , and lowest in nonsarcopenic patients who were overweight or obese (p = 0.03) (24). These results suggest that sarcopenia in mRCC is a predictor of sorafenib-induced toxicity. Since sorafenib promotes muscle loss severe adverse events may be more frequent in sarcopenic patients. A future area of research would be to individualize the dose of a targeted therapy based on a patient's skeletal muscle mass, in order to decrease dose-limiting toxicities and optimize clinical outcomes (23).
A similar interaction between sarcopenia and toxicity was observed in a retrospective analysis of 112 mRCC patients treated with mTOR inhibitors, immunotherapy, VEGF inhibitors, Tyrosine Kinase Inhibitors (TKIs), and best supportive care (26). The prevalence of sarcopenia was 20.5% at baseline and increased to 38.4% at the end of the evaluation. Sarcopenia was independently associated with increased frequency of severe (common toxicity criteria grade > 2) treatment toxicity (Pearson chi-square value 12.82; p = 0.001).
Sunitinib was permanently discontinued during the first cycle in 30% of sarcopenic patients, as compared with 2.4% of the remaining patients (p = 0.01). On multivariate analysis, the combination of sarcopenia and BMI<25 kg/m 2 was the only independent predictor of early DLTs (p = 0.04). However, the presence of sarcopenia had no significant impact on OS (p = 0.75) and PFS (p = 0.071).
Cushen et al. (28) investigated the impact of fat-free mass and skeletal muscle mass (SMM; metabolic tissues such as the liver and kidney, intracellular and extracellular water, and bone) on DLTs in mRCC patients treated with sunitinib. Sarcopenia was present in 33% of the patients (18/55). DLTs were inversely associated with SMM; 92% of the patients with SMM < 25th percentile experienced DLTs, in contrast to 57% of those patients with SMM > 75th percentile (p = 0.05). Patients with low fatfree mass (n = 4) experienced significantly more DLTs than those with high fat-free mass (n = 2, p = 0.002), but it is unclear what cutoff was used to determine the differentiation between these two groups.

Discussion
The IMDC database has the largest series of patients evaluating the impact of BMI on outcomes in mRCC patients treated with targeted therapy. OS was significantly improved in overweight patients (BMI ≥ 25 kg/m 2 ), and this observation was externally validated in patients who participated in Pfizer trials (17,18). FASN gene is an emerging oncogene in mRCC, and high BMI may be a surrogate for low FASN levels. If this finding is externally validated prospectively, future studies should optimize the FASN assay and determine whether inhibition of this pathway has the potential to improve outcomes for mRCC.
CT may provide a more refined description of body composition than relatively crude measurements such as BMI. The impact of adipose tissue on mRCC outcomes is unclear. High VFA and SFA were associated with improved OS in studies performed by Steffens et al. (20) and Gu et al. (22). Conversely, increased VFA was associated with worse outcomes were in Ladoire's cohort (21). Antoun et al. (23) found no association between VFA or SFA and outcomes. Sarcopenia was associated with worse outcomes by Antoun et al. (23), but not in the cohort evaluated by Gu et al. (22). These four cohorts did not observe a significant association of BMI with outcomes, but were all smaller than the IMDC cohort (17). These investigators used different combinations of imaging software programs and anatomical landmarks. Further refinement of this technology is required, and the cutpoints for sarcopenia, VFA, and SFA require further validation in a larger cohort of patients. These studies also focused on baseline measurements of BMI, and SFA and VFA (17,18,(20)(21)(22)(23). Further studies evaluating the paradox of obesity being a risk factor as wells a prognostic marker of response to targeted therapy are needed.
Toxicity from targeted therapy appears to be independent of BMI (17,18). Retrospective series evaluating sarcopenia consistently demonstrate a relationship between sarcopenia and increased toxicity from targeted agents (24,(26)(27)(28). These studies were small, and the ideal method for evaluating sarcopenia as well as the optimum cutoff point has yet to be established.
What is the impact of sarcopenia on mRCC patients who are treated with targeted therapy?
Sarcopenia is consistently associated with increased toxicity to inhibitors of angiogenesis and mTOR. A structured exercise program and dietary intervention may strengthen patients with mRCC and improve response to targeted therapy. Prospective studies evaluating the impact of diet and exercise on targeted therapy tolerance, quality of life, and body composition are warranted.

Future
longitudinal studies should evaluate the prognostic impact of sarcopenia, VAT, SFA, BMI, and BSA in mRCC patients treated with targeted therapy. This would require controlling for other prognostic variables in a very large patient cohort to fully address this issue. This may facilitate the development of more refined prognostic models of mRCC treated with targeted therapy.