The association between renal sinus fat area and the progression-free survival in Chinese non-metastatic clear-cell renal cell carcinoma patients

In this retrospective study, we evaluated the association between renal sinus fat area (RSFA) and survival in 268 Chinese non-metastatic clear-cell renal cell carcinoma (ccRCC) patients. Patients with high RSFA exhibited better progression-free survival than those with low RSFA in both univariable (HR: 0.240; 95% CI: 0.119–0.482; p < 0.001) and multivariable (HR: 0.432; 95% CI: 0.369–2.749; p = 0.027) analyses. A propensity-score matched (PSM) analysis using Kaplan-Meier curves confirmed our findings (log-rank test; p = 0.028). Based on the multivariable analysis, we constructed a prognostic nomogram with 4 factors, namely, RSFA, Fuhrman grade, AJCC stage and sarcomatoid component. The c-index values for the Leibovich scoring system and the nomogram were 0.762 (95%CI, 0.688–0.835) and 0.823 (95%CI, 0.759–0.888), respectively. These findings demonstrate that high RSFA is associated with better progression-free survival in non-metastatic ccRCC.


INTRODUCTION
Renal cell carcinoma (RCC) represents 2-3% of all human malignances and the most common histological subtype is clear-cell renal cell carcinoma (ccRCC) [1][2]. Many post-operative clinicopathological features are associated with overall survival of RCC patients and several predictive score models have been established to improve the risk stratification [3][4][5]. Metabolic syndrome (MS) is one of the pre-operative parameters that have been recently reported that predict survival outcomes of RCC patients. Two major features of MS, namely, visceral obesity and hypertension have been established as etiological factors of RCC [6][7]. However, contradictory results have been reported regarding association between visceral fat area (VFA) and the prognosis of RCC [8][9][10][11][12][13]. Recently, renal sinus fat accumulation (RSFA) or ectopic adipose tissue deposits have been associated with many MS features including hypertension [14][15]. However, the role of RSFA in RCC is unknown. Therefore, we studied the association between renal sinus fat area (RSFA) and the progressionfree survival (PFS) of ccRCC. In addition, we constructed a nomogram to predict the PFS of non-metastatic ccRCC.

Baseline characteristics
We included 268 patients with pathologically confirmed ccRCC (126 located in the left kidney; 142 in

Research Paper
Oncotarget 65482 www.impactjournals.com/oncotarget the right) that underwent partial (n = 82) and radical (n = 186) nephrectomy in the present study. Patients with high RSFA values correlated with high SFA, VFA and BMI value and showed a high proportion of smaller tumor size, lower AJCC stage and lower fuhrman grade. Patients with low RSFA were more likely to have renal sinus invasion. We performed a one-to-one PSM analysis because the two study cohorts were not fully comparable. In the matched cohorts (n = 162; 81 pairs), the variables were fully comparable without statistical significance (Table 1). In addition, 43 patients were diagnosed as DM cases. The distribution of pre-operative HbA1c values (ranging from 5.8% to 10.5%) were as follows: 2 above 8%, 12 between 7-8%, 9 between 6.5-7% and 20 cases less than 6.5%. Further evaluation of the tumor grade was performed according to the WHO/ISUP grading system. There were 67, 128, 55 and 18 patients that were identified as grade 1, grade 2, grade 3 and grade 4, respectively. Patients with high RSFA values also correlated with lower WHO/ISUP grade than those with low RSFA values (p = 0.007). We identified 39, 71, 19 and 5 patients as grade 1, grade 2, grade 3 and grade 4, in the high RSFA group, respectively. We also identified 28, 57, 36 and 13 patients as grade 1, grade 2, grade 3 and grade 4 in the low RSFA group, respectively.

Survival outcomes
The median follow-up of the 268 patients was 38 months with 48 patients (16 with AJCC stage I, 15 with stage II, and 17 with stage III) experiencing progression (distant metastasis or local recurrence) after surgery. Patients with high RSFA showed longer progression-free survival (PFS) than those with low RSFA (p < 0.001; Figure 1A). The 3-year PFS was 71.3% and 91.8% for patients with low and high RSFA. In the univariable Cox analysis, high SFA, VFA, and RSFA values, lower Fuhrman and WHO/ISUP grades, lower AJCC stage, smaller tumor size, absence of sarcomatoid component and renal sinus invasion were associated with higher PFS. However, multivariable analysis demonstrated that high RSFA, lower Fuhrman and WHO/ISUP grades, lower AJCC stage and absence of sarcomatoid component were associated with high PFS (Tables 2 and 3).
Further, our analysis demonstrated that larger RSFA, lower Fuhrman and WHO/ISUP grades, lower AJCC stage and the presence of sarcomatoid component were independent predictive factors of the PFS of nonmetastatic ccRCC. Leibovich scoring system used Fuhrman grade as one of the predictors. Thus, we established a prognostic nomogram with factors such as Fuhrman grade, RSFA, AJCC stage and the presence of sarcomatoid component ( Figure 2). The calibration curves demonstrated good consistency in bootstrap analysis between the calculated and actual 3-year PFS ( Figure 3). The c-index values of Leibovich scoring system and our new nomogram were 0.762 (95%CI, 0.688-0.835) and 0.823 (95%CI, 0.759-0.888), respectively. In matched cohorts, 20 patients experienced tumor progression with the 3-year PFS values for patients with low and high RSFA being 83.5% and 93.0%, respectively. Further, survival analysis revealed that patients with larger RSFA value were associated with better PFS compared to patients with low RSFA (p = 0.028; Figure 1B).

DISCUSSION
Metabolic syndrome (MS) is primarily associated with development of cardiovarcular disease, hypertension and type 2 diabetes. Recently, its role in many cancers including RCC has been recognized. Among the many features of MS, obesity and hypertension have been identified as etiological factors worldwide [6][7]. Obesity is associated with the prevalence of a variety of cancers, including RCC (especially ccRCC). It is postulated that adipose tissue derived hormones in obesity regulate angiogenesis, epithelial-mesenchymal transition, and inflammation during tumor progression [16].
However, contradictory results have been reported regarding the association of visceral adipose tissue accumulation with RCC progression. Lee and colleagues reported that larger visceral fat area (greater than 50% in both sexes) was associated with longer survival of advanced RCC patients of Korean origin [11]. Likewise, Kaneko and colleagues reported similar findings in their study [10]. In contrast, Mano and colleagues recently reported that neither SFA nor VFA were associated with the survival of non-metastatic ccRCC in a western population [9]. Furthermore, Ladoire and colleagues suggested that high VFA was a predictive biomarker for lower survival in metastatic RCC patients that were given first-line antiangiogenic drugs [12]. In our study, high SFA and VFA values showed better PFS compared to corresponding low values in the univariable analysis. However, they were not associated according to multivariable analysis. Therefore, the role of VFA and SFA in the development and progression of ccRCC remains unclear and further studies with larger cohorts are necessary.
Epidemiologic studies have demonstrated that hypertension is one of the strongest risk factors for RCC in western populations [17][18]. A recent investigation in Chinese population demonstrated that hypertension played a significant role in the etiology of RCC [19]. Also, patients with higher renal sinus fat accumulation (measured as area or volume) were associated with increased risk of hypertension [14][15]. Therefore, we assessed the association between RSFA values and the PFS of ccRCC by evaluating the CT scans. In our study, patients with high RSFA showed better PFS than those with low RSFA in both univariable and multivariable Cox analysis. Further, we used the PSM analysis to confirm our findings because it reduces bias due to confounding Oncotarget 65483 www.impactjournals.com/oncotarget baseline factors [20]. PSM analysis also showed a strong relationship between high RSFA and better RCC survival. In addition, the nomogram including RSFA, AJCC stage, Fuhrman grade, sarcomatoid differentiation parameters demonstrated a more accurate prognosis for PFS than the Leibovich scoring system, which is the most commonly used predictive model to determine PFS of ccRCC.
The major branches of the renal artery and vein along with the major and minor calices of the collecting system, ureters and lymph vessels are located within the renal sinus. Therefore, we hypothesized that excess fat tissue deposits in the renal sinus may act as a physical barrier and prevent tumor cells from migrating to distant organs through veins and lymph vessels. Moreover, Zi and colleagues demonstrated that among perineoplasm, renal sinus, and adipose tissue conditioned media, only perineoplasm conditioned medium enhanced the migration of ccRCC cells (CaKi-2 cells) due to enhanced WNT signaling [21]. Therefore, we hypothesized that the association of RSFA with PFS of ccRCC might be due to physical factors and not due to biological factors. Among the 14 patients that showed pathological renal         Oncotarget 65488 www.impactjournals.com/oncotarget groups based on the median value. Further, since adipose tissue distribution is different between males and females, use of sex-specific median value was recommended [11]. In our study, we sub-divided the patient cohorts based on sex-specific median values. Also, we took the side of the kidney into consideration in regard to RSFA, tumor location and their association with PFS of ccRCC patients. However, we were limited by the lack of standard normal reference value of RSFA in regard to gender and race. Mano and colleagues reported the median values of sexadjusted VFA as 218.62 cm 2 and 156.49 cm 2 for male and female, respectively in a western population with nonmetastatic ccRCC [9]. However, the median values of VFA in our study were 123.11 cm 2 and 80.92 cm 2 for males and females, respectively. Lee and colleagues reported sexspecific median values of VFA as 117.07 cm 2 and 79.28 cm 2 for male and females, respectively, which were similar to our reported values [11] Foster and colleagues evaluated the RSFA of 92 participants with 49% women (100% right kidney) and reported a median RSFA value of 0.43 cm 2 [15]. However, we reported larger RSFA values (1.28 cm 2 for males and 0.59 cm 2 for females) than those reported by Foster and colleagues (0.43 cm 2 ).
The limitations of our investigation include the inherent bias due to the retrospective nature of our study, relative small sample size, and short follow-up periods. Because of a relatively shorter follow-up period and a small number of deaths in that period, we were unable to evaluate the association between RSFA and overall survival and cancer-specific survival. Furthermore, we only evaluated non-metastatic ccRCC. Therefore, predictive value of RSFA in other histological types of RCC or metastatic RCC is unknown. Also, the lack of clear-cut mechanistic details and standard RSFA cut-off values limit the utility of RSFA in clinical practice.
In conclusion, our study demonstrates that high RSFA is associated with increased progression-free survival of ccRCC in Chinese patients.

Patient enrollment
We retrospectively enrolled 268 patients that underwent nephrectomy at our center between December 2009 and December 2015 and were post-operatively diagnosed as non-metastatic ccRCC. This study was approved by the institutional ethics committee of The First Affiliated Hospital of Xiamen University. A waiver of written informed consent was granted by the ethics committee for this retrospective analysis. Patient records/ information was treated anonymously for this analysis.
Patients were excluded from the analyses if (1) they were operated in our center but had received CT scans in other hospitals before operation and therefore digital CT images were not available for analysis; (2) they did not undergo operation (partial or radical nephrectomy) as the main treatment; and (3) they had surgery elsewhere. All surgical procedures were performed by the director (JCX) of our center.

Clinicopathological parameters
We assessed clinicopathological factors such as gender, age at the diagnosis (> 60 y vs. < 60 y), body mass index (BMI, ≥ 25 kg/m 2 vs. < 25 kg/m 2 ), kidney side where tumors were located (right vs. left), nephrectomy type (partial vs. radical), VFA and subcutaneous fat area (SFA) at the level of the umbilicus, RSFA, pathological stage based on the 2010 American Joint Committee on Cancer (AJCC), tumor size (< 5 cm, 5-10 cm vs. > 10 cm), Fuhrman grading (grades 1-4), sarcomatoid component, pathological renal sinus invasion, hypertension and diabetes mellitus (DM). As the WHO/ISUP grading system has recently been proposed to replace the Fuhrman grading system, we further identified the grade of our cohort according to the International Society of Urological Pathology (WHO/ISUP) grading system.
Hypertension was defined as systolic blood pressure ≥ 140 mm Hg, diastolic blood pressure ≥ 90 mm Hg. Diabetes was defined as a fasting plasma glucose ≥ 126 mg/dL, 2-hour plasma glucose value ≥2 00 mg/dL after 75 g oral glucose tolerance test (OGTT). The pre-operative HbA1c data was also collected from DM patients. The tumor size of patients was classified based on SSIGN [3] and Leibovich [4] scores, namely, < 5 cm, 5-10 cm or > 10 cm. Progression-free survival was defined as the time from nephrectomy to distant metastasis or local recurrence.

Measurements of SFA, VFA and RSFA
Routine CT scans quantified the area of adipose tissue, including SFA, VFA and RSFA [8][9][10][11][12][13][14][15]. Abdominal CT scans (Somaris/7 CT 2012B, Siemens AG, Germany) were performed for all the enrolled patients before surgery. The contours of the abdominal muscular wall separating the visceral from the subcutaneous compartment at the level of umbilicus was manually outlined ( Figure 4A-4B). Then, using an image display window width of -195 to -45 Hounsfield units, pixels containing fat were identified. The VFA and SFA values were then calculated with Syngo multimodality workplace, version VE31A (Siemens AG, Germany). Figure 5 shows the quantification of RSFA on the side of the lesion as described previously [15].
Measurements of the adipose tissue area were performed by one radiologist blinded to the clinical and pathological data. The sex-specific median value (123.11 cm 2 for male; 80.92 for female) was used as the cut-off to distinguish patients with high and low VFA values [11]. Likewise, sex-specific median values of SFA (128.05 cm 2 for male; 184.81 for female) were used to determine patients with high and low VFA. Since RSFA is potentially www.impactjournals.com/oncotarget associated with gender and kidney side, the median sexand side-specific median values (males: 1.28 cm 2 and 0.91 cm 2 for right and left kidneys; females: 0.59 cm 2 and 0.46 cm 2 for right and left kidneys) were used to distinguish high and low renal sinus fat accumulation.

Statistical analysis
SPSS 22.0 (version 13; SPSS Inc., IL, USA) was used for all statistical analyses. We performed one to one propensity-score matched (PSM) analysis to reduce the effect of confounding factors and bias caused by different baseline distribution of demographic and clinicopathologic factors between patients with high or low RSFA.
Numerical variables were compared with t-test and categorical variables with chi-square test. Survival curves were plotted by the Kaplan-Meier method and assessed by the Log-rank test. Subsequently, Cox regression analysis was performed to identify potential prognostic factors for survival. Characteristics with p < 0.1 were further evaluated