Prognostic Significance of Adipose Tissue Distribution and Metabolic Activity in PET/CT in Patients with Metastatic Colorectal Cancer

In this study, we aimed to evaluate the prognostic significance of adipose tissue distribution and metabolic activity in PET/CT to predict survival in patients with metastatic colorectal cancer (mCRC). The volume, density (HU), and FDG uptake (standardized uptake value (SUV)) of visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) and maximum FDG uptake of the tumor tissue were measured. Subcutaneous adipose tissue of volume-to-density ratio (SAT ratio) was calculated. The median OS for the patients with SAT ratio value <  −1.1 and ≥  −1.1 were 38.5 (95% CI 31.54–45.58) and 24.5 (95% CI 14.13–34.93) months, respectively (p = 0.05). During follow-up, 69 patients experienced disease progression. The median progression-free survival (PFS) was 11.03 months (95% CI: 9.11–12.95). Median PFS for patients with tumor SUV max value < 11.5 and ≥ 11.5 were 9.2 (95% CI 7.25–11.27) and 12.6 (95% CI 10.02–15.27) months, respectively (p = 0.14). Forty-eight patients received bevacizumab therapy. VAT SUV mean (HR: 0.09; 95% CI 0.01–0.52, p = 0.008) was significantly associated with PFS in patients receiving bevacizumab. SAT ratio was the significant parameter for the OS (HR: 0.58; 95% CI 0.33–1.01, p = 0.05) and PFS (HR: 1.99; 95% CI 1.02–3.91, p = 0.043). SAT ratio was an independent prognostic factor for survival in patients with mCRC. Higher SAT volume is correlated with longer survival in mCRC patients.

prognostic factors is not clear [3]. Potential mechanisms that have been shown to be effective in clinical studies are adipose tissue dysfunction, chronic inflammation, and insulin resistance [4].
Besides being a lipid and passive energy storage of adipose tissue; it is also known to be a highly active metabolic and endocrine organ. Although the majority of adipose tissue consists of adipocytes, there are also different cell types including fibroblasts, endothelial cells, macrophages, and pluripotent stem cells. Various factors are secreted from adipocytes and other cells. While some of the factors have a pro-inflammatory effect, others have anti-inflammatory, antitumoral, protective effects against diabetes and cardiovascular diseases [5].
Depending on the distribution of adipose tissue, it is classified as visceral or subcutaneous adipose tissue [6]. In addition to anatomical differences in SAT and VAT distribution, there are also differences in cellular, physiological, and clinical aspects. Absorption and accumulation of free fatty acids and triglyceride circulating in the subcutaneous adipose tissue are more [7]. Visceral fat tissue has a higher metabolic activity. Lipolytic activity is increased; free fatty acid production is high. The number of inflammatory cells in visceral adipose tissue is higher than in subcutaneous adipose tissue. Increased production of pro-inflammatory cytokines such as tumor necrosis factoralpha (TNF-a) and interleukin-6 (IL-6) is associated with the exaggerated inflammatory state [4,7]. Insulin sensitivity is less in visceral adipose tissue. Insulin resistance, hyperglycemia, and the formation of hyperinsulinemia are associated with an increased risk of diabetes [6,7].
Obesity, especially visceral fat, causes adipose tissue dysfunction. The secretion of adipokines becomes erratic. While the secretion of pro-apoptotic, anti-inflammatory adipokine such as adiponectin decreases the secretion of proinflammatory cytokines such as TNF-a and IL-6 increases [8,9]. Increase of pro-inflammatory cytokines together with adipose tissue hypoxia caused by obesity triggers chronic inflammation [10]. Obesity itself and adipokines secreted from dysfunctional adipose tissue also cause insulin resistance [11]. All these changes are thought to play a role in the development of carcinogenesis and metastasis due to obesity [12].
BMI is commonly used to define obesity. However, BMI, which fat and muscle mass contribute together, is insufficient to show total body fat and distribution. It does not fully reflect the fat tissue activity [13]. Although FDG uptake in PET/CT was generally used for tumor and metastasis evaluation in previous studies, it has also been used for the evaluation of adipose tissue in various studies recently. FDG uptake, density, and volume of adipose tissue were measured in PET/CT in order to evaluate adipose tissue in malignant diseases and determine its relation to tumor development and prognosis [14][15][16].
Few studies have evaluated the relationship between adipose tissue volume and activity with prognosis in patients with colorectal cancer [15,[17][18][19]. However, the prognostic value of adipose tissue in colorectal cancer is still uncertain. In this study, it was aimed to determine the distribution and activity of adipose tissue from PET/CT images of patients followed up for metastatic colorectal cancer and to evaluate the relationship between the obtained data and survival.

Patients
Permission for this study was obtained from the Pharmaceuticals and Non-Medical Device Research Ethics Committee of Necmettin Erbakan University.
The medical records of 250 patients, who were followed up at the Medical Oncology Clinic of our university with the diagnosis of metastatic colorectal cancer between January 2010 and December 2018 and who underwent PET/CT, were analyzed retrospectively. Eighty-three of these patients were included in the study. Patients who did not have PET/ CT imaging at the time of the diagnosis of metastasis, had secondary malignancies, and had an incomplete file, and laboratory information was excluded from the study.

F-FDG PET/CT
All patients fasted for at least 6 h prior to 18F-FDG (FDG) injection. FDG injection at a dose of 0.140 mCi/kg was administered to patients with blood glucose level < 200 mg/ dL. After the injection, the patient was kept in the room inside the PET/CT unit for about 1 h for the bio-distribution of FDG. Later, imaging was performed with Siemens Biograph 6 TruePoint PET/CT (Siemens, Germany). The body region between the vertex and proximal thigh was taken as the FDG PET/CT imaging area.
First, non-contrast CT imaging (130 kV, 50 mA, 5-mm slice thickness) was performed. Subsequently, PET imaging was performed in 3D mode, with an imaging area of 7-8 beds for each patient and 3 min of scanning time for each bed. PET images were reconstructed with the Iterative method.

FDG PET/CT Image Analysis
PET/CT images were retrospectively evaluated using TrueD software on the Multimodality Workplace (Syngo MMWP, Version VE25A, Siemens, Germany) workstation.
Extraperitoneal adipose tissue between subcutaneous and muscles was defined as subcutaneous adipose tissue (SAT); intraabdominal adipose tissue was defined as visceral adipose tissue (VAT).
The area of interest (ROI) was drawn for SAT and VAT at the level of the L4 vertebra in computed tomography (CT) and applied to three consecutive sections. A range of −200 to −50 Hounsfield units (HU) was chosen for the determination of adipose tissue (AT) in areas of interest (ROIs). SAT volume (SAT cm 3 ), SAT average density (SAT HUavg), VAT volume (VAT cm 3 ), and VAT average density (VAT HUavg) were calculated for adipose tissue (AT) within SAT and VAT areas of interest (ROIs) (Fig. 1).
Standardized uptake value (SUV) for metabolic activity was calculated based on injected FDG dose and body weight.
Areas of interest (ROIs) drawn on CT for SAT SUV mean and VAT SUV mean metabolic activity measurements were automatically inserted into PET/CT fusion images. For precise metabolic activity measurement, regions of interest (ROIs) were carefully evaluated and detected ones were manually removed in order to avoid high FDG uptake from adjacent tissues (e.g., intestine, vascular, urinary tract, muscle).
SAT ratio value was obtained by the proportion of the calculated volume and density values for subcutaneous adipose tissue. The highest SUV value in the tumor tissue (tumor SUV max) was measured.  SPSS 22.0 package program was used in the computer environment for data entry and statistical analysis. Ratio and median values were calculated for descriptive data. A p value of < 0.05 was considered statistically significant. Cox regression analysis was used to evaluate the survival effects of clinical and PET/CT data. The Kaplan-Meier method was used to determine survival times.

Results
Data of 83 patients were analyzed in the study. Of the patients in the group, 51 (61.5%) were male and 32 (38.5%) were female. The median age of the patients at the time of diagnosis was 61 (21-83), and their body mass index (BMI) was 26.4 (13.8-39.1). The primary tumor site was in the right colon in 13 patients (15.6%) and in the left colon in 70 patients (84.4%). After the diagnosis of metastatic disease, there were 48 (57.83%) patients receiving bevacizumab therapy and 8 (9.63%) patients receiving EGFR-targeted therapy. Chemotherapy regimens without targeted agents were applied to 27 (32.53%) patients. The clinical characteristics of the patients are shown in Table 1.

Discussion
In this study, the relationship between adipose tissue parameters in PET/CT and survival in mCRC patients was evaluated. A significant relationship was found between SAT ratio and overall survival and progression-free survival. In addition, in the subgroup of patients who received bevacizumab therapy, a higher VAT SUV mean value was better in terms of progression-free survival.
Fat tissue is divided into two regions as visceral and subcutaneous fat tissue according to its distribution. Apart from the anatomical location, these two adipose tissues differ in terms of endocrine function, lipolytic activity, and molecular aspects. Pro-inflammatory cytokine production and inflammatory cell number are higher in visceral adipose tissue [6,7,20]. We measured the volume and density values in PET/CT to evaluate the prognostic effect of the differences between VAT and SAT in colorectal cancer. Previous studies on various cancer types have found different results in the effect of VAT and SAT volume on survival. In studies conducted on patients with head and neck cancer and non-small cell lung cancer, those with high BMI have been shown to have better survival [16,21]. In this condition, which is called the obesity paradox, VAT and SAT volumes were measured to illuminate the protective effect of obesity. While better survival was observed in patients with high VAT volume in head and neck cancer, better survival was observed in patients with high SAT volume in non-small cell lung cancer. A similar relationship between high SAT volume and better survival has been demonstrated in renal cell carcinoma and prostate cancer studies [22,23]. In these studies, it was thought that high SAT volume can prevent cancer cachexia against lipolysis and increased energy consumption as the protective mechanism of adipose tissue.
In our study, no relationship was found between VAT and SAT volume and survival. However, the ratio obtained by dividing the volume of SAT by density showed a significant correlation with overall survival in all patients, and  progression-free survival and overall survival in patients receiving bevacizumab therapy. At this ratio, higher SAT volume was associated with better survival. Due to adipose tissue dysfunction, the secretion of proinflammatory cytokines and chemokines increases and macrophage infiltration occurs [24]. As a result of the changes, the adipose tissue microenvironment becomes suitable for cancer development and tumor growth. In this environment, with the effect of cancer cells, adipocytes lose their lipid content and differentiate into fibroblast-like cells [25]. These adipocytes associated with differentiating cancer can increase tumor progression and metastasis. Fat tissue density on CT images was used to evaluate these changes in the adipose tissue environment. In a study conducted on non-human primates, it was found that high adipose tissue density was associated with adipocytes with lower lipid content and fibrotic change [26]. In previous studies conducted with pancreatic cancer, head and neck cancer, and stomach cancer, it has been shown that high-fat tissue density is worse in terms of prognosis [14,21,27]. Unlike the results of these studies conducted on various cancer types, in our study, no significant association was found between SAT and VAT density values and survival. However, it was shown that higher SAT density at SAT ratio may be associated with better survival.
FDG uptake of adipose tissue reflects glucose metabolism and inflammatory status. Conflicting results have been found in previous studies evaluating the relationship between FDG uptake of VAT and SAT with clinical results. In studies conducted with colon cancer, pancreatic, and stomach cancer, it has been shown that high FDG uptake in VAT is worse in terms of survival [14,15,27]. It has been suggested that high FDG intake in these studies is associated with the inflammatory response that develops due to the increase in the expression of pro-inflammatory cytokines and macrophage infiltration in visceral adipose tissue. Studies on head and neck cancer and lung cancer did not show a significant relationship between fat tissue FDG uptake and survival [16,21]. Apart from these, two studies have shown that low FDG uptake in adipose tissue is worse in terms of prognosis. One of the studies was conducted on pancreatic cancer patients, and it was found that SAT FDG uptake is decreased in advanced stage and tumors with high FDG uptake [28]. As an explanation, it has been shown that inhibitors released from tumor cells reduce fatty acid uptake in adipose tissue and inhibit lipoprotein catabolism. The other study is evaluating the relationship between the presence of colorectal adenoma and adipose tissue metabolism [29]. In this study, it was shown that the probability of having a colorectal adenoma was higher in those with low FDG uptake on VAT, and it was stated that this situation was caused by insulin resistance due to obesity. Decreased glucose metabolism in insulin-resistant adipocytes and the increase in this insulin-resistant adipocyte mass and low FDG uptake in VAT were explained. In our study, a significant association was found between VAT SUV  and progression-free survival in patients receiving bevacizumab therapy. Survival was better in patients with high VAT SUV. Previous studies have been shown that VEGF release is higher in visceral adipose tissue [30,31]. It has also been found that increased VEGF release is associated with pro-inflammatory cytokines and inflammation [32,33]. The high VAT SUV mean may be associated with increased VEGF release. This may affect the response of bevacizumab therapy. PET/CT is one of the imaging methods used for diagnosis, staging, and follow-up of treatment response of colorectal cancer. The standard uptake value (SUV) is the parameter used to evaluate the degree of FDG uptake in PET/CT. Tumor SUV value is thought to be related to tumor aggressiveness and prognosis [34]. Previous studies on colon cancer, non-small cell lung cancer, and nasopharyngeal cancer have shown that tumor SUV max value is associated with prognosis and a higher SUV max value is associated with poorer survival [34][35][36]. In our study, PFS was found to be better in patients with high tumor SUV max value in the survival analysis performed in all patients and the group receiving bevacizumab therapy. Perhaps anti-VEGF therapy may be more effective in these patients.
Our study has some limitations. The study was conducted retrospectively on patients who were followed up in a single center. In addition, there were limitations in the comparisons due to the lack of an established method for the evaluation of adipose tissue metabolism and volume measurement.

Conclusıon
SAT ratio obtained by the ratio of SAT volume and SAT density values were found to be an important prognostic factor for PFS and OS in patients with metastatic colorectal cancer. High SAT volume was associated with longer survival in mCRC patients. In patients who received bevacizumab therapy, the VAT SUV mean value was associated with PFS. Progression-free survival was better in those with high VAT SUV mean than those with low. Further studies are needed to confirm the results of the study, as well as to explain the links between colorectal cancer and obesity.

Competing Interests
The authors declare no competing interests.