Preoperative PET/CT 18F-FDG Standardized Uptake by Lymph Nodes as a Significant Prognostic Factor in Patients with Colorectal Cancer

Purpose We evaluated the prognostic value of preoperative 18F-FDG uptake by suspected lymph nodes (LNs) using 18F-FDG PET/CT in colorectal cancer patients. Methods Patients with CRC underwent 18F-FDG PET/CT before radical surgery. We used Cox proportional hazards regression to examine the relationship between recurrence and the 18F-FDG maximum standardized uptake value (SUVmax) in the suspected LNs (SUVLN) on 18F-FDG PET/CT. Results Clinical data, treatment modalities, and results from 90 CR C patients were reviewed. The median follow-up was 19 months (range 3 to 72 months). Receiver operating characteristic analysis identified SUVLN 1.15 was the optimal cut-off value for predicting recurrence. SUVLN correlated with tumour size (P=0.045), lymph node metastasis (P=0.03), and recurrence (P < 0.0001). Univariate analysis showed significant associations between recurrence and SUVLN (P=0.017), and tumour grade (P=0.013). Multivariate analysis identified SUVLN (P < 0.0001), and tumour grade (P=0.005) as independent risk factors for recurrence. Patients with SUVLN ≤ 1.15 and SUVLN > 0.15 differed significantly in terms of recurrence (P < 0.0001). Conclusion Preoperative SUVLN measured by 18F-FDG PET/CT was significantly associated with recurrence and had significant prognostic value for recurrence-free survival in patients with colorectal cancer.


Introduction
Colorectal cancer (CRC) is one of the most common malignant tumours worldwide [1,2]. Surgery and chemotherapy are the main two strategies for its treatment. However, treatment outcomes for CRC remain unsatisfactory, because of recurrence and metastasis, particularly in patients with advanced CRC [3][4][5][6]. e TNM classification has been widely used to estimate prognosis and to decide on treatment of malignant tumours [7]. Lymph node (LN) metastasis is one of the most important prognostic factors in CRC because the survival rate among CRC patients with lymph node metastasis was significantly lower the rate among those without lymph node metastasis [8][9][10][11]. Traditional imaging methods play an important role in detecting lymph node metastases of malignant tumours [12,13]. However, these methods only reflect the size, density, and morphology of the lymph nodes; the biologic activity and aggressiveness of lymph nodes cannot be determined by traditional imaging methods. us, alternative imaging methods that better reflect the biologic behaviour of lymph nodes in CRC are of great importance.
Fluorine 18 ( 18 F) fluorodeoxyglucose (FDG)-combined positron emission tomography (PET) and computed tomography (CT) (PET/CT) is based on the abnormally high rate of glucose metabolism found in cancer cells. It is widely used in diagnostic imaging of many malignant tumours [14][15][16]. Although previous studies showed that 18 F-FDG PET had relatively low sensitivity in the assessment of nodal status in early-stage CRC, it showed better performance than conventional imaging methods, including CT [17,18]. 18 F-FDG PET/CT was recommended as a routine procedure in the management of CRC. However, most studies have focused on the relationship between the metabolic activity of the primary tumour and prognosis in CRC [19][20][21]. e metabolic activity of lymph nodes, measured as 18 F-FDG uptake, possibly better reflecting the biologic behaviour or aggressiveness in CRC, has been rarely evaluated. We hypothesized that the metabolic activity of lymph nodes in patients with CRC may reflect the biologic behaviour or aggressiveness of the primary tumour and may have prognostic importance.
We investigated the relationship between 18 F-FDG uptake by lymph nodes and clinicopathological characteristics in patients with CRC, and evaluated the significance of 18 F-FDG uptake by lymph nodes for predicting recurrence in CRC patients.

Study Population.
We included 90 patients (56 men and 34 women; age range, 43-87 y) with colorectal cancer. All had undergone 18 F-FDG PET/CT before radical resection between 2011 and 2017. Patients were included when they met the following criteria: they had been treated by radical resection of colorectal cancer with lymphadenectomy; the diagnosis of colorectal cancer had been confirmed by histopathologic examination; complete case records, including data on age, sex, tumour location, tumour size, T-stage, lymph node metastasis, lymphovascular invasion, tumour grade, and adjuvant treatment, were available. e study was approved by the institutional review board of the Shanghai Jiaotong University-affiliated Ren Ji Hospital and was in accordance with the 2013 revision of the Declaration of Helsinki. Informed consent was waived in this study. 18 F-FDG PET/CT was performed using a whole-body scanner (Biograph mCT; Siemens Medical Systems). All patients received an intravenous 3.7 MBq/kg injection of 18 F-FDG after having fasted for at least 6 h and rested for 1 h. e mean uptake time was 50 ± 6 min. Blood glucose levels were measured and were found to be less than 140 mg/dL at the time 18 F-FDG was administered. e CT component of the scan was performed without contrast administration at 120 kV, 140 mA, and a section thickness of 5.0 mm to match the thickness of the PET images. PET image datasets were reconstructed iteratively, with the CT data applied for attenuation correction.

Image Analysis.
For quantitative analysis, irregular regions of interest were placed over the most intense area of 18 F-FDG uptake. SUVmax was calculated as (maximum pixel value with the decay-corrected region-of-interest activity (MBq/mL))/(injected dose (MBq)/body weight (kg)). SUV LN was defined as the maximum standardized uptake value of suspected lymph nodes. As for the SUV LN , a single lymph node with the most avid FDG uptake was chosen for analysis and blood pool activity was used as the representative value of SUV LN, if there were no visible lymph nodes on CT scan and there was no avid 18 F-FDG uptake on PET [22]. e PET/CT images were evaluated by two experienced nuclear medicine physicians.

Lymph Node Dissection and Histologic Evaluation.
All patients were treated with radical surgery and lymphadenectomy, according to tumour location. Primary tumour and each lymph node were sliced and stained with haematoxylin and eosin and examined microscopically by a pathologist. e numbers of lymph nodes retrieved in each area and the presence or absence of metastases were recorded.

Clinical Endpoints and Follow-Up.
After surgical resection, all patients underwent clinical follow-up that included diagnostic imaging methods and blood tests after surgical resection. During follow-up, clinical assessment, including serum CEA, CA199, and CA724 levels, was performed every 2-3 months. Enteroscopy, contrast-enhanced CT, or MRI scans were performed every 6-8 months. Median follow-up was 19 months (range, 3-72 months). In addition, 18 F-FDG PET/CT was performed if the clinical assessment or studies performed during follow-up showed an abnormal finding. Recurrent tumour and distant metastasis were diagnosed based on either a positive biopsy or unequivocal clinical or radiographic evidence of progression. e time to recurrence was defined as the time from the date of surgery to the date of recurrence.

Statistical Analysis.
e relationship between the clinicopathological characteristics and recurrence were analysed by the χ 2 test, the unpaired two-tailed test and the Mann-Whitney U test, where applicable. e Kaplan-Meier method with the log-rank test was used to explore the relationship between SUV LN and recurrence. e Cox proportional hazard model was used to evaluate prognostic variables. Receiver operating characteristic (ROC) curve analysis was performed to determine the cut-off values for predicting recurrence. e receiver operating characteristic curve was used to assess the optimal threshold of SUV LN with which to predict recurrence. Pearson correlation coefficient was used to measure the correlations between SUV LN and clinicopathological characteristics. e data are represented as mean ± standard deviation. P < 0.05 was indicated significant difference. All statistical analyses were performed with SPSS software (SPSS, version 13.0; SPSS, Chicago, Ill).

Patient Characteristics.
e clinicopathological characteristics of the 90 enrolled patients are shown in Table 1.
Among the 19 patients with recurrence, 17 (89.5%) experienced recurrence within the first two years. e most frequent site of recurrence was distant metastasis (n � 9, 47.4%), followed by locoregional recurrence (n � 6, 31.6%), and peritoneal recurrence (n � 4, 21.0%). e most common site of distant metastasis was liver (n � 4), followed by the lung (n � 3), distant lymph node (n � 1), and bone (n � 1). Table 2 depicts the patient characteristics and 18 F-FDG PET/CT imaging grouped by CRC patients with or without recurrence. No significant differences between these groups were found in terms of age, sex, tumour location, tumour size, lymph node metastasis, lymphovascular invasion, adjuvant treatment, or SUV Tumour . However, a significant difference in SUV LN , T-stage, and tumour grade was found between these groups. Table 3 summarizes the relationship between SUV LN and site of recurrence in 19 patients with recurrence. In the 19 patients with recurrence, the Kruskal-Wallis test showed that there were no significant differences in SUV LN between nine patients with distant metastasis (median, 3.42; range, 0.7-10.2), four patients with peritoneal recurrence (median, 3.0; range, 2.0-4.0), and six patients with locoregional recurrence (median, 3.59; range, 0.7-15.2; P � 0.841).

Measurement of SUV LN Cut-Off
Value. ROC analysis identified a cut-off value of 1.15 as significant for SUVLN (area under the curve 0.683; P � 0.015; 95% CI 0.562-0.803; Figure 1). e sensitivity and specificity at this value were 84.2% and 59.2%, respectively. Based on the ROC curve analysis, the patients could be divided into two groups: SUV LN ≤ 1.15 vs. SUV LN > 1.15. Table 4 summarizes the results of the Cox proportional hazard model of prognostic factors for recurrence-free survival. Optimal cut-off values were 4.25 cm for tumour size and 10.65 for SUV Tumour , as determined by receiver operating characteristic curve analysis. Among clinicopathological characteristics and 18 F-FDG PET/CT parameters, SUV LN (P � 0.017), and tumour grade (P � 0.013) were risk factors for recurrence at univariate regression analysis. T-stage showed marginal significance (P � 0.054). Multivariate analysis showed that SUV LN (P < 0.0001) and tumour grade (P � 0.005) were the independent risk factors for recurrence. e patient group categorized by SUV LN showed a significant difference in PFS (log-rank test, P < 0.0001) as shown in the Kaplan-Meier survival curves (Figure 2).

Correlations between SUV LN and Clinicopathological
Characteristics. Table 5 depicts the correlation between SUV LN and clinicopathological characteristics in patients with CRC.

Discussion
e purpose of our study was to determine if preoperative metabolic activity in the lymph node as measured in terms of SUVmax on 18 F-FDG PET/CT had prognostic significance in patients with CRC. e preoperative evaluation of the metabolic activity of suspected LN by 18 F-FDG PET/CT was found to be a highly accurate prognostic tool for predicting recurrence in colorectal cancer. To the best of our knowledge, this study was the first to investigate the prognostic value of preoperative SUV LN for recurrence risk in colorectal cancer treated with radical surgery.
Several studies have evaluated 18 F-FDG PET in prognosis of patients with CRC. 18 F-FDG uptake of tumour lesions was a significant prognostic factor for CRC patients who underwent curative surgical resection [19,23]. In the present study, SUV Tumour was not associated with recurrence in patients with CRC. is may be due to the small number of enrolled patients. e principle finding was that preoperative SUV LN measured on 18 F-FDG PET/CT was the most significant factor for predicting recurrence in colorectal cancer.
ough the TNM classification has been widely used in the clinical setting to estimate prognosis, and previous studies have demonstrated that lymph node    metastasis was a well-known risk factor for disease-free survival and overall survival in patients with CRC [7][8][9][10], contrary to these prognostic factors, 18 F-FDG PET was advantageous in providing prognostic information even before surgery. Although several studies have evaluated the relationship between the recurrence and the SUV of lymph node in patients with CRC pathologically confirmed nodepositive disease [24,25], our study showed that preoperative SUV LN measured on 18 F-FDG PET/CT could predict the recurrence and it is not necessary to consider whether lymph node is positive.
In our study, the Pearson correlation coefficient provided strong evidence that SUV LN was positively associated with lymph node metastasis, tumour size and recurrence. ese results suggest that the metabolic activity of lymph node can be a useful functional marker of tumour aggressiveness. In this respect, the preoperative metabolic activity of suspected lymph node in patients with CRC, as measured by 18 F-FDG PET/CT, may become a novel and promising functional biomarker for predicting recurrence before surgery. e size, density, and morphology of the lymph nodes in malignant patients with early-stage may not change significantly, and thus, in such patients CT or MRI may prove less accurate in detecting metastatic lymph nodes. erefore, care is necessary in interpreting 18 F-FDG PET/CT results, especially when the lymph node size is small and the lymph node morphology is normal. Falsenegative findings are mainly due to the limited spatial resolution of 18 F-FDG PET/CT. erefore, preoperative metabolic activity of suspected lymph nodes in CRC patients measured by 18 F-FDG PET/CT scan may be a novel biomarker for predicting recurrence. When the SUV LN in CRC patients increased, the recurrence risk increased significantly. is result emphasized the importance of the metabolic activity of lymph node and highlights the possibility that preoperative SUV LN could be a promising prognostic marker before radical surgery. More intensive attention should be paid to high SUV LN group patients. Although the clinical benefit of the intensive surveillance strategy was not evaluated, earlier detection of recurrence might influence decision-making for treatment and may affect prognosis.
is strategy should be confirmed in further prospective studies.
is study has several limitations. First, our study was in part limited by its retrospective design at a single institution with  a small sample size. Further large prospective studies in different institutions are needed to validate the prognostic value of SUV LN in patients with CRC. Nevertheless, our study is noteworthy because it is the first study to show the importance and prognostic value of preoperative SUV LN in patients with CRC. Second, the area under the curve in the ROC analysis was 0.683, and the specificity using the cut-off of SUV LN was only 59.2%. Although the sensitivity was 84.2%, low specificity at the SUV LN cut-off value may be a problem in clinical practice. ird, we did not perform overall survival analysis because there were only six patients who were died during follow-up. Future large prospective studies are needed to investigate the relationship between SUV LN and overall survival in CRC patients.

Conclusion
In conclusion, we provide for the first time the importance of metabolic activity of lymph nodes, showing that preoperative SUV LN was an intuitive and simple method, significantly associated with recurrence in patients with CRC. erefore, preoperative assessment of SUV LN may be a promising prognostic marker to identify patients with a high risk of recurrence of CRC.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Ethical Approval
All procedures performed in studies involving human participants were in accordance with the principles of the institutional review board of Shanghai Jiao Tong University-affiliated Ren ji Hospital and the 1975 Declaration of Helsinki, as revised in 2013. is article does not contain any studies with animals performed by any of the authors.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
Ruohua Chen and Yining Wang contributed equally to this work.