It is very common to use magnetic resonance imaging (MRI), computed tomography (CT) to evaluate the pre-operative lymph node status in colorectal cancer patients. However, the low sensitivity can lead to bias and incorrect evaluation.1, 2 The latest technology, which uses positron emission tomography/computed tomography. 18FDG PET/CT is becoming more popular, especially with regard to using the surveillance and pre-operative staging to assess recurrence of disease and distant metastases.
Yoshiyuki’s research3 reported that the using of PET/ CT scan for pre-operative diagnosis of lymph mode metastasis of colorectal cancer gives a sensitivity of around 51.2%, specificity of around 85.1% and accuracy of around 69.3% for the regional node group.The purpose of this research was to study the association between the diagnostic value of 18FDG PET/CT scan (maximum standardized uptake value - SUV) for pre-operative lymph nodes status and the results shown by the final histopathology.
This was a retrospective study which gathered the information from patient medical records at Bangkok Hospital Medical Center. The patients were examined for colorectal cancer between May 2007 and November 2009; they received the pre-operative 18FDG PET/CT scan before having oncologic colorectal surgery. The subjects in this study numbered 30 patients, 20 males and 10 females. The average age was 63, and 64.4 years old respectively for male and female patients. Table 1 shows the locations of primary tumors revealed in the colon of 20 patients and the rectum of 10 patients. However, in this sample, the primary tumor of one patient was not detected by 18FDG PET/CT scan (post incomplete polypectomy) and the lymph node status of 6 patients was not revealed either.
Inclusion / Exclusion Criteria
The records of inpatients with carcinoma of rectum and/or colon (code ICD10), who were examined and operated on between May 2007 and November 2009, showed that every patient was examined by 18FDG PET/ CT scan in order to evaluate distant metastases within the 2 week period prior to their operations, and had tissue histopathology to confirm the diagnosis before surgery. Diabetic patients fasting blood sugar (FBS) should have been lower than 200 mg%, or have foregone insulin injection before having 18FDG for PET/CT scan.
PET/CT scans Technique
The 18FDG PET/CT, Gimini GXL was used for this study. The scintillator was gadolinium oxyorthosilicate (GSO) with a detector dimension of 4x6x30 mm. The axial field of view was 16.2 cm with sensitivity of 8.3 cps/kBq. Peak noise equivalent count rate (NECR) was 70 kcps. The transverse and axial resolutions were 5.1 and 5.5 respectively. The CT mode was used for attenuation correction. All patient fasted 6 hours; fasting blood sugar was less than 200 mg%. Patients were injected with a dosage of 0.14 mCi/Kg of 18FDG. Then after resting for 1 hour, the study was begun. The field of study included the base of the skull down to the upper thighs. CT scan technique: Detector type was cadmium tungstate (CdWO4) with 16 slices; 120 kV, 250 mA.; slice thickness 5 mm. Immediately after CT scan was done, PET scan was performed with total 6 - 7 bed positions. The raw data of PET study was attenuation corrected by using the CT scan, then reformatted into trans axial, coronal and sagittal views.
Interpretations of PET/CT scan
The PET/CT scan was reviewed by one nuclear medicine professional, who had no prior patient detail such as diagnosis of any previous PET/CT scans.The report was interpreted from the SUV of lymph nodes. Moreover, we also had the report on the SUV of primary tumor(s).
An association between SUV of lymph node and result from final histopathology was calculated using by Student’s t-test. A corrected p < 0.05 was considered evidence of statistical significance. Comparison of SUV between each cell types of primary tumor was done, using one-way analysis of variance (ANOVA).4
Table 1: Demographic data of colorectal cancer patients.
From the review, we have found that the CT scan could identify 24 out of 30 patients with pre-operative lymphadenopathies. The final histopathology showed 14 out of 24 patients (58.3%) with metastatic lymphadenopathies (Figure 1). In addition, there were 6 out of 30 patients whereby the 18FDG PET/CT scan was unable to identify pre-operative lymphadenopathies. The final histopathology showed that 2 out of those 6 patients (33.3%) did have metastatic lymphadenopathies, but the PET could not detect them (Figure 2). The imaging of CT showed no demonstrated regional node enlargement (Figure 3). The mean of the SUV in malignant lymphadenopathies (1.18 ± 0.69) together with benign lymphadenopathies (0.59 ± 0.54) in primary colorectal cancer patient had the significant differentiation (p = 0.014) which was shown in Table 2. The mean total of malignant lymph nodes found per patient during colorectal surgery was 13 (range 3 - 25).
The correlation between SUV and cell type of primary tumor, we found that there was no significant difference in the statistic between cell types (well, moderately and poorly differentiated) and SUV (p = 0.304) (Table 3). There was only one patient whose the final histopathology of primary tumor revealed lymphoma and we did not include this patient’s results for the evaluation.
Figure 1 : True positive PET-CT. This case of primary CA Colon with regional node metastasis SUV 8.0 at right pelvic cavity shows increased uptake.
Figure 2 : False negative PET. This case of primary CA Splenic flexor shows at primary site but regional node has not increased uptake SUV 0.8.
Figure 3 : False negative CT. This case of primary CA Splenic flexor, only CT study shows no demonstrated regional node enlargement.
Table 2: Pathological lymph nodes detected by PET/CT.
Table 3: Comparison of standardized uptake value (SUV) between each cell type of primary tumor.
The hypothesis is that cancer cells will have a higher glycolysis metabolic rate than normal cells.5 Given that assumption, using PET scan for 18F-fluorodeoxyglucose (18FDG) detection to count the glucose uptake on the cellular level will help the detection of primary tumors and metastatic lesions. However, different primary tumors have different uptakes of 18FDG. It is believed that important enzymes for glucose uptake in cancer cells are hexokinase and glucose transporters (Gluts), which are located on the cell membrane.6 Gauthier’s study of mitrochondrial hexokinase found that it accounted for 75-80% of total intracellular enzyme activity.7 Different cell types in colorectal cancer tumors-well, moderate and poorly differentiated - have different enzyme activities, meaning there is a diversity of 18FDG uptake.
Our study however, did not show significant difference of 18FDG uptake between cell type (p = 0.304 as shown in Table 3), presumably because of our small sample size. Furthermore, there are many other factors that impact on and account for the variation in 18FDG uptake, which include 18FDG quantity, the time period between 18FDG injections and actual time of PET scan, base line blood sugar before injection of 18FDG, as well as patients’ weight and PET scan model used. Other enzyme activities can affect 18FDG uptake, for example glucose phosphatase. Due to the abovementioned factors, it is important to set up the criteria which take these factors into account, in order to reduce the deviation on the 18FDG uptake (SUV) calculation.
Yoshioka’s study8 found that the trend of 18FDG uptake in histological cancer cell line (which is less differentiated in terms of using PET/CT scan to evaluate lymph node metastasis) is close to the primary tumor, if using SUV for the detection. Yoshiyuki also found that there should be more than 1.5 on cut-off point in metastatic lymph node in colorectal cancer.3 PET/CT scan results in this study, then, also showed the trend of higher uptake of 18FDG in metastatic lymph nodes, compared with non-metastatic lymph nodes. This can be confirmed by the significant result from final histopathology (p = 0.014). However, as study was retrospective, it still had the constraint of not being able to exactly confirm that the lymph nodes identified as positive by the 18FDG PET/CT scan, actually matched those lymph nodes detailed in the final histotathology. The size of lymph node is another factor that affects the evaluation of 18FDG PET/CT scan results. It was found that the lymph nodes that may cause metastasis are usually larger than 1 cm.9, 10
We show a clear correlation between 18FDG PET/CT diagnoses of suspicious malignancies and final histopathological lymphadenopathies. It is therefore clearly beneficial to further document the pre-operative predictive role of the 18FDG PET/CT scan in diagnosing lymph node metastasis in colorectal cancer patients. One of the constraints of this study, being retrospective, is that we could not confirm the exact correlation between the location of the malignant nodes identified by PET/CT, and those in the final histopathology report.
The pre-operative 18FDG PET/CT scan shows the tendency of malignant lymph nodes to have a higher uptake of FDG than is seen in benign lymph nodes. However, due to our small sample size, the cutoff point of significant SUV could not be exactly calculated. This report will be the basis for further studies in the future.