ABSTRACT
Conclusion:
In conclusion, any unexpected focal 18F-FDG uptake in 18F-FDG PET/CT studies is suspicious for malignancy and should be clarified by colonoscopy. The intensity of 18F-FDG uptake does not preclude the application of colonoscopy and histopathological verification of the lesion if there is any.
Results:
Fifty two patients (83.8%) had at least 1 corresponding lesion on colonoscopy, whereas in 10 patients no lesion was detected. Subsequent histopathological examinations revealed no corresponding lesion in 13 (13.7%), a benign lesion in 18 (18.9%), hyperplastic polyp in 10 (10.5%), low-grade polyp in 16 (16.8%), high-grade polyp in 29 (30.5%) and malignant lesion in 9 (9.5%) of the focal 18F-FDG uptake sites. According to histopathology results, statistically no significant difference was found between the SUVmax measurements of malignant and benign cases (p>0.05) but the average SUVmax measurements of malignant cases were found to be significantly higher than lower + high-grade cases (p<0.05) and hyperplastic polyp cases (p<0.01).
Methods:
We retrospectively reviewed the files of 8,017 patients with known or suspected malignancy, who underwent whole-body 18F-FDG PET/CT at our institution during the period November 2017 to November 2019. Patients showing a single site of focally increased colonic 18F-FDG uptake that was more intense compared to liver uptake on 18F-FDG PET studies and referred to colonoscopy were enrolled in the study.
Objectives:
The present study aimed to identify the prevalence of focal uptake in the colon on 18fluorine-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) studies performed for the evaluation of malignancies other than colon, to detect the rate of malignancy in incidental focal 18F-FDG avid colonic lesions and to investigate if any possible role of maximum standardized uptake value (SUVmax) values in the discrimination of malignant lesions from premalignant and benign ones exist.
Inroduction
In imaging studies, an incidental finding, which is commonly named as “incidentaloma,” is a lesion which is detected serendipitously and is of indeterminate clinical significance. 18Fluorine-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) is increasingly being used as an imaging modality in oncology and this has led to an increasing number of focal 18F-FDG-avid lesions in several organs including the thyroid gland, adrenal gland gastrointestinal tract, pituitary gland, prostate gland (1,2,3,4,5). Although identification of some of these findings may provide a chance to treat a secondary primary malignancy, in many cases, further studies done for exploration of these lesions might cause unnecessary anxiety in patients, complications from additional medical interventions and economic burden (6,7).
Physiologic colonic 18F-FDG uptake is a commonly seen variant on PET scans and can be distinguished from malignant processes with its diffuse pattern. Segmental involvement of the colon in 18F-FDG PET studies is suggestive for inflammatory process. Focal involvement of the colon leaves the interpreter with a dilemma since it has the potential for malignancy, while some of the benign lesions also show 18F-FDG uptake. The prevalence of focal colonic incidentalomas detected by 18F-FDG PET or PET/CT was found as 3.6% pooled risk of malignant or premalignant lesions was 68% (8). It can be deduced from all these facts that it is crucial to know the malignancy rates of these lesions and distinguishing features on 18F-FDG PET/CT, in order to make further management properly.
The aims of the present study were to identify the prevalence of focal uptake in the colon on 18F-FDG PET/CT studies done for the evaluation of malignancies other than colon, rate of malignancy in incidental focal 18F-FDG avid colonic lesions and to investigate any possible role of maximum standardized uptake value (SUVmax) values in the discrimination of malignant lesions from premalignant and benign ones.
Materials and Methods
Results
Of the 8,017 PET/CT scans performed during the study period, 95 focally increased colonic 18F-FDG uptake was found in 62 (0.77%) patients. Among these 62 patients showing focal 18F-FDG uptake, 52 patients (83.8%) had at least 1 corresponding lesion in colonoscopy, whereas in 10 patients no lesion was detected. Of the 95 hypermetabolic foci, 7 were in the cecum, 13 in the ascending colon, 12 in the transverse colon, 13 in the descending colon, 22 in the sigmoid colon, 24 in the rectum and 4 in the anal region. Subsequent histopathological examinations revealed no corresponding lesion in 13 (13.7%), a benign lesion in 18 (18.9%), hyperplastic polyp in 10 (10.5%), low-grade polyp in 16 (16.8%), high-grade polyp in 29 (30.5%) and malignant lesion in 9 (9.5%) of the focal 18F-FDG uptake sites. So a premalignant lesion (high grade polyp + low grade polyp) and a malignant lesion were detected in totally 54 (56.8%) of the suspicious hypermetabolic foci.
Discussion
Focal incidental 18F-FDG uptake is a commonly encountered finding in PET/CT studies done of various oncologic diseases. Sone et al. (9) reported that in 6.7% of the PET/CT studies incidental finding is seen and in 2.2% of all patients, histopathology revealed malignancy in incidental lesions. The most commonly detected sites for these incidental lesions were the colon, lung and stomach (9).
In patients with more than one malignancies, although the prognosis becomes poorer, early detection of the second primary cancer improves the overall outcome (10,11). Adams et al. (6) reported that at least 1 incidental finding was reported in the findings section of 74.9% of PET/CT reports, resulting in a substantial additional cost per PET/CT study. So it becomes crucial to decide when to make suggestions for further investigations in the evaluation of incidental findings instead of follow-up and this requires knowledge of its differential diagnosis, the possibility of malignancy and other PET/CT criteria, which can be used in the characterization of the finding (12).
In our study, incidental colorectal uptake of 18F-FDG was found in 62 of 8,017 patients who underwent 18F-FDG PET/CT studies. In a meta-analysis comprising 32 studies, a pooled prevalence of 3.6% (0.4-16.3%) was reported for focal incidental colonic uptake (8). Initial detection rate of incidental colonic findings might be as high as 10% due to the high mean age of the patients included (13). Lower levels of incidental detection rate of 1.17% was also reported, which is attributed to the inclusion criteria of the study that restricted patients only to meet the strict criteria for focal colonic uptake (14).
In our study, among the patients showing focal 18F-FDG uptake, 83.8% had at least 1 corresponding lesion in colonoscopy, either benign or malign, whereas in 10 patients (16.2%) no lesion was detected. On lesion-based analysis 13 of the 95 (13.7%) focal uptake sites represented physiologic accumulation with no corresponding lesion. Although physiological activities are generally seen in diffuse pattern and therefore not included in studies on incidental 18F-FDG uptake, they can also be seen as a focal involvement. In the past studies, no correlative lesions were detected at endoscopy in 13.7-56% of patients with focal colorectal uptake of 18F-FDG who underwent colonoscopy (13,15,16,17,18). Although the mechanisms of this physiologic intestinal activity are not established clearly, several studies have suggested that the gut microbiota, peristaltic activity in muscles, presence of lymphoid tissue, high concentration of WBCs in the bowel wall, 18F-FDG secreting cells in the wall of the intestines, is responsible (18,19,20,21,22).
A premalignant lesion (high grade polyp + low grade polyp) and a malignant lesion were detected in totally 54 (56.8%) of suspicious hypermetabolic foci. This finding was consistent with the literature; Treglia et al. (8) conducted a systematic review and meta-analysis of incidental focal colonic uptake among 89,061 patients evaluated by 18F-FDG PET or PET/CT and they found the pooled risk of malignant or premalignant lesions as 68% [95% CI (60-75%)]. Another study conducted by Treglia et al. (17), reported malignant or premalignant lesions in 64% of the patients who underwent further investigations. The rate of false positivity can be related to the reference activity chosen for comparison to accept a lesion as positive. Liu et al. (23) reported colon cancer in 10 out of 24 (PPV: 42%) patients who showed incidental focal 18F-FDG uptake in 18F-FDG PET/CT studies, which means that 58% of lesions were false positive. They explained this relatively high false positivity by taking the comparison of focal 18F-FDG activity with blood pool activity as criteria for positive focal 18F-FDG (23). The authors compared their results with Cho et al.’s (24) study in which also low levels of PPV were reported (51.6%) and they attributed this to the definition of a positive criterion on imaging findings with SUVmax >3.5.
In our study, focal suspicious colorectal 18F-FDG uptake sites showing intense activity compared with the liver were accepted as positive focal 18F-FDG. So given that mean SUVmax value of liver is 2.89±1.26 (1.30-5.2) in our study, we can postulate that the false positivity rate (43.2%) could be lower if we accepted a higher reference SUVmax value.
As an indicator of metabolic activity, SUVmax value of lesions has been proposed as a semiquantitative index that can be helpful in the discrimination of benign and malignant lesions in oncologic PET/CT studies. When the colonic incidentalomas concerned, previous studies report that SUVmax value is unreliable enough to be used in the differentiation of malignant, premalignant and benign lesions because of significant overlap between SUVmax of benign and malignant lesions. Indeed, some of these previous studies have reported differences in SUVmax measurements between malignant and benign lesions or physiologic uptake, but none of them claimed that this difference in SUVmax precludes colonoscopy (18,25,26,27,28). Several studies have shown no statistical differences in SUVmax between true and false positive uptake sites (17,29,30). There have been made to determine a certain cut-off value for SUVmax in the discrimination of malignant and benign lesions; Luboldt et al. (28) proposed optimal SUVmax threshold of 5. However, their results were not confirmed by a study by Rigault et al. (31) in which they reported 14 advanced neoplasias with SUVmax values ≤5.
Hoeij et al. (25) reported sensitivity 80%, specificity 82%, PPV 34%, negative predictive value 98% in the discrimination of malignant and benign incidental colonic lesions when the optimal cut-off value was taken as 11.4. Although the authors stated that all incidental focal lesions showing 18F-FDG uptake with a SUVmax ≥11.4 should be examined by colonoscopy without delay, they concluded that SUVmax alone is not sufficiently discriminative to differentiate malignant, premalignant and benign lesions (25).
We could not find any significant difference between the SUVmax measurements of malignant and benign cases. When these benign lesions were overviewed; activated ulcerative colitis was the cause in 2 lesions of a patient, radiation colitis was detected in two patients who had undergone radiotherapy for gynecological malignancies and diverticula were detected in the descending colon of 2 patients. Although these lesions are not discriminated from malignancies due to their metabolic activity, when the morbidity of these lesions is regarded, early detection of them with 18F-FDG PET/CT becomes crucial (32).
Depending on the statistically significant high values of SUVmax in malignant cases compared with low + high grade cases and hyperplastic polyp cases, we tried assessing the threshold of SUVmax to differentiate the malignant cases and we found that above 5.2 the malignancy is detected with a sensitivity and specificity of 88.9% and 62.2% compared with low + high grade groups and 88.9% and 90%, when compared with hyperplastic polyp groups. Although these values seem to be high, given that 12 out of 18 benign lesions and 10 out of 13 physiologic uptake sites exhibit SUVmax values more than 5.2, in line with the literature, this cut-off value cannot be set as a strict threshold for the discrimination of malignant lesions.
Conclusion
In conclusion, any unexpected focal 18F-FDG uptake in 18F-FDG PET/CT studies is suspicious for malignancy and should be clarified by colonoscopy. The intensity of 18F-FDG uptake does not preclude the application of colonoscopy and histopathological verification of the lesion if there is any. SUVmax values more than 5.2 might only alert the physician to the higher risk of malignancy and force for urgent intervention.
Patient Population
We retrospectively reviewed the files of 8,017 patients with known or suspected malignancy, who underwent whole-body 18F-FDG PET/CT at our institution during the period November 2017 to November 2019. Patients with a previous history of colorectal cancer and inflammatory bowel disease, were excluded from our study. Patients showing a single site of focally increased colonic 18F-FDG uptake that was more intense compared to liver uptake on 18F-FDG PET studies and referred to colonoscopy were enrolled in the study. Of the 8,017 patients, 62 (30 men, 32 women; age range, 19-88 y; mean age, 63.66±10.09 y) met these criteria. The type and frequencies of primary malignancies are given in Table 1. Ethical approval was obtained from the Ethics Committee of University of Health Sciences Turkey, Prof. Dr. Cemil Tascioglu City Hospital (protocol number: E-48670771-514.10).
PET/CT Protocol: Imaging and Interpretation
Patients were imaged using an integrated PET/CT scanner that consisted of a full-ring HI-REZ LSO PET and a six-slice CT scanner (Siemens Biograph 6, Chicago, IL, USA).
Patients were instructed to fast, for 4-6 h before the injection of 370-555 MBq (10-15 mCi) of 18F-FDG. All patients were administered oral contrast starting 4 h before the study. Blood glucose levels were measured before the study and 18F-FDG was injected only when the blood glucose level was below 11.11 mmol/L. At 60 min post-injection, PET/CT scan was conducted with an emission time of 3 min per bed position from the vertex to the upper thigh. Before emission images, a low-dose CT scan was performed for attenuation correction and anatomical localization with the following parameters: 50 mA, 140 kV, and 5 mm section thickness. Image analysis was carried out on the Esoft multimodality computer platform (Siemens Medical Solutions, Erlangen, Germany). All images were reassessed by two experienced nuclear medicine physicians who were unaware of the endoscopic and histopathologic results. Focal suspicious colorectal 18F-FDG uptake sites showing intense activity compared with the liver were recorded, whereas diffuse and segmental uptake sites were excluded.
Regions of interest were manually drawn in transaxial slices encircling the focal activity to measure the SUVmax as a semiquantitative index. The colon was divided into 7 anatomical segments as the rectum, sigmoid colon, descending colon, transverse colon, ascending colon, cecum, anal region. The lesions are classified according to their locations in these segments by using the body low-dose CT component.
We accepted a findings in PET/CT results as true-positive when focal 18F-FDG uptake corresponded to a certain lesion in endoscopic or surgical evaluation. When no solid lesion was detected with these evaluations, the cause of the 18F-FDG uptake was attributed to physiologic accumulation of activity and the result was interpreted as false-positive. The true-positive lesions were further categorized as benign, premalignant and malignant.
Colonoscopic and Histopathological Evaluation
Histopathologic evaluation of the lesions following colonoscopy was used as the gold standard and performed in all patients within 60 days of the PET/CT scan. Biopsy or excision of the 95 lesions corresponding to the focal 18F-FDG uptake sites was performed. The descriptions of the lesions were also done morphologically during colonoscopy and the lesions were reported as polyp, mass lesion, diverticulum, hemorrhoid, ulcerovegetative mass, radiation colitis, rectovaginal fistula and ulcer. Any 18F-FDG uptake focus without a corresponding lesion on colonoscopy and negative histopathological result was considered as physiological.
On histopathological evaluation, the lesions are categorized as physiological; benign; hyperplastic polyp; low-grade polyp; high-grade polyp and malignant. The lesions were also categorized according to their dimensions as 1 cm >; 1-3 cm; 3-5 cm; 5 cm <.
Statistical Analysis
Number Cruncher Statistical System 2007 & PASS (Power Analysis and Sample Size) 2008 Statistical Software (Utah, USA) program was used for statistical analysis. While evaluating the study data, in addition to descriptive statistical methods (mean, standard deviation, median, frequency, ratio), Shapiro-Wilk test and box plot graphs were used for the normal distribution of variables. Mann-Whitney U test was used for intergroup comparisons of parameters not showing normal distribution. Spearman’s correlation analysis was used to evaluate the relationships between variables. Receiver operating characteristic (ROC) curve analysis and diagnostic screening tests were used to determine the cut off for the SUVmax value. Significance was evaluated at the p<0.05 level.
Malignant and Premalignant Lesions
The malignant lesions were found in the descending colon in 3 (33.3%), transverse colon in 3 (33.3%), rectum in 2 (22.2%) and the sigmoid colon in 1 (11.1%) cases (Figure 1). SUVmax in malignant lesions was 14.41±14.4 (0-49.8) on average. The size of the malignant lesions was 4.56±1.32 (2.5-5.5) cm on average. The distribution of 45 premalignant lesions detected on colonoscopy were 14 in sigmoid (31.1%), 7 in the rectum (15.5%), 8 in the ascending colon (17.7%), 7 in the descending colon (15.5%), 6 in transverse colon (13.3%), 3 in the cecum (6.6%). Histopathologic examination revealed 29 high-grade and 16 low-grade polyps (n=24 tubuler adenoma; n=21 tubulovillous adenoma) showing 18F-FDG uptake with an average SUVmax of 6.54±7.87 (0-27) and size of 1.41±1.13 (0.5-5.5) (Figure 2).
Hyperplastic Polyps
The distribution of 10 hyperplastic polyps among the colon segments was as follows: 5 lesions in the rectum (50%), 2 in the ascending colon (20%), 1 in the cecum (10%), one in the descending colon (10%) and 1 in sigmoid (10%). SUVmax in hyperplastic polyps was 2.76±5.12 (0-16.3) on average.
Physiologic Uptake and Benign Lesions
Histopathologic examinations revealed benign inflammatory pathologies in 18 of the lesions. Activated ulcerative colitis was the cause in 2 lesions (11.1%). Granulation was detected in the ulcerous ground in 5 lesions (27.7%). In 2 patients who had undergone radiotherapy for gynecological malignancies, radiation colitis was detected (11.1%). Diverticula was detected in the descending colons of 2 patients (11.1%). Hemorrhoids in the lower rectum was seen in 4 patients and rectovaginal fistulas in 1 patient (5.5%). Tuberculous ulcers were the cause in two incidental 18F-FDG uptake sites (11.1%) (Figure 3). SUVmax in these benign lesions was 7.88±5.09 (2.9-24.6) on average. In 13 of the hypermetabolic foci, colonoscopy revealed no corresponding lesions and the activity accumulations at these sites are attributed to physiologic uptake. The average SUVmax values at these false positive uptake sites were 9.22±4.88 (3.8-18.4). The average SUVmax in all lesions was 7.51±8.05 (0-49.8) (Figure 4).
According to histopathology results, statistically no significant difference was found between the SUVmax measurements of malignant and benign cases (p>0.05). The average SUVmax measurements of malignant cases were found to be significantly higher than low + high-grade cases (p<0.05) and hyperplastic polyp cases (p<0.01).
Based on this significance, it was considered to determine the cut-off point for SUVmax in detecting malignant cases. ROC analysis was used to determine the cut off point according to the groups.
For the 5.2 cut-off value of SUVmax measurement; sensitivity, specificity, positive predictive value (PPV), negative predictive value and accuracy in the discrimination between malignant and low + high-grade groups were 88.9%, 62.2%, 32%, 96.6%, 85.2% respectively. The ODDS rate for SUVmax measurement is 12.00 [95% confidance interval (CI): 1.38-104.3] that means that the risk of malignancy is 12 times higher in patients with SUVmax level of 5.2 and above.
When discrimination between malignant and hyperplastic polyp groups was concerned, for the 5.2 cut-off value of SUVmax measurement; sensitivity, specificity, PPV, negative predictive value and accuracy were 88.9%, 90%, 88.9%, 90%, 84.2% respectively.
A statistically significant difference was found between SUVmax measurements according to lesion size (p<0.01). SUVmax measurement of cases with a lesion size less than 1 cm was found to be statistically significantly lower than other dimensions.
According to histopathology results, size measurements of malignant cases were found to be significantly higher than premalignant cases (p<0.05).
Study Limitations
Our study has some limitations. First, it was limited by its retrospective and single center design. We couldn’t detect the sensitivity of 18F-FDG PET/CT in the identification of colonic neoplasms since whole lesions found in colonoscopy whether they were 18F-FDG avid or not are not recorded. All patients were administered oral contrast before the study as a part of routine applications in our PET/CT studies in our department. This might be mentioned as a limitation since it might have caused artefacts in the PET images, but these artefacts were distinguished from unusual focal 18F-FDG uptake with their diffuse patterns, and any misinterpretations were avoided. Administration of negative contrast material like water to improve bowel distention might have been more appropriate in this kind of study.