| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Nuclear Medicine, Rene Huguenin Cancer Research Center, 35 Rue
Dailly, Saint-Cloud, France 92210.
2 Department of Nuclear Medicine, Henri Becquerel Cancer Research Center, Rouen,
France.
3 Department of Nuclear Medicine, Hospital Foch, Suresnes, France.
Received August 26, 2004; accepted after revision October 19, 2004.
Address correspondence to J.-L. Alberini
(jl.alberini{at}stcloud-huguenin.org).
OBJECTIVE. The aim of this study was to assess the performance of FDG PET/CT for the detection of colonic lesions, especially advanced neoplasms (villous or >10-mm adenomas, carcinomas). Because of 18F FDG accumulation in adenomatous polyps, PET using FDG can detect early premalignant colorectal lesions.
MATERIALS AND METHODS. FDG PET/CT studies performed for a 1-year period in 1,716 consecutive patients with various malignant diseases, except colorectal cancer, were retrospectively reviewed. PET images obtained 1 hr after FDG injection and noncontrast CT images used for attenuation correction were fused for analysis. Of 45 patients showing intense focal colonic FDG uptake, 20 patients (with 21 foci) underwent a colonoscopic investigation, and, when necessary, polyp resection. The intensity of FDG uptake was quantified using the standardized uptake value (SUVmax).
RESULTS. The FDG colonic foci were associated with 18 colonoscopic abnormalities in 15 patients, with no colonic abnormality detected in five patients (false-positive [FP] results). Histopathologic findings revealed advanced neoplasms in 13 patients (13 villous adenomas and three carcinomas) and two cases of hyperplastic polyps. A difference in the mean SUVmax was found between FP and true-positive colonic FDG foci but was not statistically significant (p = 0.14).
CONCLUSION. Presence of a focal colonic FDG uptake incidental finding on a PET/CT scan justifies a colonoscopy to detect (pre-)malignant lesions. The fusion of PET and CT images allows an accurate localization of the lesions. PET/CT is a useful tool to differentiate pathologic from physiologic FDG uptake.
Colorectal cancer is the second leading cause of cancer-related deaths in the Western world. However, its mortality can be reduced by the screening of asymptomatic persons over 50 years old [1]. Colonic adenomas are considered a precursor of colorectal cancer [2]; the adenoma carcinoma sequence is an established theory for colorectal cancer [3]. Detection and removal of adenomatous polyps are crucial for prevention and a potential cure. The clinical importance of small adenomas (< 5 mm) is controversial because only a few of them will acquire the additional genetic alterations predisposing to cancer. Advanced adenomas are defined as larger adenomas (> 1 cm) or adenomas with appreciable villous tissue or high-grade dysplasia. Prevention of colorectal cancer is now mainly based on reliable tests able to detect and remove the advanced adenomas before they become malignant [4–6].
PET imaging using 18FDG (FDG PET) is well accepted in the imaging workup of various malignancies. PET is recognized as a useful tool to manage colorectal cancer and was shown to have an additional value in the detection of colorectal cancer recurrence [7]. The usefulness of FDG PET for incidental premalignant colonic lesion detection has been previously reported [8–11] but to our knowledge, no study has evaluated combined PET/CT for the same purpose.
The aim of this retrospective study was to assess the value of PET/CT to detect colonic lesions, especially advanced neoplasms, compared with colonoscopic findings in patients referred for PET/CT studies for different malignant lesions with colonic nodular FDG foci.
Materials and Methods
Patients
We retrospectively analyzed the records of 1,716 consecutive patients who
underwent an FDG PET/CT scan for a malignant disease during a 1-year period
(October 2002–October 2003) at our center. Our institutional review
board does not require their approval or informed consent for review of
patient files and images. Patients with a previous history of colorectal
cancer were excluded from analysis. Forty-five patients with PET/CT scan
reports mentioning a colonic nodular FDG uptake were selected. In each report
conclusion, the possibility of a preneoplastic colonic lesion was stated and a
colonoscopy was recommended. Among these 45 patients, 20 patients (12 men,
mean age 65 ± 9 [SD] years old) with 21 nodular FDG foci underwent a
colonoscopy within 3 months after the PET/CT study. The primary malignancies
of these patients were lung carcinoma (n = 7), breast carcinoma
(n = 6), carcinoma of unknown primary (n = 3), Hodgkin's
disease (n = 1), melanoma (n = 1), pancreatic cancer
(n = 1), and ethmoidal cancer (n = 1). Despite a mean
follow-up of 7.8 ± 3.5 months and a mail and telephone contact with
each patient's referring physician, 25 patients did not undergo a
colonoscopy.
PET/CT Acquisitions
PET/CT studies were performed using a combined PET/CT scanner (Discovery
LS, GE Healthcare). Patients were asked to fast for at least 6 hr before FDG
injection; their blood glucose levels were determined in capillary blood
samples before FDG injection. In our department, the cutoff blood glucose
level that contraindicates FDG injection is 8 mmol/L. PET images were acquired
1 hr after injection of 4–5 MBq/kg of FDG on a 2D mode, from the skull
to the midthigh, with 5–7 bed positions of 5 min each. CT images were
used for attenuation correction and fusion; no contrast medium was used.
Helical CT was acquired first with the following parameters: 40 mAs, 140 kV,
5-mm section thickness, 0.8 sec per CT rotation, and 22.5 mm/sec table speed.
Whole-body CT was performed in a craniocaudal direction. Immediately
afterward, PET data were collected in a caudocranial direction. The CT data
were resized from a 512 x 512 matrix to a 128 x 128 matrix to
match the PET data to fuse the images.
Colonoscopy and Surgery
A rectosigmoidoscopy was performed in the five patients who had a focal FDG
uptake located within the rectosigmoid on PET/CT fused images. The 15
remaining patients underwent a colonoscopy. All the lesions that appeared
suspicious of malignancy were biopsied and removed by polypectomy and/or
surgery, and their localization was recorded. Colonoscopy with pathology
examination was considered the gold standard in our study.
Data and Statistical Analysis
Attenuation-corrected and noncorrected PET images, together with the
coregistered CT data, were reviewed. Our viewing facilities (eNTEGRA
workstation, GE Healthcare) included a triangulation tool; coronal, sagittal,
and transverse slices visualization; and PET/CT image fusion. The PET/CT image
review was performed without knowledge of the colonoscopy results. The FDG
uptake pattern was specified as nodular-focal or nodular-multifocal. The colon
was divided into five segments on PET/CT images (ascending colon, transversal
colon, descending colon, sigmoid, and rectum) for a more precise localization
of the FDG uptake, allowing comparison with the colonoscopic findings. A
PET/CT result was considered a true-positive (TP) when an FDG focus and a
colonoscopic abnormality were situated in the same segment. A false-positive
(FP) PET/CT result was defined as a focal FDG uptake without matching
colonoscopic abnormality. When the colonoscopy detected premalignant or
malignant lesions with no corresponding focal FDG uptake, the PET/CT result
was interpreted as false-negative (FN).
Per-patient and per-lesion analyses were performed to evaluate the FDG PET/CT positive predictive value for the detection of colonic abnormalities. We focused on the ability of PET/CT to detect advanced adenomas (e.g., adenomas >10 mm in diameter, adenomas with a villous component or moderate to severe dysplasia) and carcinomas.
A semiquantitative analysis of FDG activity was obtained by calculating the
maximum standardized uptake value (SUVmax) for each focal FDG
uptake in the attenuation corrected PET data. The SUVmax is a
glucose metabolism index derived from the FDG PET data. It represents the
ratio of a region of interest uptake to the average whole-body uptake
[12,
13].
 |
The SUVmax (mean and SE values) was correlated with the different histologic findings using a t test.
Our image review also included a record of the colon CT appearances in the region of the FDG uptake, taking into account the peristaltic intestinal motions that could have occurred between the CT and PET data acquisitions. We were specifically looking for the following features: fecal stasis, wall thickening, intraluminal mass, and pericolonic infiltration. Malignancies were suspected with these two latter criteria.
Results
PET/CT and colonoscopic findings are presented in Table 1.
|
Among the 20 patients included, 19 had one focus of FDG colonic uptake and one patient (patient 19) had two foci on the PET/CT scan. The FDG findings were located within the ascending colon (n = 7), descending colon (n = 3), and the rectosigmoid (n = 11).
For three patients (patients 8, 9, and 16), our image review led to a different conclusion from that of the scan report regarding the nature of the FDG focal colonic uptake. The colonic lesion was changed into a physiologic activity because the FDG uptake was located in the colonic lumen without bowel-wall involvement, in an area of fecal stasis. However, we consider this result as obtained after the fact. Consequently, our results are based on the study inclusion criteria as defined initially (that is, the PET/CT scan report), unless stated otherwise.
Per-Patient Analysis
Among the 20 patients who underwent a colonoscopy, the results were normal
in five cases (25%) and showed lesions in 15 patients (75%). In these 15
patients, 18 lesions were found: two benign hyperplastic polyps in the
rectosigmoid and 16 advanced neoplasms. The advanced neoplasms consisted of 13
villous adenomas and three adenocarcinomas. The size of the villous adenomas
ranged from 4 to 35 mm. The carcinomas were 10, 25, and 48 mm, respectively.
The FDG findings matched the colonoscopic abnormalities in 14 patients (TP
rate of 70%). In one patient (patient 5), the FDG finding and the polyp
detected by the colonoscopy were not located in the same site; thus, the PET
finding was considered an FP result. Among the 15 patients who underwent a
total colonoscopy, four FN PET/CT results were found in three patients
(patients 2, 5, and 6): one 5-mm hyperplastic polyp and three adenomas (two
mild to moderate grade dysplasia adenomas of 15 and 20 mm, respectively,
located in the ascending colon and rectosigmoid; and one high-grade dysplasia
adenoma [20 mm] in the rectosigmoid).
Per-Lesion Analysis
Fourteen FDG findings matched the colonoscopic abnormalities, which
consisted of 10 adenomas, three carcinomas, and one hyperplastic polyp. The
adenomas' dysplasia was characterized as mild to moderate in six cases and
high in four cases. The positive predictive value of PET/CT to detect
colonoscopic abnormalities was 67% (14/21)
(Table 2). Of the 16 advanced
adenomas diagnosed by the colonoscopy, 13 (81%) were detected on the FDG
PET/CT scan. Seven FDG findings situated in the ascending colon (n =
4), descending colon (n = 1), and rectosigmoid (n = 2) did
not match any colonoscopic abnormality, and thus, the PET/CT scan results were
interpreted as FP (33%).
|
When excluding the three FP lesions that were reassigned as TN findings after review and for which the colonoscopy was negative, the FP rate would have been 22% (4/18).
SUVmax ranged from 3.1 to 25 (mean value 10.5 ± 6.6). Figure 1 shows the correlation between the SUVmax and the histologic findings for four groups: carcinomas, high-grade dysplasia adenomas, mild-to-moderate-grade dysplasia adenomas, and negative colonoscopy. SUVmax was 15 ± 11.6 for carcinomas (n = 3), 12 ± 3.7 for high-grade dysplasia adenomas (n = 4), 8.8 ± 4.9 for mild-to-moderate-grade dysplasia adenomas (n = 6), 25 for the benign hyperplastic polyp (n = 1), and 7.1 ± 3.3 in the negative colonoscopy group (n = 7). Excluding the case of the only hyperplastic polyp, we calculated the mean SUVmax for the TP and FP results. It was higher in the case of advanced neoplasms (TP, 11.2 ± 6.5, n = 13) than the negative colonoscopy (FP, 7.1 ± 3.3, n = 7) but this value was not statistically significant (p = 0.14).
|
Discussion
From the per-patient analysis, it can be inferred that a focal colonic finding on FDG PET/CT should systematically lead to performing a colonoscopy. Considering the perlesion analysis, the main result of our study is that among the 21 FDG findings detected in 20 patients, 13 advanced neoplasms (10 adenomas and three carcinomas) were found.
Our results are consistent with two large studies showing that FDG PET is a sensitive tool to detect premalignant lesions [8, 9]. In a series of 3,210 PET scans performed for screening, focal FDG uptake was found in 20 patients and corresponded to 12 villous adenomas, six carcinomas, and two tubular adenomas in the colonoscopy [8]. In another series of 3,000 patients with no previous history of colorectal cancer, 13 focal or multifocal intense FDG uptakes matched 13 abnormalities in the colonoscopy (seven adenomatous polyps and six carcinomas) [9]. In another study, FDG PET had a sensitivity of 74% and a specificity of 84% for the detection of colonic abnormalities in 39 patients referred for a clinical FDG PET scan [10]; however, some patients who had colorectal cancer were included. FDG PET sensitivity to detect adenomas seems to be related to their size, as shown in a cohort of 110 asymptomatic patients who underwent a systematic PET scan and colonoscopy. The TP rate of FDG PET for the detection of adenomas larger than 13 mm was 90%; the FP rate was 5.5% [11].
In our study, the mean SUVmax of FDG colonic findings was higher in the TP PET/CT results than in the FP (11.2 ± 6.5 vs 7.1 ± 3.3). We showed that FDG uptake intensity was correlated with the severity of the lesions, although the cutoff value between benign polyps and advanced neoplasms was not clear. We observed that our mean SUVmax per category of lesion was higher than that found by Chen et al. [8]. Mean SUVmax of carcinomas and adenomas was 5.74 ± 2.26 and 3.56 ± 0.68, respectively. The SUVmax of the unique benign hyperplastic polyp detected in our study does not appear consistent with the rest of the SUVmax. We could not find a satisfactory explanation for such a discrepancy, which is not in agreement with a previous study that showed no FDG accumulation in 35 hyperplastic polyps [14].
|
|
In our study, only focal or multifocal FDG findings were considered significant. Despite this criterion, FP results represent 33% (7/21). This rate is lower than the one found by Zhuang et al. [22], who published 9/14 (64%) FP PET scan results for nodular FDG uptake. Nevertheless, the discrepancy can be explained by the higher accuracy of FDG uptake localization allowed by combined PET/CT. CT helps to localize more precisely the FDG findings, particularly when there is a spatial congruence with the bowel wall on the fused images (Figs. 2 and 3). Combining information from anatomic (CT) and functional (PET) imaging techniques improves the scans' diagnostic value compared with both techniques taken separately [23, 24]. A recent study [25] showed that combined PET/CT provided more information than the separate interpretation of PET and CT in 99/204 patients and affected the patients' management in 14% of the cases. The additional value of combined FDG PET/CT versus FDG PET alone was studied in patients with colorectal carcinoma [26]. FDG PET/CT halved the number of equivocal PET findings and increased the overall correct staging from 78% to 89%. Generally, the CT pattern of carcinoma includes pericolonic infiltration and colonic wall thickening [27], which may be an additional sign of malignancy but is not specific. On the contrary, FDG uptake delineating the bowel lumen on the fused PET/CT images indicates physiologic activity. In our study, an abnormal colonic pattern on the CT slices, such as an intraluminal mass or a pericolonic infiltration, proved to be predictive of a positive colonoscopy for eight of nine lesions. A review of PET/CT slices specifically focused on the colonic FDG uptake led to an improvement in the FP rate (22% vs 33%) (Fig. 4).
|
A proper interpretation of PET/CT implies PET and CT data analysis in combination, keeping in mind the possible inaccuracies of FDG uptake location generated by peristaltic motions occurring between CT and PET image acquisitions. These acquisitions are not simultaneous, and the time interval between them can be as long as 10–15 min in the abdominal area, generating misregistration. Administering an antiperistaltic agent, such as scopolamine, may be useful in diminishing these motion artifacts [28]. Glucagon has also been used to minimize the smooth muscle spasm and colonic peristalsis [29], but its use remains controversial [30] and it could increase blood glucose level, thus impairing the FDG PET image quality.
The use of oral contrast agents can improve PET/CT specificity by a better delineation of intestinal structures, thus allowing a better discrimination between physiologic and abnormal FDG uptake. An overestimation of the FDG activity measured on the attenuation-corrected PET images due to oral contrast agents has been shown [31]. However, this phenomenon does not seem to have much clinical effect: no significant difference in the distribution of intense FDG uptake in the intestinal regions was shown in two groups of 30 patients each [32]. To minimize these artifacts, low-density barium appears suitable for clinical use [33]; administration of a negative oral contrast agent using mannitol is also promising [34].
Avoiding confusion between a colonic FDG finding and physiologic activity in the urinary tract due to the FDG urinary elimination is an additional advantage of PET/CT coregistration. When the cause of an FDG uptake is in doubt, a delayed acquisition step focused on the area, usually the pelvis, can be performed after micturation [35].
Oral or IV contrast agents have to be evaluated to facilitate CT analysis, keeping in mind the artifacts that can be generated by miscalculation of attenuation coefficients and misregistration related to peristaltic motion.
Our study showed that colonic neoplasia and particularly advanced incidental neoplasm are suspected when colonic nodular-focal FDG findings are detected by PET/CT. Consequently, such findings have to be clearly mentioned in the reports. In a cohort of 20 patients with 21 areas of focal colonic uptake on a routine PET/CT scan, clinically significant lesions were found in 75% of the patients and 67% of the FDG findings. Coregistered PET/CT improves the analysis of colonic FDG uptake by avoiding confusion between abnormal focal uptake and physiologic activity due to fecal stasis, smooth muscular activity, or segmental inflammation. Measurement of SUVmax does not allow the differentiation between benign and (pre-)neoplastic lesions, but together with visual analysis, it can decrease the rate of FP results. Detection by PET/CT of high-intensity focal FDG uptake indicative of advanced adenoma should lead to a colonoscopy.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. B. Roedl, R. R. Colen, K. King, A. J. Fischman, P. R. Mueller, and M. A. Blake Visual PET/CT Scoring for Nonspecific 18F-FDG Uptake in the Differentiation of Early Malignant and Benign Esophageal Lesions Am. J. Roentgenol., August 1, 2008; 191(2): 515 - 521. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. De Wever, Y. Vankan, S. Stroobants, and J. Verschakelen Detection of extrapulmonary lesions with integrated PET/CT in the staging of lung cancer Eur. Respir. J., May 1, 2007; 29(5): 995 - 1002. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Gollub, T. Akhurst, A. J. Markowitz, M. R. Weiser, J. G. Guillem, L. McG. Smith, S. M. Larson, and A. R. Margulis Combined CT Colonography and 18F-FDG PET of Colon Polyps: Potential Technique for Selective Detection of Cancer and Precancerous Lesions Am. J. Roentgenol., January 1, 2007; 188(1): 130 - 138. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Czernin, M. Allen-Auerbach, and H. R. Schelbert Improvements in Cancer Staging with PET/CT: Literature-Based Evidence as of September 2006 J. Nucl. Med., January 1, 2007; 48(1_suppl): 78S - 88S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. B. Prabhakar, D. V. Sahani, A. J. Fischman, P. R. Mueller, and M. A. Blake Bowel Hot Spots at PET-CT RadioGraphics, January 1, 2007; 27(1): 145 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Veit-Haibach, C. A. Kuehle, T. Beyer, H. Stergar, H. Kuehl, J. Schmidt, G. Borsch, G. Dahmen, J. Barkhausen, A. Bockisch, et al. Diagnostic Accuracy of Colorectal Cancer Staging With Whole-Body PET/CT Colonography JAMA, December 6, 2006; 296(21): 2590 - 2600. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= mild-to-moderate-grade dysplasia adenoma (n = 6); 
=
negative colonoscopy (n = 7, false-positive FDG PET/CT result).
