| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Nuclear Medicine |
1 From the Departments of Diagnostic and Interventional Radiology (A.G., G.A., H.K., T.E., E.H., S.G., J.D.), Oto-Rhino-Laryngology (M.F.), and Nuclear Medicine (A.B., L.F.), University Hospital of Essen, Hufelandstrasse 55, 45122 Essen, Germany. Received September 25, 2003; revision requested December 5; revision received March 11, 2004; accepted April 8. Address correspondence to A.G. (e-mail: Andreas.Gutzeit@ksa.ch).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Institutional review board approval and informed consent were obtained. Forty-five patients (26 men and 19 women) with metastatic cervical adenopathy (n = 18) or extracervical metastases (n = 27) of unknown primary tumor were included. The mean age of the patients was 57 years (range, 29–95 years). PET/CT imaging was performed in all patients 1 hour after administration of 350 MBq of fluorodeoxyglucose with a whole-body field of view. Contrast agents were administered orally and intravenously in all patients to ensure diagnostic CT data. PET/CT data sets were evaluated for the primary tumor, and imaging results were compared with those of CT, PET, and PET and CT side-by-side evaluation. Differences in diagnostic performance were assessed by using the McNemar test with Bonferroni correction, which accounts for multiple comparisons.
RESULTS: PET/CT depicted the primary tumor in 15 (33%) of 45 patients. In 30 (67%) patients, the primary tumor site remained occult (P > .05). PET and CT side-by-side evaluation depicted 13 (29%) of 45 tumors (P > .05). PET alone revealed the primary tumor in 11 (24%) of 45 patients (P > .05), while CT alone helped in the correct diagnosis in eight (18%) of 45 patients (P > .05). There were no significant differences between the diagnostic accuracies of PET/CT and the other imaging modalities.
CONCLUSION: PET/CT was able to depict more primary tumors, though not significantly, than either of the other imaging modalities, but larger patient cohorts are required to finally judge its value for revealing the primary tumor site.
© RSNA, 2004
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
To date, fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET) has been demonstrated to be the most efficient method to localize unknown primary tumors, with detection rates between 24% and 53% (3). Thus, the indication for FDG PET in patients with cancer of an unknown primary tumor is graded Ia according to the German consensus conference. Grade Ia indicates scientifically proved benefit and an established clinical use (3). The main limitation of FDG PET relates to its limited anatomic information. Frequently, subtle FDG tracer accumulations cannot be precisely pinpointed. In such cases, a combination of functional PET imaging with a high-spatial-resolution morphologic display, as provided with computed tomography (CT) or magnetic resonance (MR) imaging, would be desirable. While image fusion algorithms for functional PET and morphologic CT or MR data sets have been successfully implemented in the brain, motion-induced data misregistration hampers computer-aided fusion of separately collected data in other body regions (4).
Recently, combined in-line PET/CT has become commercially available (5). By providing coregistered morphologic and functional data sets as part of a single examination, faulty coregistration owing to motion-induced misalignment is minimized. Initial studies in which the diagnostic effect of dual-modality PET/CT was assessed in patients with various malignant diseases have revealed promising results (6–9).
The aim of this study was to retrospectively evaluate the benefit of fused PET/CT in depicting the primary lesion in cancer of an unknown primary tumor, as compared with PET, CT, and PET and CT side-by-side evaluation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
PET/CT Imaging
Dual-modality PET/CT was performed by using a Biograph scanner (Siemens Medical Solutions, Hoffmann Estates, Ill). The CT component is based on dual-section helical CT (Somatom Emotion; Siemens Medical Solutions, Erlangen, Germany), while the full-ring PET component is based on ECAT HR+ (Siemens Medical Solutions, Hoffmann Estates). The CT component provides a minimum gantry rotation time of 800 msec and a maximal scan time of 100 seconds. CT images were acquired with 130 mAs, 130 kv, section width of 5 mm, and table feed of 8 mm per rotation. To ensure diagnostic CT image quality, 140 mL of a contrast agent containing 300 mg iodine per milliliter (Xenetix 300; Guerbet, Sulzbach, Germany) was administered intravenously according to a standardized protocol with use of an automated injector (LF; Medical Supplies, Kerpen, Germany) (10). For intestinal delineation, either 1000 mL of barium or a combination of 1500 mL of water containing 0.2% locust bean gum and 2.5% mannitol was administered (11). A limited breath-hold technique was used to avoid motion-induced artifacts in the area of the diaphragm (12). The scanning area for CT and PET was defined on a CT topogram. Single-section whole-body spiral CT was performed starting with the head and subsequently covering the neck, thorax, abdomen, and pelvis.
The PET component represents a full-ring tomograph with an in-plane spatial resolution of 4.6 mm and a transverse field of view of 15.5 cm for each bed position. PET images were corrected for attenuation on the basis of the CT data, and iterative reconstruction algorithms with two iterations and eight subsets were performed (13). PET imaging was performed 60 minutes after the administration of 350 MBq of FDG. Patients had been instructed to fast for a minimum of 4 hours prior to starting the examination. Blood samples collected before the injection of the radioactive tracer ensured blood glucose levels in the normal range.
Data Analysis
The Biograph scanner provides separate CT and PET data sets that can be accurately coregistered on a workstation featuring Syngo software (Siemens Medical Solutions, Erlangen). Both PET and CT data sets can, however, be evaluated separately without image fusion or they may be viewed side by side. To avoid potential errors owing to an effect of intravenous and oral contrast agents on PET emission data, we evaluated the PET data sets with and without attenuation correction. PET data sets were evaluated by two nuclear medicine physicians (A.B. and L.F. with PET experience of 20 and 6 years, respectively) in consensus, while CT data sets were evaluated by two radiologists (H.K. and E.H. with CT experience of 11 and 6 years, respectively) in consensus. Both the nuclear medicine physicians and the radiologists had 2 years of PET/CT experience.
On CT images, the primary tumor assessment was based on the detection of a contrast material–enhanced mass, whereas on PET images, a focally increased glucose metabolism with a standard uptake value exceeding 2.5 was considered pathologic. The evaluating physicians were informed about the clinical background but were blinded to the results of the other imaging modality. After separate image evaluations, PET and CT data sets were evaluated side by side without image fusion by one nuclear medicine physician (L.F.) and one radiologist (E.H.) in consensus. Side-by-side image evaluation was performed on two different screens after PET and CT images had been manually misregistered in all three dimensions by a third person not involved in the evaluation process. Finally, PET/CT images were evaluated by the same two physicians. In the evaluation of PET and CT images side by side and of the PET/CT images, morphologic and functional information was considered for primary tumor detection, and a consensus was reached on a case-by-case basis in patients with discordant findings at CT and PET. There were no cases of divergent results. To avoid recognition bias, we chose an interval of 4 weeks between the reading sessions and a randomized order of the images. All potential sites of the primary tumor depicted by the different imaging procedures were histologically verified.
Statistical Analysis
Data analysis was performed on the basis of all patients and separately for patients with extracervical lymph node involvement (group 1) and patients with cervical lymph node metastases (group 2). Sensitivity and positive predictive values regarding the detection of the primary tumors were calculated for the different imaging procedures.
The four modalities (PET, CT, PET and CT evaluation side by side, and PET/CT) were compared with each other, which resulted in six 2 x 2 contingency tables. Differences in diagnostic performance were assessed by using the McNemar test with Bonferroni correction, which accounts for multiple comparisons. For this purpose, two thresholds were tested for each diagnostic modality. The first threshold was set as the number of true-positive and false-positive versus false-negative findings, and a second threshold was set as the number of true-positive versus false-positive and false-negative findings. A confidence level of less than .05 was considered to be statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
PET imaging alone correctly depicted the primary tumor in 11 (24%) patients, with six (13%) false-positive findings (P > .05). With CT imaging alone, eight (18%) primary tumors were correctly diagnosed, with three (6.7%) false-positive findings (P > .05). Figures 1 and 2 demonstrate cases in which the primary tumor was missed at CT. Sensitivities and positive predictive values for the different imaging procedures are listed in Table 2.
|
|
|
There were two patients (patients 11 and 35) with false-negative findings at CT, PET, and PET and CT side-by-side evaluation. PET/CT, however, helped determine the correct diagnosis and helped characterize two lesions as the primary tumors in the submandibular gland and in the right main bronchus.
Group 1: Extracervical Metastases
When only patients in group 1 were considered, PET/CT depicted nine (33%) of 27 primary tumors (P > .05), with two false-positive findings. PET and CT side-by-side evaluation helped identify the primary tumor site in eight (30%) patients (P > .05), with two false-positive findings. On PET images, the primary tumor was identified in seven (29%) patients (P > .05), with three false-positive findings. CT images revealed the location of four (15%) primary tumors, with one false-positive finding (P > .05). Sensitivities and positive predictive values for assessment of patients in group 1 are summarized in Table 3.
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
CT-based diagnostic work-up of patients with cancer of an unknown primary tumor has been shown to be rather limited in identification of the primary tumor sites (18–20). Compared to most reports, the CT results of this study were favorable: Eight (17%) of 45 primary tumors were identified when interpretation was based on CT images alone. To some degree, these results reflect the high diagnostic standard achieved with the whole-body CT protocol. The oral administration of either barium or a water-based contrast agent ensured adequate delineation of bowel loops. Even though authors of studies addressing the issue of oral contrast agents in PET/CT have not found a substantial effect on PET image quality and PET tracer quantification (21), local accumulation of CT contrast agents may introduce artifacts into the PET emission data (22). Intravenous contrast material was injected in a standard dynamic fashion. Although intravenous contrast agents have also been shown to go along with artifacts on PET images in selected cases, these artifacts seem to be limited to areas of a contrast agent bolus passage in thoracic veins (23). The effect of intravenous contrast agents on PET tracer quantification was found to be of minor issue in the clinical setting (24). On the basis of these data, a substantial effect of these contrast agents on PET image interpretation was not to be expected.
By mirroring other published data (19), the diagnostic utility of PET alone was superior to that of CT when searching for the primary tumor in patients with known cervical metastases. In our study, 11 primary tumors were correctly identified, which corresponded to a success rate of slightly better than 24%. Although these data must be viewed in conjunction with a considerable number of false-positive findings, they compare rather favorably with data of other studies in which sensitivities of FDG PET were reported to range between 8% and 33% for depicting primary tumor sites in patients with cancer of an unknown primary tumor (18,25–31).
The simultaneous availability of functional PET and morphologic CT data sets enhanced the diagnostic performance with regard to the number of true-positive and false-positive findings. Side-by-side viewing of the two data sets enhanced the diagnostic performance considerably over that of CT or PET alone. Thus, compared with PET, two additional primary lesions could be identified with PET and CT side-by-side evaluation, which raises the sensitivity rate above 30% and at the same time lowers the number of false-positive findings by a third. Coregistration of PET with CT images further improved the detection rate of the primary lesion. PET/CT led to the diagnosis of carcinoma of the submandibular gland (Fig 3), which had been misinterpreted as a lymph node metastasis at PET and CT side-by-side evaluation. Furthermore, PET/CT depicted a bronchial carcinoma that was misinterpreted as a lymph node metastasis at PET and CT side-by-side evaluation (Fig 4). In these cases, the simultaneous availability of fused CT data helped reduce the number of false-positive PET findings that related to the erroneous interpretation of lymph node metastases as primary tumors. These findings stress the advantage of accurate fusion of functional and morphologic data for optimized detection of the primary tumor and for guidance of any subsequent surgical intervention (32,33).
When PET and PET/CT were compared, PET/CT was able to correctly depict the primary tumor site in four patients in whom findings of PET alone had been negative. Furthermore, the number of false-positive findings was reduced by three with use of PET/CT. Though these differences did not demonstrate statistical significance, further differences between the two imaging modalities have to be discussed when a potential benefit of PET/CT over that of PET alone is assessed. A great advantage of combined PET/CT over PET alone is the vast shortening of the examination time induced by the CT-based PET attenuation correction that is inherent to the in-line PET/CT system. Thus, the mean examination time of PET/CT amounted to less than 30 minutes, which compares favorably with the time requirements of approximately 45 minutes for conventional whole-body PET.
Similar assessment was made for patients with extracervical lymph node metastases (group 1) and patients with cervical metastatic disease (group 2). All patients had undergone panendoscopy for search of the primary tumor before undergoing PET/CT. Only those with negative results were referred to as patients with an unknown primary tumor, which may explain the inferior performance of PET and PET/CT when compared with data from the literature. In both groups, detection rates in excess of 33% were achieved. At the same time, the number of false-positive findings was further reduced with PET/CT, increasing the positive predictive value to 83% for all patients.
Equally encouraging results regarding the diagnostic power of PET/CT data sets collected with a dual-modality system have been reported for other diseases (8,9). Nevertheless, the failure rate of 67% for nondetectable primary cancers points to how much room there is for improvement in optimizing diagnostic strategies in patients with cancer of an unknown primary tumor. In most cases, the primary tumor remains unidentified during the lifetime. If comprehensive postmortem diagnostic work-up is performed, the most frequent primary tumor site is the pancreas (34). Pancreatic carcinomas are, however, difficult to detect with diagnostic imaging procedures, especially if they are small and do not cause biliary obstruction. Findings of other studies (35,36) showed that PET is a feasible diagnostic modality for depicting pancreatic carcinoma, with a sensitivity between 82% and 85%, but these studies do not reveal the patients with an unknown primary tumor site. In comparison with the high failure rate of the currently available radiologic procedures, findings of only postmortem autopsies have been shown to provide a higher diagnostic detection rate of the primary tumor, which was detected in up to 57% of patients with an unknown primary tumor (37).
The difference in diagnostic performance between PET/CT data sets and side-by-side image evaluation did not prove to be statistically significant. This may reflect a study bias related to the acquisition of PET and CT data with the same imaging system by using dedicated PET/CT protocols. Usually, PET and CT are performed with different imaging systems and will, therefore, differ in the field of view, the respiration state, and the location of movable organs. Furthermore, organ shift may become more profound as the interval between the two imaging procedures increases. Clearly, identical patient positioning and spatial imaging parameters for both PET and CT minimized misregistration but can only be partially accounted for as a result of manual misregistration of PET and CT data sets.
Finally, a dedicated CT image will usually be acquired at end inspiration, while PET data are acquired during shallow breathing. The ensuing difference in respiratory states has been demonstrated to cause considerable misalignment between the two data sets (38–39). In this study, patients were examined according to an optimized PET/CT protocol designed to minimize motion-induced artifacts by using a limited breath-hold technique (12). Thus, CT data of the chest and upper abdomen were collected at end expiration, while PET data were acquired during shallow breathing. Thus, differences between PET/CT and side-by-side image evaluation may be more profound if PET and CT images had been acquired with different scanners. Separate PET and CT images in addition to the combined PET/CT images would, however, expose patients to additional radiation, rendering such a study design ethically problematic.
In conclusion, PET/CT was able to depict more primary tumors than did CT, PET, and PET and CT side-by-side evaluation. In addition, the number of false-positive findings was reduced with PET/CT. However, data analysis did not reveal a statistically significant improvement when patients with cancer of an unknown primary tumor were imaged with PET/CT compared with the other imaging procedures. While this finding may be a result of the limited number of patients included in the data analysis, larger patient cohorts are required to further assess the value of accurately coregistered PET/CT images over that of PET, CT, and PET and CT images evaluated side by side when assessing patients with cancer of an unknown primary tumor.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, all authors; study concepts, A.G., G.A., H.K.; study design, A.G., G.A., H.K., T.E.; literature research, E.H., S.G., L.F.; clinical studies, A.G., G.A., T.E., E.H., J.D.; data acquisition, T.E., M.F., E.H.; data analysis/interpretation, S.G., L.F., J.D.; statistical analysis, S.G., L.F.; manuscript preparation, A.G., G.A., E.H., S.G., J.D., T.E., M.F.; manuscript definition of intellectual content, editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. M. Blodgett, C. C. Meltzer, and D. W. Townsend PET/CT: Form and Function Radiology, February 1, 2007; 242(2): 360 - 385. [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]
|
|
|
|
 |
|
 J. D. Berry and G. J. R. Cook Positron emission tomography in oncology Br. Med. Bull., December 14, 2006; (2006) ldl013v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Mukherji and C. R. Bradford Controversies: Is There a Role for Positron-Emission Tomographic CT in the Initial Staging of Head and Neck Squamous Cell Carcinoma? AJNR Am. J. Neuroradiol., February 1, 2006; 27(2): 243 - 245. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
