| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Investigations |
1 Department of Nuclear Medicine and Radiation Biology, Faculty of Medicine, Université de Sherbrooke, Sherbrooke, Quebec, Canada
2 Department of Medicine, Faculty of Medicine, Université de Sherbrooke, Sherbrooke, Quebec, Canada
3 Metabolic and Functional Imaging Center, Clinical Research Center, Centre Hospitalier Universitaire de Sherbrooke, Sherbrooke, Quebec, Canada
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Key Words: Hodgkin’s disease • 18F-FDG PET • lymphoma
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
With modern treatment protocols, tumor burden (i.e., stage) rather than histologic subtype is the most important prognostic variable. The current 5-y survival is close to 90% for stages I and IIA. Even in cases of advanced disease, stages IVA and IVB, the 5-y survival can be as high as 70% (1,2). Consequently, long-term effects of chemotherapy and radiotherapy (secondary cancers, pulmonary fibrosis, and accelerated atherosclerosis) are now becoming a major concern. Judicious use of chemotherapy and radiotherapy can be greatly facilitated by modern diagnostic imaging modalities.
18F-FDG PET is now a well-recognized diagnostic tool used for staging and monitoring response to therapy in lymphomas (3–6). The advantages of 18F-FDG PET technology over conventional diagnostic imaging methods are particularly obvious in the postchemotherapy period. Up to 64% of patients demonstrate residual abnormalities on CT after completion of therapy, but only 42% of those patients will relapse on follow-up (7–9). This is primarily caused by inability of CT to distinguish viable tumor tissue from fibrosis.
Several studies have been published that address the effectiveness of therapy monitoring and assessment for tumor recurrence with 18F-FDG PET in HD and NHL (3,4,6,9). However, most of these studies included cases of HD and NHL without distinction, which is of utmost importance to the referring oncologists, considering their differing treatment and clinical outcome. Two studies analyzed the role of 18F-FDG PET in cases of HD as a single population after therapy, but neither of these studies assessed the importance of the time interval between the end of treatment and 18F-FDG PET (10,11).
In this study, our objectives were to determine the value of 18F-FDG PET in predicting relapse in relation with the time intervals between the end of chemotherapy and the PET study. We also compared these results to those obtained with conventional diagnostic imaging (CT). Disease-free intervals were also calculated with respect to the standardized uptake values (SUVs) of the most active lesion in patients with positive PET studies.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
PET
PET was performed using an ECAT EXACT HR+ scanner (Siemens Medical Systems) with attenuation correction using rotating 68Ge rod sources. Two-dimensional acquisitions were made from the neck to the upper part of the thighs 60 min after intravenous administration of 7.5 MBq/kg of 18F-FDG. All patients had to fast at least 6 h before the study. The capillary blood glucose level was monitored for hyperglycemia in all patients. Diabetic patients were asked not to use subcutaneous rapid-onset insulin on the day of the study. In hyperglycemic patients (with capillary glucose levels, >7.5 mmol/L), an intravenous bolus of Humulin R insulin (Eli Lilly) was injected according to a preestablished chart at least 60 min before the 18F-FDG injection to reduce the blood glucose level without increasing muscle uptake. The acquisition times were 5–7 min for the emission scan and 3 min for the transmission scan for each bed position, with a typical body coverage of 5–7 bed positions from the neck to the upper femurs. Scans were reconstructed using the ordered-subset expectation maximization algorithm (2 iterations/16 subsets) in a 128 x 128 matrix with a gaussian filter (8-mm full width at half maximum) and were reoriented in transverse, coronal, and sagittal planes. All images were qualitatively reviewed by an experienced investigator with access to clinical data. Intravenous furosemide (40 mg) was given to patients when suspicious foci of 18F-FDG uptake were visible in the pelvic area to minimize interfering urinary activity. These patients had a second imaging study centered on the pelvis approximately 45 min after the furosemide administration. All foci of elevated 18F-FDG uptake not explainable by physiologic uptake (i.e., brain, myocardium, urinary tract, digestive tract, muscles) were considered to represent viable lymphoma. The PET studies were classified as positive or negative. The standardized uptake value (SUV) was measured at the most active lesion in patients with positive PET studies. A region of interest (ROI) with a diameter corresponding to 75% of the maximal lesion activity was used. This threshold was used to reflect the most active part of the tumor and to ensure consistent ROI drawings. The SUV measurements were also normalized for actual body weight.
CT scans of the pelvis, abdomen, or thorax were obtained using a SOMATOM Plus 4, fourth-generation helical scanner (Siemens Medical Systems), in 32 patients. Slices were obtained at 8-mm intervals for thoracic studies and 5-mm intervals for abdominal and pelvic studies. Intravenous contrast material (100–120 mL of nonionic contrast medium) was used in all patients. Experienced radiologists interpreted the scans. Any lymph node >1 cm in diameter was considered positive for residual lymphoma.
Statistical Analysis
Disease-free intervals and proportions were calculated with the Kaplan–Meier method. Comparisons between groups were made using the log- rank test. All other comparisons were made using the Fisher exact test. The analyses were performed with GraphPad Prism version 3.00 for Windows (GraphPad Software).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
Relapses were significantly more frequent in positive 18F-FDG PET patients than in negative 18F-FDG PET patients (2-tailed P < 0.0001). As shown in Figure 1 with the Kaplan–Meier method, the median disease-free interval was 79 d for patients with positive 18F-FDG PET, whereas the disease-free proportion was 89% at 1,861 d (after treatment) for patients with negative 18F-FDG PET (P < 0.0001). The hazard ratio (slope ratio) between the 2 curves was 19.5.
|
|
SUVs of the most active lesion in patients with positive 18F-FDG PET ranged from 1.78 to 14.09. Only 1 patient with a positive PET study could not be included in the analysis. The lower part of his PET study, which demonstrated recurrent disease in the right lower limb, was acquired without attenuation correction. The mean SUV of all patients was 6.06 ± 4.21. The median SUV was 4.16 and was chosen as a cutoff value to create 2 groups: a high-SUV (>4.16) group and a low-SUV group (
4.16). Using the Kaplan–Meier and the log-rank tests, no statistical difference (P = 0.6656) was observed between the 2 groups relating to the rapidity of relapse. The patient whose SUV of the most active lesion was 4.16 was arbitrarily included in the low-SUV group. Including this patient in the high-SUV group rather than the low-SUV group did not generate any significant difference between the 2 curves.
Comparison Between 18F-FDG PET and CT
Sensitivity, specificity, and negative and positive predictive values for 18F-FDG PET and CT are listed in Table 3. In 48 patients, the sensitivity of PET was 79% and the specificity 97%, with an overall accuracy of 92%. In a subset of 32 patients who also had CT after chemotherapy, the sensitivity of CT was 83%, the specificity 40%, and the accuracy 56%. Note that 16 patients had PET, but not CT, as the sole imaging modality at the end of chemotherapy. If we excluded those patients from the calculations, the sensitivity of PET would have been 92% and the specificity 95%, with an overall accuracy of 94% (30/32). A comparison between patients with positive 18F-FDG PET and patients with positive CT (Fig. 3) demonstrates a significantly lower disease-free proportion for patients with positive 18F-FDG PET (by the Kaplan–Meier method): 8% disease-free proportion at 15 mo with positive 18F-FDG PET and 52% disease-free proportion at 38 mo with positive CT (P = 0.0046). However, no significant difference was observed between disease-free proportions for patients with negative 18F-FDG PET or negative CT: Disease-free proportions at 62 mo after chemotherapy were 89% and 80%, respectively (P = 0.2075).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Anatomic imaging with CT, in pretherapy assessment, has a high sensitivity (91%) and specificity (100%) (19). In posttherapy assessment, a study published by Cremerius et al., in which 72 patients were included, reported a sensitivity of 84% and a specificity of 31% (20). This dramatic decrease in specificity in assessment after therapy results from the use of lymph node size as a criterion to determine whether active lymphoma is still present. This shortcoming of morphologic imaging reflects the longer time period required for nodes to normalize in size after therapy compared with glucose metabolic changes observed with 18F-FDG PET. In the neck, in the thorax, and at the aortic bifurcation, nodes of >1 cm are considered abnormal, whereas in the pelvis and the upper retroperitoneum the upper limit is 0.8 cm (21). In our study, we have used an upper limit of 1 cm in all regions of the body, which could lead to an overestimation of specificity and an underestimation of sensitivity for CT. In spite of that, we observed a rather poor specificity (40%) with a positive predictive value of 45%. Hence, positive CT after chemotherapy in patients with HD can hardly be used as a reliable diagnostic tool to predict relapses and the need for therapy consolidation. In comparison, 18F-FDG PET had a much higher specificity (97%) and positive predictive value (92%) and a significantly shorter median disease-free interval (79 d). The sensitivity of CT appeared to be higher (83%) than its specificity, with a negative predictive value of 80%. These values were not statistically different from those obtained with 18F-FDG PET (sensitivity, 79%; negative predictive value, 92%). However, the overall diagnostic accuracy clearly remained in favor of 18F-FDG PET (92% vs. 56%).
Not all patients had CT concurrent with PET. This might be due to the fact that referring physicians were satisfied with the PET results and did not feel compelled to order an additional diagnostic procedure. This raises the issue of possible bias in the CT data. If all patients had undergone CT, the comparison would have been more robust by increasing the confidence in the CT data. Excluding the patients who had PET but not CT did not change significantly the overall accuracy of the PET data. In the subgroup of patients without concurrent CT data, 2 of 16 patients relapsed on follow-up. At best, if CT was accurate in all of those patients (an unlikely event), the specificity of CT would remain relatively poor at 65% and the sensitivity would not change significantly. Therefore, the specificity of CT would still be much lower than the specificity achieved with PET. In all cases of falsely positive or negative 18F-FDG PET, CT, when available, did not provide additional information in helping to predict residual disease or relapse. In a posttreatment context, it appears that 18F-FDG PET has a better diagnostic accuracy than CT.
All false-negative 18F-FDG PET studies (3/3) that we observed occurred in patients who underwent their PET study within the first 49 d after the end of chemotherapy. Sensitivity and negative predictive values were both equal to 100% in the subgroup of patients who had their PET study later after the end of therapy. Possible explanations for this observation could be the important decrease in lesion size after chemotherapy to a point where the recovery of radioactivity concentration is severely impaired by partial-volume effects. Römer et al. also showed in a group of 11 patients with NHL that the metabolic rate of 18F-FDG in tumor could be significantly altered by chemotherapy treatments independently of tumor size (22). This change was visible as early as 7 d after initiation of treatment but was a better predictor of long-term clinical outcome at 42 d. There is also evidence, in different tumor types, that this effect can be observed after completion of chemotherapy (23,24). In addition, no imaging modality can detect microscopic residual disease and this might be the primary explanation for the negative PET studies, given that the time lag between the negative PET scans and the relapse was close to 1 y in all cases. The small numbers of patients with negative PET studies that will relapse preclude definite conclusions about the optimal timing for PET studies after chemotherapy.
SUV measurements of the most active lesion did not appear to be a useful predictor of long-term clinical outcome in patients with positive 18F-FDG PET studies after treatment. The disease-free interval was not significantly different between the high- and the low-SUV groups. In a published study that addressed the value of SUV parameters in posttherapy HD, Dittman et al. reported that the SUV was not superior to visual assessment to detect the presence of residual disease (11). However, it is interesting to note that in the only false-positive PET study we obtained, the SUV measurement was the lowest of all positive PET studies. Thus, in doubtful cases or when a false-positive result is strongly suspected, a cautious approach would be to obtain either a biopsy confirmation or a close repeated PET scan, as occurred in that particular case.
Only 2 patients with stage I disease were included in this study. Considering the excellent prognosis and the somewhat different treatment protocols used in patients with stage I disease, they can be viewed as a specific subgroup. The underrepresentation of this subgroup makes it difficult to extrapolate our results to patients with early-stage HD. In our data, we found a true relapse proportion of 9% (2/23) in stage II, 27% (4/15) in stage III, and 50% (4/8) in stage IV. Most likely, in patients with stage I, one would expect to observe a very low relapse rate associated to a high negative predictive value with 18F-FDG PET, reflecting the favorable outcome associated with early-stage HD (10,25).
All retrospective studies are subject to potential biases in patient selection and study interpretation and can be limited because intraobserver reproducibility cannot easily be assessed. Another limitation is that the PET and CT results may influence the outcome and, therefore, overestimate accuracy of these tests. However, oncologists rarely rely on a single modality in the decision-making process, and an integration of several data sources with confirmatory studies is generally used to guide therapy. Nevertheless, this remains a significant limitation of all retrospective imaging studies that cannot be overcome without a prospective trial in which the oncologist is unaware of the test results. In addition, it may sometimes also be difficult to identify what subset of the true incident patient population was studied, because inclusion criteria are defined a posteriori. Despite a small number of patients with early-stage disease, we believe that the patient population we studied fairly reflects the presentation and evolution of HD in our area because the majority of patients with HD undergo PET scanning after treatment. Furthermore, our hospital is the only center within a large geographic area to provide care for these patients, ensuring consistency in referral patterns as well as therapeutic and diagnostic approaches. The CT data would be subject to the same limitations as the PET data; however, our study clearly shows that morphologic imaging has a very poor specificity to predict an unfavorable outcome after treatment. Finally, not all patients in this study were examined with 18F-FDG PET immediately after chemotherapy. The rate of residual disease in patients studied >90 d after chemotherapy was lower than the rate of patients studied at 0–90 d (19% and 37%, respectively). This reduces concerns that patients imaged at the longer intervals might have been referred because of signs or symptoms suggesting recurrence—therefore, introducing a selection bias that could favor the test under assessment.
Radiation therapy was used in a substantial proportion of patients (31/48), not always before PET, because some patients had radiation treatment after a positive PET result. Despite this, the accuracy of PET was not affected by the timing of radiation therapy. In patients who had radiation therapy after PET, only 4 patients had residual disease on PET and all of these patients eventually relapsed. From these very limited data, one should not construe that radiation therapy is ineffective in this setting because our study was not designed to assess the efficacy of this treatment. Chemotherapy is highly effective in HD to eradicate malignant cells. This is probably the main explanation for the lack of observed influence of radiation therapy on PET accuracy, because most patients with a favorable outcome responded well to chemotherapy. Refractory HD is extremely difficult to treat, whether by chemotherapy or radiation therapy, alone or in combination, and this probably explains the relapse in PET-positive patients who underwent radiation therapy afterward.
A potential research avenue that remains to be explored is determination of the 18F-FDG PET value as a prognostic indicator very early after initiation of treatment. This information could guide important therapeutic decisions, such as decreasing the number of chemotherapy cycles administered to a patient or switching to a more aggressive protocol. Available published data originating from NHL patients in a limited number of patients suggest that this approach is promising (22). Furthermore, Kostakoglu et al. recently published a mixed series of 13 HD and 17 NHL patients with rapid tumor response assessment after only 1 cycle of chemotherapy using a coincidence camera and achieved good results to predict treatment failure with early scanning (26). This approach warrants a thorough comparison with scans obtained a few weeks after the end of chemotherapy.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
For correspondence or reprints contact: François Bénard, MD, Clinical Research Center, Centre Hospitalier Universitaire de Sherbrooke/Hôpital Fleurimont, 3001, 12th Ave. N., Sherbrooke, Quebec, Canada, J1H 5N4.
E-mail: Francois.Benard{at}USherbrooke.ca
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D.J. Sher, P.M. Mauch, A. Van Den Abbeele, A.S. LaCasce, J. Czerminski, and A.K. Ng Prognostic significance of mid- and post-ABVD PET imaging in Hodgkin's lymphoma: the importance of involved-field radiotherapy Ann. Onc., June 18, 2009; (2009) mdp071v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hutchings and S. F. Barrington PET/CT for Therapy Response Assessment in Lymphoma J. Nucl. Med., May 1, 2009; 50(Suppl_1): 21S - 30S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Terasawa, T. Nihashi, T. Hotta, and H. Nagai 18F-FDG PET for Posttherapy Assessment of Hodgkin's Disease and Aggressive Non-Hodgkin's Lymphoma: A Systematic Review J. Nucl. Med., January 1, 2008; 49(1): 13 - 21. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Lin, E. Itti, C. Haioun, Y. Petegnief, A. Luciani, J. Dupuis, G. Paone, J.-N. Talbot, A. Rahmouni, and M. Meignan Early 18F-FDG PET for Prediction of Prognosis in Patients with Diffuse Large B-Cell Lymphoma: SUV-Based Assessment Versus Visual Analysis J. Nucl. Med., October 1, 2007; 48(10): 1626 - 1632. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Advani, L. Maeda, P. Lavori, A. Quon, R. Hoppe, S. Breslin, S. A. Rosenberg, and S. J. Horning Impact of Positive Positron Emission Tomography on Prediction of Freedom From Progression After Stanford V Chemotherapy in Hodgkin's Disease J. Clin. Oncol., September 1, 2007; 25(25): 3902 - 3907. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Dann, R. Bar-Shalom, A. Tamir, N. Haim, M. Ben-Shachar, I. Avivi, T. Zuckerman, M. Kirschbaum, O. Goor, D. Libster, et al. Risk-adapted BEACOPP regimen can reduce the cumulative dose of chemotherapy for standard and high-risk Hodgkin lymphoma with no impairment of outcome Blood, February 1, 2007; 109(3): 905 - 909. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Reinhardt, C. Herkel, C. Altehoefer, J. Finke, and E. Moser Computed tomography and 18F-FDG positron emission tomography for therapy control of Hodgkin's and non-Hodgkin's lymphoma patients: when do we really need FDG-PET? Ann. Onc., September 1, 2005; 16(9): 1524 - 1529. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. P. Beal, H. W. Yeung, and J. Yahalom FDG-PET scanning for detection and staging of extranodal marginal zone lymphomas of the MALT type: a report of 42 cases Ann. Onc., March 1, 2005; 16(3): 473 - 480. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Kumar, Y. Xiu, S. Potenta, A. Mavi, H. Zhuang, J. Q. Yu, T. Dhurairaj, S. Dadparvar, and A. Alavi 18F-FDG PET for Evaluation of the Treatment Response in Patients with Gastrointestinal Tract Lymphomas J. Nucl. Med., November 1, 2004; 45(11): 1796 - 1803. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Torabi, S. L. Aquino, and M. G. Harisinghani Current Concepts in Lymph Node Imaging J. Nucl. Med., September 1, 2004; 45(9): 1509 - 1518. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Meyer, R. F. Ambinder, and S. Stroobants Hodgkin's Lymphoma: Evolving Concepts with Implications for Practice Hematology, January 1, 2004; 2004(1): 184 - 202. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| JOURNAL OF NUCLEAR MEDICINE TECHNOLOGY | THE JOURNAL OF NUCLEAR MEDICINE |
