|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Ahmanson Biological Imaging Clinic (Drs. Filmont, Czernin, Silverman, Quon, and Phelps, and Ms. Yap), Department of Molecular and Medical Pharmacology; and Department of Medicine (Dr. Emmanouilides), Division of Hematology/Oncology, UCLA School of Medicine, Los Angeles, CA.
Correspondence to: Johannes Czernin, MD, UCLA School of Medicine, Nuclear Medicine, AR 277A CHS, 10833 Le Conte Ave, Los Angeles, CA 90095-6942; e-mail: jczernin{at}mednet.ucla.edu
| Abstract |
|---|
|
|
|---|
Patients and methods: Forty-three patients with lymphoma who underwent ASCT with FDG-PET evaluation were studied. Group 1 (n = 20) patients (6 patients with Hodgkin disease [HD], and 14 patients with non-Hodgkin lymphoma [NHL]) underwent PET 2 to 5 weeks after initiation of salvage chemotherapy, prior to ASCT. Group 2 (n = 23) patients (6 patients with HD, and 17 patients with NHL) underwent PET within a median interval of 2.4 months (range, 2 to 6 months) after ASCT.
Measurements and results: Study end points were complete remission, relapse, or death. In group 1, 8 of 20 patients (40%) were disease free after a median follow-up of 13.3 months; 12 patients relapsed or died. PET findings were true-negative in 7 of 8 patients and true-positive in 11 of 12 patients who relapsed after ASCT. In group 2, 9 of 23 patients (39%) were disease free after a median follow-up of 16.5-months; 14 patients relapsed. PET findings were true-negative in 8 of 9 patients and true-positive in 13 of 14 patients who relapsed. Positive and negative predictive values of PET were 92% and 88% (group 1) and 93% and 89% (group 2), respectively. Predictive accuracy values of PET were 90% and 91% for group 1 and group 2, respectively, vs 58% and 67% for CT (p < 0.05).
Conclusions: PET findings but not CT results were strongly correlated with disease-free survival (p < 0.01). Our results show that FDG-PET can be used to predict the post-ASCT outcome of lymphoma patients with high accuracy.
Key Words: aggressive lymphoma autologous stem-cell transplantation positron emission tomography prognostic value
| Introduction |
|---|
|
|
|---|
The success rate of ASCT depends on a number of factors.8 These include age, concurrent medical conditions, the Karnofsky global performance status, the histologic type of lymphoma, and the number of extranodal disease sites.9 However, the most important factor influencing the success rate of ASCT is the tumor response to salvage chemotherapy.6 7 10 Currently, this response is evaluated by conventional imaging modalities such as CT or MRI. Further prognostic information following ASCT is obtained from a variety of imaging tests.
Whole-body positron emission tomography (PET) using F-18 fluorodeoxyglucose (FDG) restages lymphoma after standard chemotherapy with a higher accuracy than conventional imaging.11 12 13 It is unknown, however, whether FDG-PET findings can be used to predict the outcome of patients with lymphoma during salvage chemotherapy or after ASCT. The purpose of this study was to compare the predictive values of CT and PET in patients with lymphoma who underwent salvage chemotherapy and ASCT.
| Materials and Methods |
|---|
|
|
|---|
|
Group 2: In group 2 (n = 23) [6 patients with HD, and 17 patients with NHL], FDG-PET was performed within a median interval of 2.4 months (range, 2 to 6 months) after ASCT. There were 12 male and 11 female patients, with a median age of 42 years (range, 17 to 65 years).
Patients in groups 1 and 2 underwent PET and CT imaging within median intervals of 5 days and 19 days. The two study groups, consisting of different patients, had similar demographic characteristics (Table 1) and the same incidence of clinical relapse (60%).
Determination of Clinical Outcome
All patients were followed up for at least 6 months after ASCT. The treating oncologist (C.E.) classified patient outcome based on conventional imaging findings (n = 43), biopsy (n = 14), and clinical examination (n = 43) as complete remission, disease progression/relapse, or death related to the primary disease. Follow-up FDG-PET studies were excluded from the clinical outcome assessment. FDG-PET and CT findings were later correlated with clinical outcome at the time of the last known follow-up in patients who remained in complete remission, and at the first evidence of relapse determined by conventional imaging and/or biopsy in patients with recurrent disease.
Imaging Protocol
Patients fasted for at least 6 h before being injected with 370 to 550 MBq of FDG. Serum glucose levels averaged 93 ± 21 mg/dL at the time of injection (± SD). None of the patients received IV insulin prior to PET. PET imaging was performed with an ECAT EXACT or HR+ system (CTI/Siemens; Knoxville, TN).14
15
The standard clinical imaging protocol started 45 to 60 min after tracer injection. Whole-body images were acquired over six to eight bed positions. Attenuation correction was performed in 22 patients (51%). Corrected images were reconstructed using iterative reconstruction algorithms. The remaining 21 patients were studied prior the implementation of these procedures. Standard filtered back-projection was employed in these patients.
Image Interpretation
PET images were re-read by an experienced nuclear medicine physician (J.C.) without knowledge of conventional imaging findings and/or clinical history. Written reports were used to establish the predictive value of CT imaging. PET and CT findings were classified as positive or negative for residual/recurrent disease. It should be noted that all patients underwent CT scanning; however, CT reports were not available in four patients (two patients from group 1, and two patients from group 2).
Statistical Analysis
Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated by standard methods, using the clinical outcome as the "gold standard." Categorical data were examined using the
2 test; the Fisher exact test was used if expected cell counts were < 5. The probability of disease-free survival was computed by Kaplan-Meier analysis. The log-rank test was used to evaluate the difference between Kaplan-Meier curves; p < 0.05 was considered significant.
| Results |
|---|
|
|
|---|
|
Biopsy was performed in seven patients. PET findings were true-positive in two patients and true-negative in five of these patients (in whom conventional imaging had suggested residual disease).
Group 2
Nine of 23 patients (39%) who underwent FDG-PET early after ASCT were disease free after a median follow-up of 16.5 ± 11 months. The remaining 14 patients relapsed after a median follow-up of 7.3 months. FDG-PET findings were true-negative in 8 of 9 patients and true-positive in 13 of 14 patients who relapsed after ASCT. The predictive sensitivity and specificity of PET were 93% and 89%, respectively. The PPV, NPV, and predictive accuracy were 93%, 89%, and 91%, respectively. PET-negative patients had a significantly better clinical outcome than PET-positive patients (p = 0.003; Fig 1
, bottom, B). The predictive sensitivity of CT was similar to that of PET at 92% (p = not significant), but its specificity was only 25% (p < 0.05). The predictive accuracy of CT was also significantly lower than that of PET (67%; p < 0.05). The CT findings were not useful for predicting patient outcome (p = 0.55; Fig 1
, bottom, B).
One false-positive PET finding occurred in a patient with avascular necrosis of the left hip misinterpreted as bone involvement by PET. The patient underwent hip replacement and remained in complete remission 9 months after ASCT. The false-negative PET finding occurred in a patient in whom mild linearly increased tracer uptake was interpreted as radiation pneumonitis. In this patient, disease progression occurred 9 months later, as evidenced by new mediastinal and retroperitoneal lymph nodes on CT. The patient was further scheduled to undergo an allogeneic stem cell transplantation.
Biopsy was performed in seven patients. PET predicted the biopsy findings correctly in all seven patients. PET findings were true-positive in four patients and true-negative in three of the seven patients (in whom conventional imaging had suggested residual disease).
| Discussion |
|---|
|
|
|---|
Salvage cytoreductive chemotherapy, usually administered for two cycles, followed by HDC supported by ASCT, successfully treats aggressive disease in approximately 40% of patients with lymphoma.16 The success of this regimen is currently determined by prospective follow-up of patients with clinical examination, serum markers, and imaging tests such as CT, MRI, and whole-body 67Ga scanning.6 7 10 The specificity of these tests for differentiating between benign posttreatment tissue alterations and residual/recurrent disease is, however, low at approximately 50%.17 18 19
FDG-PET is more accurate than CT, MRI, or 67Ga imaging for restaging and treatment monitoring of lymphoma after conventional therapy.12 20 21 22 The value of PET for predicting the clinical outcome of patients with lymphoma who undergo standard treatment is high and is superior to that of anatomic imaging.11 12 13 23 Accordingly, PET yielded a higher NPV than CT in the current population of patients with aggressive lymphoma (88% vs 50%, p < 0.05). Importantly, this prognostic information was already derived after a few weeks of salvage chemotherapy prior to ASCT. No additional gain in prognostic accuracy was derived from a posttransplant PET study.
It could be argued that early identification of treatment failure has only minimal therapeutic consequences. This is because HDC supported by ASCT, regardless of whether tumor response to treatment is complete or partial, may be considered the only available therapeutic option. This is the case even in patients without satisfactory response to salvage treatment because disease progression might be delayed; however, new therapeutic concepts such as nonmyeloablative allogeneic transplants (minitransplant) are now emerging in the field of lymphoma treatment and are about to change this paradigm.24 25 26 These therapeutic strategies use lower and less toxic doses of chemotherapy or radiation to suppress the patients immune system and allow for donor cell engraftment. The success of nonmyeloablative transplant are thought to be increased after the cytoreduction achieved by HDC and ASCT.27
Group 1 of the current study population was studied early after cytoreductive treatment, ie, 2 to 5 weeks after initiation of therapy. Even at this early point in time, PET was predictive of patient outcome. This is consistent with the notion that PET predicts the tumor response to treatment after just a few cycles of chemotherapy in patients undergoing standard chemotherapy for lymphoma.13 23 Thus, FDG-PET imaging might be useful to identify patients who are likely to achieve a satisfactory treatment response and for differentiating them from those who do not remain in remission after ASCT. The latter might be candidates for nonmyeloablative allogeneic transplant following maximum tumor reduction achieved by HDC and ASCT, or other experimental approaches.28
Several limitations of this study warrant discussion. First, the population included patients with HD and NHL. It is well known that these disease entities carry different prognoses; however, relapse rates after ASCT are similar for the aggressive forms of these diseases. Thus, inclusion of patients with both diseases is justified.
As another limitation, patients were studied retrospectively. The potential bias of such design is obvious. In this study, 60% of all patients who underwent salvage chemotherapy and ASCT between 1998 and 2001 at our institution were enrolled; 6 of the 12 patients (50%) with HD remained in complete remission while 21 of 31 patients (67%) with NHL relapsed. These relapse rates after ASCT are consistent with those reported in the current literature1 7 29 30 ; therefore, the current population likely represents the typical population of patients with aggressive lymphoma.
The time interval between ASCT and FDG-PET evaluation for patients in group 1 ranged from 2 to 6 months (median interval, 2.4 months), with one patient being evaluated by PET 6 months after receiving ASCT. While a 6-month time interval may be sufficiently long for some aggressive lymphomas to relapse, the retrospective nature of this study did not permit regulation of the interval between ASCT and PET evaluation.
PET images were re-read in a blinded fashion, while written reports of clinical CT studies were used to establish the predictive values of these tests. This approach was chosen because clinical PET images are usually obtained after CT images and are thus interpreted with knowledge of the CT findings. Using written reports to establish the clinical stage would have put PET at an inappropriate advantage relative to conventional imaging. It should, however, be mentioned that the written clinical PET findings were concordant with the blinded interpretation in 100% of the patients.
Written reports were used to establish the predictive value of CT imaging. Some patients might have exhibited partial responses to treatment; however, all written reports clearly stated whether findings were positive or negative for residual/recurrent disease. It was this conclusion of individual reports that was used to establish the predictive value of CT. Moreover, such a classification would be more consistent with that of PET since all PET studies were classified as either positive or negative for disease. As a technical limitation, about 50% of PET images were acquired without attenuation correction. This is because attenuation correction and iterative image reconstruction were not routinely performed in our clinic before July 2000; however, despite the lack of attenuation correction, the predictive accuracy of FDG-PET was high at approximately 90% in the current study. Further, the predictive accuracy of corrected and noncorrected PET images were identical. This supports the notion that attenuation correction does not result in additional improvements in the diagnostic accuracy of PET.31 As a further limitation, no comparison between whole-body 67Ga and PET was performed because FDG-PET has largely replaced whole-body 67Ga in our institution; therefore, no meaningful comparison between these two tests could be performed.
| Conclusion |
|---|
|
|
|---|
| Footnotes |
|---|
Received for publication August 15, 2002. Accepted for publication January 22, 2003.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
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 |