Fax: (011) 31-204442601
18F-fluoro-deoxyglucose positron emission tomography in assessment of myeloma-related bone disease: A systematic review
Version of Record online: 1 SEP 2011
Copyright © 2011 American Cancer Society
Volume 118, Issue 8, pages 1971–1981, 15 April 2012
How to Cite
van Lammeren-Venema, D., Regelink, J. C., Riphagen, I. I., Zweegman, S., Hoekstra, O. S. and Zijlstra, J. M. (2012), 18F-fluoro-deoxyglucose positron emission tomography in assessment of myeloma-related bone disease: A systematic review. Cancer, 118: 1971–1981. doi: 10.1002/cncr.26467
- Issue online: 6 APR 2012
- Version of Record online: 1 SEP 2011
- Manuscript Accepted: 7 JUL 2011
- Manuscript Revised: 1 JUL 2011
- Manuscript Received: 28 APR 2011
The goal of this study was to conduct a comparative analysis of whole body X-ray (WBXR) and 18F-fluoro-deoxyglucose positron emission tomography (18FDG PET) in staging and response assessment of multiple myeloma.
We performed a systematic review of studies comparing 18FDG PET with WBXR and/or magnetic resonance imaging in terms of sensitivity for myeloma-related bone disease at staging and during follow-up.
Eighteen studies involving 798 patients met the inclusion criteria. The mean Quality Assessment of Diagnostic Accuracy Studies (QUADAS) score, expressed as a percentage of the maximum score, was 61%. In 7 studies (n = 242 patients), concordance assessment between WBXR and 18FDG PET scan was possible, showing a higher sensitivity of the 18FDG PET in the detection of myeloma bone lesions in 6 studies. The only study reporting on the prognostic value of 18FDG PET at staging found that the number of FDG-avid focal lesions was an independent prognostic parameter. In addition, the limited studies on response monitoring showed that normalization of 18FDG PET during treatment correlated with a superior clinical outcome.
In general, 18FDG PET has a superior sensitivity for myeloma bone lesions compared with WBXR. Future studies have to validate the additive value of myeloma-related bone disease detected on 18FDG PET–computed tomography (CT) in predicting outcome. Response monitoring with the use of 18FDG PET-CT during treatment is promising, allowing more precise prediction of prognosis compared with the standard response monitoring. In view of the expanding treatment options for multiple myeloma, this may provide important information for treatment decisions in the future. Cancer 2012. © 2011 American Cancer Society.
Bone disease is common in patients with multiple myeloma, developing in 90% of patients during the course of the disease. Due to the resulting pain, spinal cord compression and hypercalcemia, this disease has a major impact on quality of life. Accurate investigation of bone disease in multiple myeloma is needed to prevent complications through initiation of anti–multiple myeloma therapy. Recently, the International Myeloma Working Group (IMWG) reported a consensus statement on the role of imaging techniques in multiple myeloma1 in which whole body X-ray (WBXR) was considered the gold standard in initial staging. However, the sensitivity of WBXR for bone disease is low, because at least 30% of trabecular bone substance must be lost to give rise to visible lytic lesions.2 This finding underscores the need for new, more sensitive imaging techniques. In general, over the last few years, small studies in heterogeneous patient groups using different techniques have reported an increased sensitivity of such new techniques. In addition, there were some reports on the prognostic value of 18F-fluoro-deoxyglucose positron emission tomography (18FDG PET)–computed tomography (CT) in newly diagnosed patients that appeared superior to WBXR.3, 4 Moreover, the predictive value of normalized 18FDG uptake during therapy was associated with improved survival outcomes. This finding is intriguing because in contrast to CT and magnetic resonance imaging (MRI), 18FDG PET or PET-CT quantifies metabolic function due to increased glycolysis and thereby active disease.5, 6 In addition, information on extramedullary disease becomes available by 18FDG PET or PET-CT. In contrast, MRI will generally only be performed on the spine, thereby lacking this information.
Thus, 18FDG PET or PET-CT is a promising diagnostic tool in multiple myeloma. However, in the recent IMWG consensus statement, 18FDG PET or PET-CT was not recommended for routine use in the management of myeloma patients, although the relevant literature was available but not reviewed in a systematic manner. Moreover, the large prospective trial of Bartel et al3 has been published after this consensus statement.
Because of this omission, we performed a systematic review of the literature to determine the accuracy of dedicated (full ring) PET using 18FDG at staging. In addition, studies on the value of 18FDG PET or PET-CT in response assessment were analyzed. To this end, all relevant scientific reports were identified using a comprehensive search strategy. Accepted methodological standards for the evaluation of diagnostic tests were applied. We here summarize the existing data on the relevance of PET(-CT) scanning in multiple myeloma as a first step toward the development of guidelines for the effective use of 18FDG PET-CT.
MATERIALS AND METHODS
Literature Search for the Identification of Studies
A search of the bibliographic databases PubMed/MEDLINE and EMBASE (through EMBASE.com) was conducted up to December 2010, without language restrictions. The search strategy for the identification of primary studies regarding diagnostic tests was run in conjunction with a specific search for PET and multiple myeloma, adapted for each database. All searches were performed using controlled indexing terms (MeSH in MEDLINE and EMTree in Embase) and free text words. A detailed account of the searches can be obtained from the first author. We augmented this search by cross-referencing. Unpublished data and conference proceedings were not included in the review.7
The criteria for inclusion of studies were 1) multiple myeloma or plasmacytoma according to Salmon-Durie criteria and 2) the use of dedicated 18FDG PET scan complementary to standard diagnostic tools. Exclusion criteria were 1) use of radiopharmaceuticals other than 18FDG; 2) animal studies; 3) abstracts, reviews, case reports, editorials, and comments; and 4) publication language other than English.
Two reviewers (D.v.L. and J.M.Z.) independently selected the studies for possible inclusion by checking titles and abstracts. Full texts of all studies considered (potentially) eligible were retrieved; the final decision was made on the full article. Disagreement was resolved by consensus.
Methodological Quality Assessment
For the results of an accuracy study to be valid, an independent, blind comparison with a valid reference test to avoid review bias is essential. This reference test has to be measured in all patients independent of the results of the PET scan to avoid verification and work-up bias. The reference test has to be applied in a standardized manner.
In oncology, histological proof of presence or absence of viable tumor is generally considered an accurate reference test. However, in a potentially widespread disease such as multiple myeloma, this approach is not possible; biopsy of all possible myeloma bone lesions is impossible, and because conventional staging systems may underestimate the extent of the disease, it would only be partially representative. Because of this limitation, we chose the best alternative and compared the 18FDG PET or PET-CT scan to WBXR and, if performed, MRI of the spine and pelvis (T1-weighted, T2-weighted with and without fat suppression and short inversion time inversion recovery).
The same 2 reviewers independently assessed the methodological quality of the selected studies using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) criteria.8 Some items were modified for this specific review (Table 1). Each item was scored as “positive” or “negative. ” If the information on a specific item was insufficient, it was scored as “negative. ” Standard performance of the 18FDG PET or PET-CT scan was scored as “positive” if the type of PET scanner, the scanned trajectory and the criteria of PET positivity were described. If semiquantitative measurement of 18FDG-uptake (using standardized uptake values [SUVs]) was used in addition to visual assessment, the dose of 18FDG, the time interval between injection and scanning, and the methods of image reconstruction had to be mentioned.
|1. Was the spectrum of patients representative of the patients who will receive the test in practice?||MM/plasmacytoma (plasmacelldyscrasias)||Other|
|2. Were selection criteria clearly described?||Inclusion and exclusion criteria described||Not described|
|3. Is the reference standard likely to correctly classify the target condition?||WBXR (=specificity) and MRI spine and pelvis (=sensitivity) (and histology)||Other tests|
|4. Is the time period between reference standard and index test short enough to be reasonably sure that the target condition did not change between the two tests?||<6 weeks||>6 weeks|
|5. Did the whole sample or a random selection of the sample receive verification using a reference standard of diagnosis?||Diagnostic criteria for MM (ie, Salmon-Durie criteria)||Not done|
|6. Did patients receive the same reference standard regardless of the index test result?||Yes||No|
|7. Was the reference standard independent of the index test (ie, the index test did not form part of the reference standard)?||Yes||No|
|8. Was the execution of the index test described in sufficient detail to permit replication of the test?||Type of scanner (PET vs PET-CT, full ring)||No|
|Criteria for positivity of PET described|
|Correction for body weight, glucosea|
|Postreconstruction image resolutiona|
|9. Was the execution of the reference standard described in sufficient detail to permit its replication?||WBXR: all bones described||WBXR: bones not described|
|MRI: type of scanner, scanned trajectory, criteria positivity described||MRI: not described|
|10. Were the index test results interpreted without knowledge of the results of the reference standard?||Yes||No|
|11. Were the reference standard results interpreted without knowledge of the results of the index test?||Yes||No|
|12. Were the same clinical data available when test results were interpreted as would be available when the test is used in practice?||Salmon-Durie criteria||No SD criteria|
|13. Were uninterpretable/intermediate test results reported?||Yes||No (also when not mentioned)|
|14. Were withdrawals from the study explained?||Yes||No (also when not mentioned)|
First, to determine the value of the 18FDG PET or PET-CT scan in the detection of myeloma-related bone disease at diagnosis, we performed a concordance assessment. The results of the 18FDG PET or PET-CT scan were compared with the standard diagnostic tools (ie, the WBXR) and, if performed, the MRI of the spine and/or pelvis. Second, we summarized the existing data on response assessment after first-line treatment. Finally, we searched the literature for existing data on the value of the 18FDG PET or PET-CT scan in evaluation and follow-up of nonsecreting and extramedullary multiple myeloma.
The search strategy yielded 274 publications in EMBASE and in MEDLINE. Of these studies, 229 were excluded after reviewing the information provided in the title and abstract. Reviewing the full articles of the 45 remaining studies resulted in exclusion of another 27 articles due to an overlap in study population (1 study) not using the 18FDG PET or PET-CT scan complementary to conventional staging (1 study), being a review article, case report, or editorial (16 studies), using another pharmaceutical (5 studies) or language restriction (4 studies). Finally, 18 studies involving 798 patients were included in this systematic review (Figure 1).3, 4, 9-25 There was no disagreement between the reviewers regarding the inclusion of the articles. The characteristics of the included studies are presented in Table 2. The total number of patients per study ranged from 6 to 303, and the patients' age ranged from 23 to 85 years. Two studies included only plasmacytoma patients, 13 included only multiple myeloma patients, and the remaining studies included a combination of multiple myeloma and plasmacytoma patients.
|Study||Study Design||No of Patients||Diagnosis||Reference Test||Endpoints|
|Durie, 2002||Retrospective, 18FDG PET||66||MM (52), MGUS amyloidosis (14)||WBXR (all), MRI, CT, PA||Accuracy|
|Jadvar, 2002||Prospective, 18FDG PET||6||MM||WBXR (6), CT spine (2), MRI spine (4), bone scan(3)||Accuracy|
|Schirrmeister, 2002||Prospective, 18FDG PET||43||MM (28), plasmacytoma (15)||WBXR (43), bone scan (21), CT (21), MRI spine (19), extravertebral (13)||Accuracy|
|Schirrmeister, 2003||Prospective, 18FDG PET||15||Plasmacytoma (15)||WBXR(all)||Accuracy|
|Mileshkin, 2004||Retrospective, 18FDG PET||69||MM (62), MGUS (3), plasmacytoma (4)||WBXR (all), additional sites confirmed with PA+/- CT or MRI 10/11; 99mTc-MIBI||Accuracy|
|Bredella, 2005||Retrospective, 18FDG PET||13||MM||MRI (16), CT (4), WBXR (6)||Accuracy|
|Hung, 2005||Prospective, 18FDG PET||12||MM||WBXR (all)||Accuracy|
|Breyer, 2006||Retrospective, 18FDG PET-CT||16||MM/MGUS (1)||CT(25), MRI (22), WBXR(13)||Accuracy, change in therapy|
|Adam, 2007||Retrospective, 18FDG PET||49||MM (39), MGUS (6), plasmacytoma (2)||WBXR (all)||Accuracy|
|Fonti, 2007||Prospective, 18FDG PET-CT||33||MM||MRI and 99mTc-MIBI(all)||Accuracy|
|Hur, 2007||Retrospective, 18FDG PET||10||MM||MRCT and MRI(all)||Accuracy|
|Zamagni, 2007||Prospective, 18FDG PET-CT||46||MM||MRI and WBXR(all)||Accuracy|
|Hur, 2008||Retrospective, 18FDG PET||22||MM||MRI (all)||Accuracy|
|Salaun, 2008||Prospective, 18FDG PET-CT||24||Plasmacytoma||MRI (all)||Accuracy|
|Bartel, 2009||Prospective, 18FDG PET-CT||303||MM||WBXR and MRI (all)||Prognostic value|
|Shortt, 2009||Prospective, 18FDG PET-CT||24||MM||MRI (all)||Accuracy|
|Dimitrakopoulou- Strauss, 2009||Prospective, 18FDG PET||19||MM||MRI (all)||Prospective value|
|Castellani, 2010||Prospective, 18FDG PET||18||MM||WBXR (all)||Prognostic value|
The definition of a positive PET scan differed in the included studies (Table 3), and 9 studies used semiquantitative measurement of 18FDG uptake in addition to visual assessment. The reference test consisted of WBXR in 11 studies with 10 studies with complementary MRI of spine and pelvis. Incidentally, biopsies were performed when the diagnosis was plasmacytoma.
|Breyer, 2006||All focal areas of abnormal 18FDG uptake were correlated with the other imaging studies to determine clinical significance. A SUV ≥2.5 was considered to be abnormal and to indicate a site of active disease.|
|Fonti, 2007||All focal areas visible on at least 2 contiguous PET slices showing a maximal SUV of ≥2.5 and corresponding CT abnormalities not attributable to benign bone pathologies.|
|Zamagni, 2007||All nonphysiologic areas of increases 18FDG uptake if SUV max based on body weight was >2.5 and a lytic bone lesion was recognized on the corresponding CT images.|
|Salaun, 2008||Not mentioned|
|Bartel, 2009||Hypermetabolic FLs (PET-FL) were defined as being more intense than background marrow uptake and described by location, number, size, and associated standardized uptake values (SUV-FL). Intramedullary PET-FL were considered metabolically active disease, and when resolved, inactive or treated disease. EMD was defined as the presence of 18FDG-avid tissue that, according to CT examination, was not contiguous to bone and arose in soft tissue sites (eg, lymph nodes, liver, pleura, testis, skin). If present, EMD was described by location, size, number, and SUV (SUV-EMD).|
|Shortt, 2009||Disease activity was assessed on attenuation-corrected PET images with a maximal SUV of >2.5 denoting 18FDG-avid lesions.|
|Durie, 2002||All areas of focal uptake were interpreted as positive for myeloma unless they were on sites of known accumulation, including the kidney and bladder, gastrointestinal tract, and certain skeletal areas showing symmetric joint uptake. A diffuse increase in bone marrow activity was interpreted as positive for myeloma if the intensity of activity relative to the background was considered greater than that seen using comparable imaging in patients without bone marrow disease.|
|Jadvar, 2002||Not mentioned|
|Schirrmeister, 2002||All foci presenting with a significant increase in 18FDG uptake located in the skeleton were considered indicative of myeloma. Focal increase in tracer uptake in extraskeletal regions was interpreted as EMD. Homogeneous moderate increase in 18FDG uptake in bone marrow was interpreted as probably indicative of MM. Intense 18FDG uptake in the complete skeleton was interpreted as definitely indicative for bone marrow involvement|
|Schirrmeister, 2003||All foci presenting with significant increase in 18FDG uptake located in the skeleton were considered as indicative for myeloma. Nonphysiologically, focal increased tracer uptake at extraskeletal regions was interpreted as EMD.|
|Mileshkin, 2004||A lesion is reported to be positive and consistent with myeloma if the reporting nuclear medicine doctor says so.|
|Bredella, 2005||Not mentioned|
|Hung, 2005||Not mentioned|
|Adam, 2007||Not mentioned|
|Hur, 2007||All areas of diffuse or focal increases in 18FDG uptake in bone marrow if the intensity was determined greater than the uptake of adjacent or contralateral normal tissue not located in a region of physiologically increased uptake.|
|Hur, 2008||All focal uptake unless they were at sites of known accumulation or diffuse uptake if the intensity of activity relative to the background was considered greater than that seen using compatible imaging in patients without bone marrow disease.|
|Dimitrakopoulou- Strauss, 2009||Not mentioned|
|Castellani, 2010||Not mentioned|
Methodological Quality Assessment
Methodological quality was assessed by 14 items for each of the 18 selected studies. The scores for the QUADAS tool are presented in Table 4. All studies had a representative patient cohort, although in 7 studies (39%) the inclusion and exclusion criteria were not described sufficiently. All studies had a valid reference test (WBXR and/or MRI spine/pelvis) independently of the index test. However, in only 7 studies (39%) the execution of the reference test was described in sufficient detail to permit its replication. Moreover, statements about blinding of clinicians for PET scan results were often lacking. The main problem was lacking the description of a positive PET scan; in 7 studies the definition of a positive scan was unclear. In addition, 7 studies using semiquantative measurements provided insufficient information on image reconstruction methods. In only 9 studies (50%), the clinicians were blinded from the results of the reference test. The mean QUADAS score, expressed as a percentage of the maximum score, was 61% (range, 29%-86%). The studies with a methodological quality above average were the studies by Bredella et al,10 Fonti et al,14 Hur et al,16, 17 Schirrmeister et al,23, 24 and Zamagni et al.4, 21
|Study||1||2||3||4||5||6||7||8||9||10||11||12||13||14||Total||Percentage of Maximum Score|
|Breyer, 2006||+||+||+||−||+||+||+||Visual −||−||−||−||+||−||−||7||50|
|Fonti, 2007||+||+||+||+||+||+||+||Visual +||−||+||+||+||−||−||11||79|
|Zamagni, 2007||+||+||+||+||+||+||+||Visual +||+||+||+||+||−||−||12||86|
|Salaun, 2008||+||−||+||−||−||−||−||Visual −||−||−||−||+||−||+||4||29|
|SUV not done|
|Bartel, 2009||+||−||+||−||+||+||+||Visual +||+||−||−||+||+||−||9||64|
|Shortt, 2009||+||+||−||+||+||+||+||Visual +||+||−||−||+||−||−||8||57|
|Durie, 2002||+||−||+||−||+||+||+||Visual +||−||−||−||+||−||−||7||50|
|SUV not done|
|Jadvar, 2002||+||−||+||−||−||−||+||Visual –||−||+||−||+||−||−||5||36|
|SUV not done|
|Schirrmeister, 2002||+||+||+||−||+||+||+||Visual +||+||+||−||+||−||+||11||79|
|SUV not done|
|Schirrmeister, 2003||+||+||+||−||+||+||+||Visual +||+||+||−||+||−||+||11||79|
|SUV not done|
|Mileshkin, 2004||+||+||+||−||+||+||+||Visual +||−||−||−||+||−||−||8||57|
|SUV not done|
|Bredella, 2005||+||+||+||+||+||+||+||Visual +||−||+||−||+||−||−||10||71|
|Hung, 2005||+||−||+||−||+||+||+||Visual −||−||+||−||+||−||−||7||50|
|SUV not done|
|Adam, 2007||+||+||+||−||+||+||+||Visual −||−||−||−||+||−||−||7||50|
|Hur, 2007||+||+||+||+||+||+||+||Visual +||+||+||+||+||−||−||12||86|
|SUV not done|
|Hur, 2008||+||+||+||−||+||+||+||Visual +||+||+||+||+||−||+||11||79|
|SUV not done|
|Dimitrakopoulou– Strauss, 2009||+||−||−||−||−||−||−||Visual +||−||−||−||−||+||+||4||29|
|Castellani, 2010||+||+||+||+||+||+||+||Visual −||−||−||−||+||−||+||9||64|
Comparison of 18FDG PET or PET-CT and Conventional Imaging at Staging
Of the 18 studies conducted, concordance assessment between WBXR and 18FDG PET scan was possible in 7 (concerning 242 patients). In 6 out of these 7 studies, 18FDG PET (with or without CT) scan showed more lytic lesions then conventional WBXR with the exception of lytic lesions located in the skull4, 9-11, 19, 21, 23, 24 (Table 5).
|Study||Study Design||Diagnosis (No. of Patients)||Reference Test (No. of Patients)||Results (Accuracy), Percentage of Patients (No. of Patients)|
|Breyer, 2006||Retrospective||MM (16)||WBXR (13), MRI (22), CT (25)||WBXR = PET-CT: 12.5% (2/16)|
|WBXR > PET-CT: 37.5% (6/16)|
|WBXR < PET-CT: 50% (8/16)|
|MRI = PET-CT: 25% (4/16)|
|MRI > PET-CT: 31% (5/16)|
|MRI < PET-CT: 44% (7/16)|
|Fonti, 2007||Prospective||MM untreated (33)||MRI of spine and pelvis (33), 99mTc-MIBI (33)||PET/CT: overall, 32/33 (97%)|
|PET/CT: spine and pelvis, 21/33 (63%)|
|MRI: 27/33 (81%)|
|Zamagni, 2007||Prospective||MM untreated (46)||WBXR (46), MRI of spine and pelvis (46)||WBXR = PET-CT: 46% (21/46) WBXR > PET-CT: 8% (4/46)|
|WBXR < PET-CT: 46% (21/46)|
|MRI = PET-CT: 70% (32/46)|
|MRI > PET-CT: 30% (14/46)|
|MRI < PET-CT: 0%|
|35% (16/46); PET/CT detected lesions outside range of MRI|
|Salaun, 2008||Prospective||Plasmacytoma (24)||MRI (24)||MRI = PET-CT: 60% (12/20)|
|50% (10/20) on PET-CT lesions outside range of MRI|
|Durie, 2002||Retrospective||MM (52), MM untreated (16), MGUS/amyloidosis (14)||WBXR (66)||Positive PET: 100% (16/16)|
|Negative WBXR: 25% (4/16)|
|MGUS PET-negative: 100% (14/14)|
|Jadvar, 2002||Prospective||MM (3), MM untreated (3), response evaluation||WBXR (6), CT of spine (2), MRI (4), radionuclide bone scan (3)||All PET scans were positive|
|2/3 posttreatment scans improved and were concordant with clinical improvement|
|1/3 showed new lesions concordant with clinical deterioration|
|Schirrmeister, 2002||Prospective||MM (28), plasmacytoma (15)||WBXR (43), MRI spine (19), extravertebral (13), CT (21), radionuclide bone scan (21)||WBXR = PET: 34% (38/112)|
|WBXR > PET: 3% (3/112)|
|WBXR < PET: 63% (71/112)|
|In 23 patients with known osteolytic lesions|
|Schirrmeister, 2003||Prospective||Plasmacytoma (15), plasmacytoma untreated (11)||WBXR (15), CT (5), MRI of spine (2), radionuclide bone scan (5)||WBXR = PET: 60% (9/15)|
|WBXR > PET: 7% (1/15)|
|WBXR < PET: 33% (5/15)|
|Mileshkin, 2004||Retrospective||MM (62), MGUS (3), plasmacytoma (4)||WBXR (69)||36 patients with PET|
|WBXR = PET: 50% (18/36)|
|WBXR > PET: 0%|
|WBXR < PET: 50% (18/36)|
|Bredella, 2005||Retrospective||MM (4), MM untreated (9), response evaluation||WBXR (6), MRI (16), CT (4)||PET = radiographs: 50% (2/4)|
|PET < radiographs: 50% (2/4)|
|Hung, 2005||Prospective||MM (12)||WBXR (12), 99mTc-MIBI (12)||WBXR: 80% (12/15); bone lesions|
|PET: 93% (14/15); bone lesions|
|99mTc-MIBI: 80% (12/15); bone lesions|
|Adam, 2007||Retrospective||MM (39); MM untreated (13), MGUS (6), plasmacytoma (2)||WBXR||In MM: WBXR = PET: 76% (10/13); WBXR < PET: 24% (3/13)|
|In MGUS: 2/6 PET-positive; other malignancies|
|Hur, 2007||Retrospective||MM untreated (10)||MDCT (10), MRI of spine (10)||In a total of 140 vertebrae MDCT detected 102, MRI 95 and 18FDG PET 84 lesions (MDCT = MRI; MDCT > PET)|
|Hur, 2008||Retrospective||MM untreated (22)||MRI of spine (22)||Stage 1 and 2: PET = MRI: 78% (29/37) vs 86% (32/37); lesions|
|Stage 3: PET< MRI: 81/101 (80%) vs 93/101 (92%); lesions|
|Shortt, 2009||Prospective||MM (24)||Whole body MRI (24)||18FDG PET and MRI were concordant in 62% (21/34)|
|18FDG PET and MRI were discordant in 54% (13/24): 8 were correct on MRI (62%) and 5 were correct on 18FDG PET (38%)|
Two studies compared 18FDG PET-CT scan with WBXR. In a study by Zamagni et al4, 21 on 46 newly diagnosed multiple myeloma patients, 18FDG PET-CT was superior in detecting lesions in 46% of the patients. In only 8% of patients, WBXR showed more lesions mainly localized in the skull. In agreement with these findings, a retrospective study by Breyer et al.11 showed that 18FDG PET-CT detected more lesions in 8/16 (50%) newly diagnosed patients. However, this superiority was only accounted for by extramedullary lesions in 11 patients. When these findings are excluded, 18FDG PET-CT was even inferior to WBXR.
Five studies compared 18FDG PET with WBXR. In a prospective study, Schirrmeister et al.23, 24 reported that in 28 multiple myeloma patients and 15 plasmacytoma patients 18FDG PET detected more lesions in 63% of patients. In 3% of patients, WBXR showed more lesions, but the location of those lesions was not mentioned. Presumably, the 15 patients with plasmacytomas have been reported separately, showing a superior detection by 18FDG PET in 33% of patients. Mileshkin et al.19 showed in a retrospective study that 50% of the 36 PET-evaluated patients had more lesions on 18FDG PET scan than on WBXR. In addition, Adam et al9 described in 13 previously untreated myeloma patients in 24% more lesions on the 18FDG PET. In contrast with the above studies, Bredella et al.10 showed in a small group of newly diagnosed multiple myeloma patients that the WBXR detected more bone lesions than the 18FDG PET scan in 2/4 (50%) of the patients.
In 5 studies, concordance assessment between 18FDG PET scan and MRI spine and/or pelvis was possible.4, 11, 14, 17, 21, 18 FDG PET was inferior to MRI in detecting myeloma bone disease, especially in case of diffuse bone infiltration. In 4 of these 5 studies, MRI was superior; in the other study, the results were equal (Table 5).
Four studies compared 18FDG PET-CT with MRI of the spine and/or pelvis, and 1 study compared 18FDG PET with MRI of the spine. In the study by Zamagni et al,4, 2118FDG PET-CT was inferior to MRI in 14/46 (30%) patients, of whom 10 had a diffuse pattern and 4 had a focal pattern of bone involvement on MRI. In agreement with the findings of Zamagni et al, Breyer et al.11 showed that 18FDG PET-CT detected equally numbers of focal lesions as the MRI; however, MRI was superior in detecting diffuse pattern of bone involvement. Fonti et al.14 compared prospectively MRI of the spine and pelvis to 18FDG PET-CT; MRI showed lesions in 81% of patients versus 63% on 18FDG PET-CT. Also in this study, diffuse pattern of bone involvement was missed on 18FDG PET-CT. In a retrospective study by Hur et al,17, 18 FDG PET detected 80% of the total bone lesions versus 92% detected by MRI. The only study showing similar sensitivity of PET and MRI was a study by Salaun et al.22 One study compared 18FDG PET-CT to whole body MRI. In this study by Shortt et al,25 PET and whole body MRI findings were concordant in 62% of cases, whole body MRI faring slightly better in nonconcordant cases (62% whole body MRI correct vs 38% PET correct).
18FDG PET or PET-CT in Response Monitoring During and After Treatment
Only 4 prospective studies by Zamagni et al,4 Jadvar et al,18 Bartel et al,3 and Dimitrakopoulou-Strauss et al20 examined the value of 18FDG PET or PET-CT scan in response evaluation after initial therapy. The first indications that midtreatment scans reflected treatment response came from the casuistic study of Jadvar et al, in which 3 multiple myeloma patients were followed before and after therapy; in 2 of the patients, there was a decline in metabolic activity on 18FDG PET scan concordant with clinical improvement. In the remaining patient, the appearance of new hypermetabolic lesions and higher metabolic activity in the previously known lesions was concordant with clinical deterioration. In the study of Zamagni et al., evaluation of 23 multiple myeloma patients 3 months after autologous stem cell transplantation showed that in 65% of patients, 18FDG PET-CT scans normalized, and this corresponded with a marked (≥90%) decrease in M-protein concentration. Interestingly, in the 6 of 8 patients in whom no normalization of the 18FDG PET-CT scan was observed, a complete remission or near complete remission was diagnosed when only the M-protein levels were taken into account. In the study by Bartel et al, 239 patients treated according to “Total Therapy 3” were evaluated with 18FDG PET-CT and MRI at baseline and before autologous stem cell transplantation. The presence of more than 3 18FDG-avid focal lesions at baseline (data-driven threshold) was an independent parameter associated with inferior overall and event-free survival (30-month estimate of event-free survival of 66% in patients with >3 lesions vs 87% in patients with ≤3 lesions). In addition, complete normalization of 18FDG PET uptake before autologous stem cell transplantation correlated with better overall and event-free survival (30-month estimate of event-free survival in patients with complete suppression of 18FDG uptake of 89% vs 63% in patients with still 18FDG-avid lesions), even in patients with poor cytogenetics. Dimitrakopoulou-Strauss et al. demonstrated that baseline SUV is helpful for the prediction of progression-free survival after first-line anthracycline-based chemotherapy. On the contrary, they did not find any correlation between the decrease of SUV after 1 chemotherapy cycle and progression-free survival. This correlation between the height of the SUV at baseline and progression-free survival was also found in the prospective study by Castellani et al.12 In several retrospective studies, response evaluation with 18FDG PET scan after initial therapy was examined. In a study by Kim et al,26 15 plasmacytoma patients were treated with radiotherapy; 11 18FDG PET scans were performed after radiotherapy. Seven of these scans normalized (64%) and stayed negative during follow-up. One scan was still positive 3 months after therapy but turned negative and stayed negative during follow-up. In the study by Mileshkin et al,19 6 multiple myeloma patients after (radio)therapy were evaluated; 6 out of 9 scans were negative, 3 correlated with complete remission, and the other 3 had progressive disease on routine disease markers, but the majority of the known disease sites were not 18FDG-avid.
In the study by Bredella et al,10 3 patients underwent serial 18FDG PET scans to assess response to chemotherapy and bone marrow transplant. In 2 patients, 18FDG PET scan showed a decline in metabolic activity concordant with clinical improvement. In 1 patient, a new focus of abnormal uptake was detected consistent with recurrent disease.
18FDG PET or PET-CT Scan in Nonsecreting and Extramedullary Multiple Myeloma
There is only sparse information on the value of 18FDG PET or PET-CT scan in evaluating nonsecreting and extramedullary multiple myeloma. In 2 studies, nonsecreting multiple myeloma was mentioned; Breyer et al11 described 5 patients with nonsecreting multiple myeloma, 2 of which were 18FDG-negative. Durie et al13 reported 6 patients with nonsecreting multiple myeloma, all of whom were 18FDG-positive. In 9 studies, extramedullary disease was detected with 18FDG PET, ranging from 2%-27%. It was not clear whether the extramedullary lesions were only detected with 18FDG PET or were also present on conventional radiological examinations.
This systematic review analyzed the diagnostic accuracy of dedicated (full ring) 18FDG PET or PET-CT scan for initial staging of multiple myeloma and response monitoring. The studies included in our analysis were of moderate methodological quality, with a median score in the QUADAS tool of 61%. One of the limitations was unclear blinding of PET and reference test readers. In general, this tends to overestimate the diagnostic accuracy. Another type of bias is related to the patient selection. In most studies, interpreted test results and withdrawals were not mentioned. The main problem, however, was that the criteria for PET positivity were not completely consistent among the studies. Several used quantitative measures (in particular, the more recent 18FDG PET-CT studies), and others used variable definitions of focal and diffuse skeletal and extramedullary pathology. With standardized uptake values (SUV) and skeletal pathology, 2.5 was the typical cut-off value among the studies. Although visual inspection of 18FDG PET or PET-CT images remain very important for diagnosis and response assessment, SUVs allow for an objective assessment for treatment response by measuring relative change in SUV, thereby eliminating observer variations. However, because of the disperse variability in methodology of acquisition, image reconstruction, and data analysis, such values are difficult to generalize. Implementation of guidelines such as the Netherlands Society of Nuclear Medicine (NEDPAS)27 or European guidelines for standardization of the acquisition and interpretation of 18FDG PET or PET-CT images28 will overcome these difficulties in the future and are essential for the use of 18FDG PET-CT in staging and response monitoring of multiple myeloma in multicenter trials and meta-analyses.
Recently, the IMWG reported a consensus statement on the role of imaging techniques in multiple myeloma, in which the WBXR was considered to be the gold standard. Our review demonstrated that for staging purposes, 18FDG PET or PET-CT detects more myeloma bone lesions than the gold standard. However, lytic skull lesions are often missed, probably obscured by intense physiological cerebral uptake, and some studies showed only superiority of 18FDG PET due to the detection of extramedullary lesions. In contrast, the 18FDG PET scan seems to be less sensitive than MRI for detecting bone disease in the spine and pelvis, especially when the disease is spread diffusely through the spine instead of localized lytic lesions. Even though we were unable to analyze the specificity of MRI in this context, the impression is that for evaluating the extent of bone involvement at baseline, MRI seems to be a better choice than 18FDG PET. The potential effect of whole body MRI techniques, as well as the impact of the added effect of using the CT information that automatically comes with modern 18FDG PET-CT technology, remain to be studied. The clinical relevance of these new findings is yet unclear. With new and more sensitive imaging techniques, it is necessary to examine the clinical implications of early treatment. Currently, according to the guidelines of the IMWG, only symptomatic multiple myeloma patients are treated. Using 18FDG PET-CT or MRI in initial staging patients with smoldering myeloma will probably be upstaged. Whether subsequent treatment upon those findings will improve prognosis is not clear and was not the scope of the above studies. We did not pool the data on concordances because the heterogeneity between studies was too large to allow for a meaningful meta-analysis.
The study by Bartel et al3 was one of the few studies—and also the largest study—in which the prognostic value of the baseline 18FDG PET-CT scan was investigated. The investigators showed that in addition to osteolytic lesions detected on WBXR, the presence of more than 3 focal 18FDG-avid lesions at baseline adversely affected event-free and overall survival. If these findings and threshold are prospectively validated, this strongly suggests that it is justifiable to start earlier treatment based on 18FDG PET-CT scan at baseline.
When 18FDG PET was the only imaging technique, no information was obtained on cortical bone structure or fracture risk. Therefore, in clinical settings, complementary CT scan is needed. There were no studies comparing 18FDG PET with 18FDG PET-CT; furthermore, in the studies using 18FDG PET-CT, the PET images were not interpreted independently of the CT scan.
For response monitoring, however, 18FDG PET-CT is relevant because of detection of metabolic activity in pre-existent bone lesions, whereas WBXR and standard MRI are not expected to show normalization. This restaging after initial therapy for response monitoring with 18FDG PET or PET-CT has been performed in a very small number of studies. Because imaging with 18FDG PET-CT combines functional characteristics of tissue with anatomy, both for bones and soft tissue, vital myeloma cells can conceptually be discriminated from inactive lesions. The limited studies show a correlation between the response on the 18FDG PET or PET-CT scan and clinical outcome. Zamagni et al4, 21 illustrated that normalization of 18FDG PET-CT occurred in more patients than normalization of MRI after autologous transplantation. Interestingly, in the 6 of 8 patients in whom no normalization of the 18FDG PET-CT scan was observed, a complete remission or near complete remission was diagnosed when only the M-protein levels were taken into account. This finding suggests that normalization of the 18FDG PET-CT scan might add in predicting prognosis; however, this was not evaluated.
Bartel et al3 studied the predictive value of 18FDG PET-CT earlier during treatment. They showed that complete normalization of 18FDG PET uptake before autologous stem cell transplantation correlated with better overall and event-free survival of 92% and 89% versus 71% and 63% in patients with still 18FDG-avid lesions at 30 months. For defining complete remission, Bartel et al stated that normalization of 18FDG PET uptake is a better prognostic marker than M-protein, but more evidence on this topic has to be documented.
The presently available evidence on the diagnostic performance of 18FDG PET-CT in the initial staging and response monitoring of multiple myeloma is promising. In general, 18FDG PET has a higher sensitivity for myeloma bone lesions compared with WBXR, although head-to-head comparisons suggest that MRI may surpass PET. Future studies have to validate the additive value of myeloma-related bone disease detected on 18FDG PET-CT in predicting outcome. However, it is essential that these studies use uniform guidelines for standardization of the acquisition and interpretation of 18FDG PET-CT images to generalize the results. Response monitoring with the use of 18FDG PET-CT during treatment is promising, allowing more precise prediction of prognosis compared with the standard response monitoring. In view of the expanding treatment options for multiple myeloma, this may provide important information for treatment decisions in the future.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 4A prospective comparison of 18F-fluorodeoxyglucose positron emission tomography-computed tomography, magnetic resonance imaging and whole-body planar radiographs in the assessment of bone disease in newly diagnosed multiple myeloma. Haematologica. 2007; 92: 50-55., , , et al.
- 6fluorination with molecular fluorine. a convenient synthesis of 2-deoxy-2-fluoro-D-glucose. 1977; 43: 2341-2342., , .
- 8Development and validation of methods for assessing the quality of diagnostic accuracy studies. Health Technol Assess. 2004; 8: iii, 1-234., , , , , .