| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Imaging, Diagnosis, Prognosis |
Authors' Affiliations: 1 The Jerome Lipper Multiple Myeloma Center, Dana-Farber Cancer Institute, 2 Massachusetts General Hospital, and 3 Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts
Requests for reprints: Noopur Raje, Massachusetts General Hospital and Dana-Farber Cancer Institute, P.O. Box 218, 55 Fruit Street, Boston, MA 02114. Phone: 617-726-0711; Fax: 617-724-3166; E-mail: nraje{at}partners.org.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Experimental Design: Eleven multiple myeloma (MM) patients with ONJ were included in the study. Patients underwent clinical, biochemical, radiographic, and molecular profiling. Ten MM patients on aminobisphosphonates without ONJ and five healthy volunteers were used as controls for biochemical and molecular studies.
Results: MM patients with ONJ were treated with either pamidronate (n = 3), zoledronate (n = 4), or both agents sequentially (n = 4) for a mean of 38.7 months. Radiographic studies showed bone sclerosis and fragmentation on plain films and computerized tomography. Quantitative regional analysis of NaF-PET and FDG-PET scans confirmed an increased standardized uptake value (SUVmax) in areas of ONJ. The target to background ratio of SUVmax was significantly greater for NaF-PET compared with FDG-PET scan. Biochemical bone marker data and transcriptional profiling studies showed that genes and proteins involved in osteoblast and osteoclast signaling cascades were significantly down-regulated in patients with ONJ.
Conclusions: ONJ was associated with a mean duration of 38.7 months of aminobisphosphonate exposure. Radiographic and functional imaging confirmed sites of clinically established ONJ. Gene and protein studies are consistent with altered bone remodeling, evidenced by suppression of both bone resorption and formation.
Bisphosphonates have potent antiresorptive activity that have remarkably reduced bone disease–related morbidity. Prospective randomized placebo-controlled studies in patients with cancer, osteoporosis, and Paget's disease have shown that bisphosphonates decrease SREs (4–6). Both pamidronate and zoledronic acid are second-generation nitrogen-containing aminobisphosphonates with the additional nitrogen side chain conferring greater potency with respect to bone resorptive effects. Alendronate, risedronate, and ibandronate are other aminobisphosphonates commonly used to prevent and/or treat osteoporosis and Paget's disease (7).
Clinical trials have validated the use of intravenous pamidronate or zoledronic acid in patients with breast cancer and MM (8, 9), and have provided the basis for guidelines for their use by the American Society of Clinical Oncology. For patients with MM, current revised American Society of Clinical Oncology guidelines recommend the use of either pamidronate or zoledronic acid for the treatment of bone disease for up to 2 years (10). This is largely based on increasing concerns of osteonecrosis of the jaw (ONJ). The recommendations for the use of bisphosphonates in MM stem from two pivotal studies. First, a placebo-controlled randomized study of pamidronate versus placebo in 392 stage III MM patients with at least one osteolytic lesion showed a reduction in the mean number of SREs in the treatment versus placebo groups (4, 11). This led to the Food and Drug Administration approval for pamidronate in the treatment of MM bone disease in 1995. Zoledronic acid was approved by the Food and Drug Administration in 2002 (12) and was incorporated into American Society of Clinical Oncology guidelines for treatment of MM bone disease based on two randomized studies demonstrating noninferiority when compared with pamidronate with respect to number and time to development of SREs (6, 13).
ONJ has been reported in patients treated with aminobisphosphonates (14). Most case reports and case series have included patients with either MM or metastatic breast and prostate cancer, although several patients with osteoporosis have also been reported (15–17). Most of the studies are retrospective case series or case reports with unclear incidences and frequency estimates. These studies have suggested an association with bisphosphonates; however, their role in the pathogenesis of ONJ remains to be determined. Here, we studied 11 affected patients with MM in order to define ONJ as well as gain insights into its pathophysiology. This is the first report with comprehensive clinical, radiographic, and biochemical characterization of ONJ in MM. Importantly, it provides insights into the potential pathophysiology of ONJ suggesting that altered bone remodeling may be critical to the development of ONJ. These clinical and laboratory end points will be validated in future prospective clinical trials.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
All patients underwent a comprehensive oral examination independently by two oral medicine specialists (S-B. Woo and N. Treister). Clinical data were validated on at least two separate dental visits separated by 8 weeks, and intraoral photographs were obtained.
Anatomic imaging. Patients underwent conventional anatomic imaging including digital panoramic and periapical radiographs, dedicated high-resolution head and neck computerized tomography (CT) scans, and magnetic resonance imaging (MRI) of the head. Intravenous contrast was not given during either the CT or the MRI scans because all patients had MM. Controls were not evaluated radiographically.
PET/CT imaging. Patients were evaluated on separate days with [18F]fluorodeoxyglucose (FDG) PET/CT and [18F]sodium flouride (NaF) to evaluate glucose metabolism and bone mineralization, respectively. All PET/CT scans were done on a GE Discovery ST 16 PET/CT scanner in two-dimensional PET acquisition mode. All CT scans performed as part of the PET/CT exam were acquired without the use of contrast media and with a weight-based reduced dose protocol for attenuation correction and anatomic localization. Patients were instructed to avoid vigorous exercise the day before the FDG-PET/CT scan and fast for at least 6 h prior to the FDG injection. No special patient preparations were required for the NaF PET/CT scan. For both the NaF- and FDG-PET/CT scans, the patient was injected with 
20 mCi of either NaF or FDG and the patient was instructed to rest quietly for 1 h. At 60 min after injection of the radiopharmaceutical, a dedicated head and neck PET/CT scan was done to evaluate areas of ONJ followed by a whole-body PET/CT scan covering the rest of the body to evaluate other potential sites of disease. For the FDG-PET/CT scans, the head and neck scan was acquired over 2 bed positions for 10 min/bed, whereas the whole-body scan was acquired over 11 to 13 bed positions for 5 min/bed. For the NaF-PET/CT scans, the head and neck scan was acquired for 7 min/bed, whereas the whole-body scan was acquired over 3 min/bed. FDG- and NaF-PET scans were reconstructed into 128 x 128 images using two-dimensional OSEM with two iterations and 21 subsets. All PET images were corrected for photon attenuation, radioactive decay, detector efficiency, and random coincidences to allow absolute quantification of tissue radioactive concentration (mCi/cc).
Image analysis. Qualitative clinical interpretation was done by a radiologist for the diagnostic CT and MRI images of the head. The radiologist was initially blinded to the clinical details in each case. The mouth was divided into four quadrants: left mandible, right mandible, left maxilla, and right maxilla. The radiologist was asked to state first whether each quadrant was definitely normal or abnormal for each modality and then define the abnormalities, if any. The clinically uninvolved areas of the mouth would then serve as a form of internal control for each patient.
The PET/CT images were reviewed by a nuclear medicine physician and quantitative standardized uptake value (SUV) analyses were done on the head and neck PET/CT images. Areas of increased FDG or NaF uptake were manually identified and a region-of-interest was defined using a 70% threshold of the maximum uptake. For each lesion, a background region was manually identified on normal contralateral tissue or on similar normal tissue in the case of bilateral disease. SUV were calculated to account for differences in patient size and injected activity according to the following formula:
 |
The maximum SUV uptake (SUVmax) was calculated for each lesion and background region. The ratio of lesion-to-background SUVmax was also calculated to allow relative comparisons between FDG and NaF uptake and image contrast.
Laboratory data. Aerobic and anaerobic bacterial cultures were obtained from exposed necrotic bone, or purulent discharge in cases in which suppuration was present. Routine laboratory data included complete blood count, chemistry and liver function tests, and serologic markers of MM. Additionally, serum calcium, urinary N-telopeptide, creatinine clearance, serum intact parathyroid hormone, and 1,25(OH)2 vitamin D levels were done. Special tests included ELISA done according to manufacturer guidelines to measure serum osteoprotegrin, receptor activator of nuclear factor 
B ligand (RANK-L), osteopontin, dickkopf homologue 1 (DKK-1), and macrophage inflammatory protein-1
(R&D Systems), as well as bone alkaline phosphatase (Immunodiagnostic Systems). These special tests were also done on 10 age-matched control MM patients receiving aminobisphosphonate therapy but without ONJ, and 5 healthy volunteers.
Gene expression profiling. Exploratory transcriptional profiling of peripheral blood mononuclear cells was done using the Affymetrix U133Plus 2.0 Gene Chip (Affymetrix) which analyses 54,600 genes in all 11 patients and compared with the profiles of 10 MM patients on bisphosphonate therapy without ONJ and 5 healthy volunteers. Total RNA was isolated from peripheral blood mononuclear cells using TRIzol Reagent (Invitrogen). The DNA chips were then analyzed using a Gene Array Scanner (Affymetrix). CEL files were obtained using Affymetrix Microarray Suite 5.0 software, a DNA Chip Analyzer (DChip; ref. 18) normalized all CEL files to a baseline array with overall median intensity, and the model-based expression (perfect match only) was used to compute relative gene expression values. The raw and transformed data sets for each microarray experiment was submitted to the GEO repository and is accessible via the accession number GSE7116.
Statistical considerations. All 54,600 genes were first screened to select 21,335 genes that were modified in >30% of the samples. Gene expression analysis identified signals varying by 
1.5-fold (lower bound) with a 90% confidence interval. Hierarchical clustering was subsequently done on genes of functional interest.
All other statistical analysis was exploratory. Variables were summarized as mean (SD) and median (range). Pair-wise comparisons of the levels of bone markers in patients with MM and ONJ versus those without ONJ and healthy volunteers used the Wilcoxon rank-sum test. Bonferroni analysis was applied to control for a family wise error rate of 0.05 and P < 0.0167 (0.5/3, two-sided) was considered statistically significant.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
Clinical imaging. Figures 1 and 2 are representative examples of MM patients with ONJ. The digital panoramic and periapical radiographs were abnormal in 12 of 17 involved quadrants. The abnormalities included sclerotic changes with mottling and bone fragmentation and sequestrum formation. Three patients had persistent extraction sockets. Of the 27 clinically unaffected quadrants, four had abnormalities on panorex similar to the affected quadrants.
|
|
All the patients had FDG-PET. One patient did not have NaF-PET. On the FDG-PET, there was abnormal focal uptake in 12 of 17 involved quadrants. On the NaF-PET, there was abnormal focal uptake in 13 of 16 quadrants. In quadrants in which there was increased uptake on both FDG- and NaF-PET, the NaF uptake was more intense and extensive by visual analysis. Of the 27 clinically unaffected quadrants on FDG-PET, two had focal abnormal tracer uptake. Of the 24 clinically unaffected quadrants on NaF-PET, three showed focal abnormal tracer uptake.
One of the patients, who had involvement of one segment, did not have MRI. The MRI scans were reported as abnormal in 6 of 16 involved quadrants. The abnormalities on MR included loss of T1 signal intensity in bone marrow and cortical fragmentation. The areas of low signal corresponded with sclerotic findings of radiograph and CT. Of the 24 clinically unaffected quadrants, 2 had similar MRI signal abnormalities.
Assuming that the clinically uninvolved sites of disease were truly normal allows us to calculate values for sensitivity, specificity, and accuracy for detecting clinical sites of ONJ using the different modalities as shown in Table 2A . It is important to note that because there were no control patients and the presence of subclinical disease could not be known, these variables do not necessarily characterize the performance of each modality for the diagnosis of ONJ. However, it does show that CT, NaF-PET, FDG-PET, and oral radiography were able to differentiate, with reasonable accuracy, between clinically involved and clinically uninvolved segments. One patient (patient 8 in Table 1) did not have any radiological abnormalities on any of the imaging modalities assessed. Semiquantitative analysis of the NaF- and FDG-PET images showed that the target-to-background ratio of SUVmax for the NaF-PET was significantly greater than the FDG-PET (Table 2B).
|
, and DKK-1). Exploratory molecular profiling showed that genes involved in osteoclast signaling, activation, or differentiation were significantly down-regulated in patients with ONJ. Genes involved in osteoblast signaling, differentiation, and activation were similarly down-regulated, whereas osteoclast-inhibiting factors were up-regulated, consistent with a global inhibition of osteoblastogenesis (Table 3 ). No differences in genes associated with angiogenesis were noted (data not shown). Hierarchal clustering of NFAT transcription factors and protein kinase C family proteins, which are involved in osteoclast and osteoblast signaling, respectively, were significantly down-regulated in patients with ONJ compared with healthy volunteers and age-matched MM controls (Fig. 3 ).
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Bisphosphonates are pyrophosphate analogues with the addition of side chains. One side chain is usually a hydroxyl moiety with a high affinity for hydroxyapatite crystals and calcium; the other allows for the addition of a nitrogen atom, as in the aminobisphosphonates, which increases potency. They bind to resorbing bone and are not metabolized; therefore, high concentrations are maintained in bone for prolonged periods. Although most pivotal studies in oncology have shown the efficacy of these compounds, long-term follow-up data is limited. Importantly, the development of more effective treatment strategies in MM such as the use of novel agents like thalidomide, lenalidomide (19), and bortezomib (20) is prolonging survival. As a consequence, patients are now exposed to longer periods of supportive care measures such as bisphosphonates, highlighting the need to evaluate their long-term sequelae.
Although bisphosphonates have been associated with ONJ, their role in its pathophysiology remains to be defined (21). Interference of normal bone homeostasis by aminobisphosphonates may result in the accumulation of microdamage, thereby affecting the mechanical integrity of the bone. The antiangiogenic effects of bisphosphonates might also be contributory. However, the predilection of ONJ for the maxilla and the mandible cannot be due solely to the vascularity of the bone because the maxilla, unlike the mandible, has an abundant vascular supply and is often involved. Instead, the lack of repair of physiologic microdamage, high masticatory forces, the demand placed on bone remodeling because of infection and following extractions, and exposure of tooth socket to a contaminated environment may all contribute to the development of ONJ. Current recommendations to avoid ONJ therefore focus on the avoidance of invasive dental procedures and careful maintenance of dental and periodontal health in patients on bisphosphonates (22).
We studied 11 MM patients with ONJ clinically, biochemically, and radiographically. Our patient group showed abnormalities in multiple different imaging modalities in established clinical ONJ. However, there are limitations that must be considered in the interpretation of these findings. First, there was no control group. We have attempted to partially address this by using the clinically unaffected regions of the mandibles and maxillae as internal controls. This approach has been validated, in part, because with the exception of MRI, all of the imaging modalities showed reasonable accuracy in distinguishing the affected from the unaffected regions. Second, the documented findings are nonspecific and may occur with a number of pathologic conditions. This was of particular concern in our group of patients who all had MM, a condition that could result in multiple abnormalities on all of the imaging modalities used in the study. Although using clinically unaffected areas as an internal control may address some of these issues, in any given case, it is not possible to definitively state whether FDG uptake, for example, was due to inflammation related to osteonecrosis or to other dental conditions such as periodontal infections which are associated with bone sclerosis and increased bone turnover. MM could however be excluded as all these patients were in remission or had stable disease at the time of ONJ diagnosis and did not have evidence of MM in the jaw. Lastly, the patient group was comprised of patients with established ONJ. Therefore, the radiological findings in our group may not necessarily apply as methods of screening patients for subclinical disease.
The hallmarks of the findings in our patient group with advanced ONJ seemed to be bone sclerosis and fragmentation. CT seemed to be more sensitive to radiography in our group and permitted a more accurate delineation of pathology and complications such as fracture or sequestra formation. NaF uptake was increased at sites of clinically apparent disease, as typically seen in areas with increased bone turnover and mineralization. FDG uptake usually present at sites of bone healing seen in our patients may reflect an attempted healing response or may be related to inflammation and/or infection at the site of necrosis. In our group of patients with ONJ, between the two PET agents, NaF seemed to be superior at identifying clinical disease compared with FDG. MRI was surprisingly poor at identifying sites of disease and may have been related in part to the lack of gadolinium contrast in our study. Osteonecrosis in other regions, such as the hip, have been studied using MRI which was noted to be very sensitive in early disease (23). Our data shows bone sclerosis and fragmentation, which is present, radiologically, at the end stage of osteonecrosis elsewhere in the body. It is therefore possible that MRI may be useful in earlier stage disease.
A number of radiological studies of ONJ have been recently published. Bianchi et al. (24) studied 32 patients with both CT and oral radiography. They found CT to be highly sensitive for abnormalities in the mouth and accurate in delineation of pathology with 30 of 32 patients detected. However, in their group, oral radiography was not found to be sensitive with only 16 out of 32 patients detected. Chiandussi et al. (25) described 11 patients, who were studied with CT, MRI, and Tc99m bone scintigraphy. They described signs of marrow edema on MRI and osteolytic lesions on CT which is concerning for involvement with MM. All of these studies did not include a control group and therefore possess some of the inherent limitations present in our study. In addition, these studies included patients with different malignancies, unlike our study, which only examined patients with MM. Nonetheless, in general, our findings concur with other studies which illustrate that advanced, clinically apparent ONJ may be radiologically described with reasonable accuracy.
Our study is the first to show that bone homeostasis is affected in ONJ, evidenced by decreased bone formation and resorption markers at both the transcriptional and protein levels. Our biochemical data suggested decreased urinary N-telopeptide in our patients with ONJ, suggesting that active MM bone disease was not an ongoing issue. Other markers of bone formation such as bone alkaline phosphatase were decreased in both MM patients with and without ONJ. To gain further insights into the pathophysiology of ONJ, we did exploratory gene expression studies. We chose peripheral blood mononuclear cells for molecular profiling, which were readily obtained and contained adequate numbers of cells of osteoclast and osteoblast lineage (26, 27). We included matched MM patients without ONJ to control for the effect of other MM therapies. Our data show a global inhibition of genes involved in bone remodeling in patients with ONJ. In the future, it will be critical to perform similar studies in patients that develop ONJ while on bisphosphonates for nonmalignant diseases in order to avoid confounding effects of chemotherapy regimens. Moreover, prospective trials are needed to assess the predictive value of biomarkers in early diagnosis and progression of ONJ.
Importantly, prospective trials are needed to both define the true incidence of ONJ and to evaluate strategies for its prevention. A prospective Greek study evaluated 202 patients with MM treated with either pamidronate or zoledronic acid (28). Fifteen (7.4%) patients developed ONJ with a median bisphosphonate exposure time of 39 months, suggesting a higher cumulative hazard of developing ONJ after zoledronic acid exposure. Another study evaluating 303 patients with MM identified an 11% ONJ incidence after bisphosphonate exposure, with a higher rate in patients receiving zoledronic acid (29). In contrast, in a retrospective analysis of 4,000 patients which identified ONJ in 34 cases, the generic type of bisphosphonate exposure did not emerge as a significant risk variable in a multivariate logistic regression analysis (30).
Although bisphosphonates clearly benefit patients with MM, ONJ presents a clinically significant adverse event in a subset of patients, and two potential strategies are ongoing to decrease the risk of ONJ. First, several novel therapies, i.e., targeting macrophage inflammatory protein-1, RANK-L, and OPG are undergoing clinical evaluation for the treatment of bone disease (31, 32). Second, alternate regimens of bisphosphonate therapy are being recommended to limit drug exposure. For example, a Mayo Clinic consensus statement has recently advocated the use of bisphosphonates in MM patients for 2 years; beyond 2 years, bisphosphonates are recommended at a reduced dosing frequency only in those patients with active and/or significant bone disease (33). Importantly, clinical studies are needed to determine whether less frequent dosing schedules will maintain the benefits of bisphosphonates without untoward side effects like ONJ. For example, bone marker data from randomized studies of pamidronate and zoledronic acid show a 2-fold increased rate of SREs in patients with elevated urinary N-telopeptide levels (34); using this biomarker to tailor therapy may therefore allow for amelioration of bisphosphonate-related toxicity. To this end, we are launching a prospective study of bisphosphonate therapy in patients with MM which will compare alternate dosing schedules based on urinary N-telopeptide levels. We will monitor for ONJ prospectively in order to both validate our current clinical, radiographic, and biomarker findings as diagnostic criteria, as well as to determine their value for predicting the development of ONJ. Importantly, this study will address whether less frequent bisphosphonate dosing will reduce the number of SREs while avoiding ONJ.
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Presented at The American Society of Hematology 2006.
Received 6/11/07; revised 11/ 7/07; accepted 11/28/07.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
B ligand inhibitor, denosumab, in patients with multiple myeloma or bone metastases from breast cancer. Clin Cancer Res 2006;12:1221–8.This article has been cited by other articles:
|
|
 |
|
  A. Badros, E. Terpos, E. Katodritou, O. Goloubeva, E. Kastritis, E. Verrou, K. Zervas, M. R. Baer, T. Meiller, and M. A. Dimopoulos Natural History of Osteonecrosis of the Jaw in Patients With Multiple Myeloma J. Clin. Oncol., December 20, 2008; 26(36): 5904 - 5909. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kyrgidis and C. Andreadis Clinical Characterization Might Help in Preventing Osteonecrosis of the Jaw Clin. Cancer Res., December 15, 2008; 14(24): 8321 - 8321. [Full Text] [PDF]
|
|
|
|
|
|
N. Raje and K. C. Anderson Reply to the Letter to the Editor from Noopur Raje and Kenneth C. Anderson Clin. Cancer Res., December 15, 2008; 14(24): 8321 - 8321. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
