Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print September 26, 2007, 10.1148/radiol.2452061877
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2452061877v1
245/3/732    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pakos, E. E.
Right arrow Articles by Ioannidis, J. P. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pakos, E. E.
Right arrow Articles by Ioannidis, J. P. A.
(Radiology 2007;245:732-741.)
© RSNA, 2007

Evidence-based Practice

Osteomyelitis: Antigranulocyte Scintigraphy with 99mTc Radiolabeled Monoclonal Antibodies for Diagnosis—Meta-Analysis1

Emilios E. Pakos, MD, Haris D. Koumoulis, MD, Andreas D. Fotopoulos, MD, and John P. A. Ioannidis, MD

1 From the Clinical Trials and Evidence-Based Medicine Unit, Department of Hygiene and Epidemiology (E.E.P., H.D.K., J.P.A.I.), and Department of Nuclear Medicine (A.D.F.), University Hospital of Ioannina School of Medicine, University Campus, Ioannina, Epirus 45110, Greece; and Institute for Clinical Research and Health Policy Studies, Department of Medicine, Tufts-New England Medical Center, Tufts University School of Medicine, Boston, Mass (J.P.A.I.). Received November 2, 2006; revision requested January 3, 2007; revision received February 3; final version accepted March 25. Address correspondence to J.P.A.I. (e-mail: jioannid{at}cc.uoi.gr).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Purpose: To perform a meta-analysis of the sensitivity and specificity of antigranulocyte scintigraphy with monoclonal antibodies (MoAbs) in the diagnosis of osteomyelitis across different patient groups and clinical settings.

Materials and Methods: MEDLINE and EMBASE searches were conducted. Data on the diagnostic performance of antigranulocyte scintigraphy with MoAbs were combined. Weighted sensitivities and specificities were estimated by using a random-effects model that incorporated between-study heterogeneity and by constructing summary receiver operating characteristic (ROC) curves. The weighted positive and negative likelihood ratios (LRs) across studies were estimated. Data syntheses were performed for all patients and for various subgroups. The reference standard used in each individual study was accepted.

Results: Nineteen nonoverlapping studies with a total of 714 examinations and reference standards of cell culture, histologic examination, clinical follow-up, and radiologic examination were eligible. The independent random-effects summary estimates of sensitivity and specificity were 81% (95% confidence interval [CI]: 70%, 88%) and 77% (95% CI: 66%, 86%), respectively, with statistically significant between-study heterogeneity (exact P < .001 for both metrics). In the summary ROC curve, a sensitivity of 81% corresponded to a specificity of 86%, and a specificity of 77% corresponded to a sensitivity of 87%. The weighted positive LR was 3.02 (95% CI: 2.07, 4.42), and the weighted negative LR was 0.26 (95% CI: 0.17, 0.39), with statistically significant between-study heterogeneity (exact P < .001 for both metrics). Sensitivity was better for peripheral than for axial skeleton lesions (87% vs 53%).

Conclusion: Antigranulocyte scintigraphy with MoAbs has a sensitivity of 81% and a specificity of 77% in the diagnosis of osteomyelitis.

© RSNA, 2007


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Accurate diagnosis of osteomyelitis is important for early initiation of antimicrobial and surgical treatment, with critical consequences for patient outcome (1). Diagnosis is mainly based on results of histologic examination and cell cultures from the infected area and is supported by imaging options such as radiography, computed tomography (CT), magnetic resonance (MR) imaging, and nuclear medicine. The inflammatory process involves the infiltration of cells, including leukocytes, in the affected area. Leukocytes pass through the gap between the endothelial cells as a response to inflammatory mediators and migrate up the chemotactic gradient to the site of inflammation (2). Therefore, the use of radiolabeled leukocytes ex vivo (3,4) and radiolabeled antigranulocyte monoclonal antibodies (MoAbs) in vivo allows visualization of the inflammatory process. This might be feasible even in the early stages of the inflammatory process, when there are no major morphologic changes yet.

Antigranulocyte scintigraphy with the use of MoAbs or antibody fragments is a promising diagnostic tool that has been widely used during recent years for the diagnosis of infection in several clinical settings, including osteomyelitis. Several studies have been performed to evaluate the accuracy of antigranulocyte scintigraphy with a variety of technetium 99m (99mTc)–labeled MoAbs in osteomyelitis (5–23). The most commonly used MoAbs are an immunoglobulin G antibody against normal cross-reactive antigen-95 (BW 250/183) and the Fab fragment of immunoglobulin G antibody directed against the glycoprotein cross-reactive antigen-90 (sulesomab, LeukoScan; Immunomedics, Morris Plains, NJ). Other MoAbs that have been used for the diagnosis of osteomyelitis include anti–stage-specific embryonic antigen–1 (anti-SSEA-1) and fanolesomab (LeuTech), two antibodies directed against the glycoprotein lacto-n-fucopentaose CD15.

However, single published studies may leave considerable uncertainty as to the diagnostic accuracy of these methods. Therefore, the purpose of our study was to perform a meta-analysis of the sensitivity and specificity of antigranulocyte scintigraphy with MoAbs in osteomyelitis across different patient groups and clinical settings.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Identification and Eligibility of Relevant Studies
We identified published studies that involved the use of antigranulocyte scintigraphy with MoAbs or antibody fragments labeled with 99mTc for the diagnosis of osteomyelitis. Relevant studies were considered to be eligible regardless of the type of monoclonal antibody used (sulesomab, BW 250/183, fanolesomab, anti-SSEA-1), the type of osteomyelitis evaluated (acute or chronic), and the location of the osteomyelitis (axial or peripheral skeleton). We considered studies if they had included both patients who were eventually given a diagnosis of osteomyelitis and patients for whom osteomyelitis was eventually excluded. Studies that enrolled at least 10 subjects were considered eligible. We set no language restrictions. We excluded nonhuman studies.

One author (E.E.P.) performed MEDLINE and EMBASE searches (the last search was performed in June 2006). The search strategy was based on the following combination: ["antigranulocyte scintigraphy" OR "WBC-scintigraphy" OR "leukocyte scintigraphy"] AND ["monoclonal antibodies" OR "sulesomab" OR "LeukoScan" OR "Fab" OR "BW 250/183" OR "LeuTech" OR "anti-SSEA-1"] AND ["osteomyelitis" OR "bone infection"], where WBC stands for white blood cell.

References of retrieved articles were also screened for additional studies by two authors (E.E.P. and H.D.K.) in consensus. Investigators of eligible studies were contacted (E.E.P.) and asked to supplement additional data when key information relevant to the meta-analysis was missing. Whenever reports pertained to overlapping study populations, we retained only the largest study to avoid duplication of information. Two reviewers (E.E.P. and H.D.K.) evaluated titles for eligibility in consensus.

Data Extraction
Three investigators (E.E.P., H.D.K., and A.D.F.) extracted data from eligible studies independently. Discrepancies were discussed and consensus was reached on all items, with the exception of four articles (5,11–13), for which a fourth investigator (J.P.A.I.) also arbitrated the data.

We extracted data on characteristics of studies and patients, measurements performed, and results. For each report we recorded author names, journal and year of publication, number of patients examined (overall and by disease location [axial or peripheral]), study design (prospective, retrospective, or unclear), demographic characteristics of patients (age and sex), type of osteomyelitis (acute or chronic), antibiotic therapy, type of MoAb or antibody fragment used, technical characteristics of antigranulocyte scintigraphy (dose and timing of scan after injection), the number of readers who assessed the results of scintigraphy, methods used for the interpretation of the scintigraphic results (qualitative, semiquantitative, quantitative), whether any blinding of the readers to the final diagnosis was reported, and the definition of the reference standard for the diagnosis of osteomyelitis. Finally, we assessed the reported quality of each study by using the Quality Assessment of Studies of Diagnostic Accuracy included in Systematic Reviews, or QUADAS, tool (24).

For each report, we recorded the number of true-positive, false-positive, true-negative, and false-negative results for antigranulocyte scintigraphy with MoAbs in the diagnosis of osteomyelitis. We accepted the reference standard that was used in the individual studies for the documentation of osteomyelitis (histologic examination, cell culture, follow-up, MR imaging, "other," and combinations thereof). We also reported separate subgroup data, whenever these were pertinent, according to type of osteomyelitis (acute, chronic), location (axial, peripheral), and reference standard.

Statistical Analysis and Data Synthesis
Data on the diagnostic performance of antigranulocyte scintigraphy with MoAbs were combined quantitatively across eligible studies (E.E.P., with the supervision of J.P.A.I.). Four approaches were used: First, we independently combined sensitivities and specificities across studies. Between-study heterogeneity was assessed with the Fisher exact test. We estimated the weighted sensitivities and specificities by using a random-effects model that incorporated between-study heterogeneity. Second, we constructed summary receiver operating characteristic (ROC) curves. Third, we estimated the weighted positive and negative likelihood ratios (LRs) across studies by using random-effects calculations. Weights were the inverse of the random-effects variance.

Independent combination of sensitivity and specificity has the advantage of enabling the ready derivation of point estimates and confidence intervals (CIs). However, for a diagnostic or predictive test, the sensitivity (number of true-positives) and specificity (1 – number of false-positives) are related to each other. Evaluating these two quantities independently may lead to some underestimation of diagnostic performance.

To bypass this problem, one may use the summary ROC method. The summary ROC curve is estimated by the regression D = a + bS, where D is the difference of the logits of the true-positive and false-positive rates and S is the sum of these logits (25). Both weighted and unweighted regressions were estimated. The summary ROC curve shows the tradeoff between sensitivity and specificity across the included studies. The relative disadvantage is that no single pair of sensitivity and specificity values is obtained, and the simple regression outlined above does not lead to calculation of CIs. Therefore, we report representative values of pairs of sensitivities and specificities from the resulting summary ROC curve.

LRs are also metrics that combine both sensitivity and specificity in their calculation. The positive LR is defined as the ratio of sensitivity over 1 – specificity, while the negative LR is defined as the ratio of 1 – sensitivity over specificity. When a diagnostic test has absolutely no discriminating ability, both likelihood ratios equal 1. The discriminating ability is better with a higher positive LR and a lower negative LR. Although there is no absolute cutoff, a good diagnostic test may have a positive LR of greater than 5 and a negative LR of less than 0.2 (26). The presence of between-study heterogeneity (diversity) across all studies was assessed with the {chi}2-based Q statistic, and heterogeneity was considered significant for P values of less than .10. To examine whether results were also similar in a calculation approach involving computation of positive and negative LRs jointly rather than independently, we also used the proposed hierarchical summary ROC analysis method (27,28), which we implemented with software (WinBugs; David Spiegelhalter, Cambridge, England).

Finally, we also calculated the diagnostic log odds ratio from each study and combined the log odds ratios by using random-effects calculations to obtain a summary log odds ratio (29).

Bias, Subgroup, and Sensitivity Analyses
Funnel plot approaches have been proposed for determining whether small studies give different results from those of larger studies. Funnel plot asymmetry could reflect publication bias but also a multitude of other effects ("small study effects"), and inferences should be made cautiously (30). For diagnostic test accuracy, we examined funnel plot asymmetry by calculating the correlation between the log odds ratio and the inverse of the effective sample size in each study, where effective sample size is given as in the method of Deeks et al (31).

Subgroup analyses were performed according to localization of osteomyelitis, type of study design, type of osteomyelitis, type of MoAb used, time of antigranulocyte scintigraphy after injection, methods used for antigranulocyte scintigraphy interpretation, and blinding. LR estimates were compared between subgroups by considering the difference of their natural logarithms divided by the square root of the sum of their variances. We also evaluated whether the exclusion of studies in which a single method was used as the reference standard altered the results.

Analyses were performed with software (SPSS, SPSS, Chicago, Ill; Meta-Test, Joseph Lau, Boston, Mass; and StatXact, version 3.0, Cytel, Boston, Mass). P values were two tailed.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Eligible Studies
Thirty potentially eligible studies were initially considered. Eleven of these were eventually excluded, leaving 19 nonoverlapping studies with a total of 714 examinations in 683 patients eligible for the present meta-analysis (Table 1). The reasons for exclusion included fewer than 10 patients in five of 11 studies (32–36) and missing data in three of 11 studies (37–39). The studies with missing data had a total of 40 subjects. Finally, three of 11 studies (40–42) overlapped with three larger series (5,16,21) and were thus also excluded.


View this table:
[in this window]
[in a new window]

  		Table 1. Characteristics of Studies Included in Meta-Analysis

 
A total of 562 antigranulocyte scintigraphy examinations (in 17 of 19 studies) were performed for evaluation of peripheral lesions, and 85 examinations (in eight of 19 studies) were performed for evaluation of axial lesions. Reports of two of 19 studies did not indicate the location of the examined lesions (9,17).

Reports of eight studies stated them to be prospective, while four studies were retrospective and the design was not stated for seven studies. One study (16) was in a language other than English. Mean or median age ranged between 45 and 61 for studies that reported this information, but information was often missing on age range; there was always a preponderance of men (Table 1).

Five studies (9–11,13,15) evaluated chronic osteomyelitis, one study addressed acute osteomyelitis (17), and one study had mixed nonseparable populations with both chronic and acute disease (12). The definitions of the type of osteomyelitis were not specified in most studies. Despite contacting the authors, we did not get information on the duration of osteomyelitis symptoms and signs for the other studies. All but two studies (5,19) did not report any information on antibiotic use. One study (5) reported that patients who received antibiotic therapy for at least 7 days were excluded, and the other (19) stated that no patient was taking antibiotics during the period when scintigraphy was performed.

Among the 19 eligible studies, nine (with 402 examinations) evaluated the diagnostic performance of sulesomab (6–8,14,17,19–21,23), nine (with 287 examinations) used the antibody BW 250/183 (9–13,15,16,18,22), and one study (with 25 subjects) used the antibody fanolesomab (5). Varying doses of 99mTc were used, and the timing of the scans also varied (Table 2). In 12 studies (8–11,13,15–20,23), it was reported that at least two experts interpreted the results of antigranulocyte scintigraphy. Interpretation of scintigraphic results was based on qualitative methods in seven studies (5–8,14,17,19), on semiquantitative methods in nine studies (10,11,13,15,16,18,20,21,23), and on quantitative methods in two studies (9,22), while one study did not report the type of interpretation (12). Reports of seven studies (5,9,11,13,15,17,20) indicated that the experts were blinded to the clinical data of patients.


View this table:
[in this window]
[in a new window]

  		Table 2. Characteristics of Antigranulocyte Scintigraphy in Included Studies

 
Different reference standards were used to establish the diagnosis of osteomyelitis across eligible studies and included different combinations of histopathologic examinations (15 studies), cell culture (16 studies), clinical follow-up (13 studies), and radiologic evaluations (10 studies) (Table 2).

Qualitative Appraisal
Spectrum of disease differed across studies on the basis of the inclusion of different proportions of peripheral and axial lesions and acute and chronic cases, as described above. All studies, with one exception (12), described selection criteria. The adequacy of the reference standard as a disease classifier was variable because different combinations of evaluations were used, as described above. No studies gave detailed information on the exact timing of scintigraphy compared with the timing of the evaluations comprising the reference standard. However, all patients who underwent antigranulocyte scintigraphy also underwent some reference standard evaluation. There was no apparent verification bias (the reference standard was not different depending on the results of scintigraphy) in any study. The reference standard was always independent of antigranulocyte scintigraphy (ie, scintigraphy was not included in the reference standard). All studies described the performance of antigranulocyte scintigraphy in sufficient detail as to permit its replication. Conversely, the reference standard was variable, as described above, and this would cause difficulty in its replication. Although in six studies, antigranulocyte scintigraphy expert readers were blinded to reference standard results and to clinical data, no study reported whether interpretation of the reference standard was also performed with blinding to the results of antigranulocyte scintigraphy. One study (13) also reported intermediate results, and its investigators decided to classify them together with the negative results. Finally, four studies (15,16,21,23) specifically reported that the number of examined patients was different from the number of eligible patients and reported the specific reasons for the exclusions.

Data Synthesis
The sensitivity rates of antigranulocyte scintigraphy with MoAbs for the diagnosis of osteomyelitis ranged between 35% and 100% across eligible studies, while the respective specificity rates ranged from 30% to 100% (Table 3). The independent random-effects summary estimates of sensitivity and specificity were 81% (95% CI: 70%, 88%) and 77% (95% CI: 66%, 86%), respectively. There was statistically significant between-study heterogeneity in both the sensitivity and specificity estimates (exact P < .001 for both metrics).


View this table:
[in this window]
[in a new window]

  		Table 3. Sensitivities and Specificities of Eligible Studies Included in Meta-Analysis

 
In the summary ROC curve, a sensitivity of 81% corresponded to a specificity of 86%, and a specificity of 77% corresponded to a sensitivity of 87% (Fig 1) in the unweighted analysis. The respective weighted pairs were 81% and 84% and 77% and 85%. Both curves (weighted and unweighted) were symmetric (the b coefficient was not significantly different from 0, suggesting that these studies may still be operating at different thresholds on the same ROC curve).


Figure 1
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1: Summary ROC curve analysis results for diagnostic accuracy of antigranulocyte scintigraphy with MoAbs for osteomyelitis. Both weighted (thick line) and unweighted (thin line) calculations are shown. Each study is indicated by an ellipse with a diameter proportional to the weight of the sensitivity and specificity of the study. Also shown are summary estimates (x) of sensitivity and specificity estimates independently combined by using fixed- and random-effects calculations, along with their 95% CIs (gray boxes): Each x refers to sensitivity in vertical line and 1 – specificity in horizontal line; the two x marks refer to different combinations of sensitivity and specificity according to weighted (light gray box) or unweighted (dark gray box) analysis. Note that the independently combined estimates slightly underestimate the diagnostic performance.

 
LR syntheses yielded a weighted positive LR of 3.02 (95% CI: 2.07, 4.42) and a weighted negative LR of 0.26 (95% CI: 0.17, 0.39), with statistically significant between-study heterogeneity for both metrics (Fig 2). When the positive and negative LRs were estimated jointly by using a hierarchic summary ROC analysis, the summary values were 4.08 (95% CI: 3.36, 4.80) and 0.34 (95% CI: 0.10, 0.58), respectively.


Figure 2A
View larger version (15K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a: Forest plots for (a) positive LR and (b) negative LR for accuracy of antigranulocyte scintigraphy with MoAbs in diagnosis of osteomyelitis. Each study is shown with point estimate and 95% CIs. Also shown is summary estimate (ALL) according to random-effects calculations. The study of Seybold et al (12) had a positive LR of 59.1 and lies outside the depicted range.

 

Figure 2B
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b: Forest plots for (a) positive LR and (b) negative LR for accuracy of antigranulocyte scintigraphy with MoAbs in diagnosis of osteomyelitis. Each study is shown with point estimate and 95% CIs. Also shown is summary estimate (ALL) according to random-effects calculations. The study of Seybold et al (12) had a positive LR of 59.1 and lies outside the depicted range.

 
The summary odds ratio obtained with random-effects calculations was 19.4 (95% CI: 9.1, 41.3). There was no evidence of funnel plot asymmetry (correlation between log odds ratio and inverse of effective sample size, –0.19; P = .26).

Subgroup Analyses
Comparison of negative LR values showed a significantly lower negative LR for the diagnosis of axial versus peripheral lesions (P = .001), and this was also reflected in a lower sensitivity for axial lesions. There was also better performance, with higher positive LRs, in studies that reported blinded interpretation of antigranulocyte scintigraphy than in those that did not report on blinding (P = .023). Otherwise, no significant differences were observed according to study design, type of osteomyelitis, type of MoAb, time of injection, and method of scan interpretation, but the number of evaluations was small for many of the examined subgroups (Table 4). Two studies (14,16) used a single method as a reference standard (radiologic examinations and cell culture, respectively). The exclusion of these studies did not alter the results (82% and 76% for overall sensitivity and specificity, respectively).


View this table:
[in this window]
[in a new window]

  		Table 4. Subgroup Analyses of Diagnostic Performance of Antigranulocyte Scintigraphy with MoAbs

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
The results of the present meta-analysis indicate that antigranulocyte scintigraphy with MoAbs has relatively high, but not perfect, diagnostic performance for the diagnosis of osteomyelitis. The sensitivity and specificity rates were at the level of 80% or slightly higher in both the overall analysis and the various subgroup analyses of the type of MoAbs used, the study design, and the type of antigranulocyte scintigraphy interpretation. There was significant between-study heterogeneity in most of the analyses. The LRs yielded similar inference, falling short of passing the thresholds of a positive LR of greater than 5 and a negative LR of less than 0.2 supportive of strong diagnostic evidence in the majority of analyses. We found evidence that sensitivity is worse for axial than for peripheral lesions. Diagnostic performance tended to be better when it was stated that antigranulocyte scintigraphy interpretation was blinded. No statistically significant differences were observed in the various other subgroup analyses, although these may not be conclusive because of relatively small numbers.

The higher sensitivity rates we observed in studies with peripheral lesions compared with those in studies with axial lesions warrants further evaluation. Lower sensitivity rates for axial lesions of osteomyelitis have been also reported for leukocyte scintigraphy, in which autologous granulocytes are administered to the patient after ex vivo radiolabeling (38,43,44). This decreased diagnostic performance has been attributed to the accumulation of radiolabeled leukocytes in marrow. Therefore, diagnostic performance may be limited in the axial skeleton. However, the difference between axial and peripheral lesions could be due to between-study differences, because for both locations, the combined sensitivity rates showed significant between-study heterogeneity.

Another remarkable finding in the present meta-analysis is that studies that reported blinded reading had higher specificity rates than other studies. However, we should acknowledge that studies that did not report blinding may also have used blinding but simply did not report it in the printed article.

It has been reported that, due to the relatively slow blood clearance of sulesomab, a 24-hour postinjection scan is generally necessary for correct localization of the inflammatory focus (45). However, a subgroup analysis did not reveal substantial differences in sensitivity and specificity rates according to the timing of antigranulocyte scintigraphy.

Different diagnostic methods such as cell culture, histologic examination, follow-up, and radiologic examinations have been used across eligible studies in various combinations for the establishment of osteomyelitis. Most of these reference standards are not completely reliable, especially when used in isolation. There can be some subjectivity and/or error in the interpretation of these evaluations, and this is occasionally further compounded by lack of details on the methods used. The combination of different reference standards in the majority of included studies probably precluded serious misclassification. Some misclassification may nevertheless be inevitable, because the lowest sensitivity rate (35%) among the eligible studies was observed by Gratz et al (14), who used only radiologic examinations for establishing the diagnosis of osteomyelitis. Allowing for misclassification in the reference standard may mean that the true diagnostic accuracy of antigranulocyte scintigraphy may even be slightly better than what we estimated.

Limitations of the meta-analysis included the relatively small number of patients in the individual studies and the remarkable heterogeneity in the characteristics of antigranulocyte scintigraphy (different doses and different scanning times) across eligible studies. Only one study used fanolesomab. Moreover, although we aimed to be inclusive and tried to retrieve additional data from investigators, it is unavoidable that some missing and unpublished data may still exist. The extent of publication and reporting biases in this literature remain unknown, and efforts are being made to improve the transparency of diagnostic study reporting (46). Furthermore, as stated above, the eligible studies used different reference standards. It was not possible to perform sensitivity analyses according to the type of reference standard used because most studies did not provide separate results on the basis of each reference standard. It was also not possible to perform analyses according to age or sex subgroups, because this information was not available for every study. Finally, the interpretation of antigranulocyte scintigraphy results in most studies was based on the more broad and subjective qualitative and semiquantitative methods rather than on quantitative methods. Nevertheless, there is no solid evidence that quantitative methods have better accuracy than qualitative methods.

In summary, and allowing for the foregoing caveats, our meta-analysis suggests that antigranulocyte scintigraphy with MoAbs can be used as a major diagnostic method in patients suspected of having osteomyelitis but that it cannot completely replace traditional methods such as histologic examination and cell culture.


   ADVANCES IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
IMPLICATION FOR PATIENT CARE
 References
 


   IMPLICATION FOR PATIENT CARE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
References
 


   ACKNOWLEDGMENTS
 
We thank Dr H. J. Gallowitsch for contributing additional data clarifications on his study.


   FOOTNOTES
 

Abbreviations: anti-SSEA-1 = anti–stage-specific embryonic antigen–1 • CI = confidence interval • LR = likelihood ratio • MoAb = monoclonal antibody • ROC = receiver operating characteristic

Author contributions: Guarantor of integrity of entire study, J.P.A.I.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, E.E.P., H.D.K., A.D.F.; clinical studies, E.E.P.; statistical analysis, E.E.P., J.P.A.I.; and manuscript editing, all authors

Authors stated no financial relationship to disclose.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 

  1. Haas DW, McAndrew MP. Bacterial osteomyelitis in adults: evolving considerations in diagnosis and treatment. Am J Med 1996;101:550–561.
  2. Lazzarini L, Mader JT, Calhoun JH. Osteomyelitis in long bones. J Bone Joint Surg Am 2004;86:2305–2318.
  3. Seabold JE, Ferlic RJ, Marsh JL, Nepola JV. Periarticular bone sites associated with traumatic injury: false-positive findings with In-111-labeled white blood cell and Tc-99m MDP scintigraphy. Radiology 1993;186:845–849.
  4. Unal SN, Birinci H, Baktiroglu S, Cantez S. Comparison of Tc-99m methylene diphosphonate, Tc-99m human immune globulin, and Tc-99m-labeled white blood cell scintigraphy in the diabetic foot. Clin Nucl Med 2001;26:1016–1021.
  5. Palestro CJ, Caprioli R, Love C, et al. Rapid diagnosis of pedal osteomyelitis in diabetics with a technetium-99m-labeled monoclonal antigranulocyte antibody. J Foot Ankle Surg 2003;42:2–8.
  6. Becker W, Palestro CJ, Winship J, et al. Rapid imaging of infections with a monoclonal antibody fragment (LeukoScan). Clin Orthop Relat Res 1996;329:263–272.
  7. Becker W, Bair J, Behr T, et al. Detection of soft-tissue infections and osteomyelitis using a technetium-99m-labeled anti-granulocyte monoclonal antibody fragment. J Nucl Med 1994;35:1436–1443.
  8. Von Rothenburg T, Schaffstein J, Ludwig J, Vehling D, Koster O, Schmid G. Imaging osteomyelitis with Tc-99m-labeled antigranulocyte antibody Fab' fragments. Clin Nucl Med 2003;28:643–647.
  9. Horger M, Eschmann SM, Pfannenberg C, et al. The value of SPET/CT in chronic osteomyelitis. Eur J Nucl Med Mol Imaging 2003;30:1665–1673.
  10. Kaim A, Ledermann HP, Bongartz G, Messmer P, Muller-Brand J, Steinbrich W. Chronic post-traumatic osteomyelitis of the lower extremity: comparison of magnetic resonance imaging and combined bone scintigraphy/immunoscintigraphy with radiolabelled monoclonal antigranulocyte antibodies. Skeletal Radiol 2000;29:378–386.
  11. Kaim A, Maurer T, Ochsner P, Jundt G, Kirsch E, Mueller-Brand J. Chronic complicated osteomyelitis of the appendicular skeleton: diagnosis with technetium-99m labelled monoclonal antigranulocyte antibody-immunoscintigraphy. Eur J Nucl Med 1997;24:732–738.
  12. Seybold K, Frey LD, Locher J. Immunoscintigraphy of infections using 123I and 99mTc-labeled monoclonal antibodies: advanced experiences in 230 patients. Angiology 1992;43:85–90.
  13. Guhlmann A, Brecht-Krauss D, Suger G, et al. Fluorine-18-FDG PET and technetium-99m antigranulocyte antibody scintigraphy in chronic osteomyelitis. J Nucl Med 1998;39:2145–2152.
  14. Gratz S, Schipper ML, Dorner J, et al. LeukoScan for imaging infection in different clinical settings: a retrospective evaluation and extended review of the literature. Clin Nucl Med 2003;28:267–276.
  15. Sciuk J, Brandau W, Vollet B, et al. Comparison of technetium 99m polyclonal human immunoglobulin and technetium 99m monoclonal antibodies for imaging chronic osteomyelitis: first clinical results. Eur J Nucl Med 1991;18:401–407.
  16. Gallowitsch HJ, Heinisch M, Mikosch P, et al. Tc-99m ciprofloxacin in clinically selected patients for peripheral osteomyelitis, spondylodiscitis and fever of unknown origin: preliminary results. Nuklearmedizin 2002;41:30–36.
  17. Vicente AG, Almoguera M, Alonso JC, et al. Diagnosis of orthopedic infection in clinical practice using Tc-99m sulesomab (antigranulocyte monoclonal antibody fragment Fab'2). Clin Nucl Med 2004;29:781–785.
  18. Dominguez-Gadea L, Martin-Curto LM, de la Calle H, Crespo A. Diabetic foot infections: scintigraphic evaluation with 99Tcm-labelled anti-granulocyte antibodies. Nucl Med Commun 1993;14:212–218.
  19. Devillers A, Garin E, Polard JL, et al. Comparison of Tc-99m-labelled antileukocyte fragment Fab' and Tc-99m-HMPAO leukocyte scintigraphy in the diagnosis of bone and joint infections: a prospective study. Nucl Med Commun 2000;21:747–753.
  20. Rubello D, Casara D, Maran A, Avogaro A, Tiengo A, Muzzio PC. Role of anti-granulocyte Fab' fragment antibody scintigraphy (LeukoScan) in evaluating bone infection: acquisition protocol, interpretation criteria and clinical results. Nucl Med Commun 2004;25:39–47.
  21. Harwood SJ, Valdivia S, Hung GL, Quenzer RW. Use of Sulesomab, a radiolabeled antibody fragment, to detect osteomyelitis in diabetic patients with foot ulcers by leukoscintigraphy. Clin Infect Dis 1999;28:1200–1205.
  22. Ruther W, Hotze A, Moller F, Bockisch A, Heitzmann P, Biersack HJ. Diagnosis of bone and joint infection by leucocyte scintigraphy: a comparative study with 99mTc-HMPAO-labelled leucocytes, 99mTc-labelled antigranulocyte antibodies and 99mTc-labelled nanocolloid. Arch Orthop Trauma Surg 1990;110:26–32.
  23. Delcourt A, Huglo D, Prangere T, et al. Comparison between Leukoscan (Sulesomab) and Gallium-67 for the diagnosis of osteomyelitis in the diabetic foot. Diabetes Metab 2005;31:125–133.
  24. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25.
  25. Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med 1993;12:1293–1316.
  26. Jaeschke R, Guyatt G, Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid? Evidence-Based Medicine Working Group. JAMA 1994;271:389–391.
  27. Rutter CM, Gatsonis CA. A hierarchical regression approach to meta-analysis of diagnostic test accuracy evaluations. Stat Med 2001;20:2865–2884.
  28. Harbord RM, Deeks JJ, Egger M, Whiting P, Sterne JA. A unification of models for meta-analysis of diagnostic accuracy studies. Biostatistics 2007;8:239–251.
  29. Littenberg B, Moses LE. Estimating diagnostic accuracy from multiple conflicting reports: a new meta-analytic method. Med Decis Making 1993;13:313–321.
  30. Lau J, Ioannidis JP, Terrin N, Schmid CH, Olkin I. The case of the misleading funnel plot. BMJ 2006;333:597–600.
  31. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58:882–893.
  32. Ryan PJ. Leukoscan for orthopaedic imaging in clinical practice. Nucl Med Commun 2002;23:707–714.
  33. Gratz S, Behr T, Herrmann A, et al. Intraindividual comparison of 99mTc-labelled anti-SSEA-1 antigranulocyte antibody and 99mTc-HMPAO labelled white blood cells for the imaging of infection. Eur J Nucl Med 1998;25:386–393.
  34. Peltier P, Potel G, Lovat E, Baron D, Chatal JF. Detection of lung and bone infection with anti-granulocyte monoclonal antibody BW 250/183 radiolabelled with 99Tcm. Nucl Med Commun 1993;14:766–774.
  35. D'Amico P, Lastoria S, Caccavella N, Salvatore M. Radiolabelled granulocytes in inflammatory bone disease. Int J Rad Appl Instrum B 1991;18:145–147.
  36. Thakur ML, Marcus CS, Henneman P, et al. Imaging inflammatory diseases with neutrophil-specific technetium-99m-labeled monoclonal antibody anti-SSEA-1. J Nucl Med 1996;37:1789–1795.
  37. Kroiss A, Bock F, Perneczky G, et al. Immunoscintigraphy for the detection of inflammation foci in bone and joint diseases [in German]. Wien Klin Wochenschr 1990;102:713–717.
  38. Gratz S, Braun HG, Behr TM, et al. Photopenia in chronic vertebral osteomyelitis with technetium-99m-antigranulocyte antibody (BW 250/183). J Nucl Med 1997;38:211–216.
  39. Vasiliou O, Ballaris V, Alexandris R, Gerasimou G, Liaros G, Konidari-Dedousi E. Comparative study of immunoscintigraphy with (99m)Tc-antibodies (Fab-fragment-leukoscantrade mark) and the X-ray and CT findings in patients with subacute osteomyelitis [in Greek]. Hell J Nucl Med 2004;7:83–87.
  40. Lind P, Langsteger W, Koltringer P, Dimai HP, Passl R, Eber O. Immunoscintigraphy of inflammatory processes with a technetium-99m-labeled monoclonal antigranulocyte antibody (MAb BW 250/183). J Nucl Med 1990;31:417–423.
  41. Hakki S, Harwood SJ, Morrissey MA, Camblin JG, Laven DL, Webster WB Jr. Comparative study of monoclonal antibody scan in diagnosing orthopaedic infection. Clin Orthop Relat Res 1997;335:275–285.
  42. Palestro CJ, Kipper SL, Weiland FL, Love C, Tomas MB. Osteomyelitis: diagnosis with (99m)Tc-labeled antigranulocyte antibodies compared with diagnosis with (111)In-labeled leukocytes—initial experience. Radiology 2002;223:758–764.
  43. Termaat MF, Raijmakers PG, Scholten HJ, Bakker FC, Patka P, Haarman HJ. The accuracy of diagnostic imaging for the assessment of chronic osteomyelitis: a systematic review and meta-analysis. J Bone Joint Surg Am 2005;87:2464–2471.
  44. Schauwecker DS. Osteomyelitis: diagnosis with In-111-labeled leukocytes. Radiology 1989;171:141–146.
  45. Love C, Palestro CJ. 99mTc-fanolesomab Palatin Technologies. IDrugs 2003;6:1079–1085.
  46. Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Radiology 2003;226:24–28.




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2452061877v1
245/3/732    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pakos, E. E.
Right arrow Articles by Ioannidis, J. P. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pakos, E. E.
Right arrow Articles by Ioannidis, J. P. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE