Diagnostic Performance of PET or PET/CT Using 18F-FDG Labeled White Blood Cells in Infectious Diseases: A Systematic Review and a Bivariate Meta-Analysis

Background: Diagnostic performance of positron emission tomography using white blood cells labeled with fluorine-18-fluorodeoxyglucose (18F-FDG-WBC PET or PET/CT) in patients with suspicious infectious diseases has been evaluated in several studies; however, there is no consensus about the diagnostic accuracy of this method. Therefore, a systematic review and meta-analysis was carried out on this topic. Methods: A comprehensive computer literature search screening PubMed/MEDLINE, Embase and Cochrane library databases through March 2019 was performed. Pooled sensitivity, specificity, positive and negative likelihood ratios (LR+ and LR−), and diagnostic odds ratio (DOR) of 18F-FDG-WBC PET or PET/CT in patients with infectious diseases were calculated. Results: Eight studies on the use of 18F-FDG-WBC PET or PET/CT in suspicious infectious diseases were discussed in the systematic review. The meta-analysis of seven studies (236 patients) provided these pooled results on a per patient-based analysis: sensitivity was 86.3% [95% confidence interval (95%CI) 75–92.9%], specificity 92% (95%CI 79.8–97.1%), LR+ 6.6 (95%CI: 3.1–14.1), LR− 0.2 (95%CI: 0.12–0.33), DOR 43.5 (95%CI: 12.2–155). A statistically significant heterogeneity was not detected. Conclusions: Despite limited literature data, 18F-FDG-WBC PET or PET/CT demonstrated a good diagnostic accuracy for the diagnosis of infectious diseases; nevertheless, larger studies are needed.


Introduction
Infectious diseases are a frequent cause of morbidity and mortality worldwide [1]. Early and accurate diagnosis of infectious diseases can be difficult and time-consuming, whereas a delayed diagnosis can be life-threatening. Accurate and early detection and localization of infectious diseases is crucial for patient management and treatment, as well as for the cost containment [2,3].

Search Strategy
The co-authors performed a comprehensive computer literature search of PubMed/MEDLINE, Cochrane library and Embase databases to find relevant published studies on the diagnostic accuracy of 18 F-FDG-WBC PET or PET/CT in patients with suspicious infectious diseases.
This search algorithm based on a combination of terms was created and used: (A) "FDG" OR "fluorodeoxyglucose" AND (B) "label*" AND (C) "leukocyte*" OR "leucocyte*" OR "white blood" OR "granulocyte*". No beginning date limit nor language restrictions were used. The literature search was updated until 26 March 2019. References of the retrieved articles were also screened to search for possible additional articles.

Study Selection
Studies assessing the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients with suspicious infectious diseases were eligible for inclusion in the qualitative (systematic review) and quantitative analysis (meta-analysis). The exclusion criteria for the systematic review were: (a) articles not within the field of interest; (b) editorials or letters, review articles, comments, conference proceedings; and (c) case reports. If studies with possible patient data overlap were found, only the Diagnostics 2019, 9, 60 3 of 13 article with more complete information was included in the meta-analysis, whereas all of them were included in the systematic review.
All the co-authors independently screened the abstracts of the retrieved articles, applying the predefined inclusion and exclusion criteria. Subsequently, the researchers independently reviewed the full-text of the selected articles to assess their eligibility for inclusion. Disagreements were solved through a consensus meeting among all co-authors.

Data Extraction
For each selected study, information was collected on basic study characteristics (authors, year of publication, country, study design), patient characteristics (type and number of patients evaluated, age and sex ratio), technical details (type of hybrid imaging used, activity used for WBC labeling, WBC labeling efficiency, injected activity, time interval between radiotracer injection and image acquisition, image analysis and other imaging methods performed for comparison), and data on diagnostic accuracy on a per patient-based analysis (including true positive and true negative findings, false positive and false negative findings, sensitivity, specificity, positive and negative predictive values and accuracy).

Quality Assessment
The quality of the studies included in this systematic review was critically appraised using the revised "Quality Assessment of Diagnostic Accuracy Studies" tool (QUADAS-2) [10]. QUADAS-2 includes four domains (patient selection, index test, reference standard, and flow and timing) and each domain was assessed in terms of risk of bias, and the first three domains were also assessed in terms of concerns regarding applicability [10].

Statistical Analysis
Sensitivity and specificity, positive and negative likelihood ratios (LR+ and LR−) and diagnostic odds ratio (DOR) of visual analysis of 18 F-FDG-WBC PET or PET/CT in patients with suspicious infectious diseases were obtained from individual studies on a per patient-based analysis. A bivariate random-effects model was used for statistical pooling of data about sensitivity and specificity. This statistical approach takes into account any correlation that may exist between sensitivity and specificity [11]. A random-effects model was used for statistical pooling of LR+, LR− and DOR. Pooled data were presented with 95% confidence intervals (95%CI) and displayed using forest plots. Heterogeneity has been estimated by using the I-square index (I 2 ) [12]. Publication bias was assessed through the Egger's test [13].
Statistical analyses were performed using OpenMeta[Analyst]®software (version 0.1503) funded by the Agency for Healthcare Research and Quality (AHRQ) (Rockville, Maryland, United States).

Literature Search
Literature search results are summarized in Figure 1. A total of 160 records were identified through the comprehensive computer literature search of PubMed/MEDLINE, Cochrane library and Embase databases. Screening 160 abstracts, 148 records were excluded: 130 because they were not in the field of interest, 11 as editorials, reviews or letters, seven as case reports. Twelve articles were selected and retrieved in full-text. No additional records were found screening the references of these articles, whereas four articles were excluded after the analysis of the full text. Therefore, eight articles were included in the qualitative analysis (systematic review) [14][15][16][17][18][19][20][21]. One article was excluded from the meta-analysis for possible patient data overlap [16]. Overall, seven articles (236 patients with suspicious infectious diseases) were included in the quantitative analysis (meta-analysis) [14,15,[17][18][19][20][21]. The characteristics of the studies included in the systematic review  Tables 1 and 2. Diagnostic accuracy data from these articles are showed in Table 3, whereas the overall quality assessment of the studies is reported in Figure 2. A total of 160 records were identified through the comprehensive computer literature search of PubMed/MEDLINE, Cochrane library and Embase databases. Screening 160 abstracts, 148 records were excluded: 130 because they were not in the field of interest, 11 as editorials, reviews or letters, seven as case reports. Twelve articles were selected and retrieved in full-text. No additional records were found screening the references of these articles, whereas four articles were excluded after the analysis of the full text. Therefore, eight articles were included in the qualitative analysis (systematic review) [14][15][16][17][18][19][20][21]. One article was excluded from the meta-analysis for possible patient data overlap [16]. Overall, seven articles (236 patients with suspicious infectious diseases) were included in the quantitative analysis (meta-analysis) [14,15,[17][18][19][20][21]. The characteristics of the studies included in the systematic review are presented in Tables 1 and 2. Diagnostic accuracy data from these articles are showed in Table 3, whereas the overall quality assessment of the studies is reported in Figure 2.  Legend: 18     Legend: NA = not available; * = excluded from the meta-analysis for possible data overlap; NPV = negative predictive value; PPV = positive predictive value.

Basic Study and Patient Characteristics
Screening the selected databases, eight articles evaluating the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients wih suspicious infectious diseases were selected (Table 1) [14][15][16][17][18][19][20][21]. All the selected articles were prospective single-center studies published from 2006 to 2018 by research groups of different countries from Europe, Asia and America. The mean age of patients included in these studies ranged from 41 to 61 years and the percentage of male patients (sex ratio) ranged from 22% to 96%.

Technical Aspects
Technical details about 18 F-FDG-WBC PET or PET/CT in the included studies are summarized in Table 2. Hybrid PET/CT was performed in 87.5% of the studies, without contrast-enhanced CT in the majority of cases. Mean WBC labeling efficiency ranged from 70% to 81%.
The injected activity and the time interval between injection and PET acquisition were quite different among the studies. The analysis of PET and PET/CT images was performed by using qualitative criteria (visual analysis) in all the studies. Some authors have used visual scores for PET images, comparing the activity in suspicious lesions and in background regions. Additional

Basic Study and Patient Characteristics
Screening the selected databases, eight articles evaluating the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients wih suspicious infectious diseases were selected (Table 1) [14][15][16][17][18][19][20][21]. All the selected articles were prospective single-center studies published from 2006 to 2018 by research groups of different countries from Europe, Asia and America. The mean age of patients included in these studies ranged from 41 to 61 years and the percentage of male patients (sex ratio) ranged from 22% to 96%.

Technical Aspects
Technical details about 18 F-FDG-WBC PET or PET/CT in the included studies are summarized in Table 2. Hybrid PET/CT was performed in 87.5% of the studies, without contrast-enhanced CT in the majority of cases. Mean WBC labeling efficiency ranged from 70% to 81%.
The injected activity and the time interval between injection and PET acquisition were quite different among the studies. The analysis of PET and PET/CT images was performed by using qualitative criteria (visual analysis) in all the studies. Some authors have used visual scores for PET images, comparing the activity in suspicious lesions and in background regions. Additional semi-quantitative criteria, i.e., through the calculation of the maximal standardized uptake values (SUV max ), were used in 62% of studies. At visual analysis all the areas of focal increased radiotracer uptake greater than the surrounding tissue and not judged as physiological activity were considered to be abnormal. Normal physiologic biodistribution of 18 F-FDG-WBC includes the reticuloendothelial system (liver, spleen and bone marrow). Minimal radiotracer activity may be detected in brain, myocardium and urinary bladder (due to 18 F-FDG eluted from radiolabeled WBC). Compared to 18 F-FDG PET, no physiological activity is usually demonstrated in bowel, kidneys or ureters by using 18 F-FDG-WBC PET [14][15][16][17][18][19][20][21].
Histopathological/microbiological results and/or clinical/imaging/biochemical work-up and follow-up were used as reference standard in the included studies.
In most of the included studies a good diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients wih suspicious infection (including musculoskeletal, vascular or soft-tissue infections) was reported (Table 3).
False negative findings of 18 F-FDG-WBC PET or PET/CT were due to poor host immune reaction, low virulence or chronic infections, vertebral osteomyelitis, infections in site of physiological radiotracer uptake, prior antibiotic therapy or immunosuppressant drugs administration. False positive findings of 18 F-FDG-WBC PET or PET/CT were due to hemorrhagic lesions, some cases of aseptic/sterile inflammation, recent surgery or eluition of 18 F-FDG from radiolabeled WBC over time [14][15][16][17][18][19][20][21].
Histopathological/microbiological results and/or clinical/imaging/biochemical work-up and follow-up were used as reference standard in the included studies.
In most of the included studies a good diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients wih suspicious infection (including musculoskeletal, vascular or soft-tissue infections) was reported (Table 3).
False negative findings of 18 F-FDG-WBC PET or PET/CT were due to poor host immune reaction, low virulence or chronic infections, vertebral osteomyelitis, infections in site of physiological radiotracer uptake, prior antibiotic therapy or immunosuppressant drugs administration. False positive findings of 18 F-FDG-WBC PET or PET/CT were due to hemorrhagic lesions, some cases of aseptic/sterile inflammation, recent surgery or eluition of 18 F-FDG from radiolabeled WBC over time [14][15][16][17][18][19][20][21].

Discussion
Osman and Danpure first described the in vitro labeling of WBC with 18 F-FDG, showing that 87% of 18 F-FDG is associated with labeling of granulocytes [22]. Subsequently, several studies have evaluated the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in different types of infectious diseases [14][15][16][17][18][19][20][21], but these studies have limited power due to the small number of patients enrolled. Therefore, we have pooled data reported in the published studies through a bivariate meta-analysis to derive more robust estimates on the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in this setting.
Overall, despite the relatively limited data, our systematic review and bivariate meta-analysis demonstrated a good diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients with suspicious infectious diseases without significant adverse effects. To this regard, possible false negative findings (i.e., due to poor host immune reaction, low virulence or chronic pattern of disease, vertebral osteomyelitis, infections in site of physiological radiotracer uptake, prior antibiotic therapy or immunosuppressant drugs administration) and false positive findings (i.e., due to hemorrhagic lesions, some cases of aseptic/sterile inflammation, recent surgery or eluition of 18 F-FDG from radiolabeled WBC over time) should be taken into account when nuclear medicine physicians interpret 18 F-FDG-WBC PET or PET/CT images [14][15][16][17][18][19][20][21].
About the comparison of 18 F-FDG-WBC PET or PET/CT with other imaging modalities in patients with suspicious infectious diseases, a higher diagnostic performance was found using

Discussion
Osman and Danpure first described the in vitro labeling of WBC with 18 F-FDG, showing that 87% of 18 F-FDG is associated with labeling of granulocytes [22]. Subsequently, several studies have evaluated the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in different types of infectious diseases [14][15][16][17][18][19][20][21], but these studies have limited power due to the small number of patients enrolled. Therefore, we have pooled data reported in the published studies through a bivariate meta-analysis to derive more robust estimates on the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in this setting.
Overall, despite the relatively limited data, our systematic review and bivariate meta-analysis demonstrated a good diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients with suspicious infectious diseases without significant adverse effects. To this regard, possible false negative findings (i.e., due to poor host immune reaction, low virulence or chronic pattern of disease, vertebral osteomyelitis, infections in site of physiological radiotracer uptake, prior antibiotic therapy or immunosuppressant drugs administration) and false positive findings (i.e., due to hemorrhagic lesions, some cases of aseptic/sterile inflammation, recent surgery or eluition of 18 F-FDG from radiolabeled WBC over time) should be taken into account when nuclear medicine physicians interpret 18 F-FDG-WBC PET or PET/CT images [14][15][16][17][18][19][20][21].

Discussion
Osman and Danpure first described the in vitro labeling of WBC with 18 F-FDG, showing that 87% of 18 F-FDG is associated with labeling of granulocytes [22]. Subsequently, several studies have evaluated the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in different types of infectious diseases [14][15][16][17][18][19][20][21], but these studies have limited power due to the small number of patients enrolled. Therefore, we have pooled data reported in the published studies through a bivariate meta-analysis to derive more robust estimates on the diagnostic performance of 18 F-FDG-WBC PET or PET/CT in this setting.
Overall, despite the relatively limited data, our systematic review and bivariate meta-analysis demonstrated a good diagnostic performance of 18 F-FDG-WBC PET or PET/CT in patients with suspicious infectious diseases without significant adverse effects. To this regard, possible false negative findings (i.e., due to poor host immune reaction, low virulence or chronic pattern of disease, vertebral osteomyelitis, infections in site of physiological radiotracer uptake, prior antibiotic therapy or immunosuppressant drugs administration) and false positive findings (i.e., due to hemorrhagic lesions, some cases of aseptic/sterile inflammation, recent surgery or eluition of 18 F-FDG from radiolabeled WBC over time) should be taken into account when nuclear medicine physicians interpret 18 F-FDG-WBC PET or PET/CT images [14][15][16][17][18][19][20][21].
About the comparison of 18 F-FDG-WBC PET or PET/CT with other imaging modalities in patients with suspicious infectious diseases, a higher diagnostic performance was found using 18 F-FDG-WBC PET/CT compared to conventional imaging methods (CT or MRI) and 18 F-FDG PET/CT [14][15][16][17]19].
The normal biodistribution of 18 F-FDG includes the brain and the genitourinary tract and variably high activity in the myocardium, bone marrow, gastro-intestinal tract and liver. 18 F-FDG PET/CT may have a reduced performance for the detection of infection in these sites [7]. Conversely, the physiological uptake of 18 F-FDG-WBC essentially occurs within the reticuloendothelial system. The absence of gastrointestinal and renal uptake and the faint brain and myocardial uptake makes 18 F-FDG-WBC PET/CT a suitable imaging method for the assessment of intra-abdominal, renal, cardiac and cerebral infectious diseases [14][15][16][17][18][19][20][21]. Another difference between 18 F-FDG-WBC PET/CT and 18 F-FDG PET/CT relates to the different cellular types involved in the signal detected in infected sites. 18 F-FDG reveals different types of inflammatory cells, including macrophages, whereas 18 F-FDG-WBC essentially reveals active diapedesis of granulocytes through chemotactic processes [20,23], and this could explain the higher diagnostic performance obtained by using 18 F-FDG-WBC PET/CT compared to 18 F-FDG PET/CT in patients with suspicious infectious diseases [17,19].
Only one study compared 18 F-FDG-WBC PET and 111 In-WBC scintigraphy in patients with suspicious infectious diseases reporting a similar diagnostic performance of these imaging methods [21]. However, further considerations should be added to this regard. The labeling efficiency of 18 F-FDG-WBC is variable and significantly lower than that of WBC radiolabeled with gamma-emitting tracers [15,[17][18][19][20][21]; although not examined in detail, the leukocyte glucose transporter expression, the serum glucose levels, and the presence of intrinsic proteins have been thought to affect the labeling efficiency of 18 F-FDG-WBC [3]. Nevertheless, the relative low labeling efficiency seems to not affect the diagnostic accuracy of 18 F-FDG-WBC PET or PET/CT in detecting infectious diseases. On the other hand, the mean cell viability of 18 F-FDG-WBC was very high and comparable to that of radiolabelled WBC used for scintigraphy [18,21]; thus, the 18 F-FDG labeling procedure does not affect the WBC viability. Another factor that should be considered is the short physical half-life of 18 F which excludes imaging at 24 h after radiotracer injection (as usually performed for radiolabeled WBC scintigraphy); the WBC labeling and loading time may also have an impact due to the fast radiotracer decay. Some advantages of 18 F-FDG-WBC PET/CT compared to radiolabelled WBC scintigraphy or SPECT are the better image quality and resolution and the reduced imaging time [21]. Dosimetry of 18 F-FDG-WBC PET for activities of 225-315 MBq was found to be comparable to results with 111 In-WBC scintigraphy [24].
Although the short half-life of 18 F excludes imaging at 24 h after radiotracer injection (which is known to increase the specificity of radiolabeled WBC scintigraphy), a high sensitivity and specificity for infection was found in our pooled analysis on 18 F-FDG-WBC PET or PET/CT; the better resolution of PET imaging compared to planar scintigraphy and SPECT may have contributed to the high sensitivity and the high level of anatomic detail provided by the co-registered CT images may have contributed to the high specificity [20].
Contrast enhancement could further improve the diagnostic performance of 18 F-FDG-WBC PET/CT for detection of infection, in particular for visceral localizations, but this should be better evaluated in further studies [18,20]. Moreover, the use of semi-quantitative criteria of interpretation of 18 F-FDG-WBC PET/CT (by using SUVmax thresholds) seems to be not sufficiently accurate to make a differential diagnosis between infectious and non-infectious conditions, but semi-quantitative analysis can be used as an adjunct tool to visual analysis for PET interpretation, even if a possible effect of cell division on the SUVmax within the imaging time frame cannot be excluded [17].
Diagnostic accuracy of an imaging test is not a measure of clinical effectiveness and high diagnostic performance does not necessarily result in improved patient outcomes. Other factors beyond the diagnostic performance should influence the choice of an imaging modality in patients with suspicious infectious disease (i.e., availability, radiation dose, safety, examination time, legal, organization, economic aspects). Overall, our systematic review and meta-analysis demonstrated a good diagnostic performance of 18 F-FDG-WBC PET/CT in patients with suspicious infectious diseases, but larger prospective and multicenter studies on this topic are needed, in particular, more comparison studies with radiolabeled WBC scintigraphy are warranted. There are some drawbacks of 18 F-FDG-WBC PET/CT that should be underlined, including the relatively long time for WBC labeling, the need for high 18 F-FDG activities for labeling, variable labeling efficiency, risk of contamination by direct contact with blood products and radiation exposure [3,6,7,19]. Furthermore, the possibility of an increased risk of lymphoid malignancies associated with the administration of 18 F-FDG-WBC is a very controversial subject that needs further investigation [25]. Lastly, there are still insufficient in vitro data on putative elution of 18 F-FDG from radiolabeled WBC over time [7].
Some limitations and biases of our meta-analysis should be taken into account. First of all, a limited number of studies were available for the meta-analysis. Moreover, as a composite reference standard was used in most of the studies a possible verification bias could not be excluded. Furthermore, based on the information provided in the studies selected for the meta analysis, a selection bias could be present. Heterogeneity among studies (i.e., due to baseline differences among the patients included, diversity in methodological aspects, and different study quality) may represent a potential source of bias in a meta-analysis [11]. We have not detected a statistically significant heterogeneity among the included studies in our meta-analysis, although the significant differences about the patient population evaluated and the technical details in the included studies.

Conclusions
Based on available literature data, 18 F-FDG-WBC PET or PET/CT seem to demonstrate a good diagnostic performance in detecting infectious diseases.
The literature on this topic is still limited and further investigations on 18 F-FDG-WBC PET/CT in patients with suspicious infectious diseases are warranted. Based on available data, this imaging method should not be considered as standard clinical practice, but it could be used in a research setting.