1

Title: Whole Body 18F-FDG-PET and 18F-FDG-PET/CT in Patients with Suspected

Paraneoplastic Syndrome: A Systematic Review and Meta-analysis of Diagnostic

Accuracy

Running title: FDG-PET/CT in Paraneoplastic Syndrome

Authors: Sara Sheikhbahaei1, Charles V Marcus1, Roberto S Fragomeni1, Steven P

Rowe1, Mehrbod S Javadi1, Lilja B Solnes1

Affiliations:

1. The Russell H. Morgan Department of Radiology and Radiological Science,

Johns Hopkins University School of Medicine, Baltimore, MD, US

Corresponding author: Lilja B Solnes

The Russell H. Morgan Department of Radiology and Radiological Science, JHOC 3009,

601 N Caroline St, Baltimore, MD 21287,

Ph: 410- 955 6989, Email: lsolnes1@jhmi.edu

First author: Sara Sheikhbahaei, Research Fellow

Department of Radiology and Radiological Sciences, JHOC 4230, 601 N. Caroline St,

Baltimore, MD, 21287, Email: ssheikh6@jhmi.edu

Word count: 4,822

Disclaimer: None, Financial support: None

Journal of Nuclear Medicine, published on December 15, 2016 as doi:10.2967/jnumed.116.183905

2

ABSTRACT

The purpose of this study was to assess the diagnostic performance of whole body 18F-

FDG-PET or 18F-FDG-PET/CT for detection of underlying malignancy in patients with

clinically suspected neurological and non-neurological paraneoplastic syndromes.

Methods: A systematic search was performed in PubMed (Medline), Embase, Scopus

(last updated November 2016) to identify relevant published studies reporting the

performance of 18F-FDG-PET or 18F-FDG-PET/CT in patients with suspected

paraneoplastic syndrome. Histopathologic confirmation or clinical follow-up was

considered as the reference standard. Pooled estimates (95% confidence intervals) of

sensitivity, specificity, diagnostic odds ratio (DOR) were calculated. A summary receiver-

operating-characteristics curve was constructed, and the area under the curve (AUC)

was determined along with the Q* index.

Results: Twenty-one studies including a total of 1,293 individual patients suspected of

having a para-neoplastic syndrome and who underwent 18F-FDG-PET or 18F-FDG-

PET/CT examinations met our inclusion criteria. There was moderate to high

heterogeneity among the included studies. The pooled sensitivity, specificity and DOR of

18F-FDG-PET or 18F-FDG-PET/CT for the detection of underlying malignancy were

0.81 (0.76 to 0.86), 0.88 (0.86 to 0.90) and 34.03 (18.76 to 61.72), respectively. The

AUC and the Q* index were 0.916 (standard error= 0.018) and 0.849, indicating

excellent diagnostic accuracy. The diagnostic accuracy was slightly improved after

excluding studies with high applicability concerns (AUC of 0.931, standard error= 0.020).

In a subgroup analysis, 18F-FDG-PET/CT was found to have significantly higher

specificity (0.89 versus 0.79) compared to 18F-FDG-PET alone, with no evidence of

significant difference in the overall performance (AUC of 0.930 versus 0.891, two-tailed

P-value for difference= 0.31).

3

Conclusion: This meta-analysis of available studies demonstrates that whole body 18F-

FDG-PET or 18F-FDG-PET/CT has high diagnostic accuracy and moderate to high

sensitivity and specificity for detection of underlying malignancy in patients suspected of

having a paraneoplastic syndrome.

Key-words: Paraneoplastic syndrome, 18F-FDG-PET, Sensitivity and Specificity, meta-

analysis

4

INTRODUCTION

Paraneoplastic syndromes are a rare systemic complication of malignancy that is not

directly caused by local tumor extension or metastases 1, 2. The pathogenesis of

paraneoplastic syndromes is not completely understood, yet, it is believed to be

mediated by an altered immune response to the presence of cancer1. Paraneoplastic

syndromes comprises a heterogeneous group of disorders that can affect any organ

system including the central and peripheral nervous systems as well as the

musculoskeletal, dermatologic, hematologic, endocrine or gastrointestinal systems2, 3.

The nervous system involvements are the most commonly reported paraneoplastic

syndrome 2, 4. An immunologic response to an antigenic target that is shared between a

component of the nervous system and tumor cells may be the underlying pathology1, 4.

Paraneoplastic symptoms can be the initial or the most prominent manifestation

of malignancy 2. Timely recognition of these symptoms and detecting the underlying

malignancy can play a critical role in the early treatment of tumor and can guide further

management and improve patients’ survival. Paraneoplastic signs and symptoms may

regress by treating the underlying malignancy. Thus, accurate imaging examinations are

essential in identifying the underlying malignancy whenever a paraneoplastic syndrome

is suspected 5. Anatomical cross-sectional imaging including computed tomography (CT)

and magnetic resonance imaging may fail to detect the underlying malignancy as the

majority of associated malignancies are small, often with only lymphatic metastases 3, 5.

2-deoxy-2-[18F]fluoro-D-glucose positron emission tomography (18F-FDG-PET)

and 18F-FDG-PET/CT have emerged as practical imaging modality that allows for the

detection of metabolically active tumors6. Several studies have attempted to address the

role of whole body 18F-FDG-PET and 18F-FDG PET/CT in the workup of patients with

suspected paraneoplastic manifestations 3, 5, 7-17. Although some of these studies have

shown promising results, they were mostly conducted in relatively small and pre-

5

selected patient populations. Variations existed in their methodological designs, patient

populations, and conclusions. Currently, there is no consensus on the value of whole

body 18F-FDG-PET and 18F-FDG-PET/CT in patients with suspected paraneoplastic

syndromes, warranting the need for a meta-analysis.

This study aims to review the literature systematically and to extract the summary

estimates of the diagnostic performances of whole body 18F-FDG-PET or 18F-FDG-

PET/CT for detection of underlying malignancy in patients with clinically suspected

neurological and non-neurological paraneoplastic syndromes. The results of this study

can provide further insight into the usefulness of whole body 18F-FDG-PET and 18F-

FDG PET/CT in patients with suspected paraneoplastic syndromes.

MATERIALS AND METHODS

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses

(PRISMA) statement was followed18.

Search Strategy

Systematic electronic searches of Medline (PubMed), Embase, Scopus and

abstract proceedings of major scientific meetings (SNMMI, RSNA, EANM) were

performed to identify relevant published studies evaluating the diagnostic performance of

whole body 18F-FDG-PET or 18F-FDG-PET/CT in detecting primary malignancies or

metastasis in patients presenting with a paraneoplastic syndrome. The search strategy

was based on the combination of following keywords; A) “paraneoplastic” OR “para

neoplastic” OR “PNS”; AND B) “PET” OR “Positron emission tomography” OR “18 f

fluorodeoxyglucose” OR “FDG”; AND C) “diagnostic” OR “sensitivity” OR “specificity”

A comprehensive search strategy was used without any restrictions on language

or publication status. The search was last updated in November 2016, without beginning

date limit.

6

Criteria for Study Consideration

Patients with clinically suspected neurological and non-neurological

paraneoplastic syndromes were eligible for inclusion. Diagnostic whole body 18F-FDG-

PET or 18F-FDG-PET/CT studies performed to detect occult malignancy were used as

an index test. Histopathologic confirmation or imaging follow-up after whole body 18F-

FDG-PET or 18F-FDG-PET/CT was considered as the reference standard.

Selection of Studies, Data Extraction, and Management

Two authors (S.S., C.M.) reviewed all records identified through the electronic

search. Studies were initially evaluated on the basis of the title and abstract of the

articles. Review articles, editorials, case-reports, and irrelevant citations were excluded

in the initial assessment. The full-texts of the potentially relevant publications were

retrieved for further consideration. All potentially eligible articles were independently

checked by two authors for predefined inclusion criteria.

To avoid duplication of data, only the most recent articles with no overlapping

study period were included when there was more than one published article from the

same institution17, 19.

Two authors (S.S., C.M.) independently extracted the following data from each

included study; bibliographic details, patient sampling and characteristics, number of

patients or scans, index test, reference standard, prevalence of underlying malignancy,

and the raw data to construct 2 × 2 contingency tables including the number of true-

positive, false-positive, true-negative, and false-negative patients/scans. The diagnostic

performance of whole body 18F-FDG-PET or 18F-FDG-PET/CT for detection of

malignancy was assessed by cross-relating index test results and the reference

standards. All data extracted by two review authors were compared in each step and

any discrepancies were resolved through consensus or by a third author (L.S.).

Assessment of Methodological Quality

7

The methodological qualities of the included studies were assessed using a

modified version of the Quality Assessment Tool for Diagnostic Accuracy (QUADAS-2)

as recommended by Cochrane Collaborations 20.

The tool consists of four domains, each domain appraises the risk of bias and

applicability/ generalizability of the study through a series of questions (Supplemental

Table 1). Two authors (S.S., C.M.) independently assessed the quality of the studies.

Disagreements were mediated by consensus.

Statistical Analysis and Data Synthesis

Studies with adequate data to reconstruct the 2 × 2 diagnostic table were

included in the quantitative analysis. The sensitivity and specificity, along with 95%

confidence intervals (CIs), were calculated for each study. The forest plots of sensitivity

and specificity were used to display the variations in the results of the individual studies.

Homogeneity among the studies was assessed using a chi-square test and p < 0.05

indicated statistically significant heterogeneity 21, 22. The I-squared index (I2) was

measured to quantify the degree of heterogeneity in studies. I2 lies from 0 to 100, and

the respective values around 0, 25, 50, and 75, indicate no, low, moderate, and high

heterogeneity among studies 23. We calculated the pooled estimates of sensitivity,

specificity, positive likelihood ratio (LR), negative LR, and diagnostic odds ratios (DOR)

using random effect (DerSimonian–Laird) assumptions24. The diagnostic tests with a

DOR more than 25 and 100 are considered moderately and highly accurate,

respectively25. The receiver operating characteristic curve (SROC) space is defined by

sensitivity (y-axis) and 1-specificity (x-axis), and each data point represents one

particular study. The area under the SROC curve (AUC) represented an overall

summary of the diagnostic test performance and presents the tradeoff between

sensitivity and specificity. An AUC value of 1.0 (100%) denotes a perfect discriminatory

ability for a diagnostic test22. To reduce the heterogeneity, subgroup analysis was

8

performed to determine the effect of imaging modality (18F-FDG-PET alone versus.

18F-FDG-PET/CT) on the diagnostic performance. The statistical significance of the

difference between two AUCs was estimated with the Hanley JA method (Two-tailed P-

value) 26.

Publication bias was assessed using funnel plots of standard error (SE) and

Egger’s regression intercept27. To calculate the adjusted DOR, a trim and fill technique

was used to account for the potentially missed studies27. Analyses were performed using

Meta-DiSc (version 1.4; Hospital Universitario Ramon y Cajal, Madrid, Spain),

comprehensive meta-analysis (CMA version 2, Biostat, Englewood, NJ, USA) and

MedCalc statistical software (version 16.8.3, Ostend, Belgium).

RESULTS

Search Results

Figure 1 illustrates the details of the study selection. A total of 887 relevant

records were identified through a comprehensive literature search. After screening titles

and abstracts, we excluded 851 non-relevant articles. To assess the eligibility of the

remaining 36 records, we retrieved the corresponding full-texts or contacted the study

authors inquiring more information. Ultimately, a total of 24 studies were included in the

qualitative assessment and 21 studies met our predefined criteria to be included in the

quantitative analysis 3, 5, 7-17, 28-37.

Study Characteristics and Methodological Quality Assessment

The details of the study design, patient demographics and characteristics,

paraneoplastic presentation, paraneoplastic antibody status, index test, proportion of

patients diagnosed with underlying malignancy and reference standard(s) of each

included study were summarized in supplemental Tables 2 and 3 3, 5, 7-17, 28-31. Among the

24 studies included in the qualitative assessment, four studies were prospective and 20

9

studies were retrospective. All studies enrolled patients with a suspected paraneoplastic

syndrome. The index test was whole body 18F-FDG-PET (seven studies), 18F-FDG-

PET combined with CT (one study), 18F-FDG-PET or 18F-FDG-PET/CT (one study) and

hybrid 18F-FDG-PET/CT (15 studies).

Three studies did not clearly describe the reference standard used in patients

with negative 18F-FDG-PET/CT results and were not included in the quantitative

analysis 29, 30, 36.

Twenty-one studies including 1,293 individual patients suspected to have

paraneoplastic syndrome assessed the diagnostic accuracy of 18F-FDG-PET or 18F-

FDG-PET/CT in the detection of underlying malignancy and were included in the

quantitative analysis. Detailed information on true positive, false negative and false

positive sites is summarized in supplemental Table 4. Figure 2 depicted the risk of bias

and applicability concerns across the included studies according to QUADAS-2.

Findings

The proportion of patients diagnosed with underlying malignancy among the

included studies ranged between 7.5% and 90%, with the pooled estimate of 16.9%

(95% CI= 15.0% to 19.1%).

The sensitivity of whole body 18F-FDG-PET or 18F-FDG-PET/CT in the

detection of underlying malignancy in patients with a suspected paraneoplastic

syndrome ranged between 0.57 and 1.00, with a pooled sensitivity of 0.81 (95%CI= 0.76

to 0.86) (Fig 3A). The specificity of whole body 18F-FDG-PET or 18F-FDG-PET/CT

ranged between 0.25 and 1.00, with a pooled specificity of 0.88 (95% CI= 0.86 to 0.90)

(Fig 3B). The random effects model was used because of moderate to substantial

heterogeneity among the included studies (I2= 57.7% for sensitivity and I2= 81.5% for

specificity). The pooled results for the positive LR, negative LR, and DOR of whole body

10

18F-FDG -PET or 18F-FDG -PET/CT were 5.95 (95% CI= 4.01 to 8.84), 0.25 (95% CI=

0.18 to 0.35) and 34.03 (95% CI= 18.76 to 61.72), respectively.

The SROC curve summarizing the accuracy of whole body 18F-FDG -PET or

18F-FDG-PET/CT examinations across all included studies is depicted in Figure 4. The

AUC was 0.916 (SE= 0.018) and the Q* index was 0.849 indicating excellent diagnostic

accuracy.

We performed a secondary analysis after excluding studies with high applicability

concerns in all domains9, 10, 12, 13, 31. The pooled estimates of sensitivity, specificity and

DOR were 0.88 (95% CI= 0.82 to 0.93), I2 = 24.9%; 0.87 (95%CI= 0.84 to 0.89), I2 =

81.2%; and 44.86 (95%CI= 19.72 to 102.07), respectively. The diagnostic performance

was slightly improved with the AUC of 0.931 (SE= 0.020) and Q* index of 0.866.

Patients with Neurologic Paraneoplastic Syndromes

In a subgroup analysis, we tried to include only patients with exclusively

neurologic paraneoplastic symptoms, where possible. A total of twelve studies on 528

patients were included. The pooled estimates of sensitivity and specificity for 18F-FDG -

PET or 18F-FDG -PET/CT were 0.89 (95%CI= 0.81 to 0.94), I2=38.4% and 0.83

(95%CI= 0.79 to 0.87), I2= 76.9%, respectively. The pooled result for the positive LR,

negative LR, and DOR of 18F-FDG-PET or 18F-FDG-PET/CT were 4.47 (95% CI= 2.7

to 7.40), 0.25 (95% CI= 0.16 to 0.40), and 26.99 (95% CI= 11.17 to 65.23), respectively.

The AUC was 0.915 (SE= 0.028) and the Q* index was 0.848, indicating excellent

diagnostic accuracy.

Whole Body 18F-FDG -PET/CT versus 18F-FDG -PET Alone

In studies in which 18F-FDG-PET alone (7 studies, n= 273) was used, the

pooled sensitivity and specificity were 0.88 (95% CI= 0.78 to 0.95), I2= 0% and 0.79

(95% CI= 0.73 to 0.84), I2= 60.7%, respectively. In studies in which combined 18F-FDG-

PET/CT (13 studies, n= 895) was performed, the pooled sensitivity and specificity were

11

0.77 (95% CI=0.70 to 0.84), I2= 66.9% and 0.89 (95%CI= 0.87 to 0.91), I2= 76.5%,

respectively. 18F-FDG-PET/CT had significantly higher specificity compared to 18F-

FDG-PET alone (no overlap in 95% CIs), but there was no statistically significant

difference in the pooled sensitivity. The pooled result for the positive LR, negative LR

and DOR were 3.5 (95% CI= 2.06 to 5.93), 0.22 (95% CI= 0.12 to 0.41) and 19.26 (95%

CI= 8.45 to 43.92) for FDG-PET alone and 7.17 (95% CI= 4.59 to 11.21), 0.26 (95% CI=

0.17 to 0.42) and 37.34 (95% CI= 17.63 to 79.06) for FDG-PET/CT. There were no

statistically significant differences between 18F-FDG-PET/CT and 18F-FDG-PET in the

AUC with the respective values of 0.930 (SE= 0.023) vs. 0.891 (SE= 0.031); Two-tailed

P-value for difference= 0.31.

Risk of Publication Bias

Figure 5 demonstrates the funnel plot of the included studies. The log-DOR on the x-axis

is plotted against the SE of the log-DOR on the y-axis. Egger’s regression intercepts for

DOR pooling were 1.76 (95% CI= 0.62 to 2.89; two-tailed P-value= 0.004) suggesting

the presence of publication bias. After adjustment for the potential effect of publication

bias, looking for missing studies to the left of mean- trim and fill technique-, the

estimated adjusted DOR was decreased to 18.89 (95% CI= 10.01 to 35.65).

DISCUSSION

This is the first meta-analysis to evaluate the diagnostic performance of whole

body 18F-FDG-PET or 18F-FDG-PET/CT in the detection of underlying malignancy in

patients with paraneoplastic manifestations. This meta-analysis demonstrated a pooled

sensitivity of 0.81, specificity of 0.88 and moderate DOR for 18F-FDG-PET or 18F-FDG-

PET/CT in patients suspected to have a paraneoplastic syndrome. The SROC curve

analysis yielded an excellent trade-off between sensitivity and specificity with an AUC of

0.916.

12

There is considerable heterogeneity in the literature and yet no consensus on the

value of whole body 18F-FDG-PET or 18F-FDG-PET/CT in patients suspected of

harboring a paraneoplastic syndrome. Our results showed substantial difference

between the pre-pooled sensitivity and specificity of individual studies. This can be partly

explained by different patient inclusion criteria, as well as different methods used for the

reference standard. The selection bias appeared as some studies included

heterogeneous group of patients with different manifestations and with or without

inconclusive/negative prior conventional imaging. Others only included highly selective

patients with positive paraneoplastic-antibodies. The pretest-selection of patients

suspected for paraneoplastic syndrome can alter the diagnostic performance of whole

body 18F-FDG-PET or 18F-FDG-PET/CT. We tried to reduce the effect of clinical

diversity and heterogeneity by performing a secondary analysis after exclusion of studies

with high risk of bias. The performance of 18F-FDG-PET or 18F-FDG-PET/CT in the

detection of underlying malignancy was slightly improved (AUC =0.931; sensitivity of

0.88) in this setting. In addition, 18F-FDG-PET or 18F-FDG-PET/CT was found to have

fairly similar diagnostic performance when limiting the analysis to patients with

neurologic paraneoplastic manifestations.

Our results showed a false negative rate of 19% for whole body 18F-FDG-PET or

18F-FDG-PET/CT. Therefore, a negative 18F-FDG-PET/CT cannot rule out the

presence of malignancy. Patients with clinically suspected paraneoplastic syndrome can

have relatively small tumors that may be missed by 18F-FDG-PET/CT or tumors may

not be 18F-FDG avid. Therefore, careful follow-up and repeat screening by 18F-FDG-

PET/CT or other imaging modalities is necessary in these patients. The European

Federation of the Neurological Societies has addressed this by recommending repeat

screening in 3 to 6 months after an initial negative evaluation, followed by screening

every 6 months up to four years, if testing remains unrevealing38.

13

Sets of definitions are taken into consideration to establish diagnostic criteria and

classify patients with paraneoplastic manifestations into “definite” or “possible”

paraneoplastic syndrome categories, mainly based on the presenting clinical symptoms

and presence or absence of paraneoplastic antibodies39. The presence of the well-

characterized paraneoplastic antibodies provide important additional diagnostic

information to predict a specific underlying malignancy1, 39. Recent studies investigated

the performance of whole body 18F-FDG-PET/CT in relation to paraneoplastic antibody

status16. A number of studies showed that whole body 18F-FDG-PET and 18F-FDG-

PET/CT are highly useful in the screening of patients with suspected paraneoplastic

syndrome and positive paraneoplastic antibodies9, 10. However, a recent study by

Vatankulua et al suggested that paraneoplastic antibody presence should not be an

indispensable factor for performing 18F-FDG-PET/CT in patients suspected of harboring

a paraneoplastic syndrome16.

Besides, Vaidyanathan et al. showed that 18F-FDG-PET/CT can help in risk

classification of patients suspected of having a paraneoplastic syndrome. They showed

that the 18F-FDG-PET/CT results could add confidence to clinical likelihood in 28% of

patients and correctly downgrade (16%) or upgrade (12%) the clinical score5. A recent

study by Subramaniam et al. showed that 18F-FDG-PET has substantial impact in the

management of patients with suspected para-neoplastic syndrome and those with

cancer of unknown primary origin, resulted in changing the intended management in

25% and 43% of patients 40. Our meta-analysis showed that the diagnostic performance

of FDG-PET/CT for the detection of underlying malignancy in patients with suspected

paraneoplastic syndrome is comparable with its’ general performance in patients with

cancer of unknown primary origin. A previous meta-analysis of 11 studies including 433

patients with cancer of unknown primary reported a sensitivity and specificity of 84%

(95% CI 78–88%) and 84% (95% CI 78–89%) for FDG-PET/CT in the primary tumor

14

detection41. However, the translation of test performance characteristics into improved

patient management has not been addressed in this meta-analysis and needs to be

further investigated.

The development of the integrated 18F-FDG-PET and CT has helped to

precisely localized lesions on 18F-FDG-PET scans6. Over the last decade, more studies

have investigated the application of combined 18F-FDG-PET/CT in the initial evaluation

and diagnosis of paraneoplastic syndromes and occult malignancy3, 5, 14-17. In our study,

the subgroup analysis by imaging modality revealed that 18F-FDG-PET/CT showed

significantly higher specificity compared with 18F-FDG-PET alone. As 18F-FDG-PET

results are only based on the visual analysis of lesions with no anatomical correlate, any

lesions with higher metabolism than background could have been counted as positive6.

However, we found no evidence of significant difference in the sensitivity and overall

performances (AUC) of whole body 18F-FDG-PET vs. 18F-FDG-PET/CT. A possible

explanation can be the effect of publication year and the fact that older studies were

performed on smaller sample sizes.

Due to the limited number of studies with direct comparison of 18F-FDG-PET or

18F-FDG-PET/CT and conventional imaging, we were not able to report the pooled

diagnostic performance for conventional imaging. Five studies compared the diagnostic

performance of 18F-FDG-PET or 18F-FDG-PET/CT with conventional screening

methods9, 11, 13, 15, 28. In these studies, the sensitivity and specificity of conventional

screening modalities were ranged between 30- 82% and 71- 100%, respectively. Further

studies are needed to investigate the additional value of 18F-FDG-PET/CT and its cost

effectiveness over conventional screening modalities in the work-up of patients with

suspected paraneoplastic syndrome.

Limitations of this analysis include the risk of selection bias and the lack of

histopathologic confirmation of all lesions reported in the included studies. Although

15

clinical follow-up data were recorded in some studies, patients who died shortly after the

18F-FDG-PET or 18F-FDG-PET/CT without proven malignancy were considered as

false positive, which could underestimate the performance of the modality. We cannot

exclude the possibility of positive result publication bias as non-significant or unfavorable

study results tend to be discarded. Besides, most results of the meta-analysis showed

statistic heterogeneity. We analyzed the subgroups according to the index test and those

patients presenting with neurologic symptoms, and this partly eliminated the effect of

heterogeneity although further subgroup analyses were limited by the restricted original

data. Surely, larger observational studies are warranted to further determine the utility of

whole body 18F-FDG-PET and 18F-FDG-PET/CT in relation to paraneoplastic antibody

status and its effect on patient management and outcome.

CONCLUSION

This meta-analysis of available studies demonstrates that whole body 18F-FDG-

PET and 18F-FDG-PET/CT have excellent diagnostic accuracy, and moderate to high

sensitivity and specificity for detection of underlying malignancy in patients suspected of

having a paraneoplastic syndrome.

16

REFERENCES

1. Storstein A, Vedeler CA. Paraneoplastic neurological syndromes and onconeural

antibodies: clinical and immunological aspects. Adv Clin Chem. 2007;44:143-

185.

2. Pelosof LC, Gerber DE. Paraneoplastic syndromes: an approach to diagnosis

and treatment. Mayo Clin Proc. Sep 2010;85(9):838-854.

3. Kristensen SB, Hess S, Petersen H, Hoilund-Carlsen PF. Clinical value of FDG-

PET/CT in suspected paraneoplastic syndromes: a retrospective analysis of 137

patients. Eur J Nucl Med Mol Imaging. Dec 2015;42(13):2056-2063.

4. de Beukelaar JW, Sillevis Smitt PA. Managing paraneoplastic neurological

disorders. Oncologist. Mar 2006;11(3):292-305.

5. Vaidyanathan S, Pennington C, Ng CY, Poon FW, Han S. 18F-FDG PET-CT in

the evaluation of paraneoplastic syndromes: experience at a regional oncology

centre. Nucl Med Commun. Aug 2012;33(8):872-880.

6. von Schulthess GK, Steinert HC, Hany TF. Integrated PET/CT: current

applications and future directions. Radiology. Feb 2006;238(2):405-422.

7. Rees JH, Hain SF, Johnson MR, et al. The role of [18F]fluoro-2-deoxyglucose-

PET scanning in the diagnosis of paraneoplastic neurological disorders. Brain.

Nov 2001;124(Pt 11):2223-2231.

8. Berner U, Menzel C, Rinne D, et al. Paraneoplastic syndromes: detection of

malignant tumors using [(18)F]FDG-PET. Q J Nucl Med. Jun 2003;47(2):85-89.

9. Linke R, Schroeder M, Helmberger T, Voltz R. Antibody-positive paraneoplastic

neurologic syndromes: value of CT and PET for tumor diagnosis. Neurology. Jul

27 2004;63(2):282-286.

17

10. Younes-Mhenni S, Janier MF, Cinotti L, et al. FDG-PET improves tumour

detection in patients with paraneoplastic neurological syndromes. Brain. Oct

2004;127(Pt 10):2331-2338.

11. Patel RR, Subramaniam RM, Mandrekar JN, Hammack JE, Lowe VJ, Jett JR.

Occult malignancy in patients with suspected paraneoplastic neurologic

syndromes: value of positron emission tomography in diagnosis. Mayo Clin Proc.

Aug 2008;83(8):917-922.

12. Hadjivassiliou M, Alder SJ, Van Beek EJ, et al. PET scan in clinically suspected

paraneoplastic neurological syndromes: a 6-year prospective study in a regional

neuroscience unit. Acta Neurol Scand. Mar 2009;119(3):186-193.

13. Selva-O'Callaghan A, Grau JM, Gamez-Cenzano C, et al. Conventional cancer

screening versus PET/CT in dermatomyositis/polymyositis. Am J Med. Jun

2010;123(6):558-562.

14. Matsuhisa A, Toriihara A, Kubota K, Makino T, Mizusawa H, Shibuya H. Utility of

F-18 FDG PET/CT in screening for paraneoplastic neurological syndromes. Clin

Nucl Med. Jan 2012;37(1):39-43.

15. Schramm N, Rominger A, Schmidt C, et al. Detection of underlying malignancy in

patients with paraneoplastic neurological syndromes: comparison of 18F-FDG

PET/CT and contrast-enhanced CT. Eur J Nucl Med Mol Imaging. Jul

2013;40(7):1014-1024.

16. Vatankulu B, Yilmaz Aksoy S, Asa S, et al. Accuracy of FDG-PET/CT and

paraneoplastic antibodies in diagnosing cancer in paraneoplastic neurological

syndromes. Rev Esp Med Nucl Imagen Mol. Jan-Feb 2016;35(1):17-21.

17. Pena Pardo FJ, Garcia Vicente AM, Amo-Salas M, et al. Utility of 18F-FDG-

PET/CT in patients suspected of paraneoplastic neurological syndrome:

importance of risk classification. Clin Transl Oncol. 2016; [Epub ahead of print].

18

18. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for

systematic reviews and meta-analyses: the PRISMA statement. Int J Surg.

2010;8(5):336-341.

19. Garcia Vicente AM, Vega Caicedo CH, Mondejar Solis R, et al. [(18)F-FDG

PET/CT in the evaluation of patients suspected of paraneoplastic neurological

syndrome]. Rev Esp Med Nucl Imagen Mol. Jul-Aug 2015;34(4):236-243.

20. Whiting PF, Rutjes AW, Westwood ME, et al. QUADAS-2: a revised tool for the

quality assessment of diagnostic accuracy studies. Ann Intern Med. Oct 18

2011;155(8):529-536.

21. Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in

systematic reviews of diagnostic tests. Stat Med. Jun 15 2002;21(11):1525-1537.

22. Lee YH, Choi SJ, Ji JD, Song GG. Diagnostic accuracy of 18F-FDG PET or

PET/CT for large vessel vasculitis : a meta-analysis. Z Rheumatol. Dec 25 2015.

23. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in

meta-analyses. BMJ. Sep 6 2003;327(7414):557-560.

24. Borenstein M, Hedges LV, Higgins JP, Rothstein HR. A basic introduction to

fixed-effect and random-effects models for meta-analysis. Res Synth Methods.

Apr 2010;1(2):97-111.

25. Deeks JJ. Systematic reviews in health care: Systematic reviews of evaluations

of diagnostic and screening tests. BMJ. Jul 21 2001;323(7305):157-162.

26. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver

operating characteristic (ROC) curve. Radiology. Apr 1982;143(1):29-36.

27. Duval S, Tweedie R. Trim and fill: A simple funnel-plot-based method of testing

and adjusting for publication bias in meta-analysis. Biometrics. Jun

2000;56(2):455-463.

19

28. Masangkay N. AA, Torigian D., Dalmau J. Comparison of FDG-PET and CT in

detection of neoplasia in patients with paraneoplastic neurologic syndrome

(PNS). J Nucl Med. 2008;49(supplement 1):369P.

29. Bannas P, Weber C, Derlin T, et al. 18F-FDG-PET/CT in the diagnosis of

paraneoplastic neurological syndromes: a retrospective analysis. Eur Radiol. Apr

2009;20(4):923-930.

30. McKeon A, Apiwattanakul M, Lachance DH, et al. Positron emission tomography-

computed tomography in paraneoplastic neurologic disorders: systematic

analysis and review. Arch Neurol. Mar 2010;67(3):322-329.

31. Crandall J GP, Joshi K, Chaudhry M, Wahl R. Diagnostic performance of F18-

FDG PET/CT in patients with suspected paraneoplastic syndrome. J Nucl Med.

2014;55(supplement 1):1892.

32. Han S NC, Poon FW. The role of F-18 FDG PET/CT imaging in the investigation

of suspected paraneoplastic syndrome. Eur J Nucl Med Mol Imaging. 2011 38

(SUPPL. 2 (S256)).

33. Chakraborty PS SP, Soundararajan R, Karunanithi S, Singh H, Kumar R,

Malhotra A, Bal C. Yield of 18F-FDG PET/CT in patients with paraneoplastic

neurological syndrome for detection of primary malignancy. Eur J Nucl Med Mol

Imaging. 2013;40(SUPPL. 2 (S484)).

34. Fuster D VM, Depetris M, Mayoral M, Grau JJ, Muxi A, Lomeña F, Pons F. Value

of PET/CT with 18F-Fluorodeoxyglucose in patients with paraneoplastic

syndrome. Eur J Nucl Med Mol Imaging. 2013;40( SUPPL. 2 (S378)).

35. Iyngkaran G CB, Schultz C. The value of 18-FDG PET in the diagnosis of

tumours associated with paraneoplastic neurological syndromes. Intern Med J.

2009;39(SUPPL. 4 (A114)).

20

36. Gupta N BR, Sasikumar A, Sood A, Bhattacharya A, Mittal BR. Role of F-18 FDG

PET/CT in suspected cases of paraneoplastic neurological syndrome. Eur J Nucl

Med Mol Imaging. 2014;41(SUPPL. 2 (S313)).

37. Grozdic I NM, Odalovic S, Al-Nahhas A. The value of PET and PET/CT imaging

in paraneoplastic syndrome. Nucl Med Commun. 2011;32(5 (453-454)).

38. Titulaer MJ, Soffietti R, Dalmau J, et al. Screening for tumours in paraneoplastic

syndromes: report of an EFNS task force. Eur J Neurol. Jan 2010;18(1):19-e13.

39. Graus F, Delattre JY, Antoine JC, et al. Recommended diagnostic criteria for

paraneoplastic neurological syndromes. J Neurol Neurosurg Psychiatry. Aug

2004;75(8):1135-1140.

40. Subramaniam RM, Shields AF, Sachedina A, et al. Impact on Patient

Management of [18F]-Fluorodeoxyglucose-Positron Emission Tomography (PET)

Used for Cancer Diagnosis: Analysis of Data From the National Oncologic PET

Registry. Oncologist. Jul 8 2016.

41. Kwee TC, Kwee RM. Combined FDG-PET/CT for the detection of unknown

primary tumors: systematic review and meta-analysis. Eur Radiol. Mar

2009;19(3):731-744.

21

FIGURE 1. Flowchart of systematic literature review

22

FIGURE 2. The risk of bias and applicability concerns: review of authors’

judgments about each domain, presented as percentages across included

studies.

23

FIGURE 3. Forest-plots of sensitivity, specificity and diagnostic odds ratio across

the included studies (*conference abstracts)

24

25

FIGURE 4. Summary receiver operating characteristic curve and its’ two sided

95% confidence interval for assessment of the diagnostic accuracy of 18F-FDG-

PET or 18F-FDG-PET/CT in the detection of occult malignancy in patients

suspected for paraneoplastic syndrome.

26

FIGURE 5. Funnel plot of the included studies.

SUPPLEMENTAL TABLE 1. Modified QUADAS-2 questions used to assess the methodological quality of the included studies.

Risk of Bias (Low/ High/ Unclear risk)

Concerns regarding applicability (Low/ High/ Unclear risk)

Patient Selection

Could the selection of patients introduced bias? Was a consecutive or random sample of patients enrolled?

(Yes/ No) Was a case-control design avoided? (Yes/ No) Did the study avoid inappropriate exclusion? (Yes/ No)

Are there concerns that the spectrum of included patients do not match the topic of our study?

Index Test Could the conduct or interpretation of the index test have introduced bias?

Were the index test results interpreted blind to the results of the reference standard? (Yes/ No)

Are there concerns that the index test, its conduct, or interpretation differs from the topic of our study?

Reference Standards

Could the reference standard or its conduct have introduced bias? Are the reference standard likely to correctly classify the target

condition? (Yes/ No) Was the reference standards result interpreted without

knowledge of the results of the index test? (Yes/ No)

Are there concerns that the target condition as defined by the reference standard does not match the review question?

Flow and Timing

Could the patient flow have introduced bias? Was there an appropriate interval between index test and the

reference standard? (Yes/No) Did all patients receive a reference standard? (Yes/ No) Were all patients included in the analysis? (Yes/ No)

SUPPLEMENTAL Table 2. Summary of characteristics of the included studies

Author, Year Country Study design

N. patients (M: F)

Median age (range- Yrs)

Paraneoplastc presentation (N)

Patients with Positive paraneoplastic Ab (type of Ab) †

Previous history of treated cancer

Prior negative or inconclusive imaging ‡

Rees JH, 2001

UK R 43 (24:19)

62 (39- 86) Neurologic (43) 9/43 (20.9%) [8 anti-Hu, 1 anti-Yo]

3 [1 Bladder ca, 1 Melanoma, 1 Breast ca]

43/43 (100%)

Berner U, 2003

Germany R 30 (17:13)

55 (22- 76) Dermatologic (24) Neurologic (6)

NR NR

12/ 12 (100%)

Younes-Mhenni S, 2004

France P 20 (12:8) 62 (49- 78) Neurologic (20) 20/20 (100%) [12 anti-Hu, 1 anti Hu+anti CV2, 1 anti CV2, 4 anti-Yo, 1 anti Ri, 1 anti amphiphysin]

4 [2 Ovarian ca, 1 Breast ca, 1 Vocal cord ca]

20/20 (100%)

Linke R, 2004

Germany P 13 (4:9)

59 (22- 79) Neurologic (13) 13/13 (100%) [8 anti-Hu, 4 anti-Yo, 1 Anti-Tr]

6 [1 Neuroblastoma, 3 Ovarian ca, 1 SCLC, 1 Breast ca]

10/13 (77%)

Patel RR, 2008

US R 104 (46:58)

56 (17- 84) Neurologic (104) 73/104 (70.2%) [14 AchR, 19 ANNA, 1 Amphiphysin, 11 CRMP-5 IgG, 6 Striational, 9 N type Na channel, 9 VGCC, 12 K channel, 9 Glutamic acid decarboxylase]

NR 45 /71 (63.4%)

Hadjivassiliou M, 2008

UK P 80 (NA) NR Neurologic (80) 7/80 (8.7%) [5 Anti-Hu, 2 voltage-gated Ca Ab]

NR 80/80 (100%)

Masangkay N, 2008

US R 22 (8:14) NR (28- 82) Neurologic (22) NR NR 11/22 (50%)

Iyngkaran G, 2009

Australia R 24 (9: 15) 62 (36- 80) Neurologic (24) NR NR NR

Bannas P, 2010 *

Germany R 46 (22:24)

65 (21- 81) Neurologic (46) 4/40 (10%) [2 Anti-Hu, 1 anti-Ri, 1 CV2/CRMP5]

NR 38/46 (82.6%)

McKeon A, 2010 *

US R 56 (28:28)

61 (22- 80) Neurologic (56) 39/56 (69.6%) [6 ANNA-I, 2 PCA, 8 CRMP-5 IgG, 1 Amphihysin A, 9 AChR, 7 VGCC, 6 Striational, 2 VGKC, 1 Ma1, 3 unclassified]

10 [3 Breast ca, 2 Testicular ca, 1 Uterine ca, 1 Myeloma, 1 Lymphoma, 1 Lung ca, 1 Prostate ca]

56/56 (100%)

Selva-O’Callaghan A, 2010

Spain P 55 (18:37)

58 (46- 69) Dermatomyositis/ Polymyositis (55)

NR 0 46/55 (83.6%)

Grozdic I, 2011

Serbia R 124 NR NR NR NR NR

Han S, 2011 Scotland R 47 (18: 29)

58 (46- 70) All types NR NR NR

Matsushia A, 2012

Japan R 27 (16:11)

65 (41- 96) Neurologic (27) 2/7 (28.6%) [1 anti VGCC, 1 anti Hu]

2 [1 Lung ca, 1 HCC]

NR

Vaidyanathan S, 2012

UK R 68 (24:44)

58 (23- 82) Neurologic (55), Musculoskeletal (5),Endocrine (3), Constitutional (5)

34/68 (50%) [6 CA19-9, 6 CEA, 3 CA-125. 3 Anti-Hu, 3 Antiampiphysin, 2 Anti-MAG, 2 Anti-NMDA, 2 IgG, 1 Anti MA-2, 1 Anti Yo, 1 Anti CV2, 1 Anti AChR, 1 Anti PNMA, 1 Anti VGCC, 1 Anti VGKC, 1 beta HCG, 1 Anti-Ri]

8 [1 Breast ca, 1 Rectal ca, 1 Thyroid ca, 1 NHL, 1 Melanoma, 1 AML, 1 Lung ca, 1 Pancreatic ca]

NR

Chakraborty PS, 2013

India R 65 (44: 21)

65 (52- 78) Neurologic (65) 12/ 65 (18.5%) [Anti-Hu] NR NR

Fuster D, 2013

Spain R 24 (13: 11)

(38- 87) Neurologic (18), Hematologic (3), others (3)

NR NR NR

Schramm N, 2013

Germany R 66 (28:38)

60 (21- 87) Neurologic (60) 19/58 (32.7%) NR 54/66 (81.8%)

Crandall J, 2014

US R 226 (102:124)

60 (5- 86) All types NR NR NR

Gupta N, 2014 *

India R 80 (58: 22)

NR (10- 86) Neurologic (80) NR NR NR

Kristensen SB, 2015

Denmark R 137 (70:67)

61 (19- 63) Neurologic (67) Dermatologic (18) Rheumatologic(25) Nephrologic (6) Hematologic (2) others (19)

NR NR NR

Penapardo FJ, 2016

Spain R 73(47:26) 61 (NR) Neurologic (73) 14/55 (25.4%) 8 [1 Lymphoma, 2 Breast ca, 1 Prostate ca, 2 Colorectal ca, 1 Teratoma, 1 Renal ca]

NR

Vatankulu B, 2016

Turkey R 42 (26:16)

58 (14- 82) Neurologic (42) 21/34 (61.8%) [19 Anti-Hu, 18 Anti-yo, 9 Anti-Ri]

NR NR

Gupta K, 2016

India R 44 (25: 19)

NR NR NR NR NR

*Excluded from the quantitative analysis No reference standard for PET negative patients were reported

†: Proportion of patients with at least one positive paraneoplastic antibody among those with available results for paraneoplastic antibodies

‡: Proportion of patients with prior negative or inconclusive radiologic investigation among those with available prior imaging examinations

Ca, cancer; R, retrospective; P, prospective; Ab, antibody; NR, not reported; NHL, non-Hodgkin lymphoma; HCC, hepatocellular carcinoma; M:F,

male to female ratio; VGCC, Voltage gated Ca Channel; VGKC, Voltage gated K channel; CRMP-5 IgG, collapsin response-mediator protein 5

IgG; Yrs, years.

SUPPLEMENTAL Table 3. Summary of the index test, reference standard and proportion of patients diagnosed with underlying malignancy across the included studies.

Author, Year Index tests Index test interpretation

proportion of patients diagnosed with underlying malignancy

Reference standard (N) True positive PET / Total PET positive

Rees JH, 2001 18F-FDG-PET 2 NM physicians 10/43 (23.2%)6 Histo confirm (7) & FU of 18 mo (range 2 to 40 mo)

9/16

Berner U, 2003 18F-FDG-PET 2 NM physicians 7/30 (23.3%) Histo confirm (14) & FU of 43 mo (range 12 to 72 mo)

6/8

Younes-Mhenni S, 2004 18F-FDG-PET NR 18/20 (90%) Histo confirm (14) & FU of 12 mo (range 1 to 24 mo)

15/18

Linke R, 2004 18F-FDG-PET+ CT & CT

Various readers 10/13 (76.9%) Histo confirm (9) & FU of 48 to 60 mo

9/10

Patel RR, 2008 18F-FDG-PET & CT Various readers & 1 NM physician

10/104 (9.6%) Histo confirm (31) & FU of 18 mo (range 0.7 to 84 mo)

8/24

Hadjivassiliou M, 2008 18F-FDG-PET 1 radiologist 9/80 (11.2%) Histo confirm (8) & FU (NR) 9/18 Masangkay N, 2008 18F-FDG-PET & CT NR 9/22 (40.9%) Histo confirm (NR) 9/9 Iyngkaran G, 2009 18F-FDG-PET NR 3/ 24 (12.5%) Histo confirm (NR) & FU of 27 mo 3/8 Bannas P, 2010 * 18F-FDG-PET/CT 2 NM physicians &

2 radiologists 5 /46 (10.9%) Histo confirm (7) & FU (NR) 5/10

McKeon A, 2010 * 18F-FDG-PET/CT NR 10 / 56 (17.9%) Histo confirm (10) & FU (NR) 10/22 Selva-O’Callaghan A, 2010

18F-FDG-PET/CT & conventional screening#

1 NM physician & 1 radiologist

9/55 (16.4%) Histo confirm/ surgery (7) & FU of 14 mo (range 8 to 30 mo)

6/7

Grozdic I, 2011 18F-FDG-PET or PET/CT

NR 14/130 (10%) Histo confirm (9) & FU of 37 mo (range 11 to 63.5 mo)

13/13

Han S, 2011 18F-FDG-PET/CT NR 8/47 (17%) Histo confirm (16) & FU of at least 12 mo

7/13

Matsushia A, 2012 18F-FDG-PET/CT NR 6/27 (22.2%) Histo confirm/ surgery (6) & FU of 23.6 mo (range 1 to 86 mo)

5/6

Vaidyanathan S, 2012 18F-FDG-PET/CT Various readers 8/68 (11.8%) Histo confirm/ surgery (19) & FU of 8.9 mo (range 0.2 to 40 mo)

8/18

Chakraborty PS, 2013 18F-FDG-PET/CT 2 NM physicians 11/65 (16.9%) Histo confirm (NR) & FU of at least 11/14

6 mo (NR) Fuster D, 2013 18F-FDG-PET/CT NR 9/24 (37.5%) Histo confirm (9) & FU of at least 6

mo (15) 8/9

Schramm N, 2013 18F-FDG-PET/CT and CECT

1 radiologist & 1 radiology resident

9 /66 (13.6%) Histo confirm (10) & FU of 20 mo (range 3 to 57 mo)

9/10

Crandall J, 2014 18F-FDG-PET/CT Various readers 1246/226 (20.3%) Histo confirm (23) & FU of 60 mo (at least 12 mo)

26/34

Gupta N, 2014 * 18F-FDG-PET/CT 2 NM physicians 6/80 (7.5%) Histo confirm (12) 6/12 Kristensen SB, 2015 18F-FDG-PET/CT 1 NM physician 12/137 (8.8%) Histo confirm/ surgery (10) & FU of

26 mo (range 0.5 to 51.5 mo) 9/31

Penapardo FJ, 2016 18F-FDG-PET/CT 2 NM physicians 8/73 (10.9%) Histo confirm (17) & FU of 33 mo (range 12 to 96 mo)

6/7

Vatankulu B, 2016 18F-FDG-PET/CT 2 NM physicians 7/42 (16.7%) Histo confirm (7) & FU of 35 mo (range 8 to 68 mo)

6/6

Gupta K, 2016 18F-FDG-PET/CT NR 8/ 44 (18.2%) Histo confirm (25) & FU (range 13.5 to 17 mo)

8/12

*Excluded from the quantitative analysis; no reference standard for PET negative patients were reported

#Conventional screening, a combination of physical exam, laboratory tests, CT, mammography or ultrasonography

FN, False-Negative; TP, True-positive; Histo confirm, histopathology confirmation; NR, not reported; FU, follow-up; N, number; NM,

nuclear medicine physician.

SUPPLEMENTAL Table 4. Detailed information on true positive, false negative and

false positive results among the included studies.

Author, Year Detailed information on true positive, false negative and false positive sites (N) Rees JH, 2001 TP (9): SCLC (3), Lung (2), Adenoca unknown primary (1), Cecal ca (2), Colon ca (1)

FN (1): Breast ca (1) FP (7): Prostate (1), pleural fibrosis (1), hilum (1), Died w/o PM [Hilum (4)]

Berner U, 2003 TP (6): Bronch ca(3), Colon ca (1), Ovarian ca (1), Lymphoma (1) FN (1): Bronchial ca (1) FP (2): Adrenal adenoma (1), NR (1)

Younes-Mhenni S, 2004

TP (15): SCLC (9), Lung adenoca (2), Breast ca (2), Ovarian Ca (1), Thymoma (1) FN (1): Abdominal metastasis from ovarian ca (1), SCLC (1) FP (3): Parotid (1), mediastinum (2)

Linke R, 2004 TP (10): SCLC (5), HL (1), LN adenoca (1), Ovarian ca (2), Neuroblastoma (1) TN (0): None FP (1): Axillary node (1)

Patel RR, 2008 TP (8): thymoma (1), mixed adenoca/SCLC (1), mixed large cell/SCLC(1), SCLC (4), adenoca of unknown origin (1) FN (2): Serous ca of fallopian tube (1), metastatic spindle cell uterine ca (1) FP (16): NR

Hadjivassiliou M, 2008 TP (9): SCLC (4), NHL (1), Colon ca (1). Thymoma (1), Endometrial ca (1), high clinical suspicion (2) FN (3): SCLC in LN (1), high clinical suspicion (2) FP (9): Liver (1), Mediastinum (1), Hilum (1), Lung (1), Died w/o PM [Thoracic vertebra (1) and Mediastinum (4)]

Masangkay N, 2008 NR Iyngkaran G, 2009 NR Bannas P, 2010 TP (6): Neuroendocrine ca (1), SCLC (1), Adenoca (1), Gastric ca (1), Thymoma (1)

FN: Not mentioned FP (4): Vagina (1), Urinary bladder (1), Uterus (1), Maxilla (1)

McKeon A, 2010 TP (10): PC Thyroid (2), NSCLC (1), SCC of tonsil (1), Lung adenoca (1), Prostate ca (1), SCLC (2), Breast ca (1), Colon ca (1) FN: Not mentioned FP (12): Colon (1), Thyroid (2), LN (4), Lung (2), Esophagus (1), Cecum (1), Multiple foci (1)

Selva-O’Callaghan A, 2010

TP (6): Pancreatic ca (1), Lung ca (1), Colon ca (1), Breast ca (3) FN (3): Breast ca (2), Vaginal ca (1) FP (1): colon (1)

Grozdic I, 2011 TP (13): NR FN (1): cerebral tumor (1)

Han S, 2011 TP (7): PTC (1) SCLC (2), pulmonary carcinoid (1), recurrent colorectal ca (1), multiple metastases (1), active TB (1) FN (2): Lymphoma (1), brain tumor (1) FP (6): Vocal cord (1), thyroid (1), stomach (1), colon (1), ovary (1), bone marrow (1)

Matsushia A, 2012 TP (5): SCLC (1), Colon ca (1), Pancreas ca (1), Lung adenoca (1), PCT (1) FN (1): HCC (1) FP (1): lung granuloma (1)

Vaidyanathan S, 2012 TP (8): Pulmonary carcinoid (1), PCT (2), SCLC (2), Sigmoid ca (1), LN enlargement by CT treated as Ca (1), Died soon after (1) FN (0): None FP (11): NR

Chakraborty PS, 2013 TP & FP (14): Thyroid ca (4), Lung ca (2), lymphoma (2), Colon ca (2), breast ca (1), Prostate ca (1), brain metastases (1), bone metastases (1)

FN (0): None

Fuster D, 2013 TP & FN (9): Lung ca (4), CNS tumor (3), ovarian ca (1), duodenal ca (1) FP (1): NR FN (1): NR

Schramm N, 2013 TP (9): SCLC (1), Cervical ca (2), Bladder ca/bronchial ca (1), Breast ca (2), Ovarian ca (1), Lung adenoca/ tonsil ca (1), Tonsil ca (1) FN (0): None FP (5): Thymus (1), NR (4)

Crandall J, 2014 TP (26): SCLC (6), Adenoca of UO (2), Ca of UO (1), Pancreas ca (3), Ovarian ca (2), NSCLC (2), Lymphoma (2), Thyroid ca (1), Renal ca (1), Esophageal ca (1), Colorectal ca (1), Breast ca (1), high clinical suspicion (3) FN (20): NR FP (8): NR

Gupta N, 2014 TP (8): Lung ca (3), urinary bladder transitional cell ca (1), Colon ca (2), Sarcoidosis (1), Tuberculosis (1) FP (4): NR

Kristensen SB, 2015 TP (9): NSCLC (4), Breast ca (1), Lymphoma (2), Pheochromocytomoa (1), Unknown primary (1) FN (3): Gastric ca(1), Glottis ca (1), Parotid Small cell neuroendocrine ca(1) FP (22): Lymph node (5), pelvis (1), breast (3), intestine (3), rectum (4), lung (4), uterus (1), bone marrow (1)

Penapardo FJ, 2016 TP (7): Lung ca (3), Breast ca (1), Lymphoma (1), Rectal ca (1), NR (1) FN (4): Prostate ca (1), Bladder ca (1), NR (2) FP (18): NR

Vatankulu B, 2016 TP (6): NHL (1), Thymic ca (1), SCLC (3), Gallbladder ca (1) FN (1): Gastric ca (1) FP (0): None

Gupta K, 2016 TP (8): mostly SCLC, lymphoma

NR, not reported; HL, Hodgkins’ lymphoma; NHL, Non-Hodgkin lymphoma; ca, cancer; UO,

unknown origin; SCLC, small cell lung cancer; adenoca, adenocarcinoma; N, number; PCT,

papillary thyroid carcinoma; w/o PM, without proven malignancy; ca, Cancer; (N), number of

patients; LN, lymph node.