18F-FDG PET or PET/CT for detecting extensive disease in small-cell lung cancer: a systematic review and meta-analysis

18

F-FDG PET or PET/CT for detecting extensive

disease in

small-cell lung cancer: a systematic review and

meta-analysis

Yu-Yu Lu

*, Jin-Hua Chen

, Ji-An Liang

, Shannon Chu

,

Wan-Yu Lin

*

and Chia-Hung Kao

Introduction

Lung cancer is the first leading cause of cancer-related death in the male population and the second leading cause of cancer-related death in the female population worldwide. Approximately 1 608 800 new cases are

diagnosed annually and 1 378 400 deaths occur each year [1]. Small-cell lung cancer (SCLC) accounts for approximately 15% of all lung cancers and is distinguished from

non-small-cell lung cancer (NSCLC) by its short volume doubling time and propensity for widespread metastatic disease.

In clinical practice, TNM classification is usually collapsed into a simple binary classification, limited disease (LD) and extensive disease (ED), which was first proposed by the Veterans Administration Lung Study Group (VALSG) [2] and later modified by the International Association for the Study of Lung Cancer

(IASLC) [3]. The definition of LD based on VALSG

includes patients with primary tumor and nodal involvement limited to one hemithorax, which can be encompassed in a tolerable radiation field. IASLC defines LD as all cases without distant metastasis. At the time of presentation, 30–40% of patients with SCLC present with LD and 60–70% present with ED [4]. The

therapeutic strategy for SCLC is a combination of chemotherapy and thoracic irradiation for LD and

chemotherapy alone for ED. Accurate staging in patients with SCLC has a major impact on clinical management and prognosis.

Current staging modalities used for SCLC include computed tomography (CT) of the chest and abdomen, abdominal sonography, MRI of the brain, and bone scintigraphy. PET using the radiolabeled glucose analog fluorine-18 2-fluoro-2-deoxy-D-glucose (18F-FDG) is a

whole-body metabolic imaging technique that is capable of detecting a wide range of tumors that exhibit higher accumulation of 18F-FDG compared with surrounding

normal tissues. It has been proven to be valuable in the staging of and treatment planning for many malignancies [5–8].

Several studies have indicated that 18F-FDG PET is an

accurate method for the staging of patients with NSCLC [9–13]. It has been suggested that PET should routinely be added to the conventional workup of NSCLC patients and be used for the detection of distant metastases [14]. Unlike in the case of NSCLC, little is known about the diagnostic accuracy of 18F-FDG PET or

PET/CT in patients with SCLC because of small sample size. The purpose of the present study was to evaluate the diagnostic performance of 18F-FDG PET in pretherapeutic

staging of patients with SCLC by conducting a

meta-analysis of the published literature. Materials and methods

Literature search

A comprehensive computer search for relevant articles was conducted using PubMed/MEDLINE and EBM Review

search engines. The search included combinations of the following terms: (a) PET; (b) 18F-FDG, fluorodeoxyglucose;

(c) SCLC; and (d) staging. Searches were limited to studies on human subjects. Although no language restrictions were used initially, the full-text review and final analysis were limited to articles published in English. Additional studies were manually searched using the references cited in the

retrieved articles. Data selection

Studies were eligible for inclusion in the analyses based on the following criteria: (a) they evaluated SCLC for staging; (b) diagnosis was performed using 18F-FDG PET

or 18F-FDG PET/CT; and (c) 2_2 tables could be

derived from the provided data. Abstracts presented at congresses, unpublished data, case reports, meta-analyses, reviews, editorials, and comments were excluded.

To avoid missing potentially useful papers for the present meta-analysis, the abstracts were double-checked by at least two authors to determine whether the reports fitted the inclusion criteria for this study.

Quality assessment and data extraction

Two reviewers independently assessed the methodological quality of the eligible studies. The modified list used by Chen et al. [15] was further adapted for the needs of this specific review. Some items on the list were modified for this specific review. The complete list of criteria used is

presented in Table 1. Internal validity (IV) criteria were scored as ‘positive’ (adequate methods) or ‘negative’

(inadequate methods, potential bias, or insufficient information provided on a specific item). External validity (EV)

criteria were assessed to evaluate generalizability. The inclusion and exclusion criteria for EV scores were scored positive if they were mentioned in the publication. Standard performance of 18F-FDG PETor PET/CTwas scored positive

when the type of PET or PET/CT camera, the dose of

18F-FDG, the time between injection and scanning, and the

method of reconstruction were described. The criteria for EV were scored positive if sufficient information was provided to judge the generalizability of findings. Disagreements were resolved by consensus. Quality scores were expressed as a percentage of the maximum score. Subtotals were calculated for IV (maximum 6) and EV (maximum 6) separately.

For each report, we recorded the number of true-positive (ED), false-positive, true-negative (LD), and false-negative findings for 18F-FDG PET or PET/CT in diagnosing ED

of SCLC.

Statistical analysis

The sensitivity and specificity of the 18F-FDG PET or

PET/CT in diagnosing ED of staging SCLC were extracted or calculated using 2_2 contingency tables from the original numbers provided in the publications. We calculated the sensitivity, specificity, and their 95% confidence interval (CI) for pooled estimators. The pooled sensitivity and specificity estimators were weighted averages in which the weight of each study is the individual sample size. The sources of heterogeneity included the pattern of observed study results and variation introduced by the diagnostic threshold. If there was any evidence for which the diagnostic threshold varied between the studies, we considered the summary receiver operating characteristic (SROC) curve. The SROC curve shows the trade-off between sensitivity and specificity across the included studies [16]. Testing of the diagnostic threshold was performed with Spearman’s correlation test. In this study, the threshold effect

did not exist, but we show the symmetrical receiver operating characteristic curve in figures, which include values of Q* index and area under the curve.

Likelihood ratios (LRs) are also metrics that combine sensitivity and specificity in the calculations. The ratio of sensitivity over 1 – specificity is defined as LR+. The ratio of 1 – sensitivity over specificity is defined as LR– . The discrimination ability is considered to be better with a higher LR+ and a lower LR– . In previous papers, a test was

considered clinically useful when LR+ was greater than 5.0 and LR– was less than 0.2 [17]. Analyses were conducted using the free software Meta-DiSc (version 1.4, http:// www.hrc.es/investigacion/metadisc_en.htm) [18].

Results

Literature research

A total of 973 studies that investigated lung cancer

using 18F-FDG PET or PET/CT were found initially. After reviewing the titles and

excluded on the basis of the criteria listed in the Data selection subsection of the Materials and methods section. We screened the full-text of 18 articles. Four studies were excluded because of insufficient information to construct a 2_2 table [19–22]. One study was

excluded because the results of the diagnostic performance of 18F-FDG PET or PET/CT could not be

estimated [23]. One study was excluded because the results could not be classified to LD or ED in staging of SCLC [24]. Finally, a total of 369 patients from 12

eligible studies [25–36] were analyzed in the systematic review (Fig. 1).

Study characteristics

The characteristics of the eligible studies are summarized in Table 2. Five of the studies were prospective [27,30–33], and the others were retrospective. In all studies, the results of the diagnostic performance were patient-based. Nine of the studies were performed using an 18F-FDG PET

scan [25–27,29–32,34,36], two studies were performed using an 18F-FDG PET or PET/CT scan [28,35], and one was

performed using an 18F-FDGPET/CTscan [33]. Five studies

used the qualitative method [29,31–33,35] as measurement and three studies used both qualitative and quantitative methods [25,30,36].

Quality assessment

Methodological quality was assessed by 12 items for each of the 12 selected studies. The scores for IV and EV of the 12 selected studies are presented in Table 3.

Eight studies had a valid reference test (IV1). The readers were blinded to the results of the reference standard in all of the selected studies (IV2). All studies had verification bias (IV4) because patients were selected for assessment using the reference test, but this not performed independently of 18F-FDG PET results. Five

studies were prospective (IV6), and in the other seven studies patients were enrolled in the studies consecutively (EV5).

In six studies, the inclusion criteria (EV3) were described and in three studies the exclusion criteria (EV4) were described. The type of camera, the 18F-FDG dosage, the

uptake period, the time interval, and reconstruction were reported in all of the studies (EV6). The total score for the combined IV and EV, expressed as a fraction of the maximum score, ranged from 33 to 92%, with a mean of 60.3%.

Performance

The cross-tabulations from the enrolled studies in diagnosing ED of SCLC are presented in Table 4. The pooled estimated results of 18F-FDG PET or PET/CT in

the detection of ED in SCLC were patient-based. The pooled estimates of sensitivity, specificity, LR+, and LR – of 18F-FDG PET or PET/CT in diagnosing ED of

SCLC were 97.5% (95% CI, 94.2–99.2%), 98.2% (95% CI, 94.9–99.6%), 19.86 (95% CI, 9.79–40.30), and 0.06 (95% CI, 0.03–0.10), respectively. The I2 values of sensitivity

and specificity between studies were 3.0 and 0.0%, respectively. The SROC curve for diagnosing ED in

patients with SCLC before treatment has been presented in Fig. 2. The Q* index presents maximum joint

sensitivity and specificity, calculated as a global measure of diagnostic accuracy. The Q* index was 0.9456 for 18FFDG

PET or PET/CT in detecting ED in patients with SCLC.

The pooled estimates of sensitivity, specificity, LR+, and LR – of 18F-FDG PET alone in diagnosing ED of

SCLC were 98.1% (95% CI, 94.7–99.6%), 97.5% (95% CI, 93.0–99.5%), 18.41 (95% CI, 8.25–41.11), and 0.05 (95% CI, 0.02–0.10), respectively. The I2 values of sensitivity

and specificity across studies were 10.3 and 0.0%, respectively. The Q* index presents maximum joint sensitivity and specificity, calculated as a global measure of diagnostic accuracy. The Q* index was 0.9398 for

18F-FDG PET in detecting ED in patients with SCLC.

Discussion

staging of SCLC have been published earlier [37–39]; however, diagnostic accuracy of 18F-FDG PET (or PET/

CT) is not exactly known because of the lack of pooled study findings. This meta-analysis is the first to address global 18F-FDG PET diagnostic performance in staging

SCLC. The diagnostic performance of the 12 studies

discussed in the current review was patient-based. The pooled estimates of sensitivity, specificity, LR+, and

LR – of 18F-FDG PET (PET/CT) in the detection of ED

in patients with SCLC were 97.5%, 98.2%, 19.86, and 0.06, respectively. The global measure of diagnostic accuracy was 0.9456. The 18F-FDG PET or PET/CT

presented a high sensitivity and a high specificity, which

showed that there are very few false-negative and falsepositive cases. The high positive LR showed large

probability changes from pretest to post-test, and a low negative LR showed a significant contribution of the test in lowering the probability of the subject having the disease. The results of this meta-analysis indicate that

18F-FDG PETor PET/CT has a good diagnostic accuracy

for the evaluation of ED in SCLC patients.

There were no significant differences in diagnostic performance (accuracy: 0.9456 and 0.9398, respectively) between

mixture data (PET or PET/CT) and PET alone in the

detection of ED in patients with SCLC. Only one study [33] was performed with 18F-FDG PET/CT, and 20 patients were

enrolled in this study. Two studies [28,35] were performed with 18F-FDG PET or PET/CT. Data of patients who

underwent PET/CTcould not be extracted from the original data. Therefore, the exact number of patients included in the meta-analysis who had undergone 18F-FDG PET/CTwas

not available. The diagnostic accuracy of 18F-FDG PET/CT

for the detection of ED in SCLC could not be calculated. Further large studies are needed to evaluate disease status in

SCLC using 18F-FDG PET/CT. SCLC tends to disseminate early in the disease course

because of its high cellular turnover and very active metabolism. Systemic chemotherapy is the mainstay of treatment for patients at all stages of SCLC. Concurrent

chemotherapy and thoracic irradiation can promote longterm survival in patients diagnosed with LD [40,41].

However, thoracic irradiation might cause adverse events such as radiation pneumonitis and esophagitis, which worsens the patient’s quality of life and survival. Accurate staging is important both in the choice of treatment strategy and in the planning of locoregional radiation therapy target volumes. The potential of 18F-FDG PET

(PET/CT) to detect unsuspected distant metastases is an important advantage over conventional imaging.

18F-FDG PET can improve the accuracy of estimation of

the disease extent to alter the treatment plan or to better define the radiation field in patients with LD, which

could ultimately have an impact on survival. On the basis of the results of this meta-analysis we can conclude that

18F-FDG PET (PET/CT) is a powerful tool for assessing

disease extension and could potentially change patient management.

The role of 18F-FDG PET in the detection and

delineation of brain metastases has been questioned. Kamel et al. [28] revealed false-negative results in the detection of brain metastases in two of 24 patients. Brink et al. [31] showed that cranial MRI or CT had a higher sensitivity (100 vs. 46%) and specificity (100 vs. 97%) compared with 18F-FDG PET in the assessment of brain

metastasis. Vinjamuri et al. [34] showed that 18F-FDG

PET missed brain metastases found on cranial MRI or CT in five of 51 patients. These results indicate that

18F-FDG PET is inferior to MRI or CT for the detection

of brain metastases. Cranial MRI or CT is better suited for the assessment of brain metastases in patients with SCLC.

There were some potential limitations in this study. First, the presence of clinical heterogeneity in the study design may affect the generalizability of the results. It is important to note that different staging criteria exist, although the staging criteria mentioned in the majority of selected studies were based on VALSG. Second, histopathological confirmation of

some lesions was not available because systemic chemotherapy is effective in SCLC, and initiating chemotherapy

quickly to prevent disease progression is needed. Clinical follow-up was needed in some cases to confirm their true nature, including a variety of imaging modalities and clinical examinations, not all of which were performed in the same manner in the studies. Finally, all selected studies have verification bias. This is because the reference test was assessed in patients selected by the index test results, which can lead to overestimation of sensitivity. Despite these drawbacks, this meta-analysis demonstrates the diagnostic performance of 18F-FDG PET (PET/CT) in the pretherapeutic

staging of patients with SCLC.

Conclusion

The results of this systematic review and meta-analysis suggest that whole-body 18F-FDG PET or PET/CT is a

valuable imaging tool for the pretherapeutic assessment of ED in patients with SCLC, which is helpful for