Diagnostic value of thyroid transcription factor-1 for pleural or other serous metastases of pulmonary adenocarcinoma: a meta-analysis

The role of thyroid transcription factor 1 (TTF-1) in the diagnosis of metastatic pulmonary adenocarcinomas in pleural, pericardial, and peritoneal effusions has not been defined. This study aimed to assess the overall diagnostic accuracy of TTF-1 for metastatic pulmonary adenocarcinomas in pleural or other effusions. Literature search was conducted in PubMed, EMBASE, and other databases to find eligible publications. Quality was assessed according to standardized QUADAS-2 criteria. Sensitivity, specificity, positive/negative likelihood ratio (PLR/NLR), and diagnostic odds ratio (DOR) were pooled. Summary receiver operating characteristic (SROC) curves were used to assess overall performance of the TTF-1 assay. A systematic search revealed 20 studies comprising a total of 1,213 subjects in this meta-analysis. The summary estimates were listed as follows: sensitivity, 0.74 (95% CI: 0.69–0.79); specificity, 0.99 (95% CI: 0.97–1.00); PLR, 78.16 (95% CI: 27.15–225.05); NLR, 0.26 (95% CI: 0.22–0.32); and diagnostic odds ratio, 297.75 (95% CI: 104.16–851.19). Estimated positive and negative post-probability values for metastatic pulmonary adenocarcinomas prevalence of 20% were 95% and 6%, respectively. The area under the SROC curve was 0.96. TTF-1 shows significant potential as a diagnostic marker to differentiate metastatic pulmonary from non-pulmonary adenocarcinomas in pleural or other effusions. These results justify larger, more rigorous studies to confirm such a diagnostic role.


Selection of eligible studies.
A study was included if it fulfilled the following criteria: (i) it examined the ability of TTF-1 to differentiate metastatic pulmonary adenocarcinomas from non-pulmonary adenocarcinomas in humans; (ii) it analyzed pleural, pericardial, and peritoneal effusions as samples; (iii) it reported sufficient data to allow calculation of true positive (TP), false positive (FP), false negative (FN), and true negative (TN) rates; (iv) it reported definitive determination of metastatic pulmonary adenocarcinomas and non-pulmonary adenocarcinomas using gold-standard methods; and (v) it was an original research study published in English or Chinese. Conference proceedings and studies published only as abstracts were excluded. To avoid selection bias, we also excluded studies involving fewer than 20 patients. When several articles concerned the same subjects, only results from the publication with the largest sample were used. Data extraction. Two reviewers (Y. Shen and C. Pang) independently identified eligible studies and extracted data on study methodology, characteristics and test accuracy using a standardized extraction form. The data extracted were: name of first author, publication year, country, serous effusion types, sample preparation method, TTF-1 immunostaining method, antibody clone and dilution, and two-by-two tables of TP, TN, FP and FN. Detailed information about controls with non-pulmonary adenocarcinoma was also reviewed.
Assessment of methodological quality. The same two reviewers (Y. Shen and C. Pang) assessed the quality of the selected studies using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) criteria, which cover four key domains for assessing risk of bias and applicability of the study results. These domains are patients selection, index test, reference standard, and flow and timing of samples/patients through the study 16 . Any discrepancies between the two authors (Y. Shen and C. Pang) during study selection, data extraction or quality assessment were resolved by discussion with a third author (K. Shen).
Statistical analysis. We used standard methods recommended for bivariate meta-analysis of diagnostic test evaluations 17 . We descriptively analyzed study characteristics and QUADAS-2 quality assessment using Excel and Review Manager 5.2 (The Cochrane Collaboration, Copenhagen, Denmark). The following measures of test accuracy were computed for each study, together with 95% confidence intervals (95% CIs): sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and diagnostic odds ratio (DOR). A summary ROC (SROC) curve covering all the studies was plotted using the data on sensitivity and specificity for a single test threshold from each study. The area under the SROC curve (AUC) was used to summarize the overall diagnostic performance of TTF-1.
The heterogeneity effect was measured using the Q test and the inconsistency index (I 2 ). P < 0.05 or I 2 ≥ 50% indicated significant heterogeneity, which was then analyzed through meta-regression to identify potential covariates. Deeks's funnel plot was used to detect publication bias 18 . Post-test probability was calculated using the overall prevalence of 20% with Fagan nomograms. All analyses were performed using the "Midas" module in STATA 12.0 (Stata Corp., College Station, TX), and Meta-DiSc 1.4 for Windows (XI, Cochrane Colloquium, Barcelona, Spain). All statistical tests were two-sided, with P < 0.05 taken as the threshold for statistical significance.

Patient characteristics and study design.
The final set of 20 studies involved 1,213 subjects, comprising 668 patients with metastatic pulmonary adenocarcinomas and 545 controls with non-pulmonary adenocarcinomas (median 60 patients per study; range 32-113 patients). (Table 1). Thirteen studies were performed in Asia, six in the USA, and one in Europe. The most frequent cancer types among the 545 patients with non-pulmonary adenocarcinomas were breast (n = 178), gastrointestinal (n = 147) and ovary adenocarcinomas (n = 145). Eight   studies assayed pleural effusion; four studies, pleural effusion, pericardial effusion, and ascites; another four studies, pleural effusion and ascites; two studies, only mentioned as serous effusions; one study, pleural effusion and pericardial effusion; and one study, only pericardial effusion. Only two studies involved analysis of smear samples 25,34 ; the remainder relied on analysis of cell blocks. Table 2 summarizes individual study designs and results for diagnostic performance of TTF-1.

Author (Ref) Year Country Subjects Samples
Most studies detected TTF-1 using the 8G7G3/1 antibody; three used the SPT24 antibody 33,34,37 , and one study did not report this information 35 . Twelve studies used immunohistochemistry to detect TTF-1, while the remaining eight used immunocytochemistry. Antibody dilutions from 1:40 to 1:500 were used in the included studies, while five did not report dilution factors. All studies defined nuclear staining as positive. Supplemental Table 1 summarizes the clinical information of patients with non-pulmonary adenocarcinomas.
Methodological quality of the included studies. QUADAS-2 was proposed in 2011 as an improved redesign of the original QUADAS and it was integrated into RevMan 5.2 in 2012. We applied the four criteria of QUADAS-2 (patient selection, index test, reference standard, flow and timing) to the studies in our meta-analysis. A response of "Yes" was given if the criterion was fulfilled, "Unclear" if fulfillment was unclear, and "No" if the criterion was not fulfilled. Based on these responses, the risk of bias for each criterion was classified as low, high, or unclear. Based on the first three domains, the applicability of the results was also evaluated. The quality of included studies was generally good, but three studies 19,21,27 were at high risk of bias due to deficiencies in patient selection. Figure 2 shows the summary of QUADAS-2 assessments of included studies.
Diagnostic accuracy of TTF-1. Sensitivity of TTF-1 for diagnosing metastatic pulmonary adenocarcinomas in effusions was between 0.54 and 0.88, and the pooled sensitivity was 0.74 (95% CI: 0.69-0.79). Specificities of TTF-1 varied from 0.92 to 1.00, and the pooled specificity was 0.99 (95% CI: 0.97-1.00). The other pooled parameters for TTF-1, calculated over all 20 studies, were: PLR, 78. 16 (Fig. 3). Figure 4 shows a plot of the TP rate as a function of the FP rate in individual studies, as well as the corresponding SROC curve. The AUC was 0.96, indicating a high discriminatory ability for TTF-1. Fagan's nomogram for likelihood ratios (Fig. 5) indicated that using TTF-1 to detect metastatic pulmonary adenocarcinomas increased the post-probability to 95% when the results were positive, and reduced the post-probability to 6% when the results were negative.    (prospective vs. retrospective); and (vi) blinding (blind vs. other). None of these covariates was found to be a significant source of heterogeneity (all P > 0.05, Table 3). Deeks's funnel plot asymmetry test was used to assess likelihood of publication bias in the included 20 studies. The slope coefficient was associated with P = 0.11, suggesting symmetry in the data and low likelihood of such bias (Fig. 6).

Discussion
Diagnosing lung adenocarcinoma based on resection histology is normally straightforward, but diagnosing metastatic pulmonary adenocarcinomas based on effusion samples can be extremely challenging 5,6 . Available biomarkers in effusions differentiate poorly between pulmonary adenocarcinomas and non-pulmonary adenocarcinomas. TTF-1 has emerged as a promising candidate biomarker, but studies of its diagnostic performance have given conflicting results [11][12][13] . Our meta-analysis is strengthened by the use of a standard protocol, strict inclusion criteria, standardized data extraction, independent reviewers, and a bivariate random-effects model 39 .
Our meta-analysis of available evidence suggests that TTF-1 can accurately predict whether an adenocarcinoma cell originated from a pulmonary or non-pulmonary site. However, TTF-1 probably cannot stand on its own and so should be used in conjunction with other markers.
Our meta-analysis indicated that TTF-1 performed with medium sensitivity (0.74, 95% CI: 0.69-0.79); and high specificity (0.99, 95% CI: 0.98-1.00), with a relatively high rate of missed diagnoses (16%) but a low rate of misdiagnosis (1%). These findings suggest that TTF-1 is a highly specific marker of pulmonary adenocarcinoma origin in pleural and other effusions. The SROC curve, which assesses overall test performance by showing the trade-off between sensitivity and specificity 40 , had an AUC of 0.96, suggesting high overall accuracy. Another indicator of diagnostic accuracy is DOR, which combines sensitivity and specificity data into a single number ranging from 0 to infinity, with higher values indicating better discriminatory test performance. Mean DOR in  our meta-analysis was 297.75, suggesting that assaying TTF-1 should be helpful in the diagnosis of metastatic pulmonary adenocarcinomas. We further examined the diagnostic accuracy of TTF-1 by calculating PLR and NLR, which can be easier to relate to clinical practice than SROC and DOR. The pooled PLR value of 78.16 suggests that patients with metastatic pulmonary adenocarcinomas have an approximately 78-fold higher chance of giving a positive TTF-1 result than do patients without metastatic pulmonary adenocarcinomas. At the same time, the pooled NLR was 0.26, indicating that a negative TTF-1 result is still 26% likely to be a false negative, which is not low enough to rule out metastatic pulmonary adenocarcinomas.
The relatively low sensitivity of TTF-1 in identifying metastatic pulmonary adenocarcinoma cells in effusion samples means that it is probably not sufficiently reliable on its own. Instead it should be used in conjunction with other markers. For example, combining of TTF-1 and napsin A gave higher sensitivity and accuracy than TTF-1 alone in identifying metastatic pulmonary adenocarcinomas 34 . Carcino-embryonic antigen is often targeted during immunostaining of metastatic pulmonary adenocarcinoma in pleural or other effusions 41 , so, including TTF-1 within a panel of immunostaining markers such as CEA, and napsin A, may increase the overall sensitivity and specificity, thereby improving overall accuracy.
Though TTF-1 may play a role in identifying malignant effusions, comparing the diagnostic performance of TTF-1 with that of classical tumor markers such as CA15-3 and vascular endothelial growth factor (VEGF) is difficult, because the two types of biomarker serve different purposes. Immunostaining for TTF-1 is done primarily to determine the source of malignant cells. When measured by ELISA on pleural fluid supernatants, TTF-1 had a poor diagnostic accuracy for differentiating malignant and benign effusions with the sensitivity of only 9% 10 . Examining markers such as CA 15-3 and VEGF, or using other diagnostic tools such as percutaneous pleural biopsy and VATS-directed biopsy, is done to determine whether effusions are malignant or benign [42][43][44][45] .
TTF-1 is also a sensitive marker for papillary carcinoma of the thyroid, although it is estimated that fewer than 1% of patients with papillary thyroid carcinoma have malignant pleural effusions 46 . In this meta-analysis, only one case of thyroid carcinoma was reported in 545 patients with non-pulmonary adenocarcinomas. The rarity of metastatic thyroid carcinoma in serous effusions explains the nearly 100% specificity of TTF-1 in detecting metastatic lung adenocarcinoma across several studies.
Our meta-analysis results indicated an association between TTF-1 and presence of metastatic pulmonary adenocarcinomas, implying that TTF-1 may contribute to such metastasis. Winslow et al. reported that downregulation of TTF-1 is associated with loss of differentiation, enhanced tumor seeding ability and increased metastatic potential in lung adenocarcinoma 47 . Positive and partially positive TTF-1 expression in lung adenocarcinoma patients correlates with EGFR mutations (exon 19 and 21). In clinical practice, the combination of TTF-1 expression and EGFR mutations, especially mutations in exon 21, can guide timely clinical treatment for lung adenocarcinomas 48 . Future studies should examine how TTF-1 functions in lung adenocarcinoma-related regulatory and signaling pathways. At the same time, researchers and clinicians should not overextend their interpretations of TTF-1 expression, which should be taken into account only when malignant cells are present. Indeed, the requirement for malignant cells limits the diagnostic sensitivity and clinical significance of TTF-1, and distinguishes it from assays based on circulating tumor DNA or classical tumor markers.
Standardized techniques for detecting TTF-1 should be established in order to maximize the clinical utility of this biomarker. Studies should rigorously determine whether immunohistochemistry or immunocytochemistry   is superior, and the dilution of primary antibody should be optimized. Dilution factors among the studies in this meta-analysis ranged from 1:40 to 1:500. Studies should also compare the different primary antibodies available; one study has suggested that the SPT24 antibody clone is better than the 8G7G3/1 clone 49 . It may also be possible to improve sensitivity or specificity of immunohistochemical staining by optimizing antibody cut-off values 50 . The findings of this meta-analysis should be interpreted with caution because of several limitations. While our strict inclusion and exclusion criteria may have helped reduce selection bias, they led to a relatively small final set of studies for which statistical power may be inadequate for drawing definitive conclusions about the ability of TTF-1 to discriminate metastatic pulmonary adenocarcinomas from metastatic non-pulmonary adenocarcinomas in pleural or other effusions. For example, we included only studies published in English and Chinese in a relatively small number of databases. Our results may be biased by our omission of unpublished studies, studies published in other languages and studies published in journals not indexed in the databases we searched. In addition, we detected substantial heterogeneity across the included studies, for which we were unable to identify causes using meta-regression. Future studies should aim for greater rigor in order to decrease the risk of bias.

Conclusions
In summary, our meta-analysis suggests that TTF-1 may significantly aid the diagnosis of metastatic pulmonary adenocarcinomas in pleural or other effusions. Our data provide further evidence that TTF-1 is a useful marker for distinguishing metastatic adenocarcinoma of the lung from non-pulmonary adenocarcinoma in specimens of pleural or other effusions.