DNA Methylation Analysis of the SHOX2 and RASSF1A Panel Using Cell-Free DNA in the Diagnosis of Malignant Pleural Effusion

Objectives The differential diagnosis of pleural effusion (PE) is a common but major challenge in clinical practice. This study aimed to establish a strategy based on a PE-cell-free DNA (cfDNA) methylation detection system for the differential diagnosis of malignant pleural effusion (MPE) and benign pleural effusion (BPE). Methods A total of 104 patients with PE were enrolled in this study, among which 50 patients had MPE, 9 malignant tumor patients had PE of indefinite causes, and the other 45 patients were classified as benign controls. The methylation status of short stature homeobox 2 (SHOX2) and RAS association domain family 1, isoform A (RASSF1A) was detected using PE-cfDNA specimens by real-time fluorescence quantitative PCR. Total methylation (TM) was defined as the combination of the methylation levels of SHOX2 and RASSF1A. The electrochemiluminescence immunoassay was applied to evaluate the levels of multiple serum tumor markers. Results The PE-cfDNA methylation status of either SHOX2 or RASSF1A was much higher in MPE samples than in benign controls. The combination of SHOX2 and RASSF1A methylation in PE yielded a diagnostic sensitivity of 96% and a specificity of 100%, respectively. When compared with the corresponding serum tumor marker detection results, TM showed the highest diagnostic efficiency (AUC = 0.985). Furthermore, the combination of the SHOX2 and RASSF1A methylation panels using PE-cfDNA could apparently improve the differential diagnostic efficacy of BPE and MPE and could help compensate for the deficiency of cytology. Conclusions Our results indicated that SHOX2 and RASSF1A methylation panel detection could accurately classify BPE and MPE diseases and showed better diagnostic performance than traditional serum parameters. The SHOX2 and RASSF1A methylation detection of PE-cfDNA could be a potentially effective complementary tool for cytology in the process of differential diagnosis. In summary, PE-cfDNA could be used as a promising non-invasive analyte for the auxiliary diagnosis of MPE.


Introduction
Pleural efusion (PE) is a common clinical symptom caused by over 50 diferent diseases, including infection, malignant tumors, heart failure, and hypoproteinemia. According to etiology, PE is typically classifed as malignant pleural effusion (MPE) or benign pleural efusion (BPE) [1]. Lung cancer, metastatic breast cancer, malignant pleural mesothelioma, lymphoma, and gastrointestinal tumors are major causes of MPE, and the formation of MPE often predicts a worsening condition in patients [2]. With regard to BPE, tuberculous pleurisy is the most common cause, especially in developing countries [3]. Patients with MPE tend to have a relatively poor prognosis, and the appearance of MPE directly afects the evaluation of tumor stage and the selection of a therapeutic approach. However, patients with BPE can usually be clinically cured by means of antiinfection, anti-tuberculosis, and closed thoracic drainage treatments and have a relatively good prognosis [4]. In clinical practice, it remains a challenge to distinguish MPE and BPE, and cytology or pleural biopsy is the gold standard for diagnosing MPE [5]. Cytological examination of PE is fast, convenient, and noninvasive, with a specifcity close to 100%, but the sensitivity is low due to the paucity of enough cells [6]. Pleural biopsy can provide relatively highsensitivity, but such an invasive test is not appropriate for regular examination due to the high risk of complications [7,8].
Classic serum tumor markers, such as carcinoembryonic antigen (CEA), neuron-specifc enolase (NSE), progastrinreleasing peptide (Pro-GRP), cytokeratin fragment (CYFRA) 21-1, squamous cell carcinoma-related antigen (SCC-Ag), cancer antigen (CA) 19-9, CA-125, CA15-3, and alpha-fetoprotein (AFP), have been widely used in the identifcation of MPE [9][10][11]. However, the overall sensitivity is insufcient to meet the demand of the current clinical diagnostic efciency. In view of the limitations of conventional detection methods, there is an urgent need to develop a more accurate and efective method for the rapid diferential diagnosis of MPE.
DNA methylation is a major epigenetic mechanism that maintains gene expression and cell characteristics [12]. In particular, aberrant promoter CpG island hypermethylation may lead to the gene silencing of tumor suppressor genes, thus contributing to the occurrence of multiple cancers [13]. RAS association domain family 1, isoform A (RASSF1A), is a classic tumor suppressor molecule that is widely involved in signal transduction, the cell cycle, tumor metastasis and other biological pathways [14]. Short stature homeobox gene two (SHOX2) plays an important role in skeletal development, embryonic development, and cardiovascular diferentiation. Moreover, SHOX2 was found to regulate cell proliferation and diferentiation and induce epithelialmesenchymal transition, making it an important growth regulator in the body [15][16][17]. Accumulating evidence suggests that aberrant promoter CpG island hypermethylation of SHOX2 and RASSF1A can serve as a diagnostic and prognostic biomarker for lung cancer, colorectal cancers, biliary tract cancers, and head and neck squamous cell carcinomas [18][19][20][21].
Cell-free DNA (cfDNA) refers to the DNA fragments released by original cells into body fuids and was frst reported in human plasma in 1948 [22]. It was further discovered in multiple types of body fuids, including PE, ascites, amniotic fuid, cerebrospinal fuid, and urine, both under physiological and pathological conditions [23][24][25][26]. Currently, cfDNA has been widely used in prenatal diagnosis, tumor diagnosis, and the selection of targeted therapy management and has shown great signifcance in the feld of liquid biopsy [27][28][29]. Te preliminary study of our research group found that cfDNA derived from PE supernatant may serve as high-quality material for molecular detection in guiding targeted therapies [30]. Several studies have concentrated on the detection of aberrant methylation of the SHOX2 and RASSF1A genes in plasma cfDNA as well as bronchoalveolar lavage fuid specimens to help improve the diagnosis of malignant tumors, while little attention has been given to cfDNA methylation in PE [31]. Compared to the traditional specimens, cfDNA from PE represents an easily accessible and direct source for the diferential diagnosis of BPE and MPE and is not limited by the number of exfoliated cells.
In this study, we evaluated the SHOX2 and RASSF1A promoter methylation status in defnitively diagnosed PE-cfDNA from 95 patients with benign and malignant diseases using real-time fuorescence quantitative PCR technology. Furthermore, we conducted combined analysis with traditional cytology and serum tumor markers, aiming to develop a new diagnostic technique for distinguishing BPE and MPE more quickly and efciently.

Clinical Specimens.
Te study was approved by the Human Research Ethics Committee of the Beijing Chest Hospital, Capital Medical University. 104 PE specimens were collected from consenting individuals from August 2020 to July 2021. Patient information, including sex, age, smoking history, and clinical diagnosis, was recorded. Individuals with a prior history of any cancer were excluded from this study.

Cytological and Pathological
Analyses. PE samples from each patient were obtained by pleural puncture and chest drainage. After centrifugation at 2000 g for 10 min, the supernatant was carefully isolated into a new tube without disturbing the debris for cfDNA extraction, and the corresponding sediments were processed for liquid-based cytology or prepared into parafn-embedded cell blocks and stained with hematoxylin and eosin. Immunocytochemical staining examination was applied for cell block sections as a supplement. All slides were individually reviewed by two experienced pathologists. Te detection of pathological changes in malignant tumors in pleural biopsy tissue or cytology has been considered the diagnostic gold standard for MPE. Journal of Oncology methylation for each targeted gene was calculated using the following formula:

Statistical Analysis.
Statistical analyses and graphics generation were performed using IBM SPSS Statistics 21.0 software (SPSS Inc., Chicago, IL) and GraphPad Prism 8.0. Descriptive statistics (GraphPad, San Diego, CA, USA) are reported as frequencies for categorical variables and medians ± interquartile ranges for continuous variables. Te detection sensitivity, specifcity, negative predictive value (NPV), positive predictive value (PPV), and relative 95% confdence interval (CI) for diagnosing BPE and MPE by methylation tests were calculated. Te positive frequency of methylation in the SHOX2 and RASSF1A genes, serum tumor marker detection, and cytology examination were analyzed using the chi-square test. Te receiver operating characteristic (ROC) curve was used to determine the cutof values of ΔCt SHOX2 and ΔCt RASSF1A and to calculate the area under the ROC curve (AUC), aiming to evaluate the diagnostic efcacy. A two-sided p value <0.05 was considered to be statistically signifcant.

Characteristics of the Patients.
A total of 104 patients with symptomatic PE were fnally enrolled in this study. Tere were 67 males and 37 females, with a mean age of 57.9 years (range: 19∼82). Fifty-nine cases were diagnosed as defnite malignant tumors by histopathology or cytology analysis, including 52 lung cancers, 2 thymic squamous cell carcinomas, 1 ovarian cancer, 1 lymphoma, 1 mesothelioma, 1 adenoid cystadenocarcinoma, and 1 esophageal cancer. Among the cohort, 50 patients were determined to have MPE, while 9 patients had pleural efusions of indefnite causes. Te other 45 patients were classifed as benign controls, comprising 37 patients with tuberculosis, 5 patients with pulmonary infections, and 3 patients with heart failure diseases. All the demographic and clinical characteristics are summarized in Table 1.

Diagnostic Performance of the cfDNA Methylation Test for BPE and MPE.
We evaluated the diagnostic value of SHOX2 and RASSF1A methylation by real-time PCR using PE-cfDNA specimens. Ct SHOX2 <32 and Ct RASSF1A <35 criteria were set for calculating the delta cycle threshold (ΔCt) [32]. We assigned "NoCt" � 40. Te quantitative SHOX2 and RASSF1A methylation status (ΔCt SHOX2 and ΔCt RASSF1A ) of the tested specimens is plotted in detail in Figure 1. Te methylation level of SHOX2 was much higher in MPE samples than in benign controls (A, 5.7 ± 4.6 vs. 17.4 ± 3.8, p < 0.0001), and the same tendency was observed in the RASSF1A (B, 10.0 ± 7.8 vs. 19.6 ± 0.9, p < 0.0001). According to the ΔCt method, the cutof value of SHOX2 and RASSF1A Journal of Oncology 3 methylation was defned as the maximum of Youden index. ROC analysis was performed, where the AUC was calculated and the cutof value was determined accordingly. As shown in Figure 2, the AUCs of methylation SHOX2, methylation RASSF1A and the combination of SHOX2 and RASSF1A (TM, total methylation) were 0.810, 0.959, and 0.985, respectively, which showed good discriminative power in diferentiating BPE from MPE. Te optimal cutof points of SHOX2 and RASSF1A methylation in this study were set at ΔCt SHOX2 � 10.0 and ΔCt RASSF1A � 13.0. Based on this cutof, the detection sensitivities of SHOX2 and RASSF1A were 90.0% and 62.0%, respectively, while combination analysis of these two genes improved the sensitivity up to 96.0%, which was higher than that of any single parameter, while the specifcity remained unchanged at a high level of 100% (Table 2). Tese results showed that the methylation levels of SHOX2 and RASSF1A using PE-cfDNA had potential diagnostic value in the diferentiation of BPE and MPE, while the combination panel of SHOX2 and RASSF1A had the best diagnostic efciency.

Clinicopathological Correlation Analysis.
We further tested whether PE-cfDNA methylation levels were associated with the corresponding clinicopathological parameters. As suggested in    with that of multiple traditional serum tumor markers. As shown in Figure 3 and Table 4, the AUC of TM was signifcantly higher than those of each of the traditional serum tumor markers. In contrast, the detection sensitivity and specifcity of traditional serum tumor markers were relatively low, among which CA-125 exhibited the highest sensitivity of 78.6% while the specifcity was only 47.2%. TM detection achieved a sensitivity of 96% (95% CI: 78.8%-95.9%) and a specifcity of 100% (95% CI: 90.2%-100%), obviously exceeding those of the above methods.
Moreover, in our cohort, 68.0% (34/50) of MPE cases had defnite positive cell block cytology, 8% (4/50) were atypical, and 24% (12/50) were negative. In contrast, liquidbased cytology was defnitively positive in 54.1% (20/37) of MPE cases, atypia in 16.2% (6/37) of the cases, and negative in 29.7% (11/37) of the cases. A positive result was recorded when cancer cells were found by either of the above two methods. When they were combined, the positive rate of cytology increased to 74% (37/50), the atypia rate was 10% (5/50), and the negative rate was 16% (8/50). Te positive rate of the PE-cfDNA methylation assay for SHOX2 and RASSF1A was 94.6% (35/37) in the MPE group with defnite positive cytology, and 100% (5/5) and 100% (8/8) in the MPE group with atypia and negative cytology (further confrmed by cytology or pleural biopsy during the same period), respectively. Te combined methylation detection of SHOX2 and RASSF1A in PE-cfDNA yielded a diagnostic sensitivity of up to 96%, suggesting that it could be a potential efective complementary tool for cytology in the diferential diagnosis of MPE and BPE. In summary, the use of PE-cfDNA methylation of SHOX2 and RASSF1A could facilitate the early diagnosis and diferential diagnosis of MPE.

Discussion
MPE, as a common symptom of metastatic lung cancer, breast cancer, and other cancer-related pleural lesions, upgrades the tumor stage and precludes resection of the primary tumor, indicating a poor prognosis for the patient. Nevertheless, patients with BPE can usually be clinically cured if treated in a timely manner. Hence, it is of great signifcance to accurately diferentiate between BPE and MPE for therapeutic decisions and thus improve the prognosis of patients with MPE. However, to date, no   Tuberculosis  37  37  0  -37  0  -37  0  -Pulmonary infection  5  5  0  -5  0  -5  0  -Heart failure  3  3  0  -3  0  -3 0 -Journal of Oncology 5 specifc method for diferential diagnosis is readily available in clinical practice due to either its low sensitivity or specifcity. Toracoscopy exhibited an excellent diagnostic yield (>95%) for malignant pleural lesions and was considered the gold standard. However, high requirements for surgical staf and equipment, as well as high complications, limit its application [33]. Cytology analysis can identify malignancy in approximately 60% of patients but easily causes misdiagnosis [34]. In terms of morphologic imaging, CT and FDG-PET showed a potential role in identifying the nature of PE, but the false-positive rate was relatively high. When combination analysis was used, the specifcity, PPV, and accuracy reached 76%, 67%, and 84%, respectively [35]. Te test results of emerging molecular detection technologies are more sensitive and objective, which can compensate for the shortcomings of morphological diagnosis. Aberrant DNA methylation is correlated with the malignant characteristics of various tumors [36]. As one of the most promising new biomarkers in the diagnosis of cancer, aberrant methylation of a set of genes, including SHOX2, RASSF1A, and SEPT9, has been transferred into clinical application [37][38][39]. Te association of cancer diagnosis with the abnormal methylation status of the SHOX2 and RASSF1A genes has been studied in multiple sample types, including bronchoalveolar lavage fuid and formalin-fxed parafn-embedded (FFPE) tissues. Te combination of SHOX2 and RASSF1A methylation in bronchoalveolar   lavage fuid yielded a diagnostic sensitivity of 81.0% and specifcity of 97.4%, showing the highest diagnostic efciency compared with the established cytology examination and serum biomarker CEA [40]. Based on FFPE samples, the positive detection rates of the SHOX2 and RASSF1A panels in SCLC, SCC, and adenocarcinomas were 100%, 96.1%, and 82.9%, respectively. Furthermore, SHOX2 methylation, but not RASSF1A methylation, was correlated with the stages of lung cancer [32]. As promising markers of DNA methylation for the diagnosis of thoracic malignancies, few studies using PE-cfDNA have been conducted. In the feld of tumor targeted therapy, cfDNA derived from PE supernatant was reported as a high-quality material for molecular detection, which was readily available and less afected by the amount of cells shed [30].
In this study, we evaluated the promoter methylation status of SHOX2 and RASSF1A using PE-cfDNA in 104 PE cases with the Methylated Human SHOX2 and RASSF1A Gene Detection Kit (Tellgen Co. Ltd., Shanghai, China). Te results showed that the methylation levels of SHOX2 and RASSF1A were obviously higher in the MPE group than in the benign control group, which was in accordance with our understanding that hypermethylation is associated with malignancy. ROC curves were plotted to evaluate the diagnostic ability of SHOX2 and RASSF1A DNA methylation in PE, with AUC values of 0.959 for SHOX2, 0.810 for RASSF1A, and 0.985 for the combination analysis. Tis indicated that the methylation of SHOX2 or RASSF1A demonstrated an efcient diagnostic ability to distinguish MPE from BPE, with the combination performing best. Based on the calculated cutof values (ΔCt SHOX2 � 10.0 and ΔCt RASSF1A � 13.0), we found that the methylation positive rates of SHOX2 and RASSF1A were obviously higher in the MPE group (90.0% and 62.0%) than in the benign control group (0% and 0%) and the no-exact-diagnosis group (44.4% and 22.2%). Te combination of SHOX2 and RASSF1A greatly enhanced the detection positivity rate to 96%, with a high specifcity of 100%. Multiple potential diseases causing PE were included in our cohort, but a certain group had a limited number of samples. Lung cancer was the major cause of MPE (45/50). Te detection sensitivities of SHOX2, RASSF1A, and TM in lung cancer-derived MPE were 91.1%, 64.4%, and 97.8%, slightly higher than those in the overall MPE samples. In lung adenocarcinoma-derived MPE, the methylation positive rates were 91.9% for SHOX2 and 59.5% for RASSF1A, which were diferent from those in a previous study using histology specimens, 54.9% for SHOX2 and 46.3% for RASSF1A in adenocarcinoma [18]. In three lung squamous cell carcinoma-derived MPE patients, both SHOX2 and RASSF1A methylation detections were positive. Tere have also been studies about the distribution of the methylation status of SHOX2 and RASSF1A in diferent lung cancer subtypes using bronchial aspirates and FFPE samples. Te data showed that the positive detection rate of SHOX2 in SCLC was 80-100%, followed by a percentage of 63-96.1% in squamous cell carcinoma, while adenocarcinomas exhibited the lowest positive detection rate of 39-82.9% [41,42]. RASSF1A methylation was more frequently detected in adenocarcinomas (39%) than in squamous cell carcinomas (13%) [43]. In our cohort, MPE was principally adenocarcinoma-derived, and whether the methylation positive rates of SHOX2 and RASSF1A are diferent in specifc disease subtypes needs further exploration. Moreover, we applied the methylation test using PE-cfDNA and obtained ideal diagnostic results. A comparison between PE-cfDNA and the cell pellet samples will be performed in the future. In addition, four of the other fve MPE patients were positive, with single positivity for SHOX2 in ovarian and thymic squamous cell carcinoma and double positivity for SHOX2 and RASSF1A in lymphoma and adenoid cystic carcinoma. SHOX2 and RASSF1A methylation were not detected in mesothelioma, but individual samples were not representative. A larger sample size for each disease is needed to validate the results in the future. Overall, the positive detection results suggested the potential for the wide use of methylation detection in multiple malignancy-derived PE samples.
Several common serum biomarkers, such as NSE, CEA, and CYFRA 21-1, have been widely used in clinical practice [44]. In comparison with the conventional assays, the methylation analysis of the SHOX2 and RASSF1A panels using PE-cfDNA showed the best diagnostic power, with an AUC of 0.985. Furthermore, we found that it presents some advantages over traditional noninvasive cytological testing and invasive thoracoscopic as well as pleural biopsy techniques. In our study, of the thirteen MPE samples detected with atypical and negative cytology, the positive rate for combined SHOX2 and RASSF1A methylation assays using cfDNA reached 100%, which can substantially improve the diagnostic sensitivity of cytology. As a noninvasive detection method, it can be applied to most patients with PE and can provide an important basis for the determination of PE traits.
Tere were some limitations in the present study. First, the sample size was relatively small. Studies with a larger number of patients need to be conducted to validate the results in the future. Second, the MPE samples collected in the study were mainly derived from lung cancer patients, and more disease types should be included. Furthermore, whether the methylation changes of the two genes were correlated with the prognosis of patients is not clear.
In summary, the methylation analysis of the SHOX2 and RASSF1A panels using cf-DNA showed efcient diagnostic ability in diferentiating MPE from BPE and could be a potential tool to greatly enhance the results. It may have the potential to be a complementary tool for cytology in the diferential diagnosis of MPE.

Data Availability
Some or all data generated or used during the study can be obtained from the corresponding author upon request.