Prognostic Significance of Carcinoembryonic Antigen Level in Pleural Lavage Fluid for Patients With Lung Adenocarcinoma

Transcription

1 Prognostic Significance of Carcinoembryonic Antigen Level in Pleural Lavage Fluid for Patients With Lung Adenocarcinoma Masaki Tomita, MD, PhD, Tetsuya Shimizu, MD, PhD, Yasunori Matsuzaki, MD, PhD, Masaki Hara, MD, PhD, Takanori Ayabe, MD, PhD, and Toshio Onitsuka, MD, PhD Department of Surgery II, Miyazaki Medical College, University of Miyazaki, Miyazaki, Japan Background. Prognostic indicators for non-small cell lung cancer patients traditionally include TNM staging, pleural lavage cytology, and serum carcinoembryonic antigen levels. This prospective study evaluates carcinoembryonic antigen levels in pleural lavage fluid as a potential determinant for patients with lung adenocarcinoma. Methods. One hundred and fifty patients underwent thoracotomy. Pleural lavage fluid was collected, and pleural lavage cytology and lavage carcinoembryonic antigen levels were determined. The control group included 40 patients with nonmalignant disease. Results. Sixteen patients (10.7%) had positive pleural lavage cytologies. These patients and those with elevated serum carcinoembryonic antigen levels generally had a poor prognosis. Thirty-seven patients (24.7%), however, showed elevated lavage carcinoembryonic antigen levels, and a significant correlation with patient survival was demonstrated. Multivariate analysis confirmed these results. We also found a correlation between positive pleural lavage cytologies and serum carcinoembryonic antigen levels and patient survival in patients with pn0 disease but not in those with pn1-2 disease. Elevated lavage carcinoembryonic antigen levels, however, correlated significantly with survival rates in patients with pn1-2 disease. Conclusions. An elevated lavage carcinoembryonic antigen level is an independent prognostic determinant for patients with lung adenocarcinoma, even with advanced disease, and may be a more useful marker of subclinical microdissemination than pleural lavage cytology. (Ann Thorac Surg 2005;80:276 81) 2005 by The Society of Thoracic Surgeons The best predictor for patient prognosis in non-small cell lung cancer (NSCLC) is TNM classification. Although these staging methods tend to concentrate on macroscopic evidence of disseminated disease, several studies have evaluated the usefulness of pleural lavage cytology (PLC) and have suggested that positive PLC findings may indicate a more aggressive tumor biology, a higher risk of recurrence, and a poorer survival rate [1 7]. Some researchers advocate adding PLC to the TNM staging system for NSCLC in order to stratify patients more accurately [2 4]. In addition to cytology, molecular techniques, including reverse transcriptase polymerase chain reaction, have been used to detect peritoneal dissemination of gastric and ovarian cancers and have proved to be more sensitive [8]. At present, one of the most commonly used tumor markers is the serum carcinoembryonic antigen (CEA). Several reports have indicated that preoperative elevated serum CEA levels are associated with more advanced disease and with very poor survival rates after surgical resection [9 15]. Since CEA may reflect tumor production, CEA messenger ribonucleic acid (mrna) levels have also Accepted for publication Jan 7, Address reprint requests to Dr Tomita, Department of Surgery II, University of Miyazaki, Miyazaki Medical College, Kihara 5200, Kiyotake, Miyazaki , Japan; mtomita@post.miyazaki-med.ac.jp. been used to detect microdissemination or micrometastasis [16 18]. Based on these results, we hypothesized that CEA levels in pleural lavage fluid may be useful markers of microscopic advanced disease. To ascertain the frequency and to quantify the effect of this finding on patient survival, we performed a prospective study of intraoperative PLC and CEA levels in pleural lavage fluid. Earlier reports have documented that serum CEA levels in adenocarcinomas are significantly higher than in squamous cell carcinomas [10, 13, 15], and it is conceivable that adenocarcinomas may characteristically produce higher values than squamous cell carcinomas [10]. Since we previously demonstrated a prognostic significance of serum CEA levels in patients with lung adenocarcinoma but not in patients with squamous cell carcinoma [15], we limited our present study to patients with adenocarcinoma. Material and Methods This study was conducted from 1997 through Patients with macroscopic pleural effusions, extreme adhesions, or adenocarcinomas in other organs were excluded. One hundred and fifty consecutive lung adenocarcinoma patients (87 men, 63 women) who fulfilled the inclusion criteria and who underwent thoracotomy for intended surgical resection were included in this 2005 by The Society of Thoracic Surgeons /05/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg TOMITA ET AL 2005;80: LAVAGE CEA IN LUNG ADENOCARCINOMA 277 Table 1. Results of Pleural Lavage Cytology and Lavage CEA Level According to Clinicopathologic Characteristics PLC Lavage CEA Total Positive Negative p Value Elevated Normal p Value Male Female Age Age p-stage I p-stage II p-stage III pt pt pt pt pn pn pn p p p p Well differentiated Moderately differentiated Poorly differentiated Serum CEA elevated Serum CEA normal CEA carcinoembryonic antigen; PLC pleural lavage cytology. study. Ages ranged from 37 to 90 years, with an average age of years. Baseline characteristics and stage classifications are summarized in Table 1. Overall follow-up periods ranged from 24 to 83 months. The clinical investigation section of our hospital measured serum CEA levels using the twosite immunoenzymometric assay; the normal upper limit for this assay was 5 ng/ml. The time interval between serum CEA determination and staging or surgical resection was less than a month in all patients. Immediately after thoracotomy and prior to any further manipulation of the pulmonary parenchyma, the pleural cavity was carefully washed with 200 ml of physiologic saline solution. The fluid was irrigated over the visceral and parietal pleura. Special care was taken to prevent undue contamination of the lavage fluid with blood and to avoid contact with the pleural surface in order to collect only desquamated cells. The fluid was placed in a glass bottle containing heparin and centrifuged at 1,500 rpm for 5 minutes. Pleural lavage CEA levels were measured using the same procedure as for serum CEA levels. The sediment was stained using Giemsa and Papanicolaou methods. Cytologic results were divided into two categories: negative and positive. Papanicolaou classes I, II, and III were regarded as negative; classes IV and V were considered positive. The PLC and lavage CEA levels were determined for 40 control patients with benign lung disease, including 10 tuberculomas, 8 inflammatory pseudotumors, 4 hamartomas, 4 fungus infections, and others. Pathologic TNM staging was recorded for all patients. Pleural invasion (p status) was classified as follows: p0 tumor with no pleural involvement or reaching the visceral pleura but not extending beyond its elastic layer; p1 tumor extending beyond the elastic layer of the visceral pleura but not exposed on the pleural surface; p2 tumor exposed on the pleural surface but not involving the parietal pleura; and p3 tumor invading the visceral and parietal pleura. Follow-up information, including cause of death, was ascertained through a review of clinic notes and direct or family contact. Comparisons of categorical data between the two groups were made using linear regression analysis or the Mann-Whitney U test. Survival curves were calculated according to the Kaplan-Meier method, and the log-rank test was used to compare survival curves. Factors related to prognosis were analyzed by multivariate analyses according to the Cox proportional hazards model. Statistical calculations were conducted with Stat- View (Abacus Comp Inc, Berkeley, CA), and p values of less than 0.05 were considered significant. Results Positive PLC findings were present in 16 (10.7%) of the lung adenocarcinoma patients. In the 40 control patients

3 278 TOMITA ET AL Ann Thorac Surg LAVAGE CEA IN LUNG ADENOCARCINOMA 2005;80: Table 2. Relationship Between Lavage CEA Level and PLC Findings Lavage CEA Level Elevated Normal p Value PLC positive PLC negative CEA carcinoembryonic antigen; PLC pleural lavage cytology. with benign lung disease, all PLC findings were negative, and all lavage CEA levels were less than 0.5 ng/ml. From these results, we determined the normal upper limit for lavage CEA to be 0.5 ng/ml. Thirty-seven patients showed an elevated lavage CEA ( ng/ml). Using linear regression analysis, we found a significant correlation between serum and lavage CEA levels (correlation coefficient [r] 0.179; coefficient of correlation [R2] 0.032; p 0.029). We detected no correlation between PLC and lavage CEA (r 0.108; R ; p 0.189). Clinicopathologic characteristics, PLC findings, and lavage CEA levels are shown in Table 1. Positive PLC findings were significantly related to pathologic stage (p 0.013), T status (p 0.001), and p status (p ). Even in pathologic stage I disease or p0 disease, however, some patients were positive for malignancy. Although we detected a trend toward a correlation between positive PLC findings and N status, it did not reach statistical significance (p 0.183). Our data revealed a significant correlation between elevated lavage CEA levels and pathologic stage (p ), T status (p 0.001), N status (p ), and p status (p 0.039). Using the Mann- Whitney U test, we also demonstrated a correlation between lavage CEA and serum CEA (p ). The PLC findings and lavage CEA levels did not appear to be influenced by age, gender and histologic differentiation. The ratio of elevated lavage CEA levels in pathologic stage I disease or p0 disease was higher than that of positive PLC findings. Table 2 compares PLC findings and lavage CEA levels. Twenty-seven patients had a negative PLC and an elevated lavage CEA; 6 patients had a positive PLC and a normal lavage CEA. While linear regression analysis did not detect a relationship between PLC and lavage CEA, the Mann-Whitney U test showed a correlation. As seen in Figures 1A and 1B, serum CEA levels and positive PLC findings were significantly related to overall patient survival. Figure 1C, however, shows that an elevated lavage CEA is a stronger prognostic factor. Survival curves were also compared based on PLC findings and lavage CEA levels. As shown in Figure 2A, patients with a negative PLC and an elevated lavage CEA had a poorer prognosis than patients with a positive PLC and a normal lavage CEA. A comparison of survival curves based on serum and lavage CEA levels is shown in Figure 2B. When both the serum and the lavage CEA were elevated, patients had an unfavorable prognosis. When both CEAs were normal, patients had a much more favorable prognosis. When only one marker was Fig 1. Overall survival of patients according to serum CEA level (thick line serum CEA normal; thin line serum CEA elevated) (A), the findings of PLC (thick line PLC negative; thin line PLC positive) (B), and lavage CEA level (thick line lavage CEA normal; thin line lavage CEA elevated) (C). (CEA carcinoembryonic antigen; PLC pleural lavage cytology.) elevated, the patient survival rate was an intermediate value. Multivariate Cox analysis was also performed. As shown in Table 3, lavage CEA and pn status were independent prognostic factors (p and p 0.009,

4 Ann Thorac Surg TOMITA ET AL 2005;80: LAVAGE CEA IN LUNG ADENOCARCINOMA 279 Fig 3. Survival of patients with pn1-2 disease according to lavage serum carcinoembryonic antigen (CEA) level. (thick line lavage CEA normal; thin line lavage CEA elevated.) respectively). In patients with pn0 status, serum CEA levels and positive PLC findings were related to patient survival (p and p 0.001, respectively; data not shown). In pn1-2 patients, however, serum CEA levels and positive PLC findings did not reach statistical significance (p and p 0.247, respectively; data not shown). In contrast, as shown in Fig 3, lavage CEA proved to be a significant prognostic factor in pn1-2 patients (p 0.003). Fig 2. Survival of patients based on the findings of pleural lavage cytology (PLC) and lavage carcinoembryonic antigen (CEA) level (A) and based on serum and lavage CEA level (B). (Fig 2A: bold line PLC negative/lavage CEA normal; thin line PLC positive/lavage CEA normal; bold gray line PLC negative/lavage CEA elevated; thin gray line PLC positive/lavage CEA elevated. Fig 2B: bold line serum CEA normal/lavage CEA normal; thin line serum CEA normal/lavage CEA elevated; bold gray line serum CEA elevated/lavage CEA normal; thin gray line serum CEA elevated/ lavage CEA elevated.) Comment In our study, 10.7% (16 of 150) of the patients had a positive PLC. Previous studies have reported the rate of positive PLC findings to be 4.5% to 38% [1 7]; a wide range attributable to differences in patient stages or to the use of different techniques for collecting the samples or different diagnostic standards for evaluating them. These studies further indicate that the volume of lavage fluid ranged from 50 to 1,500 ml and that different carrier solutions including saline and Ringer s lactate had been employed. It is not surprising, then, that the relationship between positive PLC findings and clinicopathologic factors remains controversial. Most previous studies, how- Table 3. Multivariate Analysis Unfavorable Favorable Hazard Ratio 95% CI p Value Age Higher Lower Gender Male Female p-stage II-III I pt status pt2-4 pt pn status pn1-2 pn p status p1 3 p Histologic differentiation Moderately/poorly Well serum CEA level Elevated Normal PLC Positive Negative Lavage CEA level Elevated Normal CEA carcinoembryonic antigen; CI confidence interval; PLC pleural lavage cytology.

5 280 TOMITA ET AL Ann Thorac Surg LAVAGE CEA IN LUNG ADENOCARCINOMA 2005;80: ever, do conclude that patients with a positive PLC, a possible indicator of the initial stage of carcinomatous pleuritis, have a poor survival rate because a small amount of malignant pleural effusion or a few minute pleural dissemination nodules are probably overlooked at thoracotomy [1 7]. Our present study confirms this as well even though some of our results, including the relationship between PLC and N status, were not consistent with previous studies and may be attributable to different patient populations or to the use of different techniques or diagnostic standards. In addition to positive PLC findings, our data demonstrated a significant correlation between elevated lavage CEA levels and clinicopathologic factors, suggesting that an elevated lavage CEA may be a useful prognostic determinant. Despite this statistically significant correlation, 10 patients showed normal serum but elevated lavage CEA levels. In other words, among 100 patients with normal serum CEAs, 10 patients could have a poorer prognosis based on their elevated lavage CEA levels. In contrast, 23 patients had an elevated serum CEA but a normal lavage CEA. Although the reason for these discrepancies is not known, there is the possibility that extraction mechanisms of CEA in serum and pleural lavage fluid may differ. The combined use of serum and lavage CEA levels, however, may prove a useful prognostic determinant because patients with elevated CEA levels in both serum and lavage had a poor prognosis. Furthermore, an elevated lavage CEA level may be a marker of advanced disease and reflect subclinical microdissemination that is not always detected by PLC. We detected a positive PLC in only 10.7% (16 of 150) of our patients as opposed to an elevated lavage CEA in 24.7% (37 of 150). These findings suggest that an elevated lavage CEA is a more sensitive marker of microdissemination than PLC. Six patients, however, had normal lavage CEA levels but positive PLC findings. These false-negatives may be due to the potential limitation of the techniques used in this study and could possibly be reduced by using other procedures such as molecular techniques [8, 16 18]. Further study is required to reduce the falsenegative ratio. Twenty-seven patients showed an elevated lavage CEA and a negative PLC. The survival rate of these 27 patients was significantly poorer than that of the 6 patients with a normal lavage CEA and a positive PLC, indicating that an elevated lavage CEA level is a stronger prognostic determinant than PLC. Although we demonstrated a significant correlation between a positive PLC and p status, we also found a positive PLC in some patients with p0-1 disease, in whom the tumor was covered by a mesothelial layer and not exposed to the pleural cavity. Additionally, some of these patients with p0-1 disease also had elevated lavage CEA levels. Therefore, in addition to p status, it appears that other factors may produce malignant pleural effusion. Previous researchers have speculated that a positive PLC results from a combination of impaired lymphatic drainage through the intrapulmonary lymphatic channel to the mediastinal nodes and exfoliation of tumor cells [1 7]. As reported earlier, positive PLC findings and serum CEA levels are significant prognostic determinants [1 7, 9 15]. Although our data support these findings, we found that positive PLC findings and serum CEA levels are not prognostic factors for pn1-2 patients. In fact, previous studies reporting on the prognostic significance of positive PLC findings or serum CEA levels were based on early stage or overall patients, and none of these studies were limited to pn1-2 patients [1 7, 9 15]. It is possible that our data showing positive PLC findings and normal serum CEA levels in pn1-2 disease may be due to the small number of patients in our study. Interestingly, however, our data do indicate that an elevated lavage CEA is a significant prognostic determinant for pn1-2 patients; ie, a useful prognostic predictor in patients with advanced disease. On the basis of our results, pn1-2 patients with normal lavage CEA levels may be among a favorable prognostic subgroup even when treated with surgery alone. Thus, we believe that, based on lavage CEA levels, different postoperative therapeutic approaches may be available to pn1-2 patients. The measurement of lavage CEA levels is an inexpensive method and routinely available. Although some studies [16 18] have investigated the use of the more sensitive CEA mrna to detect microdissemination or micrometastasis, these molecular detections of tumor cells are expensive and available only as research tools. In conclusion, an elevated lavage CEA level appears to be an independent prognostic determinant in patients with lung adenocarcinoma, even with advanced disease, and may be a more useful marker of subclinical microdissemination than pleural lavage cytology. When planning postoperative adjuvant therapies, we believe that lavage CEA levels should be considered. Unfortunately, there is no evidence that adjuvant therapy would be useful in patients with elevated lavage CEA levels, but this may be a question for future studies. The authors wish to thank Nancy Folks for her support in preparing this manuscript. References 1. Okumura M, Ohshima S, Kotake Y, Morino H, Kikui M, Yasumitsu T. Intraoperative pleural lavage cytology in lung cancer patients. Ann Thorac Surg 1991;51: Kondo H, Asamura H, Suemasu K, et al. Prognostic significance of pleural lavage cytology immediately after thoracotomy in patients with lung cancer. J Thorac Cardiovasc Surg 1993;106: Buhr J, Berghauser KH, Gonner S, Kelm C, Burkhardt EA, Padberg WM. The prognostic significance of tumor cell detection in intraoperative pleural lavage and lung tissue cultures for patients with lung cancer. J Thorac Cardiovasc Surg 1997;113: Dresler CM, Fratelli C, Babb J. Prognostic value of positive pleural lavage in patients with lung cancer resection. Ann Thorac Surg 1999;67: Riquet M, Badoual C, Le Pimpec Barthes F, et al. Visceral pleura invasion and pleural lavage tumor cytology by lung cancer: a prospective appraisal. Ann Thorac Surg 2003;75: Okada M, Sakamoto T, Nishio W, Uchino K, Tsuboshima K, Tsubota N. Pleural lavage cytology in non-small cell lung

6 Ann Thorac Surg TOMITA ET AL 2005;80: LAVAGE CEA IN LUNG ADENOCARCINOMA 281 cancer: lessons from 1000 consecutive resections. J Thorac Cardiovasc Surg 2003;126: Lim E, Ali A, Theodorou P, Nicholson AG, Ladas G, Goldstraw P. Intraoperative pleural lavage cytology is an independent prognostic indicator for staging non-small cell lung cancer. J Thorac Cardiovasc Surg 2004;127: Nakanishi H, Kodera Y, Yamamura Y, et al. Molecular diagnostic detection of free cancer cells in the peritoneal cavity of patients with gastrointestinal and gynecologic malignancies. Cancer Chemother Pharmacol Suppl 1999;43:S Ford CHJ, Stokes HJ, Newman CE. Carcinoembryonic antigen and prognosis after radical surgery for lung cancer: immunocytochemical localization and serum levels. Br J Cancer 1981;44: Vincent RG, Chu TM, Fergen TB, Ostrander M. Carcinoembryonic antigen in 228 patients with carcinoma of the lung. Cancer 1975;36: Concannon JP, Dalbow MH, Hodgson SE, et al. Prognostic value of preoperative carcinoembryonic antigen (CEA) plasma levels in patients with bronchogenic carcinoma. Cancer 1978;42: Icard P, Regnard JF, Essomba A, Panebianco V, Magdeleinat P, Levasseur P. Preoperative carcinoembryonic antigen level as a prognostic indicator in resected primary lung cancer. Ann Thorac Surg 1994;58: Rubins JB, Dunitz J, Rubins HB, Maddaus MA, Niewoehner DE. Serum carcinoembryonic antigen as an adjunct to preoperative staging of lung cancer. J Thorac Cardiovasc Surg 1998;116: Okada M, Nishio W, Sakamoto T, et al. Prognostic significance of perioperative serum carcinoembryonic antigen in non-small cell lung cancer: analysis of 1,000 consecutive resections for clinical stage I disease. Ann Thorac Surg 2004;78: Tomita M, Matsuzaki Y, Edagawa M, Shimizu T, Hara M, Onitsuka T. Prognostic significance of preoperative serum carcinoembryonic antigen level in lung adenocarcinoma but not squamous cell carcinoma. Ann Thorac Cardiovasc Surg 2004;10: Castaldo G, Tomaiuolo R, Sanduzzi A, Ponticiello A, Marchetiello I, Salvatore F. Carcinoembryonic antigen mrna analysis detects micrometastatic cells in blood from lung cancer patients. Eur Respir J 2003;22: Kurusu Y, Yamashita J, Ogawa M. Detection of circulating tumor cells by reverse transcriptase-polymerase chain reaction in patients with resectable non-small-cell lung cancer. Surgery 1999;126: D Cunha J, Corfits AL, Herndon JE II, et al. Molecular staging of lung cancer: real-time polymerase chain reaction estimation of lymph node micrometastatic tumor cell burden in stage I non-small cell lung cancer preliminary results of Cancer and Leukemia Group B Trial J Thorac Cardiovasc Surg 2002;123: Member and Individual Subscriber Access to the Online Annals The address of the electronic edition of The Annals is If you are an STS or STSA member or a non-member personal subscriber to the print issue of The Annals, you automatically have a subscription to the online Annals, which entitles you to access the full-text of all articles. To gain full-text access, you will need your CTSNet user name and password. Society members and non-members alike who do not know their CTSNet user name and password should follow the link Forgot your user name or password? that appears below the boxes where you are asked to enter this information when you try to gain full-text access. Your user name and password will be ed to the address you designate. In lieu of the above procedure, if you have forgotten your CTSNet username and/or password, you can always send an to CTSNet via the feedback button from the left navigation menu on the homepage of the online Annals or go directly to We hope that you will view the online Annals and take advantage of the many features available to our subscribers as part of the CTSNet Journals Online. These include inter-journal linking from within the reference sections of Annals articles to over 350 journals available through the HighWire Press collection (HighWire provides the platform for the delivery of the online Annals). There is also crossjournal advanced searching, etoc Alerts, Subject Alerts, Cite-Track, and much more. A listing of these features can be found at We encourage you to visit the online Annals at ats.ctsnetjournals.org and explore by The Society of Thoracic Surgeons Ann Thorac Surg 2005;80: /05/$30.00 Published by Elsevier Inc