HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Michael A. Maddaus Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rubins, J. B. Articles by Niewoehner, D. E. Search for Related Content PubMed PubMed Citation Articles by Rubins, J. B. Articles by Niewoehner, D. E.

J Thorac Cardiovasc Surg 1998;116:412-416
© 1998 Mosby, Inc.

General Thoracic Surgery

Serum carcinoembryonic antigen as an adjunct to preoperative stagingof lung cancer

Supported by the Veterans Affairs Department of Research.

Received for publication Nov. 24, 1997. Revisions requested Jan. 30, 1998; revisions received May 14, 1998. Accepted for publication June 3, 1998. Address for reprints: Jeffrey B. Rubins, MD, Pulmonary (111N), VAMedical Center, One Veterans Dr, Minneapolis, MN 55417.

	Abstract

Top Abstract Introduction Patients and methods Results Discussion References

 
Objective: To determine whethermeasurement of preoperative serum carcinoembryonic antigen concentrations addsuseful prognostic data to current preoperative staging of lung cancer bycomputed tomography, bronchoscopy, and mediastinoscopy.
Methods: A prospective cohort study of 130 consecutivepatients was evaluated for suspected lung cancer from July 1991 through December1992 at a university-affiliated Veterans Affairs Medical Center. Serumconcentrations of carcinoembryonic antigen were measured before diagnosis,staging, or resection of cancer.
Results:Malignant disease was diagnosed by bronchoscopy, needle biopsy, mediastinoscopy,or resection in 111 of 130 patients. In the 50 patients undergoing resectionwith curative intent, multivariate analysis indicated that carcinoembryonicantigen was a significant predictor of survival independent of patient age,pathologic stage, histologic type, and tumor size (P =.0357).
Conclusions: Elevated preoperativeserum concentrations of carcinoembryonic antigen predict a poor prognosis forlung cancer independent of other conventional staging parameters and have anadjunctive role in the staging of lung cancer. (J Thorac Cardiovasc Surg1998;116:412-6)

	Introduction

Top Abstract Introduction Patients and methods Results Discussion References

 
Approximately 50% of patients die within 5 years of resection oflung cancer with curative intent, suggesting that the laboratory tests andimaging modalities currently used for preoperative staging have limitedsensitivities to detect unfavorable tumor biology and disseminated malignantdisease. Although many serologic and tissue tumor markers have been reported toidentify unfavorable tumor biology, carcinoembryonic antigen (CEA) is a serummarker that has the potential to augment current staging modalities byidentifying advanced lung cancer. ¹CEAs represent a heterogeneous group of oncofetal glycoprotein antigens thatcirculate in high concentrations in patients with certain malignant tumors,especially epithelial tumors. Although serum CEA assays have low sensitivity andspecificity for diagnosing lung cancers, ¹several reports have indicated that elevated preoperative CEA levels areassociated with more advanced disease and with very poor survival afterpresumptively curative surgical resection. ^2-6

Despite this evidence, the preoperative measurement of serum CEA has notbeen incorporated into the staging evaluation of patients considered forresection of lung cancer. Because assay of CEA is an inexpensive, readilyavailable preoperative serum test that may correlate with aspects of tumorbiology not measured by conventional staging modalities, we sought to re-examinein our clinical practice whether preoperative measurement of serum CEA addeduseful prognostic data to the current preoperative staging algorithm includingcomputed tomography (CT), bronchoscopy, and mediastinoscopy. In addition, weattempted to characterize the biologic plausibility of CEA as a predictive toolby correlating elevations in serum CEA with tumor size, histologic type, andstage.

	Patients and methods

Top Abstract Introduction Patients and methods Results Discussion References

 
Selection of patients
From July 1991 through December 1992, 142 consecutive patients wereevaluated by the Pulmonary Medicine Service at the Minneapolis Veterans AffairsMedical Center for lung cancer suspected on the basis of abnormal chestradiographs or unexplained hemoptysis. Of these, 130 patients gave informedconsent to phlebotomy of a single blood sample for determination of serum CEAbefore bronchoscopy, which was the first invasive study performed in allpatients. This study was approved by the Institutional Review Board on HumanStudies at the Minneapolis Veterans Affairs Medical Center.

In all 130 patients, the disease was staged by history, physicalexamination, laboratory tests including liver function and serum calcium, CTfrom the chest through the adrenal glands, and bronchoscopy. Patients in whomadvanced disease was suspected on the basis of systemic symptoms (especiallyweight loss), abnormal laboratory tests (especially elevations in levels ofalkaline phosphatase or calcium), or who had larger or more central tumors hadadditional radiographic studies including CT of the head and abdomen andradionuclide bone scans. Solitary lesions suggestive of metastatic spread ofcancer from distant sites were evaluated by CT-guided needle biopsy. Allpatients with potentially resectable lung cancer underwent exploration bymediastinoscopy and intraoperative bronchoscopy before thoracotomy.Contralateral lymph nodes were sampled during mediastinoscopy, and mediastinallymph node dissection was performed for pneumonectomy and lobectomy cases.Thoracic surgeons and patients were kept blinded to the results of CEAdeterminations throughout the study. All thoracotomies were performed within 4weeks of the serum CEA measurements.

Patient demographic data and pathology reports were obtained from thehospital computer database. Date of last contact or date of death was obtainedby scanning the hospital computer database and tumor registry for admissions,outpatient visits, and dates of death. For patients who could not be clearlyidentified as alive (outpatient or inpatient visit within the past 2 months) ordead from the hospital computer records, the Veterans Administration BeneficiaryIdentification and Records Locator System database was used to ascertain datesof death.

CEA assay
Blood was allowed to clot, and serum was separated by centrifugation andfrozen at –20°C until assayed. Serum CEA was measured by meansof the Tandem solid-phase 2-site radioimmunoassay (Hybritech Inc, San Diego,Calif). The detection limit of the assay was 1 ng/mL, with within-run andbetween-run coefficients of variation of 5.9% and 6.9%,respectively. According to the manufacturer, the upper limit of normal for thisassay for healthy individuals was 3 ng/mL; serum CEA values exceeded 5 ng/mL and10 ng/mL, respectively, in 11% and 7% of patients with malignantpulmonary disease.

Statistical analysis
The following analytic approach was taken to assess whether preoperativeCEA levels add useful prognostic information about length of survival beyondthat supplied by the standard prognostic variables of pathologic stage, tumorsize, histologic type, and patient age at the time of operation. This analysisincluded only those patients who underwent surgical resection of their tumor (n = 50). Pathologic stage was dichotomized intostage 3 or higher versus stages 1 and 2. Histologic type was categorized asadenocarcinoma, squamous cell carcinoma, or other. The date of operation wasincluded in the models to control for any possible bias that would be introducedas a result of the differing lengths of time at risk. Survival was calculatedfrom the date of operation to the date of death or last contact. A Coxproportional hazards model was fitted to the data incorporating the covariates(stage, tumor size, histologic type, age, and date of operation) other than CEA.CEA was then added to the model, after diagnostics examining plots of residualsand tests for power transformations indicated that the only usefultransformation was rounding the CEA level to the nearest integer (CEA values >25 were coded as 25). The likelihood ratio test was used to compare thedifference in the two models (with and without CEA). Risk ratios and 95%confidence intervals were derived from the model. Interaction terms between CEAand histologic type, size, and stage were subsequently introduced into the modelbut were not significant and are not presented here. From the model we alsoderived estimates of the effect of various CEA levels on 3-year survival forpatients with stage 1 and 2 squamous cell carcinoma and adenocarcinomaseparately, using the observed mean value for the other covariates.

To study the biologic plausibility of CEA as a prognostic indicator inlung cancer, we investigated univariate associations between CEA and histologictype, surgical stage, tumor size, and liver function tests. Because of theskewed distribution of serum CEA concentrations, median values are presented andassociations of serum CEA concentrations with histologic type and TNM stage wereanalyzed by Kruskal-Wallis 1-way analysis of variance with post hoc comparisonsby Wilcoxon rank sum tests. Associations between serum CEA concentrations andtumor size were by Pearson bivariate correlation. Analyses were performed bymeans of SPSS for Windows (release 6.1, Chicago, Ill) and SAS (release 6.11, SASInc, Cary, NC). All P values are two-tailed.

	Results

Top Abstract Introduction Patients and methods Results Discussion References

 
Between July 1991 and December 1992, 130 of 142 consecutive patientsevaluated for suspected lung cancer consented to have serum CEA concentrationsmeasured before diagnostic procedures. In 111 (109 male, 2 female) of these 130patients, a diagnosis of malignant disease was made by bronchoscopy, needlebiopsy, mediastinoscopy, or resection. Sixteen patients were considered not tobe surgical candidates because of preoperative histologic studies showing smallcell (13 patients), lymphoma (2 patients), or metastatic adenocarcinoma (1patient). An additional 43 patients had advanced non–small cell lungcancer determined by the staging methods described under Patientsand methods (unresectable stage 3A, 2 patients; stage 3B, 23patients; stage 4, 18 patients). Two additional patients were precluded fromsurgery because of severe lung disease. Thus 50 of 130 study patients withsuspected malignancy had apparently limited lung cancer after a full stagingevaluation and were taken to thoracotomy for resection. These 50 resectionsincluded 8 pneumonectomies, 39 lobectomies, and 3 wedge resections. Resectionmargins were histologically free of tumor in all cases. All 3 patientsundergoing wedge resections (because of limited cardiopulmonary reserve) had T1lesions with no evidence of nodal disease.

In a multivariate analysis of the 50 resected cases, we evaluated theutility of preoperative CEA as a predictor of long-term survival, as describedunder Patients and methods. The likelihoodratio test comparing the Cox proportional hazards model, first without and thenwith CEA added as a covariate to the model, indicated that CEA was a significantpredictor of survival independent of patient age, pathologic stage, histologictype, and tumor size (P = .0357). Themodel-estimated risk ratio for CEA level is 1.086 (95% CI: 1.009, 1.169),indicating that for each increase of CEA of 1 ng/mL, the hazard (risk) of deathincreases by 8.6%. Estimated probabilities of 3-year survival by CEAlevel for stage 1 and 2 adenocarcinoma and squamous cell carcinoma are shown inTable I. For example, the estimated probability of surviving 3years after surgery for patients with stage 1 or 2 squamous cell carcinoma isapproximately 60% for those with preoperative CEA levels less than 2ng/mL, but only 40% for those with CEA levels of 7 ng/mL, and 20%for those with CEA levels of 14 ng/mL. For patients with stage 1 or 2adenocarcinoma, the 3-year survival probabilities for these three levels of CEAare approximately 45%, 30%, and 10%, respectively. However,the confidence intervals around these survival probability estimates are widebecause of the relatively small number of cases available for analysis.

	Discussion

Top Abstract Introduction Patients and methods Results Discussion References

The biologic basis for the association between serum CEA and survival inlung cancer remains obscure. Serum CEA is related to tumor histologic type. Wefound significantly higher serum CEA concentrations in adenocarcinomas and smallcell carcinomas than in large cell and squamous cell carcinomas, in agreementwith previous reports. ^7,8,10,11Interestingly, tumor production of CEA measured by immunocytochemical stainingor assay of tumor cytosolic concentrations does not correlate with serum CEAconcentrations and does not predict survival in lung cancer, even whensimultaneously measured serum CEA levels do. ^2,9,12 In addition to the weakrelationship between cellular production of CEA and serum levels, we and othersfound no significant association with primary tumor size and serum CEAconcentrations. ^4,5 Although serum CEA may reflectproduction by tumor at other sites, we found no significant association betweenTNM stage and serum CEA concentrations in 130 cases, in agreement with otherstudies. ^4,6,7 In addition, one autopsy study ofpatients with lung cancer concluded that serum CEA concentrations are notrelated to the number of organs involved or the sites of metastatic cancer. ⁵ Serum CEA is cleared primarily by theliver and is thought to be elevated in colorectal carcinoma in cases of livermetastases. However, we found no association between liver function tests andserum CEA concentrations. Given the size of our study, we cannot definitivelyexclude a contribution of some of these factors to net serum CEA concentrations.At present, we can only conclude from our study and from publishedinvestigations that the exact mechanism of elevated serum CEA concentrations inlung cancers remains unclear.

In summary, serum CEA has all of the properties desired for a biologicmeasure to be used as a prognostic indicator in the clinical evaluation of lungcancer. It is objective, reproducible, and can be evaluated before theoperation. Although a number of other serum tumor markers, including tissuepolypeptide antigen and CA-125, appear to also have prognostic utility in thepreoperative evaluation of lung cancer, ⁹currently serum CEA is the only one of these that is inexpensive and routinelyavailable. Serum CEA also has distinct advantages over the variety of othertumor markers that are being investigated to better understand lung cancerbiology. Markers of histologic features, such as tumor grade, mitotic index, andvascular invasion, and immunocytochemical assays of oncogene products such asp53, ras, and others, ¹³ furnish at most semiquantitativeinformation, are subject to interobserver variation, and require resection of afinite volume of tumor for studies to be accomplished. Direct genetic studies,such as DNA sequencing and polymerase chain reaction single-strandedconformational polymorphism, are presently available only as research tools andhave the same limitations.

Taking together our data and previously published studies, we concludethat routine preoperative serum CEA measurements are an important adjunct toconventional anatomic tumor staging algorithms, particularly for patients withadenocarcinoma. Marked elevations of serum CEA are indications for full andcareful staging despite apparent localized disease on chest radiographs and CT.In such patients, especially those who are marginal surgical candidates due tounderlying illnesses, the poor prognosis implied by the test may be factoredinto decisions regarding surgical resection. Further studies are required todetermine whether immediate or delayed postoperative adjuvant treatment mightimprove the outcome for these patients. Although such therapies have not hadconclusive benefit for patients with non–small cell lung cancer, the useof serum tumor markers such as CEA may be a superior means of identifying suchpatients for these adjuvant therapies.

	Acknowledgments

 
We thank Ms Patti Baribeau, Minneapolis Veterans AffairsTumor Registry, and Ms Mary Macauley, Minneapolis Veterans Affairs Department ofThoracic Surgery, for assistance in obtaining patient demographic data,pathology reports, and outcomes. We thank David Nelson, PhD, for invaluableassistance in statistical data analysis.

	References

Top Abstract Introduction Patients and methods Results Discussion References

 

1. NIH. Carcinoembryonic antigen: its role as amarker in the management of cancer. BMJ 1981;282:373-5.
   
2. 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;44:145-53.[Medline]
   
3. Dent PB, McCulloch PB, Wesley-James O,MacLaren R, Muirhead W, Dunnett CW. Measurement of carcinoembryonic antigen inpatients with bronchogenic carcinoma. Cancer 1978;42:1484-91.[Medline]
   
4. Stokes TC, Stevens JFS, Long P, Lockey E,Miller AL. Preoperative carcinoembryonic antigen and survival after resection oflung cancer. Br J Dis Chest 1980;74:390-4.
   
5. Vincent RG, Chu TM, Fergen TB, Ostrander M.Carcinoembryonic antigen in 228 patients with carcinoma of the lung. Cancer 1975;36:2069-76.[Medline]
   
6. Concannon JP, Dalbow MH, Hodgson SE, et al.Prognostic value of preoperative carcinoembryonic antigen (CEA) plasma levels inpatients with bronchogenic carcinoma. Cancer 1978;42:1477-83.[Medline]
   
7. Niklinkski J, Furman M, Laudanski J, KozlowskiM. Prognostic value of pretreatment CEA, SCC-Ag and CA 19-9 levels in sera ofpatients with non–small cell lung cancer. Eur J Cancer Prevent 1992;1:401-6.[Medline]
   
8. Niklinkski J, Furman M, Laudanski J,Palynyczko Z, Welk M. Evaluation of carcinoembryonic antigen (CEA) andbrain-type creatine kinase (CK-BB) in serum from patients with carcinoma of thelung. Neoplasma 1991;38:129-35.[Medline]
   
9. Diez M, Torres A, Maestro ML, et al.Prediction of survival and recurrence by serum and cytosolic levels of CEA,CA125 and SCC antigens in resectable non-small-cell lung cancer. Br JCancer 1996;73:1248-54.
   
10. Icard P, Regnard J, Essomba A, Panebianco V,Magdeleinat P, Levasseur P. Preoperative carcinoembryonic antigen level as aprognostic indicator in resected primary lung cancer. Ann Thorac Surg 1994;58:811-14.[Abstract]
    
11. Ratto GB, Mereu C, Rovida S. Multiple tumourmarkers for diagnosis, management and follow-up of potentially resectable lungcancer. Panminerva Med 1993;35:186-92.[Medline]
    
12. Graziano SL, Tatum AH, Newman NB, et al. Theprognostic significance of neuroendocrine markers and carcinoembryonic antigenin patients with resected stage I and II non–small cell lung cancer.Cancer Res 1994;54:2908-13.[Abstract/Free Full Text]
    
13. Greatens TM, Niehans GA, Rubins JB, et al. Domolecular markers predict survival in non–small cell lung cancer? Am JRespir Crit Care Med 1998;157:1093-7.[Abstract/Free Full Text]
    

This article has been cited by other articles:

				K. Matsuoka, S. Sumitomo, N. Nakashima, D. Nakajima, and N. Misaki Prognostic value of carcinoembryonic antigen and CYFRA21-1 in patients with pathological stage I non-small cell lung cancer Eur. J. Cardiothorac. Surg., September 1, 2007; 32(3): 435 - 439. [Abstract] [Full Text] [PDF]

				R. Fukai, Y. Sakao, M. Sakuraba, S. Oh, K. Shiomi, S. Sonobe, Y. Saitoh, and H. Miyamoto The prognostic value of carcinoembryonic antigen in T1N1M0 and T2N1M0 non-small cell carcinoma of the lung Eur. J. Cardiothorac. Surg., September 1, 2007; 32(3): 440 - 444. [Abstract] [Full Text] [PDF]

				M. Tomita, T. Shimizu, Y. Matsuzaki, M. Hara, T. Ayabe, and T. Onitsuka Prognostic Significance of Carcinoembryonic Antigen Level in Pleural Lavage Fluid for Patients With Lung Adenocarcinoma Ann. Thorac. Surg., July 1, 2005; 80(1): 276 - 281. [Abstract] [Full Text] [PDF]

				Y. Sakao, S. Tomimitsu, Y. Takeda, M. Natsuaki, and T. Itoh Carcinoembryonic antigen as a predictive factor for postoperative tumor relapse in early-stage lung adenocarcinoma Eur. J. Cardiothorac. Surg., April 1, 2004; 25(4): 520 - 522. [Abstract] [Full Text] [PDF]

				N. Sawabata, H. Hirano, M. Inoue, Y. Okumura, H. Asada, S.-i. Takeda, and H. Maeda Immunohistochemical analysis of resectedclinical Stage I pulmonary adenocarcinomas withhigh preoperative levels of serumcarcinoembryonic antigen Ann. Thorac. Surg., July 1, 2003; 76(1): 203 - 207. [Abstract] [Full Text] [PDF]

				T. Aoki, M. Tsuchida, T. Watanabe, T. Hashimoto, T. Koike, T. Hirono, and J.-i. Hayashi Surgical strategy for clinical stage I non-small cell lung cancer in octogenarians Eur. J. Cardiothorac. Surg., April 1, 2003; 23(4): 446 - 450. [Abstract] [Full Text] [PDF]

				N. Sawabata, M. Ohta, S.-i. Takeda, H. Hirano, Y. Okumura, H. Asada, and H. Maeda Serum carcinoembryonic antigen level in surgically resected clinical stage i patients with non-small cell lung cancer Ann. Thorac. Surg., July 1, 2002; 74(1): 174 - 179. [Abstract] [Full Text] [PDF]

				K. Takamochi, K. Nagai, J. Yoshida, K. Suzuki, Y. Ohde, M. Nishimura, K. Takahashi, and Y. Nishiwaki THE ROLE OF COMPUTED TOMOGRAPHIC SCANNING IN DIAGNOSING MEDIASTINAL NODE INVOLVEMENT IN NON-SMALL CELL LUNG CANCER J. Thorac. Cardiovasc. Surg., June 1, 2000; 119(6): 1135 - 1140. [Abstract] [Full Text] [PDF]

				K. Takamochi, K. Nagai, K. Suzuki, J. Yoshida, Y. Ohde, and Y. Nishiwaki Clinical Predictors of N2 Disease in Non-small Cell Lung Cancer Chest, June 1, 2000; 117(6): 1577 - 1582. [Abstract] [Full Text] [PDF]

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Michael A. Maddaus Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rubins, J. B. Articles by Niewoehner, D. E. Search for Related Content PubMed PubMed Citation Articles by Rubins, J. B. Articles by Niewoehner, D. E.