|Year : 2014 | Volume
| Issue : 2 | Page : 239-243
The role of serum carcinoembryonic antigen in predicting responses to chemotherapy and survival in patients with non-small cell lung cancer
Hongbing Liu1, Xiaoling Gu1, Tangfeng Lv1, Ying Wu2, Yongying Xiao1, Dongmei Yuan1, Yufeng Li1, Yong Song1
1 Department of Respiratory Medicine, Jinling Hospital, Nanjing University School of Medicine, Nanjing, China
2 Jiangsu Province Geriatric Hospital, Jiangsu Province Geriatric Institute, Nanjing, China
|Date of Web Publication||14-Jul-2014|
Department of Respiratory Medicine, Jinling Hospital, Nanjing University School of Medicine, 305 East Zhongshan Road, Nanjing 210002 Jiangsu Province
Source of Support: None, Conflict of Interest: None
Background: Previous studies indicated that carcinoembryonic antigen (CEA) could predict the therapeutic objective response (OR) and overall survival (OS) of patients with cancers, including non-small cell lung cancer (NSCLC). However, the role it could play in evaluating therapeutic responses and OS in patients with NSCLC requires further elucidation. Herein, we investigated the potential role of CEA in predicting OR and OS in patients with NSCLC.
Materials and Methods: For this retrospective study, the medical records of 689 patients with NSCLC who were treated at Nanjing Jinling Hospital between January 2000 and August 2011 were reviewed. Serum levels of CEA of these patients were measured before and after chemotherapy. The relatedness between CEA levels and OR, and between CEA and OS were investigated for correlations via a series of statistical analyses.
Results: The baseline serum CEA level of 689 patients was 54.18 ± 143.45 ng/mL. Serum CEA significantly decreased after two cycles of chemotherapy (t = 2.196, P = 0.031). The receiver operator characteristic (ROC) curve analysis showed that a 5.28% reduction in CEA level was an appropriate cut-off value for predicting the OR to chemotherapy, with a sensitivity of 61.3% and a specificity of 62.4%. The Kaplan-Meier survival analysis indicated no significant correlation between baseline CEA and OS (P = 0.079).
Conclusion: Our study shows that while the baseline level of CEA was not a prognostic factor, the post-treatment reduction of CEA can predict the OR in patients with NSCLC.
结果：689例病人的基线血清CEA水平为54.18±143.45ng/ml。化疗2个周期后血清CEA明显降低（t = 2.196，P = 0.031）。ROC曲线分析表明，CEA水平降低5.28%是一个适当的截断值，在预测化疗的治疗反应(OR)方面，敏感性为61.3%，特异性为62.4%。Kaplan-Meier生存分析显示基线CEA和OS无显著相关性（P = 0.079）。
Keywords: Carcinoembryonic antigen, non-small cell lung cancer, objective response, overall survival
|How to cite this article:|
Liu H, Gu X, Lv T, Wu Y, Xiao Y, Yuan D, Li Y, Song Y. The role of serum carcinoembryonic antigen in predicting responses to chemotherapy and survival in patients with non-small cell lung cancer. J Can Res Ther 2014;10:239-43
|How to cite this URL:|
Liu H, Gu X, Lv T, Wu Y, Xiao Y, Yuan D, Li Y, Song Y. The role of serum carcinoembryonic antigen in predicting responses to chemotherapy and survival in patients with non-small cell lung cancer. J Can Res Ther [serial online] 2014 [cited 2019 Nov 14];10:239-43. Available from: http://www.cancerjournal.net/text.asp?2014/10/2/239/136541
| > Introduction|| |
Non-small cell lung cancer (NSCLC) afflicts approximately 85% of patients diagnosed with lung malignancies. The majority of NSCLC patients are diagnosed at late or advanced stages, and systemic combination chemotherapy has become the standard approach in these patients. , Unfortunately, an objective response (OR) is seen in only about 30% of NSCLC patients treated with chemotherapy. Furthermore, despite the progress in diagnosis and treatment of these patients, the overall 5-year survival rate remains unchanged (15%).  Thus, more studies exploring the potential role of tumor markers in predicting OR and overall survival (OS) in patients with NSCLC are greatly needed.
Several studies have shown that tumor markers may be able to predict the OR to chemotherapy and OS of cancer patients. ,,, Tumor markers in NSCLC include soluble fragment of cytokeratin 19 (CYFRA21-1), neuron-specific enolase (NSE), cancer antigen 125 (CA-125), squamous cell carcinoma antigen (SCC-Ag), and carcinoembryonic antigen (CEA).  Among these, CEA has been considered as an important prognostic indicator in several cancers including lung. ,, This study investigated the association between serum CEA and OR to chemotherapy or OS in patients with NSCLC.
| > Materials and methods|| |
Patient samples and treatment evaluation
Six hundred and eighty-nine patients with NSCLC diagnosed in our institution between January 2000 and August 2011 were enrolled retrospectively in this study. Diagnoses of all patients were confirmed by pathologic examination. Patients were excluded if they had medical conditions other than NSCLC that show high serum CEA levels, or suspected or confirmed co-existing gastrointestinal tumors (colon cancers) or benign diseases (colitis and pancreatitis).
The clinical stages of the enrolled NSCLC patients ranged from stage I to IV based on the guidelines of the tumor-node-metastasis staging system of the Union for International Cancer Control (7 th edition).  The stage assessment was based on computed tomography scans of the thorax and upper abdomen, magnetic resonance imaging scans of the brain, and emission computed tomography scan of bone.
All the patients received more than one cycle of standardized combination chemotherapy and were reassessed after two cycles of chemotherapy. The rules established by the Response Evaluation Criteria in Solid Tumors (RECIST) were used to evaluate tumor responses to treatment.  Patients were considered to have achieved an objective response (OR) if they had either a complete response or a partial response according to RECIST. Stable disease or progressive disease was considered a non-response (NR). The patient OS was calculated and patients were censored at the last follow-up.
The Medical Ethics Committee of Jinling Hospital approved this study.
Measurement of serum CEA levels
The serum levels of CEA were measured with Beckman Coulter's DxI 800 Immunoassay System (USA) before and after chemotherapy. The upper limit of the CEA normal value was 9.7 ng/mL.
All statistical analyses were performed with SPSS 17.0 software. The Chi-squared (χ2 ) test was used to evaluate the relatedness between serum CEA and baseline clinical characteristics, and the patient's OR after two cycles of chemotherapy. Logistical regression was used to evaluate clinical characteristics as risk factors for the basal level of serum CEA. The statistical significance of the difference between serum CEA levels at baseline and post-chemotherapy was examined using the paired-samples t-test. A receiver operator characteristic (ROC) curve was constructed to reveal an association between the proportion of decline in CEA after two cycles of chemotherapy and the OR. The Kaplan-Meier method was used calculate the survival probability estimates, and the log-rank test was performed to assess heterogeneity in survival within each prognostic factor. A Cox regression model was used for the independent prognostic risk factor analyses. A value of P < 0.05 was considered significant.
| > Results|| |
Associations between basal levels of serum CEA and the clinical characteristics of NSCLC patients
The mean age of the 689 enrolled patients was 64 years (range: 8-89 years). The basal levels of serum CEA were significantly associated with gender (male: χ2 = 11.939; P = 0.001), smoking (χ2 = 20.130; P = 0.000), histology (χ2 = 66.718; P = 0.000), stage (χ2 = 69.298; P = 0.000), pleural effusion (χ2 = 8.698; P = 0.003), and the number of distant metastatic organs (χ2 = 72.785; P = 0.000) [Table 1]. Higher basal levels of CEA were associated with the clinicopathological features: Being male, smoking less than 20 years, histology showing adenocarcinoma and adenosquamous carcinoma, pleural effusion, or more than one distant metastatic organ. No associations were found between the basal levels of CEA and age (χ2 = 2.238; P = 0.135), metastatic locations (χ2 = 5.592; P = 0.348), or original tumor location (χ2 = 0.226; P = 0.635).
The risk factors for high basal levels of serum CEA among the clinicopathological characteristics of NSCLC patients were determined by logistic regression analysis. We showed that only the adenocarcinoma and adenosquamous histological subtypes appeared to be independent risk factors for the high baseline levels of serum CEA [for adenocarcinoma, P = 0.014, hazard ratio (HR) =1.950, 95% CI = 1.145-3.321; for adenosquamous carcinoma, P = 0.017, HR = 4.497, 95% CI = 1.315-15.380). Other characteristics, namely gender, smoking, clinical stages, pleural effusion, and metastasis, had no such correlations with baseline serum CEA levels.
|Table 1: The baseline levels of serum CEA and clinical characteristics of NSCLC patients|
Click here to view
Serum CEA and ORs in patients with NSCLC
Among the 689 NSCLC patients in this study, 276 received more than two cycles of chemotherapy and the overall OR rate (those patients achieving either partial or complete response) was 34.97%. In the 166 patients whose basal serum CEA level was <9.7 ng/mL, 59 patients (35.54%) experienced an OR. In those whose basal serum CEA was ≥9.7 ng/mL, the OR was observed in 41 of 110 patients (37.27%). Thus, no significant association was observed between the baseline serum CEA levels and OR to two cycles of chemotherapy.
Commonly, the CEA levels in large numbers of patients have a normal distribution, and the aim of the study was to evaluate the changes in CEA levels in patients before and after chemotherapy. Thus, the paired samples t-test was used to assess the significance of differences in serum CEA levels between baseline and after two cycles of chemotherapy [Table 2]. There was no significant difference in the mean serum CEA levels between baseline and after two cycles of chemotherapy when considering all patients taken together (t = 0.092, P = 0.927) or only those patients in whom there was NR (t = -0.987, P = 0.325). In contrast, the patients with an OR showed a significant reduction in serum CEA after two cycles of chemotherapy (t = 2.196, P = 0.031). This result suggests that a decrease in serum CEA after chemotherapy may predict an OR in NSCLC patients.
ROC curves were constructed to determine the relatedness between CEA decline after chemotherapy and OR. The area under the curve was 0.615 (95% CI = 0.535-0.695), and a 5.28% reduction in CEA level was an appropriate cut-off value for predicting an OR to chemotherapy, with a sensitivity of 61.3% and a specificity of 62.4%. At this cut-off value, the sum of the sensitivity and specificity was the highest [Figure 1].
|Figure 1: The ROC curves for the proportion of decline in CEA and OR. Serum CEA levels were measured using a Beckman Coulter immunoassay system. The ROC curve was calculated to assess the association between sensitivity and specificity of serum CEA and the OR to chemotherapy in NSCLC patients. The area under the curve is 0.615 (95% CI = 0.535-0.695), and a 5.28% reduction of CEA levels was an appropriate cut-off value in predicting an OR to chemotherapy, with a sensitivity of 61.3% and a specificity of 62.4%|
Click here to view
Effect of serum CEA, treatment, and metastasis on OS
OS of 253 patients was obtained, with a median OS of 473 days. According to results of the Kaplan-Meier survival analysis, the OS of the 152 patients with a baseline serum CEA <9.7 ng/mL was longer than that of the 101 patients whose baseline was ≥9.7 ng/mL [Figure 2]. However, the log-rank test unveiled no strongly significant difference between these two groups (P = 0.079). Interestingly, the OS in 97 patients with >4 chemotherapy cycles was significantly longer than that of 156 patients with fewer chemotherapy cycles (log-rank test, P = 0.030) [Figure 3]. The independent risk factors for OS in these NSCLC patients were evaluated using Cox regression analysis [Table 3]. We found that the baseline serum CEA was not a prognostic factor for OS in NSCLC patients (P = 0.079, HR = 1.418, 95% CI = 0.960-2.094).
|Figure 2: The correlation between OS of patients and the basal serum CEA level. Two hundred fifty-three patients were followed up for OR. The median OS was 473 days. The Kaplan-Meier survival analysis and log-rank test were performed to examine the correlation between OS of patients and the basal serum CEA level. The upper limit of the CEA normal value was 9.7 ng/mL. The blue curve indicates the baseline serum CEA <9.7 ng/mL, while the green curve indicates the baseline serum CEA >9.7 ng/mL. Log-rank test unveiled no strongly significant difference between these two (P = 0.079)|
Click here to view
|Figure 3: The association between OS of patients and the number of chemotherapy cycles. Two hundred fifty-three patients were followed up for OR. The median OS was 473 days. The Kaplan-Meier survival analysis and log-rank test were performed to examine the correlation between OS of patients and the number of chemotherapy cycles. The blue curve indicates <4 chemotherapy cycles, while the green curve indicates ≥4 chemotherapy cycles. The OS in 97 patients with ≥4 chemotherapy cycles was significantly longer than for 156 patients with <4 chemotherapy cycles (log-rank test, P = 0.030)|
Click here to view
|Table 3: The Cox regression analysis of independent risk factors for OS in NSCLC patients|
Click here to view
| > Discussion|| |
CEA is an important and well-known tumor biomarker for a wide range of malignancies including NSCLC. In clinical practice, the assessment of serum CEA is feasible and convenient. The present study investigated correlations between serum CEA and OR to chemotherapy and OS in patients with NSCLC. Our results suggest that while the baseline level of CEA was not a prognostic factor, the post-treatment reduction of CEA level could predict the OR in patients with NSCLC.
Our analysis of serum CEA in NSCLC patients as an indicator of OR to chemotherapy and OS included the construction of an ROC curve to determine the appropriateness of the associations and the Kaplan-Meier method for estimating a population survival curve. The ROC curve analysis showed that a 5.28% reduction in CEA level was an appropriate cut-off value for predicting an OR to chemotherapy, with a sensitivity of 61.3% and a specificity of 62.4%. Consistent with this, Ardizzoni et al.  found that the CEA responses (20% reduction over baseline level; assessed after the second course of chemotherapy) were 38% and statistically significant correlations were observed between CEA and OR (P = 0.01). The Kaplan-Meier survival analysis in the present study demonstrated that the OS of the patients with a baseline serum CEA <9.7 ng/mL was longer than that of the patients whose baseline was ≥9.7 ng/mL. However, the log-rank test unveiled no strongly significant difference between these two groups. More studies are needed to confirm this.
The present finding that a reduction in CEA after treatment could predict the OR may be because serum CEA is secreted by tumor cells and therefore correlates with the number of tumor cells. Other studies have shown similar results. , Through the ROC curve analysis, we were able to calculate a more reliable cut-off value for predicting the OR to chemotherapy. However, one study reported finding no correlation between CEA expression (either prior to or after treatment) and the response to chemotherapy; in patients with IIIA NSCLC; specimens after chemotherapy displayed a greater intensity of staining and a higher percentage of CEA-positive cells.  The causes of the discrepancy between this study and ours are unclear. It may be due to differences in statistical methods, the methods used to detect CEA, or both, and needs to be further determined. It may be that changes in serum CEA directly reflect the effect of chemotherapy on tumor cells.
We used Kaplan-Meier survival analysis to assess the association between serum CEA and OS. OS data were obtained in only 253 of the 689 patients because many patients reside in villages far from our hospital and could not be reached for follow-up. However, our results suggest a correlative trend between OS and baseline serum CEA levels, although no statistical significance was found. The role that baseline serum CEA levels could play in predicting survival of NSCLC patients has been reported in several other studies. ,, We reasoned that the differences found among studies might be due to differences in patient populations and the adjusted CEA cut-off value used. This needs to explored further. The measure of baseline serum CEA might be point-based rather than a dynamic process of changes. It may be that a tumor marker index rather than a single serum tumor marker such as CEA would be more valuable and useful for predicting the prognosis of NSCLC patients. ,,
The present study enrolled a larger number of NSCLC patients compared to others ,,,, that sought to determine the feasibility of CEA as a predictor of OR and OS in cancer patients. The serum CEA cut-off value that could predict OR to chemotherapy with the greatest sensitivity and specificity was defined by constructing an ROC curve for NSCLC patients undergoing chemotherapy. Therefore, our study should be more reliable and applicable in the clinical setting.
RECIST is well accepted for the evaluation of tumor response in clinical practice. However, these criteria cannot be applied in cases with no measurable lesions or when there is a lack of early response after the initial treatment. Our findings suggest that serum CEA is a potential indicator for evaluating the OR of NSCLC patients, and may enhance or supplement RECIST.
Because serum CEA was not measured in many of our patients after the first cycle of chemotherapy, we cannot conclude that serum CEA might be more valuable than RECIST for early detection of lung cancer and predicting response to treatment. Although the baseline serum CEA level was not a prognostic factor of OS in this study, combining CEA with several cancer markers such as CYFRA21-1 and CA-125 may better predict survival in NSCLC patients. 
| > Conclusion|| |
Our study shows that while the baseline level of serum CEA is not a prognostic factor, the reduction of serum CEA level after chemotherapy can predict therapeutic efficacy in NSCLC patients. In addition, the number of chemotherapy cycles was an independent protective factor and the number of distant metastatic organs was an independent risk factor for OS in NSCLC patients.
| > References|| |
|1.||Molina JR, Yang P, Cassivi SD, Schild SE, Adjei AA. Non-small cell lung cancer: Epidemiology, risk factors, treatment, and survivorship. Mayo Clin Proc 2008;83:584-94. |
|2.||Wakelee HA, Bernardo P, Johnson DH, Schiller JH. Changes in the natural history of nonsmall cell lung cancer (NSCLC)--comparison of outcomes and characteristics in patients with advanced NSCLC entered in Eastern Cooperative Oncology Group trials before and after 1990. Cancer 2006;106:2208-17. |
|3.||Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin 2009;59:225-49. |
|4.||Molina R, Auge JM, Farrus B, Zanón G, Pahisa J, Muñoz M, et al. Prospective evaluation of carcinoembryonic antigen (CEA) and carbohydrate antigen 15.3 (CA 15.3) in patients with primary locoregional breast cancer. Clin Chem 2010;56:1148-57. |
|5.||Louhimo J, Carpelan-Holmström M, Alfthan H, Stenman UH, Järvinen HJ, Haglund C. Serum HCG beta, CA 72-4 and CEA are independent prognostic factors in colorectal cancer. Int J Cancer 2002;101:545-8. |
|6.||Park YA, Sohn SK, Seong J, Baik SH, Lee KY, Kim NK, et al. Serum CEA as a predictor for the response to preoperative chemoradiation in rectal cancer. J Surg Oncol 2006;93:145-50. |
|7.||Basbug M, Arikanoglu Z, Bulbuller N, Cetinkaya Z, Aygen E, Akbulut S, et al. Prognostic value of preoperative CEA and CA 19-9 levels in patients with colorectal cancer. Hepatogastroenterology 2011;58:400-5. |
|8.||Schneider J. Tumor markers in detection of lung cancer. Adv Clin Chem 2006;42:1-41. |
|9.||Hackbarth JS, Murata K, Reilly WM, Algeciras-Schimnich A. Performance of CEA and CA19-9 in identifying pleural effusions caused by specific malignancies. Clin Biochem 2010;43:1051-5. |
|10.||Hanagiri T, Sugaya M, Takenaka M, Oka S, Baba T, Shigematsu Y, et al. Preoperative CYFRA 21-1 and CEA as prognostic factors in patients with stage I non-small cell lung cancer. Lung Cancer 2011;74:112-7. |
|11.||Yang X, Wang D, Yang Z, Qing Y, Zhang Z, Wang G, et al. CEA is an independent prognostic indicator that is associated with reduced survival and liver metastases in SCLC. Cell Biochem Biophys 2011;59:113-9. |
|12.||Goldstraw P, Groome P. Lung. In: Sobin LH, Gospodarowicz MK, Wittekind C, editors. TNM Classification of Malignant Tumors. International Union Against Cancer (UICC). 7 th ed.: Wiley-blackwell; 2009. p. 138-46. |
|13.||Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228-47. |
|14.||Jin B, Huang AM, Zhong RB, Han BH. The value of tumor markers in evaluating chemotherapy response and prognosis in Chinese patients with advanced non-small cell lung cancer. Chemotherapy 2010;56:417-23. |
|15.||Ardizzoni A, Cafferata MA, Tiseo M, Filiberti R, Marroni P, Grossi F, et al. Decline in serum carcinoembryonic antigen and cytokeratin 19 fragment during chemotherapy predicts objective response and survival in patients with advanced nonsmall cell lung cancer. Cancer 2006;107:2842-9. |
|16.||Graziano SL, Kern JA, Herndon JE, Tatum A, Brisson ML, Memoli V, et al. Analysis of neuroendocrine markers, HER2 and CEA before and after chemotherapy in patients with stage IIIA non-small cell lung cancer: A Cancer and Leukemia Group B study. Lung Cancer 1998;21:203-11. |
|17.||Arrieta O, Saavedra-Perez D, Kuri R, Aviles-Salas A, Martinez L, Mendoza-Posada D, et al. Brain metastasis development and poor survival associated with carcinoembryonic antigen (CEA) level in advanced non-small cell lung cancer: A prospective analysis. BMC Cancer 2009;9:119. |
|18.||Kulpa J, Wójcik E, Reinfuss M, Ko³odziejski L. Carcinoembryonic antigen, squamous cell carcinoma antigen, CYFRA 21-1, and neuron-specific enolase in squamous cell lung cancer patients. Clin Chem 2002;48:1931-7. |
|19.||Cedrés S, Nuñez I, Longo M, Martinez P, Checa E, Torrejón D, et al. Serum tumor markers CEA, CYFRA21-1, and CA-125 are associated with worse prognosis in advanced non-small-cell lung cancer (NSCLC). Clin Lung Cancer 2011;12:172-9. |
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3]