Tumor markers CYFRA21-1, CEA, NSE, CA125, pro-GRP and SCC are routinely used for lung cancer. However, there has been no systematic evaluation of these markers in the same cohort. The aim of this study was to evaluate the diagnostic and therapeutic monitoring value of these markers.
The levels of 6 serum tumor markers were measured in 392 patients, including 308 patients with non-small cell lung cancer (NSCLC) and 84 with small cell lung cancer (SCLC), and 116 patients with benign lung diseases and 144 healthy controls. 34 patients were followed up after operation and chemotherapy. Multiple logistic models and receiver operating characteristic (ROC) curves were used to evaluate their diagnostic value.
CEA, NSE, CA125 and pro-GRP in SCLC, and CYFRA21-1 as well as CEA in NSCLC, were higher than those in control groups. The level of CEA and CA125 were related to the clinical stages of NSCLC. Pro-GRP was significantly increased in extensive disease (ED) compared with limited disease (LD) in SCLC. CYFRA21-1 was reduced after the third and fifth treatment cycle respectively in patients who undergoing operation and without operation. NSE and pro-GRP were reduced significantly after the second and third treatment cycles, respectively.
CEA, NSE, CA125 and pro-GRP could serve as biomarkers for SCLC, and CEA and CYFRA21-1 could serve as biomarkers for NSCLC. Pro-GRP, CA125 and CEA were related to the clinical stages of lung cancer. CYFRA21-1, NSE and pro-GRP could be used for monitoring the effect of chemotherapy.
Int J Biol Markers 2016; 31(1): e80 - e87
Article Type: ORIGINAL RESEARCH ARTICLE
AuthorsLishui Wang, Ding Wang, Guixi Zheng, Yongmei Yang, Lutao Du, Zhaogang Dong, Xin Zhang, Chuanxin Wang
- • Received on 30/01/2015
- • Accepted on 07/10/2015
- • Available online on 06/11/2015
- • Published online on 28/02/2016
This article is available as full text PDF.
Lung cancer is one of the most common respiratory tract malignancies, with high morbidity and mortality worldwide (1, 2). Patients with lung cancer usually do not exhibit noticeable symptoms, particularly in the early stage. More than 60% of the patients with lung cancer are diagnosed at an advanced stage and miss the optimal opportunity for surgery (3). The corresponding 5-year survival rates decrease significantly, from 60%-80% in stage I to less than 10% in advanced stage patients (4). Therefore, the current high mortality would be remarkably decreased if lung cancer were detected at an early stage. Among the available lung cancer testing methods, fiberoptic bronchoscope is recommended as the most reliable tool for high-risk patients. However, its wide application is hampered because it cannot be readily repeated, and it causes physical harm to patients. Conventional imaging examinations such as chest radiographs, computed tomography and cytological examination are not sensitive enough for effective early diagnosis (5).
Serum tumor markers have been shown to be excellent candidates and are widely used by clinicians in the early diagnosis of cancer. Several studies have also demonstrated the significance of serum tumor markers in the diagnosis, prognosis and follow-up of lung cancer (6-7-8-9). National Academy of Clinical Biochemistry (NACB) Laboratory Medicine Practice Guidelines recommend carcinoembryonic antigen (CEA), squamous carcinoma antigen (SCC), neuron specific enolase (NSE), cytokeratin fragment (CYFRA21-1), carbohydrate antigen 125 (CA125) and pro-gastrin-releasing peptide (pro-GRP) as routine markers of lung cancer. However, there is no systematic evaluation of these 6 markers in the same cohort.
In the present study, we measured the levels of these 6 tumor markers in patients with lung cancer, patients with benign lung diseases and healthy controls, and analyzed their diagnostic value in different types of lung cancer. The relationship between makers and clinical stage of lung cancer was also studied. Meanwhile, we collected the serum of patients before and after operation or chemotherapy to evaluate the value of tumor markers in the therapeutic monitoring of lung cancer.
Materials and methods
Patients and control subjects
Written informed consent was obtained from each participant for the use of venous blood samples. All of the lung cancer and benign lung disease patients were recruited from the Department of General Surgery and the Department of Gastroenterology and Chemotherapy, Qilu Hospital of Shandong University, between May 2010 and September 2013. The healthy controls were enrolled from the Department of Physical Examination Center, Qilu Hospital of Shandong University.
All lung cancer patients were confirmed by histopathology or cytology. Tumors were staged according to the Union for International Cancer Control (UICC) 2009 lung cancer classification. All follow-up patients were confirmed as in remission by clinicians. Preoperative blood samples were collected before any therapeutic procedures such as surgery, chemotherapy or radiotherapy. Postoperative blood samples were collected 7 days after each operation and after subsequent chemotherapy. Patients with lung benign diseases were diagnosed with pulmonary tuberculosis, pleural effusion, pneumonia, phthisis and bronchiectasis by standard diagnostic methods or histological examination. Age- and sex-matched healthy controls were recruited from a large pool of healthy individuals seeking a routine health check-up. Detailed patient characteristics are described in Supplementary Table I (Available online at
Sample collection and measurement
Briefly, 5 mL of venous blood was collected from each participant. The whole blood was separated into serum and cellular fractions within 2 hours by centrifugation at 4,000 rpm for 10 minutes. The whole process was strictly controlled to avoid hemolysis. The obtained serum was stored at -80°C before further analysis.
The CEA, NSE, CA125, CYFRA21-1 and pro-GRP concentrations were measured using electrochemiluminescence immunoassay method on a Roche Cobas e601 analyzer. The SCC concentration was measured using chemiluminescence immunoassay method on an ARCHITECT i2000SR analyzer.
Statistical analysis was performed using SPSS 17.0 (SPSS Inc., Chicago, IL, USA). The data were described as median, minimum, 25th, 75th, 95th percentiles and maximum. The Mann-Whitney test was used to compare the difference in the levels of serum tumor markers between the 2 groups. The signed-rank test was used to compare the difference in the levels of tumor markers preoperatively, postoperatively and after chemotherapy. Multiple logistic regression analysis was used to select the optimal combination of tumor markers. Receiver operating characteristic (ROC) curves was constructed, and the area under the ROC curve (AUC) was used to evaluate the diagnostic performance of tumor markers. MedCalc software was used to generate the ROC curves.
Comparison of tumor markers in lung cancer, benign lung disease and healthy control group
Expressions of CYFRA21-1, CEA, NSE, CA125 and pro-GRP were significantly higher in lung cancer than in benign lung disease patient and healthy control groups (p<0.05) (
Concentration of tumors markers in different groups
|Healthy controls||Patients with benign lung disease||Patients with lung cancer||Patients with SCLC||Patients with NSCLC|
|NSCLC = non-small cell lung cancer; SCLC = small cell lung cancer.|
|*p<0.05, vs. healthy control group.|
|†p<0.05, vs. lung benign disease group.|
|‡p<0.05, vs. NSCLC group.|
The lung cancer group was classified into those with non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC). The results indicated that CEA, NSE, CA125 and pro-GRP levels in the SCLC group, as well as CYFRA21-1 and CEA in the NSCLC group, were higher than those in the benign lung disease and healthy control groups (p<0.05). The expression of SCC (median 0.8, range 0.1-57.3) was higher in the squamous cell lung cancer group than in the benign lung disease and healthy control groups (p = 0.001 and p = 0.024, respectively). The diagnostic value of SCC in the squamous cell lung cancer group is shown in
Diagnosis value of tumor markers indicated by ROC curves
|Tumor markers||AUC||95% CI||Sensitivity (%)||Cutoff values|
|AUC = area under the curve; CI = confidence interval; NSCLC = non-small cell lung cancer; ROC = receiver operating characteristic; SCLC = small cell lung cancer.|
|Sensitivities and cutoff values of all tumor markers are for 95% specificity.|
|Lung cancer group||CYFRA21-1||0.690||0.621-0.753||37.4||3.83 ng/mL|
|SCLC group||CEA||0.682||0.604-0.754||28.1||5.12 ng/mL|
|NSCLC group||CYFRA21-1||0.779||0.715-0.835||40.8||3.83 ng/mL|
|Squamous cell lung cancer group||SCC||0.642||0.570-0.710||33.3||1.3 ng/mL|
Diagnostic efficiency of tumor markers in different types of lung cancer
The diagnostic values of CYFRA21-1, CEA, NSE, CA125 and pro-GRP in lung cancer, NSCLC and SCLC groups were evaluated using the ROC method. The AUC, sensitivity (at 95% specificity) and cutoff value of the 6 tumor makers above are shown in
Optimal combination of tumor markers in different types of lung cancer
The stepwise backward method of logistic regression analysis was used to select the optimal combination for NSCLC and SCLC, and to establish the diagnostic panel. NSE, CA125 and pro-GRP were considered as the optimal combination markers for SCLC patients. Meanwhile, CYFRA21-1 and CEA were the optimal combination markers for NSCLC patients.
The predicted probability of diagnosis for SCLC was calculated using the following equation: logit(P) = -10.483 + 0.184*NSE + 0.074*CA125 + 0.107 * pro-GRP. ROC analysis indicated that AUC and sensitivity (at 95% specificity) of the SCLC panel were 0.979 and 91.4%, respectively. The diagnostic value of the SCLC panel was superior to that for CA125 (p<0.05), but there was no statistically significant difference between NSE and pro-GRP (p>0.05).
The predicted probability of diagnosis for NSCLC was calculated using the following equation: logit(P) = -2.271 + 0.475*CYFRA21-1 + 0.288*CEA. The AUC and sensitivity (at 95% specificity) of the NSCLC panel were 0.797 and 43.6%, respectively. The diagnostic value of the NSCLC panel was superior to that for CEA (p<0.05), but there was no statistically significant difference between the panel and CYFRA21-1 (p>0.05).
Differential diagnosis value of tumor markers in NSCLC and SCLC
The expressions of the 6 tumor markers in the NSCLC and SCLC groups were compared. The results showed that the levels of NSE and pro-GRP were significantly higher in the SCLC group than in the NSCLC group (
Differential diagnosis value of NSE, pro-GRP and a combination of the 2 markers in non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC).
Relationship between tumor marker and clinical stage of lung cancer
Expression of CEA was significantly higher in stages II-IV than in stage I (p = 0.003, p = 0.005 and p = 0.009, respectively). The level of CA125 was higher in stage IV than in stage I (p<0.001), stage II (p = 0.018) and stage III (p = 0.037); and in stage III, it was higher than in stage I (p<0.001). The expression of pro-GRP was higher in extensive disease (ED) for SCLC than in limited disease (LD) (p = 0.009). However, there were no statistically significant differences in the levels of CYFRA21-1 and NSE among the different stages of lung cancer (p>0.05) (
Concentration of tumor markers in different stages of lung cancer
|Stage I||Stage II||Stage III||Stage IV||LD||ED|
|ED = extensive disease; LD = limited disease; NSCLC = non-small cell lung cancer; SCLC = small cell lung cancer.|
|*p<0.05, vs. stage I.|
|†p<0.05, vs. stage II.|
|‡p<0.05, vs. stage III.|
|§p<0.05, vs. LD.|
Therapeutic monitoring value of tumor markers in lung cancer
The patients followed up were divided into 2 groups: those undergoing an operation before chemotherapy and those not undergoing an operation before chemotherapy. The monitoring values of tumor makers were assessed in the 2 groups (
The concentration of tumor markers in pretreatment, 7 days after operation and 1, 2 and 3 cycles after chemotherapy in lung cancer patients. CYFRA21-1 was significantly decreased after the third cycle of chemotherapy in patients who experienced operation before chemotherapy (p<0.05). CYFRA21-1 median: 3.88, 2.33, 2.23, 2.01 and 1.78; CEA median: 5.12, 2.67, 2.74, 2.41 and 2.21; CA125 median: 18.27, 24.36, 18.73, 16.30 and 17.83; NSE median: 13.94, 13.40, 11.21, 12.32 and 10.89; pro-GRP median: 14.36, 14.28, 11.36, 11.78 and 10.86, respectively.
The concentration of tumor markers in pretreatment and 1, 2, 3, 4, 5 and 6 cycles of chemotherapy in lung cancer patients. CYFRA21-1 was significantly decreased after the fifth cycle of chemotherapy in patients without operation (p<0.05). NSE and pro-GRP were remarkably decreased after the second and third cycle of chemotherapy, respectively (p<0.05). CYFRA21-1 median: 3.61, 2.59, 2.14, 2.44, 2.11, 1.71 and 1.23; CEA median: 5.39, 4.29, 3.32, 2.48, 2.18, 1.88 and 1.65; CA125 median: 29.70, 18.42, 14.03, 13.20, 12.98, 12.89 and 11.87; NSE median: 43.98, 13.12, 13.09, 13.11, 12.33, 11.45 and 10.35; pro-GRP median: 353.2, 167.5, 60.37, 46.78, 24.87, 23.67 and 18.56, respectively.
Lung cancer has a high mortality rate worldwide (1). Early detection and timely treatment can significantly improve the prognosis and prolong survival times for patients with lung cancer. Detection of tumor markers has proven to be useful for early diagnosis of lung cancer, and these can be indicators for monitoring treatment effect. Currently, tumor markers, including CYFRA21-1, CEA, NSE, CA125, SCC and pro-GRP, are routinely used in clinical laboratories. However, a comprehensive evaluation of these 6 tumor markers in the diagnosis and therapeutic monitoring of lung cancer is lacking.
In the present study, 6 serum markers were evaluated individually and in combination, in lung cancer, benign lung disease and healthy controls. CEA is mainly used in the diagnosis and therapeutic monitoring of gastrointestinal cancer. Several studies have indicated increased CEA levels in lung adenocarcinoma and advanced bronchogenic cancer of various histological types (10, 11). The sensitivity of CEA is around 35% to 77% (12). Our study showed that the expression of CEA was significantly higher in both NSCLC and SCLC groups than in the benign lung disease and healthy control groups, and that it could be a useful marker for both NSCLC and SCLC patients. The sensitivity (at 95% specificity) of CEA in diagnosing NSCLC and SCLC was 30.3% and 28.1%, respectively. Furthermore, our results indicated that CEA was correlated with TNM stage. In stages II-IV, the level of CEA was higher than in stage I.
CYFRA21-1 is considered a sensitive tumor marker for lung cancer, especially for NSCLC. The sensitivity of CYFRA21-1 ranges from 55% to 70% in lung cancer (8, 13). Our results indicated that the level of CYFRA21-1 was higher in the NSCLC group than in the benign lung disease and healthy control groups. Meanwhile, we found that CYFRA21-1 and CEA were the optimal combination markers for NSCLC patients, which was similar to what was reported by Molina et al (14). The sensitivity of CYFRA21-1 for lung cancer and NSCLC was 37.4% and 40.8%, respectively. There was no correlation between CYFRA21-1 and TNM stage in the present study.
SCC is elevated in many types of squamous cell carcinoma, including uterine cervix, bronchus, nasopharynx and lung cancer. Currently, the results of SCC in serum of lung cancer patients have been controversial. Several studies have reported that SCC is valuable for the early diagnosis of lung cancer (15, 16). Other researchers have found higher SCC levels in patients with lung squamous cell carcinoma than patients with benign lung diseases and healthy controls (5). Our result was consistent with the latter. Therefore, SCC may not be a valuable marker for lung cancer diagnosis, but it may be useful for distinguishing between lung squamous cell carcinoma and other histological subtypes.
CA125 is a glycoprotein expressed in fetal epithelial cells of the body cavity. It has mainly been used as a specific marker for ovarian cancer. Recently, CA125 has gradually been introduced in lung cancer diagnosis, treatment and prognosis. There has been no report on the correlation between CA125 and tumor size, stage and histological type. It can be used as an independent prognostic indicator for patients with lung cancer (17). We found that CA125 increased significantly in the SCLC group, but there was no significant difference in the level of CA125 between the NSCLC group and the lung benign disease and healthy control groups. Furthermore, CA125 levels were associated with TNM stage. The level of CA125 was higher in stage IV than in stages I-III, and was also higher in stage III than in stage I of NSCLC.
NSE and pro-GRP are useful markers with high diagnostic value for SCLC patients (18, 19). In the present study, the expression levels of NSE and pro-GRP were higher in the SCLC group than in the lung benign disease and healthy control groups. The sensitivity of NSE and pro-GRP was 78.1% and 90.6%, respectively. Pro-GRP was correlated with the stage of SCLC. The expression of pro-GRP was higher in ED than in LD. However, no difference was observed between LD and ED in NSE level. Our results also showed that NSE and pro-GRP could be used as markers for distinguishing SCLC and NSCLC with high sensitivity and specificity.
The usual method for assessing operation or chemotherapy response in cancer patients is through the change of tumor size as measured by imaging examination. However, these measurements are often inconsistent with clinical status. Several studies have demonstrated that serum markers play an important role in therapeutic monitoring. Pang et al (8) have demonstrated that a significant reduction in CYFRA21-1 level is associated with a positive response after 2 cycles of chemotherapy in NSCLC patients. CYFRA21-1 is a useful surrogate marker for predicting responses to chemotherapy. A retrospective report by Ardizzoni et al (20) showed that a significant decrease of serum CEA level after 2 cycles of chemotherapy predicts the patient’s response as accurately as an ROC curve analysis of 0.65. Ebert et al (21) indicated that after curative surgery of NSCLC, only CYFRA 21-1 levels dropped to within the normal range within 1 week. During the monitoring of response to chemotherapy and/or radiotherapy, concordant results were obtained in 59.4% of the cases for CYFRA21-1 (TPA-M: 63.3%, TPS: 65.5%, CEA: 54.8%, NSE: 68.9%). The most discordant results were obtained in tumor remission due to an insufficient decrease in the markers. Progressive disease was most effectively indicated by CYFRA 21-1 in NSCLC (60%) and by NSE in SCLC (70.0%). Our results indicated that CYFRA21-1 was significantly decreased after 3 cycles in patients who underwent an operation before chemotherapy and 5 cycles of chemotherapy in patients without an operation. NSE and pro-GRP were remarkably decreased after 2 and 3 cycles of chemotherapy, respectively. There were no significant differences in the levels of CEA and CA125 before and after treatment. Therefore, CYFRA21-1, NSE and pro-GRP were sensitive markers for therapeutic monitoring of lung cancer.
- Wang, Lishui [PubMed] [Google Scholar]
- Wang, Ding [PubMed] [Google Scholar]
- Zheng, Guixi [PubMed] [Google Scholar]
- Yang, Yongmei [PubMed] [Google Scholar]
- Du, Lutao [PubMed] [Google Scholar]
- Dong, Zhaogang [PubMed] [Google Scholar]
- Zhang, Xin [PubMed] [Google Scholar]
- Wang, Chuanxin [PubMed] [Google Scholar] , * Corresponding Author (email@example.com)
Department of Clinical Laboratory, Qilu Hospital of Shandong University, Jinan - PR China
Article usage statistics
This article has supplementary materials available to download.
Download any of the following attachments: