| Research Background and ObjectiveLung cancer had the highest incidence and mortality in all of the malignant, replaced liver cancer before2008. The mortality rate of lung cancer in our country increased by464.84%over the past30years. More than two-thirds patients are diagnosed at a relatively late stage. Besides many patients of lung cancer relapse even received the best treatments. Conventional diagnostic techniques, including chest x-ray, bronchoscopy, and sputum cytology, demonstrated a limited ability to detect early lung cancer in old randomized studies. Recent trials with low-does spiral computed tomography(CT) have demonstrated that this technique can detect smaller tumors, but it remains unclear whether annual CT monitoring can reduce lung cancer mortality. It is so important in the diagnosis and treatment of lung cancer that the development of serological detection of lung cancer resulted from its mangy advantages, such as its convenient, non-invasive and low cost.Traditional lung cancer tumor markers, such as carcinoembryonic antigen (CEA), cytokeratin19fragment21-1(CYFRA21-1), neuron-specific enolase (NSE) which have greater utility in the diagnosis and treatment of lung cancer. Besides, as we know, other common tumor markers such as CA125, CA153and CA199whose value are lower than CEA, CYFRA21-1and NSE. Why is it? It may be that because their sensitivity, specificity, positive rate were very low. Even so, how could we use these six tumor markers effectively in the field of lung cancer? This is the first motivation that we intend to do this research.After analysis of the literature, it was considered serum amyloid A (SAA) also likely to become tumor marker in proteomics. In the1970s and1980s, scholars have found that serum amyloid A not only related to amyloidosis but also related to the malignancies. At that time, some scholars have found that serum levels of SAA in carcinoma were much higher than in healthy people, especially in patients with lung cancer whose SAA levels were highest among different types of cancer. But the method of testing SAA in majority of these experiments were radioimmunoassay(RIA) or radiation-immunodiffusion(RID). And in the1980s, with the development of proteomics, especially surface-enhanced laser desorption ionization time-of-flight mass spectrometry(SELDI-TOF-MS) techniques and matrix-assisted laser desorption ionization time-of-flight mass spectrometry(MALDI-TOF-MS) technology development, making the SAA in the field of cancer research revitalize. Song-Wei Dai made a decision tree model using mass spectrometry techniques which proper division rate of lung cancers and healthy people95%and81%, respectively. However, the human body has a certain degree of radiation damage because that RIA or RID method contains radioactive elements. And there are also many deficiencies about mass spectrometry techniques such as complicated operation, can not be directly measured potential markers of identity, measuring instruments expensive. ELISA detection technology is mature, easy to operate, inexpensive, no radiation damage.Are the conclusions of.using this method to detect serum SAA levels of lung cancers and healthy people similar to those conclusions in proteomics reported in the literature? It is rarely reported. This is the second point that we intend to do this research.If there are significant differences in lung cancer patients and healthy people SAA levels, then whether there are differences between SAA levels in patients with lung cancer and benign lung disease, can SAA be used to identify lung cancer and benign lung disease? Summarizes the existing literature on the answer to this question at present, there are still different views. One view is that SAA as one of the acute phase proteins, it also elevated in benign lung diseases, and SAA levels are higher in benign lung disease than lung cancers; another view is that, SAA levels of benign lung disease are far less than the lung cancers. And the U.S. National Cancer Institute Web site in2011lists out existing or potential lung cancer tumor marker of about180targets, in so many indicators did not mention serum amyloid A (SAA). Therefore, the value of this indicator in the identification of lung cancer and benign lung disease, even in diagnosis of clinical stage, histological type, distant metastasis of lung cancer is valuable, this is the third point why we start to this study.At the same time, no matter using RIA, RID method to test SAA, or using time of flight mass spectrometry to test SAA, we found no one article analysised the clinical significance of SAA compared to the traditional tumor markers such as CEA, CYFRA21-1and NSE. Therefore, is it that the SAA is better than the above mentioned tumor markes?. This is the fourth starting point of our study.Based on the above four reasons, preliminary experiments was done before this study, we firstly retrospective analysis the relation between existing lung cancer tumor markers (CEA, CYFRA21-1, NSE, CA125, CA153and CA199) and lung cancer, then select the good tumor markers which are high-value in lung cancer as reference index, finally analysis the clinical significance of SAA and the reference index in lung cancer.Part1Preliminary ExperimentsClinical DataThis research study contains a total of212cases. There were64healthy people, 73patients of benign lung disease group,75cases of lung cancer including59cases of NSCLC,16cases of SCLC. Lung cancer patients with pathologically confirmed, did not receive any therapy. Staging of lung cancer according to International Union Against Cancer TNM staging system developed in2009. The age, sex and smoking status distribution in three study groups were nearly the same.MethodDetections:CEA by enzyme immunoassay chemiluminescence, CYFRA21-1, NSE, CA125, CA153and CA199were using electrochemiluminescence. Positive values defined:CEA≥5ng/mL, NSE≥48ng/mL, CYFRA21-1>3.3ng/mL, CA125>35U/mL, CA153>25U/mL,CA199>39U/mL.Statistical methods:The serum concentrations of these six tumor markers showed a skewed distribution, expressed as median and range of concentration trend and dispersion trend, respectively. Levels between two groups were compared using nonparametric rank sum test, Kruskal-Wallis test was used to compare among groups, Mann-Whitney test was used to compare between the two groups. Correlation analysis using Spearman correlation analysis, discriminant analysis using Binary Logistic Regression.χ2test was used to compare rates among groups. With P<0.05was considered statistically significant. Using SPSS19.0software for statistical analysis.Testing Results-11.1All indicators of lung cancers were significantly higher than healthy people. CEA, CYFRA21-1, NSE levels were significantly increased only in the lung cancer group, CA125, CA153levels in lung cancers were higher than in benign lung disease, in benign lung disease were higher than in healthy people, the difference of CA199levels between lung cancer and benign lung disease, benign lung disease and healthy were not statistically significant. 1.2Take benign lung disease as a control group:CEA, CYFRA21-1, NSE, CA125and CA153all have high specificity which were72.6%-98.6%. NSE and CA153sensitivity were poor, only10.7%and22.7%. The accuracy of CEA, CYFRA21-1and CA125for lung cancer were70.9%,67.6%, and60.1%. For SCLC, the area under the curve(AUC) of NSE was0.939(95%CI,0.863-1.000), AUC of CEA was0.796(95%CI,0.677-0.916). For NSCLC, AUCs of CEA and CYFRA21-1were0.762(95%CI,0.681-0.843) and0.733(95%CI,0.646-0.821). AUCs of CA125and CA153in NSCLC and SCLC were not more than0.7.1.3The levels of CEA, CYFRA21-1and CA125were positively correlated with the clinical stage of lung cancer, and the correlation of CEA was highest (r=0.495).1.4Use CEA, CYFRA21-1, NSE, CA12-5and CA15-3established a discriminant, the accuracy only61.3%to discriminate patients who occurred distant metastasis.Part2Research about SAA, CEA, CYFRA21-1and NSE in Lung Cancer Clinical DataCollected all patients highly suspected lung cancer:clinical symptoms (cough, bloody sputum, weight loss, etc.) and chest CT lung shadows have performance. When to discharge,58patients were all definite diagnosed, including31cases of lung cancer,27patients with benign lung diseases. The data of all of these cases were documented completely. Lung cancer patients with pathologically confirmed, did not receive any therapy. Take22cases as control group. Staging of lung cancer according to International Union Against Cancer TNM staging system developed in2009. The age, sex and smoking status distribution in three study groups were nearly the same.MethodSpecimens collection:All the test subjects’peripheral blood were taken at the beginning of the fasting admission, naturally incubated at room temperature for half an hour,3000r/min centrifuge for20minutes. The levels of CEA, CYFRA21-1and NSE were tested on the day of admission, another part of serum specimens were stored at-80℃refrigerator. And the serum of lung cancers who after treated also were collected. Healthy human serum samples were provided by medical center in our hospital. At the end of this experimental program, SAA were tested on one day.Detections:SAA by enzyme-linked immunosorbent assay, CEA,CYFRA21-1and NSE were tested by using electrochemiluminescence. Positive values defined: CEA>5ng/mL, NSE>24ng/mL, CYFRA21-1≥3.3ng/mL, the cut-off of SAA were taken according to the receiver operating characteristic(ROC)curve.Statistical methods:The serum concentrations of these four items showed a skewed distribution, expressed as median and range of concentration trend and dispersion trend, respectively. Levels between two groups were compared using nonparametric rank sum test, Kruskal-Wallis test was used to compare among groups, Mann-Whitney test was used to compare between the two groups. With P<0.05was considered statistically significant. Using SPSS19.0software for statistical analysis.Testing Results-22.1The median levels of SAA in lung cancer group, benign lung disease group and the control group were32.4ng/mL,31.62ng/mL and9.89ng/mL, respectively. The concentrations of SAA in lung cancer group and benign lung disease group were higher than the control group (P<0.05), whereas there is not any difference of SAA levels between lung cancer group and benign lung disease group.2.2Compared to benign lung disease, AUCs of SAA, CEA, CYFRA21-1and NSE were0.456(95%CI,0.304-0.608, p>0.05),0.738(95%CI,0.612-0.865, p<0.05),0.829(95%CI,0.719-0.940, p<0.05) and0.694(95%CI,0.556-0.832, p>0.05) respectively. When taken healthy people as a control, the area under the ROC curve of these markers were higher(Figure4).(Area under the ROC curve between0.5-0.7namely lower value of diagnosis, when the diagnostic value between0.7-0.9namely moderate, high diagnostic value is greater than0.9).2.3Compared to healthy people, the sensitivity of SAA, CEA, CYFRA21-1and NSE were96.8%,41.9%,83.9%,58.1%, specificity were90.9%,100%,86.4%,90.9%. Compared to benign lung disease, the sensitivity of SAA, CEA, CYFRA21-1and NSE were48.4%,41.9%,83.9%,80.6%, specificity were59.4%,96.3%,77.8%,59.3%.2.4SAA levels in different pathological types of lung cancer from highest to lowest were SCLC, squamous cell carcinoma and adenocarcinoma, SAA levels of these three groups were41.43ng/mL,32.4ng/mL,25.54ng/mL, but the differences among the groups were no statistical significantly.2.5SAA levels of lung cancers at different stages were not different. SAA levels were higher in patients without distant metastases than with metastasis, the medium of SAA in these two groups were40.90ng/mL and23.77ng/mL, respectively (p<0.05).2.6The levels of SAA in12lung cancers after treatment were lower than before, the medium of SAA in patients after treatment and untreated were30.3ng/mL and17.6ng/mL, respectively.Testing Conclusions1The value of CEA is highest for lung cancer among the conventional tumor markers such as CEA, CYFRA21-1, NSE, CA125, CA153and CA199, and the value of CA199is lowest. The best tumor marker for NSCLC is CYFRA21-1, while for SCLC is NSE among these six tumor markers. CA125and CAI53for the presence of false positive diagnosis of lung cancer.2SAA levels in lung cancer patients was significantly higher than in healthy people, the median of SAA levels in lung cancers is about4times of healthy people. It is not significant that the use of SAA in the field of identifying lung cancers and benign lung disease.3The value of SAA in the judgment of lung cancer clinical stage, histological type, and whether the occurrence of distant metastasis is worthless., as studies showed no significant difference, it is worthless.4The value of CEA, CYFRA21-1and NSE is higher than SAA in the identification of lung cancer and benign lung disease. When taken healthy people as a control group, four indicators all have high specificity, however, only SAA and CYFRA21-1have high sensitivity. When taken benign lung disease as a control group, the value of CYFRA21-1is better than the other three indicators.5The joint detection improves the sensitivity, specificity and accuracy of diagnosis of lung cancer. In different combinations, the best combination for the diagnosis of lung cancer is SAA+CYFRA21-1, the sensitivity, specificity and accuracy of this combination fou lung cancer are74.19%,100.00%and86.21%.6The levels of SAA in lung cancers after treated is lower than before, SAA may be used to assess the efficacy of treatment for lung cancers. |
PDF Full Text Request