Relationship between serum tumor markers and EGFR gene mutation in NSCLC based on logistic regression and ROC curve analysis

ObjectiveTo evaluate the correlation between serum tumor markers and epidermal growth factor receptor(EGFR) gene mutation in non-small cell lung(NSCLC) cancer by logistic regression and receiver operating characteristic(ROC) curve.

MethodsA total of 128 hospitalized patients with lung adenocarcinoma were collected and 5 mL of venous blood was collected at the initial diagnosis, 3 mL serum was separated, and serum tumor markers were detected. According to the results of EGFR gene detection, they were divided into EGFR gene mutation group and EGFR gene non-mutation group. The general clinical characteristics of two groups of patients with NSCLC were compared. The ROC curve was drawn based on logistic regression model, and the area under the curve(AUC) was calculated to evaluate the value of each index and the combined diagnosis.

ResultsEGFR mutation was associated with gender, smoking history, SCC and CA125(P < 0.05 to P < 0.01). The mutation frequency of EGFR in female patients with NSCLC was higher than that in male(P < 0.01). The mutation frequency of EGFR in patients with smoking history was lower than that in patients without smoking history(P < 0.01). The mutation frequency of EGFR in patients with negative serum SCC was higher than that in patients with positive serum SCC(P < 0.01). The mutation frequency of EGFR in patients with negative serum CA125 was higher than that in patients with positive CA125(P < 0.05). Among the serum tumor markers, gender, CA125, SCC and smoking history had higher diagnostic value with EGFR gene mutation(P < 0.05). At the same time, the combined diagnosis of these indicators, such as gender+SCC, gender+CA125+SCC, gender+SCC+smoking history, CA125+SCC+smoking history, gender+CA125+SCC+smoking history, which all had AUC value more than 0.75(P < 0.05), had high diagnostic value.

ConclusionsThere is a good correlation between SCC and CA125 in serum tumor markers and EGFR mutation in NSCLC.The combination of SCC and CA125 with other clinical data(gender, smoking history) can enhance the predictive discrimination of EGFR, which is worthy of further clinical application.