ABSTRACT
Objective To evaluate the diagnostic value of CT and PET-CT characteristics in predicting the presence of epidermal growth factor receptor(EGFR) gene mutations in lung adenocarcinomas. Methods One hundred and sixty-eight lung adenocarcinomas cases confirmed by pathology were enrolled in our study. They were divided into EGFR gene mutations group (89 cases) and wild types group (79cases) according to whether EGFR gene mutation occurred. All patients underwent CT examination. Seventy-five patients underwent PET-CT examination, including 37 gene mutationsand 38 wild types.The demographic (the patients' age, the gender and smoking history), CT characteristics(lobulation, cavitation, spiculation, pleural-indentation, air-bronchogram, ground glass opacity/tumor ratio(G/T) and the maximum diameter of tumor(Dmax)) and PET-CT characteristics(the maximum standardized uptake value, (SUVmax))between these two groups were retrospectively compared. The independent sample t test was used to analyze the difference between these two groups regarding the patients' age,Dmax,SUVmax. The χ2 test was used to demonstrate the difference between these two groups regarding the gender, smoking history and CT features including lobulation, cavitation, spiculation, pleural-indentation, air-bronchogram and G/T.The trend analysis between SUVmax and EGFR gene mutations was performed by usingχ2 test for trend.Results No significant difference was found regarding partial CT characteristics of lesions including lobulation, cavitation, spiculation, pleural-indentation, air-bronchogram (P>0.05),however, the Dmax of EGFR gene mutations group and wild types group were(2.53±1.39),(3.00±1.77)cm, respectively. The amount of G/T>50%in EGFR gene mutations group and wild types group was 21 and 5, respectively. Significant differences were found regarding the G/T and Dmax(χ2=9.538, P50%as diagnostic criterion. Receiver operating characteristic (ROC) results indicated Dmax=1.85 cm was the optimal value in predicting EGFR gene mutations, with the sensitivity and specificity of 76% and 42%, respectively. Meanwhile, SUVmax=6.85 was the optimal value, with the sensitivity and specificity of 71% and 73%, respectively. Moreover,χ2 test for trend showed that an obvious trend was found to associate SUVmax with the incidence of EGFR gene mutations (χ2=15.755, P<0.05). Conclusion SUVmax may be helpful in predicting EGFR gene mutations in lung adenocarcinomas with relatively high diagnostic value.