The Role Of Metabolic Parameters Estimated By PET/CT In EGFR Target Therapy Options Of NSCLC

BackgroundThe treatment for NSCLC,the most predominant type of lung cancer,has developed rapidly thanks to the discovery and further study of gene drives such as EGFR-activating mutations.Mutations in EGFR act as both biomarkers and rational targets for treatment.Patients with EGFR mutations exhibit better response to EGFR targeted therapy.As a result,the genotype of an EGFR mutation is important when making treatment decisions.Nonetheless,some problems need our attention.First,no unified mutation detection approach exists for when a patient is being considered for an EGFR mutation test.Many other problems(inaccessible tumor site,shortage of tissues to test,tumor heterogeneity or patient refusal of invasive detection)also impact our decisions.Therefore,the development of noninvasive methods to help identify the status of the EGFR gene is necessary.As a noninvasive method,PET/CT has many applications in patients with NSCLC.The basis of PET/CT is that the glucose metabolism of a tumor could partly be reflected by FDG uptake.The processes downstream of the EGFR gene could influence glucose metabolism,whichcorrelates with FDG uptake.Therefore,FDG uptake might be associated with EGFR gene mutation in tumors and then the mutation status of EGFR might be predicted by these metabolic parameters estimated by PET/CT.Aim and scoreWe explored the role of metabolic parameter estimated by FDG PET/CT for predicting EGFR mutation status in NSCLC and further establish a useful prediction model to help in screening and identification of EGFR mutation status.MethodsBetween January 2013 and September 2016,a retrospective analysis was carried out for all newly diagnosed patients with pathologically confirmed NSCLC with further testing performed for EGFR mutation analysis and PET/CT examination less than 2 weeks before any treatment in Shandong Cancer Hospital Affiliated with Shandong University.We excluded the following patients:(1)patients who had any therapy prior to PET/CT,(2)patients whose specimens were inadequate for mutation analyses,(3)patients with acute and chronic pneumonia or other infections that might interfere with the imaging of PET/CT,(4)patients who had other cancers previously,(5)patients who had double or multiple primary cancers,and(6)patients whose primary lesion measured less than 1 cm in diameter,which might cause an error in PET/CT imaging due to a partial volume effect.Ultimately,the clinical data and PET/CT imaging of 87 patients with EGFR test were analyzed.Age,gender,smoking history,TNM stage,location of primary tumor,pre-therapy level of serum CEA and metabolic parameters of PET/CT(SUVmax,SUVmean and MTV of the tumors)were analyzed.Metabolic parameters estimated by PET/CT were compared by EGFR mutation status through the Wilcoxon rank-sum test.The difference in categorical variations including clinical parameters and metabolic parameters of PET/CT between different mutation statuses were analyzed through chi-squared test.An ROC curve was applied to obtain the cutoff of continual variables to predict mutation status.Multivariate logistic regression analysis was used to analyze the independent predictors of an EGFR+ vs.EGFR-status.Finally,the predictive value of the model based on the independent predictors was assessed by analyzing the area under the ROC curve(AUC).Two-sided p values<0.05 were referred to as statistically significant.All statistical analyses were performed using SPSS(version 20.0).Results1.Comparison of metabolic parameters between different EGFR mutation statusesThe MTV and SUVmean were significantly lower in cases of EGFR mutations than in cases of wild-type EGFR(p =0.001 and p=0.031,respectively).No statistically significant correlation with SUVmax was found between the patients with EGFR mutation and wild-type EGFR.2.EGFR status and related variablesThe ROC curve revealed that the pretreatment CEA cutoff and tumor diameter were 15.0(ng/mL)and 3.5(cm).The ROC curve revealed that the cutoff points forSUVmax,SUVmean and MTV were 10.4,6.0 and 11.0(cm3),respectively.We dichotomized the patients according to these thresholds.Besiedes,a significant correlation was found between EGFR status and MTV,smoking history,pathology and location of the tumor.Other characteristics including age,gender,and TNM stage did not differ significantly by mutation status..3.Multivariable analysis and Predicting EGFR mutation A multivariable logistic regression analysis revealed that only non-smoker status,peripheral tumor location and lower MTV were significant predictors of EGFR mutation.ROC curve analysis was then performed to validate the predictive model ofthese factors and the AUC was found to be 0.805,which suggests good discrimination.ConclusionsThe results of the present study suggest that EGFR mutation in NSCLC is correlated with clinical and metabolic parameters including MTV,smoking status and tumor location.The combined evaluation of these three factors may be helpful in discriminating mutation status,especially in cases of inadequate sampling or when genetic testing is not available.However,a larger,multi-institutional prospective study is needed for further validation of the current results.