Predictive value of ¹⁸F-FDG PET/CT for EGFR mutations in patients with lung adenocarcinoma

Objective: To investigate the predictive value of ¹⁸F-fluorodeoxyglucose (FDG) PET/CT imaging for the epidermal growth factor receptor (EGFR) mutations in patients with lung adenocarcinoma. Methods: From January 2013 to December 2017, a total of 146 patients (83 males, 63 females, age: (60.2 ± 10.3) years) who were confirmed as lung adenocarcinoma by pathology and were examined by ¹⁸F-FDG PET/CT imaging and EGFR mutation testing in Shanxi Cancer Hospital were retrospectively analyzed. The differences of clinical characteristics (age, gender, smoking, tumor diameter, loymph node metastasis, distant metastasis, stage, thyroid transcripition factor-1 (TTF-1), NapsinA, cyiokeratin (CK)-7, Ki-67) and PET/CT parameters (maximun standardized uptake value (SUV_max) of the primary tumor (pSUV_max), SUV_max of lymph node (nSUV_max) and SUV_max of distant metastasis (mSUV_max)) between patients of EGFR mutation and EGFR wild type were analyzed using independent-sample t test, X² test and Fisher exact test. The predictors for EGFR mutation were analyzed by logistic regression analysis. The predictive value of pSUV_max and pSUV_max combined with gender, smoking and tumor diameter was determined by receiver operating characteristic (ROC) curve analysis. Results: There were 46.58% (68/146) patients with EGFR mutations and 53.42% (78/146) patients with wild type. Gender, smoking, lymph node metastasis, tumor diameter, pSUV_max, nSUV_max, TTF-1, NapsinA and Ki-67 were significantly different between patients with EGFR mutations and those with wild type (t values: from -3.023 to -2.032, X² values: 4.725-33.749, all P < 0.05). Female (odds ratio (OR) = 3.236, 95% CI: 1.213-8.779; P = 0.029), non-smoker (OR = 4.947, 95% CI: 1.796-13.621; P = 0.019), tumor diameter < 3.5 cm (OR = 2.750, 95% CI: 1.109-6.818; P = 0.001) and pSUV_max < 9.1 (OR = 2.960, 95% CI: 1.227-7.141; P = 0.016) were predictors of EGFR mutations in lung adenocarcinoma. The area under the curve (AUC) of pSUV_max was 0.640 with the specificity of 43.6% (34/78) and the sensitivity of 27.9% (19/68), while the AUC of the four independent factors was 0.83 with the specificity of 71.8% (56/78) and the sensitivity of 83.8% (19/68). Conclusions: pSUV_max is associated with mutant EGFR status. Moreover, the combination of pSUV_max, gender, smoking and tumor diameter can enhance the predictive value on EGFR mutation status in patients with lung adenocarcinoma. (authors)