RT Journal Article
SR Electronic
T1 <strong>SUV-derived Parameters Assessed on F-18-FDG PET/CT Predict EGFR</strong> <strong>Mutation in Lung Adenocarcinoma Patients: the Construction of CART Predictive Model</strong>
JF Journal of Nuclear Medicine
JO J Nucl Med
FD Society of Nuclear Medicine
SP 295
OP 295
VO 61
IS supplement 1
A1 Rongfu Wang
A1 Xuhe Liao
A1 Si Chen
A1 Meng Liu
A1 Xueqi Chen
A1 Lei Yin
A1 Dafang Chen
YR 2020
UL http://jnm.snmjournals.org/content/61/supplement_1/295.abstract
AB 295Objectives: To investigate the potential relationship between Epidermal growth factor receptor gene (EGFR) mutation status and standardized uptake value (SUV)-derived parameters from 18F-fluorodeoxyglucose (F-18-FDG) positron emission tomography/ computed tomography (PET/CT) examinations combining with other clinical characteristics through classification and regression trees (CART) in patients with lung adenocarcinoma (ADC), in order to obtain the noninvasive predictive model for EGFR mutation. Methods: Data of 192 ADC patients pre-treatment, who underwent F-18-FDG PET/CT scans, EGFR gene mutations test for newly diagnosed ADC patients from December 2011 to April 2018, were retrospectively collected. Then a series of clinical parameters including EGFR mutation status, SUV-derived features of primary tumor [maximum standardized uptake value (SUVmax), average of standardized uptake value (SUVmean), metabolic tumor volume (MTV), and total lesion glycolysis (TLG)], serum tumor makers and so on were gathered, which were analyzed through CART to build the model for EGFR mutation prediction. Predictive effectiveness of the model was validated by 1000-time Bootstrap. Results: Ratios of EGFR mutation were 33.3% (64/192). Age, smoking status, SUVmean, pMTV (primary MTV), pTLG (primary TLG), CEA, SCC, NSE, TPA, and proGRP could be independently and significantly associated with EGFR mutation of ADC patients. The area under the curve (AUC) which for the predictive value of these factors were 0.785 (95%CI: 0.743-0.827); Sensitivity and specificity were 85.5% and 71.4% respectively. Conclusions: SUVmean, pMTV and pTLG were a set of independent predictors and could be integrated with other clinical factors (age, smoking status, CEA, SCC, NSE, TPA and proGRP) to enhance the discriminability on the EGFR mutation status in ADC patients. Keywords: Epidermal growth factor receptor (EGFR); Maximum standardized uptake value (SUVmax); metabolic tumor volume (MTV); total lesion glycolysis (TLG); classification and regression trees (CART); lung adenocarcinoma (ADC)