TY - JOUR T1 - The maximum standardized uptake value in <sup>18</sup>F-deoxyglucose (FDG) PET imaging able to increase the effectiveness of CT radiomics for distinguishing malignant from benign ground-glass nodules—a case-control study based on propensity score matching JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1046 LP - 1046 VL - 62 IS - supplement 1 AU - Rong Niu AU - Xiaonan Shao AU - Xiaoliang Shao AU - Jianfeng Wang AU - Yuetao Wang Y1 - 2021/05/01 UR - http://jnm.snmjournals.org/content/62/supplement_1/1046.abstract N2 - 1046Objectives: Lung cancer is the leading cause of cancer-related deaths worldwide, especially in China, where the incidence of lung cancer is increasing rapidly. The diagnosis and differential diagnosis of ground-glass nodules (GGNs) have become a huge challenge for clinicians. There is an urgent need to develop a reliable and practical non-invasive imaging method to accurately identify GGNs to guide the individualized clinical management of GGNs. CT radiomic features have some value for the qualitative diagnosis of GGNs, but the diagnostic performance still needs to be improved. The purpose of this study is to explore whether the maximum normalized uptake value (SUVmax) in 18F-deoxyglucose (FDG) PET imaging can increase the diagnostic efficiency of CT radiomics-based predictive models for distinguishing malignant from benign GGNs. Methods: A retrospective collection of 190 indeterminate GGNs from 165 patients underwent 18F-FDG PET/CT examinations in our hospital from January 2012 to March 2020. Propensity score matching (PSM) was used to construct GGNs with similar baseline characteristics. The LIFEx software was used to extract 49 CT radiomic features, and the least absolute shrinkage and selection operator (LASSO) algorithm was used for parameter selection and the establishment of a radiomic signature (Rad-score). Furthermore, the multiple logistic regression was used to develop a CT radiomics model combined with semantic features, and a PET/CT radiomics model combined with SUVmax. The receiver operating characteristic (ROC) curve was used to compare the diagnostic performance of different models. Results: 23 in the benign group and 92 in the adenocarcinoma group after 190 GGNs were subjected to 1:4 PSM. After texture analysis, a Rad-Score containing three CT texture features (GLRLM_LRE, GLRLM_LRLGE and NGLDM_Contrast) was developed for each nodule. Compared with Rad-score and CT radiomics model (AUC: 0.704 (95%CI: 0.562-0.845) and 0.908 (95%CI: 0.842-0.975)), the diagnostic efficiency of the PET/CT radiomics model was the best (AUC: 0.940 (95%CI: 0.889-0.990)), and there was a significant difference in pairwise comparison (P=0.001-0.030). Conclusions: SUVmax could effectively improve the diagnostic efficiency of CT radiomics model for distinguishing malignant from benign GGNs. PET/CT dual-modal radiomics is expected to become a non-invasive and reliable method for differential diagnosis of indeterminate GGNs. ER -