RT Journal Article SR Electronic T1 The Relationship between the Metabolic Parameters of Primary Tumors on PET-CT and Lymph Node Metastasis in Resectable Peripheral Non-small-cell Lung Cancer: a case-control study JF Journal of Nuclear Medicine JO J Nucl Med FD Society of Nuclear Medicine SP 1033 OP 1033 VO 62 IS supplement 1 A1 Xiao-feng LI A1 Yuetao Wang YR 2021 UL http://jnm.snmjournals.org/content/62/supplement_1/1033.abstract AB 1033Objectives: To explore the dose-effect relationship between primary tumor metabolic parameters and peripheral non-small-cell lung cancer (NSCLC) lymph node metastasis (LNM). Methods: The patients underwent 18F-FDG PET-CT examinations in our center and were suspected or diagnosed with lung cancer from Jan. 2017 to Mar. 2020. The dependent variable was the LNM status. The independent variable was primary tumor tumor-to-liver standardized uptake value ratio (PT.TLR). According to the sample distribution, the enrolled cases were stratified into three group (<4, 4-9, ≥9) based on PT.TLR. Then, the Wilcoxon-Mann-Whitney U test compared the differences among the three groups, the chi-square test or Fisher’s exact test was used for the comparison of categorical variables, and Student's t test was used to determine differences in normally distributed continuous variables.The confounding covariates were filtered by univariate analysis, then odds ratio (OR) values were compared through stratified analysis, the relationship between PT.TLR and peripheral NSCLC LNM was determined by a fitting curve and threshold effect analysis. Finally, unadjusted, simple adjusted and fully adjusted logistics models were established to determine the independent effect size. Results: Of 255 cases were finally enrolled (Fig. 1). The baseline results of different PT.TLR groups showed that, except for age, BMI, pleural indentation and lateral (p-values were 0.501, 0.197, 0.252, 0.062, respectively), the other variables all showed signifinmmncant differences. A univariate analysis of all the exposure covariates showed that PT.TLR, PT.SUVmax, MTV, TLG, PT.CTmean, lobulation sign, PT.size, and grade had statistically significant differences in the OR values of LNM (p-values were <0.001, <0.001, 0.016, 0.004, 0.002, 0.004, <0.001, and <0.001, respectively). Confounding covariates included sex, age, BMI, grade, PT.CTmean, PT.size, and MTV. These covariates were taken into account for the stratified analysis in the different PT.TLR groups. When the PT.TLR <4, the OR values were all >1. A smooth fitting curve (Fig. 2) between PT.TLR and LNM and the subgroups stratified by grade (Fig. 3) also verified the results. The threshold effect analysis showed that the curve of PT.TLR and LNM had two break-point values. When PT.TLR ≤5.07 or ≥10.52, the OR values were >1. Whereas PT.TLR was 5.07-10.52, the OR value was <1, especially in small size (<3cm) adenocarcinoma. In the moderate/well-differentiated subgroups of three multiple regression models, the OR values of the PT.TLR and 95% confidence intervals were 1.24, (1.07, 1.44), 1.28, (1.09, 1.51) and 1.18, (0.99, 1.41), respectively. And they were all greater than those of the PT.SUVmax. There were also significant differences in the trend test (p-value <0.05). However, in the poorly differentiated group, after considering confounding covariates, there was no correlation between the PT.TLR and LNM. Conclusions: In this retrospective single-center study, PT.TLR ≤5.07 is an independent risk factor for peripheral NSCLC LNM. For increasing every one value of the PT.TLR, the risk of LNM increases by 18%. When the PT.TLR is 5.07-10.52, for additional one value of PT.TLR, the risk of LNM decreased 15%, especially in small size (<3cm) adenocarcinoma, confirming the saturation and segmentation effects in the dose-effect relationship between PT.TLR and LNM. Therefore, PT.TLR could be recognized as a supplementary indicator to assess LNM, especially when PT.SUVmax is the same. The results are more significant in moderate/well-differentiated peripheral adenocarcinoma. A large-scale multicenter prospective cohort study is needed to verify our results.$$table_{D97A885D-F146-4F1C-8D89-DF335A2E70FD}$$