Premium
Correlation between tumor glucose metabolism and multiparametric functional MRI (IVIM and R2*) metrics in cervical carcinoma: Evidence from integrated 18 F‐FDG PET/MR
Author(s) -
LiOu Zhang,
HongZan Sun,
XiaoXi Bai,
ZhongWei Chen,
ZaiMing Lu,
Jun Xin,
QiYong Guo
Publication year - 2019
Publication title -
journal of magnetic resonance imaging
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.563
H-Index - 160
eISSN - 1522-2586
pISSN - 1053-1807
DOI - 10.1002/jmri.26557
Subject(s) - medicine , nuclear medicine , correlation , rank correlation , spearman's rank correlation coefficient , immunohistochemistry , carcinoma , standardized uptake value , cervical carcinoma , cervical cancer , positron emission tomography , cancer , pathology , mathematics , statistics , geometry
Background Multiparameter, multimodality 18 F‐FDG PET/MRI holds great potential for the diagnosis of cervical cancer based on the correlation between tumor glucose metabolism and imaging parameters. Purpose To characterize the heterogeneity of tumor glucose metabolism by evaluating the correlation between 18 F‐FDG uptake parameters and multiparametric functional MRI metrics in cervical carcinoma. Study Type Retrospective. Population Fifty‐four patients with cervical carcinoma. Field Strength/Sequence Hybrid PET/MR (3T), multi‐b DWI, and R2* mapping. Assessment The maximum and mean standardized uptake values (SUV max and SUV mean , respectively) from PET and functional MRI metrics (D, D*, f, and R2*) were obtained. Cervical carcinoma tissues also underwent HIF‐1α, VEGF, and GLUT‐1 immunohistochemical staining. Statistical Tests Single‐factor Spearman rank and Pearson correlation analysis and multiple linear regression (MLR) analysis were applied. Results R2*, D, and f have different degrees of correlation (moderate, weak, moderately strong correlation, respectively) with SUV max and SUV mean ( r = 0.530 and 0.527, and P < 0.001 for R2*; r = –0.292 and –0.291, and P < 0.05 for D; r = 0.539 and 0.520, and P < 0.001 for f, respectively). Immunohistochemical staining showed that HIF‐1α expression has a moderate degree of correlation with R2* ( r = 0.491; P < 0.001); GLUT‐1 expression was significantly correlated with SUV max and SUV mean ( r = 0.633 and 0.622; P < 0.001), and VEGF expression had a moderately strong correlation with f ( r = 0.457; P = 0.001). If SUV max is the dependent variable, MLR yields an R‐squared value after adjustment (adjusted R‐squared) = 0.358, and F = 10.833 ( P < 0.001), and the fitting linear equation is Y (SUV max ) = 9.184 + 0.161X1 (R2*)+50.343X2 (f)‐4.780 (D). Otherwise, MLR yields the adjusted R‐squared = 0.342, and F = 10.187 ( P < 0.001), and the linear regression equation is Y (SUV mean ) = 5.925 + 0.102X1 (R2*)+28.029X2 (f)‐2.907X3 (D). Data Conclusion The functional MRI sequence parameters R2*, f, and D can provide information on the hypoxic condition, blood perfusion, and molecular diffusion of the tumor. 18 F‐FDG PET/MR multi‐imaging technique can be adopted to evaluate the heterogeneity of glucose metabolism in cervical carcinoma. Level of Evidence: 3 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2019;49:1704–1712.