TY - JOUR T1 - Correlation between apparent diffusion coefficients (ADC) on diffusion-weighted MRI and standardized uptake value (SUV) on FDG-PET/CT in patients with cancer: a meta-analysis JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1887 LP - 1887 VL - 57 IS - supplement 2 AU - Guohua Shen AU - Shuang Hu AU - Anren Kuang Y1 - 2016/05/01 UR - http://jnm.snmjournals.org/content/57/supplement_2/1887.abstract N2 - 1887Objectives Diffusion-weighted magnetic resonance imaging (DW-MRI) and 2-deoxy-2-[18F]fluoro-D-glucose-positron emission tomography/computed tomography (PET/CT) is increasingly recognized as important for assessing tumor malignancy in oncology. The apparent diffusion coefficient (ADC) measured by DWI and the standardized uptake value (SUV) from 18F-FDG PET have similar clinical applications. Results from validation studies have so far been conflicting with some studies confirming a good correlation between ADC and SUV and others presenting negative results. The purpose of this study was to assess the correlation between the ADC and SUV in various cancers.Methods We searched PubMed, EMBASE, and Cochrane Database of Systematic Review, and selected research articles on the relationship between ADC and SUV (published between July 1, 2000-October 1, 2015) with language limited to English. According to the literature inclusion and exclusion criteria, the potential eligible articles were selected. The quality of included articles was evaluated according to the Quality Assessment of Diagnosis Accuracy Studies-2 (QUADAS-2). Correlation coefficient (r) values were extracted from each study and 95% confidence intervals (CIs) were calculated after applying Fisher’s z transformation. The r values and the 95% confidence interval (CI) were pooled with STATA 12.0 software under a random-effects model. Begg's test was used to analyze the existence of publication bias and draw funnel plot, and the sources of heterogeneity were explored by subgroup analysis based on the study design, modality and cancer type.Results Forty-three articles involving 2040 lesions were identified as eligible for the meta-analysis. Twenty-two studies were designed prospectively; other 21 studies were retrospectively designed. With regard to scanning modality, the integrated PET/MRI was used in 11 studies; in the remaining 32 studies two separate devices PET/CT and MRI were used. In all included studies, the mostly involved tumor type was lung cancer, analyzed in a total of 11 studies, followed by head and neck cancer in 6 studies. Breast tumors and cervical tumors were analyzed in 4 studies, respectively. The results of QUADAS-2 indicated that 12 studies showed low risk of bias in all the aspects assessed, and 30 studies showed low concern for applicability. The pooled r value was - 0.33 (95% CI, -0.37, -0.29), with high heterogeneity (I2=63.9%, p<0.001), indicating a relatively poor negative correlation (Figure 1). The results of subgroup analyses were presented in Table 1. Both study design and scanning modality cannot explain the heterogeneity. Subgroup analysis for different tumor types indicated the tumor type was a main source of heterogeneity. Lung cancer had a stronger correlation with r= - 0.415, compared those of head and neck cancer (r= - 0.288), cervical cancer (r= - 0.082), and breast cancer (r= - 0.233). The pancreatic tumor had the strongest correlation with r = - 0.584. As for other tumors such as lymphoma, rectal cancer, the sample size was too small to perform further analysis. Based on the Begg’s test, there was no significant publication bias in included studies.Conclusions Although both SUV and ADC are helpful parameters in oncology, the overall correlation between them is relatively poor regardless of tumor type. Different tumor types exhibited varied degree of correlation. The true correlation needs further validation by trials with a large sample of different tumor types. This study was supported by National Natural Science Foundation of China (Grant No. 81471692). View this table:Table 1. Subgroup analyses for the correlation between SUV and ADC. ER -