Correlation of the apparent diffusion coefficient and the standardized uptake value in neoplastic lesions: a meta-analysis

Objective Diffusion-weighted imaging and fluorine-18-fluorodeoxyglucose PET are increasingly being recognized as feasible oncological techniques. The apparent diffusion coefficient (ADC) measured by diffusion-weighted imaging and the standardized uptake value (SUV) from fluorine-18-fluorodeoxyglucose PET have similar clinical applications. The aim of this study was to assess the correlation between these two parameters in various cancers. Materials and methods Several major databases were searched for eligible studies. The correlation coefficient (rho) values were pooled in a random-effects model. Begg's test was used to analyze the existence of publication bias and the sources of heterogeneity were explored in subgroup analyses on the basis of study design, diagnostic method, scanning modality, and tumor type. Results Thirty-five articles were accepted. The pooled rho value of all of the accepted studies was -0.30 (95% confidence interval: -0.33 to -0.27), and notable heterogeneity was present (I-2=69.4%, P < 0.001), which indicated a relatively weak negative correlation. The pooled rho values were -0.26, -0.33, -0.32, and -0.33 for the SUVmax/ADC(mean), SUVmax/ADC(min), SUVmean/ADC(mean), and SUVmean/ADC(min) relationships, respectively. The study design and diagnostic method were potential sources of heterogeneity. Lung cancer showed a stronger correlation (rho=-0.42) than head and neck cancer (rho=-0.27), cervical cancer (rho=-0.21), and breast cancer (rho=-0.23). A Begg's test indicated no significant publication bias among the accepted studies (P>0.05). Conclusion The two functional parameters of ADC and SUV showed a very weak inverse correlation, which may contribute toward a sophisticated characterization of tumor biology. However, the findings require further validation with trials with large samples and different tumor types. Copyright (C) 2017 Wolters Kluwer Health, Inc. All rights reserved.