Meta-analysis of diffusion-weighted imaging in the differential diagnosis of renal lesions

To assess the diagnostic value of diffusion-weighted imaging (DWI) in distinguishing between renal malignant and benign lesions.

Electronic databases were systematically searched to identify original studies evaluating DWI findings on renal lesions from January 2000 through January 2018. Pooled weighted estimates of sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR) were calculated. A summary receiver operator characteristic (sROC) curve was constructed to calculate the area under the sROC curve (AUC). Publication bias was assessed by using Deeks' asymmetry test.

A total of 15 studies including 1386 renal lesions were eligible in the meta-analysis. The pooled sensitivity and specificity with a corresponding 95% confidence interval (CI) were 0.83 (95% CI: 0.80-0.86) and 0.74 (95% CI: 0.71-0.78), respectively. The PLR, NLR, and DOR were 3.21 (95% CI: 2.39-4.32), 0.24 (95% CI: 0.18-0.30), and 15.95 (95% CI: 11.19-22.71), respectively. The AUC was 0.87 (95% CI: 0.84-0.90). Significant heterogeneity was observed between the included studies. Reference standard, country, and gradient factor were identified as the most important variable sources. No evidence of notable publication bias was reported.

DWI is an informative MRI modality in discriminating benign and malignant lesions and exhibits moderately high diagnostic accuracy. However, it remains inconclusive and limited in the absence of an optimal b value and ADC cutoff value. High-quality prospective studies regarding DWI have yet to be conducted to explore optimal imaging parameters and diagnostic thresholds.