Diagnostic accuracy of Liver Imaging Reporting and Data System locoregional treatment response criteria: a systematic review and meta-analysis

Objective There is increasing adoption of Liver Imaging Reporting and Data System (LI-RADS) treatment response (LR-TR) criteria. However, there is still a relative lack of evidence evaluating the performance of these criteria. We performed this study to assess the diagnostic accuracy of LI-RADS LR-TR criteria. Methods A thorough search of PubMed, Embase, Scopus, and Cochrane Central Register of Controlled Trials for studies reporting diagnostic accuracy of LI-RADS LR-TR criteria was conducted through 30 June 2020. The meta-analytic summary of sensitivity, specificity, and diagnostic odds ratio of LI-RADS LR-TR criteria was computed using explant histopathology as the reference standard. The quality of the studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. Results Four studies were found eligible for meta-analysis. The total number of LR-TR observations was 462 (240 patients, 82.5% males). Different locoregional therapies (LRTs), including bland embolization, chemoembolization, radiofrequency ablation, and microwave ablation, had been used. The mean time interval between LRT and liver transplantation ranged from 181 to 219 days. There was a moderate to good inter-reader agreement for LR-TR criteria. The pooled sensitivity and specificity of LR-TR criteria for viable disease were 62% (95% CI, 49–74%; I 2 = 69%) and 87% (95% CI, 76–93%; I 2 = 57%), respectively. The pooled diagnostic odds ratio and area under the curve were 9.83 (95% CI, 5.34–18.08; I 2 = 19%) and 0.80. Conclusions LI-RADS LR-TR criteria have acceptable diagnostic performance for the diagnosis of viable tumor after LRT. Well-designed prospective studies evaluating criteria of equivocal lesions and effect of different LRTs should be performed. Key Points • The pooled sensitivity and specificity of LI-RADS LR-TR criteria for the diagnosis of viable tumor were 62% and 87%, respectively. • The pooled diagnostic odds ratio and area under the curve were 9.83 and 0.80. • LR-TR criteria had a moderate to good inter-reader agreement.