Diagnostic performance of Liver Imaging Reporting and Data System treatment response algorithm: a systematic review and meta-analysis.

Objective: To systematically determine the accuracy of Liver Imaging Reporting and Data System treatment response (LR-TR) algorithm for diagnosing the viability of hepatocellular carcinoma (HCC) treated with locoregional therapy (LRT).

Methods: Original studies reporting the diagnostic accuracy of LR-TR algorithm on dynamic contrast-enhanced computed tomography or magnetic resonance imaging (MRI) were identified in MEDLINE and EMBASE up to June 1, 2020. The meta-analytic summary sensitivity and specificity of LR-TR algorithm were calculated using a bivariate random-effects model. Subgroup analyses and meta-regression analysis were performed to explore study heterogeneity.

Results: We found six studies reporting the accuracy of LR-TR viable category (601 observations in 453 patients). The meta-analytic pooled sensitivity and specificity of LR-TR viable category were 63% (95% confidence interval [CI], 39-81%; I = 88%) and 96% (95% CI, 91-99%; I = 76%), respectively. The meta-analytic pooled sensitivity and specificity of LR-TR viable or equivocal category combined were 71% (95% CI, 55-84%; I = 89%) and 87% (95% CI, 73-94% I = 80%), respectively. Studies which used only MRI showed a trend towards higher sensitivity (71% [95% CI, 46-88%]) with a comparable specificity (95% [95% CI, 86-99%]) of LR-TR viable category compared to the whole group. The type of reference standard and study design were significantly associated with study heterogeneity (p ≤ 0.01).

Conclusions: The LR-TR viable category had high specificity but suboptimal sensitivity for diagnosing the viability of HCC after LRT. Substantial study heterogeneity was noted, and it was significantly associated with the type of reference standard and study design.

Key Points: • The meta-analytic pooled sensitivity and specificity of LR-TR viable category were 63% (95% CI, 39-81%) and 96% (95% CI, 91-99%), respectively. • The meta-analytic pooled sensitivity and specificity of LR-TR viable or equivocal category combined were 71% (95% CI, 55-84%) and 87% (95% CI, 73-94%), respectively. • The type of reference standard and study design were the factors significantly influencing study heterogeneity (p ≤ 0.01).