Diagnostic performance of liver fibrosis assessment by quantification of liver surface nodularity on computed tomography and magnetic resonance imaging: systematic review and meta-analysis

Abstract

Objectives: To determine the diagnostic performance of the liver surface nodularity (LSN) score for assessment of hepatic fibrosis.

Methods: Major databases were searched to identify studies reporting the diagnostic performance of the LSN score published from January 2011 to April 2021. The pooled sensitivity and specificity of the LSN score in the diagnosis of significant fibrosis (F2-4), advanced fibrosis (F3-4), and liver cirrhosis (F4) were calculated using a bivariate random-effects model, and the summary areas under the hierarchical summary receiver operating characteristics curve (AUC) for each were calculated. Subgroup analysis was performed to explore the causes of study heterogeneity.

Results: Of 168 screened studies, six were included in the analysis. The pooled sensitivity, specificity, and AUC of the LSN score were 68% (95% confidence interval [CI], 45-84%), 89% (95% CI, 83-93%), and 0.90 (95% CI, 0.87-0.92), respectively, for the diagnosis of significant fibrosis; 88% (95% CI, 60-97%), 78% (95% CI, 57-90%), and 0.89 (95% CI, 0.86-0.91), respectively, for the diagnosis of advanced fibrosis; and 83% (95% CI, 71-91%), 80% (95% CI, 72-85%), and 0.87 (95% CI, 0.83-0.89), respectively, for the diagnosis of liver cirrhosis. Cutoff values of the LSN score affected study heterogeneity in the diagnosis of advanced fibrosis (p = 0.03) and liver cirrhosis (p = 0.01). Imaging modality (p = 0.01) and reference standard (p = 0.01) affected study heterogeneity in the diagnosis of liver cirrhosis.

Conclusions: The LSN score showed good diagnostic performance in assessing the hepatic fibrosis stage. Additional investigations are needed to improve and standardize the LSN score. Key Points • The liver surface nodularity (LSN) score showed overall good diagnostic performance in assessing hepatic fibrosis. • The pooled sensitivity of the LSN score was highest in the diagnosis of advanced fibrosis (88%; 95% confidence interval [CI], 60-97%), and the pooled specificity was highest in the diagnosis of significant fibrosis (89%; 95% CI, 83-93%). • The diagnostic performance of the LSN score differed according to the cutoff value, imaging modality, and reference standard.

Keywords: Computed tomography; Diagnosis; Fibrosis; Liver; Magnetic resonance imaging.