Ultrasound normalized local variance to assess metabolic dysfunction-associated steatotic liver disease

Lauren D Hagenstein,Joseph Jenkins,Colby Adamson,Jourdain Dong,John Moore,Jing Gao
DOI: https://doi.org/10.1016/j.clinimag.2024.110326
Abstract:Purpose: Increased prevalence of metabolic dysfunction-associated steatotic liver disease (MASLD) highlights a desire for screening with liver ultrasound normalized local variance (NLV). We aimed to assess variations in NLV values measured at different sampling depths and discuss common technical considerations in measuring liver NLV. Methods: We retrospectively measured liver NLVs at variable depths on ultrasound images pre-recorded in 116 participants who underwent liver magnetic resonance imaging-proton density fat fraction (MRI-PDFF) and ultrasound to screen for MASLD. Liver NLVs were measured and differences at variable depths were tested using one-way analysis of variance (ANOVA) and multiple paired comparisons using post hoc Tukey honestly significant difference (HSD), Scheffé, Bonferroni, and Holm multiple comparisons. Diagnostic performance of NLV values were analyzed by area under the receiver operating characteristic (AUROC) curve. Results: The NLV measured at a depth of 10 cm significantly differed from those measured near the liver capsule and at depths of 6 cm and 8 cm (p < 0.001) from the skin. There was no significant difference in NLV value in other paired groups (p > 0.05). The difference in the area under AUROCs for NLVs measured at variable depths was not significant (p > 0.05). Conclusions: The best diagnostic performance of liver NLV was measured at depth of 8 cm from the skin, although NLV measured at variable depth showed similar diagnostic performance for assessing ≥ mild hepatic steatosis. The study results provide a reference that can be used in the development of standardized scanning protocols and technical considerations in measuring liver NLV.
What problem does this paper attempt to address?