Investigators in Japan developed a statistical model to identify patients with high-risk nonalcoholic steatohepatitis (NASH) using three ultrasound-based markers, which yielded an 86.5% positive predictive value.
A multiparametric ultrasound-based model showed promise for identifying patients with high-risk nonalcoholic steatohepatitis (NASH), reducing the need for liver biopsy to determine whether a patient is a candidate for pharmacologic therapy, according to a new study.
The cross-sectional secondary analysis, published in Radiology, included data collected between April 2017 and March 2019 on 111 patients at Tokyo Medical University Hospital with clinical indication for liver biopsy for suspected nonalcoholic fatty liver disease (NAFLD) to derive a model for identifying patients with high-risk NASH. High-risk NASH was defined as those with substantial liver fibrosis (F2 or greater) and an NAFLD activity score (NAS) of 4 or greater.
The results were validated in a data set from 102 patients in Korea with elevated liver enzyme levels or clinically suspected diffuse liver disease, who were referred for liver biopsy between January 2018 and July 2019.
"In this study, we present a score that uses only US-based markers (liver stiffness, attenuation coefficient, and dispersion slope) that can help identify patients with high-risk nonalcoholic steatohepatitis (NASH) who are candidates for inclusion in clinical trials and for pharmacologic therapy," the authors, led by Katsutoshi Sugimoto, MD, PhD, of Tokyo Medical University, wrote in the study. "Our choice of NASH with a nonalcoholic fatty liver disease activity score of greater than or equal to 4 and a fibrosis score of greater than or equal to F2 is based on the literature, including many therapeutic studies showing that the presence of elevated necroinflammatory activity is related to progressive injury and pharmacologic response."
The combination of the three US parameters, which the study authors called the LAD-NASH score, had a positive predictive value of 86.5% (32 of 37) (95% CI: 71.2, 95.5) for identifying patients with high-risk NASH and a negative predictive value of 87.5% 87.5% (35 of 40) (95% CI: 73.2, 95.8) for ruling out high-risk NASH.
The area under receiver operating characteristic (AUC) was 0.86 in the Japan derivation group and 0.88 in the Korea validation group.
"Although the two study samples had significantly different distributions of fibrosis, steatosis, and lobular inflammation activity, as well as different incidence rates of high-risk NASH, the LAD-NASH score was judged to be useful in both samples, providing similar levels of diagnostic performance," the study authors wrote. "We therefore believe that the LAD-NASH score is useful for efficiently identifying patients who are suitable for clinical trials and emerging therapies and can also help to reduce the number of unnecessary liver biopsies."
Strengths of the study included the use of a prospective development population and a validation population from another country and the use of standardized US techniques and histopathologic reference standards, Mark E. Lockhart, MD, MPH, of the University of Alabama, noted in an associated commentary.
"By using the combined scoring system, the results of this research infer that 59% of liver biopsies could have been avoided and this represents a higher level of performance than the previously published Fibroscan-AST algorithm," Lockhart wrote.
Limitations include that the LAD-NASH score requires a specific US scanner, skilled personnel and a complicated equation.
Larger, multicenter studies are necessary to validate the results.
“(W)e believe that use of the model combining attenuation coefficient, dispersion slope, and liver stiffness for identification of nonalcoholic steatohepatitis (LAD-NASH) score could lead to the more efficient identification of patients with progressive nonalcoholic steatohepatitis who should be considered for further treatment and the LAD-NASH score should be applied to patients with highly suspicious for liver disease, including nonalcoholic fatty liver disease,” the study authors wrote.