Tracking the evolution of NAFLD through a non-invasive lipidomic test that accurately discriminates NASH from steatosis

Objectives Nonalcoholic fatty liver disease (NAFLD) includes a spectrum of histological phenotypes including steatosis, steatohepatitis (NASH) and fibrosis. While liver biopsy is the reference for diagnosis, it is invasive and associated with procedural risks and sampling variability. There is urgent need for a noninvasive diagnostic procedure. Recently, we have described a serum-based lipidomic signature associated with NAFLD able to fulfill these unmet diagnostic needs by: differentiating NAFLD from healthy cohort; discriminating between steatosis and NASH. The aim of this study is to validate this non-invasive assay using blind-histology as a reference standard and apply it in the follow-up of the patients. Method Thirty patients were enrolled as a blind, biopsy-proven NAFLD cohort, collecting serum at the time of liver biopsy. Patients were prescribed hypocaloric diet (1500kcal/day) and aerobic exercise (30-60min/day), monitored for 2-5 years, when a serum sample was collected. Metabolic syndrome was assessed based on the presence of at least three of the conditions listed by the NCEP ATPIII. The lipidomic test was based on 467 biopsy-proven patients and two BMI-dependent logistic regression algorithms discriminating between: NAFLD-healthy liver (assay name: OWLiver Care); and NASH-steatosis (OWLiver). The diagnostic performances of both assays were assessed by the ROC curve: 0.90±0.02 and 0.95±0.01, respectively. Results Applied to the independent biopsy-proven cohort (33%female, weight=86±15kg; BMI=32±5kg/m 2 ), the test diagnosed correctly 28 out of 30 patients, misclassifying one patient having NASH with NAS score=2, but presenting metabolic syndrome; and one patient as having steatosis with NAS score=5, although without ballooning. Once validated, the test was applied to the follow-up of the patients (weight=84±14kg; BMI=31±4kg/m 2 ). 31% of the patients lose at least 5% of baseline body weight. Among those responders, 50% improved their diagnosis presenting positive post-interventional shifts from NASH to steatosis or steatosis to healthy liver. The original diagnosis remained unchanged for the 95% of the non-responder patients. Conclusions The results obtained in the independent cohort support the feasibility of these lipidomic tests as a noninvasive tool for NAFLD diagnosis and to monitor the disease progression/regression while circumventing the need for repeat liver biopsy.