Non-invasive tests in metabolic dysfunction-associated steatotic liver disease: the LITMUS prospective multicentre diagnostic accuracy study.
Abstract
BACKGROUND
The advent of regulatory-approved pharmacological treatments for metabolic dysfunction-associated steatotic liver disease (MASLD) calls for precise, non-invasive biomarkers that facilitate early diagnosis, support therapy selection, and aid clinical trial enrolment. The European multicentre LITMUS study aimed to evaluate the diagnostic performance of a broad range of non-invasive biomarkers for identifying patients with metabolic dysfunction-associated steatohepatitis (MASH), at-risk MASH, advanced fibrosis, and cirrhosis.
METHODS
In this prospective, multicentre, diagnostic accuracy study, adults with biopsy-confirmed MASLD were prospectively recruited across 13 countries in Europe and the USA. We assessed 12 single markers and 18 multimarker panels. The reference standard was centrally scored liver histology according to the non-alcoholic steatohepatitis Clinical Research Network Scoring System. Target conditions included MASH (Non-Alcoholic Fatty Liver Disease Activity Score ≥4 with ≥1 point per component), at-risk MASH (MASH plus significant fibrosis [fibrosis stage F≥2]), advanced fibrosis (F≥3), and cirrhosis (F4). Performance was evaluated by area-under-the-receiver operating characteristic curve (AUC), with the minimum acceptable criterion (MAC) of AUC higher than 0·80 to indicate potential biopsy replacement.
FINDINGS
Between Jan 1, 2018, and June 30, 2023, we analysed data from 1590 participants with centrally scored liver biopsies and at least one biomarker measurement; 609 (38%) had biopsy-confirmed at-risk MASH, 678 (43%) had MASH, 585 (37%) had advanced fibrosis, and 164 (10%) had cirrhosis. For detecting at-risk MASH, SomaSignal showed the highest diagnostic performance (AUC 0·83 [95% CI 0·80-0·85]), significantly exceeding MAC, whereas no other test reached the MAC for MASH only. For advanced fibrosis, SomaSignal showed the highest performance (AUC 0·92 [0·90-0·94]), followed by FibroMeter-VCTE (AUC 0·87 [0·83-0·92]), Agile 3+ and TSP2 (both AUC 0·84 [0·82-0·86]), and NIS4 (0·82 [0·80-0·85]). For cirrhosis, FibroMeter-VCTE showed the highest diagnostic performance (AUC 0·94 [0·90-0·97]) followed by SomaSignal (AUC 0·92 [0·90-0·94]), followed by Agile 3+ (0·88 [0·85-0·91]), Agile 4 (0·87 [0·84-0·90]), ADAPT (0·85 [0·81-0·88]), and ELF, LSM-VCTE, and TSP2 (all 0·84 [0·81-0·87]). Sequential two-step approaches using gamma-glutamyl transferase as a first-line test met the MAC for both advanced fibrosis and at-risk MASH.
INTERPRETATION
Although single markers resulted in variable and limited performance when compared to histology for detecting at-risk MASH, multimarker panels showed great potential, with a proteomic panel showing strong diagnostic performance. These findings could support the selection and implementation of non-invasive testing strategies in MASLD and inform future clinical guidelines, regulatory evaluation, and clinical trial design.
FUNDING
Innovative Medicines Initiative 2 Joint Undertaking.
