This gets cited here weekly, usually second-hand, so it is worth setting out what it does and does not establish.
The liver data is among the strongest non-weight findings in the class. The semaglutide MASH programme reported a large advantage over placebo on MASH resolution, with a substantial minority also achieving fibrosis improvement — and fibrosis is the endpoint that predicts outcomes. Mechanistically it is reduced hepatic lipogenesis, increased fatty-acid oxidation, less hepatic inflammation, and possibly a direct effect on stellate-cell activation.
Where I think it is weakest: the population was selected and supported in ways a real cohort is not, so I would read the effect size as a ceiling rather than an expectation.
What would genuinely help is knowing whether normalised enzymes tell you anything about fibrosis, and what the right follow-up measurement is. Not looking for reassurance. Looking for the part I have got wrong.
Note on sourcing:
Figures above are from the primary publication rather than the press summary. If a number here disagrees with one you have, post yours and we will work out which of us is reading a secondary source.