This conversation between Scott Friedman, Laurent Castera, Louise Campbell and Roger Green covers what the panelists found exciting, important or striking about the presentations covering NITs, artificial intelligence and bariatric surgery at TLM2023. The conversation starts with Roger discussing a presentation in which Stephen Harrison shared HistoIndex data from the MAESTRO-NASH trial for resmetirom. HistoIndex’s proprietary methods produced results suggesting 34% improvement in the placebo group vs. 54% in the treatment group. More interesting, Dean Tai from HistoIndex suggested to Roger in a private conversation that this 1/3 placebo success rate holds fairly common across trials using these methods. Given that bariatric surgery studies suggest a five-year gap from successful surgery to fibrosis regression, the one-year standard requested by FDA for conditional approval might lead us to understate drug efficacy for resmetirom and later New Drug Applications submitted under these criteria.
Laurent discusses a study from Phillipe Mathurin’s group in France that demonstrates among bariatric surgery patients, MASH resolution and fibrosis regression lead to improved outcomes. His point: it is important that this study confirms the idea that MASH resolution and fibrosis regression do, in fact, lead to fewer Major Adverse Liver Outcomes, or MALO.
The conversation shifts to NITs and the value of biomarkers in proving that a drug is working. Laurent notes that no single market is sufficiently reliable and that we will need a cluster of tests to evaluate any given drug. As an example, he cites the impact of BMI on liver stiffness as a confounder that does not exist in other tests like ELF. Scott asks about HepQuant; Laurent has limited experience but finds the idea conceptually appealing. Scott notes that the slow rollout and low budget might be responsible for the lack of market and academic interest.
Laurent focuses on a specific poster looking at a database of 16,000 patients over two years. The study shows that an increase in liver stiffness, AGILE-3 or AGILE-4 score all correlate with MALOs. What the study doesn’t address, and Laurent points up will be key to learn, is whether declines in these measures correlate with improved outcomes among declining patients.