Sagimet adopts AI-based pathology for analysis





Sagimet adopts AI-based pathology for more consistent NASH analysis
In NASH, digital pathology involves the histological imaging of biopsy samples, which are then analyzed by artificial intelligence (AI) to detect changes in fibrosis. Credit: Nephron

Sagimet Biosciences is using digital pathology in its Phase IIb non-alcoholic steatohepatitis (NASH) trial with the goal of obtaining a more consistent biopsy analysis, said CEO George Kemble. In NASH, digital pathology involves the histological imaging of biopsy samples, which are then analyzed by artificial intelligence (AI) to detect changes in fibrosis.

In NASH clinical trials, classifying patients between different stages of fibrosis can be tricky due to pathologist biases. To account for inconsistent biopsy readings between pathologists, some trials have three different experts reviewing samples, with the third pathologist breaking the tie. There are five stages of fibrosis, from F0 (no fibrosis) to F4 (cirrhosis or advanced scarring).

Because of the risk of analysis inconsistencies, the NASH space is starting to embrace digital pathology, says Kemble. Sagimet is working with HistoIndex, a Singapore-based digital pathology company, on the Phase IIb FASCINATE-2 (NCT04906421) trial of TVB-2640. HistoIndex also previously worked with Novartis on its FLIGHT-FXR Phase II trial (NCT02855164) which investigated tropifexor. US companies PathAI and CRO Summit Clinical Research are working together to develop AI-based tools for the assessment of liver disease.

NASH
George Kemble, CEO of Sagimet Biosciences.

That said, Kemble adds that digital pathology currently only complements the analysis performed by pathologists, with FASCINATE-2 collecting data from digital readings and pathologists. FASCINATE-2 would provide an exploratory reading for digital pathology in human volunteers, Sagimet having previously used this approach in mouse models, he notes.

FASCINATE-2 has a primary endpoint of a histological reduction of at least two points in the activity score (NAS) in non-alcoholic fatty liver disease (NAFLD). NAS ranges from 0 to 8 and covers steatosis, hepatocellular bloating, and lobular inflammation.

Sagimet’s asset has an upstream NASH target

Sagimet’s oral candidate TVB-2640 works by inhibiting fatty acid synthase (FASN), which is a key enzyme related to fat synthesis. A diet high in fructose leads to an excessive production of palmitate, which leads to a metabolic cascade that can lead to NASH. Palmitate contains palmitic acid, which is the main type of fatty acid. TVB-2640 essentially blocks the upstream elements of what would trigger fatty liver disease, Kemble explains.

Sagimet is expected to attend the next American Association for the Study of Liver Disease (AASLD) meeting in November. He will present consolidated data from his Phase II trial FASCINATE-1 (NCT03938246), which recruited patients in the United States and China.

The rationale behind the data teasing between the two countries is that there have been questions about the influence of diet and lifestyle in NASH, Kemble notes. So far, the data shows that TVB-2640 has a similar impact on reducing liver fat at three months in the two separate populations, he added. Sagimet worked with its China-based partner Ascletis Pharma on the trial.





Leave a Comment