The Human Use Medicines Committee (CHMP) of the European Medication Agency (EMA) has issued the first qualification opinion (OC) on an innovative development methodology based on artificial intelligence (AI). The tool, called AIM-NASH, is developed by the American Path AI and helps pathologists to analyze liver biopsies to identify the severity of non-alcoholic steatohepatitis (MASH) in clinical trials.
The Mash is a condition in which fat accumulates in the liver, causing inflammation, irritation and healing over time, without significant consumption of alcohol or other causes of liver damage. The Mash is related to obesity, type 2 diabetes, arterial hypertension, high levels of cholesterol and abdominal fat. If not, it can lead to an advanced liver disease.
The AIM-NASH tool is expected to improve the reliability and efficiency of clinical trials for new MASH treatments by reducing variability in the measurement of disease activity (inflammation and fibrosis).
After a public consultation, the CHMP issued an opinion to qualify this method, which means that the Committee can accept the evidence generated by the tool as scientifically valid for future applications. The CHMP agreed that the tool can increase the reproducibility and repeatability in the evaluations of new MASH treatments. It can help researchers obtain clearer evidence about the benefits of new treatments in clinical trials with less patients. Ultimately, this may allow patients to access effective treatments more quickly.
The tests of new treatments for the MASH are usually based on liver biopsies, on which small fragments of liver tissue are taken to confirm inflammation and healing. These biopsies are the reference method to demonstrate the effectiveness of new drug medications. However, the high variability in the clinical trials of Mash/Nash is a challenge, since the specialists who review the biopsy samples do not always coincide in the severity of inflammation or healing.
The evidence presented to the CHMP shows that the AIM-NASH biopsy readings, verified by an expert pathologist, can reliably determine the activity of the MASH disease with less variability than the current standard used in clinical trials, which is based on a consensus of three independent pathologists.
AIM-NASH is a IA-based system that uses an automatic learning model trained with more than 100,000 annotations of 59 pathologists who evaluated more than 5,000 liver biopsies in nine large clinical trials.
The qualified tool is blocked, which means that the automatic learning model cannot be modified or replaced. The CHMP encourages the optimization of the model, recognizing which important changes may require the requalification of the tool.
All the activities of the EMA in the field of AI are coordinated within the framework of the Plurianual Work Plan on the ema and the heads of the medication agencies, with the objective of guaranteeing a safe and responsible use of the AI in the entire regulatory network of European drugs.
#EMA #enables #tool #diagnose #hepatic #inflammatory #disease
