Technology

AI-developed cancer drug shows effectiveness in human trials for the first time

An AI-developed anti-cancer drug has shown effectiveness in human trials for the first time. In 28% of participants who took the maximum oral dose, tumors shrank by more than a third.

The drug IAM1363, created entirely with artificial intelligence, has shown effectiveness against cancer for the first time in clinical trials without causing serious side effects. The American biotechnology company that developed it, Iambic Therapeutics, presented the results of the I/Ib phase in a press release.

The sample included 64 people who had unsuccessfully undergone several courses of treatment with various drugs. In 5 of the 18 (28%) subjects who received the maximum oral dose of 960 mg, tumors shrank by more than 30%. The drug proved effective against lung, stomach, and breast tumors, as well as rare forms of cancer. Among participants with tumors that had spread to the brain, the effect was observed in one in three.

As noted, machine learning technologies made it possible to discover the active ingredient and conduct animal studies in less than two years. This usually takes four to seven years.

IAM1363 blocks HER2, which normally controls cell division. However, when this protein mutates or is overproduced, cells begin to multiply uncontrollably, causing cancer to develop. Most often, this occurs in the breast (15-30% of cases), stomach (up to 34%), gastroesophageal junction (29-31%), salivary ducts (up to 43%), gallbladder (up to 31%), and less often in the bladder (up to 18%), ovaries, uterus (6%), lung (2-4%), esophagus, intestine, pancreas, and prostate gland.

It is also reported that the drug can cross the blood-brain barrier, a filter in the brain that prevents most drugs from entering the bloodstream. This is particularly important for the treatment of HER2-positive tumors, which spread to brain tissue in 30-50% of cases in the later stages.

Iambic was founded in 2020. The startup is developing two more drugs using neural networks: the first is for tumors with chromosomal instability (when cells lose or gain extra chromosomes during division; human trials are scheduled to begin before the end of this year), and the second for the treatment of solid tumors (preclinical trials have not yet been completed). Since last year, the company has raised more than $150 million from Mubadala Capital, NVIDIA, Abingworth, Coatue, and the Qatar Investment Authority.

 

Source: Business Wire