AI in Medicine and the Future of Drug Discovery: The Case of ISM001-055

Artificial Intelligence has completely changed every field of research, especially in the field of drug discovery. Despite developing a new medicine on average taking around 14 years and billions of euros being spent on research, Insilico Medicine in only 30 months, through the use of AI, discovered Rentosertib (ISM001-055), a drug for idiopathic pulmonary fibrosis, which is a chronic disease that develops when the lung tissue becomes thicker and stiffer for unknown reasons. These changes can cause scarring to the lungs, which makes it progressively harder to breathe properly and causes 40,000 deaths a year in America alone.

What makes this truly a breakthrough is how both the target and the drug molecule were completely designed and identified by AI, replacing the usual years of trial and error it would take researchers to do. In an early stage of studying the medicine with 71 patients, it was shown that the lung function depended on the dose, with higher doses showing a better effect. The group that received the highest dose could breathe better after about 12 weeks, gaining about 98 mL of lung capacity (Source 1, Figure 1).

How does it work?
Insilico used two main AI systems to design the drug: PandaOmics, which analyzes a large amount of biological data to identify which new proteins or “targets” are likely to cause diseases, and Chemistry42, a generative AI that designs molecules to fight the targeted protein effectively while considering that it must be safe and possible to be made in a lab.

These are essential, as they allow for drug development to be considerably quicker and cheaper, and allow for new possibilities in the medical field by enabling a deeper understanding of disease and the creation of treatments that were previously too complex or costly to develop and study.

Figure 1 shows how patients’ lung function improved depending on the dose they received. The groups taking 30 mg QD, 30 mg BID, and 60 mg QD of Rentosertib all showed better lung capacity compared with the placebo group (Source 1).

Figure 2 shows how AI can speed up drug discovery. Traditional drug development usually takes 10 to 15 years, costs over $2.5 billion, and only has about a 10% success rate. With AI, the process can take just 2 to 5 years, cost around $0.5 to 1 billion, and reach a 30% success rate, showing how AI can save time, money, and reduce the number of failed drugs (Source 2).

With smarter algorithms and more biological data, we may see personalized medicines made specifically for individual patients, faster responses to emerging diseases, and entirely new ways to understand and treat illnesses. And despite AI currently being used mainly as a tool by researchers, it is and will quickly reshape the future of healthcare.

Bibliography:

National Heart, Lung, and Blood Institute. “Idiopathic Pulmonary Fibrosis.” NHLBI, U.S. Department of Health and Human Services, https://www.nhlbi.nih.gov/health/idiopathic-pulmonary-fibrosis. Accessed 2 Dec. 2025.

Insilico Medicine. “Insilico Medicine Announces Phase 2a Trial Results for TNIK Inhibitor in IPF.” Insilico, https://insilico.com/news/tnik-ipf-phase2a?utm. Accessed 2 Dec. 2025.

Zhavoronkov, Alex, et al. “A Generative AI-Discovered TNIK Inhibitor for Idiopathic Pulmonary Fibrosis: A Randomized Phase 2a Trial.” PubMed, https://pubmed.ncbi.nlm.nih.gov/40461817/. Accessed 2 Dec. 2025.

Pharmaphorum. “Insilico Raises $110M to Expand AI-Designed Drug Pipeline.” Pharmaphorum, https://pharmaphorum.com/news/insilico-raises-110m-ai-designed-drug-pipeline?utm. Accessed 2 Dec. 2025.

Insilico Medicine. “Insilico Announces Nature Medicine Publication of Phase IIa Results of Rentosertib, the Novel TNIK Inhibitor for Idiopathic Pulmonary Fibrosis Discovered and Designed with a Pioneering AI Approach.” Insilico, 2025, https://insilico.com/tpost/tnrecuxsc1-insilico-announces-nature-medicine-publi. Accessed 2 Dec. 2025. Insilico Medicine

Dunford, Maria Chatzou. “AI-driven Drug Discovery: Complete Guide.” Lifebit.ai Blog, 2025, https://lifebit.ai/blog/ai-drug-discovery-complete-guide/. Accessed 2 Dec. 2025.