Noam Solomon serves as the co-founder and CEO of Immunai. For numerous years, the sector has structured itself based on a solitary inquiry: how do we stumble upon additional, and improved, medication possibilities? The process of identifying targets is becoming more intelligent, while the speed of generating molecules is accelerating.
Additionally, AI is now involved in almost all aspects of early discovery. Although a molecule’s effectiveness depends on the patients it is tested on, and this aspect has not received much investment, we can see this imbalance in the progress made in non-small cell lung cancer treatments last year.
In 2025, seven new approvals were granted by the FDA, all linked to distinct genomic modifications rather than generalized diagnoses. These approvals encompassed various categories such as antibody-drug conjugates, bispecific antibodies, and targeted kinase inhibitors. The wave did not occur due to a modification in the fundamental biology of lung cancer.
The situation occurred because the field now has the resources to recognize that “lung cancer” was never a single disease, and to pair each treatment with the group of patients whose tumors are genuinely influenced by the mechanism it targets. On the other hand, a 2023 JAMA Internal Medicine study of drugs approved by the FDA between 2018 and 2021 discovered that approximately one in ten drugs did not meet its primary effectiveness goal in the critical trial that led to approval.
Despite their initial failure, certain of those medications eventually proved beneficial to actual patients.
