A new AI model developed by UC San Diego researchers could make it possible to predict treatment response based only on a tumor's genetics.

Esophageal squamous cell carcinoma (ESCC) represents a significant global health burden characterized by a high mortality rate, primarily due to uncontrolled tumor proliferation and the prevalence of distant metastases at the time of diagnosis. Despite significant advances, there remains a critical need to identify the underlying molecular mechanisms to establish new therapeutic approaches.

An artificial intelligence–driven transfer learning strategy enabled the discovery of a novel indolopyridine-based small molecule (Compound 8a) that directly targets gp130, potently suppresses the JAK2/STAT3 signaling pathway, and effectively inhibits colorectal cancer growth in vitro and in vivo, offering a promising lead and a feasible computational paradigm for developing gp130‑targeted anticancer agents.

A novel mechanism based on Sanguinarine (SAG) targeting the chaperone protein BiP to synchronously trigger apoptosis and ferroptosis in lung squamous cell carcinoma (LUSC) is provided, and a new strategy for treating this aggressive cancer via Endoplasmic Reticulum Stress (ERS) is opened.

A novel mechanism based on targeting MARK2 to destabilize mutant p53 (mutp53) and suppress triple-negative breast cancer (TNBC) progression, independent of its kinase activity, is provided, and a new strategy for treating mutTP53-driven TNBC is opened.

Researchers at Nanjing University developed platinum(IV)-antibody conjugates that enhance tumor immunogenicity without relying on strong cancer cell killing. The targeted system delivers low-dose platinum selectively to tumors, increasing MHC-I expression and improving immune recognition while reducing systemic toxicity. In mouse models, the strategy significantly enhanced the efficacy of anti-PD-1 immunotherapy.