New approaches target KRAS‑mutant cancer relapse and AI improves tumor variant detection

Researchers at Chiba University in Japan discovered that drug‑tolerant persister cells arising after KRAS‑inhibitor treatment in non‑small cell lung cancer and pancreatic ductal adenocarcinoma become dependent on glutamine metabolism and lysosome‑associated functions. Blocking these pathways may eliminate residual cancer cells and prevent disease relapse, said associate professor Shigeki Aoki.

In parallel, Google Research with the University of California, Santa Cruz and other partners released DeepSomatic, an open‑source convolutional neural‑network tool that accurately identifies somatic small‑variant mutations in tumor sequencing data across multiple platforms. The system can operate with or without matched normal samples, offering a flexible solution for precision oncology.

Both developments aim to enhance therapeutic strategies and diagnostic accuracy for cancers driven by genetic mutations.