Cancer cachexia atlas links weight-loss episodes to clinical and genomic patterns
Published in Bioinformatics, Diseases!, Journals.
Published October 18, 2026.
A Nature Medicine study published October 12 uses routine clinical records and tumor sequencing to map cancer-associated wasting over time. Researchers analyzed 52,236 patients treated at Memorial Sloan Kettering Cancer Center, with data spanning 2010–2024.
The team identified 30,938 distinct episodes meeting consensus weight-loss criteria for cachexia, affecting 46.63% of the cohort. Episodes often occurred near cancer progression and were associated with poorer survival. Their frequency varied substantially between cancer types.
Across malignancies, these periods showed recurring blood-test changes, including lower albumin and hemoglobin and higher liver enzymes. Certain tumor mutations, notably TP53 mutations in several cancers, were associated with increased risk.
The researchers also developed models for impending weight loss in lung, colorectal and prostate cancer, with external validation using a Dana-Farber cohort of 5,785 patients.
The findings suggest that information already collected during care could support earlier recognition and research on wasting. However, this observational analysis cannot establish that the associated mutations cause cachexia, or that using the prediction models improves outcomes. Its episode definitions track weight loss rather than directly measuring every component of muscle wasting.
Source: Nature Medicine, original research published October 12, 2026.