This study will include 1500 patients with advanced lung, pancreatic, or colorectal cancer at high risk of cachexia for detailed measurement via biospecimen collection, remote monitoring, questionnaire, and medical record review. The aims are to: identify distinct cachexia subtypes with data-driven clustering; operationalize cachexia subtype definitions, understand their severity, and examine how each subtype is associated with time to death, functional decline, and/or treatment tolerance.; inform treatment decisions for patients likely to progress to cachexia; and validate mechanistic hypotheses and biomarkers developed in related CanCan Projects.