Abstract: Methods for predicting a survival metric to aid health care providers in determine a first line treatment are described. The methods may comprise, for example, receiving sequence read data associated with a sample from the individual, selecting a plurality of sequence reads from the sequence read data, determining one or more characterization features based on the plurality of sequence reads, receiving one or more digital pathology features, inputting, using the one or more processors, the one or more characterization features and the one or more digital pathology features into one or more statistical models, and predicting a survival metric indicative of a likelihood of survival of the individual, based on outputs of the one or more statistical models.

Abstract: Methods for selecting an anti-cancer therapy for or treating a subject having cancer with an anti-cancer therapy comprising determining risk scores that predict the likelihood of response to the treatment are described. The methods may comprise, for example, obtaining genomic data comprising aneuploidy status or loss of heterozygosity data for one or more subgenomic intervals in a sample from the subject; analyzing the genomic data for the subject using a model configured to receive genomic data comprising aneuploidy status or loss of heterozygosity data for the one or more identified subgenomic intervals in the subject and output a risk score for the subject, wherein the risk score predicts the subject's response to a selected treatment. Also described are biomarkers for specific diseases (e.g., metastatic pancreatic cancer) and methods for treating subjects having cancer based on the determined risk scores.