Cancer comprehensive genomic profiling tests are increasingly used, but drug response rates remain limited. Simulations forecasting cancer progression could aid targeted therapies; however, existing simulations focus mainly on basic biology. We present tugMedi, a cancer-cell evolution simulator designed for cancer genome medicine. By integrating patient-specific genomic and imaging data, tugMedi reconstructs each tumor\'s genomic features and growth behavior, enabling real-time predictions of clonal dynamics under virtual drug regimens. tugMedi explicitly models copy number alterations and SNVs on parental chromosomes in recessive and dominant modes, capturing loss-of-heterozygosity and yielding precise variant allele frequencies and tumor contents. It handles mutations in cancer-related genes with real exon-intron structures by a fast algorithm. For TCGA samples, it provided ensemble predictions of clonal dynamics, allowing extraction of drug response, tumor-size shrinkage rates, and time to recurrence under specific drug conditions. tugMedi represents a first step toward simulation-driven genome medicine based on a patient-derived virtual tumor.