The development of molecularly targeted therapies for certain types of cancers (e.g., Vemurafenib for advanced melanoma with mutant BRAF; Cetuximab for metastatic colorectal cancer with KRAS wild type) has led to the consideration of population enrichment designs that explicitly factor-in the possibility that the experimental compound might differentially benefit different biomarker subgroups. In such designs, enrollment would initially be open to a broad patient population with the option to restrict future enrollment, following an interim analysis, to only those biomarker subgroups that appeared to be benefiting from the experimental therapy. While this strategy could greatly improve the chances of success for the trial, it poses several statistical and logistical design challenges. Since late-stage oncology trials are typically event driven, one faces a complex trade-off between power, sample size, number of events and study duration. This trade-off is further compounded by the importance of maintaining statistical independence of the data before and after the interim analysis and of optimizing the timing of the interim analysis. This talk will highlight the crucial role of simulation-guided design for resolving these difficulties while nevertheless maintaining strong control of the type-1 error.