There is an enormous gap between the antiproliferative and in vivo antitumor efficacy of gemcitabine in cell line-based models and its clinical efficacy. This may be due to insensitiveness of the precursor, cancer stem cell (CSC) compartment to cytotoxic agents. The hedgehog pathway is associated with CSC signaling and control. We used a direct xenograft model of pancreatic cancer and a two-stage approach was used to test the hypotheses that targeting CSC could increase the efficacy of gemcitabine. Tumors from a gemcitabine-sensitive xenograft were treated with gemcitabine first, and randomized, after tumor regression to continuing treatment with gemcitabine, a hedgehog inhibitor alone or in combination with gemcitabine. We tested markers described as associated with CSC such as CD24, CD44, ALDH, nestin, and the hedgehog pathway. After induction with gemcitabine, treated tumor showed an enrichment in CSC markers such as ALDH and CD24. Subsequently, a release from gemcitabine prompted a repopulation of proliferating cells and a decrease in such markers to equilibrate from pretreatment levels. Combined treatment with gemcitabine and cyclopamine induced tumor regression and decrease in CSC markers and hedgehog signaling. Cytoplasmic CD24 and ALDH were inversely and strongly associated with growth and were expressed in a minority of cells that we propose constitute the CSC compartment. Hedgehog inhibitors as part of a dual compartment therapeutic approach were able to further reduce tumor growth and decreased both static and dynamic markers of CSC. Direct tumor xenografts are a valid platform to test multicompartment therapeutic approaches in pancreatic cancer.