Epidemiological and preclinical data suggest that antioxidants are protective against prostate cancer whose pathogenesis has been linked to oxidative stress. However, the selenium and vitamin E Cancer Prevention Trial (SELECT), found no efficacy for selenium in reducing prostate cancer incidence while vitamin E was associated with an increased risk of the disease. These results have called in to question the models used in preclinical chemoprevention efficacy studies and their ability to predict in vivo outcomes. Chemoprevention agents have traditionally been tested on two dimensional monolayer cultures of cell lines derived from advanced prostate cancers. But as SELECT demonstrates, results from advanced disease models were not predictive of the outcome of a primary chemoprevention trial. Additionally, lack of cell-matrix interactions in two dimensional cultures results in loss of biochemical and mechanical cues relevant for native tissue architecture. We use recent findings in three dimensional organoid cultures that recapitulated the SELECT trial results to argue that the organoid model could increase the predictive value of in vitro studies for in vivo outcomes.