The combination of the results of mouse antipapilloma tests with those from hypervitaminosis A tests in mice as the basis for calculating a therapeutic index has been used for more than 20 years in the search for retinoids as useful drugs in human dermatology. A number of retinoids identified as active or inactive when administered systemically in these mouse systems have gone into clinical trials; clinical results on 11 retinoids were available for a retrospective analysis on the predictive relevance of the mouse models for retinoid activity in human psoriasis. This analysis revealed that the therapeutic index in mice correctly identified eleven compounds and differentiated them into markedly active moderately active or inactive retinoids when subsequently used clinically in the treatment of various forms of psoriasis. Acidic retinoids were more difficult to assess than nonacidic ones and the therapeutic index appeared to underestimate their potency in humans. One retinoid, motretinide, showed a favorable therapeutic index but failed to demonstrate antipsoriatic activity in the clinic. The reason for this discrepancy is that humans and mice metabolize this compound differently. Thus, although chemically induced skin papillomas in mice reflect only certain analogies to human psoriasis and other keratinizing dermatoses, they may be considered a useful tool in the search for retinoids for the treatment of keratinization disorders.