The Tumorigenic dose 50 (TD50) is a widely used measure of carcinogenic potency representing the dosage at which 50 % of animals exposed to a compound will develop tumours. The popularity of the TD50 is in part due to the large amount of publicly available data. TD50s were calculated for a large number of compounds in the Carcinogenic Potency Database, which has since been extended in the freely available Lhasa Carcinogenicity database, containing TD50s from over 7500 studies and 1700 compounds. However, due to the age of these studies many are of low quality, often comprising only a single test dosage, therefore raising questions about the applicability of such TD50 sfor toxicological risk assessment. We investigate whether the lower 99 % confidence interval is sufficient to produce conservative TD50 estimates for these studies.