Lea I, Yang L, Rashid A, Fostel JM. Benchmark dose analysis of Chemical Effects in Biological Systems (CEBS) datasets. Abstract 2918-P412. Presented at Society of Toxicology 56th Annual Meeting, Baltimore, MD, March 2017.
Abstract
The Chemical Effects in Biological Systems (CEBS: http://cebs.niehs.nih. gov) database is a unique resource that compiles individual animal data from the National Toxicology Program (NTP) testing program. Guided Searches have been built to assist users to find specific types of data in CEBS. One Guided Search, the ‘Look Up BMD Values’ Guided Search tool enables users to review the dose response relationships between exposure and histopathology findings in NTP studies with statistically significant incidences of neoplastic or non-neoplastic lesions. Benchmark dose (BMD) analysis is a method for assessing the risks associated with potentially hazardous compounds. Software (BMDS) developed by the US EPA provides a method to estimate the dose of a chemical associated with a specific response. The CEBS Look Up BMD Values allows the user to find pre-computed BMD values for significant findings from chronic studies. The CEBS Look Up BMD Values tool was developed using BMDS and provides access to dose response analysis of NTP toxicity data. Users select the study of interest by reviewing a list of key study components such as chemical name, CASRN, and species. Once a study has been selected, all statistically significant neoplastic and non-neoplastic findings for the study are shown. Users are able to evaluate lesion incidence in each dose group for each organ, morphology and sex combination by selecting the View Study Data button. The data are fitted to the multistage model for benchmark dose analysis of dichotomous data with the original and Poly-3 adjusted datasets analyzed. Poly-3 adjustment takes into account the differences in survival times of animals when comparing tumor incidence rates. Selection of a specific organ – lesion combination produces a table of BMD, BMDL (Benchmark Dose Lower Confidence Limit), BMR (Benchmark Response) and AIC (Akaike’s Information Criterion) values that enables the user to assess the models and select the one that provides an adequate fit for the dataset. A fitted dose-response curve with data points and standard error bars is also plotted with identification of BMD and BMDL values. The Look Up BMD Value search tool can be accessed at: http://cebs.niehs.nih.gov/multistage/.