Send to

Choose Destination
Arch Toxicol. 2015 Dec;89(12):2355-83. doi: 10.1007/s00204-015-1634-2. Epub 2015 Nov 26.

Bayesian integrated testing strategy (ITS) for skin sensitization potency assessment: a decision support system for quantitative weight of evidence and adaptive testing strategy.

Author information

Procter and Gamble Company, 1853, Strombeek-Bever, Belgium.
Givaudan Schweiz AG, 8600, Duebendorf, Switzerland.
Procter and Gamble Company, Mason, OH, 45040, USA.
ILS/Contractor Supporting NICEATM, Research Triangle Park, NC, 27709, USA.
Shisheido Company Limited, Tokyo, Japan.
Kao Corporation, R&D Safety Science Research, Tochigi, 321-3497, Japan.


The presented Bayesian network Integrated Testing Strategy (ITS-3) for skin sensitization potency assessment is a decision support system for a risk assessor that provides quantitative weight of evidence, leading to a mechanistically interpretable potency hypothesis, and formulates adaptive testing strategy for a chemical. The system was constructed with an aim to improve precision and accuracy for predicting LLNA potency beyond ITS-2 (Jaworska et al., J Appl Toxicol 33(11):1353-1364, 2013) by improving representation of chemistry and biology. Among novel elements are corrections for bioavailability both in vivo and in vitro as well as consideration of the individual assays' applicability domains in the prediction process. In ITS-3 structure, three validated alternative assays, DPRA, KeratinoSens and h-CLAT, represent first three key events of the adverse outcome pathway for skin sensitization. The skin sensitization potency prediction is provided as a probability distribution over four potency classes. The probability distribution is converted to Bayes factors to: 1) remove prediction bias introduced by the training set potency distribution and 2) express uncertainty in a quantitative manner, allowing transparent and consistent criteria to accept a prediction. The novel ITS-3 database includes 207 chemicals with a full set of in vivo and in vitro data. The accuracy for predicting LLNA outcomes on the external test set (n = 60) was as follows: hazard (two classes)-100 %, GHS potency classification (three classes)-96 %, potency (four classes)-89 %. This work demonstrates that skin sensitization potency prediction based on data from three key events, and often less, is possible, reliable over broad chemical classes and ready for practical applications.


Bayesian network; Integrated testing strategy; LLNA potency class; Skin sensitization

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center