SARA-ICE skin sensitization prediction tool now available

A web tool for quantitative prediction of a chemical’s potential to cause skin sensitization in humans is now available on the NTP website. The Skin Sensitization Risk Assessment – Integrated Chemical Environment (SARA-ICE) defined approach (DA) is a Bayesian statistical model developed in a collaboration between NICEATM and the consumer products company Unilever. The new web tool is an open-access resource that anyone can use to predict skin sensitization potential from an uploaded data set. It is available at https://ntp.niehs.nih.gov/whatwestudy/niceatm/test-method-evaluations/skin-sens/da/SARA-ICE.

There is an international need for non-animal approaches to identify potential skin sensitizers. While DAs are accepted for making a binary prediction of whether or not a substance might be a skin sensitizer, for some regulatory applications a quantitative prediction of skin sensitizer potency is needed. SARA-ICE is a Bayesian statistical model that estimates a human-relevant metric of skin sensitizer potency. This metric, termed ED01, is the dose with a 1% chance of human skin sensitization. SARA-ICE uses data on over 400 chemicals from the NICEATM Integrated Chemical Environment to predict the ED01 using any combination of in vivo (human predictive patch test or local lymph node assay) and in vitro (direct peptide reactivity assay [DPRA], kinetic DPRA, KeratinoSens™, human cell line activation test, or U-SENS™) data. SARA-ICE has been accepted by the Organization for Economic Co-operation and Development (OECD) as one of the defined approaches for skin sensitization described in its Guideline 497.

The SARA-ICE DA was developed in a collaboration between NICEATM and the consumer products company Unilever. Detailed information about SARA-ICE was published in a peer-reviewed paper (Reinke et al.; https://doi.org/10.1016/j.crtox.2024.100205).