Borderline Range Determined Using Data From Validation Study of Alternative Methods for Skin Sensitization: ADRA,...

Area of Science:

• Toxicology and Chemical Safety
• In vitro testing methodologies
• Regulatory science

Background:

• Predictive models for alternative methods often use binary classifications with cutoff values.
• Data variability near cutoff values can lead to ambiguous results in alternative testing.
• The OECD GL497 guideline employs a borderline range (BR) statistical method to manage this ambiguity.

Purpose of the Study:

• To clarify the specific statistical methods used for calculating borderline ranges (BRs) as per OECD GL497.
• To provide a detailed understanding of BR calculation for improved application of alternative testing methods.
• To validate the adequacy of calculated BRs by comparing them with established methods.

Main Methods:

• Utilized the scale-constant correction method to compute median absolute deviation (MAD) around the median.
• Applied the bottom-raised transformation method to handle zero values, enabling logarithmic data transformation for BR calculation.
• Calculated BRs for the amino acid derivative reactivity assay (ADRA), interleukin-8 reporter gene assay (IL-8 Luc), and epidermal sensitization assay (EpiSensA) using validation study data.

Main Results:

• Established specific BRs for ADRA (4.1–5.9) and IL-8 Luc (1.25–1.57).
• Determined BRs for four genes in EpiSensA: ATF3 (10.71–21.02), GCLM (1.64–2.45), DNAJB4 (1.61–2.52), and IL-8 (3.11–5.16).
• The deviations between BR limits and cutoff values were comparable to those of guideline methods (DPRA, KerarinoSens, h-CLAT).

Conclusions:

• The clarified methods for calculating BRs provide a robust approach to managing data variability in alternative testing.
• The established BRs for ADRA, IL-8 Luc, and EpiSensA are adequate and comparable to existing regulatory methods.
• This work enhances the reliability and applicability of alternative methods in toxicological assessments, supporting the reduction of animal testing.