PreS MD

PreS MD predicts sensitization hazard of chemical substances released from medical devices by modeling guinea pig maximization test (GPMT) outcomes to assess allergenicity of extractables and leachables for safety evaluation.

Key Features:

• Regulatory relevance: Addresses toxicity evaluation of extractables and leachables from medical devices and their potential to induce sensitization as measured by the guinea pig maximization test (GPMT).
• Computational approach: Predicts GPMT outcomes using in silico models to provide an alternative to animal testing for sensitization assessment.
• Data-driven model development: Built from the largest publicly available GPMT dataset and implements QSAR models with machine learning algorithms, including deep learning, achieving balanced accuracy of 70%–74% validated by 5-fold cross-validation and external testing on novel compounds.

Scientific Applications:

• Sensitization hazard assessment: Predicts the allergenic potential of chemical substances released from medical devices, including extractables and leachables.
• Regulatory safety evaluation support: Provides computational evidence relevant to pre-market safety assessments and regulatory decision-making regarding sensitization risk.
• Reduction of animal testing: Supports alternatives to traditional GPMT assays by enabling in silico prediction of sensitization outcomes.

Methodology:

Collection and curation of a large publicly available GPMT dataset; development of QSAR models using machine learning and deep learning algorithms; validation via 5-fold cross-validation and testing on novel compounds reporting balanced accuracy of 70%–74%.