Establishment and validation of psoriasis evaluation models.

Area of Science:

• Dermatology
• Bioinformatics
• Computational Biology

Background:

• Psoriasis is a prevalent inflammatory skin condition significantly impacting patient quality of life.
• Current diagnostic and assessment methods, like the Psoriasis Area and Severity Index (PASI), have limitations in accurately evaluating lesion severity and treatment efficacy.
• There is a need for reliable and objective tools to assess psoriasis status and response to therapies.

Purpose of the Study:

• To develop and validate novel computational models for assessing psoriasis lesion severity.
• To establish methods for evaluating the therapeutic effects of biologic treatments in psoriasis.
• To identify key genes associated with psoriasis pathogenesis and progression.

Main Methods:

• Utilized Gene Expression Omnibus (GEO) datasets to identify 17 potential model genes.
• Developed six psoriasis evaluation models using LASSO regression, linear regression, and random forest algorithms.
• Validated model performance in classifying psoriatic versus non-lesional skin and assessed correlation with clinical data from biologics-treated patients.

Main Results:

• All six developed models demonstrated high accuracy in classifying psoriatic lesions and non-lesional skin across training and testing datasets, with good Area Under the Curve (AUC) values.
• Model scores positively correlated with lesion severity and negatively correlated with treatment duration in patients receiving biologics, indicating potential for therapeutic effect assessment.
• Expression analysis confirmed increased RNA and protein levels of model genes in cytokine-stimulated keratinocytes, IMQ-induced mouse skin, and human psoriatic lesions.

Conclusions:

• The study successfully established robust computational models for evaluating psoriasis severity.
• These models show promise in objectively assessing the therapeutic efficacy of biologic treatments for psoriasis.
• The identified model genes serve as potential biomarkers for psoriasis and therapeutic response, paving the way for improved clinical management.