Comparing Clinical vs Histopathological Features in Diagnosing Erythemato-Squamous Diseases.

Area of Science:

• Dermatology
• Medical Informatics
• Pathology

Background:

• Erythemato-squamous diseases (ESDs) present overlapping clinical and histopathological features, complicating diagnosis.
• Current diagnostic methods require improvement by integrating diverse patient data.

Purpose of the Study:

• To evaluate the diagnostic value of clinical versus histopathological features in differentiating six classes of ESDs.
• To assess the predictive accuracy of integrated diagnostic models.

Main Methods:

• Analysis of the University of California, Irvine (UCI) Dermatology dataset (366 patients).
• Paired t-tests to compare clinical and histopathological feature severity.
• Multiple logistic regression (MLR) models to determine predictive strength of features for each ESD class.

Main Results:

• Clinical features showed higher average severity across ESDs, but diagnostic dominance varied by disease.
• Multiple logistic regression models achieved high predictive performance for all six ESD classes.
• Pityriasis rubra pilaris demonstrated the highest predictive accuracy (94.5%); lichen planus showed stronger prediction from histopathology.

Conclusions:

• Both clinical and histopathological data are crucial and complementary for diagnosing ESDs.
• Integrated diagnostic models can enhance classification accuracy and support data-driven dermatological strategies.