Computational study and ranking of drugs for peyronie’s disease using QSPR and MCDM techniques

Abstract Peyronie’s disease, a connective tissue disease with fibrotic plaque formation of the penile tunica albuginea, is often associated with penile deformity, pain, and erectile dysfunction. Despite the availability of a number of therapy agents, a comprehensive comparative evaluation of selected therapeutic agents according to structural and physicochemical features is still limited. In this study, an integrated framework has been developed by integrating chemical graph theory, Quantitative Structure–Property Relationship (QSPR) modeling, and Multi-Criteria Decision-Making (MCDM) techniques to analyze and rank twenty drugs used in the treatment of Peyronie’s disease. Five degree-based topological descriptors, the second Zagreb ( $$M_2$$ ), Atom Bond Connectivity (ABC), Geometric Arithmetic (GA), Sombor (SO) and Nirmala (NI) indices were computed based on the molecular graph representations of the selected drugs. Correlation analysis was performed between these descriptors and nine physicochemical properties. Further development and validation of Multiple Linear Regression models were performed using Leave-One-Out Cross Validation (LOOCV), Ridge Regression with nested LOOCV, and Y-randomization testing. Among the tested properties, molar refractivity (MR), polarizability (P), and molecular weight (MW) demonstrated the most consistent QSPR performance. The models yielded $$R^2$$ values of 0.9733, 0.9829, and 0.9768, respectively, with LOOCV $$Q^2$$ values of 0.9435, 0.9547, and 0.9500. Entropy and CRITIC techniques were employed to establish criterion weights for decision-making analysis, and then the hybrid weighting approach was used to amalgamate them. The medications studied were ranked using five MCDM methods, which are MARCOS, VIKOR, WASPAS, TOPSIS and EDAS, followed by Spearman rank-correlation analysis and consensus aggregation using average ranking and Borda count methods. The findings show a repeatable computational framework for comparing molecular compounds based on certain physical properties. The obtained rankings are based on physicochemical profile prioritizing and should not be taken as clinical efficacy or therapeutic superiority.