Background: Male androgenetic alopecia (MAGA) is the most prevalent form of hair loss in men, affecting approximately 50% of the male population by the fifth decade of life. The individual natural history is highly variable, and no validated composite prognostic instrument currently exists to stratify patients by expected disease trajectory. Surgical hair-restoration is frequently undertaken without prior quantification of individual progression risk.

Objective: To develop and clinically validate the MAGA e-Score (Male Androgenetic alopecia Evolution Score), a quantitative prognostic algorithm integrating age at onset, Hamilton-Norwood baseline staging, and three standardized trichoscopic parameters, aimed at predicting the seven-year natural progression of untreated male AGA. As a first-generation derivation study, external prospective validation in larger multicentre cohorts is planned as the direct continuation of this work.

Methods: A retrospective single-centre longitudinal observational study (STROBE-compliant) was conducted on 32 male patients with confirmed MAGA (Hamilton-Norwood stages I-V), aged 18-30 years at baseline (2016-2018), reassessed after exactly seven years (2023-2025) without any intervention. The MAGA e-Score integrates three weighted domains: age at baseline (0-3 points), Hamilton-Norwood stage (1-3 points), and trichoscopic parameters — miniaturization index, hair shaft diameter variability, and follicular unit density (0-6 points; maximum total 12 points; four prognostic classes). Predictive performance was assessed by overall accuracy, Cohen's kappa (k), and per-class F1-score.

Results: Overall accuracy 92%; Cohen's kappa 0.888 (95% CI 0.812-0.964; p < 0.001), reflecting near-perfect agreement between predicted and observed severity classes. Mean precision, recall, and F1-score each exceeded 0.90. All misclassifications involved adjacent severity classes only.

Conclusions: The MAGA e-Score is a reproducible, clinically deployable first-generation prognostic instrument for male AGA, with direct applications in surgical planning and personalized therapy stratification. Prospective multicentre validation with ethnically diverse cohorts and integration with AI-assisted trichoscopic analysis represent the priority next steps.