ALADIN (Acne Lesion And Density INdex): a Novel Tool for Automatic Acne Severity Assessment

Authors

Alfonso Medela, Alberto Sabater, Ignacio Hernández Montilla, Taig Mac Carthy, Andy Aguilar, Gerardo Fernández, Antonio Martorell, Sofia Vera Carretero, Constanza Balboni, Javier Martín Alcalde, Jose Luis Ramírez Bellver, Alejandro Martin-Gorgojo, Pedro Rodriguez Jiménez, Jose Luis López Estebaranz.

Summary

Acne severity scales are important for clinical decisions and research, but manual grading is affected by time constraints and interobserver variability. ALADIN combines automatic lesion and density analysis to estimate established acne severity measures from clinical images.

Results

ALADIN performed in line with dermatologist interobserver agreement. For Investigator Global Assessment, Pearson correlation was 0.58 for ALADIN compared with 0.56 between dermatologists on the private dataset, and 0.65 for both on the public dataset. For the Global Acne Grading System, the corresponding correlations were 0.32 compared with 0.44, and 0.43 compared with 0.44.

The authors note that the available datasets were composed primarily of White and Asian patients. Further validation in darker skin types is therefore needed.

Read full text

Published in Skin Health and Disease. Available at https://doi.org/10.1093/skinhd/vzag062.