Unsupervised Phenotypic Stratification of 22q11.2 Deletion Syndrome for Improved Clinical Screening

Latent Class Analysis of 297 suspected 22q11.2 deletion syndrome cases in Brazil identified four phenotypic subgroups with the best separability (silhouette = 0.165).

• Why it matters: Symptom heterogeneity across multiple clinical domains makes early recognition and referral for genetic testing difficult, and prior work lacked unsupervised, within-cohort stratification.
• What they did: Using binary symptom data from the Brazilian Database of Craniofacial Anomalies, the study compared K-medoids, DBSCAN, and Latent Class Analysis, then applied association rule mining and evaluated screening sensitivity/specificity by subgroup.
• The result: The current screening protocol showed adequate overall sensitivity but markedly lower performance for the Neurodevelopmental-psychiatric (C2) and Mild- craniofacial (C3) subgroups, flagging profiles more likely to be missed.

AI summary of the abstract · may contain errors · not medical advice · read the paper.