Phenotyping and hierarchical clustering in genetically unresolved DSD: identifying clinical signatures to prioritize genomic discovery

Chrysanthi Kouri; Malika Allimussina; S Faisal Ahmed; Christa E Fluck

doi:10.1530/endoabs.118.OC1.2

OC1.2

Prev Next

Section Contents Cite

Endocrine Abstracts (2026) 118 OC1.2 | DOI: 10.1530/endoabs.118.OC1.2

IDSD2026 Oral Communication Abstracts Session 1 (7 abstracts)

Phenotyping and hierarchical clustering in genetically unresolved DSD: identifying clinical signatures to prioritize genomic discovery

Chrysanthi Kouri ^1,2 , Malika Allimussina ³ , S Faisal Ahmed ^3,4 & Christa E Flück ^1,2

Author affiliations

¹Pediatric Endocrinology, Diabetology and Metabolism, Department of Pediatrics, Inselspital, Bern University Hospital, University of Bern, 3010 Bern, Switzerland; ²Department for BioMedical Research, University of Bern, 3008 Bern, Switzerland; ³Office for Rare Conditions, University of Glasgow, Glasgow G51 4TF, United Kingdom; ⁴Developmental Endocrinology Research Group, University of Glasgow, Royal Hospital for Children, Glasgow G51 4TF, UK. Correspondence to: [email protected]

Background: Individuals with genetically unsolved differences of sex development (DSD) are typically treated based on clinical and biochemical parameters related to their symptoms and signs of DSD, and other (related) health issues. However, having a clear genetic diagnosis may be essential for understanding the hereditability of the condition and predicting outcomes. A considerable subset of individuals with DSD lack an identifiable genetic aetiology. Our aim was to characterize phenotypic patterns of these unresolved cases and assess whether their multisystem anomalies resemble established genetic DSD and form distinct clusters indicative of previously unrecognized mechanotypes.

Methods: We analysed existing data on the the SDMregistrires platform for individuals with 46,XY or 46,XX DSD who lacked a molecular diagnosis. Our approach combined two complementary strategies: 1) unsupervised hierarchical clustering (Gower distance) to identify phenotypic subgroups, and 2) supervised phenotype-guided classification using reference fingerprints for NR5A1/SF-1 and POR (PORD) deficiency.

Results: Among 1,391 individuals screened, 345 (24.8%) met inclusion criteria based on the presence of additional organ anomalies. The cohort was predominantly 46,XY (83.8%). The most prevalent anomalies were small for gestational age (28.1%), cardiac (21.4%), CNS (15.7%), and renal (14.5%). Unsupervised clustering resolved five phenotypic groups, including a novel CNS-dominant cluster (10% of unclassified cases) and a craniofacial-dominant group. Phenotype-guided classification showed that 11% of cases displayed an SF-1-like pattern (blood/CNS/adrenal involvement), while 22% exhibited a PORD-like profile (adrenal/skeletal/craniofacial involvement).

Conclusions: Genetically unresolved DSD frequently presents in recognizable phenotypic clusters mimicking established syndromes or suggesting novel or known underlying regulators and genes. Systematic phenotyping is essential for identifying known signatures such as SF-1-like or PORD-like and provides a roadmap for targeted genomic re-evaluation, including searching for regulatory variants or related genes. It allows to bridge the diagnostic gap, refine clinical classification, and support individualized management.

Acknowledgements: We would like to acknowledge the help of all the centres that have contributed cases to the I-DSD Registry.