The present retrospective observational case series included 10 individuals with molecularly confirmed KBGS evaluated at the Clinical Genetics Service of Fundación Valle del Lili (Cali, Colombia) between January 2017 and February 2025. Inclusion criteria comprised a compatible clinical phenotype and the identification of a pathogenic or likely pathogenic ANKRD11 variant or a chromosomal rearrangement involving the 16q24.3 region. Patients without molecular confirmation of KBGS or with insufficient clinical or photographic data for analysis were excluded. Clinical, demographic and molecular data were obtained from medical records and standardized clinical genetic evaluations. The patient group comprised six males and four females, with an age range of 6-37 years at the time of evaluation.

To assess facial dysmorphology in KBGS, the present study included a comparative sample of 10 sex- and age-matched controls recruited during data collection sessions at Universidad Icesi and the Clinical Research Center of Fundación Valle del Lili, Cali, Colombia, between January and March 2021. The control group comprised six males and four females, with an age range of 6-37 years. Control individuals were included only for the 2D facial morphometric analysis if they were from the same geographic region, had no known history of neurodevelopmental disorder, craniofacial anomaly or genetic syndrome, and had standardized frontal facial photographs of sufficient quality for landmark annotation. No genetic testing or screening was performed in the control group.

The present study was approved by the Ethics Committee of Hospital Universitario Fundación Valle del Lili (approval no. 2025.140). Written informed consent was obtained from all patients or their legal guardians for the publication of clinical data and photographic images. The present study was conducted in accordance with the ethical principles of the Declaration of Helsinki, ensuring confidentiality and appropriate use of identifiable information for research and publication.

Phenotypic assessment included growth parameters, neurodevelopmental profile, behavioral manifestations, craniofacial dysmorphism, skeletal anomaly and involvement of other organ systems. Clinical features were described using standardized dysmorphology terminology. Developmental delay, intellectual disability and neuropsychiatric features were recorded based on multidisciplinary clinical and neuropsychological evaluation.

Molecular diagnosis and evaluation of clinically relevant differential diagnoses were performed using clinical exome sequencing. In total, four exome analyses were performed institutionally, whereas the remaining studies were performed by external diagnostic laboratories as part of the patient clinical diagnostic workflow. For externally performed tests, sequencing and bioinformatic procedures were performed out according to validated protocols, and the results were reviewed based on the clinical reports available.

For institutionally performed exome analyses, genomic DNA was extracted and purified from peripheral blood samples, followed by preparation of genomic fragment libraries using the KeyExome-GeneSGKit X. 48 Reactions kit, reference LV4283 (Genomic Systems). DNA quantity and purity were assessed using NanoDrop spectrophotometry, including the 260/280 nm ratio, and DNA quantification was performed using Qubit and Quantus fluorometers. Library quality was evaluated using TapeStation capillary electrophoresis, with library fragments showing a Gaussian distribution and a peak of approximately 320 bp. Clonal amplification and sequencing of the selected regions were performed on an Illumina NextSeq 500 platform using Illumina bridge sequencing technology and a paired-end sequencing strategy. Sequencing was performed using the NextSeq 500/550 High Output kit v2.5, 300 cycles, supporting 2x150 bp reads (cat. no. 20024908; Illumina Inc.). Mean sequencing coverage was recorded according to the clinical laboratory report. Bioinformatic analysis of the DNA sequences was performed by comparison with the reference genome sequence GRCh38 using the GeneSystems^® platform v4.0.1 (IVD-CE; https://platform.genesystm.com/services) (21). Variants were considered for analysis when they showed a read depth ≥20X and a variant allele fraction/read ratio ≥0.2.

ANKRD11 variants were described according to the Human Genome Variation Society (HGVS) nomenclature using the NM_013275.6 reference transcript and classified following the American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines (22,23). Variant interpretation was supported by variant interpretation platforms, including VarSome (varsome.com/) and Franklin (Genoox/QIAGEN; franklin.genoox.com/clinical-db/home), together with genotype-phenotype association and available clinical evidence. In silico predictive analyses were performed for variants of uncertain significance using VarSome (24,25).

Variants classified as pathogenic or likely pathogenic were confirmed by Sanger sequencing in four patients for whom confirmation was available as part of the clinical diagnostic workflow. For the remaining six patients, Sanger sequencing chromatograms were generated by external diagnostic laboratories as part of the clinical diagnostic workflow (data not available).

Parental segregation studies were performed when parental samples were available. Novelty was assessed based on the absence of previous reports in VarSome, Franklin and the available literature. Novel variants are currently in the process of being submitted to ClinVar (26).

Two-dimensional facial morphometric analysis was performed in patients and controls using standardized 2D frontal photographs. Images were acquired with participants in an upright head position and neutral expression, with open eyes and closed mouth.

A set of 21 anatomical facial landmarks were manually annotated using the tpsDig2 software version 2.32, 64-bit executable (27) and quantitative shape analyses were conducted to identify craniofacial patterns associated with KBGS. Facial shape variation was assessed using geometric morphometrics, a robust set of statistical tools designed for measuring and comparing 2D/3D shapes with high precision and efficiency (27,28-31). Facial landmark configurations were aligned in a common morphospace using generalized Procrustes analysis, removing non-shape variance due to translation, rotation and scale. Multivariate linear regression was performed to correct for age-associated facial shape differences. The regression residuals were analyzed using principal component analysis (PCA) to explore facial differences between patients with KBGS and matched controls. Morphological variation was visualized in the morphospace defined by the first two principal components. Group differentiation was assessed using Procrustes distances between mean shapes, calculated as the square root of the sum of squared differences between homologous landmarks. All statistical analyses were performed in MorphoJ 1.08.01(32) and R 4.4.2 using the geomorph package v4.0.10(33). Statistical significance was determined using 1,000-permutation tests, with P<0.05 considered statistically significant.

A total of 10 individuals with molecularly confirmed KBGS were included, comprising six males and four females. Age at molecular diagnosis ranged from 4 to 36 years, with a mean age of 16.4 years. All individuals were followed at Fundación Valle del Lili (Cali, Colombia) and underwent comprehensive clinical genetic evaluation.

Developmental delay was present in all patients (10/10, 100%) and all met criteria for intellectual disability, with severity ranging from mild to severe. Growth parameters were variable. Current height ranged from 99 to 171 cm, and birth length data were available for 4/10 individuals, ranging from 47 to 52 cm.

All phenotypic and molecular characteristics are described in Tables I and II.

Semiological findings and phenotypic abnormalities were defined using terms from Human Phenotype Ontology (34).

Craniofacial findings. Craniofacial dysmorphism was identified in all individuals, although the specific combination and severity of features varied (Fig. 1). The most frequent findings included low anterior hairline (8/10, 80%), bulbous nasal tip (8/10, 80%), synophrys (6/10, 60%), facial asymmetry (7/10, 70%) and prominent ears (6/10, 60%). Additional recurrent features included full or arched eyebrows (7/10, 70%), triangular face, brachycephaly, coarse facial features, hypertelorism, long philtrum and macrodontia of the permanent upper central incisors (all 5/10, 50%) (Fig. 2). Less frequent findings included microcephaly (3/10, 30%), downslanting palpebral fissures (2/10, 20%), upslanting palpebral fissures (3/10, 30%), macrostomy (2/10, 20%), cleft lip and/or palate (1/10, 10%), prognathism (1/10, 10%) and micrognathia (4/10, 40%).

Skeletal manifestation. Skeletal anomalies were commonly observed. Scoliosis was the most frequent finding (6/10, 60%). Digital anomalies were prominent, including clinodactyly and/or campodactyly (7/10, 70%), short fourth and fifth fingers (5/10, 50%) and persistent fetal fingertip pads (4/10, 40%). Less frequent skeletal features included cubitus valgus (2/10, 20%), short neck (2/10, 20%), single transverse palmar crease (1/10, 10%), broad thumbs (1/10, 10%), and broad hallux (1/10, 10%).

Neurological and behavioral features. All individuals exhibited developmental delay and intellectual disability. Hypotonia was observed in half of the cohort (5/10, 50%). Behavioral and neuropsychiatric manifestations were frequent, including behavioral disorder (5/10, 50%), attention-deficit hyperactivity disorder (4/10, 40%) and autistic features (4/10, 40%). Seizures were reported in four individuals (40%). Brain magnetic resonance imaging, when available, demonstrated variable findings, including periventricular white matter signal abnormalities compatible with mild leukoencephalomalacia in one patient.

Sensorineural or conductive hearing loss was identified in eight individuals (80%), making it one of the most prevalent non-neurological manifestations in the cohort. Gastrointestinal involvement was frequent, including feeding difficulties (6/10, 60%), gastroesophageal reflux (3/10, 30%), low weight (3/10, 30%) and isolated cases of anal anomaly (1/10, 10%) and intestinal abnormality (1/10, 10%). Congenital cardiovascular defects were observed in two individuals (20%) and renal anomalies were identified in one patient. Cryptorchidism was documented in one male individual.

All 10 individuals carried heterozygous ANKRD11 variants, consistent with autosomal dominant KBGS. The molecular spectrum was predominantly composed of loss-of-function variants, primarily frameshift and nonsense mutations, supporting ANKRD11 haploinsufficiency as the underlying pathogenic mechanism (Table I; Fig. 3). In addition to the ANKRD11 variants reported, no other variants were identified during exome analysis that were considered likely to contribute to the clinical conditions.

Most variants were located in exon 9, a known mutational hotspot, and a single missense variant was identified in one patient (2,6). There were no recurrent variants or 16q24.3 chromosomal rearrangements. Eight variants were considered novel based on the absence of previous reports in variant interpretation platforms, including VarSome and Franklin (24,25), and available literature: c.3600dup (p.Val1201SerfsTer3), c.7753C>T (p.Arg2585Cys), c.741C>A (p.Tyr247Ter), c.7744_7745dup (p.Asn2583Serfs20), c.6517delG (p.Val2173Serfs2), c.5397dup (p.Glu1800ArgfsTer150), c.3499dup (p.Ser1167PhefsTer) and c.3055_3059del (p.Arg1019GlyfsTer14) (Table I; Fig. 3). These variants are currently in the process of being submitted to ClinVar (26).

According to ACMG/AMP guidelines, three variants were classified as pathogenic and seven as likely pathogenic, with strong phenotypic concordance (23). Pathogenic or likely pathogenic were confirmed by Sanger sequencing in four patients for whom confirmation was available as part of the clinical diagnostic workflow (Figs. S1, S2, S3 and S4). Parental studies, when available, confirmed multiple de novo occurrences. Detailed molecular data are summarized in Table I and Fig. 3.

The PCA of 2D facial landmarks revealed partial overlap among diagnostic groups. In total, two patients fell within the range of variation of control individuals, whereas the remaining eight exhibited broad dispersion along PC1, which accounted for 34.7% of the total variance. PC2 accounted for 16.1% of variance and revealed a distinct separation between patients and controls (Fig. 4). Permutation tests based on Procrustes distances confirmed that facial shape differences between controls and patients with KBGS were statistically significant (Procrustes distance=0.045; P=0.0288).

At the extremes of PC1, individuals with the most severe dysmorphic features exhibited pronounced facial asymmetries, affecting either the left or the right side of the face (Fig. 4). At the positive extreme, patients had narrower faces, upward displacement of the right orbit, downward nasal tip displacement, increased lower lip thickness and lateral chin deviation to the right. At the negative extreme, patients exhibited wider faces, downward displacement of the right orbit, upward nasal tip displacement, increased upper lip thickness and lateral chin deviation to the left (Fig. 4).

PC2 highlighted characteristic KBGS facial features, including increased forehead height, nasal prominence, ticker lower lip and reduced chins (Fig. 4).

Patients displayed heterogeneous craniofacial phenotypes (Fig. 4), but no significant associations were detected between facial variation and mutation type (frameshift, nonsense, missense) or exon location (11,14,20).