Brain MRI Findings and Factor Associated With Clinical Significant CNS Lesions in Girls Aged 6–8 With Central Precocious Puberty

BACKGROUND: Brain magnetic resonance imaging (MRI) is recommended for selected girls with central precocious puberty (CPP). Routine brain magnetic resonance imaging (MRI) in girls aged 6-8 years with central precocious puberty (CPP) remains clinically controversial. Neurological features may help identify girls at higher risk, but whether basal luteinizing hormone (LH), a routinely measured marker of hypothalamic-pituitary-gonadal axis activation, provides additional information beyond age at diagnosis remains uncertain. METHODS: In this retrospective cross-sectional study, 848 girls aged 6-8 years diagnosed with CPP underwent brain MRI at Vietnam National Children's Hospital between April 2021 and June 2026. Brain MRI findings were classified as clinically significant/pathological, uncertain/questionable significance, or incidental. Neurological characteristics were compared across the MRI categories. Logistic regression was used to identify factors associated with clinically significant MRI abnormalities. Sequential models evaluated whether basal LH provided additional information beyond age at diagnosis, with model fit and discrimination assessed using likelihood-ratio tests, Akaike information criterion (AIC) and area under the receiver operating characteristic curve (AUC). RESULTS: Any brain MRI abnormalities were identified in 86 girls (10.1%), including 30 (3.5%) with clinically significant central nervous system (CNS) lesions, 14 (1.7%) with questionable findings and 42 (5.0%) with incidental findings. Neurological symptoms, signs or previous CNS history were more frequent among girls with clinically significant abnormalities than among those with uncertain or incidental findings (30.0%, 7.1% and 4.8%, respectively; p = 0.012). Basal LH was independently associated with clinically significant MRI abnormalities (adjusted OR 1.30 per IU/L, 95% CI 1.14-1.47; p < 0.001. Adding basal LH to age at diagnosis significantly improved model fit (AIC 247.79 vs. 259.78; likelihood-ratio p = 0.0002), although discrimination remained modest (AUC 0.654 vs. 0.550 for age alone, p = 0.055). Bone-age advancement did not provide significant additional information. CONCLUSIONS: In girls aged 6-8 years with CPP, brain MRI abnormalities were identified in approximately 1 in 10 girls, but most were incidental or of uncertain clinical significance. Clinically significant abnormalities were enriched among girls with neurological features. Basal LH provided incremental information beyond age at diagnosis but had limited discriminatory performance. These findings support a risk-stratified approach to brain MRI in girls aged 6-8 years with CPP, with careful neurological assessment remaining central to imaging decisions.