Cranial MRI Abnormalities and Long-term Follow-up of the Lesions in 770 Girls With Central Precocious Puberty

Abstract

Context: Central precocious puberty (CPP) may arise from central nervous system (CNS) lesions in a few affected girls. Recently, the incidence of girls with CPP has increased mostly in 6-8 year olds, in whom the necessity of magnetic resonance imaging (MRI) is debated. Objective: To investigate the frequency, long-term outcome and potential predictors of CNS lesions in a large cohort of girls with CPP. Methods: A multicenter cohort of 770 Turkish girls with CPP who had systematic cranial MRI between 2005 and 2017. Age at puberty onset was <6 years in 116 and 6-8 years in 654. CNS lesions were followed until final decision(6.2 +/- 3.1 years). Potential predictors of CNS lesions were evaluated by univariate analyses. Results: A total of 104/770 (13.5\%) girls had abnormal brain MRI. Of these, 2.8\% were previously known CNS lesions, 3.8\% had newly detected and causally related CNS lesions, 3.1 \% were possibly, related and 3.8\% were incidental. Only 2 (0.25\%) neoplastic lesions (1 low grade glioma and 1 meningioma) were identified

neither required intervention over follow-up of 6 and 3.5 years respectively. Age at breast development <6 years (odds ratio {[}OR] 2.38

95\% CI 1.08-5.21) and the peak luteinizing hormone/follicle-stimulating hormone (LH/FSH) ratio >0.6 (OR 3.13

95\% CI 1.02-9.68) were significantly associated with CNS lesions. However, both patients with neoplastic lesions were >6 years old. Conclusion: Although age and LH/FSH ratio are significant predictors of CNS lesions, their predictive power is weak. Thus, systematic MRI seems to be the most efficient current approach to avoid missing an occult CNS lesion in girls with CPP, despite the low likelihood of finding a lesion requiring intervention.