性早熟
医学
中枢性早熟
主成分分析
内分泌学
儿科
激素
计算机科学
人工智能
作者
Amanda Cleemann Wang,Casper P. Hagen,Trine Holm Johannsen,André Madsen,Line Cleemann,Peter Munk Christiansen,Katharina M. Main,Anders Juul,Rikke Beck Jensen
标识
DOI:10.1210/clinem/dgad535
摘要
Nonprogressive premature thelarche (PT) is a self-limiting variant of early puberty, while idiopathic central precocious puberty (ICPP) is a disorder that causes progressive development of secondary sexual characteristics and often requires treatment. The diagnostic differentiation between these conditions is important but can be challenging since they often both initially present clinically with isolated breast development.To describe relevant clinical variables in a large cohort of girls referred for early puberty, and to evaluate clinical and biochemical parameters to distinguish between girls with ICPP and PT.This retrospective study included 1361 girls referred with signs of early puberty to a single, tertiary center from 2009 to 2019. We evaluated clinical presentation, medical history, growth velocity, bone age, hormonal serum concentrations, and gonadotropin-releasing hormone (GnRH) test results.Central precocious puberty was diagnosed in 11% (ICPP: n = 143, organic CPP: n = 11) girls, whereas 8% (n = 91 girls) presented with PT. Receiver operating characteristic (ROC) analysis showed several biochemical and anthropometric markers as potential parameters to differentiate between ICPP and PT; however, none were individually adequate. Principal component analysis (PCA)-derived clinical and hormone profiles could predict girls with ICPP from girls with PT with a specificity of 90% and sensitivity of 84%, outperforming any single marker.Differentiation of girls with ICPP and PT can be supported by individual clinical and biochemical parameters. However, dimension reduction of clinical and hormonal profiles by PCA improved the diagnostic value, which in the future may support the diagnostic process as a supplement to the GnRH test in evaluation of pubertal disorders.
科研通智能强力驱动
Strongly Powered by AbleSci AI