Role of Breast Ultrasonography in the Diagnosis of Precocious Puberty

The aim of this study was to determine the role of ultrasound evaluation of breast
tissue, singularly and in a multivariate model, in the early differentiation between
incomplete precocious puberty (IPP) and complete precocious puberty (CPP). In a
retrospective analysis, we evaluated 52 girls (CPP in 26 cases and IPP in 26) aged
1.35 - 9.59 yrs (mean± SD 7.55±1.47 yrs). Diagnostic evaluation of pts included:
breast Tanner stage, bone age, FSH and LH response to LHRH, basal 17-β-estradiol
(E2), pelvic and breast ultrasound. Logistic regression models were fitted to
identify possible diagnostic factors for IPP (as compared to CPP). All variables
with a p-value<0.1 at univariate analysis were included in a multivariate model to
assess their independent prognostic value. An additional model was fitted with
mammal gland volume forced into the model, due to its clinical relevance. The
area under the model ROC curve (AUC) was computed to assess model discrimination.
The data indicate that in our pts uterine volume ≥ 5cc, LH peak ≥7UI/L and
breast volume ≥ 1,1cc are hightly associated to CPP, bone age >2DS and E2 level
≥ 15μg/ml are only marginally associated, while FSH peak ≥10UI/L and diameter
of the dominant follicle ≥ 1cm are not associated . Uterine volume, LH peak, bone
age and diameter of the dominant follicle were included in a multivariate model,
that showed an independent prognostic role of uterine volume and LH peak. The
model had a high discriminant ability (AUC ROC 0.89) and permits to classify
correctly 77.78% of cases. When adding breast volume to model, this also proved
independent prognostic role for IPP, with further increase in discriminating ability
(AUC ROC 0.91). This second model permits to classify correctly 82.22% of
cases, so leading to better discrimination between IPP and CPP.