ESPE Abstracts

Background and aim: Phenotypes of Hypothalamo-Pituitary (H-P) dysfunction are heterogenous and unpredictable. Through novel neuroimaging segmentation, we quantified anatomical alterations of each discrete component (hypothalamus, stalk, and pituitary) to better elucidate the origin and phenotype of different congenital and acquired HP disorders.

Methods: We compared 66 patients (35M/31F) aged 12.4 ± 3.1 years, followed up for 8.2 ± 4.0 years - 35 tumours/31 congenital - with 25 age/gender-matched controls (13.6±3.1 years, 13F/12M). Neuroendocrine dysfunction was graded as number of anterior (HAP-EMS) and posterior (HPP-EMS) H-P deficits [Endocrine Morbidity Scores (EMS)]. Early morning basal pituitary hormone concentrations were taken before any medication and paired with high-resolution MR images acquired on the same day, to measure hypothalamic/pituitary volumes and stalk sizes (upper/middle/lower). Normative hypothalamic and pituitary SDS (-/+ 1.5, -/+ 2) were derived from controls for 3 age groups: 6-10y (prepubertal), 11-14y (pubertal), 15-18y (post-pubertal).

Results: Across the whole cohort, the smaller the pituitary volume SDS (P = 0.001) and stalk size (ie middle P = 0.003), the greater the HAP-EMS. In the congenital group, the smaller the hypothalamic volume SDS, the greater the HPP-EMS (P = 0.012), but not the HAP-EMS. In tumours, the lower (damaged post-treatment) or non-measurable (invaded) hypothalamic volume SDS, the greater the HAP-EMS (P = 0.004) and HPP-EMS (P = 0.013). Across the whole cohort, GH peaks and TSH/LH/FSH/ACTH/cortisol concentrations decreased with smaller pituitary volume (P = 0.045, CC=0.413; P <0.001, CC=0.507; P = 0.025, CC=0.331; P = 0.009, r =0.380; P = 0.097, CC=0.248; P = 0.004, CC=0.414, respectively) and stalk size (i.e. middle P = 0.030, CC=0.409; P = 0.004, CC=0.471; P = 0.044, CC=0.337; P = 0.010, CC=0.423; P = 0.015, CC=0.404; P = 0.006, CC=0.453, respectively). In the congenital group, GH peaks and prolactin concentrations correlated with hypothalamic volume, positively (P = 0.009, CC=0.743) and negatively (P = 0.023, CC-0.45), respectively. The 3 patients with ROHHAD(NET) and hyperprolactinaemia had reduced hypothalamic volumes. Urine osmolalities positively correlated with both pituitary (whole cohort, P = 0.019, CC=0.348) and hypothalamic (congenital P = 0.012, CC=0.514; tumours P = 0.002, CC=0.638) volume.

Conclusion: The study demonstrates for the first time that the severity of anterior and posterior H-P deficits is reflective of the degree of anatomical changes not only in the pituitary, but also in the stalk and the hypothalamus, and quantifiable on neuroimaging by volumetry. We have provided insight into the differing hypothalamic and pituitary contributions underlying the aetiology of neuroendocrine phenotypes and demonstrated that hypothalamic volume loss, identified through novel imaging tools, is predictive of PP disorders, GH deficiency and hyperprolactinaemia.