SAT-086 Pituitary Volume as a Diagnostic Measure for Determining the Etiology of Short Stature in Children: Receiver Operating Characteristic Curve Analysis and Cutoff Values

Abstract

Background: It is speculated that pituitary volume (PV) is a marker of chronic growth hormone (GH) secretion. In previous studies, we determined that children with GH deficiency (GHD) and idiopathic short stature (ISS) had significantly smaller PVs than normal controls (NCs). Cutoff values for small PVs are needed to improve the clinical utility of PV in determining children who qualify for GH therapy.Objective: To define the cutoff between pathologic and nonpathologic PV in prepubertal and pubertal children with short stature (SS). Patients and Methods: The SS group was selected from the database of a pediatric endocrinology center, which was queried for siblings (SBs) aged 6–18 yrs who underwent a GH stimulation test and MRI between 2013–2019. All 77 SBs had SS, defined as 2 SDs below mean height for age, subnormal growth velocity for at least 6 months, or predicted height at least 2 inches discrepant from midparental height. The NC group was selected from the database of a neuroradiology center; these NCs consisted of 170 randomly selected subjects aged 6–18 yrs. Patients with MRI abnormalities were excluded. PVs were calculated using the ellipsoid formula (LxWxH/2). ROC curve analysis was utilized to generate cutoff values. The diagnosis of short stature was the dependent variable and PV was the independent variable. The PV with the highest Youden index was selected as the definitive cutoff for a small PV. Results: The mean (MN) and median (MD) age of SBs was 11.6 ±2.2 and 11.9 yrs, respectively, and the MN and MD age of the NCs was 12.6 ±3.4 and 13.2 yrs, respectively. The MN and MD age of prepubertal SBs (n=29) and NCs (n=58) were 9.3 ±1.2 and 9.7, and 8.6 ±1.4 and 8.6 yrs, respectively. The MN and MD age of pubertal SBs (n=48) and NCs (n=112) were 13.0 ±1.4 and 12.7, and 14.7 ±1.9 and 14.6 yrs, respectively. The difference in MN age between SBs and NCs was significant (p<0.05). For prepubertal subjects, sensitivity was 86.21% and specificity was 68.97%. The distance to corner was 0.3396, and the highest Youden index was 0.5517, corresponding to a PV of 215.02 mm3. The Area Under the Curve (AUC) was 0.8395 with a standard error of 0.0426 (p<0.001). For pubertal subjects, sensitivity was 81.25% and specificity was 79.46%. The distance to corner was 0.2781, and the highest Youden index was 0.6071, corresponding to a PV of 315.0 mm3. The AUC was 0.8460 with a standard error of 0.0337 (p<0.001). Conclusion: To our knowledge, we present the first study on the sensitivity and specificity of PV in determining the etiology of SS. Our data suggest that prepubertal patients with a PV<215.02 mm3 and pubertal patients with a PV<315.00 mm3 have small pituitary glands. Statistically calculated cutoffs are necessary to accurately diagnose pituitary hypoplasia and should be utilized to determine the etiology of SS. Future studies should include children with Tanner staging and height SDs to generate more accurate PV cutoffs.