Pituitary Volume Cutoffs as Another Tool for Determining Growth Hormone Treatment Eligibility

Abstract

\n Background: The GH stimulation test (GHST) is the gold standard for the diagnosis of GH deficiency (GHD), yet a significant number of short children fail to be diagnosed as GHD. We have speculated that pituitary volume (PV) could be used in conjunction with results from the GHST to diagnose GHD; however, cutoff values for low PVs need to be further explored.\n Objective: To define a diagnostic cutoff value of PV for determining GH treatment eligibility for patients (PTs) with short stature.\n Patients and Methods: The database of GHST results at a Pediatric Endocrinology center was queried for PTs aged 6-18 yrs who underwent a GHST, MRI, and blood work between 1/2018 - 6/2019. PTs with relevant comorbidities were excluded. Clonidine and L-dopa were used to induce GH secretion during the GHST. GHD was defined as a peak GH ≤ 10 ng/mL. MRIs were acquired on a Philips 1.5 or 3.0 T scanner (1mm slices) and PV was calculated using the ellipsoid formula (LxWxH/2). 144 PTs were the subjects of this study. ROC curve analysis was utilized to generate cutoff values. PV was used to predict GHD in prepubertal (age < 11 yrs) and pubertal (age > 11 yrs) children. The value with the greatest Youden index (J) was selected as the definitive cutoff.\n Results: The mean (MN) and median (MD) ages of PTs were 12.2 ± 2.2 and 12.3, respectively. The MN and MD ages of prepubertal PTs (n=43) were 9.4 ± 1.1 and 9.7, respectively. The MN and MD ages of pubertal PTs (n=103) were 13.4 ± 1.4 and 13.2, respectively. Initially, 10 ng/mL was utilized as the cutoff for GHD. For predicting GHD from PV in prepubertal PTs, sensitivity was 89.47% and specificity was 66.67%. The distance to corner was 0.3488, and the highest J was 0.5641, corresponding to a PV of 240.00 mm3. The Area Under the Curve (AUC) was 0.6581 with a standard error (SE) of 0.2429 (p>0.05). For predicting GHD from PV in pubertal PTs, sensitivity was 72.94% and specificity was 81.25%. The distance to corner was 0.3292, and the highest J was 0.5419, corresponding to a PV of 275.00 mm3. The AUC was 0.7901 with a SE of 0.0687 (p<0.05). Further analysis was done to explore the use of 7 ng/mL as the cutoff for GHD. For predicting GHD from PV in prepubertal PTs, sensitivity was 25.00% and specificity was 90.91%. The distance to corner was 0.7555, and the highest J was 0.1591, corresponding to a PV of 133.66 mm3. The AUC was 0.4989 with a SE of 0.0931 (p>0.05). For predicting GHD from PV in pubertal PTs, sensitivity was 57.89% and specificity was 63.64%. The distance to corner was 0.5563, and the highest J was 0.2153, corresponding to a PV of 240.00 mm3. The AUC was 0.6112 with a SE of 0.0584 (p<0.05).\n Conclusion: PVs ≤ 275.00 mm3 in pubertal PTs should be considered low; however, cutoffs for prepubertal PVs were not significant in this study. To our knowledge, we present the first study to generate a PV cutoff based on the GHST. Future studies including more PTs and tanner staging will further improve the accuracy of PV cutoffs for GHT eligibility.