SUN-038 High Hormone Values May Exclude Pituitary Source in MRI Negative Cushing’s Disease

Abstract Y.B. Omotosho: None. R. McGlotten: None. P. Chittiboina: None. L. Nieman: None. Background: Extremely high 24h urine cortisol excretion (UFC) is seen in patients (pts) with Cushing’s disease (CD) and ectopic ACTH secretion. Previous literature suggests that CD macroadenomas have higher UFC than microadenomas. Objective: To determine whether hormonal factors or sex predict pituitary MRI lesions in adults with CD. Methods: We retrospectively evaluated UFC and MRI in 98 patients (77 females) with CD confirmed by ACTH-positive immunohistochemistry of tumor. Hormone levels (UFC, 24h serum cortisol (F) and ACTH) and the largest diameter of a pituitary lesion on MRI were noted. Lesions were grouped by size (0, 2– 5, >5– <10mm, and macronodular (Macro) > 10mm). UFCs were expressed as the fold-difference from the assay upper limit of normal (ULN). Differences in variable concentrations or frequency were examined by Mann-Whitney and Chi-square tests, respectively. Logistic regression tested the effects of hormone levels on the likelihood of having a Macro or absent lesion. Spearman r, sensitivity and specificity were calculated to assess factor relatedness or diagnostic utility. Statistical analyses were done by PRISM statistical software. P value <0.05 was considered statistically significant. Results: Lesion size was distributed unequally: 0mm (n=8), 2-5 (36), 5-10mm (36) and Macro (18). There was no difference in lesion size by sex (women: median 6mm IQR 4.5– 9, maximum 30 mm; vs men: median 5mm, IQR 3 – 7, maximum 16, P=0.07) or the frequency of Macro in men (2/21) vs women (16/77), P=0.22. Lesion size correlated weakly with ACTH (r=0.21, P=0.034) and 24h F (r=0.20, P=0.038) but not with UFC (r=0.09, P=0.34).UFC predicted the likelihood of Macro (P=0.0031). However, the best UFC criterion (13xULN) for Macro had poor sensitivity (30%; 95%CI 15-52%) with 88% specificity (95%CI 78 – 94%). UFC levels for Macro (median 11.8ULN, range 3.1-147) and unseen lesions (median 5ULN, range 0.7 – 33.4, P=0.18) overlapped, and UFC did not predict an absent lesion (P= 0.9276).ACTH did not predict Macro (P=0.20) or an absent lesion (P=0.053). ACTH values were significantly higher in pts with visible lesions vs unseen lesions (median 63.9 pg/mL, range 11-527 vs median 46.2 pg/mL, range 12.9 - 103; P=0.045). 24h F did not predict Macro (P=0.1277) or an absent lesion (P=0.764). The highest value in unseen lesions was 21.2ug/dL. Due to overlap of ACTH and 24h F values across lesion sizes, no criterion aided diagnosis. Discussion: Although logistic regression by UFC predicted a Macro lesion, no criterion had acceptable sensitivity for its diagnostic identification. However, based on highest results when no lesion was identified, we speculate that hormonal values may help to identify patients with ectopic ACTH secretion if no tumor is visible on MRI and ACTH is > 105 pg/ml, 24h F is >21.2ug/dL or UFC is >33.4ULN. Sunday, June 2, 2024