SUN-298 Spectrum of Imaging in Immune Checkpoint Inhibitor Induced Hypophysitis

Abstract Background: Hypophysitis (HP) is a known immune related adverse event of immune checkpoint inhibitors (CPIs), commonly associated with CTLA-4 inhibitors and rarely with PD-1/PD-L1 inhibitors. Prior studies of MRIs at HP diagnosis noted pituitary enlargement with resolution within a few weeks. In this study, we examine MRI changes in patients with CPI-induced HP. Methods: Subjects with biochemical evidence of central hypothyroidism or central adrenal insufficiency and MRIs were reviewed by endocrinology and neuroradiology. MRIs were classified relative to HP diagnosis: baseline (at least 21 days prior), diagnosis (within 21 days), and follow up (over 21 days). Patient characteristics included age at CPI initiation, sex, race/ethnicity, personal and family history of autoimmunity, type of cancer and CPI. Results: Twenty-six subjects met the inclusion criteria. The mean age was 59 years; 62% were male and 86% were non-Hispanic white. Nineteen percent had a personal history and 31% a family history of autoimmunity. Fifty percent had melanoma. At HP diagnosis, 46% were on PD-1/PD-L1 inhibitors, 42% were on combination PD-1/CTLA-4 inhibitors and 12% were on CTLA-4 inhibitors. Median time from CPI initiation to HP diagnosis was 95 days. Time to HP was shorter on a CTLA-4 inhibitor combination or monotherapy (median 82 days) compared to a PD-1/PD-L1 inhibitor monotherapy (median 220 days; Wilcoxon rank sum, p <0.01). Central adrenal insufficiency was present in all patients not yet on steroids. Central hypothyroidism was common (10/19) in those without primary thyroid disease and was not associated with type of CPI (Fisher’s exact, p=0.18). Thirteen subjects had baseline MRIs, 18 had MRIs at HP diagnosis and 13 had MRIs in the follow up period. Baseline MRIs were normal in 12/13; one subject had an enlarged pituitary. At diagnosis, 10 had an enlarged pituitary, 7 a normal pituitary and 1 a partially empty sella. CTLA-4 inhibitor exposure was associated with pituitary enlargement at diagnosis: 9/11 compared to 1/7 on PD-1/PD-L1 inhibitor (Fisher’s exact, p <0.04). Of the subjects who had follow-up MRIs, 3 had an enlarged pituitary, 7 a normal pituitary and 3 a partially empty sella. Follow up imaging did not differ between treatment types (Fisher’s exact, p >0.05). Timing of MRI was significantly associated with pituitary appearance (Fisher’s exact, p <0.01). Conclusion: The MRI appearance of HP presents as a spectrum, from a partially empty sella, normal pituitary to an enlarged pituitary. HP diagnosed in the setting of CTLA-4 inhibitor treatment occurs earlier and is more likely to induce an enlarged pituitary gland compared to PD-1/PD-L1 monotherapy, which occurs later and is associated with a normal appearing MRI at diagnosis. This suggests that the pathogenesis of HP following CPI exposure may vary depending on the type of CPI.