FDG PET in the evaluation of immune-related hypophysitis and thyroiditis following combination ipilimumab and nivolumab in advanced melanoma

482 Objectives: Hypophysitis and thyroiditis are among the most commonly reported immune-related adverse events (irAEs) following combined ipilimumab/nivolumab therapy for melanoma. The role of 18F-FDG PET/CT (FDG-PET) in the evaluation of these endocrinopathies has not been systematically assessed. Methods: Between 2016 to 2019, all patients (pts) with advanced melanoma who received combined ipilimumab/nivolumab therapy were reviewed. Pts with a pre-treatment and post-treatment FDG-PET were included. On FDG-PET, PET-hypophysitis was defined as a discernable new uptake in the pituitary fossa and PET-thyroiditis as new diffuse uptake in the thyroid. Pre- and post-treatment SUVmax of pituitary and thyroid gland was measured. ROC analysis was used to derive the optimal threshold for metabolic changes on FDG-PET for distinguishing endocrinopathy. FDG-PET, clinical data and brain MRI were reviewed independently by a Nuclear Medicine physician, endocrinologist, and radiologist, respectively, and then findings were correlated. Results: Of 162 pts, 133 and 134 had assessable FDG-PET for hypophysitis and thyroiditis, respectively, with post-treatment FDG-PET performed at a median 76 days (IQR 52-83, range 18-225) from the start of immunotherapy. Overall 41/133 (29%) pts had PET-hypophysitis, of which 18 were clinically-confirmed, 3 were false-positive and 20 were not clinically-assessable due to receiving high-dose glucocorticoids for a concurrent irAE at the time of imaging, although 6 of these also had supportive contemporaneous MRI findings. For PET-hypophysitis pts, median pre- and post-treatment pituitary SUVmax were 2.7 (IQR 2.5-2.9, range 1.9-3.9) and 4.7 (IQR 3.6-5.5, range 2.6-16.2), with a percentage increase of 63% (IQR 39-94%, range 13-431%). The abnormal PET findings preceded the clinical diagnosis in 7/18 pts by a median of 16 days (range 5-50). FDG-PET was negative for hypophysitis in 12/29 pts with a prior or subsequent clinical diagnosis of hypophysitis. Where the clinical presentation was not masked by high-dose glucocorticoids, the positive and negative predictive value of FDG-PET for hypophysitis was 86% and 87%, respectively. Based on ROC analysis the optimal percentage change in SUVmax was 30% for distinguishing hypophysitis. PET-thyroiditis was detected in 30/134(22%) pts. The pre- and post-treatment SUVmax were 2.1 (IQR 1.7-2.3, range 1.3-3.3) and 4.8 (IQR 3.8-5.9, range 2.8-9.1), respectively, with an increase of 116% (IQR 84-177%, range 52-300%). Overall 41/134 (31%) pts had documented biochemical evidence of thyroiditis. The positive and negative predictive value of PET was 97% and 89%, respectively. Based on ROC analysis, the optimal percentage change in SUVmax for distinguishing thyroiditis was 42%. Further follow-up FDG-PET (30/39 pts with PET-hypophysitis and 25/30 pts with PET-thyroiditis) revealed resolution of SUVmax to baseline in all cases by a median of 104 days (IQR 77-133, range 40-484) and 32 days (IQR 79-194, range 49-1045), respectively. Conclusions: FDG-PET detects transient increases in FDG uptake in the pituitary and thyroid gland following combined ipilimumab/nivolumab which appears to be highly predictive of the development of these endocrinopathies, therefore prompting more stringent monitoring. A high incidence of uninterpretable biochemical assessment of the pituitary-adrenal axis likely contributed to the underestimation of hypophysitis incidence. A multimodality approach is important in the timely diagnosis of immune-related endocrinopathies.