THU538 Use Of Multiple Functional And Structural Imaging Modalities Improves Detection Of Ectopic ACTH Tumors

Abstract

Abstract Disclosure: H. Elenius: None. R. McGlotten: None. A.M. Gharib: None. L.K. Nieman: None. Introduction: In Ectopic ACTH Syndrome (EAS), Cushing's syndrome (CS) is caused by ACTH secretion from a non-pituitary tumor. Tumor resection reduces CS morbidity and mortality, but up to 38% of tumors remain occult despite repeated structural imaging. We prospectively studied whether functional PET imaging with somatostatin receptor ( 111In-pentetreotide Octreoscan [OCT], vs 68Ga DOTA0-Tyr3 octreotate DOTATATE [DOTA]) or amine precursor (18F-DOPA [F-D]) ligands would improve tumor detection. Methods: From 2014-2022, 22 EAS patients underwent imaging with F-D and DOTA PET/CT and/or MRI (1.5 and/or 3 Tesla, T) of chest/abdomen/pelvis (C/A/P), gated cardiac CT (GC-CT) and/or MRI (GC-MRI), and/or OCT before tumor resection. Pre-surgical imaging results were confirmed by pathology or biochemical cure. Results: Tumors included 16 pulmonary (pulm) neuroendocrine tumors (NET), one thymic NET, one appendiceal NET, one pheochromocytoma and one metastatic NET in mediastinal lymph nodes; two patients had tumorlets. Surgery was done if at least CT or MRI was positive; some lesions were seen only after concurrent PET scan review. Shown in decreasing order, the modalities performed as follows (number of patients = N): Sensitivity (S): CT C/A/P 100%, N 22 ; FDG-PET 100%, N 2 ; 3T MRI C/A/P 93%, N 14 ; F-D 91%, N 22 ; DOTA 86%, N 22 ; 1.5T MRI C/A/P 71%, N 21 ; GC-CT 67%, N 6 ; GC-MRI 50%, N 8 ; OCT 31%, N 13. 95% Confidence intervals (CI) overlapped for all except OCT. Positive predictive value (PPV): GC-MRI 100% ; 3T MRI C/A/P 88% ; 1.5T MRI C/A/P 84% ; DOTA 83% ; CT C/A/P 81% ; GC-CT 80% ; F-D 75% ; OCT 67% ; FDG-PET 67%. All CIs overlapped.1.5T MRI, F-D and DOTA missed a 12 mm pulm NET (false negative, FN) only seen on CT. 1.5T MRI missed three other pulm NET (6, 8 and 10 mm) seen on 3T MRI. 1.5T and 3T MRI missed a 13 mm pulm NET that abutted the heart and was seen on both gated studies - this was the only FN lesion in the 3T MRI cohort. F-D missed the 12 mm pulm NET seen on CT and the retrocardiac NET, which was DOTA +. DOTA missed three other pulm NET (6, 8 and 10 mm); all were positive on F-D. OCT detected 4/10 pulm NET, 6-25 mm, all positive on F-D and DOTA, and missed 9 tumors seen on F-D (N=8) or DOPA (N=8). F-D detected false positive (FP) lesions in seven patients while DOTA detected FP lesions in three, mostly infections or thyroid nodules. All but one patient were localized by DOTA (N=19) and/or F-D (N=20). Conclusions: As reported previously, most EAS tumors were intrathoracic. Among structural modalities, CT was the most sensitive (100%) while cardiac MRI and 3T MRI gave the highest PPV (100% and 88%). Dedicated cardiac imaging found pericardiac lesions missed on routine modalities due to motion artefacts. F-D and DOTA were significantly more sensitive (91% and 86%) than OCT (31%). DOTA detected fewer FP lesions, providing a higher PPV (83%) than F-D (75%), but with lower sensitivity. Use of multiple functional and structural scans improved tumor detection. Presentation: Thursday, June 15, 2023