In a study of 100 healthy volunteers, 10 % had a pituitary lesion on T1 SE MRI imaging , with a 3–6 mm diameter [12]. In a study of 201 patients with Cushing’s disease who had surgical confirmation of the location of the tumor, 14 % had a false positive lesion on MRI [13]. Similarly, in a study of 66 patients with ectopic ACTH secretion, 17 (26 %) had an abnormal pituitary MRI, 13 of whom had previous unsuccessful pituitary exploration [14]. In another study, 6 of 26 patients with ectopic ACTH secretion had a lesion on pituitary MRI, but only one had a diameter >6 mm (96 % specificity for 6 mm criterion) [15]. Taken together, these data indicate that a lesion on pituitary MRI does not necessarily correspond to a corticotrope adenoma. Such a lesion does provide a location to target during transsphenoidal surgery , however.

When reviewing imaging studies to identify corticotrope tumors, it is important to recognize that these may occur rarely in a non-pituitary location along the developmental path of Rathke’s pouch: in the nasal cavity [16], the sphenoid sinus [17], and clivus. They may also occur in locations proximal to, but outside of the anterior pituitary gland, including the infundibulum [18], parasellar location [19], posterior pituitary [20], and cavernous sinus. The imaging results for these areas should be reviewed in patients in whom biochemical data suggest Cushing’s disease but the pituitary MRI is negative and in those with unsuccessful transsphenoidal exploration.

A few studies have evaluated the use of ¹¹C-methionine or ¹⁸F-FDG positron emission tomography (PET) for the localization of pituitary adenomas . The essential amino acid methionine is taken up into tissues that have increased protein synthesis. Physiologic uptake is present in normal pituitary. In one study, 7 of 10 patients with Cushing’s disease had asymmetric uptake in the pituitary gland at the site of a lesion seen by SPGR MRI. These were all confirmed to be ACTH-secreting tumors after surgical resection [21]. Another study compared the ability of ¹⁸F-FDG to image metabolically active tissue with the sensitivity of T1 SE or SPGR MRI. ¹⁸F-FDG PET localized tumor in 4 patients, all of whom had a less than 180 % increase in ACTH after CRH stimulation. ¹⁸F-FDG PET also detected two adenomas not identified by T1 SE, but did not improve the sensitivity of SPGR MRI [22].

Having assigned a diagnosis of presumed ectopic ACTH secretion based on biochemical testing , the next challenge is to locate a possible tumor. Although biochemical tumor markers are not uniformly helpful, they may suggest what to image first. For example, elevated calcitonin or plasma free metanephrines may point to the thyroid or adrenal gland; on the other hand, chromogranin A is not specific, and urinary 5-HIAAA is often not abnormal in patients with foregut carcinoids, perhaps because these often do not express the enzyme aromatic l-amino-acid decarboxylase needed for serotonin synthesis.

Although the initial description of the ectopic ACTH syndrome highlighted overt and metastatic tumors, slow growing, often occult tumors represent the majority of cases in 2016. As a result, imaging identification and surgical removal of the tumor are critical to successful treatment [23]. Despite the use of anatomical imaging techniques like computed tomography (CT) and magnetic resonance imaging (MRI), up to 50 % of ectopic ACTH-secreting tumors are not found on initial imaging [24].

If no biochemical marker suggests an anatomic source, given that about 50 % of these tumors arise in the chest (Table 2), computed tomography (CT) of the thorax, using thin slice thickness (1–2 mm), is a cost-effective initial imaging strategy. If a clear-cut lesion is identified, then additional imaging may not be needed. However, in many series, tumors remain occult, or occur elsewhere, and additional imaging with different modalities over time is needed [25].

Additional imaging includes CT imaging of the neck, abdomen, and pelvis, as well as MRI of these areas and the chest. Neuroendocrine tumors may be “bright” on T2 sequences that utilize fat-suppression techniques, making these sequences an important part of an MRI protocol [26]. The use of “triple phase” CT imaging may improve detection of intestinal and pancreatic tumors and hepatic metastases. This involves imaging before injection of iodinated contrast, followed by three phases after contrast injection at a rapid rate (2–3 mL/s). These phases include a late arterial phase of enhancement, at 20–45 s after the start of the injection, followed by a third imaging at 60–70 s after the start of injection, for the portal venous phase [27]. A delayed phase scan may also be obtained at 3 min to better characterize liver lesions if present.

MRI and CT provide the best anatomic/structural resolution of tumors, and are complementary, having about 90 % combined sensitivity [14,24].

Functional imaging, also called “molecular imaging,” reduces false positive results because it relies on the specific properties of tumor cells, not just their anatomic characteristics . However, tumors lacking the relevant somatostatin receptors, increased metabolic rate (FDG-PET), or amine precursor uptake (F-DOPA) have false negative results [28].

In the United States, somatostatin receptor scintigraphy is commercially available using [¹¹¹In-DTPA-d-Phe]-pentetreotide (Octreoscan™, OCT) at a 6 mCi dose. The ability of OCT to identify the tumors depends on multiple factors, including the dose of the radiopharmaceutical, the type and degree of somatostatin receptor expression, and tumor size [28–30]. Relatively small case series report that OCT detects 4/12 [31], 6/6 [32], 10/18 [33], 12/20 [34], and 5/16 tumors [35]. A larger series of 39 patients found a sensitivity of 41 %, but with a false positive rate of 27 % [36]. A systematic review of the literature found an overall OCT detection rate of 48.9 % (84/172) [24].