Recognition and Management of Ectopic ACTH Secreting Tumors

Abstract

Ectopic ACTH syndrome (EAS), in which Cushing syndrome is caused by excessive ACTH secretion from a tumor located outside of the pituitary, is associated with an impaired quality of life and an increased mortality rate. Outcomes can be improved with successful tumor localization and resection, which often proves difficult. In order to distinguish EAS from Cushing disease, a significantly more common condition where excessive ACTH is secreted from a pituitary tumor, bilateral inferior petrosal sinus sampling (IPSS) is often necessary. Correct performance and interpretation of IPSS hence becomes crucial to avoid inappropriate future interventions, including surgical procedures. Once an ectopic source of ACTH is confirmed biochemically, identifying the causative tumor is often challenging since they can be located in unexpected areas and potentially be very small. Additionally, EAS carries a risk of severe hypercortisolism, which sometimes needs urgent treatment to avoid disastrous outcomes. The cases here illustrate pitfalls in diagnostic biochemical testing, describe helpful imaging strategies to improve the chances of tumor detection, and review available options to rapidly normalize severe hypercortisolism in critical situations.

In ectopic ACTH syndrome (EAS), Cushing syndrome (CS) is caused by excessive secretion of ACTH from a tumor located outside of the pituitary gland. The overall incidence of endogenous CS is estimated at around 2.5 cases/million/year, with EAS making up around 15% of these cases [1, 2]. Pulmonary neuroendocrine tumors (NETs) are the most common cause of EAS, representing about half of identified tumors in descriptive case series [3-12]. Historically, small cell lung carcinoma was described as one of the most common etiologies [13-15] and is likely underreported in more recent publications. With thymic NETs causing an additional 5% to 10% of cases [16], it becomes clear that most tumors causing EAS are intrathoracic. The remaining tumors have a large variety of etiologies and can be located anywhere from the neck to the pelvis. Adding to their unpredictable locations, causative tumors also can be as small as a few millimeters in size, sometimes making localization incredibly difficult and resulting in around 20% of lesions remaining occult despite extensive imaging [1].

Patients with EAS have an increased mortality rate and an impaired quality of life from CS manifestations. Prognoses vary significantly based on tumor etiology, ranging from a median survival of less than 1 year for small cell lung carcinoma to a 5-year survival rate of 86% for pulmonary NETs [1, 7]. In patients without aggressive cancers, hypercortisolism underlies the excess mortality, leading to cardiovascular disease, thromboembolic events, and an increased susceptibility to infections. Prompt identification and resection of the causative tumor improves outcomes by preventing tumor spread and restoring the normal hypothalamic-pituitary-adrenal axis. The following cases illustrate (1) the pitfalls in biochemical testing to distinguish EAS from Cushing disease (CD), (2) imaging strategies to maximize the chance of tumor detection, and (3) medical management of cases with extreme hypercortisolism.

 

Case 1

A 52-year-old woman was evaluated after 6 years of progressive symptoms including worsening glycemic control, weight gain, wide violaceous striae, proximal muscle weakness, and brain fog. CS was confirmed by abnormal low-dose 1 mg overnight dexamethasone suppression test (LDDST) [cortisol 9.6 mcg/dL; upper limit of normal (ULN) < 1.8], and increased 24-hour urine free cortisol (UFC) (133.4 mcg/d; ULN ≤45 mcg/d). Morning ACTH was 85 pg/mL (reference range 15-66 pg/mL), confirming ACTH dependence; pituitary magnetic resonance imaging (MRI) was unremarkable. EAS was diagnosed by bilateral inferior petrosal sinus sampling (IPSS) with desmopressin stimulation, which did not detect a positive central/peripheral ACTH ratio (Table 1). A chest computed tomography (CT) revealed right lower lobe nodules up to 7 mm in size; a later [⁶⁸Ga-DOTA⁰-Tyr³] octreotate (DOTATATE) positron emission tomography (PET) scan only showed physiologic tracer uptake. Hypercortisolism was controlled on ketoconazole, 200 + 400 mg daily. Annual imaging with pituitary MRI and CT chest/abdomen/pelvis in the following 2 years was unchanged.

Table 1.

Bilateral IPSS results for case 1

Time (mins)	ACTH Peri	ACTH Right	ACTH Left	ACTH R/P	ACTH L/P	PRL Peri	PRL Right	PRL Left	PRL R/P	PRL L/P
IPSS #1 (without PRL)
−5	44	42	44	1.0	1.0	—	—	—	—	—
+3	61	95	133	1.6	2.2	—	—	—	—	—
+5	107	123	131	1.2	1.2	—	—	—	—	—
+10	120	140	143	1.2	1.2	—	—	—	—	—
+15	112	119	125	1.1	1.1	—	—	—	—	—
IPSS #2 (with PRL)
−5	31.9	27.7	425	0.9	13.3*	44.2	51.3	102.6	1.2	2.3*
0	26.5	23.9	440	0.9	16.6*	39.6	53.0	103.8	1.3	2.6*
+3	45.8	56.7	4250	1.2	92.8*	36.3	49.9	110.6	1.4	3.0*
+5	54.7	69.9	3315	1.3	60.6*	37.3	44.5	104.9	1.2	2.8*
+10	69.5	80.5	945	1.2	13.6*	34.3	42.2	77.6	1.2	2.3*

Central/peripheral ACTH ratios of ≥2 and ≥3 before and after desmopressin stimulation, respectively, are indicative of a pituitary source of excessive ACTH production. Central/peripheral prolactin ratios ≥1.8 are indicative of adequate sampling. Positive ACTH and PRL ratios are marked with an asterisk (*).

ATCH is pg/mL; PRL is mcg/L.

Abbreviations: IPSS, inferior petrosal sinus sampling; L/P, left/peripheral ratio; Peri, peripheral; PRL, prolactin; R/P, right/peripheral ratio.

She was referred to us for a second opinion before possible bilateral adrenalectomy. Ketoconazole was paused for 6 weeks to confirm hypercortisolism, and in the month prior to admission weekly UFCs were consistently elevated (85-125 mcg/d). A high dose 8 mg overnight dexamethasone suppression test (HDDST) showed 86% cortisol suppression, and a 10 mcg desmopressin stimulation test showed increases from baseline in the mean 15 + 30 minute ACTH (+145%) and 30 + 45 minute cortisol (+87%) values, consistent with a pituitary source of excess ACTH, ie, CD. Repeat IPSS (Table 1) showed a positive central/peripheral ACTH ratio before and after desmopressin stimulation, also consistent with CD. Concurrent prolactin levels indicated adequate sampling of the left, but not the right, petrosal sinus. A pituitary MRI showed a left-sided 6 mm hypo-enhancing focus, and the patient underwent transsphenoidal resection of an ACTH-positive microadenoma, confirmed on pathology. Cortisol and ACTH were undetectable within 1 and 4 days, respectively.

Case 2

A 55-year-old woman presented with rapid weight gain, new hypertension, proximal muscle weakness, hirsutism, depression, and increased aggression. After admission for acutely worsening muscle weakness and psychosis, cortisol was 209 mcg/dL, ACTH 316 pg/mL, and potassium 1.8 mmol/L. Metyrapone (250 mg every 8 hours) and octreotide (50 mcg subcutaneously every 6 hours) were started but discontinued when she was transferred to our institution to undergo IPSS. On the morning of IPSS, 12 days after the initial hospitalization, cortisol was 5.1 mcg/dL. Pituitary MRI was unremarkable, while IPSS (Table 2) suggested a pituitary source of ACTH. Five days later serum cortisol was 93.5 mcg/dL and ACTH was 114 pg/mL. A subsequent CRH stimulation test and HDDST both suggested EAS. Imaging revealed a 1.4 cm retrocardiac nodule on CT and MRI, which was not avid on OctreoScan or ¹⁸F-dihydroxyphenylalanine (DOPA) PET. After the nodule was resected, ACTH and cortisol decreased to undetectable and 2.4 mcg/dL, respectively, within 72 hours. Pathology revealed a pulmonary NET with ACTH-positive immunohistochemistry. By postoperative day 10, her cortisol levels had regained a normal diurnal pattern (18.6 mcg/dL in the morning and 4.6 mcg/dL at midnight) and ACTH was normal (23.7 pg/mL).

Table 2.

Bilateral inferior petrosal sinus sampling results for case 2

Time (mins)	ACTH Peri	ACTH Right	ACTH Left	ACTH R/P	ACTH L/P
−5	18.3	25.5	25.1	1.4	1.4
0	18.7	24.5	26.8	1.3	1.4
+3	19.5	68.3	106.0	3.5*	5.4*
+5	22.4	128.0	212.0	5.7*	9.5*
+10	33.8	115.0	161.0	3.4*	4.8*

Central/peripheral ACTH ratios of ≥2 and ≥3 before and after CRH stimulation, respectively, are indicative of a pituitary source of excessive ACTH production. Positive ACTH ratios are marked with an asterisk (*).

ATCH is pg/mL.

Abbreviations: L/P, left/peripheral ratio; Peri, peripheral; R/P, right/peripheral ratio.

Discussion Cases 1 and 2—IPSS Pitfalls

As per the Endocrine Society guideline for CS diagnosis, hypercortisolism is first confirmed with ≥2 screening modalities (LDDST, UFC, and/or bedtime salivary cortisol) [17]. Then ACTH dependence is assessed: an ACTH >20 pg/mL confirms ACTH dependence and need for a pituitary MRI. If MRI shows a ≥ 6 mm adenoma, the patient may proceed to transsphenoidal resection or undergo additional noninvasive tests [18, 19]. Smaller pituitary lesions or unremarkable pituitary imaging warrant further testing to distinguish CD from EAS. Classically, compared to CD, EAS is associated with higher cortisol and ACTH levels, more severe clinical manifestations, and a more rapid onset. However, many EAS patients present with a more indolent clinical course with a biochemical picture that overlaps the typical presentation of CD [20].

Diagnostic tests should be highly specific to enhance the high pretest probability of CD (∼85%) compared to EAS. Among noninvasive options for diagnosing CD, the CRH stimulation test is the most accurate [20] but has been unavailable after CRH was withdrawn from the market [21]. The 10 mcg desmopressin stimulation test, performed like the CRH stimulation test, is another option that has been less extensively validated. Using the same response criteria as the CRH test, this test has shown promise [sensitivity 77-92%; specificity 40-100%; positive predictive value (PPV) 82-92%; negative predictive value (NPV) 40-61%] [22-25], but the optimal response criteria remain unclear [26, 27]. The HDDST performs well when positive (PPV for CD 96-98%) with optimized cortisol suppression criteria (≥−52.7-69%) but is not helpful when negative (NPV for EAS 49-55%) [20, 28].

Successful performance of the diagnostic gold standard, IPSS, requires a high level of technical expertise. It is also costly and invasive, with a small risk of complications. In experienced hands, IPSS with CRH stimulation performs very well (diagnostic accuracy 98%; PPV 99%; NPV 95%) [20]. Desmopressin has been substituted for CRH with promising results, although specificity needs further validation with larger numbers of EAS patients [29, 30]. Notably, it was recently shown that IPSS can be avoided in up to 47% of patients if imaging for an ectopic tumor is performed in addition to CRH and desmopressin stimulation tests [31], but this approach is currently not feasible due to the unavailability of CRH.

Cases 1 and 2 demonstrate 2 pitfalls in IPSS interpretation that, respectively, cause false-negative and false-positive results. In case 1, IPSS initially suggested EAS (probably due to inadequate catheterization) but subsequently correctly diagnosed CD. Successful sampling requires both correct catheter placement (assessed with fluoroscopy) and normal pituitary venous drainage. When IPSS shows a positive central/peripheral ACTH ratio (≥2 before and ≥3 after desmopressin stimulation), adequate sampling can be assumed since this would be the only explanation for higher ACTH levels in the petrosal samples. However, when a positive ratio is not seen, adequate sampling needs to be confirmed by appropriate catheter placement and measurement of prolactin, which reflects pituitary venous drainage. A central/peripheral ratio of prolactin ≥1.8 confirms adequate cannulation and venous flow [32]. It should also be noted that peripheral ACTH levels increased by +173% after desmopressin stimulation during the initial false-negative IPSS, a response indicative of CD.

Testing for the cause of ACTH-dependent CS requires the normal pituitary corticotropes to have been suppressed by long-standing exposure to the negative feedback effects of hypercortisolism (Fig. 1). Case 2 highlights the consequences of incomplete suppression that may occur in cyclic or medically controlled CS patients. In cyclic CS, periods of hypercortisolism are interspersed with periods of normal or even low cortisol, the latter due to persistent suppression of normal corticotropes [33]. If the tumor is inactive and suppression is incomplete, any response will come from the normal corticotropes, resulting in a positive central/peripheral ACTH ratio during IPSS, a positive ACTH/cortisol response to CRH stimulation, and successful cortisol suppression with dexamethasone. These normal pituitary responses would then be incorrectly interpreted as confirming CD. Notably, desmopressin only increases ACTH/cortisol in a minority of healthy controls [25, 34], which should reduce the false-positive rate in this situation. To avoid having to repeat IPSS, we routinely obtain once-weekly UFCs for 3 to 6 weeks before the procedure to rule out cyclicity or potential residual medication effects. This is particularly important after osilodrostat use, as it is associated with unexpected prolonged adrenal insufficiency after discontinuation [35-37]. We also evaluate periprocedural cortisol levels, ideally with a fast turnaround assay such as bedtime serum cortisol, to inform a decision to proceed or cancel the IPSS.

Figure 1.

Tests to determine the source of excessive ACTH-secretion require corticotrope suppression. In normal physiology (left), hypothalamic CRH induces ACTH secretion from pituitary corticotropes. In Cushing disease (center) and ectopic ACTH syndrome (right), tumoral sources of ACTH induce excessive adrenal cortisol production, which inhibits healthy CRH- and ACTH-secreting neurons in the hypothalamus and pituitary, respectively. The width of the arrows and font sizes indicate the greater amount of hormone secretion, adrenal stimulation (by ACTH), and negative feedback to the hypothalamus and pituitary (by cortisol) in patients with Cushing syndrome.

In case 2 the IPSS was falsely positive because it was performed during a period of low disease activity, evidenced by the morning cortisol (5.1 mcg/dL) and the baseline peripheral ACTH, which had dropped to 18.7 from 316 pg/mL 12 days earlier. Importantly, the decrease in cortisol was greater than would be expected with small, short-acting doses of metyrapone and octreotide, indicating that it was due to cyclic CS rather than medication effects. Cyclicity can occur with any form of CS and is not pathognomonic for EAS [33]. Another indicator that should raise suspicion of false positivity in case 2 is the relatively low petrosal sinus ACTH values, which generally are much more robust in CD—typically ≥400 pg/mL [38]. Luckily, this patient avoided unnecessary pituitary exploration since the suspicion of EAS was very high due to the rapid onset of severe manifestations and extremely high cortisol, especially in the setting of an unremarkable pituitary MRI.

Case 3

A 49-year-old man was hospitalized after a syncopal event. Eighteen months earlier, his pre-existing hypertension had worsened, and he developed type 2 diabetes mellitus and hypokalemia, with later proximal muscle weakness and fatigue. On admission, hypokalemia was noted, despite supplementation, and UFC was elevated. Because repeat testing for hypercortisolism was unremarkable, biweekly screening was initiated and returned elevated cortisol levels 4 months later. By this time the patient had developed facial rounding, flushing, and mood lability. The following month, cortisol was again unremarkable.

The patient noted a cyclical pattern of symptoms every 3 months, including obsessive/compulsive behavior, increased energy and appetite, weight gain, worsened glycemic control, and proximal muscle weakness. On admission to our institution, midnight cortisol (55.4 mcg/dL, ULN 7.5), morning cortisol (69.7 mcg/dL, ULN 25.0), and ACTH (532 pg/mL, ULN 46) were markedly increased. Pituitary MRI showed a small area of hypo-enhancement on the right. A CRH stimulation test and IPSS (Table 3) both indicated EAS, although concurrent prolactin levels suggested inadequate sampling. Imaging was pursued to identify the source of ectopic ACTH secretion. CT chest/abdomen/pelvis only showed a 2 mm left upper lobe lung nodule that was not seen on MRI. ⁶⁸Ga-DOTATATE PET scan showed 2 foci of increased avidity, 1 in the left neck and 1 in the right lung. Gated cardiac CT and MRI both showed a 9 mm nodule in the corresponding area of avidity in the right middle lobe, abutting the heart (Fig. 2). After a right middle lobe wedge resection, morning cortisol dropped from 35.4 mcg/dL preoperatively to a nadir of 4.2 mcg/dL on postoperative day 5. ACTH dropped from 86.5 pg/mL to a nadir of <5.0 pg/mL on postoperative day 1 but was normal (11.2 pg/mL) on day 4. Pathology showed a well-differentiated, typical pulmonary NET that stained positively for ACTH. A normal 1 mg LDDST on postoperative day 7 established remission.

Table 3.

Bilateral inferior petrosal sinus sampling results for case 3

Time (mins)	ACTH Peri	ACTH Right	ACTH Left	ACTH R/P	ACTH L/P	PRL Peri	PRL Right	PRL Left	PRL R/P	PRL L/P
−5	597.0	659.0	659.0	1.1	1.1	25.8	26.5	24.5	1.0	0.9
0	579.0	629.0	638.0	1.1	1.1	23.9	31.0	25.2	1.3	1.1
+3	509.0	523.0	526.0	1.0	1.0	21.9	25.4	21.8	1.2	1.0
+5	484.0	510.0	504.0	1.1	1.0	20.8	28.9	21.1	1.4	1.0
+10	429.0	442.0	421.0	1.0	1.0	19.9	23.8	18.1	1.2	0.9

Central/peripheral ACTH ratios of ≥2 and ≥3 before and after CRH stimulation, respectively, are indicative of a pituitary source of excessive ACTH production. Central/peripheral prolactin ratios ≥1.8 are indicative of adequate sampling. No ACTH or PRL ratios met these cutoffs.

ATCH is pg/mL; PRL is mcg/L.

Abbreviations: L/P, left/peripheral ratio; Peri, peripheral, PRL, prolactin; R/P, right/peripheral ratio.

Figure 2.

Use of gated cardiac and functional imaging to detect ectopic sources of ACTH. The right-sided pericardiac culprit lesion (arrows) was missed on chest computed tomography (A) but detected on later gated cardiac magnetic resonance imaging (B), with avidity being confirmed on a ⁶⁸Ga-DOTATATE positron emission tomography scan (C).

Case 4

A 45-year-old woman with inconsistent lab results was referred for evaluation of suspected CS. She reported 3 years of progressive symptoms, starting with hypertension and weight gain, followed by type 2 diabetes mellitus, oligomenorrhea, facial rounding, dark abdominal striae, irritability, anxiety, and insomnia. She noted a cyclical pattern, with 10-day periods of worsening symptoms (weight gain, edema, bloating, depression, and proximal muscle weakness) alternating with 2 to 3 weeks of improvement, associated with frequent diarrhea. Two years after her first symptoms, she had a myocardial infarction. Interval morning cortisol levels had ranged from 8.0 to 36.8 mcg/dL, with corresponding UFCs of 1.5 to 520.6 mcg/d (ULN 50). Pituitary MRI was unremarkable.

On admission to our institution, diurnal variation was partially intact with morning and midnight cortisol 13.2 and 6.3 mcg/dL, respectively. Morning ACTH was normal (22.8 pg/mL). A CRH stimulation test and HDDST both indicated CD. A repeat pituitary MRI showed a partial empty sella. IPSS was postponed when a UFC returned at 8.1 mcg/d. She returned 1 month later during a period of worsened symptoms. This time, labs showed midnight cortisol 24.1 mcg/dL, morning cortisol 14.8 mcg/dL, midnight ACTH 60.0 pg/mL and morning ACTH 35.5 pg/mL. Repeat CRH stimulation test and HDDST were now negative, indicative of EAS. IPSS suggested EAS as well, and adequate sampling was confirmed with concurrent prolactin (Table 4).

Table 4.

Bilateral inferior petrosal sinus sampling results for case 4

Time (mins)	ACTH Peri	ACTH Right	ACTH Left	ACTH R/P	ACTH L/P	PRL Peri	PRL Right	PRL Left	PRL R/P	PRL L/P
−5	33.8	35.6	40.0	1.1	1.2	9.5	1.2	47.2	0.1	5.0*
0	36.6	40.1	39.4	1.1	1.1	9.5	31.8	51.8	3.3*	5.5*
+3	31.0	68.0	70.7	2.2	2.3	8.1	24.9	55.8	3.1*	6.9*
+5	34.6	56.6	67.2	1.6	1.9	9.0	22.6	56.7	2.5*	6.3*
+10	32.2	67.9	74.4	2.1	2.3	8.1	13.6	37.0	1.7	4.6*

Central/peripheral ACTH ratios of ≥2 and ≥3 before and after CRH stimulation, respectively, are indicative of a pituitary source of excessive ACTH production. No ACTH ratio met these cutoffs. Central/peripheral prolactin ratios ≥1.8 are indicative of adequate sampling. Positive PRL ratios are marked with an asterisk (*).

ATCH is pg/mL; PRL is mcg/L.

Abbreviations: L/P, left/peripheral ratio; Peri, peripheral; PRL, prolactin; R/P, right/peripheral ratio.

To identify the ectopic source of ACTH, CT and MRI of the chest/abdomen/pelvis were performed but only revealed bilateral adrenal hyperplasia with a 1.2 cm left-sided nodule. OctreoScan was negative. ⁶⁸Ga-DOTATATE PET scan showed a left suprahilar focus of increased avidity without a structural correlate. ¹⁸F-DOPA PET did not reveal corresponding hilar uptake. However, both ⁶⁸Ga-DOTATATE and ¹⁸F-DOPA detected a focus of increased avidity in the right lower abdomen. This was followed by a second review of structural imaging, which revealed a thickened appendix correlating to the tracer-avid focus (Fig. 3). A colonoscopy was unrevealing. She underwent a diagnostic laparoscopy, during which a 1.5 cm mass was resected from the tip of the appendix. Pathology confirmed an ACTH-positive NET. The following morning, ACTH dropped to <5.0 pg/mL but became normal (22.7 pg/mL) 1 day later. Cortisol on postoperative day 2 was 4.6 mcg/dL, down from only 8.8 mcg/dL 2 days preoperatively. The following 3 days, cortisol ranged from 5.7 to 7.9 mcg/dL in the morning and 1.3 to 2.8 mcg/dL at midnight. Cortisol was undetectable after a 1 mg LDDST on postoperative day 5.

Figure 3.

Functional imaging is critical to detect ectopic sources of ACTH. The culprit lesion (arrows) was not identified on computed tomography (A) until the images were carefully rereviewed after a focus of tracer avidity was noted on ⁶⁸Ga-DOTATATE (B) and ¹⁸F-DOPA (C) positron emission tomography scans.

Case 5

A 51-year-old man was diagnosed with CS following 9 months of symptoms including 35 lbs weight gain, insomnia, easy bruising, hypertension, and facial rounding. Initial labs showed: UFC 90.8 mcg/d, 1 mg LDDST cortisol 18.4 mcg/dL, and morning ACTH 94 pg/mL. Pituitary MRI was unremarkable. IPSS performed at an outside institution indicated EAS. Because OctreoScan and structural imaging did not identify a tumor, ketoconazole treatment (200 mg every 8 hours) was initiated, with improvement of symptoms.

Evaluation at our institution, performed after a 6-week break in ketoconazole therapy, showed UFC 265-411 mcg/day (reference range 8-77 mcg/d) and a positive CRH stimulation test but a negative HDDST and IPSS. Over the following 5 years, the patient returned for 8 imaging sessions combining CT, MRI, OctreoScan, and gated cardiac CT and MRI, without successful tumor identification. In the meantime, medical therapy required intensification: total daily doses were ketoconazole 1600 mg, metyrapone 2000 mg, and mitotane 1000 mg. At this point, pituitary MRI and IPSS were repeated after a pause in medical therapy. MRI showed a questionable 4 mm hypo-enhancement in the posterior pituitary. IPSS again indicated EAS, but fluoroscopy revealed an anomalous small left petrosal sinus, putting the results into question. Transsphenoidal resection of 2 suspected microadenomas did not induce remission, and no tumor was seen on pathology. Soon thereafter, the patient underwent bilateral adrenalectomy, which was followed by significant symptom improvement.

One and 3 years after adrenalectomy, the patient returned for repeat imaging. On the CT 1 year after adrenalectomy, an 8 mm left lower lobe lung nodule was noted, which had been reported intermittently over the 3 preceding years. Eight years after adrenalectomy—13 years after the initial CS diagnosis—the patient again returned for imaging, at which point ⁶⁸Ga-DOTATATE PET/CT had become available. Both ⁶⁸Ga-DOTATATE and ¹⁸F-DOPA PET/CT showed 3 intensely tracer-avid foci in the lower lobe of the left lung. These foci were negative on an ¹⁸F-DOPA PET/CT done once, 11 years earlier. After a left basilar segmentectomy, ACTH dropped from 371 pg/mL to 35.1 pg/mL on postoperative day 2. Pathology revealed a well-differentiated pulmonary NET with metastases to 2 adjacent lymph nodes. Three months later, ACTH remained normal at 19.0 pg/mL, measured 1 hour after a 15 mg dose of hydrocortisone.

Discussion Cases 3, 4 and 5—Imaging to Identify a Tumoral Source of Ectopic ACTH

Once EAS has been confirmed, avoiding the previously discussed IPSS pitfalls, the goal is to identify the causative tumor with imaging. Although these tumors are most commonly located in the chest, they can occur anywhere from the neck to the pelvis (Table 5). Aggressive, malignant tumors are generally easy to detect as they tend to be large with high metabolic activity, making them apparent on structural imaging and easily confirmed by FDG-PET [39]. In contrast to these aggressive tumors, the NETs causing EAS are typically slow-growing with low metabolic activity, often making FDG-PET unhelpful [39-41]. Additionally, they may be quite small and/or located in hard-to-image areas and therefore be easily disregarded or missed due to adjacent structures [1]. Because of these challenges, functional scans are essential to increase both the sensitivity and the specificity of lesions detected on structural imaging. Often, tumors are found only after an area with functional avidity undergoes a careful second review of structural imaging.

Table 5.

Approximate distribution of tumors causing ectopic ACTH syndrome [3-12]

Tumor type	%
Pulmonary NET	48
Pancreatic NET	13
Lymph node metastasis with unknown primary	12
Thymic NET	6
Small cell lung cancer Pheochromocytoma	5 4
Medullary thyroid cancer	3
Gastrinoma	3
Ileal or appendiceal NET	2
Others	<1 each

This table summarizes 10 publications reporting a total of 361 cases of ectopic ACTH syndrome. Note that 64 patients (18%) had occult tumors and are not included. The group “Others” includes pulmonary tumorlets, stomach NET, small cell colon cancer, lung cancer (unspecified), mesothelioma, disseminated NET, small cell cervical cancer, esthesioneuroblastoma, and retroperiotenal NET.

Abbreviation: NET, neuroendocrine tumor.

The most commonly used functional scans target cellular characteristics of NETs, most of which express somatostatin receptors (SSTRs) [42]. OctreoScan, previously the standard modality for this indication, uses ¹¹¹In-pentetreotide ([¹¹¹In-DTPA⁰] Octreotide), which binds to SSTRs 2 and 5 and is detected by gamma cameras [39]. OctreoScan has largely been replaced by ⁶⁸Ga-linked ligands, which have higher affinity to SSTR 2 and use PET technology, resulting in higher spatial resolution [42]. In the United States, DOTATATE is the Food and Drug Administration-approved ⁶⁸Ga-linked ligand for NET detection and is now the functional modality of choice [4]. Other ⁶⁸Ga-linked ligands commercially available in Europe have shown similar diagnostic accuracy [5, 12, 43, 44]. Although not commercially available, ¹⁸F-DOPA is another functional PET/CT ligand taken up by tumors with increased L-DOPA decarboxylase activity and has been used in research settings for NET detection with promising results [43, 45]

UFCs were alternatively normal and elevated in case 3, in parallel with variable symptoms of hypercortisolism. If there still is a high suspicion of cyclic CS, we either obtain UFCs and/or bedtime salivary cortisol at regular 1- to 2-week intervals, as in this case, or ask the patient to collect samples during periods of worsening symptoms. Worsening blood pressure or glycemic control can be particularly useful to determine appropriate times for sample collections. Case 3 demonstrates the value of gated cardiac CT and MRI, during which telemetry is paired with image acquisition, allowing for selection of reconstructed images obtained at the end of ventricular emptying when the heart is smallest. This strategy maximizes the imaged pericardiac area and minimizes motion artefacts. Combined with functional (⁶⁸Ga-DOTATATE) imaging, this approach detected and confirmed a surgical target whose resection led to remission.

Functional imaging was crucial for detection in case 4, as the tumor was obscured by bowel on the CT scan. This case highlights that NETs can appear in unexpected places: a careful review beyond the chest cavity is always indicated. A colonoscopy was performed before surgery to exclude bowel inflammation, which is an alternative explanation for DOTATATE avidity, as leukocytes and macrophages express SSTR2. When using SSTR PET ligands, it is important to consider false-positive results from physiologic and pathologic uptake in non-ACTH-producing tissue (Table 6) [46].

Table 6.

Physiologic uptake and pathology with increased somatostatin receptor ligand avidity

Physiologic uptake		Pathologic processes
Spleen	Pancreatic uncinate process	Osteoblastic activitya	Inflammation, incl. reactive lymph nodes
Adrenal glands	Salivary glands	Meningioma	Medullary thyroid cancerb
Pituitary	Thyroid	Pheochromocytoma/ paragangliomab	Small cell carcinomab
Kidneys/ureters/bladder	Liver	Esthesioneuroblastomab	Tumor-induced osteomalacia
Stomach	Small and large intestine	Neuroblastoma	Iodine-refractory thyroid cancer

When using somatostatin receptor positron emission tomography ligands, such as ⁶⁸Ga-DOTATATE, these sites can cause false positives.

^a Degenerative bone disease, fractures, fibrous dysplasia, vertebral hemangiomas.

^b Potential cause of ectopic ACTH syndrome.

In case 4, periods of decreased disease activity were accompanied by frequent diarrhea, potentially a symptom of intermittent adrenal insufficiency. IPSS during the first admission was deferred at the last minute because of cyclicity; this may have been avoided if testing for hypercortisolism was done prior to the admission, as discussed earlier. This case again highlights that dynamic tests require previous prolonged exposure to hypercortisolism, as the results of both CRH stimulation and HDDST changed between admissions.

Cyclic disease activity, or prolonged eucortisolemia during medical therapy, affects not only dynamic testing but also the postoperative biochemical assessment. When successful resection of an ACTH- or cortisol-producing tumor is achieved in more typical cases of CS with persistent hypercortisolism, the healthy corticotropes remain suppressed for at least 6 months, resulting in secondary adrenal insufficiency that necessitates physiologic glucocorticoid replacement [47]. Postoperative cortisol and ACTH levels are therefore used to establish remission. In our cases with cyclic CS (cases 2-4), ACTH and cortisol dropped postoperatively but rose to detectable levels within a week. To differentiate persistent CS from the return of normal corticotrope function, the usual screening tests for hypercortisolism (LDDST, UFC, and/or bedtime salivary cortisol) should be used in the postoperative period [17].

Case 5 shows EAS without cyclicity and with a slow progression of symptoms and biochemical changes. IPSS correctly indicated EAS on 3 separate occasions, but the results were questioned due to an anomalous left petrosal sinus on fluoroscopy, leading to an unnecessary transsphenoidal exploration. This may have been avoided had concurrent prolactin levels been available to prove adequate sampling. As shown in case 5, adrenalectomy becomes indicated due to the inability to control the hypercortisolism medically but is also indicated for intolerance to medications or patient preference when the tumor is unresectable or occult. Apart from rendering the patient permanently adrenally insufficient, adrenalectomy also makes the biochemical testing of CS very difficult as normal corticotropes are more resistant to negative feedback from exogenous glucocorticoids. Therefore, efforts should be made to complete the biochemical diagnostic workup before adrenalectomy, if possible.

Although occult ectopic tumors are typically slow-growing, there is a risk of growth and metastasis. Because of this, attempts to identify the tumor should not stop after adrenalectomy has resolved the hypercortisolism. In addition to the tumors slowly growing and potentially becoming easier to identify, imaging technology continues to evolve, further improving the chances of detection. In case 5 the culprit lesions were identified thanks to ⁶⁸Ga-DOTATATE imaging once it became available, 13 years after the initial diagnosis. In retrospect, the tumor had been visible for 11 years but was disregarded since it was <1 cm, did not display growth, and was not avid on OctreoScan.

After successful resection in patients with EAS, there is always a chance of growth and spread of unidentified micrometastases, resulting in CS recurrence. In this adrenalectomized patient, however, a recurrence would not manifest with hypercortisolism. Periodic ACTH measurements may be useful in this scenario, with a clearly increasing trend warranting reimaging to identify remaining tumor tissue.

Case 6

A 66-year-old woman with presumed EAS from an occult tumor based on IPSS, CRH stimulation, HDDST, and unrevealing imaging results was well controlled with metyrapone 500 mg every 8 hours for 4 years. However, her previous signs of CS then returned (hypertension requiring 4 agents, easy bruising, fatigue, muscle weakness, facial rubor, hirsutism, and hyperglycemia). She also experienced weight loss of 20 lbs, recurrent urinary tract infections, insomnia, and anxiety. Metyrapone was increased to 2500 mg per day but serum cortisol continued to increase to more than twice the normal.

She was admitted to attempt to achieve medical control and/or localize the causal tumor. On admission blood pressure was 164/85 mmHg, potassium 2.3 mmol/L, UFC 8000 mcg/d (ULN 77), serum cortisol 47 mcg/dL (ULN 25), and ACTH 285.4 pg/mL (ULN 52). Ketoconazole (200 mg every 6 hours) was added but liver function tests increased to 3 times ULN. Based on new imaging, a right hilar mass was resected. Unfortunately, the pathology showed lymph nodes without tumor, and cortisol levels remained elevated. Mitotane (250 mg at bedtime) was added and increased to 1000 mg within 7 days, with a decrease in cortisol (from 83.4 to 32.8 mcg/dL). She developed chylothorax and hypoalbuminemia, presumably from thoracic duct damage at surgery, and required a surgical repair on postoperative day 7. One day later, she developed delirium with incoherent speech and disorientation. Since mitotane may have exacerbated her postsurgical delirium, it was discontinued. In view of her worsening condition, and without alternative medical treatments, etomidate was initiated and bilateral adrenalectomy was scheduled.

With upward titration of etomidate, delirium resolved and cortisol was controlled, decreasing from 142 mcg/dL to between 14.6 and 24.0 mcg/dL during the 8 days before adrenalectomy. Her postoperative recovery was uneventful. A few years later, a grade 1 pulmonary NET was successfully identified and resected.

Discussion Case 6—Urgent Treatment of Severe, Refractory Hypercortisolism

This case illustrates the unpredictable nature of some EAS tumors, with loss of medical control, and the need to restore eucortisolism quickly. Severe hypercortisolism (UFC >8-10-fold ULN) is associated with atypical infections related to immunosuppression, such as P. jirovecii pneumonia [48], pulmonary embolism, and severe hypokalemia, and has an increased mortality. Therefore, hypercortisolism should be reduced as quickly as possible; prophylaxis for P. jirovecii and thromboembolism should be given; and comorbidities such as hypertension, hypokalemia, and hyperglycemia should be treated aggressively [49]. There are no prospective trials of medical treatment in this setting. However, rapid control can be achieved with combination therapy using ketoconazole and metyrapone, with the addition of mitotane if needed [50]. Although additional agents are currently available, those directed at pituitary disease would not be helpful in EAS [51], and it would be difficult to monitor treatment response to a glucocorticoid antagonist paired with a steroidogenesis inhibitor. The 2 newer agents levoketoconazole and osilodrostat mimic the actions of ketoconazole and metyrapone, respectively, without clearly offering additional benefit [52]. Although gradual titration of osilodrostat is recommended, 1 study supported rapid titration in patients with severe hypercortisolism [53]. Case 6 illustrates potential adverse outcomes of right lung resection (transection of the thoracic duct), as well as the salutary outcome of bilateral adrenalectomy; the latter also being demonstrated by case 5.

Summary

In patients with ACTH-dependent CS that lack a ≥ 6 mm lesion on pituitary MRI, an accurate determination of the etiology (CD vs EAS) is crucial to target further investigations and therapy. Unless an obvious source of ectopic ACTH production is present, this distinction is achieved through dynamic testing, recognizing that clinical and biochemical presentations overlap and that CD is significantly more common. While IPSS is the gold standard, false positives and negatives do occur and can be avoided by confirming ongoing hypercortisolism leading up to the procedure and by routinely obtaining concurrent prolactin levels to confirm adequate petrosal sampling. If IPSS is unavailable, HDDST is the second best currently available test to diagnose CD, if an optimized response criterion is used (PPV 96% for ≥69% cortisol suppression). Unlike IPSS, however, HDDST is not helpful when negative.

Once CD has been confidently excluded, structural (CT/MRI) and functional (preferably ⁶⁸Ga-DOTATATE PET/CT) imaging of the chest/abdomen/pelvis should be performed, keeping in mind that tumors with ectopic ACTH production can be located anywhere from the neck to the pelvis, even if intrathoracic lesions are most common. Functional scans are essential to increase lesion specificity and to increase sensitivity by identifying areas of avidity warranting a careful second review of structural imaging. If no lesion is found, gated imaging can be helpful to identify pericardiac lesions. If the tumor remains occult, periodic imaging should not stop even if eucortisolism is achieved, since successful resection eliminates the risk of tumor spread and restores the normal hypothalamic-pituitary-adrenal axis. Severe hypercortisolism warrants prophylaxis for infection and thrombosis, optimization of comorbidities, and rapid titration of steroidogenesis inhibitors, such as ketoconazole or metyrapone. If refractory, rapidly effective last resorts for hypercortisolism include etomidate infusion and bilateral adrenalectomy.

While EAS often poses a diagnostic dilemma, avoidance of the biochemical diagnostic pitfalls and following the imaging strategies described in our cases should reduce inappropriate surgical interventions and improve the chances of successful tumor detection.

Funding

This work was funded by the intramural division of the National Institute of Diabetes and Digestive and Kidney Disorders.

Disclosures

H.E. has nothing to disclose. L.K.N. receives royalties from UpToDate.

Data Availability

No new data were generated or analyzed in support of this research.