Management of Adrenocorticotropic Hormone-Secreting Neuroendocrine Tumors and the Role of Bilateral Adrenalectomy in Ectopic Cushing Syndrome

Abstract

Introduction:

Neuroendocrine tumors (NET) can cause ectopic Cushing syndrome (ECS) and most patients have metastatic disease at diagnosis. We identified risk factors for outcome, evaluated ECS management, and explored the role of bilateral adrenalectomy (BA) in this population.

Methods:

Retrospective study including NET-ECS treated at our quaternary referral center over a forty-year period (1980–2020).

Results:

Seventy-six patients were included. Mean age at diagnosis was 46.3±15.8 years. Most patients (N=61, 80%) had metastases at ECS diagnosis. Average follow-up was 2.9±3.7 years (range, 4 months-17.2 years). Patients with NET prior to ECS had more frequent metastatic disease and resistant ECS. Patients with de novo hyperglycemia, poor NET differentiation, and metastatic disease had worse survival. Of those with non-metastatic disease, 8 (53%) had ECS resolution after NET resection, 3 (20%) were medically controlled, and 4 (27%) underwent BA. In patients with metastatic NET, hypercortisolism was initially medically managed in 92%, 3% underwent immediate BA, 2% had control after primary NET debulking, and 2% were lost to follow-up. Medical treatment resulted in hormonal control in 7 (13%) patients. Of the 49 patients with metastatic disease and medically resistant ECS, 23 ultimately had BA with ECS cure in all.

Conclusion:

Patients with NET prior to ECS development were more likely metastatic and had worse survival. De novo hyperglycemia and poor NET differentiation were predictive of worse prognosis. Medical control of hypercortisolism is difficult to achieve in patients with NET-ECS. Well-selected patients may benefit from BA early in the treatment algorithm, and multidisciplinary management is essential in this complex disease.

Article summary

We identified predictive risk factors for outcome and showed that medical control of hypercortisolism is difficult to achieve in patients with ectopic Cushing syndrome. Well-selected patients may benefit from surgical bilateral adrenalectomy early in the treatment algorithm.

INTRODUCTION

Ectopic Cushing syndrome (ECS) is a rare and complex paraneoplastic syndrome caused by excess glucocorticoids production often associated with ectopic adrenocorticotrophic hormone (ACTH) or CRH originating outside of the pituitary gland or the hypothalamus respectively.^1–3 ECS can present 1) as the predominant clinical manifestation with an asymptomatic and sometimes difficult to localize tumor or 2) following diagnosis of a malignancy that may confound and delay the diagnosis of ECS.⁴ A wide variety of histologies have been associated with ECS, although the majority are caused by neuroendocrine tumors (NETs) (Also referred as neuroendocrine neoplasms by some groups⁵). The most common location of NETs causing ECS is intrathoracic (40–60%), with bronchial tumors, small cell carcinoma of the lung and thymic carcinomas. Other anatomic locations where these tumors can be found are pancreas, thyroid (specifically medullary thyroid carcinoma) and rare locations like prostate, rectum, ovaries, bladder, among others.^6–10

ECS can manifest with classic Cushingoid features such as truncal obesity, skin striae, hyperpigmentation, moon facies, muscle wasting, and weakness as well as metabolic derangements such as hypertension, hyperglycemia, and hypokalemia. Unfortunately, many of these signs and symptoms may occur among patients with comorbid diseases, other malignancies, or those who may be critically ill.^{3, 4} Uncontrolled Cushing syndrome leads to profound weakness and devastating decrease in quality of life. These include increased cardiovascular risk, musculoskeletal impairment, metabolic disorders, infections, thrombosis, decreased quality of life, and worse survival.^{6, 11}

The complications of ECS may be compounded in a patient already stressed affected by the disease burden of metastatic NET. The metabolic consequences often are so debilitating that management of the primary tumor can not be accomplished because of the feeble physical state. When feasible, resection of the primary NET is ideal, but many patients have unresectable or metastatic disease at diagnosis.^{12, 13} Management of ECS in the metastatic NET setting consists of upfront medical control of hypercortisolism as soon as the diagnosis is established; however, it has been reported that effective preoperative medical treatment in patients with Cushing syndrome using steroidogenesis inhibitors such as ketoconazole, metyrapone or a combination of these drugs is only successful to achieve biochemical control (urinary free cortisol within 50–280 nmol/24 hrs.) in 52% of patients, with symptoms, signs and comorbidities resolution in only 32% of patients.¹⁴ Bilateral adrenalectomy can immediately eradicate the source of cortisol and drastically improve signs and symptoms of ECS. Postoperatively it does require lifelong hormone replacement therapy to avoid adrenal insufficiency.^{9, 13, 15}

Data on prognostic risk factors and outcomes in ECS is limited due to its rarity, and life-expectancy may depend on a variety of factors such as NET type, disease status, symptom burden, and type of management.⁶ Our study aim was to evaluate clinical characteristics, prognostic risk factors, and the role of bilateral adrenalectomy (BA) in patients with NET-ECS.

METHODS

Study Population

Following approval from The University of Texas MD Anderson Cancer Center’s Institutional Review Board, we conducted a comprehensive retrospective chart review of patients with NET-ECS who were treated and evaluated at our quaternary referral center from 1980–2020 (N=76). Patients eligible for inclusion were identified through querying the (1) electronic health record for a diagnosis of ECS or a combined diagnosis of Cushing syndrome and carcinoid/NET and (2) the department of surgical endocrinology and endocrine neoplasia patient databases. Some of these patients were included in previous reports from our institution.^{1, 9} Patients with iatrogenic hypercortisolism (N=35), Cushing disease (N=103), ACTH-independent Cushing syndrome (N=107), or with a non-NET source of ectopic ACTH (N=9) were excluded.

Cushing syndrome diagnosis was based on clinical features of hypercortisolemia, and biochemical confirmation of extra-pituitary, ACTH- dependent hypercortisolism diagnosed through the accepted standard of care at the time of the diagnosis.^{16, 17} Routine screening in patients with neuroendocrine tumors is not done as the expected prevalence of ectopic ACTH secretion related to NET remains low. Alternatively, we have low threshold to test patients in the presents of triggering factors such as weight changes, new onset or worsening hypertension or hyperglycemia, or other signs or symptoms suggestive of cortisol excess.

The workup at our institution begins establishing the presence of cortisol excess, before the cause of this excess is sought. We do screening with 1-mg overnight dexamethasone suppression and late-night salivary cortisol level and less frequently with 24-hour urinary free cortisol. Once hypercortisolism is documented, the workup focuses on assessing whether excessive ACTH secretion is the cause. Abnormally low ACTH levels are consistent with ACTH-independent CS resulting from autonomous cortisol secretion of the adrenal gland. After this, abdominal imaging is obtained to evaluate the adrenal glands’ morphology. A normal or high ACTH level is consistent with ACTH-dependent CS and pituitary MRI is used as first diagnostic imaging study. Bilateral inferior petrosal sinus sampling is used as the reference standard for differentiating between pituitary and non-pituitary sources of ACTH. CRH stimulation test is used to determine whereas the presence of ectopic ACTH secretion. Radiological evaluation for ECS consists of computed tomography (CT), magnetic resonance imaging (MRI), ¹²³I-meta iodobenzylguanidine (¹²³I-MIBG), ¹¹¹In-octreotide, ⁶⁸Ga-DOTATATE, and/or ¹⁸Fluorodeoxyglucose (¹⁸FDG) scans. Unknown primary source of ECS was considered by the treating physician when, after exhaustive workup with the algorithm described before, none primary tumor was localized. Resolution of ECS was defined as undetectable/normal serum or urine cortisol levels with concomitant clinical improvement in signs and symptoms of ECS. Resistant ECS was defined as hypercortisolemia managed with medical suppression with persistently elevated 24-hour urine, midnight salivary cortisol levels and if any of the later were not available, random cortisol levels in association with clinical features attributed to cortisol excess. All patients who underwent BA were treated postoperatively with replacement mineralocorticosteroids (fludrocortisone) and glucocorticosteroids (hydrocortisone).

Statistical Analysis

Demographics and clinical characteristics were evaluated utilizing descriptive statistics. Continuous variables were compared with Mann Whitney U test and dichotomous variables were analyzed with Fisher exact test. Disease-specific survival (DSS) was calculated from the date of ECS diagnosis until death due to NET as the final outcome. Survival analysis was conducted using Kaplan-Meier curves to visualize DSS, and COX regression model to estimate hazard ratios (HRs) and 95% confidence intervals (Cis) for factors associated with an increased risk of DSS. P values less than 0.05 were considered statistically significant and P values less than 0.10 were considered trending. Statistical analyses were performed using SAS Enterprise Guide 7.15 (SAS Inc. Cary, NC).

RESULTS

Demographic characteristics

In total, 76 patients with NET-ECS were included (Figure 1). Mean age at diagnosis was 49.6±15.9 years and 57.9% (44) were female. Mean follow-up was 2.9 ± 3.7 years (range 4 months-17.2 years) after ECS diagnosis. NET origins were lung in 35% (small cell lung carcinoma in 12 patients and bronchial in 14 patients, N=26), medullary thyroid cancer (MTC) in 20% (N=15), pancreatic (panNET) in 19% (N=15), unknown in 12% (N=9), thymic in 11% (N=8), and other (uterus, prostate and bladder) in 4% (N=3). Most patients (N=61, 80%) had metastases at ECS diagnosis. Twenty-three patients (30%) were diagnosed with ECS at a median of 1.7 months (IQR_25–75 0.41–18.6) prior to finding a NET source, while 53 (70%) had NET diagnosis at a median of 16 months prior (IQR_25–75 1.03–48.83) to ECS diagnosis.

Figure 1.

Neuroendocrine tumor associated ectopic Cushing syndrome cohort inclusion and exclusion criteria

Comparison by first manifestation (ECS vs NET)

Clinical and management characteristics were compared between patients who developed ECS vs NET as their initial diagnosis (Table 1). Patients who developed ECS prior to NET diagnosis were more likely to present with obesity and and have a non-metastatic NET. Patients with NET diagnosis prior to development of ECS were more likely to have metastatic NET and persistent hypercortisolism despite medical cortisol suppression.

Table 1.

Clinical and management characteristics associated with ectopic Cushing syndrome at diagnosis for patients with neuroendocrine tumor source (N=76)

Factors	ECS diagnosis prior to NET N=23 (30%)	NET diagnosis prior to ECS N=53 (70%)	P value
Median age at ECS diagnosis, years (SD; range)	49.8 (17.1; 11.77–3.6)	53.6 (15.4; 11.7–78.3)	0.112
Median age at NET diagnosis, years (SD; range)	49.8 (15.9; 22.7–73.7)	51.4 (15.4; 11.6–78.2)	0.409
Sex
Male	9 (39.1)	23 (43.4)	0.804
Female	14 (60.9)	30 (56.6)
Signs and symptoms at ECS diagnosis
Median BMI, kg/m2 (SD; range), n=72	30.5 (7.07; 21−47)	26.0 (7.8; 17−58)	0.016
Hypertension	21 (91.3)	37 (69.8)	0.075
Hyperglycemia (de novo)	15 (65.2)	36 (67.8)	0.999
Hypokalemia	17.(73.9)	44 (83)	0.365
Weakness	18 (78.3)	49 (92.5)	0.119
Thromboembolism	3(13)	7(13)	0.985
Infections	2(9)	6(11)	0.999
Primary NET
Lung	10(43.5)	16 (30.2)	0.250
Pancreatic	4(17.4)	11 (20.8)
Medullary thyroid carcinoma	2(8.7)	13 (24.5)
Thymic	2(8.7)	6 (11.3)
Unknown primary	5(21.7)	4 (7.6)
Othera	0	3 (5.7)
Biochemical diagnosis
Median ACTH pg/mL (range)	167 (70−1317)	160.5 (32−1118)	0.589
Median UFC mcg/24 hr (range)	545 (35−16182)	900 (132−16000)	0.481
Median LDST mcg/dL (range)	29.8 (16−1200)	36.3 (19.9−140.2)	0.801
Median HDST mcg/dL (range)	29.4 (11−341)	55.2 (25.4−58.1)	1.000
Median Salivary cortisol ng/dL (range)	3450 (175−20200)	2190 (800−27900)	0.905
Median Cortisol a.m. mcg/dL (range)	48.6 (11.5−500)	45 (18.5−597)	0.801
NET Gradeb
Grade 1	9 (39)	7 (13)	0.101
Grade 2	4 (17)	11 (21)
Grade 3	5 (22)	19 (36)
Unknown	3 (13)	3 (3)
Poorly Differentiatedbc	3 (13)	16 (30)	0.235
Metastatic NET	11 (47.8)	50 (94.3)	<0.001
Hypercortisolism management
Medical therapyd	19 (82.6)	49 (92.5)	0.234
Medical therapy only	12/19 (63.1)	32/49 (65.3)	0.999
BA referral	10 (43.5)	28 (52.8)	0.618
BA candidate	10 (43.5)	24 (45.3)	0.999
Persistent hypercortisolism	3 (13)	24 (45.3)	0.009
Eucortisolism achieved by (N=49)
Medical therapy only	4 (20)	6 (20.7)	0.044
NET resection	7 (35)	2 (6.9)
Bilateral adrenalectomy	9 (45)	21 (72.4)

Abbreviations: ECS, ectopic Cushing syndrome; NET, neuroendocrine tumor; SD, standard deviation

^aprostate NET (N=1), bladder NET (N=1), and uterine NET (N=1)

^bmedullary thyroid carcinoma (N=15) were not included in NET grading or differentiation

^cdifferentiation was unknown in 7 patients

^dketoconazole (N=11), metyrapone (N=20), both (N=18)

Of those with non-metastatic NET at ECS diagnosis (N=15), 8 (53%) had ECS resolution after NET resection, 3 (20%) were medically controlled, 3 (20%) required BA for persistent hypercortisolism despite medical treatment, and 1 (7%) underwent upfront BA without medical treatment. In patients with metastatic NET at ECS diagnosis (N=61), hypercortisolism was treated with upfront medical suppression in 56 (92%), 3 (3%) underwent immediate BA, 1 (2%) had control after primary NET resection, and 1 (2%) was lost to follow-up prior to any ECS management. Upfront medical treatment of hypercortisolism resulted in hormonal control in only 7/56 (13%) metastatic NET patients. Of the remaining 49 metastatic patients with ECS resistant to medical suppression, 29 were referred for BA, but only 23 ultimately received BA with ECS control, the rest being poor surgical candidates. 19 of 20 patients not referred for surgical treatment died by end of study, with a median survival of 6 months following ECS diagnosis.

Medical therapy

The steroidogenesis inhibitors used were ketoconazole (dose range administered (DRA) 200–1600 mg/day), metyrapone (DRA 500–6000 mg/day) and mifepristone (DRA 300–1800 mg/day). Of the 68 patients who received steroidogenesis inhibitory therapy for ECS, 30 (44 %) received combined medical treatment (N=25 ketoconazole/metyrapone, N=3 Ketoconazole/mifepristone and N=2 Ketoconazole/metyrapone/mifepristone) while 38 (56 %) received only single agent (N=13 ketoconazole and N=25 metyrapone). Side effects observed were altered liver function tests associated to ketoconazole (N=6) and nausea (N=6).

Patients treated with bilateral adrenalectomy and adjuvant systemic therapy

In total, BA was performed in 30 NET-ECS patients (39%) with no recurrent ECS. Morbidity was observed in 20% of the patients (N=6) and was classified as Clavien-Dindo Grade I 7% (N=2), Grade II 10%(N=3), Grade IV 3% (N=1), with one death documented (3%) secondary to multiorgan failure. Twenty-six (87%) of the 30 patients treated with BA had metastatic NET at the time of the surgery. The NET location in these patients was panNET in 8, unknown primary in 7, lung in 6, thymus in 5, and MTC in 4. Median time to BA following ECS diagnosis was 6 months (range, 3 days to 4.6 years).

Following BA, NET-directed systemic therapy was started at a median of 2.1 months postoperatively. Adjuvant systemic therapy was utilized in 52 patients (68%), wirh multiple lines of systemic therapy used in the majority of these patients (N=33, 63%). The most common systemic therapies were 5-FU + doxorubicin + streptozocin, platinum + etoposide, temozolamide + capecitabine, monoclonal antibodies, immunotherapy, somatostatin analogues, selective RET inhibitors, multikinase inhibitors and mTOR inhibitors, depending on location of the primary tumor disease status (metastatic vs non-metastatic) and multidisciplinary consensus.

Outcomes

At the end of the study, 22 (29%) patients survived, with a median follow-up of 3.9 years (range, 4 months-17.2 years) following ECS diagnosis. Disease specific survival (DSS) following ECS diagnosis was significantly worse in patients with NET diagnosis prior to ECS development when compared to patients with ECS prior to NET diagnosis. DSS did not show statistically significant difference when compared by primary tumor location (Figure 2). Univariate logistic regression analysis indicated that several factors like age >45, de novo hyperglycemia, hypokalemia, NET grade, poorly differentiated histology, metastasis at ECS diagnosis, controlled ECS after NET resection and uncontrolled ECS were associated with decreased DSS (Table 2). On multivariate analysis, NET-ECS patients with de novo hyperglycemia (HR 2.53 (1.21–5.27) p= 0.013), poorly differentiated NET (HR 2.44 (1.33–4.50) p=0.004), and metastatic NET (HR 6.92 (2.55–18.7=6) p= <0.001) had significantly worse DSS (Table 3). In addition, patients with medically resistant ECS who were subsequently controlled with BA had significantly better DSS following ECS diagnosis than patients who did not undergo BA (HR 0.45 (0.25–0.81) p= 0.008) (Figure 3).

Figure 2.

Disease-specific survival (DSS) of patients with neuroendocrine tumor (NET) associated ectopic Cushing syndrome (ECS). a) Patients with known NET prior to ECS diagnosis had significantly worse DSS (median 1.1 years, black dashed line) following ECS diagnosis than patients with ECS prior to NET diagnosis (median 7.3 years, blue dashed line). The 95% confidence intervals are shown in the respective dotted lines. b) Median (CI 95%) DSS by NET source was MTC 1.75 years (0.84–2.65), lung 0.92 years (0.07–1.77), panNET 3.45 years (2.59–4.30), thymus 1.39 years (0–5.25), other 0.72 years (0–1.79) and unknown primary 4.84 years (4.15–5.52), p=0.28

Table 2.

Factors associated with disease-specific survival (DSS) of patients with neuroendocrine tumor associated ectopic Cushing syndrome: Univariate analysis

Factors	Univariate
	HR (95% CI)	P value
Age at ECS diagnosis (> 45 vs ≤ 45 years)	1.91 (1.04−3.52)	0.037
Sex (female vs male)	1.13 (0.66−1.95)	0.660
Primary NET
Lung	ref.	0.187
Pancreatic	0.76 (0.35−1.64)
Medullary thyroid carcinoma	1.41 (0.68−2.91)
Thymic	0.53 (0.20−1.42)
Unknown primary	0.45 (0.15−1.33)
Signs and symptoms at ECS diagnosis
Body Mass Index	0.99(0.95−1.02)	0.425
Hypertension	0.84 (0.43−1.62)	0.6
Hyperglycemia (de novo)	1.99 (1.07−3.74)	0.030
Hypokalemia	2.13 (1.02−4.44)	0.043
Weakness	2.04 (0.81−5.14)	0.133
Thromboembolism	2.05(0.95−4.40)	0.067
Infections	1.92(0.86−4.29)	0.109
NET Grade
Grade 1	ref.	0.006
Grade 2	1.38 (0.50−3.81)
Grade 3	3.54 (1.75−7.15)
Poorly vs well differentiatedab	2.63 (1.46−4.74)	0.001
Metastasis at ECS diagnosis	6.08 (2.37−15.61)	<0.0001
Eucortisolism achieved by  BA controlled ECS vs no BA  NET resection controlled ECS  Medical therapy only	ref. 0.13 (0.02−0.99) 0.2.27 (0.99−5.21)	0.014

Abbreviations: ECS, ectopic Cushing syndrome; NET, neuroendocrine tumor; BA, bilateral adrenalectomy

Table 3.

Factors associated with disease-specific survival (DSS) of patients with neuroendocrine tumor associated ectopic Cushing syndrome: Multivariate analysis

Factors	Multivariate
	HR (95% CI)	P value
Hyperglycemia (de novo)	2.53 (1.21–5.27)	0.013
Poorly differentiated	2.44 (1.33–4.5)	0.004
Metastasis at ECS diagnosis	6.92 (2.55–18.76)	<0.001
Bilateral adrenalectomy	0.45 (0.25–0.81)	0.008

Abbreviations: ECS, ectopic Cushing syndrome

Figure 3.

Risk factors for disease-specific survival (DSS) of patients with neuroendocrine tumor (NET) associated ectopic Cushing syndrome (ECS). Patients with a) de novo hyperglycemia, b) poorly differentiated NET, and c) metastatic NET had worse DSS with a median DSS of 1.8 years, 8 months, and 1.3 years following ECS diagnosis. d) Patients who underwent bilateral adrenalectomy had better DSS (median DSS of 3.5 years) than patients who did not undergo BA (median DSS of 9 months).

DISCUSSION

Bilateral adrenalectomy as source control is beneficial in treating ECS. Our study consisted of a large cohort of patients at a single institution with NET associated ECS. We were able to evaluate 1) clinical characteristics comparing patients diagnosed with ECS before or after their NET diagnosis 2) predicative risk factors affecting survival 3) management strategies and efficacy in the setting of NET-ECS and 4) the role of BA.

Clinical characteristics of ECS

Comparable to previous reports^{3, 6, 7, 18}, almost all ACTH sources in our cohort were foregut NET or MTC in origin. The most common source of ACTH in our cohort was lung NET, followed by MTC and panNET. Despite different imaging modalities, 12% of our cohort had no identifiable primary NET by last follow-up, which is similar to previous series.^{3, 6, 7, 12} NET-ECS patients with unknown primary pose a diagnostic dilemma as presentation may be similar to Cushing disease. Although hypertension is common among all types of Cushing syndrome, hypokalemia and diabetes mellitus is much more prevalent in ECS than in pituitary or adrenal dependent Cushing syndrome.^{19, 20} NET diagnosis was able to be established in these patients through a combination of biochemical evaluation, imaging findings, and biopsy pathology suggestive of metastatic NET.

Unfortunately, most patients had metastatic NET prior to ECS diagnosis and usually these patients developed ECS long after their initial NET was discovered (median 16 months). This delay of ECS following metastatic NET diagnosis may support that 1) diagnosis of ECS may be delayed due to confounding signs and symptoms of metastatic NET and/or 2) a focus of NET cells, separate from the rest of the metastases, may develop autonomous ACTH secretion which produces ECS following initial NET presentation. Hypokalemia in patients with NET, especially the metastatic setting, should prompt ECS suspicion, notably with other features such as proximal muscle weakness, hyperglycemia, hypertension, and weight loss; a high index of suspicion is necessary as NET patients may not have time to develop typical cushingoid features such as truncal obesity, skin striae, hyperpigmentation, or moon facies because of rapid progression of ECS.¹¹ This information may be useful in surveillance of patients with MTC and foregut NET histologies and discussions of using cytoreduction as a potential option of ACTH source control or medical therapy/BA for early ECS control.

Most patients in our cohort had multiple signs and symptoms of ECS. Our findings of more Cushingoid signs and symptoms in ECS patients with occult NET than patients with NETs prior to development of ECS parallel those of Wajchenberg and colleagues.²¹ Particularly, obesity and was more common in patients with ECS diagnosis with occult NET; this may be due in part to the widespread obesity within our population, making this finding, when resulting from ECS, non-specific in someone with no prior NET diagnosis or other ECS symptoms or slow tumor progression that allows the development of a more classical phenotype of Cushing syndrome. In addition to increased rates of obesity these patients usually had a non-metastatic NET compared to those with NET prior to ECS diagnosis. In contrast, patients with NETs who later developed ECS seemed to have more aggressive disease that was difficult to medically control. Our findings are also in agreement with Aniszewski and colleagues¹² that most NETs were known diagnoses prior to ECS development, especially MTC. Once Cushing syndrome was confirmed in our patients, the following diagnosis of NET association was rapid (median 2 months). This information may be useful in diagnosis and prognosis of patients with NET-ECS as well as help guide management options.

ECS management strategies

Once the primary NET source of ECS is identified, the optimal treatment is surgical resection,^{6, 9} however most patients in our cohort had metastatic NET at the time of ECS diagnosis, making surgical resection only beneficial in a minority of our patients (12%). Medical therapy and BA still have a role in non-metastatic NET as not all cases were amenable to resection and some non-metastatic NETs only became apparent after ECS diagnosis and treatment. If metastatic NET is the source of ACTH, as were most patients in our cohort, then treatment consists of medical suppression and/or BA. As noted in our cohort, up front medical therapy to decrease cortisol is a good bridge to sequence onto surgical resection. Most patients in our cohort were treated in this manner with upfront medical therapy. Avoiding surgery as the second part of that sequence rarely is effective- as it was only successful in a minority of patients with metastatic NET (13%). Of note, 6 of these 7 patients required combined lines of suppression therapy for successful control. We have learned this over time as unfortunately, most patients in our cohort who were diagnosed with resistant ECS only used one type of suppression therapy by end of study. Although the efficacy of no NET-targeted systemic therapies in reducing hypercortisolemia burden was not evaluated in our cohort, previous reports show some success, especially with peptide receptor radionuclide therapy (PRRT)^{6, 22, 23} , although this therapy was not approved by the Food and Drug Administration (FDA) until 2018. This information is useful in guiding upfront medical suppression therapy as combined medical treatment for suppression may be required to achieve ECS control, especially in the metastatic NET setting.

Bilateral adrenalectomy

When considering BA for ECS treatment, one must consider the patient’s clinical status, disease burden, and prognosis.⁹ Because of the increased risk of wound infection and poor wound healing from hypercortisolism, minimally invasive BA offers a less morbid option than open in this patient population.¹¹ BA may be used to cure ECS as a sequenced approach and it is the timing that is most important. In one clinical scenario, the provider or patient choice for upfront BA. In this case a short course of medical management is suggested to improve the overall status of the patient to be in the best possible shape for surgical intervention. This may be a 4–6-week window of treatment, but close monitoring of overall performance status is critical to avoid worsening of symptoms. Another approach is to utilize BA following ineffective first line or not well tolerated medical suppression therapy. The importance of this is not too avoid operating in the critical window when it is still an option based on performance status and technical aspects needed for safe resection (i.e.; primary pancreas tumor not invading adrenal). NET-ECS may be a more difficult form of Cushing syndrome to suppress medically compared to Cushing disease as inability to tolerate medical suppression therapy was not reported in our cohort, despite many patients also receiving NET-targeted systemic therapies as well; in addition, some patients were unsuccessfully tried on multiple lines of medical suppression therapy prior to the diagnosis of resistant ECS, with no other medical options available. Furthermore, multiple line medical cortisol suppression therapy may be considered to optimize patients prior to BA, even if their ECS is not fully controlled. More than a third of our cohort received BA for ECS cure with no recurrent ECS and 3.3% operative mortality by end of study (one patient died after surgery associated to aspiration pneumonia); BA was effective in all NET types in our cohort and most of our BA patients had metastatic NET (87%) as well NET diagnosis prior to ECS (70%). This reflects the efficacy of a well performed BA but also careful patient selection as some patients were not offered BA due to poor clinical status (N=6). Unfortunately, some patients were not referred for BA despite resistant ECS; reasons for lack of referral were not well elucidated from the electronic records but this may be due to performance status or prognosis determined by the managing medical oncology or endocrinology team. BA is a timely intervention with a small window of opportunity between ECS diagnosis and optimization of signs and symptoms;⁹ patients may also worsen clinically following ECS despite medical suppression therapy. Untreated or uncontrolled hypercortisolism has been shown to worsen survival^{6, 11} and may remove the option of BA in ECS management. Following BA, patients were able to restart NET-targeted systemic therapy at a median of 2 months postoperatively. This supports that multidisciplinary management is essential and BA referral should be early in the decision algorithm following ECS diagnosis

Based on present results and following current practice guidelines²⁴, our multidisciplinary group proposed an algorithm of treatment for patients diagnosed with ECS (Figure 4). It is important to mention that currently, some selected patients with hepatic metastasis could benefit from resection of the primary tumor and liver-directed therapy or even liver transplant; in these cases, multidisciplinary evaluation is reocmmended.²⁵

Figure 4.

Therapeutic approach for neuroendocrine tumor associated ectopic Cushing syndrome.

Abbreviations: NET, neuroendocrine tumor; ECS, ectopic Cushing syndrome

^aconsider management for tumor burden reduction if indicated: cytoreduction, liver-directed therapy, and/or systemic therapy

Survival

Survival analyses in our NET-ECS cohort showed that patients who developed ECS following known NET diagnosis had worse DSS; this is likely due to the additional burden of ECS in a patient already burdened by metastatic NET.¹¹ Worse survival in NET-ECS was associated with de novo hyperglycemia at ECS diagnosis, poor NET differentiation, and the presence of metastases. Poorly differentiated and metastatic NETs are well established risk factors for worse NET survival²⁶. These risk factors for worse prognosis likely represent an aggressive subgroup of NET (poorly differentiated and/or metastatic) or a severe form of ECS (with de novo hyperglycemia) that when paired makes a lethal combination. Prolonged DSS following ECS diagnosis was observed in patients with medically resistant ECS who were subsequently controlled with BA than patients who did not undergo BA. The importance of early ECS suppression or cure becomes substantial with longer predicted NET survival. This information is useful in guidance of prognosis and expected outcomes in NET-ECS.

Limitations

Although we have a large single center cohort of NET-ECS patients, the retrospective nature of our study resulted in unavoidable limitations. Unfortunately, due to the rarity of this disease and nature of our study spanning multiple decades, some patients were ultimately lost to follow-up, which may have affected outcomes. Retrospective chart reviews are subject to recall bias and incomplete data as not all postoperative visits were equal in laboratory and imaging workups and frequency of visits. Cushing disease and NETs are rare and the coexistence of both has a low probability, although absolute certainty of ECS is difficult. As a quaternary referral center, our cohort may have represented more severely affected and metastatic cases. Despite these limitations, we documented significant findings of prognostic risk factors and management strategies that affected outcome as well as proposed a management algorithm for this rare and complex disease. Our results will affect the way we evaluate, manage, and follow patients not only with confirmed NET-ECS, but hopefully all patients with MTC and foregut NETs, especially in the metastatic setting.

CONCLUSION

Patients with NET prior to ECS development were more likely metastatic at ECS diagnosis and had worse outcomes. De novo hyperglycemia and poor NET differentiation were also predictive of worse prognosis. Complete medical suppression of hypercortisolism is difficult to achieve in most patients with NET-ECS, especially in the metastatic setting. Well-selected patients may benefit from BA early in the decision algorithm, and multidisciplinary management is essential in this complex disease.