Original Article
Divya Yogi-Morren, MD1; Mouhammed Amir Habra, MD2; Charles Faiman, MD1; James Bena, MS1; Betul Hatipoglu, MD1; Laurence Kennedy, MD1; Robert J. Weil MD3; Amir H. Hamrahian, MD1 ABSTRACT Objective: Expert opinion and a consensus statement on Cushing syndrome (CS) indicate that in a patient with a clinical presentation and biochemical studies consistent with a pituitary etiology, the presence of a pituitary tumor ≥6 mm is highly suggestive of Cushing disease (CD). The purpose of the present study was to determine the optimal pituitary tumor size that can differentiate between patients with CD and ectopic adrenocorticotrophic hormone (ACTH) secretion (EAS) and obviate the need for inferior petrosal sinus sampling (IPSS). Methods: We performed a retrospective study of 130 patients seen between 2000 and 2012 including 104 patients with CD and 26 patients with EAS. Results: A pituitary lesion was reported in 6/26 (23%) patients with EAS and 71/104 (68.3%) patients with CD, with median (range) sizes of 5 mm (3-14) and 8 mm (2-31), respectively. All tumors in the EAS group measured ≤6 mm except for 1 that measured 14 mm. The presence of a pituitary tumor >6 mm in size had 40% sensitivity and 96% specificity for the diagnosis of CD. ACTH levels >209 pg/mL and serum potassium 3.4 times the upper limit of normal (×ULN) Conclusion: Pituitary incidentalomas as large as 14 mm in size can be seen in patients with EAS. However, the 6-mm tumor size cut-off value provided 96% specificity and may be a reasonable threshold to proceed with surgery without the need for IPSS when the biochemical data support a pituitary etiology. (Endocr Pract. 2015;21:1098-1103) Abbreviations: ACTH = adrenocorticotropic hormone; CD = Cushing disease; CRH = corticotropin-releasing hormone; CS = Cushing syndrome; EAS = ectopic ACTH secretion; IPSS = inferior petrosal sinus sampling; HDDST = high-dose dexamethasone suppression test; LDDST = low-dose dexamethasone suppression test; MRI = magnetic resonance imaging; UFC = urine free cortisol; ULN = upper limit of normal. INTRODUCTION Cushing syndrome (CS) refers to a constellation of signs and symptoms associated with excessive glucocorticoid exposure. Most cases of endogenous CS are related to excessive adrenocorticotropic hormone (ACTH) secretion from a pituitary adenoma (Cushing disease [CD]) and represents about 60 to 70 % of all cases of CS, while ectopic ACTH secretion (EAS) from a variety of tumors is less common and constitutes about 10% of cases (1-5). Establishing the etiology of ACTH-dependent CS may be challenging because of the overlap in the clinical, biochemical, and radiologic features between CD and EAS. This challenge is further complicated by the fact that 5 to 10% of the population harbor incidental pituitary lesions (6,7). Additional evaluation is often needed to distinguish between CD and EAS, including a high-dose dexamethasone test (HDDST), corticotropin-releasing hormone
Does a Large Tumor Exclude Ectopic ACTH Syndrome?, Endocr Pract. 2015;21(No. 10) 1099
(CRH) stimulation test, and inferior petrosal sinus sampling (IPSS) (8,9). One of the main questions that arises during the evaluation of patients with ACTH-dependent CS is whether or not IPSS is needed. IPSS is costly, technically difficult, and not without risk, especially in institutions that do not perform these procedures regularly. Therefore, avoiding IPSS when it is not necessary is desirable. Expert opinion as well as a consensus statement on CS recommends that in a patient with a classic clinical presentation and dynamic biochemical studies compatible with a pituitary etiology, a pituitary tumor size of 6 mm is highly suggestive of CD and obviates the need for more invasive testing (10). However, this cut-off value was derived from magnetic resonance imaging (MRI) studies in healthy controls showing that in such a population, the majority of pituitary incidentalomas are ≤6 mm in size (7). In a study by Hall et al of 100 healthy controls, 10% of the subjects had focal areas of decreased signal intensity in the pituitary gland ranging from 3 to 6 mm at the greatest diameter (6). In another study by Chong et al, 20/52 of normal subjects had a mean pituitary lesion of 3.9 ± 0.9 mm (range 2-5) on 1.5-Tesla MRI (7). As these data have not been validated in patients with ACTH-dependent CS, we report biochemical and radiologic features in a large cohort of patients with ACTH-dependent CS to assist in distinguishing CD from EAS. We focused on pituitary MRI findings in the 2 groups to determine if the presence of a specific pituitary tumor size on pituitary MRI would predict the diagnosis of CD and hence obviate the need for IPSS in the diagnostic work-up of a patient with ACTH-dependent CS. METHODS All consecutive patients with CD treated between the years 2000 and 2012 who fulfilled the diagnostic criteria were identified from the pituitary database at Cleveland Clinic. All the patients with EAS and pituitary MRI treated at MD Anderson Cancer Center between the years 2003 and 2012 were identified through chart review. We ultimately identified 104 patients with CD and 26 patients with EAS. We retrieved all available clinical and biochemical data including demographics, preoperative serum cortisol and plasma ACTH levels, low-dose dexamethasone (LDDST) and HDDST results, CRH test findings, 24-hour urine free cortisol (UFC) and late-night salivary cortisol levels, pituitary tumor dimensions on MRI, IPSS results, early postoperative cortisol and ACTH levels, and pathology reports. The choice and number of tests were at the discretion of the evaluating physician. A number of the patients did not undergo an HDDST or CRH test, which is likely related to reliance of their physicians on IPSS as the gold standard test to distinguish between CD and EAS or if there was an obvious source for EAS. In addition, some
of these studies were done by referring endocrinologists, but we were unable to retrieve the data from their medical records. In addition to review of the neuroradiologist reports, MR images were independently confirmed by the on-site endocrinologist and neurosurgeon when available. The maximum tumor dimension was used for data analysis. A tumor size of 0 was assigned to patients without a visible tumor for calculating diagnostic sensitivity and specificity. The diagnosis of ACTH-dependent CS was based on the clinical presentation, plasma ACTH level >20 pg/ mL, and 1 or more abnormal biochemical tests including elevated 24-hour UFC and/or a late night salivary cortisol, a lack of cortisol suppression after an LDDST and/or a dexamethasone suppression/CRH stimulation test. The diagnosis of CD was confirmed by the presence of positive ACTH staining on pituitary pathology specimens or development of hypocortisolemia (serum cortisol 2 before or >3 after CRH administration. The results of IPSS in most of these patients have been reported in 2 previous publications by our group (11,12). In most patients, serum cortisol levels were measured using a solid-phase competitive chemiluminescent enzyme immunoassay. Prior to 2006, 24-hour UFC in most patients was measured by immunoassay; liquid chromatography mass spectrometry was employed after 2006. The 24-hour UFC values were calculated as the fold level above the upper limit of normal (ULN). ACTH levels were measured using a solid-phase, 2-site sequential chemiluminescent assay.
1100 Does a Large Tumor Exclude Ectopic ACTH Syndrome?, Endocr Pract. 2015;21(No. 10)
Statistical Analysis Continuous measures and differences in biochemical tests between the 2 groups were compared using Wilcoxon rank sum tests, given the normality of several variables. Categorical factors were described using frequencies and percentages and were compared across the groups using Pearson c2 tests or Fisher exact tests when cell frequencies were small. Data are shown as the median (range). Receiver operating characteristic (ROC) curves were used to determine the sensitivity and specificity of different tumor size cut-off values on MRI to predict a diagnosis of CD. Analyses were performed using SAS software (version 9.2; SAS Institute, Cary, NC), and plots were created using R software (version 2.15; Vienna, Austria).
except for 1 which measured 14 mm (Fig. 1). The presence of a pituitary tumor >6 mm in size had 40% sensitivity and 96% specificity for the diagnosis of CD.
Patient Population Our study population included 130 patients with ACTH-dependent CS (104 CD and 26 EAS). The age and sex (% females) were similar between the 2 groups: 46 years (16-85) and 76% for the CD group, and 42 (11-80) years and 65.4% for the EAS group, respectively (Table 1). Information on symptom duration and the presence of cushingoid features in some of the patients were not available; therefore, a comparison between the groups could not be performed.
Biochemical Evaluation Serum cortisol levels after the LDDST and HDDST, as well as the percent suppressions in cortisol levels are shown in Table 1. Considering the relatively limited number of patients who underwent HDDST, diagnostic sensitivity or specificity was not calculated. The plasma median ACTH level in the EAS group was 161.5 pg/mL (58-671), which was higher than in the CD group value of 71 pg/mL (16-209), P≤.001 (Table 1). None of the patients in the CD group had an ACTH level >209 pg/mL, whereas 10/26 (38.5%) in the EAS group had an ACTH level above this value. The median serum potassium value of 3.0 mmol/L (1.5-4.5), P≤.001 in the EAS group was lower than in the CD group (3.8 mmol/L, 2.74.8) (Table 1). None of the patients in the CD group had a serum potassium 3.4-fold the ULN, whereas in the CD group only 30/81 (37.0%) had a 24-hour UFC above this value.
Pituitary MRI Findings A pituitary lesion was identified in 73/105 (68.6%) patients in the CD group and 6/26 (23.1%) patients in the EAS group; they measured 8 mm (2-31) and 5 mm (3-14), respectively. All tumors in the EAS group measured ≤6 mm
Biochemical Findings in the 6 Patients with EAS with Identifiable Pituitary Tumors The 24-hour UFC and plasma ACTH levels among the 6 patients in the EAS group who had an identifiable lesion on MRI were 493.5 mg/day (153-699) and 95 pg/mL
RESULTS
Table 1 Demographic and Biochemical Data of Patients with EAS and CD EAS group (n = 26)
CD group (n = 104)
P valuea
65.4
76
.32
42 (11.0-80.0)
46 (16.0-85.0)
.51
161.5 (58.0-671.0)
71.0 (3.0-209.0)
Divya Yogi-Morren, MD1; Mouhammed Amir Habra, MD2; Charles Faiman, MD1; James Bena, MS1; Betul Hatipoglu, MD1; Laurence Kennedy, MD1; Robert J. Weil MD3; Amir H. Hamrahian, MD1 ABSTRACT Objective: Expert opinion and a consensus statement on Cushing syndrome (CS) indicate that in a patient with a clinical presentation and biochemical studies consistent with a pituitary etiology, the presence of a pituitary tumor ≥6 mm is highly suggestive of Cushing disease (CD). The purpose of the present study was to determine the optimal pituitary tumor size that can differentiate between patients with CD and ectopic adrenocorticotrophic hormone (ACTH) secretion (EAS) and obviate the need for inferior petrosal sinus sampling (IPSS). Methods: We performed a retrospective study of 130 patients seen between 2000 and 2012 including 104 patients with CD and 26 patients with EAS. Results: A pituitary lesion was reported in 6/26 (23%) patients with EAS and 71/104 (68.3%) patients with CD, with median (range) sizes of 5 mm (3-14) and 8 mm (2-31), respectively. All tumors in the EAS group measured ≤6 mm except for 1 that measured 14 mm. The presence of a pituitary tumor >6 mm in size had 40% sensitivity and 96% specificity for the diagnosis of CD. ACTH levels >209 pg/mL and serum potassium 3.4 times the upper limit of normal (×ULN) Conclusion: Pituitary incidentalomas as large as 14 mm in size can be seen in patients with EAS. However, the 6-mm tumor size cut-off value provided 96% specificity and may be a reasonable threshold to proceed with surgery without the need for IPSS when the biochemical data support a pituitary etiology. (Endocr Pract. 2015;21:1098-1103) Abbreviations: ACTH = adrenocorticotropic hormone; CD = Cushing disease; CRH = corticotropin-releasing hormone; CS = Cushing syndrome; EAS = ectopic ACTH secretion; IPSS = inferior petrosal sinus sampling; HDDST = high-dose dexamethasone suppression test; LDDST = low-dose dexamethasone suppression test; MRI = magnetic resonance imaging; UFC = urine free cortisol; ULN = upper limit of normal. INTRODUCTION Cushing syndrome (CS) refers to a constellation of signs and symptoms associated with excessive glucocorticoid exposure. Most cases of endogenous CS are related to excessive adrenocorticotropic hormone (ACTH) secretion from a pituitary adenoma (Cushing disease [CD]) and represents about 60 to 70 % of all cases of CS, while ectopic ACTH secretion (EAS) from a variety of tumors is less common and constitutes about 10% of cases (1-5). Establishing the etiology of ACTH-dependent CS may be challenging because of the overlap in the clinical, biochemical, and radiologic features between CD and EAS. This challenge is further complicated by the fact that 5 to 10% of the population harbor incidental pituitary lesions (6,7). Additional evaluation is often needed to distinguish between CD and EAS, including a high-dose dexamethasone test (HDDST), corticotropin-releasing hormone
Does a Large Tumor Exclude Ectopic ACTH Syndrome?, Endocr Pract. 2015;21(No. 10) 1099
(CRH) stimulation test, and inferior petrosal sinus sampling (IPSS) (8,9). One of the main questions that arises during the evaluation of patients with ACTH-dependent CS is whether or not IPSS is needed. IPSS is costly, technically difficult, and not without risk, especially in institutions that do not perform these procedures regularly. Therefore, avoiding IPSS when it is not necessary is desirable. Expert opinion as well as a consensus statement on CS recommends that in a patient with a classic clinical presentation and dynamic biochemical studies compatible with a pituitary etiology, a pituitary tumor size of 6 mm is highly suggestive of CD and obviates the need for more invasive testing (10). However, this cut-off value was derived from magnetic resonance imaging (MRI) studies in healthy controls showing that in such a population, the majority of pituitary incidentalomas are ≤6 mm in size (7). In a study by Hall et al of 100 healthy controls, 10% of the subjects had focal areas of decreased signal intensity in the pituitary gland ranging from 3 to 6 mm at the greatest diameter (6). In another study by Chong et al, 20/52 of normal subjects had a mean pituitary lesion of 3.9 ± 0.9 mm (range 2-5) on 1.5-Tesla MRI (7). As these data have not been validated in patients with ACTH-dependent CS, we report biochemical and radiologic features in a large cohort of patients with ACTH-dependent CS to assist in distinguishing CD from EAS. We focused on pituitary MRI findings in the 2 groups to determine if the presence of a specific pituitary tumor size on pituitary MRI would predict the diagnosis of CD and hence obviate the need for IPSS in the diagnostic work-up of a patient with ACTH-dependent CS. METHODS All consecutive patients with CD treated between the years 2000 and 2012 who fulfilled the diagnostic criteria were identified from the pituitary database at Cleveland Clinic. All the patients with EAS and pituitary MRI treated at MD Anderson Cancer Center between the years 2003 and 2012 were identified through chart review. We ultimately identified 104 patients with CD and 26 patients with EAS. We retrieved all available clinical and biochemical data including demographics, preoperative serum cortisol and plasma ACTH levels, low-dose dexamethasone (LDDST) and HDDST results, CRH test findings, 24-hour urine free cortisol (UFC) and late-night salivary cortisol levels, pituitary tumor dimensions on MRI, IPSS results, early postoperative cortisol and ACTH levels, and pathology reports. The choice and number of tests were at the discretion of the evaluating physician. A number of the patients did not undergo an HDDST or CRH test, which is likely related to reliance of their physicians on IPSS as the gold standard test to distinguish between CD and EAS or if there was an obvious source for EAS. In addition, some
of these studies were done by referring endocrinologists, but we were unable to retrieve the data from their medical records. In addition to review of the neuroradiologist reports, MR images were independently confirmed by the on-site endocrinologist and neurosurgeon when available. The maximum tumor dimension was used for data analysis. A tumor size of 0 was assigned to patients without a visible tumor for calculating diagnostic sensitivity and specificity. The diagnosis of ACTH-dependent CS was based on the clinical presentation, plasma ACTH level >20 pg/ mL, and 1 or more abnormal biochemical tests including elevated 24-hour UFC and/or a late night salivary cortisol, a lack of cortisol suppression after an LDDST and/or a dexamethasone suppression/CRH stimulation test. The diagnosis of CD was confirmed by the presence of positive ACTH staining on pituitary pathology specimens or development of hypocortisolemia (serum cortisol 2 before or >3 after CRH administration. The results of IPSS in most of these patients have been reported in 2 previous publications by our group (11,12). In most patients, serum cortisol levels were measured using a solid-phase competitive chemiluminescent enzyme immunoassay. Prior to 2006, 24-hour UFC in most patients was measured by immunoassay; liquid chromatography mass spectrometry was employed after 2006. The 24-hour UFC values were calculated as the fold level above the upper limit of normal (ULN). ACTH levels were measured using a solid-phase, 2-site sequential chemiluminescent assay.
1100 Does a Large Tumor Exclude Ectopic ACTH Syndrome?, Endocr Pract. 2015;21(No. 10)
Statistical Analysis Continuous measures and differences in biochemical tests between the 2 groups were compared using Wilcoxon rank sum tests, given the normality of several variables. Categorical factors were described using frequencies and percentages and were compared across the groups using Pearson c2 tests or Fisher exact tests when cell frequencies were small. Data are shown as the median (range). Receiver operating characteristic (ROC) curves were used to determine the sensitivity and specificity of different tumor size cut-off values on MRI to predict a diagnosis of CD. Analyses were performed using SAS software (version 9.2; SAS Institute, Cary, NC), and plots were created using R software (version 2.15; Vienna, Austria).
except for 1 which measured 14 mm (Fig. 1). The presence of a pituitary tumor >6 mm in size had 40% sensitivity and 96% specificity for the diagnosis of CD.
Patient Population Our study population included 130 patients with ACTH-dependent CS (104 CD and 26 EAS). The age and sex (% females) were similar between the 2 groups: 46 years (16-85) and 76% for the CD group, and 42 (11-80) years and 65.4% for the EAS group, respectively (Table 1). Information on symptom duration and the presence of cushingoid features in some of the patients were not available; therefore, a comparison between the groups could not be performed.
Biochemical Evaluation Serum cortisol levels after the LDDST and HDDST, as well as the percent suppressions in cortisol levels are shown in Table 1. Considering the relatively limited number of patients who underwent HDDST, diagnostic sensitivity or specificity was not calculated. The plasma median ACTH level in the EAS group was 161.5 pg/mL (58-671), which was higher than in the CD group value of 71 pg/mL (16-209), P≤.001 (Table 1). None of the patients in the CD group had an ACTH level >209 pg/mL, whereas 10/26 (38.5%) in the EAS group had an ACTH level above this value. The median serum potassium value of 3.0 mmol/L (1.5-4.5), P≤.001 in the EAS group was lower than in the CD group (3.8 mmol/L, 2.74.8) (Table 1). None of the patients in the CD group had a serum potassium 3.4-fold the ULN, whereas in the CD group only 30/81 (37.0%) had a 24-hour UFC above this value.
Pituitary MRI Findings A pituitary lesion was identified in 73/105 (68.6%) patients in the CD group and 6/26 (23.1%) patients in the EAS group; they measured 8 mm (2-31) and 5 mm (3-14), respectively. All tumors in the EAS group measured ≤6 mm
Biochemical Findings in the 6 Patients with EAS with Identifiable Pituitary Tumors The 24-hour UFC and plasma ACTH levels among the 6 patients in the EAS group who had an identifiable lesion on MRI were 493.5 mg/day (153-699) and 95 pg/mL
RESULTS
Table 1 Demographic and Biochemical Data of Patients with EAS and CD EAS group (n = 26)
CD group (n = 104)
P valuea
65.4
76
.32
42 (11.0-80.0)
46 (16.0-85.0)
.51
161.5 (58.0-671.0)
71.0 (3.0-209.0)
