The CuPeR Model: A Dynamic Online Tool for Predicting Cushing’s Disease Persistence and Recurrence After Pituitary Surgery

Posted on October 30, 2025 by MaryO

Abstract

Objective

Predicting postoperative persistence and recurrence of Cushing’s disease (CD) remains a clinical challenge, with no universally reliable models available. This study introduces the CuPeR model, an online dynamic nomogram developed to address these gaps by predicting postoperative outcomes in patients with CD undergoing pituitary surgery.

Methods

A retrospective cohort of 211 patients treated for CD between 2010 and 2024 was analyzed. Key patient and tumor characteristics, imaging findings, and treatment details were evaluated. Multivariate logistic regression identified independent predictors of postoperative persistence or recurrence of CD (PoRP-CD), which were then incorporated into the CuPeR model using stepwise selection based on Akaike Information Criterion. Internal validation was performed using a testing dataset, and a user-friendly online nomogram was developed to facilitate immediate, patient-specific risk estimation in clinical practice.

Results

The final predictive model identified four key factors: symptom duration, MRI Hardy’s grade, tumor site, and prior pituitary surgery. Longer symptom duration and a history of prior surgery significantly increased the risk of recurrence, while bilateral tumor location reduced this risk. The model demonstrated an area under the receiver operating characteristic curve (AUC-ROC) of 0.70, with 83% accuracy, specificity of 96%, and sensitivity of 33%.

Conclusions

The CuPeR model may offer a practical tool for predicting PoRP-CD, enhancing preoperative decision-making by providing personalized risk assessments.

Keywords

Cushing’s disease
Transsphenoidal surgery
Nomogram
Recurrence
Disease Persistence

Abbreviations

ACTH

Adrenocorticotropic Hormone

AIC

Akaike Information Criterion

AUC

Area Under the Curve

BMI

Body Mass Index

CD

Cushing’s Disease

CI

Confidence Interval

CRH

Corticotropin-Releasing Hormone

DFS

Disease-Free Survival

DL

Deep Learning

eTSS

Endoscopic Transsphenoidal Surgery

HR

Hazard Ratio

IPSS

Inferior Petrosal Sinus Sampling

ML

Machine Learning

MRI

Magnetic Resonance Imaging

OS

Overall Survival

PoRP-CD

Persistent or Recurrent Cushing’s Disease

SIADH

Syndrome of Inappropriate Antidiuretic Hormone Secretion

TSS

Transsphenoidal Surgery

UFC

Urinary Free Cortisol

Introduction

Cushing’s disease (CD) is a rare endocrine disorder, with an annual incidence rate of approximately 0.24 cases per 100,000 individuals [1]. Transsphenoidal surgery (TSS), performed using either endoscopic or microscopic approaches, remains the cornerstone of treatment for CD. Notably, meta-analytical studies have reported that TSS achieves remission and provides long-term disease control in 71–80 % of patients [[2], [3], [4]]. The remaining cases may experience persistent disease despite surgery, while others may face disease recurrence despite initial remission. In such cases, additional treatment options include second pituitary surgery, pituitary irradiation, targeted medical therapies, and bilateral adrenalectomy, each with varying success rates ranging from 25 % for medical therapy to 100 % for bilateral adrenalectomy [5].
To date, no single predictive factor has proven effective in reliably forecasting treatment outcomes in patients with CD [6]. This underscores the critical need for developing predictive models to assess the likelihood of postoperative recurrence or persistence of Cushing’s disease (PoRP-CD). However, only a limited number of studies have addressed this gap. Notably, two studies from Peking Union Medical College Hospital attempted to tackle this issue using machine learning (ML) and deep learning (DL) approaches [6,7]. These studies utilized demographic, clinical, and paraclinical variables to construct predictive models, with DL approaches showing potential to enhance predictive accuracy [7]. While the results of these models were promising, their applicability in routine clinical practice remains limited. Both studies focused exclusively on patients undergoing their initial transsphenoidal surgery, making them less applicable for cases involving patients with a prior history of pituitary surgery or radiotherapy. Furthermore, these models incorporated both preoperative and postoperative parameters, such as changes in cortisol and adrenocorticotropic hormone (ACTH) levels. However, serum cortisol, ACTH, and comprehensive endocrine testing should be available before any treatment decisions are made, and each patient should ideally be reviewed by a multidisciplinary tumor board, including neurosurgery, radiology, endocrinology, and oncology, prior to pituitary surgery. As such, more comprehensive and practical predictive tool that can support timely clinical decision-making and accommodate a broader range of patient scenarios in the management of CD.
The current study was designed to address these critical limitations and provide a more practical solution for predicting postoperative outcomes in CD. Applying one of the largest available CD cohorts, we incorporated a wide array of patient and tumor characteristics, imaging findings, and treatment details to develop a robust and comprehensive predictive model. This model offers treating surgeons reliable insights into the likelihood of tumor recurrence or persistence. By providing individualized risk predictions, the model is intended to assist clinicians in considering different therapeutic options before pituitary surgery, complementing—but not replacing—standard multidisciplinary decision-making. To further enhance its utility in clinical practice, we also developed an interactive online dynamic nomogram, allowing individualized predictions of postoperative persistence or recurrence.

Methods

Study design, patients, and endpoints

The experimental protocol was approved by the Institutional Review Board of Shahid Beheshti University of Medical Sciences (Tehran, Iran). This retrospective study investigates the clinical outcomes of pituitary surgery in patients with CD underwent pituitary surgery between 2010 and 2024 in the neurosurgery department at Loghman Hakim Hospital. Surgeries were conducted by a group of experienced neurosurgeons under the supervision of the first author (G.S). The primary objective of this study was to develop and validate a predictive model for assessing the risk of PoRP-CD. The secondary objectives were (a) to summarize patient and tumor characteristics; (b) to report surgical outcomes and remission rates following surgery; and (c) to analyze patient survival. This study was performed in accordance with the Declaration of Helsinki, and adheres to the reporting guidelines outlined in the STROBE Statement. Due to retrospective nature of the study informed consent was waived by Shahid Beheshti University of Medical Sciences Ethics Committee. All methods were performed in accordance with the relevant guidelines and regulations.

Preoperative assessments

The “index surgery” was set to the most recent pituitary surgery. Before the index surgery, patients underwent comprehensive clinical evaluations, including biochemical and neurological assessments as well as visual field examinations. This research utilized the Endocrine Society Clinical Practice Guideline to establish the diagnosis of CD [8]. Three main steps were involved in the diagnostic process: in the first step, the focus was on detecting hypercortisolemia, which was determined by examining 24-hour urinary free cortisol levels (normal: <60 mcg/24 h), as well as plasma and salivary cortisol profiles. Low-dose dexamethasone suppression testing was performed using the 2 mg/48 h protocol, which was the standard practice in our institution during the study period (2010 onward) [8]. The second step aimed to confirm ACTH-dependent cause of hypercortisolemia, through measuring plasma ACTH levels. The final step aimed to distinguish Cushing’s disease from ectopic sources of ACTH. This was performed using a high-dose dexamethasone suppression test (8  mg overnight), with a plasma cortisol suppression exceeding 50 % typically considered indicative of a pituitary origin [9].
Next, the patients were subjected to thin-slice (3 mm) 1.5-tesla dynamic pituitary gland magnetic resonance imaging (MRI) with gadolinium contrast. The MR evaluation adhered to a strict protocol, requiring an independent agreement of treating neurosurgeon and radiologist to confirm the diagnosis. MR scans were categorized according to the Hardy and Knosp classifications [10]. Normal scans required to demonstrate the absence of direct signs, including inhomogeneity in the pituitary, as well as indirect signs such as a deviation of the pituitary stalk, bulging or erosion of the Sella contour. In cases where the CD was confirmed but pituitary MRI was inconclusive, bilateral inferior petrosal sinus sampling (IPSS) was performed per standard protocol under corticotropin-releasing hormone (CRH) stimulation [11]. Patients with macroadenoma or signs of elevating the optic chiasm were candidates for Humphrey visual field examination.

Surgical approach

Patients underwent endoscopic transsphenoidal approach using conventional “Two Nostrils–Four Hands” technique [12]. Given the diminutive size and deep-seated location of most adenomas, locating the adenoma emerged as a formidable challenge, particularly when the tumor remained not visualized in pre-operative imaging studies. The surgical procedure entailed extensive drilling of the Sellar floor laterally up to the carotid artery on both sides, providing a comprehensive view of the medial wall of the cavernous sinus and exposure of the anterior and posterior intercavernous sinuses. The exploration of the entire Sella commenced in the region where the original tumor had been localized. Upon identification of a tumor, a selective adenomectomy was performed, accompanied by a thorough inspection of the pituitary gland to detect and eliminate any potential tumor remnants. The removal of any pseudo capsule was executed meticulously.
The primary surgical objective was selective adenomectomy, with further exploration guided by the side recommended by IPSS in cases where no adenoma was initially observed. The exploration involved making a plus-like incision on the corresponding half of the gland, enabling deep exploration to leave no part unexplored. In instances where creamy material suggestive of a tumor was drained after a pituitary incision, a tissue biopsy was obtained, although it was not conclusively considered a tumor. Exploration continued on the opposite side in such cases.
When no distinct adenoma was found, a peri-glandular inspection was conducted to visualize the medial wall of the cavernous sinus, diaphragm, and Sellar floor, aiming to detect an ectopic microadenoma. If an apparent tumor remained undetected, the procedure was repeated on the contralateral side, and a vertical medial incision on the pituitary gland adjacent to the pituitary stalk and neurohypophysis was made as a final effort for tumor detection. In the absence of identified pathology during the surgical procedure, hemi-hypophysectomy was considered on the side where IPSS had detected the gradient or on the side with an apparent or suspicious MRI finding. Considering the typical central location of corticotroph cells in the pituitary gland, microadenoma exploration extended posteriorly and medially to confirm extirpation.

Postoperative assessments

In this study, the patients were closely monitored for signs of diabetes insipidus and syndrome of inappropriate antidiuretic hormone secretion (SIADH). Serum sodium levels, urine-specific gravity, and volume were checked regularly. Following surgery, morning cortisol levels were measured on the first day, and other anterior pituitary hormones were evaluated on day 3. Hydrocortisone therapy was initiated based on the patient’s symptoms, signs of adrenal insufficiency, and low cortisol levels. The first postoperative check-up occurred two weeks after surgery, followed by another at three months, which included a comprehensive assessment of pituitary hormones. This evaluation was repeated every three months for two years and then annually. Additionally, patients underwent a dynamic 1.5-Tesla pituitary MRI at six months post-surgery and annually thereafter, with a minimum follow-up period of 12 months.
Remission was defined as having low cortisol levels, indicated by early morning serum cortisol level ≤ 5 μg/dL within two days post-surgery [13]. Persistent CD was characterized by ongoing hypercortisolism, and postoperative recurrence refers to the reappearance of CD symptoms despite initial remission marked by hypercortisolemia. In case of persistence or recurrence, patients were candidates for second-line treatment options selected by their physicians, including revision surgery, targeted medical therapy, pituitary radiotherapy, or bilateral adrenalectomy. Disease-free survival (DFS) was defined as the time from the index surgery to the first occurrence of disease recurrence or death from any cause, while overall survival (OS) was defined as the time from the index surgery to death from any cause.

Statistical analysis

Categorical variables were expressed as numbers and percentages, and continuous variables as mean, range, and standard deviation. The distribution of variables was checked using the Shapiro-Wilk test, which showed a deviation from normal distribution. Contingency tables were used for categorical variables with Pearson’s Chi-squared or Fisher’s Exact test used to examine their association with outcomes for univariate analyses. For continuous variables, the unpaired t-test was applied to compare means between two independent groups when the data met the assumption of normality. Analyses were conducted with R Statistical Software v4.4.0 (“Puppy Cup”). All statistical inferences were two-sided, and P < 0.05 were considered statistically significant.

Model development and internal validation

The dataset was split by “caret package” into a training set (70 %) and a testing set (30 %) using stratified sampling to ensure representative proportions of outcomes. Binary logistic regression was used to identify predictors of PoRP-CD. Patients with adequate follow-up data were included in the analysis. The variables with a marginal level of association (P < 0.15) in the univariate analysis were further included in the multivariate logistic regression analysis to identify the independent predictors of PoRP-CD. Imported factors included demographic, medical history, imaging and pathology results, and treatment details. To identify predictors of PoRP-CD, a multivariable logistic regression model was developed using stepwise selection based on Akaike Information Criterion (AIC). Model performance, including sensitivity, specificity, and area under the receiver operating characteristic curve (AUC-ROC), was evaluated using internal validation on the test dataset.

Nomogram creation and deployment

A nomogram was constructed using the validated logistic regression model. The nomogram was then integrated into a web-based application using the “Shiny package” in R program. The dynamic nomogram allows clinicians to input patient data and obtain individualized risk predictions for PoRP-CD.

Survival analysis

Survival analysis was conducted to evaluate DFS across various patient subgroups. The log-rank test was applied to assess statistical differences in survival distributions between subgroups. Cox proportional hazard regression was used to estimate hazard ratios (HR) and 95 % confidence intervals (CI). The “survival” and “survminer” R packages were applied in this section.

Results

Patients and tumors characteristics

A total of 211 patients with CD had been treated by a group of experienced neurosurgeons under the supervision of the first author (G.S) between March 2010 and January 2024 in the neurosurgery department at Loghman Hakim Hospital. Table 1 summarizes the baseline characteristics of patients at the timepoint of index surgery. The patients had a mean age of 35.9 ± 12.1 years (range: 11–67), among which 21 patients (9.9 %) were in the pediatric age range, and 165 (78.1 %) were female. Obesity was the most common patients’ symptoms (45.9 %), and physical examination reported centripetal obesity (84.3 %), moon face (75.8 %), and striae (64.4 %) as the most common clinical manifestations. Compared to the adult patients, pediatrics had less common hypertension on physical examination (35.2 vs. 5.9 %) and medical history of diabetes mellitus (36.8 vs. 4.7 %) (P < 0.05). The majority of patients (63.9 %, 135/211) had not received any prior treatment. Among those who had, surgery alone was the most common approach (n = 57, 27.0 %), performed once in 50 patients (23.6 %), twice in 6 patients (2.8 %), and three times in a single patient.

Table 1. Baseline characteristics of adult and pediatric patients with Cushing’s disease.

Demographics Total
n = 211		 Adults
n = 190		 Pediatrics
n = 21		 P Medical Hx Total
n = 211		 Adults
n = 190		 Ped.
n = 21		 P Drug-Family Hx Total
n = 211		 Adults
n = 190		 Ped.
n = 21		 P
Age; mean-SD (y) 35.9–12.1 38.3–10.2 14.8–1.7 0<.001 Hypertension 97 (45.9) 92 (48.4) 5 (23.8) 0.31 Cabergoline 3 (1.4) 3 (1.5) 0 1.0
Sex; female 165 (78.1) a 149 (78.4) 16 (76.1) 0.78 Diabetes mellitus 71 (33.6) 70 (36.8) 1 (4.7) 0<.001 Ketoconazole 12 (5.6) 12 (6.3) 0 0.61
Marital status; married 105 (70.0) b 103 (76.8) b 2 (12.5) b 0<.001 Dyslipidemia 45 (21.3) 42 (22.1) 3 (14.2) 0.56 Metyrapone 0 − − −
Smoking status; active–passive-non 17 (10)-27(17)-113(72) b 17 (11)–23(15)-101(70) b 0–4(25)-12(75) b 0.70 Prior pituitary surgery 57 (27.0)) 51 (26.8) 6 (28.5) 1.0 Pasireotide 0 − − −
Height; mean-SD (cm) 163.9–8.7 163.8–8.9 165.1–6.6 0.59 Fatty liver 37 (17.5) 32 (16.8) 5 (23.8) 0.36 Somatostatin 0 − − −
Weight; mean-SD (Kg) 74.1–22.5 74.6–22.5 69.3–23.1 0.58 Thromboembolism 6 (2.8) 6 (3.1) 0 1.0
BMI; mean-SD (Kg/m2) 28.8–6.1 29.0–6.2 27.6–5.5 0.72 DVT 3 (1.4) 3 (1.5) 0 1.0 FH of Cushing 5 (2.3) 4 (2.1) 1 (4.7) 0.43
Symptom duration; mean-SD (m) 30.7–41.2 32.0–43.2 20.0–14.2 0.78 MEN 1 (0.4) 1 (0.5) 0 1.0 FH of MEN 1 (0.4) 1 (0.5) 0 1.0
Presenting Symptoms
Obesity 75 (45.9) b 66 (45.2) b 9 (52.9) b 0.61 Striae 10 (6.1) b 8 (5.4) b 2 (11.7) b 0.27 Headache 4 (2.4) b 3 (2.0) b 1 (5.8) b 0.35
Menstrual disorders 16 (9.8) b 13 (8.9) b 3 (17.6) b 0.22 Edema 7 (4.2) b 7 (4.7) b 0 1.0 Diabetes mellitus 3 (1.8) b 3 (2.0) b 0 1.0
Hypertension 12 (7.3) b 12 (8.2) b 0 0.61 Muscular weakness 7 (4.2) b 6 (4.1) b 1 (5.8) b 0.54 Bone fracture 3 (1.8) b 3 (2.0) b 0 1.0
Blurred vision 10 (6.1) b 9 (6.1) b 1 (5.8) b 1.0 Moon face 6 (3.6) b 6 (4.1) b 0 1.0 Other 10 (6.1) b 10 (6.8) b 0 0.60
Clinical Manifestations
Acanthosis nigricans 35 (16.5) 34 (17.8) 1 (4.7) 0.12 Easy bruising 103 (48.8) 92 (48.4) 11 (52.3) 0.91 Male pat. hair loss 111 (52.6) 100 (52.6) 11 (52.3) 1.0
Acne 68 (32.2) 58 (30.5) 10 (47.6) 0.16 Ecchymosis 58 (27.5) 50 (26.3) 8 (38.0) 0.37 dysmenorrhea 96 (45.4) 84 (44.2) 12 (57.1) 0.49
Ankle edema 105 (49.7) 96 (50.5) 9 (42.8) 0.57 Exophthalmia 50 (23.7) 47 (24.7) 3 (14.2) 0.27 Moon face 160 (75.8) 141 (74.2) 19 (90.4) 0.69
Backache 66 (31.2) 60 (31.5) 6 (28.5) 0.88 Facial plethora 97 (45.9) 85 (44.7) 12 (57.1) 0.33 Osteoporosis 25 (11.8) 25 (13.1) 0 0.14
Blurred vision 70 (33.2) 67 (35.2) 3 (14.2) 0.27 Fatigue 146 (69.2) 130 (68) 16 (76.1) 0.76 Prox. myopathy 94 (44.5) 86 (45.2) 8 (38.0) 0.63
Buffalo hump 123 (58.3) 107 (56.3) 16 (76.1) 0.43 Fracture 12 (5.6) 12 (6.3) 0 0.61 Skin atrophy 81 (38.4) 73 (38.4) 8 (38.0) 1.0
Centripetal obesity 178 (84.3) 159 (83.6) 19 (90.4) 0.50 Headache 109 (51.6) 97 (51.0) 12 (57.1) 1.0 Striae 136 (64.4) 119 (62.6) 17 (80.9) 0.55
Cerebrospinal fluid leakage 5 (2.3) 4 (2.1) 1 (4.7) 0.41 Hirsutism 104 (49.3) 92 (48.4) 12 (57.1) 0.72 Supraclav. fat pad 38 (18.0) 33 (17.3) 5 (23.8) 0.67
Cranial nerve palsy 3 (1.4) 3 (1.5) 0 1.0 Hyperpigmentation 38 (18.0) 37 (19.4) 1 (4.7) 0.12 Visual field defect 24 (11.3) 22 (11.5) 2 (9.5) 1.0
Diplopia 18 (8.5) 15 (7.8) 3 (14.2) 0.41 Hypertension 69 (32.7) 67 (35.2) 2 (9.5) 0.009 Weight gain 108 (51.1) 95 (50.0) 13 (61.9) 0.39
Prior Treatments
Treatment naïve 135 (63.9) 122 (64.2) 13 (61.9) 1.0 Pituitary surgery alone 39 (18.4) 33 (17.3) 6 (28.5) 0.23 Radiotherapy alone 6 (2.8) 5 (2.6) 1 (4.7) 0.47
Medication alone 5 (2.3) 5 (2.6) 0 1.0 Combination therapy 17 (8.1) 17 (8.9) 0 0.22 Adrenalectomy alone 11 (5.2) 10 (5.2) 1 (4.7) 1.0
Hormonal Assessments
Hypothyroidism 24 (31.1) b 24 (31.1) b 0 0.09 GH deficiency 6 (8.8) b 6 (8.8) b 0 1.0 Hypogonadism 7 (9.8) b 7 (9.8) b 0 1.0
Panhypopituitarism 2 (2.5) b 2 (2.5) b 0 1.0
Imaging Features
Hardy’s grading
(sphenoid bone invasion)
0
1
2
3
4		 37 (21.1) b
102 (58.2) b
27 (15.4) b
4 (2.2) b
5 (2.8) b		 35 (22.7) b
88 (57.1) b
23 (14.9) b
3 (1.9) b
5 (3.2) b		 2 (9.5)
14 (66.7)
4 (19.0)
1 (4.7)
0		 0.45 Hardy’s staging
(suprasellar extension)
A
B
C
D
E		 36 (20.4) b
86 (48.8) b
14 (7.9) b
4 (2.2) b
36 (20.4) b		 34 (21.9) b
73 (47.1) b
14 (9.0) b
2 (1.2) b
32 (20.6) b		 2 (9.5)
13 (62)
0
2 (9.5)
4 (19.0)		 0.07 Knosp grading

0
1
2
3
4

 152 (82.6) b
13 (7.0) b
7 (3.8) b
4 (2.1) b
8 (4.3) b		 135 (82.8) b
10 (6.1) b
6 (3.6) b
4 (2.4) b
8 (4.9) b		 17 (80.9)
3 (14.2)
1 (4.7)
0
0		 0.46
Tumor size
Microadenoma
Macroadenoma
MR-negative		 122 (58.6) b
50 (24.0) b
36 17.3) b		 111 (59.3) b
42 (22.4) b
34 (18.1) b		 11 (52.3)
8 (38.0)
2 (9.5)		 0.28 Sphenoid shape
Sellar
Presellar
Conchal		 205 (97.6) b
3 (1.4) b
2 (0.9) b		 184 (97.3) b
3 (1.5) b
2 (1.0) b		 21 (100)
0
0		 1.0 Multifocality
Unifocal
Multifocal		 113 (80.1)
28 (19.8)		 97 (79.5)
25 (20.4)		 16(84.2)
3 (15.7)		 0.79
Invasion c
No invasion
Cavernous sinus
Carotid
Dura
Clivus		 185 (88.5) b
12 (5.7) b
3 (1.4) b
6 (2.8) b
3 (1.4) b		 165 (87.7) b
11 (5.8) b
3 (1.5) b
6 (3.1) b
3 (1.5) b		 20 (95.2) b
1 (4.8) b
0
0
0		 1.0 Tumor site
Right lobe
Left lobe
Bilateral
Central
Stalk		 22 (15.6)) b
16 (11.3)) b
51 (36.1) b
49 (34.7) b
3 (2.1) b		 20 (16.2) b
13 (10.5) b
43 (34.9) b
45 (36.5) b
2 (1.6) b		 2 (11.1) b
3 (16.6) b
8 (44.4) b
4 (22.2) b
1 (5.5) b		 0.38 Empty sella
No
Yes		 207 (98.1)
4 (1.8)		 187 (98.4)
3 (1.5)		 20(95.2)
1 (4.7)		 0.34
Pituitary apoplexy
No
Yes		 185 (97.3) b
5 (2.6) b		 167 (98.2) b
3 (1.7) b		 18 (90.0) b
2 (10.0) b		 0.08 Kissing carotids
No
Yes		 209 (99.0)
2 (0.9)		 188 (98.9)
2 (1.0)		 21 (100)
0		 1.0
a
the numbers in parentheses represent the percentage for each patient group.
b
percentage after ruling out missing data.
c
one patient had invasion to cavernous sinus and carotid and another one had clivus and dural invasion.
A comprehensive preoperative hormonal assessment was conducted on 77 patients (36.4 %), revealing hormonal dysregulation in 28 patients (36.3 %). Hypothyroidism was the most common abnormality, affecting 35 % of those assessed (24 out of 77). On MRI scans, most tumors were microadenomas (58.6 %), with fewer macroadenomas (24.0 %) and some cases with no detectable tumor (17.3 %). Tumors were commonly localized bilaterally (36.1 %) or centrally (34.7 %), and most were unifocal (80.1 %). Knosp grading indicated no cavernous sinus invasion in the majority (82.6 %), with only 6.4 % showing grades 3–4. According to Hardy’s grading, most patients had mild sphenoid bone invasion, predominantly grade 1 (58.2 %). For Hardy’s staging of suprasellar extension, nearly half were at stage B (48.8 %), with smaller groups in stages A and E (20.4 % each), and fewer in stages C and D. Other MRI findings are summarized in Table 1. There was no significant difference between adult and pediatric patients in terms of hormonal and imaging findings (P > 0.05). Pathology reports were available for 36 patients. The most common finding was sparse cellularity, observed in 11 patients (30.6 %) followed by dense cellularity identified in 9 patients (25 %). Crooke cell changes were the least common, present in 7 patients (19.4 %). Nine specimens (25 %) had no tumor identified in the sample submitted to pathology.

Treatment details and outcomes

A total of 36 patients (17.1 %) underwent preoperative IPSS, among which 13 had right lateralization, 13 left, 4 bilateral, 3 central, 2 central-right, and 1 central-left. Pituitary surgery was predominantly performed using the endoscopic transsphenoidal (eTSS) approach (98.5 %, 208/211), while the transplanum approach was used in 3 patients (1.5 %). Adenomectomy was the most common surgical procedure (n = 187, 88.6 %), followed by total hypophysectomy in 17 patients (8.1 %) and hemi-hypophysectomy in 7 patients (3.3 %). In addition, four patients in the total hypophysectomy group and one patient in the adenomectomy group also underwent hypophyseal stalk resection. Information on disease persistence or recurrence was available for 204 patients. Median follow-up of patients was 58.4 months (range: 4.5–170.4 months) after index surgery. In total, 23 patients (11.2 %) experienced persistent disease following the index surgery, while 10 patients (4.9 %) had disease recurrence, with a median time to recurrence of 7 months (range: 1–78 months). The median recurrence-free interval for the entire cohort was 37 months.
The surgical complication rates were as follows (Fig. 1A): cerebrospinal fluid leaks were observed in 22 patients (10.4 %), followed by cranial nerve injury in 7 patients (3.3 %) and meningitis in 5 patients (2.3 %). Carotid injury and intracerebral bleeding each occurred in 3 patients (1.4 %). Nasal bleeding, the need for a ventriculoperitoneal shunt, and embolic events were each reported in 1 patient (0.4 %). Perioperative mortality was observed in one female patient (0.4 %) due to an iatrogenic carotid injury. This patient had previously undergone three pituitary surgeries and received radiotherapy at the pituitary site. Hormonal dysregulation following surgery included hypothyroidism in 99 patients (46.9 %), diabetes insipidus in 76 patients (36 %), hypogonadism in 28 patients (13.2 %), growth hormone deficiency in 10 patients (4.7 %), and panhypopituitarism in 7 patients (3.3 %) (Fig. 1B).

  1. Download: Download high-res image (108KB)
  2. Download: Download full-size image

Fig. 1. Rates of surgical complications. (a) Intraoperative complications; (b) hormonal dysregulation rates following surgery.

Multivariate analysis on the predictors of Persistent/Recurrent Cushing’s disease

To identify potential predictive factors for PoRP-CD, we conducted a comprehensive binary logistic regression analysis, examining key clinical and imaging variables (Table 2). In the univariate analysis, factors including symptom duration (OR [odds ratio] 1.01, 95 % CI [confidence interval] 1.00–1.02, P = 0.04), MRI Hardy’s grade (OR 1.62, 95 % CI 0.98–2.69, P = 0.05), and previous pituitary surgery (OR 3.56, 95 % CI 1.39–9.07, P = 0.007) demonstrated significant association with PoRP-CD. MR-reported tumor size showed increased odds of recurrence with an increased tumor size (OR for microadenoma vs. no tumor: 2.41, 95 % CI: 0.50–11.53; OR for macroadenoma vs. no tumor: 4.15, 95 % CI 0.80–21.42), though the effect was not statistically significant (P > 0.05). To impede missing the marginal significant factors, three factors with P values between 0.05 and 0.15 were also included in the multivariate analysis, including “MRI Knosp grading”, “MR-reported tumor site”, and “previous pituitary radiotherapy”. In the multivariate analysis, “symptom duration” was positively correlated with recurrence, with an odds ratio (OR) of 1.03 (95 % CI: 1.01–1.06, P = 0.01), indicating a higher risk of recurrence with prolonged symptoms. Additionally, a history of “previous pituitary surgery” was significantly associated with recurrence, with an OR of 4.67 (95 % CI: 1.04–20.89, P = 0.04). Other factors, including tumor grading, tumor site, and previous radiotherapy, did not reach statistical significance.

Table 2. Regression analysis of patient and tumor’s factors related to postoperative persistence or recurrence in Cushing disease.

Parameters Univariate Analysis Multivariate Analysis
OR (95 % CI) P OR (95 % CI) P
Age 0.97 (0.94–1.01) 0.23
Sex (male vs. female) 1.17 (0.39–3.50) 0.77
Smoking (active smoker vs. non) 0.78 (0.65–10.28) 0.77
Family history of CD (positive vs. negative) 0.01 (0–Inf) 0.99
Family history of MEN (positive vs. negative) 0.01 (0–Inf) 0.99
Preoperative BMI 1.03 (0.94–1.13) 0.43
Symptom duration 1.01 (1.00–1.02) 0.04 ** 1.03 (1.01–1.06) 0.01 **
Preop serum ACTH (high vs. normal) 0.88 (0.13–6.00) 0.90
Preop free serum cortisol (high vs. normal) 1.18 (0.40–3.45) 0.74
Preop urine free cortisol (high vs. normal) 0.15 (0.01–2.98) 0.21
Knosp grading (ref: grade 0) 1.41 (0.93–2.15) 0.10 * 1.56 (0.61–3.97) 0.34
Hardy’s grading (ref: grade 0) 1.62 (0.98–2.69) 0.05 ** 1.98 (0.54–7.21) 0.29
Hardy’s staging (ref: stage A) 2.97 (0.61–14.38) 0.17
Tumor size
Macro vs. no tumor
Micro vs. no tumor		 4.15 (0.80–21.42)
2.41 (0.50–11.53)		 0.17
Multifocality (multifocal vs. unifocal) 1.68 (0.44–6.42) 0.44
MR-based tumor sitea
Bilateral vs. central
Left vs. central
Right vs. central
Stalk vs. central		 0.16 (0.01–1.53)
0.82 (0.18–4.40)
0.49 (0.09–2.82)
5.33 (0.37–144.16)		 0.14 * 0.34 (0.02–3.95)
0.23 (0.01–3.12)
5.36 (0.19–146.38)

  • (0.0002–0.67)
 0.03 **
0.39
0.27
0.31
Invasion (pos. vs. neg.) 1.18 (0.31–4.51) 0.80
Surgical approach (transplanum vs. eTSS) 6.21 (0.37–103.55) 0.20
Surgical type (adenomectomy vs. hypophysectomy) 1.55 (0.46–5.22) 0.47
Histopathology
Dense type vs. Crooke’s cell adenoma
Normal appearing vs. Crooke’s cell adenoma
Sparse type vs. Crooke’s cell adenoma		 2.00 (0.09–69.06)
0.80 (0.04–23.23)
0.28 (0.01–9.45)		 0.56
Ki-67 (>3% vs. ≤ 3 %) 1.34 (0.14–12.64) 0.79
Previous pituitary surgery (yes vs. no) 3.56 (1.39–9.07) 0.007 ** 4.67 (1.04–20.89) 0.04 **
Previous pituitary radiotherapy (yes vs. no) 3.36 (0.89–12.62) 0.07 * 3.63 (0.28–46.07) 0.31
Postop decrease in BMI 0.90 (0.73–1.03) 0.22
Abbreviations: ACTH − Adrenocorticotropic Hormone; BMI − Body Mass Index; CD − Cushing’s Disease; CI − Confidence Interval; eTSS − Endoscopic Transsphenoidal Surgery; Inf − Infinity; MEN − Multiple Endocrine Neoplasia; MR − Magnetic Resonance; OR − Odds Ratio; PoRP-CD − Persistent or Recurrent Cushing’s Disease; Preop − Preoperative; Postop − Postoperative.
aMR-reported.
* Significant at the level of 0.15.
** Significant at the level of 0.05.
The stepwise selection–in both forward and backward directions–retained four predictors— symptom duration, Hardy’s grading, tumor site, and prior surgery —for the final model. The final multivariate model with four predictors of “symptom duration”, “MRI Hardy’s grading”, “tumor site”, and “previous pituitary surgery” demonstrated significant associations for “symptom duration” (OR 1.03, 95 % CI 1.005–1.05, P = 0.02), previous pituitary surgery (OR 4.61, 95 % CI 1.12–22.0, P = 0.03), and a certain tumor site; tumors located bilaterally had significantly lower odds of recurrence compared to central tumors (OR 0.01, 95 % CI 0.0002–0.45, P = 0.02). On the testing dataset, the four-factor model achieved an AUC of 0.70, specificity of 96 %, and sensitivity of 33 %. The model’s accuracy in predicting PoRP-CD is 83 %.

Predicting persistent or recurrent Cushing’s disease–The CuPeR nomogram

A nomogram was developed based on the multivariate model comprising four key predictors: “Symptom duration”, “MRI Hardy’s grading”, “Previous pituitary surgery”, and “MRI-reported tumor site” (Fig. 2). This nomogram visually represents the impact of each predictor on the likelihood of PoRP-CD. The total score derived from the nomogram aligns with the probability scales, allowing for estimation of the risk of PoRP-CD. Higher cumulative points correspond to an increased likelihood of persistent or recurrent disease. To facilitate individualized predictions of postoperative persistence or recurrence, we developed an online dynamic nomogram (link: https://cushing.shinyapps.io/cuper/).

  1. Download: Download high-res image (97KB)
  2. Download: Download full-size image

Fig. 2. Nomogram for predicting postoperative persistence or recurrence of Cushing’s disease (PoRP-CD). This nomogram visually represents the predictive model for assessing the risk of recurrence or persistence of Cushing’s disease following surgery. Each predictor variable—Symptom duration (months), Knosp grading, Hardy’s grading, previous pituitary surgery, and tumor site— contributes a point value that aligns with the “Linear Predictor” scale, which maps to the “Probability of Persistence” scale, allowing estimation of recurrence likelihood.

Survival analysis

Survival analysis demonstrated a steady, gradual decline in DFS across the entire cohort, with the median DFS not reached despite substantial follow-up (Fig. 3A). Among the predefined variables, Hardy’s Grade 3 was associated with a significantly worse DFS compared with Grade 0 (HR = 6.02, 95 % CI: 1.09–33.02, P = 0.03) (Fig. 3B), whereas other Hardy’s Grades did not reach statistical significance (P > 0.05). Regarding tumor site, no site was a statistically significant risk factor for DFS; stalk tumors showed a trend toward poorer DFS but did not reach significance (HR = 5.09, 95 % CI: 0.84–30.63, P = 0.07) (Fig. 3C). Patients with a history of previous pituitary surgery had significantly worse DFS (HR = 4.72, 95 % CI: 2.29–9.75, P < 0.01) (Fig. 3D). In contrast, symptom duration was not associated with poor DFS (HR = 1.26, 95 % CI: 0.56–2.81, P = 0.57) (Fig. 3E). A similar analysis on OS was not performed, as only five events were recorded among the 211 patients (2.36 %), rendering meaningful statistical analysis infeasible.

  1. Download: Download high-res image (250KB)
  2. Download: Download full-size image

Fig. 3. Disease-free survival (DFS) analysis. (A) Kaplan-Meier curve of DFS for the entire cohort, showing a gradual decline over time; (B) DFS stratified by Hardy’s Grade, demonstrating significant impact of grade 3 on survival outcomes (P = 0.03); (C) DFS by tumor site, highlighting no significant association between tumor site and survival care (P > 0.05); (D) DFS based on previous surgery status, indicating a higher risk of recurrence or death in patients with prior surgical interventions (P < 0.01); (E) DFS by symptom duration, highlighting no significant association (P = 0.57).

Discussion

In this large cohort study, we developed the CuPeR model, a comprehensive predictive tool for PoRP-CD, by analyzing diverse patient and tumor characteristics, imaging findings, and treatment details. This model identified four key predictors—symptom duration, MRI Hardy’s grade, tumor site, and previous pituitary surgery. Multivariate analysis revealed that longer symptom duration and a history of prior surgery significantly increased recurrence risk, while bilateral tumor location was associated with a reduced risk. Validated with an AUC of 0.70 and 83 % accuracy on the testing dataset, the model offers significant clinical utility by providing treating surgeons with valuable insights into postoperative outcomes.
This study is among the few to develop a predictive model for estimating PoRP-CD (Table 3). Previous efforts, such as those by Liu et al. [6] and Fan et al. [7], employed machine learning and deep learning methodologies, respectively, demonstrating promising results (AUCs of 0.78 and 0.86). However, both studies were limited in their applicability to many clinical settings, as they focused solely on patients undergoing initial surgeries and incorporated postoperative parameters, which are unavailable for preoperative decision-making. By addressing these gaps, our study contributes a more practical tool for use in diverse clinical scenarios. Moreover, the findings of this study align with predictors identified in prior research. For instance, factors such duration of symptoms and history of previous pituitary surgery have been highlighted as critical for recurrence [6,14]. Importantly, our inclusion of MRI-based predictors and preoperative variables ensures the model’s relevance during preoperative planning, distinguishing it from previous approaches.

Table 3. Studies on predictive models or patients and tumors predictive factors of post-operative remission of Cushing’s disease.

Empty Cell Year Country Study Size Methods Main Findings Ref.
Predictive Models
Comprising 8 factors:
age,
disease coarse,
morning serum ACTH (preop),
morning serum cortisol (preop),
urine free cortisol (preop),
morning serum ACTH nadir (postop),
morning serum cortisol nadir (postop),
urine free cortisol nadir (postop)		 2019 China 354 Machine-learning using Random Forest algorithm Sensitivity 87 %, specificity 58 %
AUC 0.78		 [6]
Comprising 5 factors:
age,
disease coarse,
morning serum ACTH (postop),
morning serum cortisol nadir (postop),
urine free cortisol nadir (postop)		 2021 China 354 Deep-learning using factorization‑machine based neural approach AUC 0.86 [7]
Predictive Factors
Serum cortisol < 35 nmol/L (6–12 w after surgery) 1993 UK 11 Prospective Favorable long-term remission rate [15]
Serum 11-deoxycortisol > 150 nmol/L after metyrapone test at 14 days post-surgery 1997 Netherlands 29 Retrospective Higher risk of recurrence
Sensitivity 100 %, specificity 75 %		 [16]
Serum cortisol < 2 μ/dL (3–8 d after surgery) 2001 Japan 49 Retrospective Recurrent disease in 4 % of patients [17]
MRI-based tumor size and cavernous sinus invasion 2003 Italy 26 Retrospective Unfavorable factors of persistent disease [18]
No histological evidence of adenoma 2007 US 490 Retrospective Lower remission rate [19]
Long-term hypocortisolism after surgery (≥13 m) 2017 India 230 Retrospective Favorable for remission
Sensitivity 46 %, specificity 100 %		 [20]
Greater decrease in BMI after surgery
Lower DHEAS before surgery		 2017 Taiwan 41 Retrospective Favorable factors for higher remission [21]
High serum ACTH/cortisol ratio before surgery 2018 Turkey 119 Retrospective Risk factor for disease recurrence [22]
USP8 mutation 2018 Germany 48 Retrospective Higher recurrence rate [23]
Serum cortisol < 107 nmol/L after betamethasone suppression test following surgery 2018 Sweden 28 Interventional Sensitivity 85 %, specificity 94 %
AUC 0.92		 [24]
Tumor visualization on MRI before surgery 2022 Spain 40 Retrospective Favorable factor for remission [25]
Abbreviations: ACTH − Adrenocorticotropic Hormone; AUC − Area Under the Curve; BMI − Body Mass Index; DHEAS − Dehydroepiandrosterone Sulfate; MRI − Magnetic Resonance Imaging; PoRP-CD − Persistent or Recurrent Cushing’s Disease; Preop − Preoperative; Postop − Postoperative; USP8 − Ubiquitin Specific Peptidase 8.
Several other studies aimed to explore the predictive value of single predictors. Braun et al. (2020) summarized the predictors for CD remission following TSS in a systematic review. Key predictors include pre-surgical identification of the tumor via MRI and the absence of adenoma invasion into the cavernous sinus. Postoperative hormonal levels, particularly low cortisol (< 2 µg/dL) and ACTH levels (< 3.3 pmol/L) as well as low cortisol levels (< 35 nmol/L) at 6–12 weeks post-surgery and sustained hypocortisolism requiring long-term replacement therapy, were significant indicators of remission. Additionally, post-surgical decreases in BMI contributed to favorable outcomes. Other reported predictors included a high level of surgical expertise, younger patient age, non-mutant USP8 corticotroph tumors, and swift recovery from postoperative adrenal insufficiency [5].
This study has certain limitations that should be acknowledged. The reliance on retrospective data may result in potential biases in variable selection and data completeness. While the model demonstrated good predictive accuracy, its limited sensitivity may restrict its ability to identify all high-risk patients. Moreover, the model has not been externally validated in independent cohorts, which limits its generalizability to other clinical settings. Despite these limitations, the study possesses significant strengths that underscore its contribution to the field. Applying one of the largest CD cohorts, it provides a robust statistical foundation and enhances the reliability of the findings. The comprehensive inclusion of diverse patient and tumor characteristics, imaging findings, and treatment details resulted in a clinically relevant and well-rounded predictive model. Notably, this model stands out for its applicability to a broader spectrum of patients, including those with prior surgeries or radiotherapy, addressing a gap left by earlier studies. Furthermore, the development of an online dynamic nomogram bridges the gap between research and clinical practice, allowing personalized predictions and aiding surgeons in making informed decisions before pituitary surgery.
Although this study incorporated long-term follow-up (median 58 months) to define persistence and recurrence and to internally validate the model, external validation in prospective, multi-institutional cohorts remains essential to confirm its broader applicability. Although the CuPeR model incorporates a wide array of clinical, radiological, biochemical, and demographic variables, other potential prognostic factors were not included and may warrant consideration in future studies. For instance, the presence of osteoporosis, degree of tumor invasion, and early recovery of the adrenal axis during the postoperative period have all been reported as relevant predictors of outcomes in Cushing’s disease [26]. Moreover, the role of surgical expertise is critical, as higher surgeon and institutional experience are strongly associated with improved remission and lower recurrence rates [27]. Incorporating novel parameters, such as genetic markers or advanced imaging techniques, could further enhance the predictive accuracy and clinical utility of the model. Prospective implementation of the nomogram in routine clinical workflows will provide valuable insights into its performance and its potential to improve patient outcomes.

Conclusions

This study introduced a practical, predictive model for estimating the risk of postoperative persistence and recurrence in Cushing’s disease, possibly offering a reliable tool for preoperative planning. By integrating key clinical predictors into an interactive online dynamic nomogram, the CuPeR model may provide surgeons with personalized risk assessments to aid in preoperative planning. Its focus on preoperative data ensures broader applicability, paving the way for tailored therapeutic strategies and improved patient outcomes in diverse clinical scenarios.

Funding details

None.

CRediT authorship contribution statement

Guive Sharifi: Supervision, Conceptualization. Elham Paraandavaji: Investigation, Data curation. Nader Akbari Dilmaghani: Investigation, Data curation. Tohid Emami Meybodi: Investigation, Data curation. Ibrahim Mohammadzadeh: Investigation, Data curation. Neginalsadat Sadeghi: Investigation, Data curation. Amirali Vaghari: Visualization. Behnaz Niroomand: Visualization. Seyed Mohammad Tavangar: Resources. Mohammad reza Mohajeri Tehrani: Validation. Zahra Davoudi: Resources. Marjan Mirsalehi: Writing – review & editing. Seyed Ali Mousavinejad: Validation, Resources. Farzad Taghizadeh-Hesary: Writing – review & editing, Writing – original draft.

Informed consent

Not applicable.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

None.
The data that support the findings of this study are available on request from the corresponding author.

References

https://www.sciencedirect.com/science/article/pii/S2214623725000353

Filed under: Cushing's, pituitary, symptoms, Treatments | Tagged: , , , | Leave a comment »

Enhanced Radiological Detection of a Corticotroph Adenoma Following Treatment With Osilodrostat

Posted on October 28, 2025 by MaryO

Abstract

In approximately 30% of patients with Cushing disease, pituitary magnetic resonance imaging (MRI) does not reliably identify a corticotroph adenoma. Importantly, surgical remission rates are >2.5 fold higher for microadenomas that are radiologically visible on preoperative imaging when compared with “MRI-negative” cases. We describe a 42-year-old woman with Cushing disease, in whom MRI findings at presentation were equivocal with no clear adenoma visualized. She was initially treated with metyrapone, which resulted in partial biochemical control of hypercortisolism. After switching to osilodrostat, there was a marked improvement in her symptoms and rapid normalization of cortisol levels. Following 3 months of eucortisolemia, [11C]methionine positron emission tomography (MET-PET) coregistered with volumetric MRI (MET-PET/MRCR) localized the site of the corticotroph tumor and the patient underwent successful transsphenoidal resection. She remains in full clinical and biochemical remission at >2 years postsurgery. This case suggests that a period of eucortisolemia induced by osilodrostat may facilitate localization of corticotroph microadenomas using functional (PET) imaging.

Cushing disease, osilodrostat, [11C]methionine PET, corticotroph adenoma
Issue Section:

Case Report

Introduction

Cushing disease, caused by an ACTH-secreting pituitary adenoma, accounts for approximately 80% of endogenous Cushing syndrome [1]. Although transsphenoidal surgery remains the preferred treatment for the majority of patients, even in expert centers recurrence rates as high as 27% have been reported [2, 3]. Surgery is preferred over medical therapy because it offers the potential for definitive cure by directly removing the pituitary adenoma. In contrast, medical therapy is typically reserved for patients in whom surgery is contraindicated, incomplete, or has failed to achieve remission. Linked to this, magnetic resonance imaging (MRI) fails to detect an adenoma in approximately one third of cases [4]. In a recent systematic review, postsurgical remission rates were 2.63-fold higher (95% CI, 2.06-3.35) for MRI-detected corticotroph adenomas when compared with “MRI-negative” cases [5]. Several alternative magnetic resonance sequences have therefore been proposed to aid tumor localization (including dynamic and volumetric [eg, gradient recalled echo MRI]), but these still fail to detect a significant proportion of microcorticotropinomas [6, 7]. Accordingly, molecular (functional) imaging with positron emission tomography (PET) radiotracers that target key properties of corticotroph adenomas (eg, [11C]methionine [MET-PET], [18F]fluoroethyltyrosine, or [68Ga]DOTA-corticotropin-releasing hormone PET) has been proposed as an additional tool for localizing corticotroph tumors that evade detection on conventional MRI [6-10].

Medical therapy is often required for patients in whom surgery is not an immediate option or when there is persistent hypercortisolism postoperatively [11]. Cortisol-lowering treatment may also be considered before surgery to reduce morbidity and perioperative complications [11]. An important recent addition to the armory of medications used to treat Cushing syndrome is osilodrostat, a potent oral inhibitor of the key adrenal steroidogenic enzyme 11β-hydroxylase [12, 13].

Here, we describe how preoperative medical therapy with osilodrostat yielded dual benefits in a patient with inconclusive primary imaging: (1) rapid and effective control of hypercortisolism and (2) facilitation of the localization of a previously occult microcorticotroph adenoma using MET-PET coregistered with volumetric MRI (MET-PET/MRCR).

Case Presentation

A 42-year-old woman presented with a 7-year history of progressive central weight gain, facial plethora, acne, worsening hypertension, depression, and proximal myopathy. Her symptoms had become more pronounced during the COVID-19 pandemic, leading to profound emotional distress and functional decline. She described feeling persistently tearful and fatigued, with markedly reduced energy levels that rendered her unable to work or care for her young child, and severely affecting her quality of life. She had no significant medical history and was taking amlodipine and the progesterone-only pill. On examination, her body mass index was 29.6 kg/m² and blood pressure was markedly elevated at 197/111 mm Hg. Clinical features consistent with hypercortisolism included easy bruising, centripetal adiposity, and proximal muscle wasting. Initial laboratory evaluation was unremarkable; however, her hemoglobin A1c was at the upper end of normal (41 mmol/mol or 5.9%).

Diagnostic Assessment

Biochemical testing confirmed ACTH-dependent Cushing syndrome (Table 1). Cortisol levels following overnight and 48-hour dexamethasone suppression were elevated at 8 µg/dL (SI: 219 nmol/L) and 16 µg/dL (SI: 434 nmol/L), respectively (reference range: < 1.8 µg/dL [SI: < 50 nmol/L]). Plasma ACTH concentrations ranged from 36 to 55 ng/L (SI: 7.9-12.1 pmol/L) (reference range: 10-30 ng/L [SI: 2.2-6.6 pmol/L]), consistent with an ACTH-driven process. Urinary free cortisol (UFC) was markedly elevated at 690.95 µg/24 hours (SI: 1907 nmol/24 hours) (reference range: 18-98 µg/24 hours [SI: 50-270 nmol/24 hours]). Late-night salivary cortisol and cortisone levels were also elevated at 0.95 µg/dL (SI: 26.2 nmol/L) (reference range: < 0.09 µg/dL [SI: < 2.6 nmol/L]) and 2.7 µg/dL (SI: 74.5 nmol/L) (reference range: < 0.7 µg/dL [SI: < 18 nmol/L]) respectively. Inferior petrosal sinus sampling excluded an ectopic source of ACTH production (central-to-peripheral ACTH ratio: baseline 18.60, 0 minutes 18.4, peak at 2 minutes 94.9, 5 minutes 42.4, 10 minutes 22.3) (Table 2). However, pituitary MRI findings were inconclusive, with no definite adenoma identified. In addition, the left intracavernous carotid artery encroached medially, creating a narrow intercarotid window with distortion of normal pituitary anatomy (Fig. 1). Given these findings, the decision was made to initiate cortisol-lowering therapy and to reassess imaging appearances after a period of biochemical normalization.

Pituitary MRI at initial presentation. No discrete adenoma is visible on T1-weighted coronal precontrast (A) and postcontrast (B), T2-weighted coronal (C), and T1-weighted sagittal postcontrast (D) sequences. The sellar anatomy appears asymmetric, consistent with a medially positioned left internal carotid artery.

Figure 1.

Pituitary MRI at initial presentation. No discrete adenoma is visible on T1-weighted coronal precontrast (A) and postcontrast (B), T2-weighted coronal (C), and T1-weighted sagittal postcontrast (D) sequences. The sellar anatomy appears asymmetric, consistent with a medially positioned left internal carotid artery.

Open in new tabDownload slide

Table 1.

Open in new tab

Biochemical investigations at diagnosis confirming ACTH-dependent Cushing syndrome

Tests Results Reference Range
Overnight dexamethasone suppression test (ONDST) Cortisol: 8 µg/dL (SI: 219 nmol/L) <1.8 µg/dL (SI: < 50 nmol/L)
48-hour dexamethasone suppression test (DST) Cortisol: 16 µg/dL (SI: 434 nmol/L) <1.8 µg/dL (SI: < 50 nmol/L)
ACTH 36-55 ng/L (SI: 7.9-12.1 pmol/L) 10-30 ng/L (SI: 2.2-6.6 pmol/L)
24-hour urinary free cortisol (UFC) 690.95 μg/24 h (SI: 1907 nmol/24 h) 18-98 µg/24 h (SI: 50-270 nmol/24 hours)
Late-night salivary cortisol
late-night salivary cortisone		 0.95 µg/dL (SI: 26.2 nmol/L)
2.7 µg/dL (SI: 74.5 nmol/L)		 <0.09 µg/dL (SI: <2.6 nmol/L) <0.7 µg/dL (SI: <18 nmol/L)

Results are reported in both conventional and SI units with reference ranges shown in parentheses.

Table 2.

Open in new tab

Results of inferior petrosal sinus sampling (IPSS)

Time Plasma ACTH
(min) Left petrosal sinus Right petrosal sinus Peripheral vein
−5 1159 ng/L (255 pmol/L) 144 ng/L (32 pmol/L) 62.3 ng/L (14 pmol/L)
0 1147 ng/L (253 pmol/L) 222 ng/L (49 pmol/L) 62.3 ng/L (14 pmol/L)
2 5257 ng/L (1157 pmol/L) 2159 ng/L (475 pmol/L) 55.4 ng/L (12.2 pmol/L)
5 3677 ng/L (810 pmol/L) 2976 ng/L (655 pmol/L) 86.8 ng/L (19 pmol/L)
10 2251 ng/L (496 pmol/L) 545 ng/L (120 pmol/L) 101 ng/L (22 pmol/L)

Time Plasma cortisol
(min) Left petrosal sinus Right petrosal sinus Peripheral vein
−5 24.94 μg/dL (668 nmol/L) 25.30 μg/dL (698 nmol/L) 23.56 μg/dL (650 nmol/L)
0 25.08 μg/dL (692 nmol/L) 24.07 μg/dL (664 nmol/L) 23.34 μg/dL (644 nmol/L)
2 23.31 μg/dL (643 nmol/L) 24.32 μg/dL (671 nmol/L) 23.78 μg/dL (656 nmol/L)
5 21.97 μg/dL (606 nmol/L) 23.67 μg/dL (653 nmol/L) 23.23 μg/dL (641 nmol/L)
10 27.62 μg/dL (762 nmol/L) 26.17 μg/dL (722 nmol/L) 25.26 μg/dL (697 nmol/L)

Time Plasma prolactin
(min) Left petrosal sinus Right petrosal sinus Peripheral vein
−5 1835 mU/L (86 μg/L) 356 mU/L (17 μg/L) 251 mU/L (11 μg/L)
0 1725 mU/L (81 μg/L) 498 mU/L (23 μg/L) 248 mU/L (12 μg/L)
2 2151 mU/L (101 μg/L) 409 mU/L (19 μg/L) 240 mU/L (11 μg/L)
5 2239 mU/L (105 μg/L) 711 mU/L (33 μg/L) 246 mU/L (12 μg/L)
10 1883 mU/L (89 μg/L) 410 mU/L (19 μg/L) 244 mU/L (11 μg/L)

Central-to-peripheral ACTH gradients before and after corticotropin-releasing hormone (CRH) stimulation support a pituitary source of ACTH secretion. Reference cutoffs: basal ACTH gradient ≥2 and/or CRH-stimulated ACTH gradient ≥3 indicate central ACTH secretion.

Treatment

The patient was started on metyrapone, but despite dose escalation up to 4000 mg daily, which was associated with significant nausea and malaise, she did not achieve eucortisolemia (Fig. 2C). She was therefore transitioned to osilodrostat, which rapidly normalized cortisol levels within 5 weeks at a maintenance dose of 6 mg twice daily (Fig. 2B and 2C). In contrast to metyrapone, osilodrostat was well-tolerated with no reported side effects. Serum cortisol and clinical status were closely monitored throughout, with no biochemical or clinical evidence of adrenal insufficiency.

Bar charts illustrating changes in urinary, salivary, and serum cortisol, as well as serum ACTH, during medical treatment. (A) A 24-hour UFC (black bars, left y-axis) normalized during osilodrostat treatment, whereas serum ACTH (gray bars, right y-axis) increased. Dotted lines represent the upper limit of normal: 59.4 µg/24 hours (SI: 164 nmol/24 hours) for UFC and 30 ng/L (SI: 6.6 pmol/L) for ACTH. X-axis labels indicate treatment week and total daily osilodrostat dose. (B) Salivary free cortisol levels, collected alongside serum cortisol during a cortisol day curve (at 09:00, 12:00, 15:00, and 18:00), fully normalized with osilodrostat therapy. Bar shading from black to light gray denotes sampling time. The dotted line indicates upper limit of normal: 9.4 ng/dL (SI: 2.6 nmol/L). (C) Serum free cortisol levels during day curves showed inadequate control on escalating doses of metyrapone, with normalization achieved following initiation of osilodrostat.

Figure 2.

Bar charts illustrating changes in urinary, salivary, and serum cortisol, as well as serum ACTH, during medical treatment. (A) A 24-hour UFC (black bars, left y-axis) normalized during osilodrostat treatment, whereas serum ACTH (gray bars, right y-axis) increased. Dotted lines represent the upper limit of normal: 59.4 µg/24 hours (SI: 164 nmol/24 hours) for UFC and 30 ng/L (SI: 6.6 pmol/L) for ACTH. X-axis labels indicate treatment week and total daily osilodrostat dose. (B) Salivary free cortisol levels, collected alongside serum cortisol during a cortisol day curve (at 09:00, 12:00, 15:00, and 18:00), fully normalized with osilodrostat therapy. Bar shading from black to light gray denotes sampling time. The dotted line indicates upper limit of normal: 9.4 ng/dL (SI: 2.6 nmol/L). (C) Serum free cortisol levels during day curves showed inadequate control on escalating doses of metyrapone, with normalization achieved following initiation of osilodrostat.

Open in new tabDownload slide

ACTH levels progressively increased as the dose of osilodrostat was escalated (Fig. 2A). After 3 months of biochemical eucortisolism, she underwent Met-PET/MRCR, which revealed a distinct methionine-avid lesion in the right posterolateral aspect of the sella (Fig. 3). Imaging was performed as previously reported [7, 8, 14]. Conventional MRI findings remained stable, with no new abnormalities. As she remained clinically and biochemically eucortisolemic on osilodrostat, glucocorticoid supplementation was not required pre- or perioperatively.

11C-Methionine PET/CT coregistered with volumetric MRI (MET-PET/MRCR) following treatment with osilodrostat. A subtle area of reduced gadolinium enhancement can now be appreciated on the right posterosuperior aspect of the gland (A-C). MET-PET/MRCR confirms focal tracer uptake at this site (yellow arrows) and also within normal gland anteriorly (white arrow) (D-F). Three-dimensional reconstruction using CT, MRI, and PET datasets demonstrating the location of the corticotroph microadenoma which was confirmed at subsequent surgery (G-H).

Figure 3.

11C-Methionine PET/CT coregistered with volumetric MRI (MET-PET/MRCR) following treatment with osilodrostat. A subtle area of reduced gadolinium enhancement can now be appreciated on the right posterosuperior aspect of the gland (A-C). MET-PET/MRCR confirms focal tracer uptake at this site (yellow arrows) and also within normal gland anteriorly (white arrow) (D-F). Three-dimensional reconstruction using CT, MRI, and PET datasets demonstrating the location of the corticotroph microadenoma which was confirmed at subsequent surgery (G-H).

Open in new tabDownload slide

Outcome and Follow-up

At transsphenoidal surgery, abnormal tissue was resected from the site identified on MET-PET/MRCR. Histological examination revealed normal anterior pituitary tissue (adenohypophysis) with no evidence of a pituitary adenoma. Occasional cells showed possible Crooke’s hyaline change. The Ki-67 proliferation index was very low (<1%). Despite the absence of histological confirmation of a corticotroph adenoma, the patient entered complete biochemical and clinical remission. Early postoperative cortisol was 3 µg/dL (SI: 82.8 nmol/L), prompting initiation of glucocorticoid replacement with prednisolone. Prednisolone was chosen for its longer half-life, enabling convenient once-daily dosing. We routinely monitor prednisolone levels to guide adjustment of replacement dosing. Prednisolone was successfully tapered over a period of 6 months, with biochemical confirmation of adrenal recovery. At 2 years postsurgery, the patient had no clinical features of hypercortisolism with sustained weight loss of >20 kg. Morning 09:00 cortisol and ACTH were consistent with ongoing eucortisolism. Serial late-night salivary cortisol and cortisone levels were normal, and cortisol was undetectable following a 1-mg overnight dexamethasone suppression test, confirming durable remission of Cushing disease.

Discussion

Early transsphenoidal surgery remains the treatment of choice for most patients with Cushing disease, with the highest chance of cure achieved following a successful first operation [11]. However, even in expert centers, persistent or recurrent disease is diagnosed during follow-up, and is more likely when initial MRI has failed to identify a clear surgical target [5]. Reoperation carries increased technical difficulty and a higher risk of iatrogenic hypopituitarism, underscoring the importance of accurate preoperative localization of corticotroph adenomas. Our case illustrates a potential novel added benefit of a trial of primary medical therapy in a patient with Cushing disease and equivocal or negative MRI findings at initial presentation. Specifically, we have shown how osilodrostat, a potent inhibitor of 11β-hydroxylase, can achieve rapid normalization of cortisol levels, consistent with the findings of the LINC (LCI699 [osilodrostat] in Cushing disease) series of studies [15-17], and at the same time help reveal the location of the occult microcorticotropinoma. An important consequence of achieving effective adrenal blockade in our patient was the more than threefold accompanying rise in plasma ACTH levels (Fig. 2). We hypothesized that such an increase in tumoral activity might facilitate its detection using molecular (functional) imaging. MET-PET has been shown in several studies to facilitate localization of de novo and recurrent corticotroph adenomas [8, 18, 19] in a significant proportion of patients with equivocal or negative MRI findings. We have now shown that such an approach could potentially be enhanced by pretreatment with the potent 11β-hydroxylase inhibitor osilodrostat.

We also considered whether the rise in ACTH during osilodrostat therapy reflected increased tumor activity alone or was associated with a change in tumor size. In our case, ACTH rose significantly, likely reflecting enhanced secretory activity, whereas repeat conventional MRI remained stable, with no new abnormalities or interval changes. In the LINC 4 study, tumor volume data were available for 35 patients at both baseline and week 48. Among these, 40.0% had a ≥20% increase, 28.6% had a ≥20% decrease, and 31.4% had <20% change in tumor volume. These outcomes were observed in both microadenomas and macroadenomas, with no clear correlation to treatment duration or osilodrostat dose [20]. This variability suggests that osilodrostat does not exert a consistent effect on tumor volume.

Interestingly, although histopathological analysis did not confirm a corticotroph adenoma, this is a well-recognized finding and has been reported in a significant proportion of patients undergoing surgery for Cushing disease [21, 22]. Nonetheless, we consider the diagnosis of pituitary-dependent Cushing syndrome was clearly established by the clinical features, results of initial laboratory testing and findings at inferior petrosal sinus sampling (which demonstrated a clear central-to-peripheral ACTH gradient). In addition, abnormal tissue was identified intraoperatively at the site visualized on MET-PET and fully resected, and no other abnormal foci of tissue were seen. The patient has subsequently achieved complete and sustained clinical and biochemical remission, consistent with successful removal of an ACTH-secreting adenoma.

Recent case reports have raised concerns about prolonged adrenal insufficiency following extended osilodrostat use—an unexpected finding given the drug’s short half-life [23-25]. Although adrenal insufficiency requiring temporary glucocorticoid replacement had been reported in clinical trials (most commonly in patients undergoing rapid dose escalation [12, 15, 16]), prolonged hypothalamopituitary-adrenal axis suppression resulting from supraphysiologic glucocorticoid replacement could also be contributory. For now, the exact mechanism of this observed phenomenon remains unclear. Our patient managed to wean glucocorticoid replacement postoperatively and did not demonstrate prolonged adrenal suppression; at the same time, clinical and biochemical testing confirmed full remission from Cushing disease.

This case supports the hypothesis that preoperative cortisol suppression may enhance the diagnostic accuracy of molecular (functional) imaging in Cushing disease, particularly in cases with inconclusive MRI findings. If validated in prospective studies, this approach could refine surgical planning and potentially lead to better surgical success and durable clinical outcomes.

Learning Points

  • Approximately 30% of corticotroph adenomas causing Cushing disease are not readily localized on conventional pituitary MRI.

  • Functional imaging modalities such as MET-PET/MRCR can improve detection of previously occult pituitary adenomas in Cushing disease.

  • A period of medical pretreatment with osilodrostat, with consequent reduction in negative feedback by glucocorticoid at the hypothalamic-pituitary level, may augment tumor localization by molecular imaging.

Acknowledgments

The authors acknowledge Debbie Papadopoulou and Niamh Martin for their contributions to clinical management. Nigel Mendoza performed the transsphenoidal surgery.

Contributors

All authors made individual contributions to authorship. Z.H., L.Y., J.M., M.G., and F.W. were involved in the diagnosis and management of this patient and manuscript submission. J.M., D.G., and M.G. performed and analyzed the patient’s functional imaging. All authors reviewed and approved the final draft.

Funding

No public or commercial funding

Disclosures

None declared.

Informed Patient Consent for Publication

Signed informed consent obtained directly from the patient.

Data Availability Statement

Original data generated and analyzed during this study are included in this published article.

© The Author(s) 2025. Published by Oxford University Press on behalf of the Endocrine Society.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. See the journal About page for additional terms.

Filed under: Cushing's, pituitary, Treatments | Tagged: , , , , | Leave a comment »

Osilodrostat for Cyclic Cushing’s Disease

Posted on October 24, 2025 by MaryO

Highlights

  • •
    Cyclic Cushing’s syndrome (CCS) is a rare entity with significant comorbidities
  • •
    It is defined by at least 3 peaks of hypercortisolism, 2 troughs of eucortisolism
  • •
    Surgical cure is preferred, and medications are second-line
  • •
    Our case is the first showing successful treatment of native CCS with osilodrostat
  • •
    Osilodrostat showed rapid onset/offset and reversible inhibition of steroidogenesis

Abstract

Background/Objective

Cyclic Cushing’s syndrome is a rare subtype of Cushing’s syndrome with episodes of hypercortisolism, followed by spontaneous remission.

Case Report

Our patient was a 68-year-old male who presented with his third cycle of cyclic Cushing’s disease with facial swelling, buffalo hump, fatigue, proximal muscle weakness, and lower extremity edema. Laboratory tests showed the following: 24-hour urine free cortisol 12030.3 mcg/d (normal <= 60.0 mcg/d), morning adrenocorticotropic hormone (ACTH) 464 pg/mL (normal 6-59 pg/mL), morning serum cortisol 91 mcg/dL (normal 8-25 mcg/dL), and potassium 3.3 mmol/L (normal 3.6-5.3 mmol/L). MRI pituitary without/with contrast showed a partially empty sella. Prior inferior petrosal sinus sampling during the second cycle indicated a potential pituitary source of increased ACTH production, localized or draining to the right side. The patient was treated with osilodrostat with improvement in laboratory values and clinical symptoms by 2-3 weeks. After development of adrenal insufficiency (AI), osilodrostat was rapidly titrated off by 2 months of treatment. Subsequently, labs after 8 days off osilodrostat confirmed clinical remission and reversibility of medication-induced AI.

Discussion

Since hypercortisolism is associated with mortality risk and comorbidities, timely management is a priority. If a surgical cure is not possible, a medication that treats hypercortisolism with rapid onset, reversible inhibition, and minimal side effects would be ideal to address the cyclicity.

Conclusion

Our case is the first to our knowledge demonstrating osilodrostat’s use for native cyclic Cushing’s syndrome treatment and highlighted its reversibility and ability to preserve normal adrenal function.

Keywords

Osilodrostat
cyclic Cushing’s disease
cyclic Cushing’s syndrome

Introduction

Cyclic Cushing’s syndrome is a rare entity that represents a clinical challenge. It is defined by at least 3 peaks of biochemical hypercortisolism, which is clinically symptomatic in the majority though rarely asymptomatic, and 2 troughs with normalized cortisol production that can last from days to years.1 The phenomenon can arise from any potential source of Cushing’s syndrome, including pituitary (54%), ectopic (26%), adrenal (11%), and unclassified (9%) sources.1 Intermittent hypercortisolism can also occur after pituitary surgery for Cushing’s disease.2
The cyclicity interferes with a straightforward diagnosis. It can lead to paradoxical results from biochemical testing and inferior petrosal sinus sampling (IPSS),3 making determination of therapeutic outcomes more complicated.3 The goal of cyclic Cushing’s syndrome management, as in all types of Cushing’s syndrome, is early diagnosis and intervention to reduce the length of hypercortisolism.4 A surgical cure is preferred, as Cushing’s syndrome is associated with a five-fold increased standardized mortality risk.4 Cardiovascular, metabolic, bone, and cognitive comorbidities may persist despite remission and must be aggressively managed.4,5 For patients in whom surgical management is not possible or has not led to remission, medical therapy has a crucial role. We describe the first case to our knowledge of native cyclic Cushing’s syndrome treated successfully with osilodrostat. A case of exogenous cyclic ACTH-independent Cushing’s syndrome from pembrolizumab, with cyclicity attributed to the infusions, also demonstrated successful treatment with osilodrostat.6

Case Report

The patient was a 68-year-old male with hypertension, hyperlipidemia, and rheumatoid arthritis with a history of cyclical episodes of weight gain and facial swelling, occurring spontaneously without steroid treatments. The initial episode occurred at age 62 for 5 months, and returned at age 64 with facial swelling, buffalo hump, fatigue, proximal muscle weakness, sleep disturbances, and lower extremity edema. Laboratory tests showed the following (Table 1): 24-hour urine free cortisol >245 mcg/d (normal 11-84 mcg/d), morning adrenocorticotropic hormone (ACTH) 528.0 pg/mL (normal 7.2-63.3 pg/mL) and morning serum cortisol 91.7 mcg/dL (confirmed on dilution; normal 6.2-19.4 mcg/dL). Laboratory tests were also notable for a mildly low potassium level, low prolactin, low testosterone, and normal thyroid hormone, insulin-like growth factor-1 (IGF-1), and dehydroepiandrosterone sulfate (DHEA-S) levels. MRI pituitary without/with contrast showed no sellar and suprasellar masses. A prior CT abdomen/pelvis with contrast at age 62 noted unremarkable adrenal glands. The patient was referred for inferior petrosal sinus sampling (IPSS) (Table 2), which indicated a potential pituitary source of increased ACTH production, localized or draining to the right side. The central to peripheral gradient was >2 in the first pre-stimulation sample and >3 in all samples after providing 10mcg of desmopressin (DDAVP). There was a >1.4/1 gradient between the right and left sides, suggesting a potential pituitary source draining to the right side (Table 2). The inferior petrosal sinuses were normal and of similar size. Cushing’s symptoms receded spontaneously in 5 months, and the patient did not follow up until recurrence at age 67.

Table 1. Labs at time of onset of cyclical episodes

Empty Cell Labs at age 64 y/o (2nd episode) Labs at age 67 y/o (3rd episode)
24hr urine free cortisol level >245 mcg/24hr (normal 11-85 mcg/24hr) 12030.3 mcg/d (normal <= 60.0 mcg/d)
24hr urine creatinine 1495 mg/24hr (normal 1000-2000mg/24hr) 1868 mg/day (normal 800-2100 mg/day)
Morning ACTH 528.0 pg/mL (normal 7.2-63.3 pg/mL) 464 pg/mL (normal 6-59 pg/mL),
Morning cortisol 91.7 mcg/dL (normal 6.2-19.4 mcg/dL) 91 mcg/dL (normal 8-25 mcg/dL)
Thyroid-stimulating hormone level (TSH) 0.452 mcIU/mL (normal 0.450-4.500 mcIU/mL) 0.08 mcIU/mL (normal 0.3-4.7 mcIU/mL)
Free thyroxine (free T4) 1.34 ng/dL (normal 0.82-1.77 ng/dL) 1.30 ng/dL (normal 0.8-1.7 ng/dL)
Prolactin <1.0 ng/mL (normal 3.0-15.2 ng/mL) 8.05 ng/mL (normal 3.5-19.4 ng/mL)
Insulin-like growth factor-1 (IGF-1) 148 ng/mL (normal 64-240 ng/mL) 128 ng/mL (normal 41-279 ng/mL)_
Testosterone panel Total 66 ng/dL(11AM)
(normal 264-916 ng/dL)
Free 9.6 pg/mL (11AM)
(normal 6.6-18.1 pg/mL)		 Total 107 ng/dL (8:30AM)
(normal 300-720 ng/dL)
Bioavailable 61 ng/mL (8:30AM)
(normal 131-682 ng/mL)
Follicle-Stimulation Hormone (FSH) 3.6 mIU/mL (normal 1.6-9 mIU/mL)
Luteinizing Hormone (LH) 1.6 mIU/mL (normal 2-12 mIU/mL)
Dehydroepiandrosterone sulfate (DHEA-S) 153 mcg/dL (normal 48.9-344.2 mcg/dL)
Potassium level 3.2 mmol/L (normal 3.4-4.8 mmol/L) 3.3 mmol/L (normal 3.6-5.3 mmol/L)
Creatinine level 0.92 mg/dL (normal 0.7-1.2 mg/dL) 0.89 mg/dL (normal 0.6-1.3 mg/dL)

Table 2. Inferior Petrosal Sinus Sampling (IPSS)

Empty Cell Time Right IPS
ACTH level (normal 6-59 pg/mL)		 Left IPS
ACTH level (normal 6-59 pg/mL)		 Inferior Vena Cava ACTH level (normal 6-59 pg/mL) Serum Cortisol (normal 8-25 mcg/dL)
Baseline 1 08:25 AM 32 23 14 7
Baseline 2 08:27 AM 19 16 13 7
Desmopressin (DDAVP) 08:30 AM
Post 2 min 08:32 AM 150 34 15
Post 5 min 08:35 AM 123 32 18
Post 10 min 08:40 AM 49 26 17
Post 15 min 08:45 AM 124 31 17
Post 30 min 09:00 AM 107 28 13
*These results may indicate a pituitary source for increased ACTH production, localized or draining to the right side. There is a Central:Peripheral gradient of >2 (right IPS) in the first pre-stimulation samples and >3 in all post-desmopressin (DDAVP) 10mcg samples. If due to an adenoma, it might drain into the right given the presence of a significant (greater than 1.4/1) gradient between right and left. The inferior petrosal sinuses were of similar size and normal. These results must take into account the patient’s clinical scenario, and there are false positives and possible overlap with normal results.
*Abbreviation: min = minutes
During the third and most recent cycle of Cushing’s syndrome, laboratory tests after 1 month of symptom development showed the following (Table 1): 24-hour urine free cortisol 12030.3 mcg/d (normal <= 60.0 mcg/d), morning ACTH 464 pg/mL (normal 6-59 pg/mL), morning serum cortisol 91 mcg/dL (normal 8-25 mcg/dL), potassium level 3.3 mmol/L (normal 3.6-5.3 mmol/L), and mild leukocytosis and erythrocytosis. Repeat MRI pituitary without/with contrast showed a partially empty sella and no pituitary mass (Figure 1).

  1. Download: Download high-res image (517KB)
  2. Download: Download full-size image

Figure 1. MRI pituitary without/with contrast at the time of the third cyclical episode of Cushing’s disease. The MRI showed a partially empty sella with no evidence of a pituitary mass. Left) Coronal view. Right) Sagittal view.

The patient was started on osilodrostat 2mg twice daily. By week 2 of treatment, the morning cortisol level improved to 9.5 mcg/dL (8-25 mcg/dL) and potassium level normalized, though facial and body swelling persisted. Significant improvement in symptoms and fatigue were noted by week 3 of treatment with the following labs: morning ACTH 145 pg/mL (normal 6-59 pg/mL), morning serum cortisol 5.4 mcg/dL (8-25 mcg/dL), and 24-hour urine free cortisol 7 mcg/d (normal 5-64 mcg/d). The osilodrostat dose was decreased to 1mg twice daily, then 1mg daily, and stopped by 2 months of treatment after development of adrenal insufficiency (AI), which was confirmed on laboratory results (Table 3), along with corresponding symptoms of nausea, abdominal pain, low appetite, and fatigue. By that time, the facial and body swelling had also resolved. Potassium levels remained normal throughout treatment. After eight days off osilodrostat, laboratory tests showed the following: Noon ACTH 67 pg/mL (normal 6-59 pg/mL), noon serum cortisol 7.24 mcg/dL (normal 8-25 mcg/dL), and 24-hour urine free cortisol 26.2 mcg/d (normal <=60.0 mcg/d). Nearly 3 months off osilodrostat, the patient had an 11 AM ACTH of 68.9 pg/mL (normal 7.2-63.3 pg/mL) and 11AM serum cortisol level of 11.0 ug/dL (6.2-19.4 ug/dL). The clinical course is summarized in Table 3 and Figure 2. A DOTATATE-PET scan was discussed, though the patient wished to reconsider in the future given clinical response.

Table 3. Labs during treatment (Tx) with osilodrostat

Empty Cell 1 month before Tx Week 2 on Tx Week 3 on Tx Week 7 on Tx Week 9 on Tx – Tx stopped Week 1 off Tx Month 3 off Tx
Treatment with osilodrostat None On 2mg BID since Week 0 of Tx Advised to decrease to 1mg BID but patient did not decrease dose. Decreased to 1mg BID Decreased to 1mg daily after serum lab resulted. Then discontinued Tx after 24hr UFC resulted in several days. None None
ACTH level (pg/mL) 464 145 126 135 67 68.9
Cortisol level (mcg/dL) 91
8:32AM		 9.5
7:04AM		 5.4
7:11AM		 3.04
11:56AM		 4.9
11:26AM		 7.24
12:14PM		 11
11:08AM
24hr urine free cortisol (UFC) level (mcg/day) 12030.3 7 14 26.2
*Normal reference ranges depending on assays:
ACTH: 6-59 pg/mL or 7.2-63.3 pg/mL
Serum morning cortisol: 8-25 mcg/dL or 6.2-19.4 mcg/dL
24hr urine free cortisol: <=60.0 mcg/day or 5-64 mcg/day
*Acronyms: Tx = treatment; BID = twice daily; UFC = urine free cortisol, ACTH = adrenocorticotropic hormone

  1. Download: Download high-res image (232KB)
  2. Download: Download full-size image

Figure 2. Trends of 24hr urine cortisol levels and serum cortisol levels with osilodrostat treatment (Tx)

Discussion

Cyclic Cushing’s syndrome is a rare subtype of Cushing’s and occurs in both ACTH-dependent and ACTH-independent cases.3,7 Cyclicity has been attributed to hypothalamic dysfunction exaggerating a normal variant of hormonal cyclicity, a dysregulated positive feedback mechanism followed by negative feedback, intra-tumoral bleeding, and ACTH-secretion from neuroendocrine tumors (ex carcinoid tumors, pheochromocytomas).7,8,9,10
Potentially curative pituitary surgery or unilateral adrenalectomy are the treatments of choice.4 For example, cases of cyclic Cushing’s in primary pigmented nodular adrenocortical disease have demonstrated cure in some patients with unilateral adrenalectomy.11 In florid Cushing’s syndrome that is not amenable or responsive to other treatments, bilateral adrenalectomy could be lifesaving, though risks significant comorbidities including Nelson’s syndrome.4,12 Pituitary radiotherapy/radiosurgery are treatment options, though risks progressive anterior pituitary dysfunction.4 Medical therapy can play an important role as a bridge to surgery or radiation, with recurrence, for poor surgical candidates, or when there is no identifiable source as in our patient.13 Cyclic Cushing’s syndrome, moreover, has a higher recurrence rate (63%) and lower remission rate (25%), compared to classic Cushing’s syndrome.8
Medical treatments of cyclic Cushing’s syndrome include steroidogenesis inhibitors (ketoconazole, levoketoconazole, metyrapone, and osilodrostat), adrenolytic agents (mitotane), glucocorticoid receptor blockers (mifepristone), and pituitary tumor-directed agents (pasireotide, cabergoline, and temozolomide).8,14,15 Treatment goal is normalization of 24-hour urine cortisol levels and morning serum cortisol levels, though block-and-replace regimens occasionally are used.13,14 A block-and-replace regimen with osilodrostat and dexamethasone was used in the case of exogenous cyclic Cushing’s from pembrolizumab, given need for the immunotherapy;6 however, this regimen would hinder assessment of remission in native cyclic Cushing’s.
As our patient had cyclic Cushing’s disease, pituitary tumor-directed medications could be used for treatment. Pasireotide and cabergoline, however, are limited by a significant percentage of non-responders, along with risk of hyperglycemia for pasireotide.15 We considered mifepristone, which is a competitive antagonist at the glucocorticoid receptor and progesterone receptor; however, mifepristone is limited by the inability to directly monitor cortisol response on labs, in addition to the risk of AI and mineralocorticoid side effects with overtreatment.16
Steroidogenesis inhibitors block one or more enzymes in the production of cortisol, with potential risk of AI. The new steroidogenesis inhibitor osilodrostat, like metyrapone, selectively inhibits CYP11B1 and CYP11B2, which are involved in the final steps of cortisol and aldosterone synthesis, respectively.13,14 Ketoconazole and levoketoconazole, on the other hand, block most enzymes in the adrenal steroidogenesis pathway, including CYP11B1 and CYP11B2, and are limited by their inhibition of CYP7A (with associated hepatotoxicity) and strong inhibition of cytochrome p450 CYP3A4 (leading to many drug-drug interactions, decreased testosterone production, and QTc prolongation).14
Osilodrostat and metyrapone do not affect CYP7A and less potently inhibit CYP3A4.13 However, they can lead to increased deoxycorticosterone levels, with associated risks of hypokalemia, hypertension, and edema, and increased androgen production (with metyrapone thus being considered second-line in women).13,14,17
Osilodrostat, compared to metyrapone and ketoconazole, has a higher potency in CYP11B1 and CYP11B2 inhibition and a longer half-life, with stronger effects in lowering cortisol levels, allowance of less frequent (twice daily) dosing, and possibly less side effects.13,14,17,18 Compared to metyrapone, studies have suggested osilodrostat leads to a lesser rise in 11-deoxycortisol levels and less hyperandrogenic effects.13,14 Osilodrostat is also rapidly absorbed with sustained efficacy up to 6.7 years.17,18 Though rare cases of prolonged AI following discontinuation exist, osilodrostat (like other steroidogenesis inhibitors) is generally considered a reversible inhibitor.19 Reversible inhibition of cortisol synthesis is particularly appealing to treatment of cyclic Cushing’s syndrome as patients will not suffer from prolonged AI after episodes subside.
We thus considered osilodrostat an attractive treatment of cyclic Cushing’s syndrome. In our patient, osilodrostat was efficacious and well-tolerated, consistent with the literature,17 with clinical effects within 2-3 weeks without significant mineralocorticoid side effects. Differentiation of AI as a side effect of osilodrostat or from remission of the cyclical episode is crucial. Our patient was carefully tapered off osilodrostat after developing AI, and reversal of AI and osilodrostat inhibition were clearly demonstrated after 8 days off osilodrostat. Off treatment, the patient demonstrated neither prolonged AI nor clinical hypercortisolism, confirming remission of cyclic Cushing’s.

Conclusion

We present the first case to our knowledge demonstrating successful treatment of cyclic Cushing’s syndrome with osilodrostat. Osilodrostat showed rapid and safe control of hypercortisolism and importantly exhibited quick reversible inhibition of steroidogenesis upon discontinuation, a virtue in cyclic Cushing’s syndrome management.

References

Cited by (0)

The authors declare the following:
This paper did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
All authors do not have any conflicts of interests regarding the manuscript.
Run Yu, MD, PhD runyu@mednet.ucla.edu
Clinical Relevance
Osilodrostat is a new steroidogenesis inhibitor. Our case demonstrates the first successful treatment of native cyclic Cushing’s syndrome with osilodrostat, which showed rapid onset/offset, clinical safety, and reversible inhibition of steroidogenesis and medication-induced adrenal insufficiency. Osilodrostat’s preservation of underlying adrenal function is key when the cyclic Cushing’s episode spontaneously remits.

Filed under: Cushing's, pituitary, Rare Diseases, Treatments | Tagged: , , , , | Leave a comment »

Osilodrostat-associated Adrenal Gland Shrinkage: a Case Series of Patients with ACTH-Dependent Cushing’s Syndrome

Posted on October 18, 2025 by MaryO

The Journal of Clinical Endocrinology & Metabolism, dgaf552, https://doi.org/10.1210/clinem/dgaf552

Abstract

Context

Medical therapy for Cushing’s syndrome (CS) is increasingly used. A potent adrenal steroidogenesis inhibitor, osilodrostat, has been rarely linked to prolonged adrenal insufficiency (AI).

Objective

We hypothesized that osilodrostat-induced adrenal insufficiency could be associated with adrenal gland shrinkage.

Design

Non-interventional, retrospective, longitudinal, IRB-approved study of patients with CS treated at Oregon Health and Science University between January 1, 2000 and July 1, 2025.

Setting

Ambulatory and inpatient, academic, quaternary medical center.

Patients or Other Participants

Patients with ACTH-dependent CS, treated with osilodrostat for >3 months, and CT imaging before and after osilodrostat available for adrenal volume (AV) measurement.

Intervention(s)

Age, sex, osilodrostat doses and duration, laboratory data and AI were recorded. AV was calculated using manual segmentation on CT images by a board-certified radiologist.

Main Outcome Measure(s)

AV before and after initiation of osilodrostat was expressed as percent reduction.

Results

10 patients (5 ectopic CS, 4 unknown ACTH source, 1 Cushing’s disease) were included. Osilodrostat mean starting, maximum and final doses: 7.7, 13.8 and 5.9 mg/day, respectively, over 23 months. Four patients received block-and-replace regimen, AI developed in 5. Adrenal gland volume decreased by 46.7±22.2% from 25.5±9.9 ml to 12.7±6.4 ml, p<0.001 over a median of 19 months. AV reduction positively correlated with maximum osilodrostat dose, r=0.626, p=0.027.

Conclusions

We found that in selected patients with ACTH-dependent CS, osilodrostat can induce significant adrenal shrinkage, with or without AI. Further confirmation by larger studies of different CS types and monitoring for AI is required for all patients.

Filed under: adrenal, Cushing's, Treatments | Tagged: , , | Leave a comment »

A Preliminary Model to Tailor Osilodrostat In Patients With Adrenocorticotropic Hormone (ACTH)-Dependent Cushing’s syndrome

Posted on October 10, 2025 by MaryO

Abstract

Over the past 10 years, osilodrostat has become one of the most commonly used steroidogenesis inhibitors in patients with Cushing’s syndrome. The starting dose is usually determined based on the product characteristics, the prescriber’s experience, and cortisol levels. However, no study has attempted to determine whether there was a dose–response relationship between osilodrostat and cortisol reduction. In this study, we developed a preliminary kinetic–pharmacodynamic model to tailor osilodrostat in patients with Adrenocorticotropin hormone (ACTH)-dependent Cushing’s syndrome. We first analyzed the decrease in cortisol 48 hours after initiation or dose change of osilodrostat in 18 patients. Simulations were then performed for different doses of osilodrostat to evaluate the variation in cortisol concentrations. Our results report the first dose–response relationship between osilodrostat dose and cortisol levels, which should be helpful in identifying the optimal dosing regimen in patients with Cushing’s syndrome and in individualizing treatment to approximate a nychthemeral rhythm.

osilodrostat, Cushing’s syndrome, Cushing’s disease, cortisol, adrenal insufficiency
Issue Section:

Brief Report

Significance

The current preliminary study is a first step in trying to better understand the effect of osilodrostat on cortisol, which should help determine the optimal dose for each patient.

Introduction

Cushing’s syndrome is a rare condition in which increased cortisol levels lead to a wide range of comorbidities and increased mortality. Surgery is usually regarded as the first-line and most effective treatment.1 In some cases, cortisol-lowering drugs are necessary, mainly after failed surgery.2,3 Among several steroidogenesis inhibitors such as ketoconazole and metyrapone,4,5 osilodrostat, which acts through inhibition of 11β-hydroxylase, is now being considered an effective drug in controlling cortisol hypersecretion. Initially designed as a CYP11B2 inhibitor, the study by Ménard et al.6 involving both animal models and healthy human subjects showed that osilodrostat reduced cortisol levels from a dose of 1 mg/day, while lower doses exerted an anti-aldosterone effect. Since then, several clinical trials and retrospective studies emphasized its efficacy in all etiologies of Cushing’s syndrome.7-9 While the usual recommended starting dose is 2 mg twice a day, precise studies on the short-term effect of osilodrostat on plasma cortisol are lacking. These data could, however, be of interest to tailor the treatment. Moreover, baseline urinary free cortisol (UFC) level is not able to predict response to osilodrostat.10 Taking advantage of serial cortisol measurements performed in inpatient clinics in our center at the time osilodrostat became available, we developed a pharmacokinetic (PK)/pharmacodynamic model of plasma cortisol variation as a function of osilodrostat dose in patients with Adrenocorticotropin-hormone (ACTH)-dependent Cushing’s syndrome.

Patients and methods

Clinical data and hormonal measurements

We retrospectively included patients with ACTH-dependent Cushing’s syndrome, who had serial measurements of plasma cortisol (every 4 hours for 24 hours) before and after the first osilodrostat dose between 2019 and 2024. These measurements were part of our standard of care approach when osilodrostat became available in our tertiary expert center as a thorough evaluation of the efficacy and tolerance of a new drug. The initial dose ranged from 2 to 15 mg/day, depending on the severity of hypercortisolism. Subsequently, osilodrostat dose was gradually adjusted based on the successive cortisol measurements described above. Sex, age at diagnosis, and etiologies were recorded, as well as plasma cortisol measurements 48 hours after the initiation or any change in the osilodrostat dose and time elapsed since change of dose and last administration were recorded. All plasma cortisol measurements were performed with the same Elecsys II Cortisol, Cobas (Roche Diagnostics) assay in the hormonal laboratory of our center; cross-reactivity with 11-deoxycortisol is 4.9%. According to our institutional policy, this retrospective study did not require specific signed informed consent from patients as the data collected were anonymized. It was thus approved by the Ethics Committee of Assistance Publique—Hopitaux de Marseille (RGPD PADS reference RUXXX2). The current study complies with the Declaration of Helsinki.

Pharmacokinetics and statistical analysis

The pharmacodynamic parameters of osilodrostat on cortisol concentrations were analyzed using a kinetic–pharmacodynamic (PD) model in the software Nonlinear Mixed Effects Modeling version 7.4 (NONMEM Icon Development Solutions, Ellicott City, MD, United States). PK analysis from a previously published study6 was used to predict plasma concentration in our patients. The PK parameters were described in the article, and mean concentration values were obtained by digitizing the graph of osilodrostat vs time using the software WebPlotDigitizer version 4.2.11 With these data, a one-compartment population PK model was used to predict osilodrostat concentrations for different dosing regimens. Direct and indirect relationship between osilodrostat-predicted concentration and variation of cortisol concentrations were evaluated to consider a delay. The variation of cortisol concentrations was calculated with reference to a session without treatment. Several functions were tested to describe the relationship such as linear and sigmoidal. Model selection and evaluation were done by the likelihood ratio test (objective function), goodness-of-fit plots (observed vs predicted variation of cortisol concentrations, observed vs individual predictions, normalized prediction distribution errors vs time and variation of cortisol predictions), bootstrap, and visual predictive checks. Graphical analysis was performed with the R software version 4.4.012 using the ggplot2 package.13 Simulations were performed for different doses of osilodrostat to evaluate the variation on cortisol concentrations using the package rxode2.14

Results

Of the patients who were prescribed osilodrostat at least once between 2019 and 2024, 18 were presenting ACTH-dependent Cushing’s syndrome, 12 women (66.6%) and 6 men (33.3%). Mean age was 53.2 ± 15 years. The cause of Cushing’s syndrome was Cushing’s disease in 16 patients (88.9%), ectopic ACTH secretion in 1 patient (5.6%), and ACTH-dependent hypercortisolism of uncertain diagnosis in 1 patient (5.6%). Clinical characteristics are presented in Table 1. It should be noted that none of the patients included were Asian.

 

 

Table 1.

Open in new tab

Clinical characteristics of patients with all included patients and differentiated according to gender.

All patientsa Women Men
Age at diagnosis 53.2 ± 15 54 ± 17.2 51.5 ± 10.5
Weight 81.7 ± 13.7 79.5 ± 12.7 86.2 ± 15.6
% of CD 88.9 83.3 100
ULN of 24 hour UFC 4.4 ± 8.3 5.5 ± 10.3 2.5 ± 1.8
Osilodrostat starting dose 3.3 ± 2.2 3.7 ± 2.4 2.5 ± 1.4
Cortisol before osilodrostat intake 422.9 ± 159.2 414.7 ± 176.6 439.4 ± 130.7
Cortisol 4 hour after osilodrostat 404 ± 165.6 408.2 ± 200.1 395.5 ± 70.8

 

Abbreviations: CD, Cushing’s disease; ULN, upper limit range; UFC, urinary free cortisol.

aOf note, none of the included patients were Asian.

In their article, Ménard et al.6 showed that the dose–exposure relationship was not strictly proportional. A one-compartment model was enhanced by increasing the relative bioavailability with the dose and was estimated that the dose resulting in a 50% increase in bioavailability was 1.06 mg. The PK parameters derived from Ménard et al.6 were fixed and used to predict osilodrostat concentration in our patients. A direct relationship between the predicted osilodrostat concentrations and variation of cortisol concentrations (%) gave a better fit than an indirect model. The drug effect was modeled with the following sigmoidal function (Eq. 1);

(1)

where Imax is the maximal inhibition and IC50 is the apparent half-maximal inhibitory concentration.

The estimated PD parameters were IC50 and Imax. Their values as well as the relative standard errors (RSE%) and the corresponding bootstrap IC50 are shown in Table 2. Final parameters were used to simulate n = 500 profiles following a single dose of osilodrostat.

 

 

 

Table 2.

Open in new tab

Pharmacodynamic parameters of osilodrostat’s effects on the variation of cortisol concentrations.

Parameters Unit Estimation RSE% Bootstrap
0.025 0.975
KA (fixed)a 1/hour 4.03 — — —
CL/F (fixed)a L/hour 18.3 — — —
V/F (fixed)a L 125 — —
Imax % 44.5 18.7 12.51 90.9
IC50 mg/L 0.011 37.4 0.0001 0.10
Interindividual variability (ω)
 Imax 0.40 30.9 0.003 1.86
 IC50 3.78 41.0 0.003 9.22
Residual unexplained variability (σ)
 Additive % 23.8 12.2 18.2 29.9

 

Abbreviations: CL/F, apparent clearance; IC50, osilodrostat concentration associated with half the maximal inhibition of the cortisol variation; Imax, maximum inhibitory effect of osilodrostat on the variation of cortisol; KA, first-order absorption rate constant; RSE, relative standard error; V/F, apparent volume of distribution.

 

aAdapted from Ménard et al.6

The effects on plasma cortisol variation are depicted in Figure 1. Cortisol concentration declines during the first hour after taking osilodrostat, from 24% for a 1 mg dose to over 42% for a 20 mg dose. Thereafter, from the first hour onward, cortisol increases progressively, with loss of treatment efficacy occurring around the 10th-15th hour for 1 and 2 mg, while for doses above 5 mg, a moderate effect persists over the following hours. Figure 2 shows the variation in cortisol concentration for a 2 mg dose, with median decrease in cortisol variation of 31%, ranging from 0% to 67.5%, with, as mentioned above, a maximum effect 1 hour after osilodrostat intake, and a progressive increase in cortisol levels, mainly during the 12 hours following treatment. The same analysis for 10 mg revealed a median reduction in cortisol of 38%, ranging from 5% to 80%. Figure 3 describes the relationship between osilodrostat concentration and cortisol variation, showing that the maximum effect corresponds to the maximum concentration and that a decrease in osilodrostat concentration results in an increase in cortisol level.

Relationship between time since last administration of osilodrostat and cortisol concentrations.

Figure 1.

Relationship between time since last administration of osilodrostat and cortisol concentrations.

Open in new tabDownload slide

Visual predictive variation on cortisol concentrations following 2 or 10 mg osilodrostat administration.

Figure 2.

Visual predictive variation on cortisol concentrations following 2 or 10 mg osilodrostat administration.

Open in new tabDownload slide

Relation between osilodrostat concentration and cortisol variation.

Figure 3.

Relation between osilodrostat concentration and cortisol variation.

Open in new tabDownload slide

Discussion

To the best of our knowledge, this is the first study that attempts to define a dose/efficacy relationship between osilodrostat dose and the variation of plasma cortisol. First, our results suggest that the effect of osilodrostat appears immediately after the peak of concentration, 1 hour after treatment intake, which highlights the parallel evolution of osilodrostat and cortisol concentrations. This is unusual, as typically effect peak takes few hours, following concentration peak.15 The relationship between osilodrostat concentration and the effect on cortisol is not linear but sigmoidal with a rapid increase in concentrations producing a rapid significant effect, leading to a maximal effect. Because elimination is a slower process than absorption, the effect’s decline will also be slower: this means that efficiency remains stable during the first 5 hours, with a further progressive increase of cortisol and a loss of efficiency around 10-15 hours after intake. This confirms the need for two intakes per day, with one early in the morning and the other 12 hours later in the evening. In addition, even if our simulation suggests a wide interindividual variability, we were able to determine the impact of different doses of osilodrostat on the percent decrease in plasma cortisol levels. For instance, 20 mg osilodrostat leads to an estimated 42% decrease in cortisol concentration. Interestingly, Ferrari et al.16 recently showed that patients controlled with two doses of osilodrostat for at least 1 month had the same efficacy with a single intake (combing both doses) at 4 or 7 Pm. This is quite surprising and will need to be evaluated in future studies: our preliminary model could give more precise information on this point.

Cushing’s syndrome is also characterized by a loss of circadian rhythm leading to increased comorbidities such as diabetes, hypertension, and cardiovascular disease.17,18 This is why 24 hour UFC can only be considered an imperfect marker of glucocorticoid overexposure even though it is an easy-to-use marker, as exemplified by its use in all the clinical trials performed on cortisol-lowering drugs.7,8,10,19 Predicting the efficacy of osilodrostat on plasma cortisol might be helpful to tailor the treatment as a titrating approach. Of note, some studies suggested that there might be an inpatient variability of cortisol secretion in Cushing’s syndrome,20 and this might account for a bias in our results. However, none of our patients had cyclical Cushing’s syndrome. Moreover, 12 patients in our cohort had at least two cortisol cycles (every 4 hours during the day) before starting treatment. A comparison of these two cycles using Student’s t-test showed no significant difference (P = .7), indicating no obvious spontaneous variability. Our preliminary report gives interesting insights into the maximal efficacy expected for a single dose of osilodrostat, thus defining the initial dosage needed to rapidly control hypercortisolism, as opposed to the dose currently recommended by the manufacturer (2 mg twice daily). Thus, our results could help define an optimal dose in the morning, but also in the evening, with the aim of re-establishing a circadian profile. This will, however, have to be confirmed on an interventional study focusing on comorbidities, quality of life and their potential improvements while using this PK model.

The main limitation of this proof-of-concept study is the large CI. This may be due to the relatively low number of patients and the fact that cortisol was measured every 4 hours instead of every hour, but also to the large variability in efficacy between subjects. Due to the number of patients included in the analysis, it was not possible to investigate further if a covariate, such as the gender, may explain these differences between individuals. It is important to highlight that although our model predicts cortisol levels 1 hour post intake as the most reliable predictor of future efficacy, cortisol measurements were taken every 4 hours. Thus, this finding should be confirmed in prospective studies with more frequent cortisol measurements, particularly 1 hour after osilodrostat administration. While the kinetic–pharmacodynamic approach used in this study can present with some inherent limitations, this type of approach is regularly used to define the modalities of use for a medication in a new indication. A nonlinear mixed-effects modeling allows the use of data from the routine clinical follow-up of patients. This method is thus effective and particularly well-suited for sparse data. Finally, a larger study could include closer measurements of cortisol. Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) is the best method for avoiding cross-reactivity with steroid precursors and could be used for these measurements. However, we used the Elecsys Cortisol II Immunoassay, which shows <5% cross-reactivity with 11-deoxycortisol; thus, our results are credible.

In conclusion, we designed a kinetic–pharmacodynamic model to adapt osilodrostat in patients with ACTH-dependent Cushing’s syndrome. Our model shows that cortisol level 1 hour after treatment is the best indicator of future efficacy. Moreover, depending on the initial cortisol level and the goal to be achieved, different doses should be prescribed. Despite wide inter-patient variability, we believe our model provides insight into the minimal dose necessary to decrease cortisol levels and the maximal efficacy expected for a given dose. Thus, it should help physicians tailor the treatment to reach maximal efficacy in the shortest possible time. The next step will be to analyze whether this percent decrease remains stable on a long-term basis or becomes more important with time, as suggested by some clinical cases showing delayed adrenal insufficiency on stable doses of osilodrostat.21

Authors’ contributions

Cecilia Piazzola (Conceptualization [equal], Formal analysis [equal], Writing—original draft [equal]), Frederic Castinetti (Conceptualization [equal], Formal analysis [equal], Writing—review & editing [equal]), Katharina von Fabeck (Conceptualization [equal], Writing—review & editing [equal]), and Nicolas Simon (Conceptualization [equal], Methodology [equal], Supervision [equal], Validation [equal], Writing—original draft [equal], Writing—review & editing [equal])

Funding

This work received an unrestricted educational grant from Recordati Rare Diseases.

To see the references and the original article, please go here: https://academic.oup.com/ejendo/article/193/4/K11/8255719?login=false

 

Filed under: adrenal crisis, Clinical trials, Cushing's, pituitary, Treatments | Tagged: , , , | Leave a comment »

« Previous Page — Next Page »