Differential diagnosis of Cushing's syndrome

Machine Learning Models for Classification of Cushing's Syndrome Using Retrospective Data

Accurate classification of Cushing's Syndrome (CS) plays a critical role in providing the early and correct diagnosis of CS that may facilitate treatment and improve patient outcomes. Diagnosis of CS is a complex process, which requires careful and concurrent interpretation of signs and symptoms, multiple biochemical test results, and findings of medical imaging by physicians with a high degree of specialty and knowledge to make correct judgments. In this article, we explore the state of the art machine learning algorithms to demonstrate their potential as a clinical decision support system to analyze and classify CS to facilitate the diagnosis, prognosis, and treatment of CS. Prominent algorithms are compared using nested cross-validation and various class comparison strategies including multiclass, one vs. all, and one vs. one binary classification. Our findings show that Random Forest (RF) algorithm is most suitable for the classification of CS. We demonstrate that the proposed approach can classify CS with an average accuracy of 92% and an average F1 score of 91.5%, depending on the class comparison strategy and selected features. RF-based one vs. all binary classification model achieves sensitivity of 97.6%, precision of 91.1%, and specificity of 87.1% to discriminate CS from non-CS on the test dataset. RF-based multiclass classification model achieves average per class sensitivity of 91.8%, average per class specificity of 97.1%, and average per class precision of 92.1% to classify different subtypes of CS on the test dataset. Clinical performance evaluation suggests that the developed models can help improve physicians' judgment in diagnosing CS.

Increased prevalence of obstructive sleep apnea in patients with Cushing's syndrome compared with weight- and age-matched controls

OBJECTIVE

Diabetes mellitus and obesity are well-known risk factors associated with obstructive sleep apnea (OSA). Cushing's syndrome (CS) is also characterized by obesity and diabetes mellitus. However, the association between CS and OSA remains unclear. Therefore, we investigated the possible associations between CS and OSA in this study.

PATIENTS AND METHODS

Thirty female patients with newly diagnosed active CS and 30 age-, gender- and body mass index (BMI)-matched controls were included in this study. All participants were evaluated by overnight polysomnography. OSA was defined as having an apnea-hypopnea index (AHI) score of ≥5 events/h. Insulin resistance was calculated by homeostasis model assessment (HOMA) scores. Fasting serum cortisol was also determined.

RESULTS

The prevalence of OSA was higher (50% vs 23%, P = 0.003) in patients with CS compared with the control subjects. The mean HOMA (P = 0.046) and AHI (P = 0.028) scores were higher in patients with CS compared with the control subjects. AHI was positively correlated with the HOMA scores (r = 0.281, P = 0.046) in both groups. Linear regression analysis showed that serum cortisol remained as an independent predictor for AHI after controlling for BMI and HOMA score (P < 0.001).

CONCLUSIONS

The prevalence of OSA increased in patients with CS compared with control subjects with similar ages and BMI levels. Hypercortisolemia is an independent risk factor for developing OSA. The presence of OSA needs to be considered in patients with CS.

Cushing's syndrome is associated with sleep alterations detected by wrist actigraphy

BACKGROUND

The association between the hypothalamic-pituitary-adrenal (HPA) axis and sleep is well described. It is also known that HPA axis disturbances have an effect on sleep. In fact, patients affected by Cushing's syndrome (CS) often complain about poor sleep quality. Our aim was to evaluate objective sleep quality and duration in patients with Cushing's syndrome in active phase, using wrist actigraphy.

PATIENTS AND METHODS

In 12 patients with active CS without ongoing specific therapy (11 F, 1 M; age 40.0 ± 10.9 years; BMI 28.4 ± 6.7 kg/m(2)) and 12 healthy control subjects (HS) (11 F, 1 M; age 44.0 ± 11.0 years; BMI 23.9 ± 4.2 kg/m(2)) an actigraphic evaluation was performed on 3 consecutive days under free living conditions. Objective measurement of sleep duration and quality was estimated by an actiwatch, which is a wristwatch-like device used to detect motor activity.

RESULTS

In CS patients, wrist actigraphy showed higher fragmented sleep (fragmentation index CS 16.2 ± 4.2, HS 13.0 ± 3.6; p = 0.034) and increased nocturnal motor activity (total activity score CS 8318 ± 4308, HS 4971 ± 2372; p = 0.020; mean activity score CS 8.7 ± 4.2, HS 5.4 ± 2.2; p = 0.030; mean score in active time CS 104.8 ± 39.2, HS 74.8 ± 23.1; p = 0.030). On the contrary, actual sleep time resulted similar in CS and HS. No correlation was found between sleep alterations and urinary free cortisol in patients.

CONCLUSIONS

The impaired actigraphic parameters described in our study suggest that hypercortisolism is associated with sleep alterations, which could contribute to the worsening of life quality and metabolic comorbidities associated with CS. These results have to be confirmed in a larger cohort of patients, using more accurate instruments for sleep assessment.

Incidence of Cushing's syndrome and Cushing's disease in commercially-insured patients <65 years old in the United States

PURPOSE

To estimate the incidence of Cushing's syndrome (CS) and Cushing's disease (CD) in the US.

METHODS

MarketScan Commercial database 2007-2010 (age <65 years) was used. CS patients were defined with ≥2 claims of CS diagnosis, while CD patients were defined with CS plus a benign pituitary adenoma diagnosis or hypophysectomy in the same calendar year. Annual incidence was calculated by dividing the number of CS or CD cases by the total number of members with the same enrollment requirement during the calendar years.

RESULTS

CS incidence rates per million person-years were 48.6 in 2009 and 39.5 in 2010. The lowest rates of CS were in ≤17-year-olds and highest rates were in 35 to 44-year-olds. CD incidence rates were 7.6 in 2009 and 6.2 in 2010. The lowest rates of CD were in ≤17-year-olds and highest rates were in 18 to 24-year-olds. The rates varied by sex (2.3-2.7 in males, 9.8-12.1 in females). In females, lowest rates ranged 2.5-4.0 in ≤17-year-olds and highest 16.7-27.2 in 18-24 year olds. In males, there were too few cases to report estimates by age.

CONCLUSIONS

In the first large US-based study, the annual incidence of CS in individuals <65 years old was nearly 49 cases per million, substantially higher than previous estimates, which were based primarily on European data. Using similar methods, we estimated the incidence of CD at nearly 8 cases per million US population. These estimates, if confirmed in other epidemiologic databases, represent a new data reference in these rare conditions.

Managing Cushing's disease: the state of the art

Cushing's disease is a rare chronic disease caused by a pituitary adenoma, which leads to excess secretion of adrenocorticotropic hormone (ACTH). The over-production of ACTH leads to hyperstimulation of the adrenal glands and a chronic excess of cortisol, resulting in the signs and symptoms of a severe clinical state (Cushing's syndrome) that leads to significant morbidity, negative impacts on the patient's quality of life, and, if untreated, increased mortality. The management of patients with Cushing's disease is complicated by the heterogeneity of the condition, with signs and symptoms that overlap with those of other diseases, and high subclinical incidence rates. Controversies surrounding the tests used for screening and identifying patients with Cushing's disease add to the challenge of patient management. Surgical intervention to remove the adenoma is the first-line treatment for patients with Cushing's disease, but medical therapies are useful in patients who relapse or are unsuitable for surgery. The recent introduction of pasireotide, the first pituitary-directed medical therapy, expands the number of treatment options available for patients with Cushing's disease. This state-of-the-art review aims to provide an overview of the most recent scientific research and clinical information regarding Cushing's disease. Continuing research into improving the diagnosis and treatment of Cushing's disease will help to optimize patient management.

Etiology and clinical profile of patients with Cushing's syndrome: A single center experience

BACKGROUND

There is little published literature on the profile of patients with Cushing's syndrome (CS) from India. The aim of this study was to compile data of CS patients treated at this hospital.

MATERIALS AND METHODS

Patients referred to the endocrine services of this hospital for diagnosis/treatment of CS from January 1985 to July 2012 were the subjects for this study. All patients had detailed medical history, physical examination and biochemical and hormonal assays (which changed with availability of tests and changing views). Assays for plasma adrenocorticotropic hormone (ACTH) (late 90s), salivary cortisol estimation, IJV sampling for ACTH and corticotrophin releasing hormone stimulation tests were added on later. Imaging included computed tomography (CT), magnetic resonance imaging (since the late 80's) and (68)Ga DOTA-TOC/FDG PET-CT (2008).

RESULTS

Three hundred sixty-four patients (250 females, 114 males, age 6 months to 65 years, mean 28 years + 12 years) were diagnosed to have CS during this period. Two hundred and ninety-three patients (80.5%) were ACTH dependent (CD 215, ectopic ACTH syndrome 22, occult ACTH source 56) while 71 (19.5%) were ACTH independent (adrenal carcinoma 36, adenoma 30, primary pigmented nodular adrenal disease 4, AIMAH 1). Pituitary macro adenoma was seen in 14% of the CD cases. The most common presenting complaints were hypertension and diabetes mellitus. A total of 63% patients complained of weight gain while 15% had lost weight. Myopathy, infections, skeletal fractures and psychiatric problems were the other common observations in our patients.

CONCLUSION

The clinical spectrum was broad. CD was the most common cause for CS.

The investigation of Cushing syndrome: essentials in optimizing appropriate diagnosis and management

The investigation of Cushing syndrome (CS) should start with careful history taking and clinical examination, and exogenous steroid usage must be excluded. It is essential to confirm hypercortisolism before further investigations are undertaken. The recommended first-line tests include midnight salivary cortisol and/or the 1 mg overnight or low-dose dexamethasone suppression tests. The next step is to differentiate adrenocorticotrophin (ACTH)-dependent from ACTH-independent CS by measuring ACTH. With ACTH-dependence, further investigations should differentiate pituitary-dependent from ectopic ACTH-dependent CS. Many dynamic tests may be considered, but we suggest that bilateral inferior petrosal sinus sampling should be performed in almost all patients with ACTH-dependent CS, except for patients with a pituitary macroadenoma. Imaging should include MR scanning of the pituitary, and CT scanning of the chest and abdomen to look for an ectopic source. Confirmation of the diagnosis of CS and accurate localization of its source are vital to optimize therapy for this complex disorder.

Diagnosis/differential diagnosis of Cushing's syndrome: a review of best practice

Endogenous Cushing's syndrome is an uncommon endocrine disorder that can often prove challenging, for both initial diagnosis and subsequent differential diagnosis of the underlying cause. Understanding the advantages and pitfalls of the diagnostic methods used in Cushing's syndrome is essential for accurate diagnosis and differential diagnosis, and guidelines outlining the most appropriate approaches have recently been published. We examine how current practice guidelines can be used in a real case-based scenario, and review the recommended strategies to achieve successful diagnosis and differential diagnosis.

Cortisol assays and diagnostic laboratory procedures in human biological fluids

The overview of cortisol physiology, action and pathology is achieved in relation to the hypothalamic-pituitary-adrenal axis alteration by laboratory investigation. The measurements of cortisol and related compound levels in blood, urine and saliva used to study the physiological and pathological cortisol involvement, are critically reviewed. The immunoassay and chromatographic methods for cortisol measurement in the various biological fluids are examined in relation to their analytical performances, reference ranges and diagnostic specificity and sensitivity. Moreover, blood, urine and saliva cortisol level measurements are described taking into account the diagnostic implications. The deduction is that each method requires the definition of its own reference range and its related diagnostic cut-off levels. Thus, this review, stressing the analysis procedures, could help to understand and compare the results of the different assays.