Automatic face classification of Cushing's syndrome in women - a novel screening approach

The state of artificial intelligence for systemic dermatoses: Background and applications for psoriasis, systemic sclerosis, and much more

Artificial intelligence (AI) has been steadily integrated into dermatology, with AI platforms already attempting to identify skin cancers and diagnose benign versus malignant lesions. Although not as widely known, AI programs have also been utilized as diagnostic and prognostic tools for dermatologic conditions with systemic or extracutaneous involvement, especially for diseases with autoimmune etiologies. We have provided a primer on commonly used AI platforms and the practical applicability of these algorithms in dealing with psoriasis, systemic sclerosis, and dermatomyositis as a microcosm for future directions in the field. With a rapidly changing landscape in dermatology and medicine as a whole, AI could be a versatile tool to support clinicians and enhance access to care.

Artificial intelligence facial recognition system for diagnosis of endocrine and metabolic syndromes based on a facial image database

AIM

To build a facial image database and to explore the diagnostic efficacy and influencing factors of the artificial intelligence-based facial recognition (AI-FR) system for multiple endocrine and metabolic syndromes.

METHODS

Individuals with multiple endocrine and metabolic syndromes and healthy controls were included from public literature and databases. In this facial image database, facial images and clinical data were collected for each participant and dFRI (disease facial recognition intensity) was calculated to quantify facial complexity of each syndrome. AI-FR diagnosis models were trained for each disease using three algorithms: support vector machine (SVM), principal component analysis k-nearest neighbor (PCA-KNN), and adaptive boosting (AdaBoost). Diagnostic performance was evaluated. Optimal efficacy was achieved as the best index among the three models. Effect factors of AI-FR diagnosis were explored with regression analysis.

RESULTS

462 cases of 10 endocrine and metabolic syndromes and 2310 controls were included into the facial image database. The AI-FR diagnostic models showed diagnostic accuracies of 0.827-0.920 with SVM, 0.766-0.890 with PCA-KNN, and 0.818-0.935 with AdaBoost. Higher dFRI was associated with higher optimal area under the curve (AUC) (P = 0.035). No significant correlation was observed between the sample size of the training set and diagnostic performance.

CONCLUSIONS

A multi-ethnic, multi-regional, and multi-disease facial database for 10 endocrine and metabolic syndromes was built. AI-FR models displayed ideal diagnostic performance. dFRI proved associated with the diagnostic performance, suggesting inherent facial features might contribute to the performance of AI-FR models.

Artificial intelligence in paediatric endocrinology: conflict or cooperation

Artificial intelligence (AI) in medicine is transforming healthcare by automating system tasks, assisting in diagnostics, predicting patient outcomes and personalising patient care, founded on the ability to analyse vast datasets. In paediatric endocrinology, AI has been developed for diabetes, for insulin dose adjustment, detection of hypoglycaemia and retinopathy screening; bone age assessment and thyroid nodule screening; the identification of growth disorders; the diagnosis of precocious puberty; and the use of facial recognition algorithms in conditions such as Cushing syndrome, acromegaly, congenital adrenal hyperplasia and Turner syndrome. AI can also predict those most at risk from childhood obesity by stratifying future interventions to modify lifestyle. AI will facilitate personalised healthcare by integrating data from 'omics' analysis, lifestyle tracking, medical history, laboratory and imaging, therapy response and treatment adherence from multiple sources. As data acquisition and processing becomes fundamental, data privacy and protecting children's health data is crucial. Minimising algorithmic bias generated by AI analysis for rare conditions seen in paediatric endocrinology is an important determinant of AI validity in clinical practice. AI cannot create the patient-doctor relationship or assess the wider holistic determinants of care. Children have individual needs and vulnerabilities and are considered in the context of family relationships and dynamics. Importantly, whilst AI provides value through augmenting efficiency and accuracy, it must not be used to replace clinical skills.

da Silva C, de Castro Michelassi G, Brentani H, Nunes FL, Machado-Lima A. Computer aided diagnosis of neurodevelopmental disorders and genetic syndromes based on facial images - A systematic literature review

Neurodevelopment disorders can result in facial dysmorphisms. Therefore, the analysis of facial images using image processing and machine learning techniques can help construct systems for diagnosing genetic syndromes and neurodevelopmental disorders. The systems offer faster and cost-effective alternatives for genotyping tests, particularly when dealing with large-scale applications. However, there are still challenges to overcome to ensure the accuracy and reliability of computer-aided diagnosis systems. This article presents a systematic review of such initiatives, including 55 articles. The main aspects used to develop these diagnostic systems were discussed, namely datasets - availability, type of image, size, ethnicities and syndromes - types of facial features, techniques used for normalization, dimensionality reduction and classification, deep learning, as well as a discussion related to the main gaps, challenges and opportunities.

White blood cell count: a valuable tool for suspecting Cushing's syndrome

PURPOSE

Simple screening tests to determine whether Cushing's syndrome (CS) should be ruled out are lacking. Tools that enable early diagnosis could reduce morbidity and associated sequelae. The potential of glucocorticoid-induced changes in the white blood cell (WBC) count for raising suspicion of CS is assessed.

METHODS

This was a retrospective case‒control study. The WBC counts of 73 cases with CS and 146 matched controls were compared. The number of leukocytes (Leu), the number and percentage of neutrophils (N, Np), the number and percentage of lymphocytes (L, Lp), neutrophil-to-lymphocyte differences in the number and percentage (N-L, Np-Lp), neutrophil-to-lymphocyte ratio in the number and percentage (NLR, NLRp), and leukocyte-to-lymphocyte differences (Leu-L) were evaluated. The area under the ROC curve (AUC) was calculated for each of these parameters. Reference values were estimated that could help disclose occult CS.

RESULTS

All ten parameters showed significant differences between cases and controls. The AUC was greater than 0.7 for all ten parameters, and was the best for the NLRp and Lp (AUC: 0.89). An Lp of 23.9% showed a diagnostic accuracy of 84.9% for the diagnosis of CS. The concordance of an Lp below 24% and more than 8000 leucocytes had a PPV of 78.2% for CS, while the pairing of an Lp over 24% and a Leu below 8000 cells had an NPV of 97.3% for CS.

CONCLUSION

WBC count assessment can be an effective tool to raise suspicion of CS, prompting diagnostic testing. This simple and universally available test may allow earlier diagnosis of CS before highly evolved phenotypes develop.

Review on Facial-Recognition-Based Applications in Disease Diagnosis

Diseases not only manifest as internal structural and functional abnormalities, but also have facial characteristics and appearance deformities. Specific facial phenotypes are potential diagnostic markers, especially for endocrine and metabolic syndromes, genetic disorders, facial neuromuscular diseases, etc. The technology of facial recognition (FR) has been developed for more than a half century, but research in automated identification applied in clinical medicine has exploded only in the last decade. Artificial-intelligence-based FR has been found to have superior performance in diagnosis of diseases. This interdisciplinary field is promising for the optimization of the screening and diagnosis process and assisting in clinical evaluation and decision-making. However, only a few instances have been translated to practical use, and there is need of an overview for integration and future perspectives. This review mainly focuses on the leading edge of technology and applications in varieties of disease, and discusses implications for further exploration.

Machine intelligence in non-invasive endocrine cancer diagnostics

Artificial intelligence (AI) has illuminated a clear path towards an evolving health-care system replete with enhanced precision and computing capabilities. Medical imaging analysis can be strengthened by machine learning as the multidimensional data generated by imaging naturally lends itself to hierarchical classification. In this Review, we describe the role of machine intelligence in image-based endocrine cancer diagnostics. We first provide a brief overview of AI and consider its intuitive incorporation into the clinical workflow. We then discuss how AI can be applied for the characterization of adrenal, pancreatic, pituitary and thyroid masses in order to support clinicians in their diagnostic interpretations. This Review also puts forth a number of key evaluation criteria for machine learning in medicine that physicians can use in their appraisals of these algorithms. We identify mitigation strategies to address ongoing challenges around data availability and model interpretability in the context of endocrine cancer diagnosis. Finally, we delve into frontiers in systems integration for AI, discussing automated pipelines and evolving computing platforms that leverage distributed, decentralized and quantum techniques.

Diurnal Salivary Cortisol Profiles in Patients with Cushing's Syndrome

BACKGROUND

Quantification of salivary cortisol is one of the highly sensitive and specific screening parameters for Cushing's syndrome (CS). However, only late-night salivary cortisol is part of the standard screening procedure. In this study, we aimed to analyze salivary cortisol day profiles in patients with different types of CS to test whether specific patterns might be relevant for diagnosis and subtyping.

MATERIAL AND METHODS

Among 428 patients including those with confirmed Cushing's syndrome (N=111, of those 75 with Cushing's disease, 27 patients with adrenal CS and nine patients with ectopic CS), autonomous cortisol secretion (N=39) or exclusion of CS (control group, N=278) salivary cortisol was measured five times a day.

RESULTS

At each of the five time points, salivary cortisol was significantly higher in patients with CS compared to the control group (p≤0.001). Using the entire profile instead of one single salivary cortisol at 11 p.m. improved diagnostic accuracy (85 vs. 91%) slightly. Patients with ACTH-dependent CS had higher salivary cortisol levels than patients with adrenal CS. Also, morning cortisol was significantly higher in patients with ectopic CS than in patients with Cushing's disease (p=0.04). Nevertheless, there was a strong overlap between diurnal profiles, and the diagnostic yield for subtyping was low.

DISCUSSION

The study results show that using diurnal salivary cortisol profiles for CS diagnosis results in a limited increase in diagnostic accuracy. With significant differences between Cushing subtypes, cortisol profiles are not useful in everyday clinical practice for subtyping of CS.

The promising role of risk scoring system for Cushing syndrome: Time to reconsider current screening recommendations

Despite the current screening approach for Cushing syndrome (CS), delayed diagnosis is common due to broad spectrum of presentation, poor discriminant symptoms featured in diabetes and obesity, and low clinical index of suspicion. Even if initial tests are recommended to screen CS, divergent results are not infrequent. As global prevalence of type 2 diabetes and obesity increases, CS may not be frequent enough to back routine screening to avoid false-positive results. This represents a greater challenge in countries with limited health resources. The development of indexes incorporates clinical features and biochemical data that are largely used to provide a tool to predict the presence of disease. In clinical endocrinology, indexes have been used in Graves' ophthalmology, hirsutism, and hypothyroidism. The use of clinical risk scoring system may assist clinicians in discriminating CS in the context of at-risk populations and, thus, may provide a potential intervention to decrease time to diagnosis. Development and validation of clinical model to estimate pre-test probability of CS in different geographic source population may help to establish regional prediction model for CS. Here, we review on the latest progress in clinical risk scoring system for CS and attempt to raise awareness for the use, validation, and/or development of clinical risk scores in CS.

Facial Recognition Intensity in Disease Diagnosis Using Automatic Facial Recognition

Artificial intelligence (AI) technology is widely applied in different medical fields, including the diagnosis of various diseases on the basis of facial phenotypes, but there is no evaluation or quantitative synthesis regarding the performance of artificial intelligence. Here, for the first time, we summarized and quantitatively analyzed studies on the diagnosis of heterogeneous diseases on the basis on facial features. In pooled data from 20 systematically identified studies involving 7 single diseases and 12,557 subjects, quantitative random-effects models revealed a pooled sensitivity of 89% (95% CI 82% to 93%) and a pooled specificity of 92% (95% CI 87% to 95%). A new index, the facial recognition intensity (FRI), was established to describe the complexity of the association of diseases with facial phenotypes. Meta-regression revealed the important contribution of FRI to heterogeneous diagnostic accuracy (p = 0.021), and a similar result was found in subgroup analyses (p = 0.003). An appropriate increase in the training size and the use of deep learning models helped to improve the diagnostic accuracy for diseases with low FRI, although no statistically significant association was found between accuracy and photographic resolution, training size, AI architecture, and number of diseases. In addition, a novel hypothesis is proposed for universal rules in AI performance, providing a new idea that could be explored in other AI applications.

Systemic Contact Dermatitis: The Routes of Allergen Entry

Systemic contact dermatitis (SCD) is a generalized reactivation of type IV hypersensitivity skin diseases in individuals with previous sensitization after a contact allergen is administered systemically. Patients with SCD may consider their dermatitis unpredictable and recalcitrant since the causative allergens are difficult to find. If a patient has a pattern of dermatitis suggestive of SCD but fails to improve with conventional treatment, SCD should be taken into consideration. If doctors are not familiar with the presentations of SCD and the possible routes of allergen sensitization and exposure, the diagnosis of SCD may be delayed. In this work, we summarized all of the routes through which allergens can enter the body and cause SCD, including oral intake, local contact (through skin, inhalation, nasal spray and anal application), implants, and other iatrogenic or invasive routes (intravenous, intramuscular, intraarticular, and intravesicular). This will provide a comprehensive reference for the clinicians to identify the culprit of SCD.

Computational diagnostic methods on 2D photographs: A review of the literature

Here we provide a literature review of all the methods reported to date for analyzing 2D pictures for diagnostic purposes. Pubmed was used to screen the MEDLINE database using MeSH (Medical Subject Heading) terms and keyworks. The different recognition steps and the main results were reported. All human studies involving 2D facial photographs used to diagnose one or several conditions in healthy populations or in patients were included. We included 1515 articles and 27 publications were finally retained. 67% of the articles aimed at diagnosing one particular syndrome versus healthy controls and 33% aimed at performing multi-class syndrome recognition. Data volume varied from 15 to 17,106 patient pictures. Manual or automatic landmarks were one of the most commonly used tools in order to extract morphological information from images, in 22/27 (81%) publications. Geometrical features were extracted from landmarks based on Procrustes superimposition in 4/27 (15%). Textural features were extracted in 19/27 (70%) publications. Features were then classified using machine learning methods in 89% of publications, while deep learning methods were used in 11%. Facial recognition tools were generally successful in identifying rare conditions in dysmorphic patients, with comparable or higher recognition accuracy than clinical experts.

A New Clinical Model to Estimate the Pre-Test Probability of Cushing's Syndrome: The Cushing Score

BACKGROUND

Hypercortisolism accounts for relevant morbidity and mortality and is often a diagnostic challenge for clinicians. A prompt diagnosis is necessary to treat Cushing's syndrome as early as possible.

OBJECTIVE

The aim of this study was to develop and validate a clinical model for the estimation of pre-test probability of hypercortisolism in an at-risk population.

DESIGN

We conducted a retrospective multicenter case-control study, involving five Italian referral centers for Endocrinology (Turin, Messina, Naples, Padua and Rome). One hundred and fifty patients affected by Cushing's syndrome and 300 patients in which hypercortisolism was excluded were enrolled. All patients were evaluated, according to current guidelines, for the suspicion of hypercortisolism.

RESULTS

The Cushing score was built by multivariable logistic regression, considering all main features associated with a clinical suspicion of hypercortisolism as possible predictors. A stepwise backward selection algorithm was used (final model AUC=0.873), then an internal validation was performed through ten-fold cross-validation. Final estimation of the model performance showed an average AUC=0.841, thus reassuring about a small overfitting effect. The retrieved score was structured on a 17.5-point scale: low-risk class (score value: ≤5.5, probability of disease=0.8%); intermediate-low-risk class (score value: 6-8.5, probability of disease=2.7%); intermediate-high-risk class (score value: 9-11.5, probability of disease=18.5%) and finally, high-risk class (score value: ≥12, probability of disease=72.5%).

CONCLUSIONS

We developed and internally validated a simple tool to determine pre-test probability of hypercortisolism, the Cushing score, that showed a remarkable predictive power for the discrimination between subjects with and without a final diagnosis of Cushing's syndrome.

What Are Important Ethical Implications of Using Facial Recognition Technology in Health Care?

Applications of facial recognition technology (FRT) in health care settings have been developed to identify and monitor patients as well as to diagnose genetic, medical, and behavioral conditions. The use of FRT in health care suggests the importance of informed consent, data input and analysis quality, effective communication about incidental findings, and potential influence on patient-clinician relationships. Privacy and data protection are thought to present challenges for the use of FRT for health applications.

Toward a Diagnostic Score in Cushing's Syndrome

Cushing's syndrome (CS) is a classical rare disease: it is often suspected in patients who do not have the disease; at the same time, it takes a mean of 3 years to diagnose CS in affected individuals. The main reason is the extreme rarity (1-3/million/year) in combination with the lack of a single lead symptom. CS has to be suspected when a combination of signs and symptoms is present, which together make up the characteristic phenotype of cortisol excess. Unusual fat distribution affecting the face, neck, and trunk; skin changes including plethora, acne, hirsutism, livid striae, and easy bruising; and signs of protein catabolism such as thinned and vulnerable skin, osteoporotic fractures, and proximal myopathy indicate the need for biochemical screening for CS. In contrast, common symptoms like hypertension, weight gain, or diabetes also occur quite frequently in the general population and per se do not justify biochemical testing. First-line screening tests include urinary free cortisol excretion, dexamethasone suppression testing, and late-night salivary cortisol measurements. All three tests have overall reasonable sensitivity and specificity, and first-line testing should be selected on the basis of the physiologic conditions of the patient, drug intake, and available laboratory quality control measures. Two normal test results usually exclude the presence of CS. Other tests and laboratory parameters like the high-dose dexamethasone suppression test, plasma ACTH, the CRH test, and the bilateral inferior petrosal sinus sampling are not part of the initial biochemical screening. As a general rule, biochemical screening should only be performed if the pre-test probability for CS is reasonably high. This article provides an overview about the current standard in the diagnosis of CS starting with clinical scores and screenings, the clinical signs, relevant differential diagnoses, the first-line biochemical screening, and ending with a few exceptional cases.

Development of a computer-aided tool for the pattern recognition of facial features in diagnosing Turner syndrome: comparison of diagnostic accuracy with clinical workers

Technologies applied for the recognition of facial features in diagnosing certain disorders seem to be promising in reducing the medical burden and improve the efficiency. This pilot study aimed to develop a computer-assisted tool for the pattern recognition of facial features for diagnosing Turner syndrome (TS). Photographs of 54 patients with TS and 158 female controls were collected from July 2016 to May 2017. Finally, photographs of 32 patients with TS and 96 age-matched controls were included in the study that were further divided equally into training and testing groups. The process of automatic classification consisted of image preprocessing, facial feature extraction, feature reduction and fusion, automatic classification, and result presentation. A total of 27 physicians and 21 medical students completed a web-based test including the same photographs used in computer testing. After training, the automatic facial classification system for diagnosing TS achieved a 68.8% sensitivity and 87.5% specificity (and a 67.6% average sensitivity and 87.9% average specificity after resampling), which was significantly higher than the average sensitivity (57.4%, P < 0.001) and specificity (75.4%, P < 0.001) of 48 participants, respectively. The accuracy of this system was satisfactory and better than the diagnosis by clinicians. However, the system necessitates further improvement for achieving a high diagnostic accuracy in clinical practice.

A Survey on Computer Vision for Assistive Medical Diagnosis From Faces

Automatic medical diagnosis is an emerging center of interest in computer vision as it provides unobtrusive objective information on a patient's condition. The face, as a mirror of health status, can reveal symptomatic indications of specific diseases. Thus, the detection of facial abnormalities or atypical features is at upmost importance when it comes to medical diagnostics. This survey aims to give an overview of the recent developments in medical diagnostics from facial images based on computer vision methods. Various approaches have been considered to assess facial symptoms and to eventually provide further help to the practitioners. However, the developed tools are still seldom used in clinical practice, since their reliability is still a concern due to the lack of clinical validation of the methodologies and their inadequate applicability. Nonetheless, efforts are being made to provide robust solutions suitable for healthcare environments, by dealing with practical issues such as real-time assessment or patients positioning. This survey provides an updated collection of the most relevant and innovative solutions in facial images analysis. The findings show that with the help of computer vision methods, over 30 medical conditions can be preliminarily diagnosed from the automatic detection of some of their symptoms. Furthermore, future perspectives, such as the need for interdisciplinary collaboration and collecting publicly available databases, are highlighted.

Diagnostic use of facial image analysis software in endocrine and genetic disorders: review, current results and future perspectives

Cushing's syndrome (CS) and acromegaly are endocrine diseases that are currently diagnosed with a delay of several years from disease onset. Novel diagnostic approaches and increased awareness among physicians are needed. Face classification technology has recently been introduced as a promising diagnostic tool for CS and acromegaly in pilot studies. It has also been used to classify various genetic syndromes using regular facial photographs. The authors provide a basic explanation of the technology, review available literature regarding its use in a medical setting, and discuss possible future developments. The method the authors have employed in previous studies uses standardized frontal and profile facial photographs for classification. Image analysis is based on applying mathematical functions evaluating geometry and image texture to a grid of nodes semi-automatically placed on relevant facial structures, yielding a binary classification result. Ongoing research focuses on improving diagnostic algorithms of this method and bringing it closer to clinical use. Regarding future perspectives, the authors propose an online interface that facilitates submission of patient data for analysis and retrieval of results as a possible model for clinical application.

Cushing's syndrome: update on signs, symptoms and biochemical screening

Endogenous pathologic hypercortisolism, or Cushing's syndrome, is associated with poor quality of life, morbidity, and increased mortality. Early diagnosis may mitigate against this natural history of the disorder. The clinical presentation of Cushing's syndrome varies, in part related to the extent and duration of cortisol excess. When hypercortisolism is severe, its signs and symptoms are unmistakable. However, most of the signs and symptoms of Cushing's syndrome are common in the general population (e.g., hypertension and weight gain) and not all are present in every patient. In addition to classical features of glucocorticoid excess, such as proximal muscle weakness and wide purple striae, patients may present with the associated comorbidities that are caused by hypercortisolism. These include cardiovascular disease, thromboembolic disease, psychiatric and cognitive deficits, and infections. As a result, internists and generalists must consider Cushing's syndrome as a cause, and endocrinologists should search for and treat these comorbidities. Recommended tests to screen for Cushing's syndrome include 1  mg dexamethasone suppression, urine free cortisol, and late night salivary cortisol. These may be slightly elevated in patients with physiologic hypercortisolism, which should be excluded, along with exogenous glucocorticoid use. Each screening test has caveats and the choice of tests should be individualized based on each patient's characteristics and lifestyle. The objective of this review is to update the readership on the clinical and biochemical features of Cushing's syndrome that are useful when evaluating patients for this diagnosis.