Uvéites et médecine interne : stratégies diagnostique et thérapeutique

Neural networks for predicting etiological diagnosis of uveitis

BACKGROUND/OBJECTIVES

The large number and heterogeneity of causes of uveitis make the etiological diagnosis a complex task. The clinician must consider all the information concerning the ophthalmological and extra-ophthalmological features of the patient. Diagnostic machine learning algorithms have been developed and provide a correct diagnosis in one-half to three-quarters of cases. However, they are not integrated into daily clinical practice. The aim is to determine whether machine learning models can predict the etiological diagnosis of uveitis from clinical information.

METHODS

This cohort study was performed on uveitis patients with unknown etiology at first consultation. One hundred nine variables, including demographic, ophthalmic, and clinical information, associated with complementary exams were analyzed. Twenty-five causes of uveitis were included. A neural network was developed to predict the etiological diagnosis of uveitis. The performance of the model was evaluated and compared to a gold standard: etiological diagnosis established by a consensus of two uveitis experts.

RESULTS

A total of 375 patients were included in this analysis. Findings showed that the neural network type (Multilayer perceptron) (NN-MLP) presented the best prediction of the etiological diagnosis of uveitis. The NN-MLP's most probable diagnosis matched the senior clinician diagnosis in 292 of 375 patients (77.8%, 95% CI: 77.4-78.0). It achieved 93% accuracy (95% CI: 92.8-93.1%) when considering the two most probable diagnoses. The NN-MLP performed well in diagnosing idiopathic uveitis (sensitivity of 81% and specificity of 82%). For more than three-quarters of etiologies, our NN-MLP demonstrated good diagnostic performance (sensitivity > 70% and specificity > 80%).

CONCLUSION

Study results suggest that developing models for accurately predicting the etiological diagnosis of uveitis with undetermined etiology based on clinical information is feasible. Such NN-MLP could be used for the etiological assessments of uveitis with unknown etiology.

[Monogenic autoinflammatory uveitis]

Monogenic autoinflammatory uveitis belongs to the spectrum of monogenic autoinflammatory diseases. When early-onset uveitis is associated with specific extra-ocular manifestations, particularly in a familial or geographical context, it guides the clinician towards a diagnosis of a monogenic autoinflammatory disease. The clinical presentation and mode of inheritance will help identify the underlying cause, and the detection of a pathogenic variant will confirm the diagnosis and guide the management approach. In this review, we outline the main monogenic autoinflammatory uveitis conditions that clinicians should be aware of: Blau syndrome, ROSAH syndrome, cryopyrin-associated periodic syndromes (CAPS), partial mevalonate kinase deficiency, A20 haploinsufficiency, and NEMO syndrome.

Etiological Diagnosis of Uveitis: Contribution of the of the Extra-Ophthalmological Clinical Examination

PURPOSE

Determining uveitis etiology is a challenge. It is based primarily on demographic data and the characteristics of eye examination. It is not clear to what extent extraocular physical signs contribute to elucidating the etiology. This study aimed to establish the contribution of the clinical extra-ophthalmological features for the assessment of the underlying etiology of uveitis.

METHODS

We retrospectively reviewed 1307 patients with uveitis referred to our tertiary center between 2003 and 2021. Uveitis was classified according to the Standardization of Uveitis Nomenclature. Clinical features were collected at diagnosis by internists before the etiological diagnosis was made. The main outcome description was the contribution of clinical features.

RESULTS

Clinical extra-ophthalmological features contributed to the assessment of the underlying etiology of uveitis in 363 (27.8%) patients. The joint and the skin examinations were the most useful for etiological investigations, respectively in 12.3% and 11.8% of patients. Five etiologies of uveitis accounted for 80% of the cases: sarcoidosis, HLA-B27-related uveitis, Behçet's disease, multiple sclerosis, and Vogt-Koyanagi-Harada disease. Clinical extra-ophthalmological features were particularly important in the etiological diagnosis of acute bilateral anterior uveitis and panuveitis.

CONCLUSION

This study suggests that clinical extra-ophthalmological features are essential for the etiological diagnosis of uveitis in more than a quarter of patients. It demonstrates once again the value of collaboration between ophthalmologists and other specialists experienced in performing extra-ophthalmological clinical examinations, particularly in patients with acute bilateral anterior uveitis and panuveitis.

Value of Chest X-Ray and Chest Computed Tomography for Systemic Sarcoidosis Diagnosis in Undifferentiated Uveitis

BACKGROUND

To evaluate the contribution of chest X-ray and chest CT for the diagnosis of sarcoid uveitis.

METHODS

Retrospective study on consecutive patients with uveitis of unknown etiology, who underwent both chest X-ray and CT during uveitis diagnosis workup in a tertiary French university hospital.

RESULTS

A total of 914 patients were included. Systemic sarcoidosis was identified in 23.1%. The probability of discordance between chest X-ray and CT increased with age at diagnosis (p < 0.001). In patients 30 years of age and younger, the probability of discordance was 5% or less, and 0.8% if the ACE level was normal. After 78.3 years of age, the probability of discordance was 20% or more.

CONCLUSION

We recommend not to perform CT in patients under 30 years of age with a normal chest X ray and ACE level, and suggest performing chest CT first in the elderly.

IgG4-related uveitis. A French cohort and literature review

INTRODUCTION

IgG4-related disease (IgG4-RD) is a fibro-inflammatory disorder that can affect almost any organ. IgG4-related ophthalmic disease is a protean condition involving the orbit and ocular adnexa. Although a few cases of uveitis have been reported, the exact pattern of IgG4-related intraocular manifestations remains unclear. Here, we report on a nationwide French multicenter cohort of patients with IgG4-RD and uveitis and conducted a literature review.

METHODS

Patients with uveitis and a concomitant definite diagnosis of IgG4-RD (Revised Comprehensive Diagnostic criteria, American College of Rheumatology/European League Against Rheumatism classification criteria for IgG4-RD, International Consensus Diagnostic Criteria for auto-immune pancreatitis, or diagnostic criteria for IgG4-related hypophysitis), were screened from our national IgG4-RD and systemic fibrosis database. Concomitantly, we conducted a PubMed literature review and selected cases of definite IgG4-RD with uveitis.

RESULTS

We reported on 16 patients (8 from our database and 8 from the literature) and a total of 30 episodes of uveitis. Uveitis cases represented 3 % of total IgG4-RD patients in the national database on IgG4-RD and systemic fibrosis. Uveitis was inaugural in IgG4-RD in 4/16 cases (25 %) (appearing before any other IgG4-related symptom, at a median of 9 months), occurred concurrently to other IgG4-related symptoms in 9/16 cases (56 %) (at a median of 15 months before IgG4-RD diagnosis), and appeared during follow up in 3/16 patients (19 %) (at a median of 57 months after first IgG4-related symptoms). When uveitis occurred during follow up, it was associated with IgG4-RD manifestations in other organs in 6/9 patients (67 %). Uveitis was bilateral in 8/16 cases (50 %) and granulomatous in 5/10 cases (50 %). It was anterior in 8/13 (62 %), intermediate in 3/13 (23 %), and global (panuveitis) in 4/13 patients (31 %). Median serum IgG4 at diagnosis was 3.2 g/L. Median follow up time was of 6 years, during which 8/16 patients (50 %) experienced at least one relapse of uveitis. Treatment data was available for 29/30 uveitis flares. Steroids were used in 28/29 episodes of uveitis (97 %), leading to remission of uveitis in 16/28 cases (57 %). Methotrexate and rituximab (in combination with systemic steroids) were administered as second- or third-line therapy in 6/29 (21 %) and 5/29 (17 %) episodes of uveitis, respectively, and led to remission of uveitis in 4/6 cases (67 %) and 4/5 cases (80 %), respectively. One third of uveitides required at least two different lines of treatment for remission induction (mainly a combination of both systemic steroids and methotrexate or rituximab).

DISCUSSION AND CONCLUSIONS

Uveitis may be one of the initial symptoms of IgG4-RD, and IgG4-RD should be considered in the diagnostic workup of uveitis. Its early onset in IgG4-RD may help with early diagnosis and treatment of the disease. Steroid monotherapy may be sufficient to treat IgG4-related uveitis, yet relapses were frequent (50 %) and ultimately a third of patients required at least two lines of treatment. Hence, steroid-sparing agents can be considered at early stages of the disease, particularly for patients with a high risk of relapse or steroid-related complications.

Diagnosis, Classification, and Assessment of the Underlying Etiology of Uveitis by Artificial Intelligence: A Systematic Review

Recent years have seen the emergence and application of artificial intelligence (AI) in diagnostic decision support systems. There are approximately 80 etiologies that can underly uveitis, some very rare, and AI may lend itself to their detection. This synthesis of the literature selected articles that focused on the use of AI in determining the diagnosis, classification, and underlying etiology of uveitis. The AI-based systems demonstrated relatively good performance, with a classification accuracy of 93-99% and a sensitivity of at least 80% for identifying the two most probable etiologies underlying uveitis. However, there were limitations to the evidence. Firstly, most data were collected retrospectively with missing data. Secondly, ophthalmic, demographic, clinical, and ancillary tests were not reliably integrated into the algorithms' dataset. Thirdly, patient numbers were small, which is problematic when aiming to discriminate rare and complex diagnoses. In conclusion, the data indicate that AI has potential as a diagnostic decision support system, but clinical applicability is not yet established. Future studies and technologies need to incorporate more comprehensive clinical data and larger patient populations. In time, these should improve AI-based diagnostic tools and help clinicians diagnose, classify, and manage patients with uveitis.

Relevance of Brain MRI in Patients with Uveitis: Retrospective Cohort on 402 Patients

Aim: To assess the diagnostic value of brain magnetic resonance imaging (bMRI) for the etiological diagnosis of uveitis and to establish predictive factors associated with its advantageous use.Methods: Retrospective study on all patients with de novo uveitis who were referred to our tertiary hospital and who underwent a bMRI between 2003 and 2018.Results: bMRI was contributive in 19 out of 402 cases (5%), among patients with a contributive bMRI, 68% had neurological signs. Univariate analysis established that neurological signs (p < .001), granulomatous uveitis (p = .003), retinal vasculitis (p = .002), and intermediate uveitis (p < .001) were all significantly associated with a contributive bMRI. Multivariate analysis confirms the significant association of neurological signs (p < .001) and intermediate uveitis (p = .01).Conclusion: bMRI appears to be a relevant exam in specific cases; intermediate/posterior uveitis or panuveitis accompanied by neurological signs, retinal vasculitis, or in patients older than 40, to rule out an oculocerebral lymphoma.Abbreviations: ACE: Angiotensin-Converting Enzyme; bMRI: Magnetic Resonance Imaging; CBC: Complete Blood cell Count; BMRI: Brain Magnetic Resonance Imaging; CT: Computerized Tomography; MS: Multiple Sclerosis; NS: Neurological Signs; OCL: Oculocerebral Lymphoma; RIS: Radiologically Isolated Syndrome.

Uveitis as an Open Window to Systemic Inflammatory Diseases

Spondyloarthritis (Spa), Behçet's disease (BD) and sarcoidosis are major systemic inflammatory diseases worldwide. They are all multisystem pathologies and share a possible ocular involvement, especially uveitis. We hereby describe selected cases who were referred by ophthalmologists to our internal medicine department for unexplained uveitis. Physical examination and/or the use of laboratory and imaging investigations allowed to make a diagnosis of a systemic inflammatory disease in a large proportion of patients. In our tertiary referral center, 75 patients have been diagnosed with Spa (n = 20), BD (n = 9), or sarcoidosis (n = 46) in the last two years. There was a significant delay in the diagnosis of Spa-associated uveitis. Screening strategies using Human Leukocyte Antigen (HLA)-B27 determination and sacroiliac magnetic resonance imaging in patients suffering from chronic low back pain and/or psoriasis helped in the diagnosis. BD's uveitis affects young people from both sexes and all origins and usually presents with panuveitis and retinal vasculitis. The high proportion of sarcoidosis in our population is explained by the use of chest computed tomography (CT) and 18F-fluorodeoxyglucose positron emission tomography CT that helped to identify smaller hilar or mediastinal involvement and allowed to further investigate those patients, especially in the elderly. Our results confirm how in these sight- and potentially life-threatening diseases a prompt diagnosis is mandatory and benefits from a multidisciplinary approach.

Uveitis: Autoimmunity… and beyond

OBJECTIVE

Uveitis is the most common ophthalmological finding in the practice of rheumatology and clinical immunology. The condition is frequently idiopathic but about 60 causes of uveitis have been described. Our aim was to analyze the clinical patterns and etiologies of uveitis in a tertiary referral center.

METHODS

The records of 912 consecutive patients referred to the department of internal medicine (Lyon University Hospital, Lyon, France) for the diagnostic work-up of uveitis were examined. Demographic, clinical, anatomical, and etiological features of uveitis were analyzed.

RESULTS

The mean age at onset was 48.8 years; 59.8% of the patients were women and 78.2% were Caucasians. Anterior uveitis was the most common type of uveitis (40.6%), followed by panuveitis (31.7%), posterior (18.75%) and intermediate uveitis (9%). 46.9% of the patients had idiopathic uveitis. The most common etiologies were systemic diseases (37.3%), such as sarcoidosis (17.1%), HLA-B27-related uveitis and/or spondyloarthritis (12.5%), and tuberculosis (7.5%).

CONCLUSION

We describe one of the largest cohorts of consecutive uveitis patients referred to a department of internal medicine. The high percentage of uveitis associated with underlying (systemic) diseases highlights the need for a multidisciplinary approach, in order to reduce the diagnostic delay.

Clinical Manifestations, Complications and Treatment of Ocular Sarcoidosis: Correlation between Visual Efficiency and Macular Edema as Seen on Optical Coherence Tomography

Sarcoidosis is a chronic systemic autoimmune disease which belongs to a group of systemic granulomatous diseases. It can be confirmed through characteristic systemic and ocular manifestations and histological findings. Biopsy is the golden standard for diagnosing sarcoidosis. Ocular sarcoidosis can be confirmed, probable, or possible. Over a two-year period, ocular manifestations were studied on a sample of 52 patients, each followed for four months and diagnosed with some form of systemic sarcoidosis. Most frequent systemic manifestations in patients with ocular sarcoidosis were pulmonary, skin, glandular, and systemic generalized sarcoidosis. The disease was diagnosed four times more frequently in females than males (42:10, respectively; p < 0.05). Most frequent, and statistically significant, manifestation of ocular sarcoidosis is anterior uveitis (64.61%; p < 0.01). Macular edema and periphlebitis associated with periarteritis were frequent, and statistically significant (43.90% and 29.26%, respectively; p < 0.05). Overall, with regards to gender and location (right eye; left eye), visual acuity was >0.5 and of statistical significance (76.92%; p < 0.01). The most common therapy consisted of systemic corticosteroids (26.67%) and/or a combination of corticosteroids and immunosuppressive drugs (23.33%). In 16 eyes treated with repeated doses of sub-Tenon's injections, both initial and control visual acuity correlated with average thickness. There was positive correlation between several optical coherence tomography findings before and after treatment.

Interleukin-2 Receptor and Angiotensin-Converting Enzyme as Markers for Ocular Sarcoidosis

PURPOSE

To study the impact of soluble IL2 receptor (sIL2R), chest x-ray (CxR), and angiotensin-converting enzyme (ACE) as markers for sarcoidosis in uveitis patients.

DESIGN

Retrospective study.

METHODS

Serum concentrations of sIL2R and ACE were measured in patients with active uveitis. Those with elevated sIL2R and /or ACE values were examined for suspected systemic sarcoidosis.

MAIN OUTCOME MEASURE

Our main outcome parameters were the specificity and sensitivity of sIL2R, CxR and ACE in screening for ocular sarcoidosis.

RESULTS

We measured 261 patients with uveitis for sarcoidosis using sIL2R and ACE between January 2008 and November 2011; sarcoidosis was been diagnosed using other tests (e.g. computer tomography, brochoalveolar lavage, biopsy) in 41 of 53 patients with elevated sIL2R values (>639 U/ml) and in one patient with normal sIL2R (582 U/ml). Their mean sIL2R value was 1310 U/ml, extending from 582 to 8659 U/ml. Only 9 patients, however, presented elevated ACE (>82 U/l). Their mean ACE value was 116.4 U/l, ranging from 84.1 to 175.5 U/l. IL2R specificity was 94% with 98% sensitivity. In contrast, ACE had a specificity of 99.5%, but a sensitivity of only 22%; the chest x-ray had a specificity of 100% with 50% sensitivity in detecting sarcoidosis. We observed the entire spectrum of uveitis: sixteen patients suffered from anterior, 8 from intermediate, 16 from posterior, and 2 from panuveitis.

CONCLUSIONS

An elevated level of soluble IL2R suggests sarcoidosis with uveitis more convincingly than ACE, making sIL2R a more effective marker parameter for sarcoidosis than ACE or chest x-ray in uveitis patients.