Leukocyte profiles in blood and CSF distinguish neurosarcoidosis from multiple sclerosis

A systematic review of the application of machine-learning algorithms in multiple sclerosis

INTRODUCTION

The applications of artificial intelligence, and in particular automatic learning or "machine learning" (ML), constitute both a challenge and a great opportunity in numerous scientific, technical, and clinical disciplines. Specific applications in the study of multiple sclerosis (MS) have been no exception, and constitute an area of increasing interest in recent years.

OBJECTIVE

We present a systematic review of the application of ML algorithms in MS.

MATERIALS AND METHODS

We used the PubMed search engine, which allows free access to the MEDLINE medical database, to identify studies including the keywords "machine learning" and "multiple sclerosis." We excluded review articles, studies written in languages other than English or Spanish, and studies that were mainly technical and did not specifically apply to MS. The final selection included 76 articles, and 38 were rejected.

CONCLUSIONS

After the review process, we established 4 main applications of ML in MS: 1) classifying MS subtypes; 2) distinguishing patients with MS from healthy controls and individuals with other diseases; 3) predicting progression and response to therapeutic interventions; and 4) other applications. Results found to date have shown that ML algorithms may offer great support for health professionals both in clinical settings and in research into MS.

Contactin-associated protein 2 autoantibodies can be associated with multifocal motor-like neuropathy: a case report

Autoantibodies against contactin-associated protein 2 (CASPR2) are usually associated with autoimmune encephalitis and neuromyotonia. Their association with inflammatory neuropathies has been described in case reports albeit all with distal symmetric manifestation. Here, we report a patient who developed distal arm paresis, dominantly of the right arm, over the course of 1 year. Electroneurography showed a conduction block of motor nerve conduction, nerve ultrasonography a swelling of the right median and ulnar nerve and flow cytometry an increase in natural killer (NK cells) in the blood and natural killer T (NKT) cells in the cerebrospinal fluid (CSF), therefore indicating a multifocal motor neuropathy-like (MMN-like) phenotype. CASPR2 autoantibodies were detected in serum and CSF. Through immunotherapy with intravenous immunoglobulins the patient showed clinical and neurographic improvement. We therefore describe the first association of CASPR2 autoantibodies with a MMN-like clinical manifestation, extending the spectrum of CASPR2-associated diseases.

Supporting the differential diagnosis of connective tissue diseases with neurological involvement by blood and cerebrospinal fluid flow cytometry

OBJECTIVE

Neurological manifestations of autoimmune connective tissue diseases (CTD) are poorly understood and difficult to diagnose. We here aimed to address this shortcoming by studying immune cell compositions in CTD patients with and without neurological manifestation.

METHODS

Using flow cytometry, we retrospectively investigated paired cerebrospinal fluid (CSF) and blood samples of 28 CTD patients without neurological manifestation, 38 CTD patients with neurological manifestation (N-CTD), 38 non-inflammatory controls, and 38 multiple sclerosis (MS) patients, a paradigmatic primary neuroinflammatory disease.

RESULTS

We detected an expansion of plasma cells in the blood of both N-CTD and CTD compared to non-inflammatory controls and MS. Blood plasma cells alone distinguished the clinically similar entities N-CTD and MS with high discriminatory performance (AUC: 0.81). Classical blood monocytes indicated higher disease activity in systemic lupus erythematosus (SLE) patients. Surprisingly, immune cells in the CSF did not differ significantly between N-CTD and CTD, while CD4⁺ T cells and the CD4⁺/CD8⁺ ratio were elevated in the blood of N-CTD compared to CTD. Several B cell-associated parameters partially overlapped in the CSF in MS and N-CTD. We built a machine learning model that distinguished N-CTD from MS with high discriminatory power using either blood or CSF.

CONCLUSION

We here find that blood flow cytometry alone surprisingly suffices to distinguish CTD with neurological manifestations from clinically similar entities, suggesting that a rapid blood test could support clinicians in the differential diagnosis of N-CTD.

The CSF in neurosarcoidosis contains consistent clonal expansion of CD8 T cells, but not CD4 T cells

The tissue-specific drivers of neurosarcoidosis remain poorly defined. To identify cerebrospinal fluid (CSF) specific, antigen-driven T and B cell responses, we performed single-cell RNA sequencing of CSF and blood cells from neurosarcoid participants coupled to T and B cell receptor sequencing. In contrast to pulmonary sarcoidosis, which is driven by CD4 T cells, we found CD8 T cell clonal expansion enriched in the neurosarcoid CSF. These CSF-enriched CD8 T cells were composed of two subsets with differential expression of EBI2, CXCR3, and CXCR4. Lastly, our data suggest that IFNγ signaling may distinguish neurosarcoidosis from other neurological disorders.

Role and Relevance of Cerebrospinal Fluid Cells in Diagnostics and Research: State-of-the-Art and Underutilized Opportunities

Cerebrospinal fluid (CSF) has recently experienced a revival in diagnostics and research. However, little progress has been made regarding CSF cell analysis. For almost a century, CSF cell count and cytomorphological examination have been central diagnostic parameters, with CSF pleocytosis as a hallmark finding of neuroinflammation and cytology offering valuable clues regarding infectious, autoimmune, and malignant aetiologies. A great deal of information, however, remains unattended as modern immune phenotyping technologies have not yet been broadly incorporated into routine CSF analysis. This is a serious deficit considering the central role of CSF cells as effectors in central nervous system (CNS) immune defence and autoimmune CNS processes, and the diagnostic challenges posed by clinically overlapping infectious and immune-mediated CNS diseases. Here, we summarize historical, specimen-intrinsic, methodological, and technical issues determining the state-of-the-art diagnostics of CSF cells and outline future perspectives for this underutilized window into meningeal and CNS immunity.

CSF-resident CD4+ T-cells display a distinct gene expression profile with relevance to immune surveillance and multiple sclerosis

The CNS has traditionally been considered an immune privileged site, but is now understood to have a system of immune surveillance, predominantly involving CD4+ T-cells. Identifying functional differences between CNS and blood CD4+ T-cells, therefore, have relevance to CNS immune surveillance as well as to neurological conditions, such as multiple sclerosis, in which CD4+ T-cells play a central role. Here, CD4+ T-cells were purified from CSF and blood from 21 patients with newly diagnosed treatment-naïve multiple sclerosis and 20 individuals with non-inflammatory disorders using fluorescence-activated cell sorting, and their transcriptomes were profiled by RNA sequencing. Paired comparisons between CD4+ T-cells from CSF and blood identified 5156 differentially expressed genes in controls and 4263 differentially expressed in multiple sclerosis patients at false discovery rate <5%. Differential expression analysis of CD4+ T-cells collected from the CSF highlighted genes involved in migration, activation, cholesterol biosynthesis and signalling, including those with known relevance to multiple sclerosis pathogenesis and treatment. Expression of markers of CD4+ T-cell subtypes suggested an increased proportion of Th1 and Th17 cells in CSF. Gene ontology terms significant only in multiple sclerosis were predominantly those involved in cellular proliferation. A two-way comparison of CSF versus blood CD4+ T-cells in multiple sclerosis compared with non-inflammatory disorder controls identified four significant genes at false discovery rate <5% (CYP51A1, LRRD1, YES1 and PASK), further implicating cholesterol biosynthesis and migration mechanisms. Analysis of CSF CD4+ T-cells in an extended cohort of multiple sclerosis cases (total N = 41) compared with non-inflammatory disorder controls (total N = 38) identified 140 differentially expressed genes at false discovery rate < 5%, many of which have known relevance to multiple sclerosis, including XBP1, BHLHE40, CD40LG, DPP4 and ITGB1. This study provides the largest transcriptomic analysis of purified cell subpopulations in CSF to date and has relevance for the understanding of CNS immune surveillance, as well as multiple sclerosis pathogenesis and treatment discovery.

Artificial Intelligence in Diagnosis and Prediction of the Multiple Sclerosis Progression: A Scoping Review (Preprint).. [DOI: 10.2196/preprints.29720] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

BACKGROUND

Multiple Sclerosis (MS) is an autoimmune disease that results from the demyelination of the nerves in the Central Nervous System. The diagnosis depends on clinical history, neurological examination, and radiological images. Artificial Intelligence proved to be an effective tool in enhancing the diagnostic tools of MS.

OBJECTIVE

To explore how AI assisted in diagnosis and predicting the progression of MS.

METHODS

We used three bibliographic databases in our search: PubMed IEEE Xplore and Cochrane in our search. The study selection process included: removal of duplicated articles, screening titles and abstracts, and reading the full text. This process was performed by two reviewers. The data extracted from the included studies have been filled in an Excel sheet. This step had been done by each reviewer accordingly to the assigned articles. The extracted data sheet was checked by two reviewers to have accuracy ensured. The narrative approach is applied in data synthesis.

RESULTS

The search conducted resulted in 320 articles Removing duplicates and excluding the ineligible articles due to irrelevancy to the population, intervention, and outcomes resulted in excluding 299 articles. Thus, our review will include 21 articles for data extraction and data synthesis.

CONCLUSIONS

Artificial Intelligence is becoming a trend in the medical field. Its contribution in enhancing the diagnostic tools of many diseases, as in MS, is prominent and can be built on in further development plans. However, the implementation of Artificial Intelligence in Multiple Sclerosis is not widespread to confirm the benefits gained, and the datasets involved in the current practice are relatively small. It is recommended to have more studies that focus on the relationship between the employment of AI in diagnosis and monitoring progression and the accuracy gained by this employment.

A systematic review of the application of machine-learning algorithms in multiple sclerosis

INTRODUCTION

The applications of artificial intelligence, and in particular automatic learning or "machine learning" (ML), constitute both a challenge and a great opportunity in numerous scientific, technical, and clinical disciplines. Specific applications in the study of multiple sclerosis (MS) have been no exception, and constitute an area of increasing interest in recent years.

OBJECTIVE

We present a systematic review of the application of ML algorithms in MS.

MATERIALS AND METHODS

We used the PubMed search engine, which allows free access to the MEDLINE medical database, to identify studies including the keywords "machine learning" and "multiple sclerosis." We excluded review articles, studies written in languages other than English or Spanish, and studies that were mainly technical and did not specifically apply to MS. The final selection included 76 articles, and 38 were rejected.

CONCLUSIONS

After the review process, we established 4 main applications of ML in MS: 1) classifying MS subtypes; 2) distinguishing patients with MS from healthy controls and individuals with other diseases; 3) predicting progression and response to therapeutic interventions; and 4) other applications. Results found to date have shown that ML algorithms may offer great support for health professionals both in clinical settings and in research into MS.

Specific B- and T-cell populations are associated with cognition in patients with epilepsy and antibody positive and negative suspected limbic encephalitis

OBJECTIVE

Neuropsychological impairments are major symptoms of autoimmune limbic encephalitis (LE) epilepsy patients. In LE epilepsy patients with an autoimmune response against intracellular antigens as well as in antibody-negative patients, the antibody findings and magnetic resonance imaging pathology correspond poorly to the clinical features. Here, we evaluated whether T- and B-cells are linked to cognitive impairment in these groups.

METHODS

In this cross-sectional, observational, case-controlled study, we evaluated 106 patients with adult-onset epilepsies with a suspected autoimmune etiology. We assessed verbal and visual memory, executive function, and mood in relation to the presence or absence of known auto-antibodies, and regarding T- and B-cell activity as indicated by flow cytometry (fluorescence-activated cell sorting = FACS, peripheral blood = PB and cerebrospinal fluid = CSF).

RESULTS

56% of the patients were antibody-negative. In the other patients, auto-antibodies were directed against intracellular antigens (GAD65, paraneoplastic: 38%), or cellular surface antigens (LGI1/CASPR2/NMDA-R: 6%). Excluding LGI1/CASPR2/NMDA-R, the groups with and without antibodies did not differ in disease features, cognition, or mood. CD4+ T-cells and CD8+ T-cells in blood and CD4+ T-cells in CSF were prominent in the auto-antibody positive group. Regression analyses indicated the role education, drug load, amygdala and/or hippocampal pathology, and CD4+ T-cells play in verbal memory and executive function. Depressed mood revealed no relation to flow cytometry results.

CONCLUSION

Our results indicate a link between T- and B-cell activity and cognition in epilepsy patients with suspected limbic encephalitis, thus suggesting that flow cytometry results can provide an understanding of cognitive impairment in LE patients with autoantibodies against intracellular antigens.

Cerebrospinal fluid flow cytometry distinguishes psychosis spectrum disorders from differential diagnoses

Psychotic disorders are common and disabling mental conditions. The relative importance of immune-related mechanisms in psychotic disorders remains subject of debate. Here, we present a large-scale retrospective study of blood and cerebrospinal fluid (CSF) immune cell profiles of psychosis spectrum patients. We performed basic CSF analysis and multi-dimensional flow cytometry of CSF and blood cells from 59 patients with primary psychotic disorders (F20, F22, F23, and F25) in comparison to inflammatory (49 RRMS and 16 NMDARE patients) and non-inflammatory controls (52 IIH patients). We replicated the known expansion of monocytes in the blood of psychosis spectrum patients, that we identified to preferentially affect classical monocytes. In the CSF, we found a relative shift from lymphocytes to monocytes, increased protein levels, and evidence of blood-brain barrier disruption in psychosis. In fact, these CSF features confidently distinguished autoimmune encephalitis from psychosis despite similar (initial) clinical features. We then constructed machine learning models incorporating blood and CSF parameters and demonstrated their superior ability to differentiate psychosis from non-inflammatory controls compared to individual parameters. Multi-dimensional and multi-compartment immune cell signatures can thus support the diagnosis of psychosis spectrum disorders with the potential to accelerate diagnosis and initiation of therapy.

CSF plasmablasts differentiate MS from other neurologic disorders

Multiparametric flow cytometry (FC) of CSF allows one to easily estimate the percentage of lymphocyte subpopulations in CSF. We hypothesized that an increased ratio of B-lineage cells in CSF of MS patients, as assessed with FC, could be useful for diagnostics. We analyzed CSF of 137 patients (70 MS, 24 infectious/autoimmune neurologic disorders (INDs), and 43 non-infectious/autoimmune neurologic disorders (NINDs)), and showed that CSF plasmablasts of >0.1% had a sensitivity of 40% for MS and specificity of 92% when comparing MS and IND, while plasmablasts of >0.25% had sensitivity of 36%, and 100% specificity.

Revived Attention for Adult Craniopharyngioma

Craniopharyngioma as a rare tumor originating from cells of rathke's pouch and representing 2-5% of all intracranial tumors is a rare and generally benign neoplasm of the central nervous system with two incidence peaks one in childhood and one after 40 years of age. Data on adult patients is scarce compared to childhood onset tumors, however the burden of disease caused by the tumors and related treatment options is significant. Clinical symptoms range from headaches, visual disability, cranial nerve affection or hypothalamic symptoms (e. g. morbid obesity) to endocrine disorders. Most symptoms are related to tumor mass effect. The current standard of diagnostics is the determination of serum hormone levels and contrast enhanced magnetic resonance imaging often resulting in surgical treatment which holds a key role in all treatment concepts and should follow a hypothalamus sparing path. Radiation therapy may prove beneficial as adjuvant therapeutic option or in recurrent tumor, especially papillary tumors may be targeted using BRAF-600 inhibitors, targeted therapies for adamantinomatous craniopharyngioma have not yet reached a stage of clinical testing. Although prognosis regarding overall survival is favorable, life expectancy may be reduced due to the tumor itself as well as due to treatment effects. An important aspect especially in the adult population is the reduction in quality of life which is comparable to primary malignant brain tumors and metastases, calling for individual patient specific treatment approaches.