Automated Analysis of Cerebrospinal Fluid Cells Using Commercially Available Blood Cell Analysis Devices-A Critical Appraisal

Routine CSF parameters as predictors of disease course in multiple sclerosis: an MSBase cohort study

BACKGROUND

It remains unclear whether routine cerebrospinal fluid (CSF) parameters can serve as predictors of multiple sclerosis (MS) disease course.

METHODS

This large-scale cohort study included persons with MS with CSF data documented in the MSBase registry. CSF parameters to predict time to reach confirmed Expanded Disability Status Scale (EDSS) scores 4, 6 and 7 and annualised relapse rate in the first 2 years after diagnosis (ARR2) were assessed using (cox) regression analysis.

RESULTS

In total, 11 245 participants were included of which 93.7% (n=10 533) were persons with relapsing-remitting MS (RRMS). In RRMS, the presence of CSF oligoclonal bands (OCBs) was associated with shorter time to disability milestones EDSS 4 (adjusted HR=1.272 (95% CI, 1.089 to 1.485), p=0.002), EDSS 6 (HR=1.314 (95% CI, 1.062 to 1.626), p=0.012) and EDSS 7 (HR=1.686 (95% CI, 1.111 to 2.558), p=0.014). On the other hand, the presence of CSF pleocytosis (≥5 cells/µL) increased time to moderate disability (EDSS 4) in RRMS (HR=0.774 (95% CI, 0.632 to 0.948), p=0.013). None of the CSF variables were associated with time to disability milestones in persons with primary progressive MS (PPMS). The presence of CSF pleocytosis increased ARR2 in RRMS (adjusted R2=0.036, p=0.015).

CONCLUSIONS

In RRMS, the presence of CSF OCBs predicts shorter time to disability milestones, whereas CSF pleocytosis could be protective. This could however not be found in PPMS. CSF pleocytosis is associated with short-term inflammatory disease activity in RRMS. CSF analysis provides prognostic information which could aid in clinical and therapeutic decision-making.

Basic CSF parameters and MRZ reaction help in differentiating MOG antibody-associated autoimmune disease versus multiple sclerosis

Background

Myelin oligodendrocyte glycoprotein antibody-associated autoimmune disease (MOGAD) is a rare monophasic or relapsing inflammatory demyelinating disease of the central nervous system (CNS) and can mimic multiple sclerosis (MS). The variable availability of live cell-based MOG-antibody assays and difficulties in interpreting low-positive antibody titers can complicate diagnosis. Literature on cerebrospinal fluid (CSF) profiles in MOGAD versus MS, one of the most common differential diagnoses, is scarce. We here analyzed the value of basic CSF parameters to i) distinguish different clinical MOGAD manifestations and ii) differentiate MOGAD from MS.

Methods

This is retrospective, single-center analysis of clinical and laboratory data of 30 adult MOGAD patients and 189 adult patients with relapsing-remitting multiple sclerosis. Basic CSF parameters included CSF white cell count (WCC) and differentiation, CSF/serum albumin ratio (QAlb), intrathecal production of immunoglobulins, CSF-restricted oligoclonal bands (OCB) and MRZ reaction, defined as intrathecal production of IgG reactive against at least 2 of the 3 viruses measles (M), rubella (R) and varicella zoster virus (Z).

Results

MOGAD patients with myelitis were more likely to have a pleocytosis, a QAlb elevation and a higher WCC than those with optic neuritis, and, after review and combined analysis of our and published cases, they also showed a higher frequency of intrathecal IgM synthesis. Compared to MS, MOGAD patients had significantly more frequently neutrophils in CSF and WCC>30/µl, QAlb>10×10-3, as well as higher mean QAlb values, but significantly less frequently CSF plasma cells and CSF-restricted OCB. A positive MRZ reaction was present in 35.4% of MS patients but absent in all MOGAD patients. Despite these associations, the only CSF parameters with relevant positive likelihood ratios (PLR) indicating MOGAD were QAlb>10×10-3 (PLR 12.60) and absence of CSF-restricted OCB (PLR 14.32), whereas the only relevant negative likelihood ratio (NLR) was absence of positive MRZ reaction (NLR 0.00).

Conclusion

Basic CSF parameters vary considerably in different clinical phenotypes of MOGAD, but QAlb>10×10-3 and absence of CSF-restricted OCB are highly useful to differentiate MOGAD from MS. A positive MRZ reaction is confirmed as the strongest CSF rule-out parameter in MOGAD and could be useful to complement the recently proposed diagnostic criteria.

Analysing cerebrospinal fluid with explainable deep learning: From diagnostics to insights

AIM

Analysis of cerebrospinal fluid (CSF) is essential for diagnostic workup of patients with neurological diseases and includes differential cell typing. The current gold standard is based on microscopic examination by specialised technicians and neuropathologists, which is time-consuming, labour-intensive and subjective.

METHODS

We, therefore, developed an image analysis approach based on expert annotations of 123,181 digitised CSF objects from 78 patients corresponding to 15 clinically relevant categories and trained a multiclass convolutional neural network (CNN).

RESULTS

The CNN classified the 15 categories with high accuracy (mean AUC 97.3%). By using explainable artificial intelligence (XAI), we demonstrate that the CNN identified meaningful cellular substructures in CSF cells recapitulating human pattern recognition. Based on the evaluation of 511 cells selected from 12 different CSF samples, we validated the CNN by comparing it with seven board-certified neuropathologists blinded for clinical information. Inter-rater agreement between the CNN and the ground truth was non-inferior (Krippendorff's alpha 0.79) compared with the agreement of seven human raters and the ground truth (mean Krippendorff's alpha 0.72, range 0.56-0.81). The CNN assigned the correct diagnostic label (inflammatory, haemorrhagic or neoplastic) in 10 out of 11 clinical samples, compared with 7-11 out of 11 by human raters.

CONCLUSIONS

Our approach provides the basis to overcome current limitations in automated cell classification for routine diagnostics and demonstrates how a visual explanation framework can connect machine decision-making with cell properties and thus provide a novel versatile and quantitative method for investigating CSF manifestations of various neurological diseases.

Technological features of blast identification in the cerebrospinal fluid: A systematic review of flow cytometry and laboratory haematology methods

BACKGROUND

Involvement of the central nervous system (CNS) by acute leukemias (ALs) has important implications for risk stratification and disease outcome. The clinical laboratory plays an essential role in assessment of cerebrospinal fluid (CSF) specimens from patients with ALs at initial diagnosis, at the end of treatment, and when CNS involvement is clinically suspected. The two challenges for the laboratory are 1) to accurately provide a cell count of the CSF and 2) to successfully distinguish blasts from other cell types. These tasks are classically performed using manual techniques, which suffer from suboptimal turnaround time, imprecision, and inconsistent inter-operator performance. Technological innovations in flow cytometry and hematology analyzer technology have provided useful complements and/or alternatives to conventional manual techniques.

AIMS

We performed a PRISMA-compliant systematic review to address the medical literature regarding the development and current state of the art of CSF blast identification using flow cytometry and laboratory hematology technologies.

MATERIALS AND METHODS

We searched the peer reviewed medical literature using MEDLINE (PubMed interface), Web of Science, and Embase using the keywords "CSF or cerebrospinal" AND "blasts(s)".

RESULTS

108 articles were suitable for inclusion in our systematic review. These articles covered 1) clinical rationale for CSF blast identification; 2) morphology-based CSF blast identification; 3) the role of flow cytometry; 4) use of hematology analyzers for CSF blast identification; and 5) quality issues. 9 /L, which is much lower than the original machine count and platelet transfusion was warranted.

DISCUSSION

1) Clinical laboratory testing plays a central role in risk stratification and clinical management of patients with acute leukemias, most clearly in pediatric ALs; 2) studies focused on other patient populations, including adults and patients with AML are less prevalent in the literature; 3) improvements in instrumentation may provide better performance for the classification of CSF specimens.

CONCLUSION

Current challenges include: 1) more precisely characterizing the natural history of AL involvement of the CNS, 2) improvements in automated cell count technology of low cellularity specimens, 3) defining the role of flow MRD testing of CSF specimens and 4) improved recognition of specimen quality by clinicians and laboratory personnel.

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.