Cerebrospinal fluid cytology: a highly diagnostic method for the detection of diseases of the central nervous system

Introduction

The cerebrospinal fluid (CSF) is a clear and colorless extracellular aqueous liquid in the cavities of the central nervous system (CNS) (brain ventricles, central canal of the spinal cord) and the subarachnoid space in vertebrates. The subarachnoid space is a fluid-filled cavity that covers the brain and spinal cord and communicates with the ventricular system of the brain. It is delimited to the outside by the arachnoid membrane against the dura mater, and inwardly by the pia mater against the surface of the brain. It is permeated with arachnoid trabeculae, which contain loose connective tissue stroma. The arachnoid membrane and pia mater form the leptomeninges. The CSF protects the CNS against shock and pressure from the outside and also acts as a transport medium for a large number of soluble substances, such as electrolytes, glucose and immunoglobulins. It is formed in the ventricles by specially differentiated epithelial cells of the choroid plexus and the ependyma. As part of the CSF circulation, it drains from the ventricular system of the base of the brain into the subarachnoid space. The secretion rate is 500 mL/day. With a total volume of cerebrospinal fluid of approximately 150–170 mL, this means that the entire CSF is renewed every 5–7 h [1]. CSF is generally obtained for diagnostic purposes by puncture of the spinal canal at LV 3/4 or 4/5. It may also be collected by suboccipital puncture or ventricular drainage.

Significance of practical (conventional) CSF cytology in the overall spectrum of CSF diagnostics

After the collection of 3–15 mL CSF by way of lumbar puncture, the basic program of tests includes the analysis of the CSF, as well as the determination of the cell count, glucose and lactate in the CSF. In addition, the albumin and immunoglobulin quotients are analyzed and represented in what is known as a CSF/serum quotient diagram. Without going into detail on these parameters, it should be mentioned that they are mostly used in the diagnostics of inflammatory diseases affecting the CNS.

In addition to determining the cell count of leukocytes and erythrocytes in the CSF, a differentiated cytological examination of at least one, or better yet, two specimens should be performed as a matter of principle. The widespread practice of creating cytology specimens without an explicit request on the part of the sender – if at all – only above a certain number of cells can lead to fatal misinterpretation with serious consequences for the patient. For example, tumor cells or parasites can be detected in specimens with normal or slightly elevated cell counts (see below and [2, 3]).

The cytological examination of cerebrospinal fluid is a technically simple procedure that is quick to implement. It is cost-effective and, under diagnostic aspects, a very productive method, but does require considerable experience on the part of cytologists. Generally, CSF cytology plays a crucial role in the following [4, 5]:

• Detection and detailed mapping of pathogen-induced inflammatory diseases of the CNS

• Detection of acute and past bleeding into the CSF space

• Basic recognition and at least preliminary typing of cells of malignant tumors with respect to their origin

CSF cytology is also often used to evaluate therapeutic response, such as in the treatment of malignant tumors by means of chemotherapy.

As concerns the diagnostic significance of CSF cytology, the Reference Institute of Bioanalysis (RfB) conducted a web-based interlaboratory test 2 years ago to enable the laboratory to perform regular checks on its own experience in evaluating cytological CSF specimens and, thus, to continuously improve the quality of CSF diagnostics.

Pre-analysis in CSF cytology

CSF punctures are difficult for patients, especially due to the post-puncture low CSF headache syndrome (headaches occurring within the first 5 days following a puncture), nor can they be repeated any number of times. Accordingly, particularly stringent requirements must be in place for the careful processing of samples at the pre-analysis stage. CSF is usually collected by means of lumbar puncture between the third and fourth or between the fourth and fifth lumbar vertebrae. Where possible, sterile polypropylene tubes should be used as sample containers. Glass tubes cause cells to adhere to the container surface, thus falsifying the cell count [6]. EDTA or sodium fluoride tubes are not suitable for a cytological diagnosis either. A minimum of 2 mL CSF is required for cell counts and the production of cytological specimens. But as much as 10 mL CSF may be recommended in connection with specific tests, such as for tumor cells.

As CSF generally has only a low buffering capacity as a result of its very small protein and cell contents, it is crucial for the CSF to be transported to the laboratory and processed there immediately after being collected by puncture. Otherwise, a cytological assessment will be rendered difficult or, in most cases, impossible due to the pH value that changes quickly from 7.32 to 7.36 (regular range in CSF) to alkaline levels of 7.8 and higher, as well as due to the concomitant autolytic changes [7]. A duration of 2 h between the puncture and the production of the cytological specimen is considered critical. If this cannot be done, the CSF may be set with buffered formalin at a ratio of 1:1. But this procedure should not be employed by way of routine because of inevitable artifacts.

Cell preparation (sedimentation) and staining of cell specimens

The cell specimen for cytological diagnostics must be produced within 2 h of the puncture. CSF cytology has become a routine method at CSF laboratories thanks to the application of the Pappenheim panoptic staining already commonly used for blood and bone-marrow cells (combined from eosin-methylene blue staining according to May-Grünwald and Azure-II-eosin staining according to Giemsa; MGG staining) [4]. Special stains, such as Gram, Prussian-blue and immunocytochemical staining, will be discussed in greater detail below.

Cell populations of a normal and activated CSF cell picture

A CSF sample obtained by lumbar puncture from a healthy patient generally contains two cell types, that is, lymphocytes and monocytes. The ratio of lymphocytes to monocytes is around 70–30. As the puncture needle passes through various tissues on its way into the subarachnoid space (skin, fat and connective tissue, striated muscle), and also comes into contact with the cartilage and bones of the spine, cells of these tissues can be found in cytological specimens. By the same token, a CSF sample taken from the ventricular system may contain choroid plexus epithelial cells, ependymal cells and fragments of cerebral parenchyma [5, 6].

Lymphocytes, at least in smaller quantities, are found in almost all CSF samples. These are small (diameter of 7–9 µm), relatively isomorphic cells. The chromatin of the cell nucleus is usually dense and homogeneous. There is only a small cytoplasm edge that is stained blue, somewhat pale, somewhat more intensively colored (Figure 1A, 1). Activated lymphocytes are a form of differentiation between normal lymphocytes and the plasma cell under inflammatory conditions. Compared to non-activated lymphocytes, they are larger (up to 25 µm in diameter) and have a wider cytoplasm edge as an expression of ribosomal immunoglobulin synthesis (Figure 1A, 2). Under normal conditions, plasma cells can never be detected in CSF. In other words, their presence is always indicative of an inflammation of the CNS. The nucleus of mature plasma cells is located eccentrically and contains chromatin with a predominantly granular structure. A crescent-shaped lightening of the cytoplasm around the nucleus is typical, but need not be present in every case (Figure 1B, 1).

Figure 1:

Cell populations of a normal and activated CSF cell picture.

(A) Normal (1) and activated (2) lymphocytes (MGG staining); (B) plasma cells (1) (MGG staining); (C) normal monocyte (MGG staining); (D) activated monocytes (MGG staining). Further explanations are included in the text. Reprinted by kind permission of labmed.

Monocytes have a diameter of 15–20 µm, a lobed or indented nucleus and blue-gray cytoplasm. Occasionally, smaller vacuoles are found there (Figure 1C). Many and/or larger vacuoles are already an indication of an activated state. In addition, activated monocytes, which are generally also larger than non-activated forms, often have a rounded nucleus (Figure 1D).

Pathological CSF cell findings in connection with inflammatory diseases of the CNS

Inflammatory diseases of the CNS usually lead to significant changes in the cellular composition of the CSF. Almost any kind of infection can affect the CNS: bacterial, viral and fungal infections, as well as parasitic infections caused by both protozoa and metazoa, which account for a significant proportion, particularly in tropical regions. Autoimmune diseases can also cause a significant cellular response within the CNS.

Acute infections and infections caused by fungi and protozoa often lead to a significant increase of polymorphonuclear (mainly neutrophils) cells in the cerebrospinal fluid (in 70% of bacterial meningitis >300 µL, in 40% even >2000/µL). It is not uncommon for intracellular bacteria to be discovered in the MGG stain (Figure 2A, *). Nevertheless, emergency protocols include an additional Gram stain if bacterial meningitis is suspected (Figure 2B, * to mark Gram-negative meningococci). This is then followed by increasingly activated lymphocytes, monocytes and also macrophages. By contrast, viral infections are characterized by a prevalence of activated lymphocytes at the early stage (Figure 2C, 1), and plasma cells to some degree (Figure 2C, 2). One may detect mitosis here, but this should not trigger a diagnosis of malignant tumor cells.

Figure 2:

Pathological CSF cell findings in connection with inflammatory diseases of the CNS.

(A) Predominantly granulocytic cell picture with intracellular bacteria (*) (MGG staining); (B) predominantly granulocytic cell picture with intracellular bacteria (*) (Gram stain); (C) lymphomonocytic cell picture with activated lymphocytes (1) and plasma cells (2) with viral infection (MGG staining); (D) eosinophilic meningitis with parasitic infection (MGG staining). Further explanations are included in the text. Reprinted by kind permission of labmed.

These typical cell pictures are associated with specific pathogens, but there are also exceptions: granulocytes may occur at a very early stage in connection with peracute viral infections, while borreliosis – even though it is a bacterial infection – presents a cell picture that is dominated by activated lymphocytes and monocytes [4].

Another characteristic cell picture emerges in the presence of eosinophils (eosinophilic meningitis), which must suggest primarily a parasitic infection (Figure 2D and [8]).

Subarachnoid hemorrhage (SAH)

Penetration of blood into the subarachnoid space and/or CSF, such as due to trauma, a ruptured aneurysm, intracerebral hemorrhage, or in connection with a tumor, causes a distinct cellular response in the leptomeninges and, thus, also in the CSF. The presence of blood as a foreign material leads to a chemical or aseptic meningitis. In this case, the cell count in the CSF may sometimes be as high as up to 1500/µL, which involves mainly granulocytes. The first reliable cytological indication of intravital bleeding (in other words, not artificial bleeding caused by the actual lumbar puncture) is the presence of erythrophages, that is, monocytes that have phagocytosed erythrocytes (Figure 3A, 1). After about 3–4 days, as a result of the degradation of hemoglobin, hemosiderin is detected in the cytoplasm of these phagocytes for the first time, which are then called siderophages (Figure 3B, 1). That this pigment contains iron, and that it does not involve a melanin pigment, such as in a tumor cell, can be proved by means of Prussian-blue staining (Figure 3C). But siderophages can be detected with sufficient certainty in routine MGG staining, so that it is not necessary to wait for the results of the Prussian-blue staining to confirm or rule out suspected SAH.

Figure 3:

Subarachnoid hemorrhage (SAH).

(A) Erythrophage (1) (MGG staining); (B) siderophage (1) (MGG staining); (C) siderophage (Prussian-blue staining); (D) siderophage with hematoidin crystal (*) (MGG staining). Further explanations are included in the text. Reprinted by kind permission of labmed.

The degradation of hemoglobin eventually produces iron-free hematoidin (crystallized bilirubin), which appears in the cytoplasm of the phagocytes around 8 days after the hemorrhage. These are diamond-shaped, yellowish to brown-yellow crystals that are often intracellular, but may also be extracellular after the degeneration of the macrophages (Figure 3D, *).

Thus, the sequence erythrophagocytosis – siderophages – hematoidin crystals indicates

• that intravital bleeding has occurred and that it does not involve artificial bleeding after the actual puncture

• the time at which the bleeding occurred. It should be kept in mind, though, that all three stages mentioned may co-exist in the case of recurrent bleeding [5, 7].

Malignant cells in the cerebrospinal fluid

The following is not to provide a classification of malignant cells in the CSF, but merely to illustrate some basic facts for everyday purposes. Apart from the serendipitous discovery of tumor cells (see below), there are two basic configurations in the analysis of CSF specimens for malignant cells: either a primary tumor is already known, or there is a clinical and/or radiological suspicion of a tumor having spread to the subarachnoid space without any known primary tumor. In the former case, CSF cytology is performed for the purposes of tumor staging and/or therapy monitoring, while in the latter case, cytology is to provide certainty and/or clarity about malignant cells in the CSF, as well as yield information about the primary tumor. Important cytological criteria of malignant cells are [3, 5]:

• Abnormal size with large, polymorphic nuclei (Figure 4B–C)

• Visible nucleoli (Figure 4D, *)

• Atypical mitoses

• High nucleus-plasma ratio (Figure 4A)

• Intensive basophilia of the cytoplasm (Figure 4D)

Figure 4:

Malignant cells in the cerebrospinal fluid.

(A) Spread of tumor cells in the CSF with highly malignant lymphoma (lymphomatous meningitis) (MGG staining); (B) tumor cell of glioblastoma (MGG staining); (C), (D) spread of tumor cells into the cerebrospinal fluid with breast cancer (carcinomatous meningitis). In (D), the prominent nucleoli (*) are noteworthy (MGG staining). Further explanations are included in the text. Reprinted by kind permission of labmed.

If a routinely-prepared specimen contains cells that meet one or several of these criteria, the specimen will definitely have to be shown to an experienced cytologist, possibly also to a pathologist or neuropathologist.

Figure 4D shows an example of most of the above criteria of malignant cells on the basis of the CSF of a 62-year-old patient, which was submitted without detailed clinical information. The CSF was inconspicuous in terms of quantification; the cell count was <5/µL (!). Two of the cyto-specimens, which had been prepared routinely, exhibited, surprisingly, the malignant cells shown, which were subsequently interpreted, based on the patient’s history, as carcinomatous meningitis in connection with clinically known breast cancer. This example illustrates clearly that the production of a cytological specimen should never be made dependent on the result of manual or automated cell counting. Furthermore, when a tumor is suspected, several additional unstained specimens should be produced for any necessary immunocytological tests so as to identify the tumor more accurately. This requires, however, that the submitter previously informed the laboratory about an existing suspicion of tumor and collected a sufficient CSF sample. In the case of a single, already MGG-stained specimen (if there is one at all), it is generally impossible to run any additional immunocytological tests, which means that the patient may have to undergo another puncture.

Summary

In order to provide answers to the most common questions in the context of CSF cytology, that is, identification and detailed mapping of pathogen-induced inflammatory diseases of the CNS, detection of acute and past bleeding into the CSF spaces, and the basic identification and, at a minimum, preliminary typing of cells of malignant tumors with respect to their origin, the following points must be observed in daily laboratory practice:

• In addition to determining the cell count of leukocytes and erythrocytes in the CSF, a differentiated cytological examination of at least one, or better yet, two specimens should be performed as a matter of principle. The production of a cytological specimen should never be made dependent on the result of manual or automated cell counting.

• A duration of 2 h between the puncture and the preparation of the cytological specimen is considered critical. Otherwise, an assessment of the specimens will no longer be possible due to the autolytic changes that have occurred.

• Siderophages can be detected with sufficient certainty in routine MGG staining, so that it is not necessary to wait for the results of the Prussian-blue staining to confirm or rule out suspected SAH.

• When a tumor is suspected, several additional unstained specimens should be prepared for any necessary immunocytological tests so as to identify the tumor more accurately.

If these fundamental aspects are observed, CSF cytology – a method that is easy to carry out and cost-efficient – can often yield an important contribution to the diagnosis of CNS diseases, based on the authors’ own experience.