International Journal of Neurooncology

: 2020  |  Volume : 3  |  Issue : 2  |  Page : 63--67

Gliomas: The history of diagnosis and classification: Part 1

Amitava Ray1, Meenakshi Swain2, Rahul Lath3,  
1 Department of Neurosurgery; Department of Pathology, Apollo Hospitals, Hyderabad, Telangana, India
2 Exsegen Research, Hyderabad, Telangana, India
3 Department of Neurosurgery, Apollo Hospitals, Hyderabad, Telangana, India

Correspondence Address:
Dr. Amitava Ray
Third Floor, Nirvanaz, 8-2-293/82/A/240, Road 36, Jubilee Hills, Hyderabad - 500 033, Telangana


Over the last few years, there has been a shift from a 100-year-old tradition in classifying and grading brain tumors. Whereas traditional teachings were based on the observation of individual cells and identification of distinct recognizable patterns formed by these cells, modern classification demands the incorporation of molecular markers into the traditional systems of grading, causing considerable disruption in conventional histopathological approaches. The objective of this article is to highlight the advancements in grading and classification, to document the evolution of glioma prognostication, nil and nil the major events that marked turning points in its history. This review article is an exhaustive review of the literature that summarizes the changes in glioma classification over the last 100 years – from the early macroscopic inspections done at autopsy in the late 19th century to the first introductions of molecular markers by the World Health Organization. At this time of change, this article gives us a glimpse of how grading and classification has evolved as science has progressed, providing a glimpse into the future. Understanding the history of grading and classification is the key to better understanding the integrated histopathological and molecular approaches of today.

How to cite this article:
Ray A, Swain M, Lath R. Gliomas: The history of diagnosis and classification: Part 1.Int J Neurooncol 2020;3:63-67

How to cite this URL:
Ray A, Swain M, Lath R. Gliomas: The history of diagnosis and classification: Part 1. Int J Neurooncol [serial online] 2020 [cited 2021 Feb 26 ];3:63-67
Available from:

Full Text


Ever since the World Health Organization (WHO) 2016[1] guidelines were published, there has been a radical change in the methods for prognostication and classification of gliomas. Classification based on the ground-breaking treatise written by Bailey and Cushing in 1926[2] is now being challenged by a new system of classification built on the molecular basis of the disease. This change has inevitably presented challenges for the clinicians, not only in terms of learning about new methodologies and investigative methods, but also in terms of the keeping pace with constant change. In the 90 years between 2016 and Bailey and Cushing's seminal treatise, there had been four attempts by the WHO to classify gliomas and unify the descriptive nomenclature of brain tumors that would allow for scientific collaboration and comparisons of clinical outcomes. Although a universally acceptable histopathological classification was finally achieved in the WHO Classification published in 1993,[3] it was also the first time that chromosomal abnormalities and gene amplifications were mentioned in the classification. This article chronicles the history of glioma diagnosis and prognostication – from the first macroscopical attempts at diagnosis at autopsy through the next almost – 200 years ending in 1993 – describing along the way the emergence of histopathology, the contributions of immunohistochemistry (IHC), and the early attempts at molecular classification. The next article will cover the changes that have taken place in the classification since the Harlem Guidelines[4] in 2013 to the formation of cIMPACT and their subsequent recommendations.

 Gross Morphological Era

The first reports of a primary central nervous system neoplasm can probably be attributed to a description of a brain tumor at an autopsy as communicated by Berns in 1800.[5] Four years later in 1804, Abernety described diffuse astrocytic tumors that did not show a distinct border from the normal brain tissue. This was called medullary sarcoma by the British, but has been referred to as Fungus Medullare in German and Encephaloide by the French. Microscopic examination of tissue was still not commonplace in pathological practice [Figure 1].{Figure 1}

 Histopathological Era

The first histopathological description of gliomas can be credited to Virchow. In a lecture in 1858, he described a homogenous population of cells which he called glia, derived from the Greek word for glue, which he presumed to be the function of these cells. In the first detailed study, he demonstrated in 1865, malignant tumors arising from these glial cells that could be structurally separated from the normal tissue.[6] He also demonstrated the invasion of such cells in the brain parenchyma and coined the name “gliomas” to describe these lesions. Depending on their histological contrast to the parent tissue, Virchow separated gliomas into high and low grades – a distinction that is still made today.

The most important work in histopathology, however, can be credited to Percival Bailey, after he arrived at the Peter Brent Brigham Hospital in Boston to train under nil Cushing.[2] Spending his initial years in Boston investigating the posterior pituitary, he soon turned his attention to brain tumors, the classification of which was in its infancy. Though glial cells had been described by Virchow and in 1875 Golgi described star-shaped neurons in the brain, Cushing and Bailey both agreed that the histopathological classification of brain tumors was inadequate. They wanted to understand the structural variability that gliomas presented and its correlation with outcome and nil to refute the ubiquitous notion that microscopic examination could not predict the course of the disease. It was just before the start of this work on gliomas, that in March 1921, Dr. A. C. Broders published an article from the Mayo Clinic that studied and histopathologically graded 256 squamous cell carcinomas of the lip into four grades: Grade 1, resembling the parent tissue maximally and so being defined as being of low grade and good prognosis (though Dr. Broders was not aware of the prognosis at the time of grading) and those at the other extreme end of the spectrum or Grade 4, with multiple mitoses, deep and dark staining nucleoli.[7]

This was the first time that it was demonstrated that a pathological grading system was possible, that was completely independent of the clinical history. How much this article influenced Drs. Bailey and Cushing is hard to tell, but this article was widely discussed across the pathological world. Hence, in 1922, Bailey started work on the classification of gliomas, based on the exhaustive analysis of case records and microscopic examination of 414 gliomas cases that he collected from Cushing's clinical practice. The results were first presented at the Cameron Prize Lectures at the University of Edinburgh in 1925. A comprehensive monograph titled “A Classification of the Tumors of the Glioma Group on a Histogenetic Basis with a Correlated Study of Prognosis,” was published by Bailey and Cushing the next year, which completely changed the antiquated thinking and established that microscopic examination of the tumor was important in prognostication.[2] However, in a significant departure from Broders, the foundation for this classification was not the histological appearance but the prognosis of the disease. Thus, two systems were born – one of pathological grading and the other of prognostic classification. Though they did overlap to a large degree – the closer the lesion is to the microscopic appearance of the parent tissue, the lower the grade and better the prognosis and vice versa – this is not inevitably true. Inconsistencies in nomenclature arise when the two are in conflict. The molecular classification only adds to the confusion that already exists by adding another layer of stratification. Histopathological examination of gliomas has formed the basis of every classification since. Bailey and Cushing initially divided brain tumors into 13 categories based on prognosis, and later reduced it to 10, based on prognosis, and firmly believed that the most malignant form of brain tumors were spongioblastoma multiforme, which had a different cell of origin, and was characterized by atypical polymorphic gigantic cells that had no resemblance to the glial cells whatsoever.[8] The first descriptions of a “spongioblastoma multiforme” was probably nil credited to Drs. JH Globus and I Strauss. In 1916 they described a tumor with multiple “spongioblasts” and ran a short and fatal clinical course that was very different to the gliomas described by Virchow a few years earlier. Gliomas at this time were believed to be slow-growing, infiltrative vascular lesions that differentiated into neuroglial cells, i.e., cells that produced neurofilament. It was then believed that the primitive neuro-ectoderm differentiated into the neuroblast and the spongioblast – the former forming neurons and the latter glial cells. The large atypical cells with giant nuclei and a short clinical course made researchers believe that the cell of origin was not glial, but a precursor – known then as a spongioblast.[9]

With the departure of Bailey as the Head of the Neurosurgery section at the University of Chicago in 1928, the focus shifted across the Atlantic to Hans-Joachim Scherer, a Prussian neuropathologist born in the town of Bromberg, Poland.[10] A rather ambitious man, his life and philosophies do seem to be an assemblage of conflicting behavior. nil, arrested by the Gestapo in 1933, for his opposition to the regime a few months after Hitler came to power, nil, he was supposed to have become a loyalist and a Nazi sympathizer - in later life. Initially forced to flee to Antwerp and then Brussels, it is from where he published some of his most well-known works. Scherer noticed that there were a number of changes in the cells of mesenchymal origin – presence of a number of cells of mesenchymal origin which included mesenchymal cell reactions to the infiltrating tumor and the presence of tufts of blood vessels that mimicked glomerular formations [Figure 2]. He argued that the diagnosis must take into consideration the morphological formations that are unique to the tumor – he described rosettes, pseudo-rosettes, and also the pathognomonic palisading necrosis of glioblastoma that bears his name – Scherer Formation[11] [Figure 2]a. He also postulated that new vessel formations were an integral part of the pathological process, something he described as angiotaxis[12] [Figure 2]b. However, possibly his greatest contribution was the differentiation of glioblastomas into primary and secondary from extensive “biological and clinical” studies. This concept was unheard of at the time. In addition, he also attributed glioblastomas to arising from glial cells, a fact not accepted even in 1979, where it had been assigned to be arising from embryonal or primitive neuroepithelial cells.[13] His burgeoning career was cut short when he was killed in an Allied bombing raid on a train station in Landshut, Bavaria, in 1945.[10]{Figure 2}

While the work linking pathological observations to outcomes was popular in Europe, Kernohan continued to advance the work done by Broders and worked on a purely pathological classification.[14] He noted the classification of Bailey and Cushing where glial tumors were divided into astrocytomas, astroblastomas, polar spongioblastomas, and glioblastoma multiforme, noting that glioblastomas did not fit into the classification with the variegated appearance and multinucleate giant cells. Based on cell culture work done by Russell and Bland, Kernohan concluded that the polar spongioblastoma was nothing but a variant of astrocytoma. Based on the histological appearance, he classified astrocytomas into four grades – progressing from the least malignant to the most. He also established a continuum which suggested that based on the degrees of anaplasia and differentiation, these grades represented the same tumor group. Kernohan's classification was widely accepted by pathologists as a grading system, the only pitfall being the inclusion of pilocytic astrocytomas in this evolving continuum.

Following the end of the Second World War, and the formation of the WHO in 1948, attempts were made to classify all malignant tumors to allow easier comparisons from around the world.[15] To this end, a referral center was set up in Koln under the aegis of the WHO, in 1958, to classify and grade brain tumors. Acknowledging that differences existed in classification and grading, the referral center collected and distributed 230 tumors to eight collaborating centers around the world. In 1970 L. J. Rubenstein and K. J. Zülch were asked to develop a consensus for histopathological classification and grading. The collaborating center reports were noted and suggestions from the referral center were sent back to the collaborating center. Study groups were held in 1974 and 1976 to build consensus. At the end of the consultative process in 1978, “Babylonic differences” remained with respect to classification; there was a reasonable consensus when it came to grading. The committee agreed that the same tumor may have different nomenclature in different parts of the world as a significant proportion of tumors will contain cells of mixed lineage. Astrocytic tumors were divided into three distinct grades – the pilocytic astrocytoma being classified as Grade 1 – the diagnosis based on the presence of fusiform cells, Rosenthal fibers, and “granular” bodies. Grade 2 astrocytomas were similarly graded by three distinct morphological entities – fibrillary, gemistocytic, and protoplasmic. There was consensus on the definition of anaplasia and that it was associated with malignant biological behavior characterized by cellular pleomorphism, increased cellularity, greater mitotic activity, de-differentiation of abnormal vascular proliferation, and abnormal stromal reaction with necrosis with or without pseudo-palisading of nuclei. Anaplastic astrocytomas were characterized as Grade 3 tumors. To describe the tumors of tuberose sclerosis, the term subependymal giant cell astrocytomas was coined. Oligodendrogliomas were classified as Grade 2, as were oligo-astrocytomas – exhibiting histopathological characteristics of both astrocytomas and oligodendrogliomas. Anaplastic oligodendrogliomas were of Grade 3. Similarly, ependymomas were classified as being characterized by ependymal rosettes, perivascular rosettes, ependymal canals, and blepharoplasts, and were classified as Grade 2. Anaplastic ependymomas were of Grade 3. Extra-ventricular ependymomas were also recognized – myxopapillary ependymomas of the filum (Grade 1), papillary ependymoma, and subependymoma. Glioblastomas, however, were not considered to be astrocytic in origin and was classified separately under “poorly differentiated and embryonal tumors.” The term gliosarcoma was also considered if the stroma showed hyperplastic elements that resembled sarcomatous change.

 Immunohistochemical Era

The principle of IHC has existed since the 1930s, but it was not until 1941 that the first IHC study was reported. Coons et al. used fluorescein isothiocyanate-labeled antibodies with a fluorescent dye to localize pneumococcal antigens in infected tissues.[16] It is not until the early 1970s though that IHC was being thought of as a technique that could be used in tumor pathology. One of the main reasons for that was the preservation of tissue in a 10% formalin solution which was thought to significantly alter the antigens of interest. A change in the methods of preservation would only lead to discarding the vast amounts of historical material. In addition, formalin was very good for the preservation of precise anatomical architecture, something that was very useful to the diagnostic process.[17] The ability to manufacture monoclonal antibodies that are specific to the altered antigens and the ability to unmask antigens using enzyme digestion or heat greatly improved its efficacy.

In 1971, upon the isolation of filaments from fibrous astrocytes, the 50 kDa Type III intermediate filament glial fibrillary acidic protein (GFAP) was identified and characterized. One year later, high expression of GFAP in glioma with astrocyte characteristics astrocytoma, was described for the first time which still forms the basis of glioma IHC. While IHC was being incorporated into the histopathological study of tumors, the second “Blue Book” of WHO classification of brain tumors published in 1993[3] was able to get a consensus on classification and grading. Based on a ground-breaking article in 1988 by Daumas-Duport,[18] the classification of gliomas was dependent on the recognition of specific histopathological observations, which could directly correlate with the grade of the tumor – completely circumscribed tumors were of Grade 1, nuclear atypia made it Grade 2, mitoses Grade 3, and the presence of all the previous features with endothelial proliferation and necrosis – Grade 4. Not only his classification went a long way in removing the ambiguity of classification of the lower grade tumors, but also demonstrated that the outcomes could also be more accurately predicted [Figure 3] and [Figure 4].{Figure 3}{Figure 4}

Noting the gradual progression of diffuse infiltrative gliomas to glioblastomas and the presence of similar genetic mutations like that in p53 and LOH in 17p, 19q, and chromosome 10, it was agreed, though not unanimously, that glioblastomas belong to the astrocytic lineage. Though it was contended that some oligodendroglial and ependymal tumors also progressed to glioblastomas, it was agreed that most of these tumors showed astrocytic differentiation as well. As the percentage of P53-positive tumors did not increase significantly in glioblastomas when compared to low-grade gliomas, it was correctly presumed that this event occurred early on in gliomagenesis. It was also recognized that glioblastomas could arise de novo, where it noted the mutations in epidermal growth factor receptor seen.[3] Thus, for the first time, molecular markers were described in the classification.


The first 200 years in glioma pathology provided the basis of future classifications and the yardstick against which all glioma classifications are judged. It is really a tribute to the innovations and originality of the clinicians and scientists of the early 20th century that even a 100 years on, they form the pathological basis of disease. It is only the discovery of the role of IDH mutations in gliomas in the early 21st century, and the incorporation of this in the 2016 classification, did we see a first real change in direction, especially in low-grade gliomas. Predictions of improved survival in this “methylated” phenotype became more evident, something that was not possible by histopathological examination alone. However, even today, in the majority of IDH wild-type gliomas of the adult, the Bailey and Cushing classification still is as accurate in predicting outcome as is the whole genome studies of today. The following part of this two-part series deals with the changes that have taken place since 2016, how they affect current clinical practice, and what the future holds for prediction and treatment.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Wesseling P, Capper D. WHO 2016 Classification of gliomas. Neuropathol Appl Neurobiol 2018;44:139-50.
2Bailey P, Cushing H. A Classification of the Tumors of the Glioma Group on a Histogenetic Basis with a Correlated Study of Prognosis. Philadelphia: Lippincott; 1926.
3Kleihues P, Burger PC, Scheithauer BW. The new WHO classification of brain tumours. Brain Pathol 1993;3:255-68.
4Louis DN, Perry A, Burger P, Ellison DW, Reifenberger G, von Deimling A, et al. International Society of Neuropathology-Haarlem Consensus Guidelines for Nervous System Tumor Classification and Grading: ISN-Haarlem Brain Tumor Classification Guidelines. Brain Pathol 2014;24:429-35.
5Stoyanov GS, Dzhenkov DL. On the concepts and history of glioblastoma multiforme-morphology, genetics and epigenetics. Folia Med (Plovdiv) 2018;60:48-66.
6Virchow R. Krankhafte geshwülte. Vorlesungen 1864;2:129.
7Broders AC. Squamous-cell epithelioma of the skin: A study of 256 cases. Ann Surg 1921;73:141-60.
8Globus JH. Spongioblastoma multiforme: A primary malignant form of brain neoplasm: Its clinical and anatomic features. Arch Neurol Psychiatry 1925;14:139.
9Svien HJ, Mabon RF. A simplified classification of the gliomas, based on the concept of anaplasia. Surg Clin North Am 1949;29:1169-87.
10Peiffer J, Kleihues P. Hans-Joachim Scherer (1906-1945). Pioneer in glioma research. Brain Pathol 2006;9:241-5.
11Scherer H. Die bedeutung des mesenchyms in gliomen. Virchows Arch 1933;291:321-42.
12Scherer H. Gliomstudien III. Angioplastische gliome. Virchows Arch 1935b; 823-86.
13Scherer H. Cerebral astrocytomas and their derivatives. Am J Cancer 1940;159-98.
14Kernohan JW, Mabon RF. A simplified classification of the gliomas. Proc Staff Meet Mayo Clin 1949;24:71-5.
15Zülch KJ. Principles of the new World Health Organization (WHO) classification of brain tumors. Neuroradiology 1980;19:59-66.
16Coons A, Creech H, Jones R. Immunological properties of an antibody containing a fluorescent group. Proc Soc Exp Biol 1941;47:200-2.
17Duraiyan J, Govindarajan R, Kaliyappan K, Palanisamy M. Applications of immunohistochemistry. J Pharm Bioallied Sci 2012;4:S307-9.
18Daumas-Duport C, Scheithauer B, O'Fallon J, Kelly P. Grading of astrocytomas. A simple and reproducible method. Cancer 1988;62:2152-65.