A Global View on the Availability of Methods and Information in the Neuropathological Diagnostics of CNS Tumors: Results of an International Survey Among Neuropathological Units

In order to establish an accurate and reproducible diagnosis, neuropathologists have continuously refined their diagnostic criteria and applied methods. This process was reflected in the four different editions of the WHO classification of brain tumors 1. The current revision of the 4th edition of the WHO classification dates back to 2007 2; its update is being currently performed and will be published in 2016. Over the years, essential ancillary techniques were developed and validated such as immunohistochemistry (IHC) for epitopes with diagnostic potential and Fluorescent in situ hybridization (FISH) for analysis of genomic changes.

In the light of the upcoming revision to the WHO classification of brain tumors and the tremendous knowledge which has been accumulated on the genetics and biology of brain tumors, the crucial question is which markers and assays should be introduced as mandatory to lead to diagnosis of a brain tumor entity according to WHO classification in a balanced way, so that distinct brain tumor entities can be securely identified in (neuro)pathological departments in most countries. A controversial issue is the incorporation of more complex molecular techniques such as genome‐wide studies on the protein, mRNA or DNA level because such methods have shown to allow a more accurate diagnosis of brain tumors, and often better correlation with outcome but have not been approved for clinical decision making. 2

During the discussion on these developments raised at the XVIII International Congress of Neuropathology in Rio de Janeiro 2014, it was realized that there are no robust data available on the current status on the techniques used by (neuro)pathologists for the diagnosis of brain tumors worldwide. Therefore, we performed an international survey within the framework of the International Society of Neuropathology (ISN) to provide these data.

Methods

We prepared an online survey on google docs:https://docs.google.com/forms/d/15f05tfQl61YZ6r8M3xrPd6OTbJg2DczS_GTEw33VLok/viewform?usp=send_form.

To collect contact information of units involved in brain tumor diagnostics the ISN provided lists of national neuropathology societies, and their contact persons. More than 1600 e‐mails were sent to (neuro)pathologists from 68 countries. Recipients were informed that requested data refer to the status in individual neuropathological units (not to individual persons) and each unit should provide only one answer. Answers were collected from November 2014 to June 2015). As access to Google docs was not available in China, data from this country was received directly by e‐mail, with kind help from Dr Hiu Ming Li (Hong Kong).

The questionnaire covered the issues of diagnosed brain tumor cases per year, availability of IHC, antibodies used, molecular methods, diagnosis according to WHO classification, access to clinical data, access to imaging information and routine of freezing tumor samples. We correlated the responses to population data and human development index (HDI), provided by GLOBOCAN (http://globocan.iarc.fr/Default.aspx) and United Nations. http://hdr.undp.org/en Threshold for defining highly developed countries was above 0.78.

In order to estimate the representativity/relative number of brain tumor cases diagnosed by (neuro)pathologists from each country/year, we divided the estimated total number of cases diagnosed by all participants in a given country by the population of each country. Population data was gathered from the GLOBOCAN 2012 study or from national offices in Wikipedia. Calculation was performed as follows: The mean value of cases diagnosed/year was used to estimate the total number of cases diagnosed per center, and these were summed to obtain a total number of cases per country.

Results

Participation

A total of 314 responses from 48 countries were received. In order to evaluate the proportion of the cases covered by the survey we calculated a ratio between the mean total brain tumor cases diagnosed by (neuro)pathologists from all centers participating from each country/year, the estimated brain tumor incidence as published in GLOBOCAN. (Figure 1).

Figure 1.

The mean number of tumors diagnosed in different centers participating from each country was related to each country's estimated incidence of brain tumor. Written numbers account for number of responses per country. Color plots account for estimated percentage of total brain tumor cases/country covered by this survey. N= 314 centers.

Most centers (80%) were public hospitals/institutions/universities, and 14% were private hospitals/institutions/universities. The remaining 6% represented private labs or other institutions. Most of the participating units recorded an important diagnostic activity on brain tumors, half of the centers diagnosing between 101 and 500 brain tumor cases/year and 27% over 500/year. The great majority of centers diagnose mostly adult and a smaller percentage of pediatric cases, with only 5/314 centers (1.9%) performing exclusively pediatric cases.

Use of WHO classification of brain tumors

The vast majority of centers (79%) use the WHO guidelines systematically to define their (neuro)pathological diagnosis and 18% use it mostly with rare exceptions (Figure 2).

Figure 2.

The figure includes all 314 centers participating in the survey.

Use of immunohistochemical methods

In 47 of 48 countries from which centers responded, there was at least one center performing IHC (Figure 3). Only 5/314 responding centers declared not to perform IHC. Apart from one country (United Arab Emirates) all countries participating perform IHC in at least one center.

Figure 3.

GFAP, glial fibrilary acidic protein; EMA, epithelial membrane antigen; CK, cytokeratin; TTF1, thyroid transcription factor; PLAP, placental alkaline phosphatase; NF70; neurofilament70; Progesterone R, progesterone receptors; IDH, isocitrate dehydrogenase; BHCG, beta human chorionic gonadotropin; Hypoph horm, hypophyseal hormones; Bcat, beta catenin; NeuN, neuronal nuclei; EGFR, epidermal growth factor receptor; OCT, octamer‐binding transcription factor; OLIG2, oligodendrocyte transcription factor; MAP2, microtubule associated protein 2; PTEN, phosphatase and tensin homolog; HistoneH3K27me3, (trimethylated form of histone H3 on lysine 27); SALL4, Sal‐like protein 4; GAB, GRB2‐associated binder; YAP1, yes‐associated protein 1; LIN28A lin‐28 homolog A; MGMT prom methyl, O(6)‐methylguanine‐DNA methyltransferase promoter methylation, FISH, fluorescence in situ hybridization; QPCR, quantitative polymerase chain reaction; RT‐PCR, reverse transcription polymerase chain reaction, copy nb arrays, copy number arrays; CISH, chromogenic in situ hybridization; MLPA, multiplex ligation‐dependent probe amplification; CGH, comparative genomic hybridization.

Molecular genetic/epigenetic analysis of brain tumors

Of 314 centers participating, 235 (74.8%) declared to have access to molecular diagnosis of brain tumors in their daily practice. In 12/48 countries (25%) participants from all centers declared not to have access to molecular analytical techniques for the diagnosis of brain tumors, or declared to perform molecular techniques but none of the exams listed in the survey. However, molecular techniques are in principle available in some centers from four of these countries, but which apparently do not use these methods specifically for the diagnosis of brain tumors. No molecular techniques are available at all in any center in six countries (12.5%) (Figure 4).

Figure 4.

Numbers outside circles correspond to total centers participating from each country. Numbers inside circles correspond to total estimated annual cases of brain tumor concerned by the survey in each country. HDI= human development index. Plots are given for individual countries and for all countries and grouped countries according to IDH.

In the 12 countries not performing molecular diagnosis of brain tumors according to this survey, a total of 14 483 patients with brain tumors/year are affected if we take into account the estimated annual incidence of brain tumors of these countries/year

235/314 (75%) centers perform at least one of the molecular biology techniques listed in the survey. From these, FISH and/or CISH are the most widely used (216 centers, 69%). One hundred and ninety four (62%) centers are able to perform other techniques than CISH/FISH (Figure 4). Of note, 12% of all participants declared not to be acquainted with the methods for molecular diagnosis.

Specific molecular exams mostly performed are 1p/19q status (72% of all centers), MGMT promoter methylation status (53%), BRAF sequencing (50%) and IDH1 mutations (47%). (Figure 3).

Centers performing more exams/year are more often to be able to perform a molecular diagnosis. Moreover, the availability of molecular exams is correlated with the human development index, as defined by the United Nations (threshold = 0.78)

Discussion

Although the survey approach we used has caveats and limitations in several aspects it provides useful information. Most participating (neuro)pathological units in this survey were located in Europe and North America. This parallels active participation of (neuro)pathologists from these countries in congresses and publications in international journals. Being reachable by e‐mail and accepting and voluntarily participating in the survey also implies being somehow connected to the scientific community. This fact may introduce a bias of this survey toward units with better equipment and working conditions.

Interestingly in many countries with a relatively small population, a single or very few centers seem to be responsible for diagnosis of most brain tumors. The diagnostic activity in these countries is mostly centralized, and the implementation of novel techniques would be potentially easier than in large countries with multiple centers.

The Haarlem meeting guidelines have proposed a layered diagnosis to enable a morphological diagnosis also in centers where molecular techniques are not established, and also to provide a scaffold of a morphological diagnosis first, later complemented with a molecular/genetic information when the appropriate analyses become available, finally leading to an integrated diagnosis in entities which the WHO definition will include basic molecular information such as diffuse gliomas 3. However, it should be pointed out that many brain tumor entities could be diagnosed without any molecular diagnostics.

Besides diagnostic parameters, clinical neuro‐oncologists will demand for information on predictive markers if corresponding treatment options are available. In developing countries testing of MGMT methylation profile of a tumor will not be requested if alkylating agents will not be available for the treatment.

It is also not a surprise that countries with lower HDI have less access to molecular techniques, which may be due to different reasons. First, any additional method introduced into diagnostic procedures lead to higher costs, not covered in many countries. In addition, the training of (neuro)pathologists might not allow the proper implementation of molecular techniques. However, this is a continuous process and many techniques such as FISH analysis have already become part of diagnostic practice.

The heterogeneity of responses concerning sampling of frozen tumor material reflects the actual state, as some techniques including MLPA and sequencing can be performed on formalin fixed paraffin embedded material, while other tests specially based on the detection of RNA species such as KIAA1549‐BRAF fusions are still more efficiently performed on frozen material. Countries with higher HDI more often store frozen tumor material than countries with lower HDI, in accordance with the fact that a molecular diagnosis is more often available in countries with highest HDI.

This survey shows that most neuropathological units worldwide have access to IHC and already perform at least some form of molecular diagnostics of brain tumors. Careful implementation of additional analyses into the requirements for proper diagnosis of some tumor entities according to the upcoming revision of the WHO classification should be, therefore, manageable for most responding centers. The necessary implementation could potentially be improved through collaboration, for instance with the creation of molecular biology platforms on a national basis.

The implementation of training courses for neuropathologists and algorithms for standardization of techniques and their interpretation are of high importance. This would also be important, as it has been shown that training in neuropathology is highly heterogeneous in different parts of the world and highly influenced by historical and economical factors 4. These activities should be guided by the national and international neuropathological societies. Especially colleagues from developing countries might be enabled to participate in these activities by specific outreach programs and collaborations.