Neuropathologist calls into question the capacity of small centers to adequately diagnose brain tumors in the molecular age

With the emphasis on molecular diagnostics in the new WHO classification of tumors, an Austrian neuropathologist has called into question the capacity of smaller community hospitals, which don't have access to molecular techniques, to provide adequate diagnoses for many primary brain tumors. In an interview with the Science Daily website, Viennese neuropathologist Johannes Hainfellner (pictured) states: "Modern, advanced medical diagnosis of brain tumors relies heavily on the expertise of academic neuropathology. For this reason, smaller centers treating smaller numbers of cases and with standard pathology facilities without the benefit of university neuropathology have neither the requisite capacity nor the necessary routine." Hainfellner was a contributor to the new WHO book and has a research focus on biomarker validation and translation at the Medical University of Vienna.

Immunohistochemical surrogate for BRAF V600E mutation

Quoted highlights  from: Tanboon J, Williams EA, and Louis DN. The Diagnostic Use of Immunohistochemical Surrogates for Signature Molecular Genetic Alterations in Gliomas. J Neuropathol Exp Neurol Vol. 75, No. 1, January 2016, pp. 4–18:

- The most common BRAF alteration is an activating mutation caused by a substitution of valine for glutamic acid at codon 600 (BRAF V600E) in exon 15.

- BRAF V600E status can be screened for using the mutation-specific BRAF V600E immunohistochemistry clone VE1... Only strong homogeneous unambiguous cytoplasmic staining should be interpreted as positive.

- While BRAFV600E can be found in a wide variety of brain tumors, BRAF fusions... are mostly limited to pilocytic astrocytomas (PAs).

- For tumor classification, aBRAF V600E mutation in a ganglion cell-rich lesion biopsied from infratentorial region favors ganglioglioma over PA.

- BRAF V600E mutations are present in pleomorphic xanthoastrocytoma (PXAs) with and without anaplasia (65%–70%), ganglioglioma (33%–60%), and [occasionally] PA, especially extracerebellar PA (6%–8%).

- In addition, BRAFV600E mutations are common in epithelioid glioblastomas (eGBM) (54%), which also tend to occur in children and relatively young adults.

- BRAF V600E mutations have also been reported in a small number of desmoplastic infantile astrocytomas, desmoplastic infantile gangliogliomas, and dysembryoplastic neuroepithelial tumors.

- A few diffuse astrocytomas in children and adults harbor BRAF V600E mutation; the tumors in adults show unusual histologic features such as partly circumscribed portions and spindle cells, and may associated with more favorable prognosis.

- Strong cytoplasmic staining patterns are often observed in PXA and eGBM; on the other hand, staining can be variable in glioneuronal tumors because BRAF can be diffusely positive in the tumor cell population or it can be limited to either a glial or neuronal component. For example, in ganglioglioma, mutant BRAF is predominantly expressed in neuronal tumor cells

- In glioneuronal tumors, BRAF V600E mutation is associated with activation of the mammalian target of rapamycin (mTOR) pathway, worse postoperative seizure outcome, and shorter recurrence-free survival.

Immunohistochemical surrogate for H3 K27M mutations

Quoted highlights on IHC mutational surrogates from: Tanboon J, Williams EA, and Louis DN. The Diagnostic Use of Immunohistochemical Surrogates for Signature Molecular Genetic Alterations in Gliomas. J Neuropathol Exp Neurol Vol. 75, No. 1, January 2016, pp. 4–18:

- K27M mutations in H3F3A [the gene encoding H3.3] and HIST1H3B [the gene encoding H3.1] occur in approximately 40%–80% of pediatric diffuse intrinsic pontine gliomas (DIPG) and 20% of nonbrainstem glioblastomas.

- The central role of these mutations in such tumors will result in these neoplasms being designated as “diffuse midline glioma, H3 K27M-mutant” in the 2016 WHO Classification of Tumours of the Central Nervous System

- The presence of K27M mutations in pediatric glioblastomas (including DIPG) is associated with shorter survival compared to the wild-type tumors... [T]he presence of K27M mutations can be used as a diagnostic marker and poor prognostic marker for pediatric high-grade astrocytoma.

- K27M mutations in both the H3.3 and H3.1 histones can be detected by immunohistochemistry using an anti-H3K27M antibody. Positivity can therefore be useful to establish a diagnosis of infiltrating glioma in lower-cellularity biopsies, and can be used to subtype and genotype these diffuse gliomas. 

- The pattern of positivity is nuclear, with positivity in most of the tumor cells; nontumor cell nuclei are negative.

Immunohistochemical surrogate for ATRX mutation

Quoted highlights on IHC mutational surrogates from: Tanboon J, Williams EA, and Louis DN. The Diagnostic Use of Immunohistochemical Surrogates for Signature Molecular Genetic Alterations in Gliomas. J Neuropathol Exp Neurol Vol. 75, No. 1, January 2016, pp. 4–18:

- Loss of nuclear staining for ATRX protein by immunohistochemistry has been used as a surrogate marker for ATRX mutations.

- Consistent and strongly positive staining in the nuclei of nonneoplastic endothelial cells and neurons is commonly used as internal positive controls.

- Unfortunately, there are no standard criteria (in terms of number of cells) for what constitutes loss of ATRX staining in gliomas.

- Because ATRX mutations are uncommon in IDH-mutant tumors with 1p/19q codeletion, it is suggested that ATRX, along with TP53 mutations, can be used as markers of astrocytic lineage.

- In adults, ATRX mutations have been reported either by sequencing or immunohistochemistry in 45%–67% of diffuse astrocytomas, 57%–73% of anaplastic astrocytomas, and 33%–57% of secondary glioblastoma; ATRX mutations are uncommon in primary glioblastoma (4%–7%)

- In children, ATRX mutations have been reported in 22% of pediatric diffuse intrinsic pontine gliomas and 48% of nonbrainstem high-grade gliomas in children;when ATRX mutation occurs in children, patients tend to be over 11 years old.

Immunohistochemical surrogate for TP53 mutation

Quoted highlights on IHC mutational surrogates from: Tanboon J, Williams EA, and Louis DN. The Diagnostic Use of Immunohistochemical Surrogates for Signature Molecular Genetic Alterations in Gliomas. J Neuropathol Exp Neurol Vol. 75, No. 1, January 2016, pp. 4–18:

- Because they are rare in nonneoplastic brain lesions, TP53 mutations can be used as a marker to differentiate glioma from gliosis.

- TP53 alterations ... are essentially mutually exclusive with 1p/19q codeletion.

- Tumorigenic TP53 mutations have been reported to be present in > 50% of gliomas with astrocytic features, including 59%–74% of diffuse astrocytomas, 53%–65% of anaplastic astrocytomas, and 62%–65% in secondary glioblastomas. In contrast, these mutations are less common in gliomas with oligodendroglial features (9%–44%), and in primary glioblastomas (23%–28%).

- Intense nuclear staining for p53 protein by immunohistochemistry in a substantial percentage of tumor cells has long been used as a surrogate marker for TP53 mutations. The underlying mechanism is abnormally elongated half-lives for the protein products of the most common TP53 mutations in gliomas. 

- Strong p53 nuclear positivity in > 10% of the tumor cells is the most accurate predictor for TP53 mutations in gliomas.

- Positivity of p53 immunohistochemistry staining can occur in [non-neoplastic] conditions of cellular stress, [causing false positive results].

Immunohistochemical surrogate for IDH1 mutation

Quoted highlights on IDH mutation IHC from: Tanboon J, Williams EA, and Louis DN. The Diagnostic Use of Immunohistochemical Surrogates for Signature Molecular Genetic Alterations in Gliomas. J Neuropathol Exp Neurol Vol. 75, No. 1, January 2016, pp. 4–18:

- IDH1 and IDH2 mutations are mutually exclusive events and indicate one of the early processes in gliomagenesis, before TP53 and ATRX mutations in astrocytic tumors, and before 1p/19q codeletion, CIC, and FUBP1 mutations in oligodendroglial tumors

- IDH mutations exist in at least 70% of diffuse gliomas, particularly World Health Organization (WHO) grade II and III astrocytomas, oligodendrogliomas, and secondary glioblastomas, and are rarely present in other types of brain tumors

- Clinically, patients with either IDH1 orIDH2 mutations are younger and have a better prognosis in terms of both overall survival and progression-free survival compared to patients carrying wild-type IDH

- Intriguingly, recent studies reveal similar age of onset and little differences in clinical outcome among IDH-mutant tumors previously classified as grade II and grade III astrocytomas by WHO 2007 criteria

- The “good effect” of having IDH mutation also applies to glioblastomas since patients with IDH-mutant glioblastomas have better clinical outcomes compared to those with grade III astrocytomas having wild-type IDH...

- The presence of IDH mutations may argue in favor of a diagnosis of anaplastic glioma over primary glioblastoma given that the latter typically does not harbor the mutation

- The most useful antibodies detect the common mIDH1 R132H mutation, which is present in 90% of IDH-mutant gliomas

- Immunohistochemistry for mIDH1 R132H clone H09 shows 88%–100% concordance rate with IDH1 R132H mutational status determined by DNA sequencing

Brain Tumor Rhapsody by Dr. Arie Perry

Arie Perry, MD

I have written about Dr. Arie Perry's incredible musical talent before, and how he has applied it to neuropathology education. Well, he's brought that talent to a whole new level through a collaboration with the San Francisco Bay Area's vocal ensemble Musaic.  With Virchow as his muse, Arie outlines major biomarkers in the diagnosis of CNS tumors. This magnum opus is called Brain Tumor Rhapsody, and its Dr. Perry's first-ever educational music video.This is an epic musical and neuropathological achievement! Many thanks to Dr. Gabrielle Yeaney of the Cleveland Clinic for alerting me to this remarkable video. Check it out on YouTube!

Gabrielle A. Yeaney, MD

The Tumor Biomarker Series: INI1

This is the last in my tumor biomarker series -- at least until future significant biomarkers are established. I conclude this series with a short description of INI1, a marker for atypical teratoid/rhabdoid tumor (AT/RT). A clinically aggressive embryonal tumor of infancy, AT/RT is characterized by mutations in SMARCB1/INI1 (HSNF5). Immunohistochemical evaluation of AT/RT for the INI1 protein using the BAF47 antibody shows a loss of labelling in tumor cell nuclei, with retention of staining in internal positive control cells such as endothelial cells. Since AT/RT has morphologic overlap with medulloblastoma, CNS PNET, choroid plexus carcinoma, GBM, and other malignant tumors of childhood, INI1 immunohistochemistry is extremely useful in arriving at a diagnosis of AT/RT. A diagnosis of AT/RT carries implications for genetic counseling as this tumor -- in about a one-third of cases -- is a component of the rhabdoid tumor predisposition syndrome (RTPS) wherein there is a germline mutation of SMARCB1/INI1. Because of the risk associated with RTPS, the germline status of SMARCB/INI1 is typically assessed for each new case of AT/RT.

The Tumor Biomarker Series: Medulloblastoma Markers

Four subgroups of medulloblastoma have been defined based on genetic alterations:

Wingless (WNT) - WNT medulloblastomas display monosomy 6 and most show nuclear accumulation of the WNT pathway protein beta-catenin, which serves as a useful immunohistochemical screen for this group. Medulloblasomas with more than 50% nuclear staining for beta-catenin have been shown to have WNT pathway activation, whereas those with only focal nuclear staining do not. Overall survival for WNT medulloblastomas are dramatically longer than those of other subtypes, and clinical practices surrounding the treatment of this subtype reflects this better prognosis.

Sonic Hedgehog (SHH) - SHH medulloblastomas often show a nodular/desmoplastic histopathology and are associated with a better prognosis in younger children and infants. 9q deletion is characteristic, and MYCN amplifications are occasionally noted. GAB1 is expressed in the cytoplasm of nearly all SHH medulloblastomas but not in other groups and can be detected imunohistochemically, making it a valuable SHH-group marker.  Targeted therapies directed at this subgroup have been established and are entering clinical practice.

"Group 3" - Group 3 medulloblastomas have the worst overall prognosis, have a high incidence of large cell/anaplastic histology, and are very frequently metastatic. This group contains the vast majority of MYC amplified tumors, with MYC amplification being a strong negative prognostic factor. It has been suggested that Group 3 tumors should perhaps be re-named MYC medulloblastomas, but wide agreement has not been reached on this designation. Group 3 tumors occur more commonly in males than females, and are found in infants and children, but almost never in adults.


"Group 4" - Group 4 medulloblastomas classically harbor isochromosome 17q; but as the molecular pathogenesis of this group is not currently clear, the generic name "Group 4" remains the consensus designation. Although isochromosome 17q is also seen in Group 3 tumors, it is much more common in Group 4. KCNA1 has been suggested as an immunohistochemical marker for this group, but this requires validation.  The only other notable cytogeneic change seen in Group 4 tumors is loss of X chromosome, which is seen in 80% of females with this tumor subtype. Group 4 patients have an intermediate prognosis, similar to patients with SHH tumors.

In conclusion, the worst prognosis is associated  with Group 3 medulloblastoma; Group 4 and SHH have an intermediate prognosis; and WNT medulloblastoma tends to have the best prognosis.

The Tumor Biomarker Series: 1p/19q co-deletion

Allelic losses on chromosomes 1p and 19q are associated with oligodendroglial phenotype. Most studies have indicated that combined loss of 1p and 19q are specific to oligodendrogliomas, with only a few astrocytomas and a small subset of oligoastrocytomas harboring these alterations. Those oligodendrogliomas with 1p/19q loss show enhanced response to chemotherapy and are associated with prolonged survival. Solitary losses of 1p or 19q are also occasionally noted within an infiltrating glioma, but are not as strongly linked to the oligodendroglioma histologic phenotype and are not predictive of enhanced response to therapy or prolonged survival. Co-deletion of 1p/19q is highly associated with the IDH1 mutation, with over 80% of 1p/19q co-deleted oligodendrogliomas also carrying the IDH1 mutation.

The Tumor Biomarker Series: IDH

Mutations in two isoforms of isocitrate dehydrogenase (IDH1 and IDH2) are relevant to clinical practice. Mutations in IDH1 are frequent (70-80%) in WHO grade II & III astrocytomas, oligodendrogliomas, oligoastrocytomas, as well as secondary glioblastomas. Mutations in IDH2 have also been detected in these tumor types, but far less frequently. The finding of an IDH mutation in an infiltrating glioma is associated with a substantially improved prognosis, grade for grade. Indeed, IDH-mutant GBMs are associated with longer survival time than IDH wild-type anaplastic astrocytomas! More than 90% of IDH1 mutations involve a base exchange of guanine to adenine within codon 132, resulting in an amino acid change from arginine to histidine (R132H). The resulting protein alteration can be detected immunohistochemically. Of course, other mutations in IDH1 and mutations in IDH2 cannot be detected in this manner, but protein sequencing can be used to pick up these less common mutations if necessary.

The Tumor Biomarker Series: EGFR

Epidermal growth factor receptor (EGFR) is the most frequently amplified oncogene in astrocytic tumors (>40% or GBMs and 5-10% of anaplastic astrocytomas). EGFR is far more often amplified in de novo GBMs as compared to secondary GBMs. About one-half of those GBMs with EGFR amplification also have specific EGFR mutations (the vIII mutant), which produce a truncated receptor with constitutive activity. Both EGFR amplification and EGFRvIII mutant are mutually exclusive with IDH mutations. So, what is the utility of EGFR testing? First, astrocytomas with EGFR amplification tend to be of higher grade. So if, for example, the diagnostician is vacillating between a WHO grade II and a WHO grade III tumor, positive EGFR amplification status would favor the latter. Secondly, EGFR amplification can also help distinguish between a small cell GBM (which would potentially harbor the amplification) from anaplastic oligodendrogliomas (which do not exhibit the amplification).

The Tumor Biomarker Series: Ki-67

An immunohistochemical marker of cellular proliferation, the nuclear antigen Ki-67 is positive in cells that are actively engaged in cell cycle (i.e., not in G0). Results are expressed as a percent index of positively staining cells. Several studies have shows a correlation between Ki-67 indices in various astrocytomas, oligodendrogliomas, and mixed gliomas. Among grade II and grade III diffuse gliomas, the Ki-67 index provides prognostic value. However, investigations have consistently shown that Ki-67 proliferation indices have no prognostic value on patient outcomes for GBM. So, if you have an unmistakable GBM under your microscope, you are not practicing evidence-based parsimonious medicine by ordering Ki-67 immunohistochemistry on that tumor.  On the other hand, if you are debating between diagnosing a grade III or a grade IV astrocytoma, Ki-67 can be helpful in swaying your decision. The Ki-67 index is not used in the WHO grading system because of the high degree of technical variability between laboratories, making standardization difficult.

The Tumor Biomarker Series: ATRX

Alpha Thalassemia/Mental Retardation Syndrome X-linked (ATRX) is a gene that encodes a protein involved in chromatin remodeling. ATRX mutations are a marker of astrocytic lineage among the IDH-mutant gliomas and are mutually exclusive with 1p/19q codeletion. Present in 57% of secondary GBMs, ATRX mutations are uncommon in primary glioblastomas. Nearly all diffuse gliomas with IDH and ATRX mutations also have TP53 mutation and are associated with the Alternative Lengthening of Telomeres (ALT) phenotype. Immunohistochemistry for ATRX demonstrates loss of protein expression in neoplastic cells harboring the inactivating mutations, while expression is retained within non-neoplastic internal controls (such as endothelial cells).

The Tumor Biomarker Series: PTEN and LOH of chromosome 10

Loss of heterozygosity (LOH) of chromosome 10 occurs in most GBMs and less frequently in grade II and III diffuse astrocytomas. The phosphatase and tensin (PTEN) gene at 10q23.3 has been most strongly implicated as a glioma-related tumor suppressor on chromosome 10q, with PTEN mutations identified in about 25% of GBMs and less frequently in grade III astrocytomas. Losses of chromosome 10 and mutations in PTEN are considered to be specific for astrocytic differentiation and are rare in oligodendrogliomas. They are also markers of high-grade progression and aggressive clinical behavior in astrocytomas.

The Tumor Biomarker Series: BRAF

Aberrant constitutive activation of BRAF tends to be seen in cerebellar and midline pilocytic astrocyomas whereas the activating point mutation at BRAF V600E is more likely to be seen in cerebral examples. The V600E point mutation is also observed in other low-grade gliomas and glioneuronal neoplasms, including approximately two-thirds of pleomorphic xanthoastrocytomas, and lower percentages of ganglioglioma, desmoplastic infantile ganglioglioma, dysembrioplastic neuroepithelial tumor, and papillary craniopharyngioma. Although less common, diffusely infiltrative gliomas including glioblastoma, particularly the epithelioid variant, may also demonstrate the V600E point mutation -- making this biomarker potentially less useful as a diagnostic tool in distinguishing low-grade gliomas from high-grade ones.

The Tumor Biomarker Series: MGMT

MGMT stands for O6-methylguanine-DNA methyltransferase. I must admit that this is my favorite biomarker only because of its cool mechanism of action.  The standard chemotherapy for gliomas is temozolomide, which acts by cross-linking DNA through alkylating multiple sites including the 06 position of guanine. Crosslinking at this site is reversed by the DNA repair enzyme MGMT. Thus, low levels of MGMT activity by GBM cells is associated with enhanced response to alkylating agents such as temozolomide. To a large degree, the activity level of MGMT is determined by the methylation status of the gene's promoter. MGMT can be epigenetically silenced by hypermethylation. About half of all GBMs are epigenetically silenced in this manner and are therefore more susceptible to the alkylating action of temozolomide.  The methylation status of MGMT can be assessed by PCR-based testing. In addition to predicting better response to temozolomide, investigators have shown that epigenetic gene silencing of MGMT is a strong predictor of prolonged survival independent of treatment.

The Tumor Biomarker Series: TP53

A few month ago, the College of American Pathologists released a Template for Reporting Results of Biomarker Testing of Specimens from Patients with Tumors of the Central Nervous System. Therefore, I thought it would be a good idea to review as succintly as possible the various tumor biomarkers one could use to interrogate CNS tumors. Not all neuropathologists would agree as to which ones, if any, are essential. So, comments are most welcome! The first biomarker I'd like to address is TP53, mainly because I have doubts about it's utility -- except for cases where there is a question regarding whether or not an oligodendroglial component is present. However, immunohistochemical p53 testing is performed on virtually all high-grade astrocytomas at many institutions. Here's a summary of the CAP consensus description of the TP53 test:

Found in a majority of high-grade astrocytic tumors, TP53 mutation is rare in oligodendrogliomas. Mutation of TP53 is highly correlated with IDH mutation. As a surrogate for testing the actual TP53 mutation, p53 immunohistochemistry is typically performed. As for the utility of this test, the CAP template makes the following statement: "[T]here is a strong association between IDH1 mutation and TP53 mutation in diffuse astrocytomas and this combination of mutations is helpful in distinguishing astrocytomas from oligodendrogliomas."

My feeling is that if you have a histomorphologically classic pure astrocytic neoplasm, there is no need for p53 immunohistochemistry. And, yet, you almost always see p53 immunohistochemistry results on reports for classic glioblastomas, anaplastic astrocytomas, and infiltrative astrocytomas. I just don't get it. Even in cases where you have a question about the presence of an oligodendroglial component, testing for IDH1 mutation and 1p/19q deletion would be more helpful than testing for IDH1 and TP53. I would love to hear from people who could dissuade me of this opinion regarding the utility of p53 immunohistochemistry. Feel free to enter comments below.

Best Post of October '10 -- Army General on Blood Test for Concussion: "This is huge"

The next in our Best of the Month series is from October 15, 2011:

USA Today ran a story today about a simple blood test that the US Army has developed which may objectively test for the presence of concussion.   The Army collaborated with Banyan Biomarkers, a Florida-based company, to develop the test. In checking Banyan's website, it looks as though the test consists of a panel of immunoassays which include SBDP145, SBDP120, UCH-L1, MAP-2, GFAP. If this test turns out to be as good as the Army is implying, the implications for those on the battlefield (and on the football playing field, I might add) are enormous. If it pans out, I would agree with Gen. Peter Chiarelli, the Army vice chief of staff, who is quoted in the article as saying: "This is huge."

Thanks to Dr. Doug Shevlin for alerting me to this news story.

"This is huge."

USA Today ran a story today about a simple blood test that the US Army has developed which may objectively test for the presence of concussion.   The Army collaborated with Banyan Biomarkers, a Florida-based company, to develop the test. In checking Banyan's website, it looks as though the test consists of a panel of immunoassays which include SBDP145, SBDP120, UCH-L1, MAP-2, GFAP. If this test turns out to be as good as the Army is implying, the implications for those on the battlefield (and on the football playing field, I might add) are enormous. If it pans out, I would agree with Gen. Peter Chiarelli, the Army vice chief of staff, who is quoted in the article as saying: "This is huge."

Thanks to Dr. Doug Shevlin for alerting me to this news story.