Case Study 18

1
Case Study 18

• Craig Horbinski, M.D., Ph.D.

2
Question 1

• The patient is a 71 year-old male with an episode
  of speech disturbance that occurred two days
  PTA.  The patient had some confusion as well as
  some speech difficulty.  MRI with contrast was
  done.
• What do you see?

3
T1
4
T2
5
FLAIR
6
DWI
7
T1 with contrast
8
Answer

• Ring-enhancing mass lesion, increased surrounding
  FLAIR signal, no diffusion abnormality.

9
Question 2

• What is the significance of the contrast image?

10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
Answer

• Gadolinium is the contrast agent used.  In normal
  brain it will not cross the blood-brain barrier,
  so brain parenchyma will not light up.  Seeing
  enhancement like this suggests a breakdown of the
  barrier.

16
Question 3

• What does the FLAIR image mean?

17
Answer

• Fluid-attenuated inversion recovery (FLAIR) is a
  T2-weighted image that dampens ventricular (free
  water) CSF signal.  Compare it with the regular
  T2, which shows CSF as bright.  FLAIR is good at
  picking up and highlighting true intraparenchymal
  edema.  Thus, this mass is producing edema.

18
Question 4

• What is DWI?  (No, the answer is not driving
  while intoxicated.)  Why is it important in
  neuroradiology?

19
Answer

• Diffusion-weighted imaging (DWI) shows areas with
  restricted water diffusion as bright.  Such areas
  have reduced water diffusion because of acute
  cytotoxic edema, which traps water molecules
  inside swollen cells.  DWI is very helpful in
  identifying acute ischemia.  Bacterial abscesses
  will also produce cytotoxic edema in the
  surrounding non-necrotic tissue, so they can also
  be bright on DWI.  Old strokes, non-bacterial
  infections, tumors, contusions, and demyelinating
  diseases (like MS) are not associated with
  cytotoxic edema, so they are not usually
  hyperintense on DWI.  In this case, the lesion is
  not bright on DWI, so its probably not an acute
  infarct or bacterial abscess.

20
Question 5

• Given all this, whats the differential diagnosis?

21
Answer

• High-grade glioma or lymphoma.  Abscess,
  ischemia, or demyelinating diseases are far less
  likely (again due to the lack of signal intensity
  on DWI).

22
Question 6

• During the excision of the mass, the neurosurgeon
  calls you for an intraoperative consultation and
  wants to know, 1. If hes in the lesion 2. What
  the diagnosis is.
• So what is it?
• Cick the following links to view slides Frozen,
  Smear

23
Answer

• High grade glioma.  The tissue is markedly
  hypercellular, far more so than normal tissue, so
  hes definitely in the lesion.  There are a lot
  of cells with nuclear atypia suggesting a
  neoplastic process.  At the edges of the smear,
  processes are seen coming off the cells,
  suggesting a glioma.  Mitoses are a clue to its
  high-grade nature, but even if you cant spot
  any, the microvascular proliferation seen on the
  frozen section (which is whats causing the
  contrast enhancement seen earlier) is good enough
  evidence that this is a high-grade glioma, i.e.
  glioblastoma (grade 4/4).

24
Question 7

• The permanent sections of the case have arrived. 
  Whats your diagnosis?
• Click here to view slide.

25
Answer

• The tumor is hypercellular, with angulated
  atypical nuclei (suggestive of astrocytic origin)
  and endothelial proliferation.  Mitoses are rare,
  but even if you dont see them the microvascular
  proliferation (often erroneously called
  endothelial proliferation) is enough to push this
  to a glioblastoma multiforme (GBM), WHO grade 4.

26
Question 8

• What immunostains do we routinely order for these
  tumors?  Why?

27
Answer

1. Ki67, which labels the nuclei of all cells not in
   Go, i.e. those in some stage of cell division.  A
   higher percent of cells with positive nuclei
   means a more rapidly growing tumor.  Calling
   something a GBM with a low Ki67 proliferation
   index (i.e. less than 5) is unusual, as most
   will be at least 10.
2. Epidermal growth factor receptor (EGFR), which is
   one of the two classic molecular pathways by
   which glioblastomas arise.  GBMs strongly
   positive for EGFR arise de novo, i.e. are primary
   GBMs.
3. P53, which labels the tumor suppressor protein
   inside the nucleus.  GBMs with p53 mutations
   arose from a prior low-grade glioma (whether that
   low-grade glioma was clinically recognized or
   not), i.e. are secondary GBMs.  An interesting
   fact about p53 mutations is that they tend to
   render p53 ineffective rather than just shutting
   down p53 production.  Thus, the cell recognizes
   that it should not be dividing and produces p53
   to stop itself.  But since the p53 doesnt work,
   and the cell doesnt realize its a waste of time
   to produce bad protein, it keeps churning out
   p53.  On immunostains of such tumors the majority
   of cells will have strongly-staining nuclei. 
   Such a finding suggests a p53 mutation, which is
   the second main pathway of gliomagenesis.
4. Glial fibrillary acidic protein (GFAP), just to
   verify that the tumor is indeed glial in origin.

28
Question 9

• The immunostains have arrived.  What is your
  interpretation of these stains?
• Click the following links to view slides Ki67,
  EGFR, P53, GFAP

29
Answer

1. GFAP is moderately positive in the tumor cells,
   confirming its glial origin
2. EGFR shows strong and diffuse staining,
   suggesting a primary GBM (consistent with the
   patients age, as primary GBMs are more common in
   older patients).
3. Only about 30 of the cells are positive for p53,
   and theyre only weakly to moderately positive,
   so its not likely a p53-driven tumor.
4. The Ki67 proliferation index is a little
   subjective, but was estimated at up to 20, well
   within the range for GBMs.

30
Question 10

• In addition to the immunostains, here at UPMC we
  do an extensive molecular workup of gliomas,
  including fluorescence in situ hybridization
  (FISH) probing for the EGFR gene (on 7p12), p16
  (9p21), and 1p/19q (on 1p36 and 19q13,
  respectively).
• What does each of these results mean?

31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
Answer

1. The ratio of EGFR to the centromeric enumeration
   probe of chromosome 7 is very high (scored at
   30), when it should be a 11 ratio.  This means
   that there are multiple copies of the EGFR gene,
   a.k.a. amplification, as shown by the numerous
   red signals compared to about 2 green signals. 
   This explains why there is so much EGFR protein
   on immunostain.  Since EGFR is a well-described
   pathway for cell proliferation, this
   amplification likely explains why this tumor
   arose in the first place.
2. P16 is a well-known tumor suppressor gene.  In
   addition to its described role in
   melanomagenesis, it has been shown to be deleted
   in higher grade gliomas but not low grade
   gliomas.  Most of the cells in this image have 2
   green dots identifying 2 copies of chromosome 9,
   but no red dots where p16 should be.  Thus, there
   is a homozygous deletion of p16, meaning that
   these cells have lost an important brake on
   division.  It lends extra support to our
   diagnosis of GBM in this case.  In some cases
   where the grade is uncertain, seeing homozygous
   deletion of p16 suggests that the glioma will
   soon explode into a higher grade even if the
   histology doesnt allow us to explicitly diagnose
   it as such, well still call the clinician and
   tell them to watch this tumor closely.
3. Deletion of the whole arms of 1p and 19q is a
   classic hallmark of oligodendrogliomas.  If this
   deletion were present in this case, we might have
   to reclassify the tumor as an anaplastic
   oligodendroglioma (WHO grade 3), which has a
   longer survival expectancy than does GBM.  The
   1p/19q codeletion renders the tumor more
   vulnerable to chemotherapy for reasons not yet
   known.  In this case, there is a 11 ratio of
   red green signal in most cells, meaning no
   selective loss of either 1p or 19q.  There is,
   however, some hyperploidy of chromosome 19, a
   common finding in high-grade gliomas that has no
   prognostic or diagnostic significance.

36
Question 11

• Another set of molecular diagnostic tests we do
  includes PCR-based microsatellite analysis for
  selected targets, the results of which are shown
  below
• 1p fractional gene loss is 0 (0 markers with
  loss/ 5 informative markers)19q fractional
  gene loss is 67 (2 markers with loss/ 3
  informative markers)9p fractional gene loss is
  0 (0 markers with loss/ 3 informative
  markers)10q fractional gene loss is 100 (2
  markers with loss/ 2 informative markers)17p
  fractional gene loss is 0 (0 markers with loss/
  2 informative markers)
• The D19S112 and D19S559 markers on 19q were the
  ones showing loss of heterozygosity (LOH).The
  D9S1748 marker on 9p corresponding to the 9p FISH
  probe for p16 (9p21) did not show LOH.
• What is the significance of each of these loci? 
  How do you interpret these results?  In
  particular, how do you reconcile the FISH results
  on 19q with the PCR LOH results on 19q (look at
  the ideogram)?  Or the FISH on 9p21 with PCR on
  9p21?

37
(No Transcript)
38
Answer

1. PCR-based microsatellite analysis is another way
   to look for the 1p/19q deletion.  In a true
   oligodendroglioma, the entire arm of 1p and the
   entire arm of 19q will be deleted, so all the
   microsatellites on both arms will show fractional
   gene loss, or LOH.  It is more encompassing than
   the FISH probes, which by their very nature
   cannot scan such a large piece of DNA.  At first
   glance, there would appear to be a discrepancy
   between the FISH and PCR results for 19q.  A
   closer look, however, reveals that the 2
   microsatellite markers on 19q showing LOH are
   centromeric to the FISH probe, which is on
   19q13.  There is no overlap between the loci
   probed and, thus, no discrepancy.  Clearly,
   though, there is an interstitial deletion of some
   genetic material on 19q centromeric to the FISH
   probe.  Such occurrences are not unusual in
   genetically unstable tumors like GBMs.
2. The 9p microsatellite marker that corresponds to
   9p21 did not show LOH, whereas FISH showed
   homozygous deletion of 9p21.  This is not
   actually discrepant, because in the case of a
   complete homozygous deletion no LOH will be
   detectedthere isnt any heterozygosity to begin
   with, so the computer analysis package is fooled
   into thinking there isnt any gene loss.  In this
   case, trust the FISH.
3. 10q contains PTEN, which is another tumor
   suppressor protein that specifically acts within
   the EGFR pathway as a brake on EGFR signaling.  A
   loss of PTEN correlates well with amplification
   of EGFR in GBMs and is yet another piece of
   evidence that this is a primary GBM.  Again, in
   cases where histological grading is equivocal,
   such a molecular aberration portends rapid
   transformation into a high grade glioma.
4. 17p contains the p53 gene.  Taking into account
   Knudsons classic two-hit hypothesis, the
   conventional dogma is that a p53-driven secondary
   GBM will have an inactivating mutation on one of
   the two p53 copies (hit 1, producing strong p53
   immunostaining as discussed earlier) plus LOH on
   17p (hit 2).  No LOH is seen in this case,
   correlating nicely with the p53 immunostain.