The Bristol Experience of

1
The Bristol Experience of Molecular Genetic
Analysis of Gliomas (LOH and MGMT) for
Optimisation of Treatment
CMGS Spring Meeting 27th March 2009
The care of all patients with brain and other
central nervous system (CNS) tumours should be
coordinated through a specific model of
multidisciplinary assessment and care
Hilary Sawyer Bristol Genetics Laboratory
2
Outline

• Background to Gliomas
• Clinical Trials suggesting prognostic indicators
• Clinical Utility of 1p19q LOH and MGMT testing
• Bristol team data
• Case studies
• How results are used clinically in Bristol
• The future and laboratory issues

3
Brain tumours
1.6 of cancers in England and Wales High
morbidity and mortality

• Common symptoms headache with cognitive or
  behavioural symptoms, epilepsy, progressive focal
  neurological deficits, raised intracranial
  pressure
• Treatment - surgery, radiotherapy and more
  recently chemotherapy
• Tumours within the brain, such as gliomas, can
  rarely be completely removed because of their
  relation to critical structures and the
  infiltrating nature of the tumour
• Many do not demonstrate metastasis but invasion
  may preclude surgical resection generating
  ongoing management issues
• Slow growing tumours may transform into more
  aggressive tumours
• Classified by cell type, grade and location
• 2007 WHO classification
• Grade I low grade, well circumscribed, slowly
  progressing, can often be cured by resection
• Grade II low grade, typically
  infiltrative, low proliferation but higher
  likelihood of recurrence Grade III and IV
  high grade, malignant, grow rapidly, aggressive
• There is considerable inter-observer variation
  in diagnosis and classification however.
• Molecular tests help understand pathogenesis and
  improve classification.
• Evidence for two predictive molecular markers
  emerging for GLIOMAS
• 1p/19q LOH and MGMT

4
Gliomas

• Tumours arising from glial cells
• Non-neuronal cells
• provide physical support and nutrition
• maintain homeostasis
• form myelin
• participate in signal transmission
• Glial cell types
• Astrocytes anchor neurons to blood supply,
  regulation and signalling ASTROCYTOMAS
• Oligodendrocytes - produce myelin sheath
  OLIGODENDROGLIOMAS
• Mixed Gliomas OLIGOASTROCYTOMAS
• Ependymocytes - lining and secret CSF EPENDYMOMAS
• Glioblastomas are the most common form of
  astrocytic tumour

Average survival times Grade II astrocytoma 7
years Anaplastic astrocytoma 3.5
years Glioblastoma 9-11 months High-grade
glioma (WHO III and IV) includes glioblastoma,
anaplastic astrocytomas, anaplastic
oligodendrogliomas and anaplastic ependymomas
well documented to progress to higher grade
malignancy
5
Genetic pathways in gliomas
Astrocytes or precursor cells
Oligodendrocyte
Low-grade astrocytoma TP53 mutation (59)
WHO grade II
LOH 1p19q 50-70
p16 30 LOH 10q 50 EGFR 20
Anaplastic Astrocytoma TP53 mutation (53)
WHO grade III
Oligodendroglioma Grade III Chemosensitive
Oligodendroglioma Grade III Chemoresistant
MGMT methylation
1 Glioblastoma de novo LOH 10q (70) EGFR amp
(36) p16INK4a deletion (31) TP53 mutation
(28) PTEN mutation (25)
2 Glioblastoma LOH 10q (63) EGFR amp
(8) p16INK4a deletion (19) TP53 mutation
(65) PTEN mutation (4)
WHO grade IV
No longer enough to define oligodendrogliomas
histologically
Adapted from Ohgaki H Neuropathology 200525 1-7
6
Oligodendroglioma
Typical fried egg appearance (fixation
artifact)

• Oligodendroglia - specialised CNS myelin-forming
  cells.
• Rare primary brain tumours, frequency 0.53
  cases/100,000
• Around 25 primary brain tumours in adults
• 20 are anaplastic

Anaplastic oligodendroglioma (WHO grade III)

• Median age of onset between 40 50 years
• Treatment by surgical resection, followed by
  radiotherapy and PCV chemotherapy on recurrence.
• Main aims of treatment and follow-up are to
  increase survival while maximising a patients
  functional capability and quality of life
• Relatively good prognosis, survival of 3-10 years
  from diagnosis.
• Recurrence is common, often leading to disease
  progression and death.

7
Molecular Markers in Oligodendrogliomas (ODs)
Various trials support the clinical utility of
analysis for 1p19 LOH
There had been general disillusionment with
chemotherapy
Various trials have showed improved outcomes in
patients with 1p19q LOH Initial trials
Cairncross G et al (1994 and 1998) 1p and 19q
deletions were observed in 66 ODs Reifenberger
et al Neuropath Exp Neurol 2003
The above were small trials or retrospective
series suggesting predictive value. Further
trials (Phase III) were needed
From McDonald et al (2005) Cancer.
1104(7)1468-77
1) RTOG Trial 9402 Pure and Mixed Anaplastic
Oligodendroglioma PCV RT vs RT
alone Progression-free survival time favoured PCV
RT but 65 of patients experienced
toxicity Tumours with LOH 1p19q have longer
median survival times (gt7 v 2.8 yrs) and longer
PFS Better prognosis, time to relapse but no
difference whether early v late chemotherapy
2) EORTC In newly diagnosed AODs and
OAs Adjuvant chemotherapy improves PFS but does
not affect OS, therefore timing is less relevant.
Oligodendrogliomas with 1p loss alone have an
intermediate prognosis. Ino Y et al. Clin Cancer
Res (2001). I
8
3) NOA-4 study Phase III trial

• Multicentre randomised trial in Germany of
  sequential radiochemotherapy
• of oligoastrocytic WHO grade III tumours with PCV
  or temozolomide
• Wolfgang WW et al (2008) J Clin Oncol 26
• Patients with any grade III tumour anaplastic
  oligodendroglioma, oligoastrocytoma or
  astrocytoma with or
• without deletions
• ? Either 6 week course RT or 4x6 week cycles of
  either PCV or temozolomide
• LOH 1p/19q and hypermethylation of MGMT ? large
  risk reduction for time to
• treatment failure (TTF) regardless of histology
• Oligodendroglial histology better than astrocytic
  (NB 1p/19q LOH is present in higher
• of cases of AOD than AOA, also low in mixed
  OAs with predominant OD)
• No difference in TTF between patients on RT v
  chemotherapy or which is used first
• MGMT methylation status may have been more
  important

These trials confirmed that 1p/19q loss improved
outcome following RT with or without
chemotherapy i.e. appears to be a prognostic
factor for survival regardless of treatment type
or timing ODs with 1p19q LOH also respond more
favourably to temozolomide (an alkylating agent)
9
Loss of Heterozygosity Analysis LOH (Bristol)
Ino Y et al. (2001) Clin Cancer Res 7,
839-845. Smith J et al (2000). J Clin Oncol 18
636-645 Cairncross et al (1998) J Natl Cancer
Inst 90, 1473-1479

• Marker profile of tumour DNA compared to blood
  DNA to determine LOH. Panel of 8 markers in 3
  multiplexes and one simplex
• Fresh/frozen tumours used in Bristol v PPFE
• As oligodendrogliomas may be diffuse tumours,
  problems with mixed tissue can occur (require
  60-90 tumour tissue)

10
Bristol LOH 1p19q Data 2005 to early March 2009
300 patients. Approx 50 per year but increasing
The majority of pure AODs carry 1p/19q
deletions predictive, prognostic
Tumour Type No LOH LOH 1p/19q LOH 1p LOH 19q Partial LOH 1p Partial LOH 19q Partial LOH 1p19q Un- Reportable Total Number
Oligodendroglioma 7.5 82.5 0 5 5 0 0 0 40
Oligoastrocytoma 34.6 46 0 7.7 3.8 3.8 3.8 0 26
Astrocytoma 56.8 9.8 5.9 7.8 7.8 5.9 5.9 0 51
Malignant astrocytic glioma 50 0 0 50 0 0 0 0 2
Infiltrating astrocytic tumour 85.7 7.1 0 0 7.1 0 0 0 14
Glioblastoma 57.6 5 1.2 16 3.8 10.1 5.1 0.6 158
Gliosarcoma 50 25 0 0 0 1 0 0 4
Neuroectodermal 50 0 0 25 0 0 25 0 4
Liponeurocytoma 0 100 0 0 0 0 0 0 1
Total 300
Some OAs have loss and respond well
Grey low no cases
There is a good response of glioblastomas with
19q deletion
Patient with rare histology with 1p/19q LOH and
good response to PCV
Literature Where co-deletion of 1p and 19q is
present this is usually found throughout the
tumour
11
Clinical Case 1

• Aug 02 Male aged 68 at presentation
• Confusion and left sided weakness
• Right parieto-occipital mass
• Partial resection
• Histology ganglioglioma with foci of
  mitotically active primitive neuroectodermal
  tumour
• Oct/Nov 02 Radical radiotherapy
• Dec 02 Massive recurrence
• Further surgery but prognosis poor

12
Clinical Case 1 (con)

• Dec 02 1p19q deletion detected and offered
  palliative PCV
• Feb/Oct 03 PCV with complete radiological
  response
• Jun 06 relapse treated with stereotactic
  radiosurgery
• Dec 06 Rapid decline
• Feb 07 Died

1p19q analysis correctly identified a
chemosensitive tumour, even when the morphology
was confusing and the clinical course appeared
aggressive. Patient had a 3.5 year remission
with a good quality of life due to this
intervention
13
Clinical Case 2

• Dec 06 40 yr old female
• Presented with headaches and drowsiness
• Extensive tumour in right hemisphere and
  thalamus
• Jan 07 Partial resection
• Histology Central liponeurocytoma
• Rare tumour- c. 25 reported cases
• no clear guidance on treatment
• 1p19q deletions detected
• Feb/Apr 07 Radical radiotherapy
• Oct 07 Recurrence further surgery
• Histology unchanged
• Nov/Jul 08 PCV chemotherapy
• Jul 08 MRI clear

1p19q encouraged the use of PCV where no data was
available It has already proved of more durable
adjuvant benefit than radiotherapy
14
Glioblastomas

• Commonest primary brain tumour 5/100,000 annum
• Either develop from lower malignancy grade tumour
  or de novo (different genes/same cell pathway)
• 50 respond to alkylating agent temozolamide

Glioblastoma (WHO grade IV)

• Responsive tumours show promoter methylation
  (inactivation) of the MGMT (O6-methylguanine-DNA
  methyltransferase) gene (10q26)
• MGMT is a DNA repair protein (suicide enzyme)
  that removes alkyl groups from guanine, reversing
  the effect of temozolamide
• MGMT methylation may predict responsiveness to
  temozolamide treatment.

15
The Stupp Trial
Recruited patients with GBM post surgery
randomised to RT alone or RT with concomitant
temozolomide
N573 MS (months) 2 year survival () 3 year survival ()
RT alone 12.1 10.4 3
RT and Temo 14.6 26.5 17

• Addition of chemotherapy to radiotherapy
  significantly prolongs survival among patients
  with newly diagnosed glioblastoma
• Increase in survival rate at 2 years

16
MGMT gene silencing and benefit from
temozolomide in glioblastoma. Hegi ME, et al
New Eng J Med 2005352997-1003.
Randomized trial comparing RT alone with RT
combined with concomitant and adjuvant
temozolamide
From Hegi
206 tumours 44.7 methylated 55.3 unmethylated
MGMT methylated tumours Median Survival 18.2
months No methylation Median
Survival 12.2 months
Irrespective of treatment, MGMT promoter
methylation was an independent favorable
prognostic factor
17
Bristol MGMT methylation analysis
Main assay is a diplex of unmeth and meth product
Much of literature as separate simplexes
Meth result
Unmeth result
Bristol MGMT analysis data to Late Jan 09
To late Jan 09 388 cases
Hegi et al 2005 N206 Methylated 45 RTOG
study Methylated 50
18
2005 MGMT Survival data n21

• Methylated (n12)
• Median Survival 15.5 mths
• Range 0-31 mths 1 patient alive
• Unmethylated (n9)
• Median survival 10 mths
• Range 1-36 months 1 patient alive
• The above date from 2005, before concomitant
  temozolomide was routinely available on the NHS
• The above did not receive concomitant
  temozolomide
• Now NICE approval obtained for concomitant
  temozolamide
• It is expected that survival gap will widen.

19
How are the clinicians using these MGMT data now?

• Subgroups in Hegi study were too small to exclude
  a possible un-seen benefit for unmethylated
  patients so NICE recommendation is that all
  patients with GBM receive concomitant treatment.
• However the outlook is clearly poor for
  patients in the unmethylated group, therefore the
  challenge is either to enhance the effectiveness
  of temozolomide or to find a better option.
• RTOG dose dense temozolomide study can MGMT
  be saturated with more prolonged treatment
  schedule?
• Bristol highest recruiters to trial in Western
  Europe

20
Clinical Management in Bristol

• All patients with grade II-IV tumours have
  1p19q analysis and MGMT analysed
• Management is not affected by morphology but
  determined by grade and genetic results
• Genetic results may be particularly useful
  when the histological diagnosis is unclear
• Grade II Surgery and follow up
• 1p19q deletion informed of better
  prognosis
• At progression offer BR13 trial
  (randomised for radiotherapy or temozolamide as
  initial treatment to look at OS,PFS and QOL )
• Trial declined? PCV if 1p deletions, RT
  if no deletions
• Grade III Surgery and immediate oncology
  treatment
• If have 1p19q deletions recommend PCV
  as first line treatment as patients with good
  prognosis may suffer late effects if cranial
  radiotherapy
• No deletions recommend initial
  radiotherapy
• Patient factors and choice important
• Grade IV Surgery and immediate radiotherapy
  with concomitant temozolamide
  Standard regimen whilst awaiting results of
  dose dense study.
• Patients uncertain about chemotherapy
  are informed of methylation status
• and Hegi data to inform choice
• Where MGMT promoter is unmethylated,
  patients advised that this tumour is
  particularly aggressive

May consider using chemotherapy in patients with
rare tumours and 1p19q LOH
21

• The Future
• LOH and MGMT testing now indicated as valuable
  predictive markers in
• RCPath Dataset for tumours of the CNS (2nd
  edition) April 2008
• Require robust funding streams and staffing
  previous from
• charitable funds/training (NICE suggest 2500
  annual tests and up to 30
• technical staff across England and Wales)
• Local clinicians are proud of the accurate
  information patients are given.
• Bristol is one of the biggest testing centres in
  UK
• It is hoped to have better therapies for all
  patients
• Bristol patients in other trials therapy for low
  grade gliomas (BR13 etc)
• 1p and 1p19q LOH appears to be a marker for
  improved PFS in grade II gliomas
• Routinely collecting fresh tissue for these
  assays has allowed Bristol to be prominent in
  national and international trials to inform
  future treatment (BR13 and CATNON trials etc).
  The local team will be able to implement any new
  guidelines speedily.
• Need to underpin this work through discussion of
  laboratory testing standards and guidelines to
  improve analysis, including tissue type (fresh v
  PPFE), choice of assay method(s), EQA and through
  availability of control reference materials

22
We are part of the wider Bristol MDT network as
outlined by NICE
23
Thanks to the Bristol Team

• BGL team Mark Greenslade
• Elena Mavraki
• Sarah Burton-Jones
• Suzanne OShea
• Laura Yarram
• Thais Simmons
• Paula Waits
• Kayleigh McDonagh
• Maggie Williams
• (Emma Ryan/Karen Meaney/Meera Parmar)
• Neuropathology Seth Love, Neuropathology,
  Frenchay
• Neurooncology Kirsten Hopkins, Bristol
  Oncology Centre
• Hugh Newman, UH Bristol