Quality of data in brain tumors

1
Quality of data in brain tumors
Jerzy Slowinski

• 1) Department of Neurosurgery in Sosnowiec,
  Medical University of Silesia
• 2) Department of Epidemiology, School of Public
  Health, Medical University of Silesia
• Poland

2
The Regional Silesia Cancer Registry

• One of the 16 population-based cancer registries
  in Poland
• Covers population living in the Silesia
  Voivodeship (the second most populated region in
  Poland)

3
The Silesia Voivodeship

• The most urbanized and industrialized region in
  Poland
• 12 294 sq. km 4.7 million people

4
WHO classification of brain tumors (2007)

• Tumors of neuroepithelial tissue
• Tumours of cranial and paraspinal nerves
• Tumours of the meninges
• Lymphomas and haematopoietic tumours
• Germ cell tumours
• Tumours of the sellar region
• Metastatic tumours

5
Biological features of brain gliomas

• Diffuse, infiltrating growth
• Progression of the malignancy grade during the
  growth of tumor
• Tendency to recurrence
• Extraneural metastases extremely rare
• No lymph node involvement (no TNM staging)

6
2 months after tumor resection
3 months later
7
Issues potentially relevant to quality of
epidemiological data on brain and CNS tumors

• Classification
• Grading
• Ascertainment
• Completeness of registration
• Microscopic verification
• Reliability of histological diagnosis
• Differences in diagnostic capabilities between
  countries and regions (detection bias)

8
Classification, criteria

• Bailey, Cushing, 1929 stages of cytogenesis
• Kernohan, Sayre, 1951 degree of
  dedifferentiation
• WHO classification
• 1979 (K. Zülch)
• 1993
• 2000
• 2007

Harvey Cushing
9
WHO classification of brain tumors
WHO 2000, 2007
WHO 1993

• Tumors of neuroepithelial tissue
• Tumors of cranial and paraspinal nerves
• Tumors of the meninges
• Lymphomas and haematopoietic tumours
• Germ cell tumors
• Tumors of the sellar region
• Metastatic tumors

Cysts and tumor-like lesions Local extensions
from regional tumors Unclassified tumors

10
Grading of astrocytic tumors

• Kernohan
• Ringertz
• St. Anne-Mayo
• WHO

four-tiered
three-tiered
11
WHO grade vs. ICD-O-3 code
WHO
ICD-O-3
0 benign 1- borderline or uncertain behaviour 2
in situ not applicable 3 - malignant
I II III IV
low-grade
high-grade
12
WHO
ICD-O-3
Choroid plexus papilloma I 0 Pilocytic
astrocytoma I 1 Central neurocytoma II 1 Diff
use astrocytoma II 3 Anaplastic
astrocytoma III 3 Glioblastoma IV 3
No strict linkage between WHO and ICD-O-3
13
Multidisciplinary agreement on a standard
definition of brain tumors for collecting and
comparing data
Neuro-Oncology 2002 4 134-145.
14
Ascertainment

• Complete ascertainment is a key issue with regard
  to data quality

UNDERASCERTAINMENT
UNDERREPORTING (the number of cases not
identified)
DEFINITIONAL DIFFERENCES (cases eligible for
inclusion)
15
Ascertainment

• Autopsy rate (incidental autopsy cases)
• Case definition
• Resolution of duplicate cases
• Inclusion of benign tumors
• Shift from in-patient to out-patient settings

16
Completeness of registration

• Alarming underregistration of brain tumors by
  many registries
• 52 of cases identified in the cancer registry in
  Devon and Cornwall (Pobereskin LH. Eur J Epidem
  2001 17 413).
• 54 in the Scottish Cancer Registry (Counsell CE
  et al. J Neurol Neurosurg Psych 1997 63 94).

17
Crude incidence and mortality rates, and
mortality/incidence index for brain tumors
18
Death Certificate Only (DCO)

• Finland, Denmark 1
• Czech Republic 10
• Poland 15
• Portugal 20

Parkin DM et al. Cancer incidence in five
continents. IARC, Lyon 2002.
19
Microscopic verification

• Europe 70
• Switzerland 100
• Finland, Sweden, Austria gt90
• Poland 52
• The Silesia Voivodeship 59 (RSCR 2005)

EUROCARE-3, 1990-19941
1Sant M et al. Ann Oncol 2003 14 (Suppl.5)
v61-v118.
20
Reasons for the lack of microscopic verification

• No follow-up
• Patient refusal
• Inoperable tumor
• Inadequate availability of stereotactic
  procedures
• Low autopsy rate

21
Reliability of histological diagnosis
Pathologist
Laboratory
Tissue
22
Reliability of histological diagnosis

• The complex nosology and pathology of the CNS
  tumors
• Definitions for certain tumor types imprecise,
  rely on subjective criteria
• Surgical pathologists not familiar with the CNS
  malignancies (RARE CANCER)
• Inter-observer concordance rate ranges from 20-86

23
Concordance rates () between neuropathologists
and cancer registry data
Nosologists agree with registry (n326) 3 of 3 neuropathologists agree with registry (n204)
Exact ICD-O histology coding 88 49
General histology grouping 96 81
Tumor type
Astrocytoma/Glioblastoma 95 52
Oligodendroglioma 92 40
Medulloblastoma 100 58
Schwannoma 100 89
Meningioma 97 95
Mixed glioma 57 0
Castillo MS et al. Neuroepidemiology
20042385-93.
24
Proposals I

• The unique case ascertainment rules
• The unique classification and grading system
  (conversion to ICD-O-3 according to the Consensus
  Criteria)
• Encourage training in diagnostic neuropathology
• Putting more efforts into obtaining the material
  for tissue diagnosis (stereotactic brain biopsy)

25
Proposals II

• Include benign tumors (consistent with EUROCARE-5
  Plenary Meeting proposal)
• Check the clinical data for the precise tumor
  site and other useful information
• Collect data on diagnostic and therapeutic
  procedures for HR studies
• In the future, better delineation of tumor
  subtypes based on a molecular profile of tumor
  cells

26
Thank you!
27
Advances in surgical therapy

• modern neuroimaging (fMRI, MRS, fiber tracking,
  PET)
• microsurgery
• skull-base surgical approaches
• intraoperative monitoring (evoked potentials,
  cranial nerves monitoring , ECoG)
• endoscopic techniques
• stereotactic techniques
• neuronavigation

28
Deletions in 1p and 19q
29
(No Transcript)
30
Astrocytoma, G I
31
Astrocytoma, G II
32
Glioblastoma, G IV
33
Glioblastoma, G IV
34
Classification and therapeutic issues

• relative RT and CHT resistance
• even low-grade tumors are life-threatening
  (biological malignancy)
• no widely accepted staging system
• no in situ tumors