Study of metabolic abnormality for newlydiagnosed glioma patients

1
Study of metabolic abnormality for
newly-diagnosed glioma patients

• Xiaojuan Li
• MRSC, UCSF
• Dec, 2000

2
Objects

• To study the spatial distribution of the
  metabolic abnormality
• To study the difference of the metabolic
  properties between grades

3
Methods (1)

• To study the volume of the metabolic abnormality
• 35 newly diagnosed glioma patients 12 grade2
  14 grade3 6 GBM
• To study the biopsy data
• 84 biopsy data 43 grade2 25 grade3 16 GBM

4
Methods (2)

• To minimize the perpendicular distance instead of
  vertical distance
• To consider the profile of press box
• To define the region of normal brain tissue as
  voxels outside T2 abnormality which allows a
  better initialization
• To normalize the data with the noise level

5
Spatial heterogenity inside Cho/NAA
abnormality(t916, t558, t648)
6
Comparison of the abnormality volume
7
Volume abnormality of Grade2 patients
8
Comparison of the outliers in Grade2(t1125,
t1203)
9
Volume abnormality of Grade3 patients
10
Non-uniform of spatial distribution of metabolic
abnormality in grade 3 patients(t1169, t843)
11
Volume abnormality of GBM patients
12
Non-uniform of spatial distribution of metabolic
abnormality in GBM patients(t944, t648, t1180
t917, t558, t608)
13
Elevated Creatine in a Grade 3 patient(t1169)
14
Metabolic distribution of the biopsy data
15
Metabolic residual and Ratio distribution of
biopsy data
16
Comparison of the volume of abnormality
17
Pairs plot for grade2 biopsy
18
Pairs plot for grade3 biopsy data
19
Pairs plot for grade4 biopsy data
20
Classification by discriminant analysis(plug-in
error)
Left with parameters of ChoNAA, ChoCr, NAA,
Lac, Age and Gender right with parameters above
plus pathologic diagnosis (A, OA or OD)
21
Classification by discriminant analysis (cross
validation error)
Left with parameters of ChoNAA, ChoCr, NAA,
Lac, Age and Gender right with parameters above
plus pathologic diagnosis (A, OA or OD)
22
Summary

• Spatial heterogeneity exists inside the T2 or
  Cho/NAA abnormality
• Each grade may be composed with subgroups in
  terms of metabolite properties, considering the
  presence of Lac/Lip, percentage of T1 abnormality
  to T2 abnormality, etc..
• Creatine can be elevated or reduced and be the
  highest in Grade 3. It also has inverse relation
  with lac/lip.
• Metabolic measurements can be applied to help
  grading the gliomas patient

23
Discussion and Futures(1)

• To consider the metabolic variation of the normal
  tissue at different locations
• To consider the spatial correlation
• To incorporate with other parameters, including
  ADC, rCBV
• To find a more reliable way to normalize data
• To evaluate the oxygenation by observing the
  treatment response

24
Discussion and futures (2)

• To evaluate the radio resistance inside the
  lesion, which is important for planning of focal
  therapy
• To find an abnormal index to help grading the
  glioma patients with metabolic information