Báo cáo ( Bằng tiếng anh) Xử lý ảnh y tế (não người) Medical imaging

AI for Medical imaging
Name

Course

Master

Student ID

Dataset

BrainTumor

1. Introduction
The main task of this project: Binary Classification
Predict the MGMT methylation status using MRI from patients with brain
tumor.
1.1 What is the MGMT?
Glioblastoma is the most frequent malignant primary tumor in the brain. It has a
very poor prognosis, with a median survival of less than a year. The current
standard if care consists of surgical resection followed by radiotherapy in
addition to alkylating chemotherapy with temozolomide.
MGMT (O[6]-methylguanine – DNA methyltransferase) is a DNA repair
enzyme. This enzyme rescues tumor cells from alkylating agent-induced
damage, leading to chemotherapy resistance with alkylating agents.
1.2 MRI and MGMT Connection
MGMT promotor methylated glioblastoma is likely to show less aggressive
imaging feature than MGMT promotor unmethylated glioblastoma.
2. Datasets
Dataset link: />e=BPWmBu
Data Description

Format
Patients in training sets
Patients in testing sets

DICOM
400
185

There are 4 sub-folders, each of them corresponding to each of the MRI scans,
in DICOM format, included:
+ Fluid Attenuated Inversion Recovery (Flair)
+ T1 – weighted pre – contrast (T1w)


+ T2 – weighted contrast enhanced (T1CE)
+ T2 – weighted (T2)
The dataset structure:
Train/Test/Validation
|___00000
|

|_____FLAIR

|

|

|Image-1.dcm

|

|

|Image-2.dcm

|

|

|…

|

|_____T1w

|

|

|Image-1.dcm

|

|

|Image-2.dcm

|

|

|…

|

|_____T1wCE

|

|

|Image-1.dcm

|

|

|Image-2.dcm

|

|

|…

|

|_____T2w

|

|

|Image-1.dcm

|

|

|Image-2.dcm

|

|

|…

train/ folder: contain the training files.
labels.csv: contain the target MGMT_value for each subject in the training data.
test/ folder: contain the testing files.


Figure 1 : The bar graph for labels.csv file

Figure 2: The bar graph for train data

Figure 3: The pie chart for labels.csv

In this project, the sub-folders FLAIR and T1wCE were used.
3. Method

The workflow:
Input image
(Dcm file)

Convert to gray

Support Vector
Machines

Result


Support-vector machines (SVMs) are supervised learning models with
associated learning algorithms that analyze data for classification and regression
analysis.
SVC is a similar method that also builds on kernel functions but is appropriate
for unsupervised learning.
I use class sklearn.svm.SVC

4. Results
For training:
The mean accuracy:

For validation:

For testing:

The probability of each patient was saved in submission_c1.csv
Classification result for test data

5. Conclusion
+ The result was generate and it is not good
+ In the future, I need to apply the deep learning method in this problem to
improve the accuracy
+ Limitation: The time of semester is limited

# Bổ sung thêm code