CN107456225B - Partial volume effect correction method for cerebral blood flow - Google Patents

Abstract

The invention relates to a partial volume effect correction method of cerebral blood flow, which aims to correct partial volume by using a CBF value obtained in an ASL technology and provide a more accurate algorithm for clear AD diagnosis. The CBF value and the T1 value obtained from ASL are used as input, and the CBF value of whole brain white matter, brain gray matter and cerebrospinal fluid is obtained through PVE algorithm. The CBF values after PVE correction according to the method are divided into two groups of pixel values of white brain matter and grey brain matter. Compared with the CBF value of the whole brain, the calculation result is more accurate, and the clinical diagnosis and analysis are facilitated.

Description

Partial volume effect correction method for cerebral blood flow

Technical Field

The invention relates to the field of data processing, in particular to a method for correcting partial volume effect of cerebral blood flow.

Background

Cerebral Blood Flow (CBF) is the amount of Blood Flow per unit time through the vascular structures of a certain amount of brain tissue. Is defined as the amount of blood delivered per unit of time per unit mass of human tissue in ml/100 g/min. Changes in CBF are an early indicator of brain activation and are an important indicator of assessing brain health. It is the main basis for diagnosing and treating cerebrovascular diseases such as Alzheimer's disease, cerebral infarction, cerebral hemorrhage, aneurysm, congenital arterial and venous vascular malformation and the like. At present, the means for measuring cerebral blood flow mainly consists of PET, SPECT, MRI techniques, and the like. With the increasing development of magnetic resonance imaging (mri) technology, the magnetic resonance Arterial Spin Labeling (ASL) technology has been gradually used for CBF measurement as a new completely non-invasive blood perfusion technology. When the CBF value obtained by the magnetic resonance ASL technology is used for carrying out Alzheimer Disease (AD) research, compared with the cognitive normal old people, the AD patients have different degrees of brain atrophy, particularly medial temporal lobe atrophy, the different degrees of brain atrophy can influence the registration of images, and most ASL images have 3-4 mm interlamellar spacing, so that a large amount of gray matter and white matter mixed voxels exist. Errors necessarily result if gray or white matter cerebral blood flow calculation methods are used for these voxels, so that Partial volume correction (PVE) is required for these regions to obtain accurate CBF values.

Disclosure of Invention

The invention aims to carry out partial volume correction by utilizing the CBF value obtained in the ASL technology, and provides a more accurate algorithm for clear diagnosis of AD. The CBF value and the T1 value obtained from ASL are used as input, and the CBF value of whole brain white matter, brain gray matter and cerebrospinal fluid is obtained through PVE algorithm. Because the CBF value is low, at the boundary position of the white matter cerebrospinal fluid of the whole brain gray matter, one pixel value may contain a plurality of components such as the white matter, the brain gray matter and the cerebrospinal fluid, and the value of the cerebral blood flow in each component (white matter, gray matter and cerebrospinal fluid) is greatly different, so that the CBF value of the boundary area is more accurate after partial volume correction. The main formula is as follows:

CBF_pve= CBF/(P_GM*ΔM_GM+ P_WM*ΔM_WM+P_CSF*ΔM_CSF).

in the above formula, CBF is the cerebral blood flow perfusion value before correction, P_GM, P_WMAnd P_CSFProbability maps of gray brain matter, white brain matter and cerebrospinal fluid are shown. Δ M_GM, ΔM_WMand ΔM_CSFThe sub-tables represent arterial spin-labeled perfusion weighted signals of gray, white and cerebrospinal fluids, where Δ M_CSFThe value is 0, since the arterial discretionary marker signal of cerebrospinal fluid is almost 0. CBF in adults_grayThe value is about WM_CBF3 times the value. The above equation can therefore be simplified to:

CBF_pve= CBF/ (P_GM*3+ P_WM).

this simplified formula is used to make PVE corrections to ASL data.

The invention is realized by the following technical scheme:

a method for correcting partial volume effect of cerebral blood flow is characterized by comprising the following steps:

step one), acquiring a cerebral blood flow value, a brain grey matter value and a brain white matter value of the whole brain according to a detection result of a nuclear magnetic artery spin labeling sequence;

step two), setting the brain gray value of an area with the brain gray value less than or equal to 0.25 as 0, and setting the area with the brain gray value not equal to 0 as GM; setting the white matter value of a region with the white matter value less than or equal to 0.25 as 0, and setting the obtained region with the white matter value not equal to 0 as WM;

step three) for each numerical point in the area GM, the correction is carried out according to the following formula

CBF_gray=CBF/（GM_gray+0.33*GM_white）

Wherein CBF_grayThe correction result of the cerebral blood flow value in the gray matter of the point is CBF which is the cerebral blood flow value of the point, GM_grayThe gray brain value, GM at that point_whiteThe white matter value of the point;

for each value point in the area WM, the correction is made as follows

CBF_white=CBF/（WM_white+3*WM_gray）

Wherein CBF_whiteIs the correction of the cerebral blood flow value in the white matter of the spot, CBF is the cerebral blood flow value of the spot, WM_whiteTo the white matter value of the point, WM_grayThe grey brain matter value of the point;

step four) according to CBF_grayMaking a brain gray matter brain blood flow chart according to the values of the above (CBF)_whiteThe cerebral leukoencephalogram is prepared by the numerical values of (A).

Compared with the prior art, the invention has the following advantages:

the CBF values after PVE correction according to the method are divided into two groups of pixel values of white brain matter and grey brain matter. Compared with the CBF value of the whole brain, the calculation result is more accurate, and the clinical diagnosis and analysis are facilitated.

Drawings

FIG. 1 is a CBF map of the whole brain;

figure 2 is a corrected gray matter CBF map.

Detailed Description

Example 1

The method can be realized by adopting matlab programming, and the matlab program is as follows:

function FG_PVE_correction_for_perfusiondata(Imgs,Grays,Whites,smooth_kernel)

defining a PVE function with four function variables (whole brain, white matter, gray matter, smooth kernel parameters)

sub = [1,2,3,4,5,6,7, …, n ] (a read image number, n is a data number of an incoming program)

for k =1: x (x represents the number of data entering the program)

n = sub (k) (n represents the number of data)

S = num2str (n); (changing from a character variable to a number variable)

Imgs = strcat ('local path \ means', S, '. img') obtaining a whole brain pixel value

Grays = stricat (' local path ' \\ ', ' c1S ', S, '. img ') acquires gray matter pixel values

Whites = strcat ('local path \', 'c2S', S, '. img') acquires a white matter pixel value

smooth _ kernel = [ 222 ], and the number of smooth cores is defined as 2

for i =1: size (Grays,1) read brain data

Vmat = spm _ vol (Imgs (i); assigning whole brain pixel values to the Vmat variable

V = spm _ read _ vols (Vmat); reading Vmat variable into V variable

GM = spm _ read _ vols (spm _ vol (Grays (i)))), gray matter data is assigned to GM

WM = spm _ read _ vols (spm _ vol (Whites (i)))), white matter data is assigned to WM

GM (find (GM < =0.25)) =0, and gray matter is 0 or less with a gray matter value of 0.25

WM (find (WM < =0.25)) =0, white matter is 0 with white matter number less than or equal to 0.25

V21= zeros (size (V); V21 variable space size is defined as equal to V (brain space size), and is assigned a value of 0

V22= zeros (size (V); V22 variable space size is defined as equal to V (brain space size), and is assigned a value of 0

V21(find(GM))=V(find(GM))./(GM(find(GM))+0.4*WM(find(GM)));

V21 (grey matter) = grey matter/(grey matter +0.33 x white matter)

V22(find (WM) = V (find (WM))/(WM (find (WM)) +2.5 GM (find (WM)); V22 (white matter) = white matter/(white matter +3 grey matter)

Vmat.fname=deblank(PVE_gray_imgs(i,:));

spm_write_vol(Vmat,V21);

Assigning the value of V21 to PVE _ gray _ imgs

Vmat.fname=deblank(PVE_white_imgs(i,:));

spm_write_vol(Vmat,V22);

The value of V22 is assigned to PVE _ white _ imgs

end

fprintf ('\n-----PVE correction is done!......\n')

GUI interface prompting that PVE calibration has been completed

end

Fig. 1 and 2 show the correction effect of the method, and the CBF values after PVE correction according to the method are divided into two groups of cerebral blood flow images of white matter and gray matter. Fig. 2 is a corrected gray matter brain blood flow chart, and compared with a whole brain CBF value, the calculation result is more accurate, and clinical diagnosis and analysis are facilitated.

Claims (1)

1. A method for correcting partial volume effect of cerebral blood flow is characterized by comprising the following steps:

step one), acquiring a cerebral blood flow value, a brain grey matter value and a brain white matter value of the whole brain according to a detection result of a nuclear magnetic artery spin labeling sequence;

step two), setting the brain gray value of a point with the brain gray value less than or equal to 0.25 as 0, and setting the obtained area with the brain gray value not being 0 as GM; setting the white matter value of a point with the white matter value less than or equal to 0.25 as 0, and taking the obtained area with the white matter value not equal to 0 as WM;

step three) for each numerical point in the area GM, the correction is carried out according to the following formula

CBF_gray=CBF/（GM_gray+0.33*GM_white）

Wherein CBF_grayThe correction result of the cerebral blood flow value in the gray matter of the point is CBF which is the cerebral blood flow value of the point, GM_grayThe gray brain value, GM at that point_whiteThe white matter value of the point;

for each value point in the area WM, the correction is made as follows

CBF_white=CBF/（WM_white+3*WM_gray）

Wherein CBF_whiteIs the correction of the cerebral blood flow value in the white matter of the spot, CBF is the cerebral blood flow value of the spot, WM_whiteTo the white matter value of the point, WM_grayThe grey brain matter value of the point;

step four) according to CBF_grayMaking a brain gray matter brain blood flow chart according to the values of the above (CBF)_whiteThe cerebral leukoencephalogram is prepared by the numerical values of (A).

Images