RU2589652C1 - Method for preoperative determination of degree of malignancy and histological subtype of tumours of brain tunic - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention relates to neurosurgery and neuroradiology. MRI images are analysed in mode of T1 with contrasting in stages. Method includes first determining intensity of each pixel in tumour area to contrast MRI T1 weighted images. Method then includes normalisation of intensity of each pixel on intact tissue white matter of brain of patient, taking into account histogram offset coefficient relative to average background colour database MRI images of brain of patients with tumours of brain membranes. Forming a histogram of normalised intensity of pixels on MRI images. Determining position of histogram peak. By comparing its value with range of values different histological types of tumours of meninges, specified in database, determine histology and corresponding degree of malignancy.

EFFECT: method provides highly accurate detection of tumours by histological type of MRI images in preoperative period.

1 cl, 7 dwg, 2 ex, 3 tbl

Description

FIELD OF TECHNOLOGY

The invention relates to medicine, namely to the sections of neurosurgery and neuroradiology.

BACKGROUND

Meningiomas are common tumors of the arachnoid membrane of the brain, the treatment of which in most cases is surgical removal. The radicalism and safety of the removal of these tumors largely depends on their consistency. Fibroplastic and mixed types of meningiomas are, as a rule, more dense than meningotheliomatous and psamomotous. In this regard, the non-invasive method of preoperative determination of the histological type and malignancy of tumors of the meninges is important for optimal preoperative planning.

There is a method of analysis based on determining the signal intensity from meningiomas (areas of interest) in various MP modalities (T1, T2, FLAIR, DWI, MD and FA cards) [1, 2, 3]. Dense tumors most often have an iso-intensive signal on MD-DWI cards and hyper-intense on FA-DWI mode.

A known method [4] measuring dynamic changes in the concentration of contrast medium during MRI of the brain with contrast. In the same area of interest, an MRI scan is performed in T1 mode at different time intervals, and then an increase in the MP signal intensity is graphically shown, which correlates with the area of blood-brain barrier disturbance and perfusion, which is indirectly associated with angiogenesis.

A known method [5] of analyzing the zone of interest of brain regions using volumetry, which uses the summation of the number of voxels in the zone of interest, presenting this through histograms. A known method [6] analysis of neoplasms on MRI images, in particular using layer-by-layer analysis of each image. The method is aimed at segmenting brain tissue by constructing a histogram of images of T1 and T2 modes, which allows you to separate the gray and white matter of the brain in each case and is not aimed at comparing images between different patients.

A known method [7] analysis of neoplasms using the construction of histograms, however, in the known method the principle of normalization of the image on the white matter of the brain is not described. White matter is not homogeneous and its intensity of the MP signal ranges from -50 to 120 (depending on the tomograph), and it can also contain vessels whose intensity is greater than 120. In addition, the boundaries of the intensity of the white matter of the brain can vary and depend on the time of the study and on tomographs of different manufacturers. Therefore, simply rationing without taking these features into account reduces the sensitivity of the methods and does not allow comparing histograms obtained from images of different tomographs, such as Toshiba, Siemens, GE.

The disadvantages of the above known methods is the inability at the preoperative stage to establish the histological type of a tumor, to differentiate meningiomas from other tumors of the meninges, a significant increase in the protocol and duration of an MRI study.

SUMMARY OF THE INVENTION

The aim of the developed method for the analysis of MRI images is the preoperative determination of the histological type, subtype, and degree of malignancy of tumors of the meninges.

The technical result is an increase in the recognition accuracy of the histological type of neoplasms by MRI images in the preoperative period.

The technical result is achieved by the method of preoperative determination of the degree of malignancy and the histological type of brain meningiomas, including the analysis of MRI images in T1 mode with contrast, containing stages in which:

- determine the intensity of each pixel in the tumor area in contrast MRI T1 weighted images;

- normalize the intensity of each pixel on the intact tissue of the white matter of the patient’s brain, taking into account the displacement coefficient of the histogram relative to the average background color of the MRI database of images of patients with tumors of the meninges of the brain;

- form a histogram of normalized pixel intensity in MRI images;

- determine the position of the peak of the histogram and based on a comparison of its value with the values of different histological types of tumors of the meninges indicated in the database, determine the histological type of tumor and the corresponding degree of malignancy.

BRIEF DESCRIPTION OF THE DRAWINGS

In FIG. 1 - FIG. Figure 3 shows the brightness variations of MRI images.

In FIG. 4 shows a preoperative MP image of T1, weighted in the axial plane, showing the meningotheliomatous subtype of meningiomas.

In FIG. 5 shows a histogram corresponding to the meningotheliamotous type of tumor, constructed from images of Patient No. 1.

In FIG. 6 shows an MP image of T1, weighted in the axial plane, representing the anaplastic subtype of meningioma.

In FIG. 7 shows a histogram corresponding to the anaplastic type of meningioma, constructed from images of Patient No. 2.

DETAILED DESCRIPTION OF THE INVENTION

The method is as follows.

Before surgery, the patient undergoes an MRI of the brain with contrast. Then, from the obtained DICOM data packet, images are exported in the JPEG Lossless extension, visualizing the tumor (region of interest) in T1 mode with contrast.

A histogram of the relative intensity of the MP signal from the region of interest summarized from all exported images is constructed. The area of interest is normalized to the white matter of the brain, taking into account the histogram displacement coefficient relative to the average background color of the MRI database of images of patients with tumors of the meninges.

The formula for calculating the intensity indicator for each pixel in the region of interest, on the basis of which the histogram will be built.

K = a-a0; X = X0-K / 10a; Y = 1000Y0 / Z.

Legend:

Y - The normalized number of identical values of X0 on all slices.

X - Relative pixel color (by additive RGB color model) in area 2.

1 area - background, normal (intact) tissue area of the white matter of the brain in the MP image

2 region - region of interest, tumor region in the MP image

Y0 - The number of identical values of X0 on all slices.

X0 - The absolute color of the pixel (according to the additive RGB color model) in area 2.

X - Relative pixel color (by additive RGB color model) in area 2

a0 - Average color (according to the additive RGB color model) of the white matter of the brain from the database

a - Average color (according to the additive RGB color model) in area 1.

K - Displacement coefficient of the histogram relative to the average background color of the database of MRI images of patients with tumors of the meninges

Z - The sum of pixels across all slices in area 2.

To calculate the coefficient K, it is necessary to have a formed database of MRI images of patients with tumors of the meninges of the brain. A database was formed from images of 50 patients, in which sections of the white matter of the brain, the “background,” were analyzed. The average color is calculated in each background and as a whole in the base (a0). The average color in the base (a0) is the color that is taken as the basis for calculating the coefficient K.

Subtracting the zone of interest K from each color of the pixel (X0), we do not change the structure of the distribution of colors, but only shift the histogram to one side, making the image brighter or darker. At the same time, we divide each background pixel of the new changed zone of interest by the average background color of the original analyzed image (a). This action is necessary to increase the sensitivity of the method, which allows you to take into account variations in the brightness of the images (Fig. 1-Fig. 3). When normalized to the most common color in the white matter of the brain in FIG. 1 and FIG. 3 without taking into account the coefficient K, we did not get an adequate result on the histological type of tumor.

We plot the function (f) X = Y. We determine the position of the peak of the graph of values along the x axis and compare it with the limits of the values of different histological types of meninges tumors indicated in the database.

The histological type and subtype of the tumor are morphological microstructural tissue characteristics of the tumor. The degree of malignancy of the tumor is a parameter indicating the aggressiveness of the behavior of the tumor. The higher the degree of malignancy, the more the tumor cell differs from the normal cell and the less favorable the prognosis of the disease.

These two concepts are related to each other (Table 1).

Table 1 presents the ratio of the type of tumors of the meninges and the degrees of malignancy according to the classification of WHO 2007 (4th edition, extreme) [8]

Based on the analysis of MRI data of a number of patients, the correspondence of the position of the peak of the histogram graph to the histological type of tumors of the meninges was formed (Table 2).

Additionally, to display an example of identifying histogram peaks, an additional table will be given (Table 3).

7

8

Based on the table, by the peak of the histogram, it is reliably possible to determine 5 groups of tumors:

Meningotheliomatous meningiomas (group 1), fibroplastic meningiomas (group 2), hemangiopericytomas (group 3), atypical, anaplastic meningiomas (malignant meningiomas, group 4) and fibrosarcomas (group 5).

Atypical and anaplastic meningiomas, despite a slight difference in the degree of malignancy of tumors, can be combined into one group because of similar biological behavior, which leads to an identical approach to their treatment. The first step is the resection of the tumor followed by irradiation. The role of chemotherapy is currently not fully understood [9, 10].

The sensitivity of this method is 93.3% of the correct conclusions.

Examples:

Example No. 1: A patient with a meningotheliomatous subtype of meningiomas. In FIG. 4 shows a preoperative MP image T1 weighted in the axial plane. On the histogram constructed using the inventive method (Fig. 5) and its subsequent processing, a peak was detected at the level of 16.5, which corresponds to the meningotheliomatous subtype of meningiomas.

Example No. 2: A patient with an anaplastic subtype of meningioma. In FIG. 6 shows a preoperative MP image T1 weighted in the axial plane. On the histogram constructed using the inventive method (Fig. 7) and its subsequent processing, a peak was revealed at level 9, which corresponds to the meningotheliomatous subtype of meningiomas.

The claimed invention can be widely used in preoperative diagnosis of malignant brain tumors.

Information sources

1. The role of advanced MRI techniques in differentiating typical, from atypical and malignant meningiomas Hazim Ibrahim Tantawy a, *, Faten Fawzy Mohamad a, 1, Ayman Mohammad Ismael b, 2 The Egyptian Journal of Radiology and Nuclear Medicine (2010) 41, 411-419.

2. Neuroradiological findings of atypical meningiomas Noriaki Tomuraa, *, Satoshi Takahashia, Ikuo Sakumaa, Koichi Omachia, Jiro Wataraia, Toshio Sasajimab, Hiroyuki Kinouchib, Kazuo Mizoib CMIG Extra: Cases 28 (2004) 33-39.

3. Romani R, Tang WJ, Mao Y. Diffusion tensor magnetic resonance imaging for predicting the consistency of intracranial meningiomas. Acta Neurochir (Wien). 2014 Oct; 156 (10): 1837-45.

4. Magnetic Resonance Imaging Workbench: Analysis and Visualization of Dynamic Contrast-enhanced MR Imaging Data. RadioGraphics 2006; 26: 621-632. Published online 10.1148 / rg. 262045187.

5. US patent No. 4856528, JOHN HOPKINS UNIVERSITY, 08/15/1989.

6. MRI Brain image segmentation using graph cuts. Mohammad Shajib Khadem, Signal Processing Group Department of Signals and Systems CHALMERS UNIVERSITY OF TECHNOLOGY Goteborg, Sweden, 2010.

7. Determination of Grade and Subtype of Meningiomas by Using Histogram Analysis of Diffusion-Tensor Imaging Metrics. Sumei Wang et al. Radiology: Volume 262: Number 2-February 2012.

8. David N. Louis, Hiroko Ohgaki, Otmar D. Wiestler, Webster K. Cavenee, Peter C. Burger, Anne Jouvet, Bernd W. Scheithauer, and Paul Kleihuescorresponding author / The 2007 WHO Classification of Tumours of the Central Nervous System. 2007.

9. Prashanth Giridhar, Supriya Mallick, K.P. Haresh, Subhash Gupta, P.K. Julka, G.K. Rath / Intracranial fibrosarcoma treated with adjuvant radiation and temozolomide: Report of a case and review of all published cases / Journal of the Egyptian National Cancer Institute. July 4, 2015.

10. Modha, Ashok M.D .; Gutin, Philip H. M. D. Diagnosis and Treatment of Atypical and Anaplastic Meningiomas: A Review Neurosurgery: September 2005 - Volume 57 - Issue 3 - pp 538-550.

Claims (1)

A method for preoperatively determining the degree of malignancy and the histological type of brain meningiomas, including the analysis of MRI images in T1 mode with contrasting, comprising the steps of:
- determine the intensity of each pixel in the tumor area in contrast MRI T1 weighted images;
- normalize the intensity of each pixel on the intact tissue of the white matter of the patient’s brain, taking into account the displacement coefficient of the histogram relative to the average background color of the MRI database of images of patients with tumors of the meninges of the brain;
- form a histogram of normalized pixel intensity in MRI images;
- determine the position of the peak of the histogram and based on a comparison of its value with the values of different histological types of tumors of the meninges indicated in the database, determine the histological type of tumor and the corresponding degree of malignancy.

Images