KR102144721B1 - How to increase the chances of clinical trial success probability by selecting gruops of patients with parkinson's disease - Google Patents

Abstract

The present invention relates to a method for increasing a clinical trial success probability by selecting a Parkinson′s disease patient group with an artificial intelligence-based Parkinson′s disease diagnosis device including an image acquisition unit, an image processing unit, an image analysis unit, a diagnosis unit, and a central management unit. The method includes: a first step of the image acquisition unit acquiring multi echo intensity- and phase-related first images from brain MRI of patients constituting a clinical trial candidate group for drug efficacy verification; a second step of the image processing unit post-processing the first images so that nigrosome-1 area and substantia nigra used as a Parkinson′s disease imaging biomarker can be observed; a third step of the image analysis unit classifying second images including the nigrosome-1 area by analyzing the post-processed first images; a fourth step of the image analysis unit detecting the nigrosome-1 area from the classified second images; a fifth step of the image analysis unit providing the diagnosis unit with information on the detected nigrosome-1 area in order to diagnose the presence or absence of Parkinson′s disease in the patient; a sixth step of the central management unit receiving information on at least one first patient having Parkinson′s disease and constituting the patients from the diagnosis unit; and a seventh step of the central management unit verifying the drug efficacy based on the clinical trial on the first patient.

Description

HOW TO INCREASE THE CHANCES OF CLINICAL TRIAL SUCCESS PROBABILITY BY SELECTING GRUOPS OF PATIENTS WITH PARKINSON'S DISEASE}

The present invention relates to a method of increasing the probability of success of a clinical trial by selecting a patient group with Parkinson's disease, and more particularly, to diagnose Parkinson's disease by analyzing Magnetic Resonance Imaging (hereinafter referred to as'MRI'). , It relates to a method to increase the probability of success in clinical trials by selecting a group of patients with Parkinson's disease.

Neurodegenerative disease refers to a disease that causes abnormal brain function as nerve cells are destroyed by some cause.

Representative neurodegenerative diseases include Alzheimer's disease, Parkinson's disease, and, rarely, Lou Gehrig's disease.

Among neurodegenerative diseases, Parkinson's disease is a representative neurodegenerative disease in which nerve cells are destroyed, and symptoms of hardening of the body, trembling hands and feet, difficulty walking, depression, and anxiety are accompanied, which greatly reduces the quality of life.

As a method of diagnosing such a neurodegenerative disease, a non-invasive method of diagnosis without contact with mucous membranes, skin destruction, or internal body cavity other than natural or artificial bodies is diagnosed.

For example, in Patent Document 1 and Patent Document 2 below, techniques for diagnosing neurodegenerative diseases according to the prior art are disclosed.

1. Korean Patent Registration No. 10-1754291 (announced on July 6, 2017) 2. Korean Patent Publication No. 10-2016-0058812 (published on May 25, 2016)

1. Sung-Min Gho et al., "Susceptibility Map-Weighted Imaging (SMWI) for Neuroimaging", Magnetic Resonance in Medicine 72:337-346(2014) 2. Yoonho Nam et al., "Imaging of Nigrosome 1 in Substantia Nigra at 3T Using Multiecho Susceptibility Map-Weighted Imaging (SMWI)", J. MAGN. RESON. IMAGING 2017; 46:528-536 3. Christian Langkammer et al., "Quantitative susceptibility mapping (QSM) as a means to measure brain iron? A post mortem validation study", Neuroimage. 2012 Sep; 62(3):1593-9. doi: 10.1016/j.neuroimage. 2012.05.049. Epub 2012 May 24.

[18F]FP-CIT Positron emission tomography (PET) using isotopes has been used as the most objective method for the diagnosis of Parkinson's disease and differential diagnosis of drug-induced Parkinsonism to date.

However, the [18F]FP-CIT PET is a very expensive inspection method, and there is a risk of exposure to radiation.

Therefore, there is a need to develop a technology capable of diagnosing Parkinson's disease by observing the Nigrosome 1 region using MRI.

An object of the present invention is to solve the above problems, and to provide a Parkinson's disease diagnosis apparatus and method capable of diagnosing Parkinson's disease by analyzing MRI.

Another object of the present invention is to provide a Parkinson's disease diagnosis apparatus and method capable of diagnosing Parkinson's disease by analyzing the nigrosome 1 region presented as an imaging biomarker of Parkinson's disease.

In addition, it is an object of the present invention to additionally perform pre-processing processes such as angle adjustment, image enlargement, and reslice for SMWI images, so that the black matter and Nigrosome 1 region presented as image biomarkers of Parkinson's disease in the obtained image It is to provide an apparatus and method for diagnosing Parkinson's disease that can make the gulf observable.

In addition, an object of the present invention is to detect red nucleus and substantia nigra that are easier to detect around through machine learning, and detect an image at the moment when red nucleu disappears after both red nucleus and substantia nigra exist, nigrosome- It is to provide an apparatus and method for diagnosing Parkinson's disease that can effectively detect one area.

In addition, an object of the present invention is to generate a plurality of learning models with the same data set (SET), and to derive a predictive diagnosis result when each of the plurality of learning models is applied, and when the diagnosis results between the plurality of learning models are different, majority vote It is to provide a Parkinson's disease diagnosis apparatus and method capable of determining the final diagnosis for normal or Parkinson's disease according to the principle of.

In addition, it is an object of the present invention to provide an apparatus, system, and method for increasing the probability of success in clinical trials by utilizing the diagnosis method of Parkinson's disease using artificial intelligence to select a patient group and a normal group.

On the other hand, the technical problems to be achieved in the present invention are not limited to the technical problems mentioned above, and other technical problems that are not mentioned are clearly to those of ordinary skill in the technical field to which the present invention belongs from the following description. It will be understandable.

Patients with Parkinson's disease using an artificial intelligence-based Parkinson's disease diagnosis apparatus including an image acquisition unit, an image processing unit, an image analysis unit, a diagnosis unit, and a central management unit, which are an aspect of the present invention for achieving the above technical problem. In the method of increasing the probability of success of a clinical trial by selecting, from MRI photographing the brains of a plurality of patients who are candidates for a clinical trial experiment for demonstrating drug efficacy, the image acquisition unit may include a plurality of images related to the size and phase of a multi echo. A first step of obtaining a first image of; A second step of post-processing the acquired plurality of first images by the image processing unit to enable observation of the nigrosome 1 region and the black matter used as image biomarkers of Parkinson's disease; A third step of classifying a plurality of second images including the nigrosome 1 region by analyzing the plurality of post-processed first images by the image analysis unit; A fourth step of detecting, by the image analysis unit, the nigrosome 1 region from the classified second images; A fifth step of providing information on the detected Nygrosome 1 region to the diagnosis unit in order to diagnose the presence or absence of Parkinson's disease in the patient; A sixth step of receiving, by the central management unit, information on at least one first patient with Parkinson's disease among the plurality of patients from the diagnosis unit; And a seventh step in which the central management unit proves the efficacy of the drug on the basis of a clinical trial on the first patient. Including, wherein in the second step, the image processing unit applies a quantitative susceptibility map mask to the plurality of first images based on a quantitative susceptibility mapping algorithm to generate a susceptibility map weighted imaging image, and the post-processing And between the second and third steps, the image processing unit further performs at least one of angle adjustment, image enlargement, and reslice on the generated susceptibility map weighted imaging image. Step 2.5; may further include.

Meanwhile, an artificial intelligence based Parkinson's disease diagnosis apparatus including an image acquisition unit, an image processing unit, an image analysis unit, and a diagnosis unit, which is another aspect of the present invention for achieving the above technical problem; And a server that communicates with the artificial intelligence-based Parkinson's disease diagnosis apparatus; in a method of increasing the probability of success of a clinical trial by selecting a patient group with Parkinson's disease, a plurality of patients who are candidates for a clinical trial for demonstrating drug efficacy A first step of obtaining, by the image acquisition unit, a plurality of first images related to the magnitude and phase of a multi-echo from the MRI photographed of the brain; A second step of post-processing the acquired plurality of first images by the image processing unit to enable observation of the nigrosome 1 region and the black matter used as image biomarkers of Parkinson's disease; A third step of classifying a plurality of second images including the nigrosome 1 region by analyzing the plurality of post-processed first images by the image analysis unit; A fourth step of detecting, by the image analysis unit, the nigrosome 1 region from the classified second images; A fifth step of providing information on the detected Nygrosome 1 region to the diagnosis unit in order to diagnose the presence or absence of Parkinson's disease in the patient; A sixth step of receiving, by the server, information on at least one first patient with Parkinson's disease among the plurality of patients from the diagnosis unit; And a seventh step of verifying, by the server, the efficacy of the drug based on the clinical trial on the first patient. Including, wherein in the second step, the image processing unit, based on a quantitative susceptibility mapping algorithm, by applying a quantitative susceptibility map mask to the plurality of first images to generate a susceptibility map weighted imaging image, the post-processing And between the second and third steps, the image processing unit further performs at least one of angle adjustment, image enlargement, and reslice on the generated susceptibility map weighted imaging image. Step 2.5; may further include.

As described above, according to the method for increasing the success probability of a clinical trial by selecting a patient group with Parkinson's disease according to the present invention, only the image containing the nigrosome 1 region in MRI is classified, and the nigrosome in the classified image The effect of being able to diagnose Parkinson's disease is obtained by analyzing the 1 area.

And according to the present invention, by applying a susceptibility map weighted imaging protocol and a quantitative susceptibility mapping algorithm to improve the visibility of the Nigrosome 1 region, and diagnose Parkinson's disease using an image with a visualized black matter structure, the universally supplied Using MRI equipment, Parkinson's disease can be accurately diagnosed, and the accuracy of diagnosis results can be improved.

In addition, the present invention further performs pre-processing such as angle adjustment, image magnification, and reslice for the SMWI image, thereby observing only the black matter and the Nigrosome 1 region presented as an image biomarker of Parkinson's disease in the obtained image. You can make it possible.

In addition, the present invention detects red nucleus and substantia nigra that are easier to detect around through machine learning, and detects an image at the moment when red nucleu disappears after both red nucleus and substantia nigra exist, and nigrosome-1 region Can be detected effectively.

In addition, the present invention generates a plurality of learning models with the same data set (SET), and derives predictive diagnosis results when each of the plurality of learning models is applied, and when the diagnosis results between the plurality of learning models are different, the principle of majority vote Depending on whether it is normal or Parkinson's disease, the final diagnosis can be determined.

In addition, in clinical trials for demonstrating drug efficacy, results are determined by showing statistical significance whether or not a predicted expected effect is achieved for clinical trial participants. When applying the Parkinson's disease diagnosis method and apparatus according to the present invention, By including only patients with Parkinson's disease targeted by the new drug as clinical trial subjects, the probability of successful clinical trials can be increased as much as possible.

That is, through the Parkinson's disease diagnosis method using artificial intelligence according to the present invention, it can be used to select patient groups and normal groups to increase the probability of success in clinical trials.

Meanwhile, the effects obtainable in the present invention are not limited to the above-mentioned effects, and other effects not mentioned will be clearly understood by those of ordinary skill in the art from the following description. I will be able to.

1 is a diagram illustrating MRI including a region of black matter and nigrosome 1 in relation to the present invention.
2 is a block diagram of a Parkinson's disease diagnosis apparatus according to a preferred embodiment of the present invention.
3 is a diagram illustrating an operation of an image acquisition unit and an image processing unit according to the present invention.
FIG. 4 is a diagram illustrating an operation of additionally performing a preprocessing process such as angle adjustment, image enlargement, and reslice for an SMWI image by an image processing unit according to the present invention.
5 illustrates an example of a result of an image processing unit adjusting an angle and expanding an image of an SMWI image in relation to FIG. 4.
6 and 7 are views for explaining the operation of the image analysis unit according to the present invention.
8 is a diagram illustrating a method in which an image analysis unit according to the present invention can more effectively detect a nigrosome-1 region.
9 is a diagram illustrating an operation of a diagnosis unit according to the present invention.
10 is a diagram illustrating a state in which a diagnosis result is applied to an input image according to the present invention.
11 is a diagram illustrating a method in which a diagnosis unit according to the present invention can improve diagnosis performance by combining multiple prediction results.
12 shows an example of a process chart for explaining step-by-step a method for diagnosing Parkinson's disease according to a preferred embodiment of the present invention.
13 is a flowchart illustrating a post-processing process so that only the black matter and the Nigrosome 1 region presented as an image biomarker of Parkinson's disease can be observed in an image obtained according to a preferred embodiment of the present invention.
14 is a flowchart illustrating a method of analyzing an image processed according to a preferred embodiment of the present invention, classifying an image including a Nigrosome 1 region, and detecting a Nygrosome 1 region in the classified image.
FIG. 15 is a flowchart illustrating a process of diagnosing Parkinson's disease by determining whether or not the Nigrosome 1 region is normal in each classified image.
FIG. 16 is a diagram illustrating a method of increasing the probability of success in a clinical trial by using artificial intelligence to select a patient group and a normal group through a Parkinson's disease diagnosis method.

Hereinafter, a preferred embodiment of the present invention will be described with reference to the drawings. In addition, the examples described below do not unreasonably limit the content of the present invention described in the claims, and the entire configuration described in the present embodiment cannot be said to be essential as a solution to the present invention.

Hereinafter, an apparatus and method for diagnosing Parkinson's disease according to a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

Criteria for diagnosing Parkinson's disease by observing the nigrosome 1 area in the black matter

First, the criteria for diagnosing Parkinson's disease by observing the nigrosome 1 region in the black matter will be described with reference to FIG. 1.

1 is a diagram illustrating an MRI including a black matter and a nigrosome 1 region.

Fig. 1(a) shows an imaging slab, and Fig. 1(b) shows an image of a normal person's substantia nigra and a nigrosome 1 area, and Fig. 1(c) ) Shows images of the nigrosome 1 region and the black matter of Parkinson's disease patients.

The nigrosome 1 area in the black matter of a normal person is displayed in black as shown in FIG. 1(b), and the nigrosome 1 area in the Parkinson's disease patient is relatively gray as shown in FIG. 1(c). It is displayed close.

Accordingly, the present invention can diagnose the presence or absence of Parkinson's disease by observing the shadow of the nigrosome 1 region presented as an imaging biomarker of Parkinson's disease in MRI.

For example, the present invention applies the susceptibility map weighted imaging (Susceptibility Map Weighted Imaging, hereinafter'SMWI') protocol and the internal quantitative susceptibility mapping (Qluantitative Susceptibility Mapping, hereinafter'QSM') algorithm in 3T MRI Parkinson's disease is diagnosed by analyzing whether the nigrosome 1 area is normal by improving the visibility of the grosome 1 area.

The substantia nigra pars compacta is a midbrain structure containing a high-density population of dopaminergic neurons. These neurons are gradually lost in idiopathic Parkinson's disease (IPD), resulting in disability. This area shows an increase in iron concentrations in IPD patients compared to healthy controls.

Recently, as a result of visualizing a small part of the dense nigrosome known as nigrosome 1 in the high-resolution susceptibility contrast image of a healthy subject, the contrast between nigrosome 1 and the substantia nigra regions around it was determined by the difference in iron concentration. As a result, it was found that the difference in susceptibility between the two sites in IPD patients was significantly reduced.

This reduction in the susceptibility difference between the two sites is used as an imaging biomarker for IPD.

Therefore, the Nigrosome 1 structure has been successfully depicted on 7T MRI using high-resolution (e.g. 0.3 mm planar resolution) T2-weighted imaging or susceptibility weighted imaging (SWI).

However, due to the limited spatial resolution and signal/contrast-to-noise ratio (SNR/CNR) in low magnetic field strength, for example, 3T MRI, a structure with significantly reduced contrast, that is, contrast, is observed in 3D high-resolution T2-weighted imaging.

These limitations hampered the reliability and applicability of Nigrosome 1 imaging in 3T MRI, despite several successful studies demonstrating the usefulness of the approach.

Recently, in order to solve the above problems, a new method for providing an improved magnetic susceptibility contrast has been proposed.

One of them is to use multi-echo gradient recall echo (hereinafter referred to as'multi-echo GRE') instead of using a single echo image to improve SNR. By combining 3T MRI, it shows a relatively high accuracy in diagnosing IPD.

An alternative to the size image is susceptibility-weighted imaging (SWI), which uses phase information as a weight mask to increase susceptibility contrast, or artifacts due to blooming of phase imaging. ) Can be created.

Other approaches related to susceptibility contrast are GRE phase (or frequency) images and quantitative susceptibility mapping (QSM), both of which are highly sensitive to susceptibility and have been widely applied in recent years.

In addition, a new method using a susceptibility weighting mask derived from QSM has been proposed for a sized image.

This approach is similar to SWI, but it solves SWI's blooming artifacts and can potentially improve the visualization of susceptibility changes.

QSM mask weighted imaging has proven useful in visualizing the Nigrosome 1 structure.

The non-patent document 1 discloses an SMWI technology for neural imaging, and the non-patent document 2 discloses an imaging technology of nigrosome 1 in the black matter using a multi-echo SMWI in 3T MRI. Patent Document 3 discloses a QSM technology for measuring brain iron.

Accordingly, the present invention diagnoses Parkinson's disease based on the change in the Nigrosome 1 region due to Parkinson's disease according to the correlation between the concentration of brain iron and the susceptibility.

Parkinson's disease diagnostic device

2 is a block diagram of an apparatus for diagnosing Parkinson's disease according to a preferred embodiment of the present invention.

Parkinson's disease diagnosis apparatus 10 according to a preferred embodiment of the present invention, as shown in Figure 2, the image acquisition unit 20 for acquiring a multi-echo size and phase image from the MRI of the patient's brain, the acquired An image processing unit 30 that post-processes so that only the nigrosome 1 area and the black matter presented as an image biomarker of Parkinson's disease can be observed in the image, and the processed image is analyzed to classify and classify the image containing the Nigrosome 1 area. An image analysis unit 40 for detecting the nigrosome 1 area in the image and a diagnosis unit 50 for diagnosing Parkinson's disease by determining whether the nigrosome 1 area in each classified image is normal.

Hereinafter, the roles and functions of each component will be described in detail with reference to the drawings.

Operation of image acquisition unit and image processing unit

With reference to FIG. 3, the configuration of the image stroke feature and the image processing unit will be described in detail.

3 is a diagram illustrating an operation of an image acquisition unit and an image processing unit. FIG. 3 shows a process of generating an SMWI image for a Nygrosome 1 structure from a multi-echo synthesis GRE image.

As shown in FIG. 3, the image acquisition unit 20 is communicatively connected to the MRI equipment 21 or a database (not shown) that stores and manages MRI photographed in the database to diagnose Parkinson's disease. Acquire the patient's MRI.

As shown in FIG. 3, the image processing unit 30 is a multi-echo GRE composite image in which a multi-echo size image and a multi-echo phase image are combined using a QSM algorithm for visualization of the Nygrosome 1 structure. Generates an SMWI image.

For example, the image processing unit 30 generates a channel-combined size image from the multi-channel composite image by the square root of the sum of the squares of the multi-channel size image, and the phase image is a complex average after correcting the global phase offset of the individual channels. Combined (step 1).

Then, the image processing unit 30 combines the size of the six echoes into a single image by the square root of the sum of squares of the images (step 2).

The image processing unit 30 calculates phase images of different TEs using a Laplacian unwrapping algorithm, and calculates a combined frequency w in each voxel (step 3).

The image processing unit 30 removes the background region from the frequency image by using the harmonic background phase removal using the Laplacian operator method (step 4).

Here, the QSM may be reconstructed using an improved sparse linear equation and least-squares (iLSQR) method.

For example, in the reconstruction parameter of iLSQR, error tolerance = 0.01, threshold D ₂ for an incomplete k-region mask, _{and thres} = 0.1.

Subsequently, the image processing unit 30 additionally processes the resulting QSM to generate a QSM mask S _mask for weights versus susceptibility (step 5).

The mask may be generated using Equation 1.

[Equation 1]

Here, X is the quantitative susceptibility value calculated in step 4 (in ppm), and X _th is the paramagnetic threshold. The threshold can later be determined using the Nigrosome 1 imaging data for the optimal CNR.

Finally, the image processing unit 30 may generate the SMWI image by multiplying the multi-echo-combined size image by the QSM mask using Equation 2 below.

[Equation 2]

Here, m is the number of multiplications for the susceptibility weight, and mag is the multi-echo size combined image of the second step.

On the other hand, in the present invention, by additionally performing pre-processing processes such as angle adjustment, image enlargement, and reslice on the generated SMWI image, only the black matter and Nigrosome 1 region presented as the image biomarkers of Parkinson's disease in the obtained image. You can make it observable.

4 is a view for explaining an operation of additionally performing a preprocessing process such as angle adjustment, image enlargement, and reslice for an SMWI image by an image processing unit according to the present invention.

Referring to FIG. 4A, an example of an SMWI image generated by multiplying a multi-echo combined size image by a QSM mask using Equation 2 is shown.

Thereafter, as shown in (b) of FIG. 4, the image processing unit 30 may perform an angle adjustment and an image enlargement operation.

First, the operation of adjusting the angle of the image processing unit 30 is performed to correct image distortion due to movement of a photographer during photographing.

In addition, the image enlargement operation of the image processing unit 30 is performed to enlarge the nigrosome-1 area.

Specifically, the angle adjustment according to the present invention is the angle in which nigrosome-1 is well seen by checking information on each field value of medical image information (dicom->image orientation patient, image position patient, slice location, slice thickness, etc.) Can be adjusted to

Further, the image magnification according to the present invention can be performed by dividing the entire image into squares into nine and expanding the center square area by three times the width and length.

5 illustrates an example of a result of an image processing unit adjusting an angle and expanding an image of an SMWI image in relation to FIG. 4.

Referring to (a) of FIG. 5, the axial image and the red line are represented by a coronal plane, and the result of the angle adjustment operation is shown in (b) of FIG. Has been.

Referring to FIG. 5(b), image distortion due to the motion of the photographer during shooting is corrected, and nigrosome-1 is adjusted to a well-visible angle, which can be provided by expanding the nigrosome-1 area through the enlargement operation. have.

Returning to FIG. 4 again, after performing the operation of adjusting the angle and expanding the image in FIG. 4(b), the thickness of the MR image is further increased as shown in FIG. 4(c). A process of reslicing may be performed to make more images including the nigrosome-1 region by fine adjustment.

This reslice process is to make more images including nigrosome-1 regions by controlling and applying the thickness of the MR image more precisely than before, as described above.

Through the reslice process of the present invention, it is possible to additionally increase about 2 to 3 images, which can generate 2 to 3 additional images in addition to the existing 2 to 3 images. It is possible to generate more images containing more than twice the nigrosome-1 region.

Typically, in the reslice process according to the present invention, a region including nigrosome-1 may be enlarged by adjusting the slice thickness of the input mr image information to 0.5 or less.

Operation of the image analysis unit

Next, the operation of the image analysis unit will be described in detail with reference to FIGS. 6 and 7.

6 and 7 are diagrams for explaining the operation of the image analysis unit. FIG. 6 illustrates a process of classifying an image including the nigrosome 1 region, and FIG. 7 shows a process of detecting the nigrosome 1 region from the classified image.

As illustrated in FIG. 6, through a post-processing process of the image processing unit 30, a plurality of SWMI images may be acquired.

In general, about 40 to 70 MRIs are taken per patient to diagnose Parkinson's disease, and among them, about 3 to 6 or less images including the Nigrosome 1 region used for Parkinson's disease diagnosis.

As shown in FIG. 6, the image analysis unit 40 analyzes the image including the Nygrosome 1 region among all MRIs through machine learning, and analyzes the image including the Nygrosome 1 region and the image not included. Images can be classified.

In addition, as shown in FIG. 7, the image analysis unit 40 designates a location where the nigrosome 1 area exists in each image including the nigrosome 1 area classified through machine learning, and the designated nigrosome 1 Area can be detected.

For example, the image analysis unit 40 may classify a region including the Nygrosome 1 region by using a one-stage detector method among methods using a deep learnig neural network of machine learning.

That is, the image analysis unit 40 uses a convolutional neural network (hereinafter referred to as'CNN') on the acquired image to provide a feature map having features of a fully convolutional layer. Is detected, and cross-scale connections are formed by applying a feature pyramid network (hereinafter referred to as'FPN') in the feature map, and classification loss and bounding for the classification result By optimizing the bounding-box regression loss and the focal loss to minimize unnecessary classification, it is possible to finally minimize only the image containing the Nigrosome 1 region in the image of various sizes.

8 is a diagram illustrating a method in which an image analysis unit according to the present invention can more effectively detect a nigrosome-1 region.

Referring to FIG. 8, the image analysis unit 40 according to the present invention detects red nucleus and substantia nigra that are easier to detect around through machine learning, and both red nucleus and substantia nigra exist, but red nucleu disappears. By detecting the instantaneous image, the nigrosome-1 region can be effectively detected.

Referring to (a) of FIG. 8, an image containing the Nygrosome 1 region among all MRIs is analyzed through machine learning to classify an image containing and without a Nygrosome 1 region (classification). )can do.

Next, referring to (b) of FIG. 8, in order to designate a region containing nigrosome-1 in an input image, deep learning through bounding boxes of object and class prediction Deep learning outputs can be presented.

In the present invention, in order to detect and analyze the location of substantia nigra, nigrosome-1 and red nucleus in the input image, a feature map in which location information is not lost is generated through a CNN neural network, and the feature map is analyzed by applying the FPN method. A method corresponding to the change in the size of the object can be applied.

That is, referring to (c) of FIG. 8, based on images associated with nigrosome-1 (Images with nigrosome 1 regions), an image that has undergone preprocessing is a specific position (R0.7 P28.4 F5.2) mm) and angles (C> T39.6> S1.4).

In other words, the preprocessed image has a specific position (R0.7 P28.4 F5.2 mm) and an angle (C> T39.6> S1.4), and thus the nigrosome required for Parkinson's diagnosis. In order to detect the -1 region, we can first detect red nucleus and substantia nigra, which are more easily detectable, through machine learning.

Specifically, in (c) of FIG. 8, images are sequentially checked in the Inferior direction, and the image at the moment when the red nucleus disappears after both red nucleus and substantia nigra exist can be detected. have.

Thereafter, assuming that the nigrosome-1 region exists in about 2 to 3 mm slice from the detected image, a number of images with a thickness of 2 to 3 mm including the nigrosome-1 region among the entire MR images (e.g., 5 to 6 Chapter) can be detected.

By selecting at least one image among images of 2 to 3 mm thickness including the nigrosome-1 region, in Figs. 8(d) and (e), each image including the classified Nigrosome 1 region is It is possible to designate a location where a little 1 area exists, and detect the designated Nygrosome 1 area.

Operation of the diagnostic unit

Next, an operation of the diagnosis unit 50 according to the present invention will be described with reference to FIGS. 9 and 10.

9 is a diagram illustrating an operation of a diagnosis unit, and FIG. 10 is a diagram illustrating a state in which a diagnosis result is applied to an input image.

As illustrated in FIG. 9, the diagnosis unit 50 may diagnose the presence or absence of Parkinson's disease by analyzing whether the Nygrosome 1 region detected by the image analysis unit 40 is normal through machine learning.

For example, the diagnosis results may be applied as shown in FIGS. 10E to 10H according to the diagnosis result of Parkinson's disease in the input images of FIGS. 10A to 10D.

Meanwhile, in the present invention, in order to improve diagnostic performance, a method in which the diagnostic unit 50 combines multiple prediction results to make a determination may be applied.

11 is a diagram illustrating a method in which a diagnosis unit according to the present invention can improve diagnosis performance by combining multiple prediction results.

Referring to FIG. 11, on the basis of the same data set (SET) displayed in (a), as illustrated in (b) of FIG. 11, several learning models may be generated.

In addition, as shown in (c) of FIG. 11, predictive diagnosis results when applying each of the plurality of learning models are derived, and in (d), when the diagnosis results between the plurality of learning models are different, the majority vote is applied. This will determine the final diagnosis for normal or Parkinson's disease.

Therefore, by determining the final diagnosis according to the principle of majority vote when the diagnosis results between models are different, it is possible to finally diagnose whether the target is normal or a patient by combining the prediction results of multiple deep learning models, thereby improving diagnosis performance. .

How to diagnose Parkinson's disease

Next, a method for diagnosing Parkinson's disease according to a preferred embodiment of the present invention will be described in detail with reference to FIG. 12.

12 is a flowchart illustrating step-by-step a method for diagnosing Parkinson's disease according to a preferred embodiment of the present invention.

In step S10 of FIG. 12, the image acquisition unit 20 acquires multiple MRIs of a patient who wants to diagnose Parkinson's disease through communication with the MRI device 21 or a database.

Here, the image acquisition unit 20 may acquire a multi-echo GRE composite image in which a multi-echo size image and a multi-echo phase image are combined in 3T MRI.

In step S12, the image processing unit 30 post-processes a multi-echo GRE composite image in which a multi-echo size image and a multi-echo phase image are combined using a QSM algorithm for visualization of the Nygrosome 1 structure, and then SMWI. Create an image.

In step S14, the image analysis unit 40 analyzes the image including the Nygrosome 1 region from the entire MRI through machine learning, and classifies the image containing the Nygrosome 1 region and the image not included. )do.

In addition, the image analysis unit 40 designates a location where the nigrosome 1 area exists in each image including the nigrosome 1 area classified through machine learning, and detects the designated nigrosome 1 area (S16).

Finally, the diagnosis unit 50 determines whether or not the Nigrosome 1 area detected in each image classified as an image including the Nigrosome 1 area by the image analysis unit 40 is normal, and diagnoses the presence of Parkinson's disease ( S18).

Hereinafter, specific processes of steps S12, S14, S16, and S18 will be described in detail with reference to the drawings.

Post-processing to generate an SMWI image (S12)

13 is a flowchart illustrating a post-processing process so that only the black matter and the Nigrosome 1 region presented as an image biomarker of Parkinson's disease can be observed in an image obtained according to a preferred embodiment of the present invention.

Referring to FIG. 13, the above-described image processing unit 30 is a multi-echo GRE composite image in which a multi-echo size image and a multi-echo phase image are combined using a QSM algorithm for visualization of the Nygrosome 1 structure. A detailed step of the S12 process of generating an SMWI image by post-processing is shown.

First, a channel-combined size image is generated from the multi-channel composite image by the square root of the sum of the squares of the multi-channel size image, and the phase image is combined with a complex average after correcting the global phase offset of individual channels (S20). It goes on.

Thereafter, a step (S21) of combining the six echoes into a single image by the square root of the sum of squares of the size of the image is performed.

After step S21, phase images of different TEs are calculated using a Laplacian unwrapping algorithm, and a combined frequency w is calculated in each voxel (S22).

In addition, a step (S23) of removing a background region from a frequency image using a harmonic background phase removal using a Laplacian operator method proceeds.

Thereafter, the QSM is further processed to generate a QSM mask (Smask) for the susceptibility versus the weight (S24), and the multi-echo combined size image is multiplied by the QSM mask to generate an SMWI image (S25). do.

In addition, after the SMWI image generated by post-processing, adjusting the image angle and magnifying (S26) is performed, the thickness of the MR image is further adjusted to increase the number of images including the nigrosome-1 region. By reslicing to make (S27), a post-processing process is performed so that only the nigrosome 1 region and the black matter presented as an image biomarker of Parkinson's disease in the acquired image can be observed.

The process of classifying the image with and without the Nigrosome 1 region (S14), and detecting the designated Nygrosome 1 region (S16)

14 is a flowchart illustrating a method of analyzing an image processed according to a preferred embodiment of the present invention to classify an image including a Nigrosome 1 region, and to detect a Nygrosome 1 region from the classified image.

Referring to FIG. 14, a detailed description is given of a process of classifying an image including and without a Nygrosome 1 region (S14) and detecting a designated Nygrosome 1 region (S16).

First of all, a step (S31) of detecting red nucleus and substantia nigra, which are easier to detect around, is performed through machine learning.

Thereafter, images are sequentially checked in the inferior direction, and an image at the moment when the red nucleu disappears after both red nucleus and substantia nigra are present (S32) is performed.

In addition, since the nigrosome-1 region exists in about 2 to 3 mm slice from the detected image, the step of detecting 2 to 3 mm thick images (5 to 6 photos) including the nigrosome-1 region of the entire MR image (S33 ) Proceeds.

After step S33, a step (S16) of detecting the nigrosome-1 region through at least one of the images derived through step S33 is performed.

Process of diagnosing the presence of Parkinson's disease by determining whether the detected Nigrosome 1 region is normal (S18)

FIG. 15 is a flowchart illustrating a process of diagnosing Parkinson's disease by determining whether or not the Nigrosome 1 region is normal in each classified image.

Referring to FIG. 15, first, a step of generating multiple learning models using the same data set (SET) (S41) is performed, and a predictive diagnosis result when each of the plurality of learning models is applied is derived (S42).

In this case, a case in which the diagnosis results between the plurality of learning models are different may occur, and in the present invention, a final diagnosis of whether it corresponds to normal or Parkinson's disease is determined according to the principle of majority vote (S43).

Therefore, by determining the final diagnosis according to the principle of majority vote when the diagnosis results between models are different, it is possible to finally diagnose whether the target is normal or a patient by combining the prediction results of multiple deep learning models, thereby improving diagnosis performance. .

A method to increase the probability of success in clinical trials by using artificial intelligence to diagnose Parkinson's disease by selecting patients and normal groups

The method and apparatus for diagnosing Parkinson's disease according to the present invention described above can be used to select a patient group and a normal group to increase the probability of success in a clinical trial.

That is, the present invention can provide an apparatus, system, and method for increasing the probability of success in a clinical trial by using for selection of a patient group and a normal group through a Parkinson's disease diagnosis method using artificial intelligence.

In clinical trials for demonstrating drug efficacy, results are determined by showing statistical significance whether or not the predicted expected effect is achieved for clinical trial participants. When applying the Parkinson's disease diagnosis method and apparatus according to the present invention, a new drug is exactly By including only target Parkinson's disease patients as clinical trial subjects, the probability of successful clinical trials can be increased as much as possible.

First, the problems of the existing clinical trials for new drugs will be explained in advance.

In clinical trials for demonstrating drug efficacy, results are judged by showing statistical significance whether or not the expected effect predicted in advance is achieved for clinical trial participants.

On the other hand, in the case of Parkinson's disease, there is no choice but to rely on UPDRS evaluation, neurological examination, Hoehn & Yarr stage evaluation (0-5 steps), etc. to measure the improvement of symptoms due to drug efficacy, and these methods are by questionnaire and data There is a problem that the scale of is not detailed.

Therefore, in order to prove statistical significance, the value of the evaluation scale must be increased statistically significantly before and after administration or compared with the sham drug group, and the higher the predicted increase value, the smaller the number of target subjects and the likelihood of achieving statistical significance increases.

At this time, if the predicted increase value is small, the number of target subjects increases and the difficulty of verifying statistics increases.

As a result, it is very difficult to increase the evaluation scale of Parkinson's disease by one level, and thus, the possibility of passing the clinical trial is very low.

In the present invention, in order to solve this problem, it is intended to increase the probability of success in the clinical trial as much as possible by including only Parkinson's disease patients targeted by the new drug as clinical trial subjects.

One of the important factors of failure in the development of new drugs for the central nervous system is the difficulty of accurately selecting subjects and selecting drug responders.

In the case of central nervous system drugs, particularly, the response rate to placebo is high, reducing the heterogeneity of the subject group, and setting up a biomarker capable of predicting drug responsiveness is an important strategy in increasing the success rate.

In addition, in the case of Parkinson's disease, since it takes a long period to be confirmed (about 3 months), there is a problem that it is very difficult to include only Parkinson's disease patients targeted by the new drug as clinical trial subjects because screening is difficult.

Therefore, it can be utilized to increase the probability of success in clinical trials by using the method for diagnosing Parkinson's disease using artificial intelligence proposed by the present invention to select patient groups and normal groups.

FIG. 16 is a diagram illustrating a method of increasing the probability of success in a clinical trial by using artificial intelligence to select a patient group and a normal group through a Parkinson's disease diagnosis method.

Referring to FIG. 16, first, a step (S1) of recruiting candidate groups for clinical trials for demonstrating drug efficacy is performed.

Thereafter, for a plurality of experimental candidate groups, the step of obtaining an image (S10), the step of post-processing the image by applying a QSM algorithm to each image (S12), and the Nygrosome 1 region using machine learning. The step of classifying the image (S14), the step of detecting the nigrosome 1 region (S16), and the step of diagnosing the presence of Pickinson's disease (S18) based on the determination of whether the nigrosome 1 region is normal are performed.

Steps S10 to S18 have been described in detail in FIGS. 12 to 15, and repetitive descriptions are omitted for the sake of simplicity of the specification.

Thereafter, when a diagnosis result is derived through step S18, a step (S50) of dividing a plurality of experimental candidate groups into an actual Parkinson's disease patient group and a normal patient group based on the diagnosis result may proceed.

At this time, through the step of conducting a clinical trial (S52) and verifying the drug efficacy based on the clinical test results (S54) based only on the subjects classified as the actual Parkinson patient group, the Parkinson's disease patient that the new drug is exactly the target of By including only as the subject of the clinical trial, the probability of success in the clinical trial can be increased as much as possible.

As a result, through the Parkinson's disease diagnosis method using artificial intelligence according to the present invention, it can be used to select patient groups and normal groups to increase the probability of success in clinical trials.

Steps S1 to S54 described above may be performed independently of the Parkinson's disease diagnosis apparatus 10, or a separate server (not shown) or a separate central management device (not shown), and the Parkinson's disease diagnosis apparatus 10 ) May be applied to perform the entire operation.

Effects according to the invention

According to the apparatus and method for diagnosing Parkinson's disease according to the present invention, there is an effect that it is possible to diagnose the presence or absence of Parkinson's disease by classifying only the image including the nigrosome 1 region in MRI and analyzing the nigrosome 1 region in the classified image. Is obtained.

And according to the present invention, by applying a susceptibility map weighted imaging protocol and a quantitative susceptibility mapping algorithm to improve the visibility of the Nigrosome 1 region, and diagnose Parkinson's disease using an image with a visualized black matter structure, the universally supplied Using MRI equipment, Parkinson's disease can be accurately diagnosed, and the accuracy of diagnosis results can be improved.

In addition, the present invention further performs pre-processing such as angle adjustment, image magnification, and reslice for the SMWI image, thereby observing only the black matter and the Nigrosome 1 region presented as an image biomarker of Parkinson's disease in the obtained image. You can make it possible.

In addition, the present invention detects red nucleus and substantia nigra that are easier to detect around through machine learning, and detects an image at the moment when red nucleu disappears after both red nucleus and substantia nigra exist, and nigrosome-1 region Can be detected effectively.

In addition, the present invention generates a plurality of learning models with the same data set (SET), and derives predictive diagnosis results when each of the plurality of learning models is applied, and when the diagnosis results between the plurality of learning models are different, the principle of majority vote Depending on whether it is normal or Parkinson's disease, the final diagnosis can be determined.

In addition, in clinical trials for demonstrating drug efficacy, results are determined by showing statistical significance whether or not a predicted expected effect is achieved for clinical trial participants. When applying the Parkinson's disease diagnosis method and apparatus according to the present invention, By including only patients with Parkinson's disease targeted by the new drug as clinical trial subjects, the probability of successful clinical trials can be increased as much as possible.

That is, through the Parkinson's disease diagnosis method using artificial intelligence according to the present invention, it can be used to select patient groups and normal groups to increase the probability of success in clinical trials.

Meanwhile, the effects obtainable in the present invention are not limited to the above-mentioned effects, and other effects not mentioned will be clearly understood by those of ordinary skill in the art from the following description. I will be able to.

The above-described embodiments of the present invention can be implemented through various means. For example, embodiments of the present invention may be implemented by hardware, firmware, software, or a combination thereof.

In the case of hardware implementation, the method according to embodiments of the present invention includes one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), and Programmable Logic Devices (PLDs). , Field Programmable Gate Arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, and the like.

In the case of implementation by firmware or software, the method according to the embodiments of the present invention may be implemented in the form of a module, procedure, or function that performs the functions or operations described above. The software code may be stored in a memory unit and driven by a processor. The memory unit may be located inside or outside the processor, and may exchange data with the processor through various known means.

Detailed description of the preferred embodiments of the present invention disclosed as described above has been provided to enable those skilled in the art to implement and implement the present invention. Although the above has been described with reference to preferred embodiments of the present invention, those skilled in the art will understand that various modifications and changes can be made to the present invention without departing from the scope of the present invention. For example, those skilled in the art may use the configurations described in the above-described embodiments in a manner that combines with each other. Accordingly, the invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The present invention may be embodied in other specific forms without departing from the spirit and essential features of the present invention. Therefore, the detailed description above should not be construed as restrictive in all respects and should be considered as illustrative. The scope of the present invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention. The invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein. In addition, the embodiments may be configured by combining claims that do not have an explicit citation relationship in the claims, or may be included as new claims by amendment after filing.

Claims (2)

In a method of increasing the probability of success in a clinical trial by selecting a patient group with Parkinson's disease using an artificial intelligence-based Parkinson's disease diagnosis apparatus including an image acquisition unit, an image processing unit, an image analysis unit, a diagnosis unit, and a central management unit,

A first step of obtaining, by the image acquisition unit, a plurality of first images related to the magnitude and phase of a multi echo from the MRI of the brains of a plurality of patients who are candidates for a clinical trial for demonstrating drug efficacy;

A second step of post-processing the obtained plurality of first images by the image processing unit to enable observation of the nigrosome 1 region and the black matter used as image biomarkers of Parkinson's disease;

A third step of classifying a plurality of second images including the nigrosome 1 region by analyzing the plurality of post-processed first images by the image analysis unit;

A fourth step of detecting, by the image analysis unit, the nigrosome 1 region from the classified second images;

A fifth step of providing information on the detected Nygrosome 1 region to the diagnosis unit in order to diagnose the presence or absence of Parkinson's disease in the patient;

A sixth step of receiving, by the central management unit, information on at least one first patient with Parkinson's disease among the plurality of patients from the diagnosis unit; And

A seventh step in which the central management unit proves the efficacy of the drug based on the clinical trial on the first patient; Including,

In the second step,
The image processor performs the post-processing by applying a quantitative susceptibility map mask to the plurality of first images based on a quantitative susceptibility mapping algorithm to generate a susceptibility map weighted imaging image, and

Between the second step and the third step,
The presence of Parkinson's disease further comprising: performing at least one of an angle adjustment, an image enlargement, and a reslice on the generated susceptibility map weighted imaging image by the image processing unit. How to increase the probability of success in clinical trials by selecting a group of patients who do.
An artificial intelligence based Parkinson's disease diagnosis apparatus including an image acquisition unit, an image processing unit, an image analysis unit, and a diagnosis unit; And a server communicating with the artificial intelligence-based Parkinson's disease diagnosis apparatus; in a method for increasing a clinical trial success probability by selecting a patient group with Parkinson's disease,

A first step of obtaining, by the image acquisition unit, a plurality of first images related to the magnitude and phase of a multi echo from the MRI of the brains of a plurality of patients who are candidates for a clinical trial for demonstrating drug efficacy;

A second step of post-processing the obtained plurality of first images by the image processing unit to enable observation of the nigrosome 1 region and the black matter used as image biomarkers of Parkinson's disease;

A third step of classifying a plurality of second images including the nigrosome 1 region by analyzing the plurality of post-processed first images by the image analysis unit;

A fourth step of detecting, by the image analysis unit, the nigrosome 1 region from the classified second images;

A fifth step of providing information on the detected Nygrosome 1 region to the diagnosis unit in order to diagnose the presence or absence of Parkinson's disease in the patient;

A sixth step of receiving, by the server, information on at least one first patient with Parkinson's disease among the plurality of patients from the diagnosis unit; And

A seventh step of verifying, by the server, the efficacy of the drug based on the clinical trial on the first patient; Including,

In the second step,
The image processor performs the post-processing by applying a quantitative susceptibility map mask to the plurality of first images based on a quantitative susceptibility mapping algorithm to generate a susceptibility map weighted imaging image, and

Between the second step and the third step,
The presence of Parkinson's disease further comprising: performing at least one of an angle adjustment, an image enlargement, and a reslice on the generated susceptibility map weighted imaging image by the image processing unit. How to increase the probability of success in clinical trials by selecting a group of patients who do.

Images