WO2019098415A1

WO2019098415A1 - Method for determining whether subject has developed cervical cancer, and device using same

Info

Publication number: WO2019098415A1
Application number: PCT/KR2017/013015
Authority: WO
Inventors: 정재훈; 최성원
Original assignee: 주식회사 버즈폴
Priority date: 2017-11-16
Filing date: 2017-11-16
Publication date: 2019-05-23
Also published as: KR20190087681A; CN110352461A

Abstract

The present invention relates to a method for determining whether a subject has developed cervical cancer, and a determination device using the same. Specifically, a determination device according to the present invention: acquires a captured image of a subject's cervix; generates analysis information relating to whether the subject has developed cervical cancer, on the basis of a machine learning model for cervical cancer from an input of the acquired captured image of the cervix; and provides the generated analysis information, so as to support a user of the determination device or another user at a remote place to read whether cervix cancer has developed, which corresponds to the analysis information, and acquire, store, and output evaluation information of the analysis information, as a result of the reading.

Description

A method for determining the onset of a subject for cervical cancer and a device using the same

The present invention relates to a method for judging whether or not a subject develops cervical cancer and a judgment apparatus using the same. More specifically, the determination apparatus according to the present invention is a determination apparatus for acquiring an image of the cervix of the subject, and acquiring, from the input of the obtained cervical-radiographic image, And provides analysis information on the incidence of the cervical cancer and provides the generated analysis information so as to allow the user of the determination apparatus or the remote user to read out the incidence of the cervical cancer corresponding to the analysis information Acquires and stores evaluation information on the analysis information as a result of the reading, and outputs the evaluation information. Further, the determination apparatus according to the present invention can re-learn the machine learning model based on the evaluation information.

Cervical cancer is the most common cancer ranking among Korean women, because it can be affected by pregnancy and childbirth due to hysterectomy, and it can cause a sense of loss as a woman. According to the statistics for 2013, the number of cervical cancer patients in Korea is 26,207, ranking 4th among female cancer (Ministry of Health and Welfare data). In addition, it is the 7th cancer recommendation in Korea, and it is included in the national cancer screening project in 1999, and the rate of early diagnosis is increasing. In recent years, cervical intraepithelial cancer (precancerous stage) called "0 period" cancer of the cervix is also on the rise, and it is recommended that women who have experience of sexual experience have annual checkups.

Looking at the market status of the screening, the proportion of cervical intraepithelial neoplasia in young women is increasing, and the screening target has been lowered from 30 to 20 years from 2016. In particular, unlike other cancers, health insurance benefits apply to 300% of the cost of screening for cervical cytology examinations. However, it is recommended that screening tests be conducted in parallel with the cervical screening test because the false negative rate (ie, false positive rate) of screening reaches 55%. As a result, the market for cervical cancer screening in the world is estimated at 6.86 trillion won Of these, the cervical dilatation test is 30%, reaching about 2 trillion won.

FIG. 1 is a conceptual diagram schematically showing a method of examining a cervical cell examination and a cervical dilatation examination, which has been conventionally performed to diagnose cervical cancer. Referring to the bottom of FIG. 1, (For example, the cervical lumen shown in FIG. 1), and analyzing the resultant image and using the result, the misdiagnosis rate of the examination for cervical cancer can be lowered.

However, when using a conventional cervical loupe, the medical staff confirms whether the cervical cancer has developed in relation to the image of the cervix in view of education and experience. This method is often repeated and obscure, It can take a long time and the accuracy can drop together.

The CDSS (Clinical Decision Support System) has been developed in various fields as a technology to support the repeated diagnostic physician's diagnosis, so that the efficiency of the doctor's decision is doubled, The global CDSS market is projected to grow at an annual rate of KRW24 trillion (up 25% on average) from KRW4.2tr in 2014.

In the present invention, CDSS which is specialized for cervical cancer is used as a CDSS method for improving the efficiency of cervical cancer screening and preventing misdiagnosis through a computing device, thereby enabling a medical staff to perform diagnosis of cervical cancer more quickly and accurately. And a judgment device for this purpose is proposed.

It is an object of the present invention to solve the above-mentioned problems and make it possible to quickly and accurately read the presence or absence of cervical cancer through a high-resolution imaging of the cervix taken by an obstetrician or the like.

The characteristic configuration of the present invention for achieving the object of the present invention as described above and realizing the characteristic effects of the present invention described below is as follows.

According to one aspect of the present invention, there is provided a method of determining whether a subject has developed cervical cancer, the method comprising: (a) the computing device acquiring an image of the cervix of the subject Supporting another device associated with the computing device to obtain; (b) the computing device generates analysis information on the incidence of the cervical cancer of the subject based on the obtained machine learning model of the cervical cancer from the input of the cervical radiographic image, To generate the generated data; (c) the computing device provides the generated analysis information or provides the other device with information, thereby allowing a user of the computing device or a remote user of the remote location to determine whether the cervical cancer corresponding to the analysis information has occurred Supporting to read; And (d) the computing device is configured to: (i) acquire and store evaluation information for the analysis information as a result of performing the step (c), or support the other device to acquire and store the evaluation information And (ii) performing a process of outputting the evaluation information or supporting the output of the other device.

Advantageously, the method further comprises the step of (e) re-learning the machine learning model or allowing the other device to re-learn the machine learning model based on the evaluation information .

According to another aspect of the present invention, there is also provided a computer program recorded on a machine-readable non-volatile storage medium, comprising instructions embodied to perform the above-described method.

According to still another aspect of the present invention, there is provided a computing device for determining whether a subject is affected by cervical cancer, the computing device comprising: a communication unit for obtaining an image of the cervix of the subject; And generating analysis information on the onset of the cervical cancer of the subject based on the obtained machine learning model of the cervical cancer from the input of the cervical radiographic image, Wherein the processor is configured to provide the analysis information or to provide the other apparatus with information on whether the cervical cancer corresponding to the analysis information is present or not, (I) acquiring and storing evaluation information on the analysis information as a result of the reading, or (ii) supporting the other device to acquire and store the evaluation information, and (ii) And outputs the evaluation information or causes the other device to output The process of supporting

Preferably, the processor of the computing device re-learns the machine learning model based on the evaluation information.

According to the present invention, compared to the conventional method in which a medical staff directly observes a cervical image obtained through a cervical loupe and confirms the state of the cervix individually based on education and experience, it is possible to quickly and accurately determine the onset of cervical cancer There is an effect that can be done.

Further, according to the present invention, by utilizing advanced artificial intelligence such as machine learning, in particular, deep learning technology, it is possible to prevent mistakes of the medical staff, that is, misdiagnosis.

According to the present invention, it is possible to facilitate the division of labor in the medical field by making it possible to read out the cervical cancer even in a remote place away from the photographing site of the cervical image.

The present invention has the effect of continuously improving the determination performance through re-learning according to the method of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention to those skilled in the art Other drawings can be obtained based on these figures without an inventive task being performed.

FIG. 1 is a conceptual diagram schematically showing a method of cervical cytology examination and cervical dilatation examination which were conventionally performed to diagnose cervical cancer.

FIG. 2 is a diagram showing a main concept for explaining a CNN (convolutional neural network) which is one of the machine learning models used in the present invention.

3 is a conceptual diagram schematically illustrating an exemplary configuration of a computing device that performs a method for determining whether a subject develops cervical cancer according to the present invention.

FIG. 4 is a flowchart illustrating an exemplary method for determining the incidence of cervical cancer according to the present invention.

5A to 5E are diagrams illustrating exemplary user interfaces (UIs) provided at respective steps of the method for determining the incidence of cervical cancer according to the present invention.

The following detailed description of the invention refers to the accompanying drawings, which illustrate, by way of example, specific embodiments in which the invention may be practiced in order to clarify the objects, technical solutions and advantages of the invention. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention.

Throughout the description and claims of the present invention, 'learning' is a term referring to performing machine learning in accordance with a procedure, and is not intended to refer to a mental function such as a human educational activity. Can be understood.

And throughout the description and claims of this invention, the word 'comprise' and variations thereof are not intended to exclude other technical features, additions, elements or steps. Other objects, advantages and features of the present invention will become apparent to those skilled in the art from this description, and in part from the practice of the invention. The following examples and figures are provided by way of illustration and are not intended to limit the invention.

Moreover, the present invention encompasses all possible combinations of embodiments shown herein. It should be understood that the various embodiments of the present invention are different, but need not be mutually exclusive. For example, certain features, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in connection with one embodiment. It should also be understood that the position or arrangement of individual components within each disclosed embodiment may be varied without departing from the spirit and scope of the present invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is to be limited only by the appended claims, along with the full scope of equivalents to which such claims are entitled, if properly explained. In the drawings, like reference numerals refer to the same or similar functions throughout the several views.

Unless otherwise indicated herein or clearly contradicted by context, items referred to in the singular are intended to encompass a plurality unless otherwise specified in the context. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, so that those skilled in the art can easily carry out the present invention.

Referring to FIG. 2, a CNN (convolutional neural network) model can be briefly described as an artificial neural network stacked in multiple layers. That is, this is referred to as a deep neural network in the sense of a network of deep structure. As shown in FIG. 2, by learning a large amount of data in a structure of a multi-layer network, It is a form that learns the network by learning the feature automatically and minimizing the error of the objective function through it. It is also expressed as a connection between the nerve cells of the human brain, and is thus a representative of artificial intelligence. In particular, as shown in FIG. 2, CNN is a model suitable for classification of a two-dimensional image such as an image. The CNN is a composite layer for creating feature maps using a plurality of filters (eg, points, lines, and surfaces) by repeating a pooling layer (a sub-sampling layer) that reduces the size of the feature map and extracts features that are invariant to changes in position or rotation It is possible to extract various levels of features from complicated and meaningful high-level features. Finally, if the feature extracted through the fully-connected layer is used as the input value of the existing model, Can be constructed.

Next, Fig. 3 schematically shows an exemplary configuration of a computing device that performs a method of determining whether a subject has a cervical cancer incidence according to the present invention (hereinafter referred to as " a method of determining whether or not a cervical cancer has occurred) It is a conceptual diagram.

3, a computing device 300 according to an exemplary embodiment of the present invention includes a communication unit 310 and a processor 320. The communication unit 310 communicates with an external computing device (not shown) Communication is possible.

In particular, the computing device 300 may include a variety of devices, such as routers, switches, and the like, which may include conventional computer hardware (e.g., a computer processor, memory, storage, input and output devices, Electronic communication devices, electronic information storage systems such as Network Attached Storage (NAS) and Storage Area Networks (SAN)) and computer software (i. E., Instructions that cause a computing device to function in a particular manner) System performance.

The communication unit 310 of such a computing device can send and receive requests and responses to and from other interworking computing devices. As an example, such requests and responses can be made by the same TCP session, For example, as a UDP datagram. In addition, in a broad sense, the communication unit 310 may include a keyboard, a mouse, and other external input devices for receiving commands or instructions.

In addition, the processor 320 of the computing device may include a hardware configuration such as a micro processing unit (MPU) or a central processing unit (CPU), a cache memory, a data bus, and the like. It may further include a software configuration of an operating system and an application that performs a specific purpose.

Now, a method for determining the incidence of cervical cancer according to the present invention will be described in detail with reference to FIG. FIG. 4 is a flowchart illustrating an exemplary method for determining the incidence of cervical cancer according to the present invention.

Referring to FIG. 4, the method for determining the incidence of cervical cancer according to the present invention is characterized in that the communication unit 310 of the computing device 300 acquires an image of the cervix of the subject, (Step S410).

In step S410, the photographed image may be acquired by a predetermined photographing module linked to the computing device 300. [

In addition, the captured image may be captured and obtained by another apparatus located far away from the place where the cervical cancer incidence determination method according to the present invention is performed by the

computing apparatus

300, 300 can acquire it.

Referring to FIG. 5A, when the photographed image 514 obtained by the other device is illustrated as an example in step S410, for example, when the 'Request' button 512 is detected to be pressed, The captured image 514 may be transmitted to the computing device 300 so that the computing device 300 acquires the captured image 514. [ 5A, the subject corresponding to the photographed image 514, that is, the subject information as the patient information, is displayed on the subject information input area 510 And information on the input time point, which is the acquired time point, can also be obtained together. The information on the subject and the input time point can be obtained from the captured image 514 and the captured image 514, May be communicated to the computing device 300 together.

4, the method for determining the incidence of cervical cancer according to the present invention is characterized in that an analysis module (not shown) implemented by the processor 320 of the computing device 300 detects the cervical cancer (S420) generating analysis information on the onset of the cervical cancer of the subject based on a machine learning model of the cervical cancer from the input of the image or supporting the other apparatus to generate the analysis information.

For this purpose, the machine learning model includes a plurality of previously entered training information, that is, (i) data of a plurality of cervical tomography images, (ii) whether or not there are lesions of cervical cancer in the plurality of cervical- Data, and (iii) if there is a lesion, the processor 320 learns the machine learning model using information including data indicating in which part of the image the lesion is present. As described above, the machine learning model may be a CNN (convolutional neural network) model, or a combination of CNN and a support vector machine (SVM).

For example, in the CNN model, learning can be performed by applying a gradient descent and a backpropagation algorithm based on an image of input training information.

To increase the accuracy and reliability of the machine learning model, a large amount of training data is required. The more training data, the better the accuracy and reliability of the machine learning model.

In step S420, the analysis information may include classification information on negative, atypical, positive, and malignant characteristics of the cervical cancer. As is well known to those of ordinary skill in the art, such classification information may include probability information indicating how accurate the classification is. More specifically, the analytical information may include negative judgment information, such as information on whether the subject is negative or whether the risk is positive or low (low cancer risk vs. high cancer risk). Acetowhite Epithelium, Mosaic , Morphological information such as Erosion or ulceration, Irregular surface contour, Punctation, Atypical Vessels, and Discolaration.

In step S420, the analysis information may be cataloged to correspond to a plurality of photographed images 514 and may be provided to the subject information 520 and the input time point information 522 as exemplarily shown in FIG. 5B. (Denoted as 'suspicious') 524 may be provided depending on the classification information and probability information calculated by the machine learning model and whether or not the onset of the cervical cancer is suspected. 5B illustrates an example in which the 'Evaluation' buttons 526 corresponding to a specific shot image are displayed in the example (FIG. 5B) so that the user can continue to perform subsequent steps following step S420 with respect to the selected specific shot image, Respectively.

The method for determining the incidence of cervical cancer according to the present invention may include a preprocessing module implemented by the processor 320 of the computing device 300, (Not shown) may perform the pre-processing on the cervical image or support the other device to perform the preprocessing (S415).

Here, the preprocessing may include at least one of image quality enhancement through blurring, histogram smoothing, etc., blurring and noise processing to perform robustness to the illuminance and noise of the photographed image. It will be understood by those of ordinary skill in the art.

Referring again to FIG. 4, a method for determining the incidence of cervical cancer according to the present invention is then performed by a read support module (not shown) implemented by the processor 320 of the computing device 300, (S430) supporting the user of the computing device or another user at the remote location to read whether or not the cervical cancer corresponding to the analysis information is caused by providing information or providing the other device with the information .

Here, for security in communication with a remote place, when the computing device is interlocked with the other device, information transmission / reception between the computing device and the other device may be performed by encryption and decryption, for example, AES 128 bit encryption and / Decoding may be applied.

As part of step S430, all or part of the photographed image may be provided on the user interface for image reading of the other user of the user or the remote site, as exemplarily shown in Fig. 5C, (For example, a rectangle, an arrow, a text input, and the like) while judging whether or not there is an unusual area.

Here, the analysis information may be processed in a manner easy for the user to understand, such as reading, and provided through a predetermined display device (display). For example, the classification information on the lesion location or lesion included in the analysis information may be displayed in a predetermined format.

As an example of the step S430 shown in FIG. 4, by using the analysis information provided through a predetermined display device, the user of the computing device 300, for example, Evaluation information on the classification information may be generated, and evaluation information on the occurrence and classification information of the classification information may be determined by the reading of the remote location. Here, the evaluation information includes information on whether the provided analysis information is correct, that is, whether or not the onset included in the analysis information is correct, and if the classification information included in the analysis information is correct or not, May be included. In the latter case, in the next step S440, the computing device 300 may acquire the evaluation information through the communication unit 310. [

In the above embodiment of the step S430, the evaluation information may include information on the quality of the photographed image, for example, information on a technical defect of the photographed image. For example, such a technical defect is that it is difficult to accurately determine the photographed image due to excessive mucus or blood in the photographed image, or it is difficult to confirm the incidence of the cervical cancer due to the angle of the photographed image or the position of the photographed region It may also be an image problem with an acetic acid reaction, insufficient acetic acid reaction, or out of focus, overexposure, or underexpression, even though there is an acetic acid reaction.

In step S430, as exemplarily shown in Fig. 5D, all or part of the photographed image 540, history information 541 of another image previously photographed with respect to the same subject, A subject information exposure area 542 indicating the subject information as inputted through the subject information input area 510, a sound positive determination information input area 543 to which sound positive determination information can be inputted, A technical defect information input area 545 in which information on technical defects of the morphological feature information input area 544 and the shot image 540 to which information can be inputted can be inputted, An artificial intelligence analysis information output area 546 representing information and a user opinion input area 547 by which the user (reader) can input the findings based on the photographed image can be provided on the user interface , Thereby making it easy for the user of the computing apparatus or the remote user to read whether or not the onset of the cervical cancer corresponding to the analysis information is read.

Next, a method for determining the incidence of cervical cancer according to the present invention includes: (i) acquiring and storing evaluation information on the analysis information as a result of performing step (S430); or And (ii) performing a process of outputting the evaluation information or supporting the output of the other device (S440).

Here, the evaluation information may be processed and provided in the form of a medical result report, for example, a user interface provided for this purpose is shown in FIG. 5E, and the medical result report 550 may include information on the onset of cervical cancer, And the like.

It should be noted that such a medical result report can be provided through another device located far away from the place where the photographed image was first obtained (i.e., the photographing place) or the step of reading the photographed image by the step S430 The person skilled in the art will understand.

As described above, the method for determining the incidence of cervical cancer according to the present invention can determine whether the cervical cancer has an onset based on a previously learned machine learning model, As an example of the method for determining the incidence of cervical cancer according to the present invention for taking advantage of this advantage, there is an advantage that the machine learning model can perform more accurate reading, A re-learning module (not shown) implemented by the processor 320 of the computing device 300 re-learns the machine learning model (S450) based on the evaluation information .

As described above, according to the present invention, in comparison with the conventional method in which the medical staff directly examines the cervical image obtained through the cervical loupe and confirms the state of the cervix individually based on education and experience, There is an effect of quickly and accurately determining the incidence of cervical cancer.

An advantage of the techniques described hereinabove with respect to the above embodiments is that it is possible to prevent mistakes, that is, misdiagnosis, of the medical personnel who must accurately determine the cervical cancer despite a large number of diagnoses. In particular, the use of machine learning technology allows the physician to analyze and learn the characteristics and forms of lesions of the cervical cancer, which are known only through education and experience of many years, by the computing device itself, It is possible to assist the judgment in cases where it is difficult to determine the incidence of cervical cancer.

Based on the description of the above embodiments, one of ordinary skill in the art can clearly understand that the present invention can be achieved through a combination of software and hardware, or can be accomplished by hardware alone. Objects of the technical solution of the present invention or portions contributing to the prior art can be recorded in a machine-readable recording medium implemented in the form of program instructions that can be executed through various computer components. The machine-readable recording medium may include program commands, data files, data structures, and the like, alone or in combination. The program instructions recorded on the machine-readable recording medium may be those specially designed and constructed for the present invention or may be those known to those of ordinary skill in the computer software arts. Examples of the machine-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like.

The hardware device may be configured to operate as one or more software modules for performing the processing according to the present invention, and vice versa. The hardware device may include a processor, such as a CPU or a GPU, coupled to a memory, such as ROM / RAM, for storing program instructions, and configured to execute instructions stored in the memory, And a communication unit. In addition, the hardware device may include a keyboard, a mouse, and other external input devices for receiving commands generated by the developers.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, And various modifications and changes may be made thereto by those skilled in the art to which the present invention pertains.

Therefore, the spirit of the present invention should not be construed as being limited to the above-described embodiments, and all of the equivalents or equivalents of the claims, as well as the following claims, I will say.

Equally or equivalently modified such methods will include logically equivalent methods which can yield, for example, the same results as those of the method according to the invention.

Claims

A method for determining the onset of a subject for cervical cancer,

(a) supporting a computing device to acquire a captured image of the cervix of the subject or acquire another device associated with the computing device;

(b) the computing device generates analysis information on the incidence of the cervical cancer of the subject based on the obtained machine learning model of the cervical cancer from the input of the cervical radiographic image, To generate the generated data;

(c) the computing device provides the generated analysis information or provides the other device with information, thereby allowing a user of the computing device or a remote user of the remote location to determine whether the cervical cancer corresponding to the analysis information has occurred Supporting to read; And

(d) the computing device is configured to: (i) acquire and store evaluation information for the analysis information as a result of performing the step (c), or support the other device to acquire and store the evaluation information; And (ii) performing a process of outputting the evaluation information or supporting the output of the other device

Of the cervical cancer.
The method according to claim 1,

(e) based on the evaluation information, the computing device re-learning the machine learning model or supporting the other device to re-learn the machine learning model

Of the cervix cancer.
The method according to claim 1,

In the step (b)

Wherein the analysis information includes classification information on the negative, atypical, positive, and malignant characteristics of the cervical cancer.
The method according to claim 1,

Prior to step (b)

(b0) further comprising the step of the computing device performing a pre-processing on the cervical image or supporting the other device to perform the preprocessing.
The method according to claim 1,

The machine learning model includes a composite neural network (CNN) model and a support vector machine (SVM) model

The step (b)

(b1) supporting the computing device to extract features or extract features from the cervical radiographic image based on the synthetic neural network model; And

(b2) the computing device generates classification information by classifying the cervical-radiographed image using the feature based on the support vector machine model or supports the other device to generate the classification information

And determining whether the cervix cancer has been diagnosed.
The method according to claim 1,

Wherein the information transmission / reception between the computing device and the other device is performed by AES 128 bit encryption and decryption when the computing device is interlocked with the other device.
A computer program comprising instructions implemented by a computing device to perform the method of any one of claims 1 to 6.
1. A computing device for determining whether a subject has a cervical cancer,

A communication unit for acquiring an image of the cervix of the subject; And

From the input of the obtained cervical tomography image, analysis information on the onset of the cervical cancer of the subject on the basis of the machine learning model of the cervical cancer, or to generate other information interworking through the communication unit Supported Processors

, &Lt; / RTI &

The processor comprising:

And provides the generated analysis information or provides the other device to provide the analysis information so that the user of the computing device or the remote user at the remote location can read whether or not the cervical cancer corresponding to the analysis information is detected,

(i) acquiring and storing evaluation information on the analysis information as a result of the reading, or (ii) supporting the other device to acquire and store the evaluation information, and (ii) And outputting the resultant cervical cancer.
9. The method of claim 8,

The processor comprising:

And re-learning the machine learning model based on the evaluation information, or causing the other apparatus to re-learn the machine learning model.
9. The method of claim 8,

The machine learning model includes a composite neural network (CNN) model and a support vector machine (SVM) model

The processor comprising:

Upon generating the analysis information,

Extracting features from the cervical radiographic image or extracting the other apparatus based on the synthetic neural network model,

Wherein the support vector machine model is used to generate categorization information by classifying the cervical radiographic image using the feature, or to assist the other device to generate categorical information.