WO2016085236A1 - Method and system for automatic determination of thyroid cancer - Google Patents

Abstract

The present invention relates to a method and system for determining, as a degree of risk, the presence or absence of thyroid cancer with high accuracy by using ultrasound images and result values of previous biopsies. Values for a plurality of factors are automatically or manually input from an ultrasound image, and result values of previous biopsies are utilized to provide a result that is highly accurate and universal.

Description

Automated thyroid cancer determination method and system

The present invention relates to a method and system for discriminating thyroid cancer as a risk with high accuracy using ultrasound images and previous biopsy results.

The thyroid gland is one of the endocrine organs that secretes thyroid hormones. Diseases in which the thyroid abnormalities occur are largely two types, thyroid dysfunction that can be cured by drug treatment, etc., and nodules or nodules of the thyroid.

In the case of thyroid nodules, about 95% are benign nodules, which are not normal cells but adenomas, clusters or blisters, which are not cancerous. About 5% is thyroid cancer, the slowest growing cancer among humans, but malignant in need of treatment.

Ultrasound diagnosis is most widely used for thyroid cancer diagnosis, and biopsy is used to improve accuracy.

Referring to Figure 1, a conventional general thyroid cancer determination method will be described.

If the attending physician treats the patient and suspects thyroid cancer, the doctor recommends taking an ultrasound image, and the ultrasound implementer takes an ultrasound of the patient using the ultrasound machine 100. The captured ultrasound image is acquired by the image extracting means 110, and the acquired ultrasound image is directly transmitted to the reader or when the PACS, i.e., a picture archiving and communication system, is constructed, the hospital database 200 Is uploaded to the PACS database 210 and delivered to the reader.

The readout person visually checks the transmitted image through the output unit 400 to determine whether thyroid cancer is present. Based on the experience and knowledge of the reader, nodules presumably cancers are found, and the shape, size, boundary, internal components, degree of calcification, and echo are determined visually. Based on the result of the decision, the clinician again estimates thyroid cancer based on experience, knowledge, and manual, and decides whether to perform biopsy or follow-up based on the malignant risk. The read result determined as described above is directly transmitted to the attending physician or uploaded to the PACS database 210 of the hospital database 200. When an EMR system, that is, an electronic medical record system, is uploaded and delivered to the EMR database 220.

On the other hand, when the patient has a biopsy results are delivered directly to the attending physician, or computerized through the biopsy record input unit 225 is uploaded to the EMR database 220 is delivered. The attending physician will examine the treatment method based on the results of the above tests, the results of the readings and the biopsy. Depending on the results of the biopsy, a retest may be performed.

This process has the following problems.

First, the image acquisition step and the readout step are dualized, which takes considerable time. Also, the interpretation of the image is very different depending on the subjective opinion of the reader. For example, looking at the deformed elongated oval nodule, some may be considered ovoids and some are irregular. In addition, since the combination of shapes, sizes, boundaries, internal components, degree of calcification, and echoes, which are the criteria for determination, varies widely, it is difficult to calculate the human reading alone by combining all of them.

Most importantly, thyroid cancer is most often a cancer that does not progress quickly. In other words, it is preferable to avoid surgery and avoid biopsies if it is not a clear diagnosis of cancer. However, according to the above-described methods, the reading result and treatment method are different for each of the physician or physician in charge so that unnecessary surgery or biopsy is performed. The disadvantage is that it can.

Review the patent literature related to this.

Korean Laid-Open Patent Publication No. 10-2014-0094760 discloses an apparatus for predicting whether malignant tumors exist in a subject using a medical image. Although it calculates the probability of being a malignant tumor, it is not accurate because the patient's medical records are not referenced, and it is difficult to apply when a new cancer discrimination method or criterion is presented because it is calculated based only on the degree of microcalcification. Because it is a method for discriminating all types of cancer, the specificity of thyroid cancer is not taken into account.

US patent application 2011-0295782 discloses a method for identifying the malignant risk of thyroid nodules. In consideration of the specificity of thyroid cancer, lymph node size is extracted and used from an ultrasound image. The accuracy is low because only the node size is used and the patient's medical records are not referenced.

US Patent No. 8,366,619 discloses an apparatus for determining whether malignant nodules by checking the stiffness index of the nodules using ultrasonic elastomeric. The prior art uses physical examination results that do not use images.

Taken together, the current method of determining thyroid cancer takes a lot of time and can lead to different results depending on the readout, so the accuracy or level of confidence is not high, and cancer surgery or biopsy is often recommended. Some patent documents provide universal criteria to prevent different readings, but are also problematic in accuracy and do not reflect the specificity of thyroid cancer.

<Private Technology>

Korean Patent Publication No. 10-2014-0094760

United States Patent Application No. 2011-0295782

US Patent No. 8,366,619

The present invention has been made to solve the above problems.

In other words, we propose a system and method for determining thyroid cancer on a fast, accurate and universal basis. In addition, it is proposed a system and method that can further increase the accuracy by reflecting the specificity of thyroid cancer, not other cancers.

In order to solve the above problems, an embodiment of the present invention, (a) the image extraction means 110 to extract the ultrasound image; (b) the image analyzing means 120 detecting a nodule in the extracted ultrasound image and calculating values of a plurality of factors of the nodule; (c) the risk calculating means (310) retrieving medical records from a database, and calculating the risk by a predetermined method using the values of the plurality of factors and the medical records of step (b); (d) determining, by the risk calculating means 310, whether the detected nodule is a nodule of a predetermined shape according to a guideline criterion when the calculated risk is greater than or equal to a predetermined criterion; (e) in the case of a nodule of a predetermined shape according to the guidelines, the diagnostic information calculating means 320 determining whether the diameter of the nodule is greater than 1 cm; (f) the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400 when the diameter of the nodule is greater than 1 cm, and outputs 400 when the diameter of the nodule is 1 cm or less. ) And outputting a "horizontal observation", wherein the medical records of step (c) include a previous biopsy result, which provides an automatic thyroid cancer determination method.

Further, after the step (d), (d1) when the calculated risk is less than the predetermined criterion, the diagnostic information calculating means 320, if the diameter of the nodule is greater than 2cm through the output unit 400 It is preferable to further include the step of outputting the "history examination", and outputting "overlook observation" through the output unit 400 when the diameter of the nodule is 2 cm or less.

Also, in the step (d), the nodule of the predetermined shape is preferably an indeterminate nodule.

In addition, after the step (d), (e1) if it is not the first nodule according to the guidelines, the diagnostic information calculating means 320 determines whether the diameter of the nodule is greater than 0.5cm; (f1) the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400 when the diameter of the nodule is greater than 0.5 cm, and the output unit when the diameter of the nodule is 0.5 cm or less. It is preferable to further include the step of outputting "observation" through (400).

In addition, after the step (f), (g) updating the medical record by inputting a biopsy result in the database; (h) extracting the second ultrasound image by the image extracting means (110); (i) the image analyzing means (120) detecting a nodule in the extracted second ultrasound image and recalculating a value of a plurality of factors of the nodule; (j) the risk calculating means 310 uses the values of the plurality of factors recalculated in the step (i) and the medical record updated in the step (g), by the predetermined method of the step (c). Calculating a risk; (k) If the calculated risk is greater than or equal to the predetermined criterion, the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400, otherwise, through the output unit 400. Preferably, the method further includes outputting an "observation".

In addition, the previous biopsy result (Previous Biopsy Result) of step (c) and the biopsy result input in step (g) are divided into the procedure method and the biopsy result, the biopsy result is a six-step diagnosis It is preferable that the result is recorded as a value according to the classification system (Bethesda System for Reporting Thyroid Cytopathology).

In addition, the plurality of factors include diameter, internal content, sponge appearance, shape, margin, echogenicity, and calcification. It is preferable.

In addition, the image analysis means 120, using the image extracted in the step (a), calculates the diameter of the nodule, the internal components of the nodule (solid), almost solid (predominantly solid) ), Almost predominantly cystic, cystic (cystic), calculates the presence of the nodule on the spongy, and the shape of the nodule (ovoid / round), irregular (irregular), rectangular (taller than) compute the boundary of the nodule as either smooth, ill-defined, or spiculated, and echo the nodules. The diagram is computed with any one of marked hypoechogenicity, hypoechogenicity, isoechogenicity, and hyperechogenicity, and the calcification of the nodule is microcalcified, rim calcification. It is preferable to operate with either macrocalcification.

In addition, the step (b), the input unit 315 is preferably a step of inputting the values of the plurality of factors of the nodule detected in the extracted ultrasound image.

In addition, the step (b), the image analysis means 120 includes a step of detecting a plurality of nodules in the extracted ultrasound image and setting the priority of the plurality of nodules based on their size. Do.

Another embodiment of the present invention to solve the above problems, the image extraction means 110 for extracting the ultrasound image; Image analysis means (120) for detecting a nodule in the ultrasound image extracted by the image extraction means and calculating values of a plurality of factors of the nodule; Retrieve a medical record from a database, calculate the risk by a predetermined method using the values of the plurality of factors and the medical record, and if the calculated risk is above a predetermined criterion, the detected nodule is based on guidelines. A risk calculating means 310 for determining whether a nodule of a predetermined shape according to; Diagnostic information calculating means (320) for determining whether the diameter of the nodule is greater than 1 cm in the case of a nodule of a predetermined shape according to the guidelines; And an output unit 400 outputting "history examination" when the diameter of the nodule is greater than 1 cm, and outputting "transspective observation" when the diameter of the nodule is 1 cm or less. The medical record provides an automatic thyroid cancer discrimination system including a previous biopsy result.

In addition, the database is preferably an electronic medical record (EMR) database 220 of the hospital database 200.

In addition, the hospital database 200, PACS (Picture Archiving and Communication System) is uploaded to the ultrasound image extracted from the image extraction means 100 and the value of a plurality of factors calculated by the image analysis means 120 It is preferable to further include a database 210.

In addition, it is preferable that a biopsy result may be further input to the EMR database 220.

In addition, the previous biopsy results (Previous Biopsy Result) and the biopsy results are divided into the procedure and the biopsy results, the biopsy results are values according to the six-stage diagnostic system (Bethesda System for Reporting Thyroid Cytopathology) The thyroid cancer automatic detection system may further include a biopsy record input unit 225 capable of inputting the previous biopsy result and the biopsy result.

In addition, it is preferable to further include a guideline database 500 in which the guideline criteria are stored.

In addition, the guideline database 500 preferably extracts information about the guideline from a predetermined website and stores the guideline information based on the guideline.

In addition, the plurality of factors include diameter, internal components, sponge phase, shape, boundary, echo, and calcification, the automatic thyroid cancer determination system, input unit 315 that can input the value of the plurality of factors It is preferable to further include.

With the present invention, it is possible to present the same or similar diagnostic results quickly and accurately and for every hospital or reading / doctor.

In addition, the specificity of thyroid cancer is reflected in 10 minutes, further improving the accuracy.

In addition, since the patient's conventional medical records are scored and reflected, the individual characteristics of the patient are also taken into consideration to further improve accuracy.

In addition, even if the guidelines change in the society, etc. can be reflected in the present system in real time, even if the standard changes, the most recent treatment tendency and diagnosis tendency can be reflected as it is.

In addition, it is possible to link with the currently used PACS and EMR, easy to manage the database, it is possible to build big data, regression analysis is performed based on the built data, and the analyzed results can be reflected as a standard again As data accumulates, the accuracy can be further improved.

Since patients do not have a very high chance of cancer, they do not have to carry out unnecessary biopsies or surgery to remove cancer, thereby reducing medical expenses and minimizing their discomfort. May affect

1 is a view for explaining a thyroid cancer determination method according to the prior art.

2 is a conceptual diagram illustrating a thyroid cancer determination system according to the present invention.

3A and 3B are flowcharts illustrating a thyroid cancer determination method according to the present invention.

4 and 5 illustrate screens output to an output unit when the thyroid cancer determination system and method according to the present invention are implemented.

Here "system" is to be understood as meaning an object that is the opposite of the method.

Here, "the size of the nodule" means the distance on the largest straight line of the nodule. For example, if the crystal is elliptical, the long diameter, not the short diameter, is the size. In the following description, the term "diameter nodule" is used, but it should be understood as having the same meaning as "size of the nodule".

Here, the "operation unit" and "operation means" means information processing means for inputting information and performing arithmetic operation by a predetermined method or algorithm to derive a result. An example may be a computer equipped with computing devices such as a CPU.

Description of Automatic Thyroid Cancer Screening System

Hereinafter, the present invention will be described in detail with reference to FIGS. 2, 4 and 5. FIG.

The automatic thyroid cancer determination system according to the present invention includes an ultrasound machine 100, a hospital database 200, an operation unit 300, an output unit 400, and a guideline database 500.

The ultrasound machine 100 may be any ultrasound machine currently available, and may use any machine provided with an image extracting means 110 capable of extracting a photographed image and transferring it to a database or a calculation unit. However, it is characterized in that the image analysis means 120 is provided.

The image analyzing means 120 of the ultrasound machine 100 of the automatic thyroid cancer determination system according to the present invention performs the following functions.

First, all the plurality of nodules may be selected and selected from the extracted ultrasound image, and each priority may be determined according to a predetermined criterion. The method of automatically selecting a nodule in the ultrasound image is a prior art and thus a detailed description thereof will be omitted.

Second, extract the desired information from each of the selected nodules. In other words, the values of the predetermined factors should be extracted automatically. A number of factors to be extracted are as follows.

Diameter: As described above, refers to the size of the nodule, for example, the value is extracted in cm.

Internal Content: Four levels of values are extracted: solid, nearly solid, almost cystic, and cystic. The image analyzing means 120 may calculate this based on the captured ultrasound image, and the criteria for distinguishing the four stages of classification, ie, "solidity" / "almost solidity" / "almost cystic" / "cystic" in advance It may be input, for example, 10%, 50%, 90% based on the ratio of the black points per pixel of the image may be distinguished. Since the image analysis method is a conventional technology, a detailed description thereof will be omitted.

■ Spongiform appearance: It is extracted with and without spongiform appearance. The sea surface refers to a case in which a plurality of microcystic components are divided by a thin diaphragm, and the image analyzing means 120 may classify them based on a predetermined number of lines per unit pixel.

■ Shape: It is extracted into three types: oval / round, irregular, and rectangular (wideler than wide). The image analyzing means 120 may determine the contour by using a contour determination technique, which is one of the conventional general techniques, and when it is included within a predetermined range of predetermined curvature, the image analysis means 120 may have a predetermined aspect ratio. If it is within the range, it is automatically classified into a rectangle if it is not rectangular, oval or rectangular.

Margin: It is extracted by dividing into smooth, ill-defined, and spiculated. The image analyzing means 120 automatically divides the shape of the boundary into the above three by using the contour determination technique.

■ Echoogenicity: Extracted into distinctly marked hypoechogenicity, hypoechogenicity, isoechogenicity, and hyperechogenicity. "Earth low echo" is less echo of the nodule than the echo of the surrounding muscles, "low echo" is less echo of the nodule than the echo of the surrounding thyroid tissue, "back echo" is the echo of the surrounding thyroid tissue and nodules If it is the same, "koeko" means that the echo of the nodule is higher than the echo of the surrounding thyroid tissue. As mentioned above, the image analyzing means 120 can distinguish between the nodular portion and the non-nodular portion, and calculates the shade of each portion to compare the four pixels with the surrounding pixel region (ie, the surrounding thyroid tissue or the surrounding muscle). It is possible to automatically extract branched echoes.

■ Calcification: Extracted into three categories: microcalcification, rim calcification, and macrocalcification. When calcification proceeds, a mark such as a white dot appears in the ultrasound image. The image analyzing means 120 may automatically calculate the degree of calcification by calculating the number of white pixels per unit area and the density frequency of the white pixels in the unit area. .

Meanwhile, in one embodiment, the values of the plurality of factors may be automatically extracted by the image analyzing means 120 of the ultrasound machine 100, but in another embodiment, the risk calculating means may be provided through a separate input unit 315. It may also be entered directly into 310. One example of a screen input by the input unit 315 is shown in FIG. 4. In the first line, the previous biopsy result is disclosed, which is a part for inputting the already obtained biopsy result, which will be described later.

The hospital database 200 includes a PACS database 210 and an EMR database 220. You can also use an existing database.

The values of a plurality of factors automatically identified by the image analyzing means 120 are uploaded to the PACS database 210 of the hospital database 200 together with the ultrasound images. At this time, an identifier for identifying the patient is uploaded together.

The values and identifiers of the multiple factors uploaded to the PACS database 210 are linked with the EMR database 220 to retrieve the existing medical records of the patient, which are stored in advance, using the identifiers. Stored. Here, the existing medical record includes a previous biopsy result for patients with a history of biopsy.

The calculation unit 300 calculates the actual thyroid cancer risk and calculates what information of diagnostic information, for example, "observation" or "tissue examination" for a suitable medical treatment method is to be output. The calculating unit 300 includes a risk calculating means 310 and a diagnostic information calculating means 320.

The risk calculator 310 retrieves a number of factors from the EMR database 220 in the ultrasound image and a previous biopsy result, which is an existing medical record of the patient.

Previous Biopsy Results include what was used as the procedure and the actual results (see Figure 5).

The procedure is classified into fine needle aspiration biopsy (FNAB) or core needle biopsy (CNB) and stored in the EMR database 220.

The biopsy results are stored in an international six-step Bethesda System for Reporting Thyroid Cytopathology.

"1" means "non-diagnostic result", and the cell number is not enough to accurately diagnose. Typically, the risk of thyroid cancer is 1 to 4%, but retesting is needed. "2" is "benign nodule" and the risk of thyroid cancer is 0 to 3%. "3" is an undetermined nodule divided into "AUS" and "FLUS". Some cells are atypical, suspected of having thyroid cancer, or difficult to diagnose with clear benign nodules. Thyroid cancer risk is 5 to 15%. "4" is "follicular tumor (or suspected follicular tumor)". Thyroid cancer risk is 15 to 30%. "5" is "suspect thyroid cancer". Thyroid cancer risk is 60 to 75% and thyroid surgery is recommended. "6" is diagnosed as "thyroid cancer" and the thyroid cancer risk is 97 to 99%.

The biopsy results are uploaded to the EMR database 220 of the hospital database 200 through a separate biopsy record input unit 225 provided in the EMR database 220.

Of course, patients who have not previously been biopsied have no previous biopsy results.

The risk calculating means 310 automatically calculates the risk of thyroid cancer by combining the values of the plurality of factors thus collected and previous biopsy results. A specific calculation method is described in detail below.

The diagnostic information calculating means 320 calculates any one of "experimental observation" or "tissue examination" as diagnostic information based on the calculation result of the risk calculating means 310 and outputs it through the output unit 400. On the other hand, the diagnostic information calculation means 320 further utilizes one criterion, which is a general guideline criterion. Guidelines criteria may be extracted from the guidelines database 500.

The guideline database 500 is a website provided with a criterion for determining whether thyroid cancer is among Internet websites. The guideline database 500 extracts information about a guideline from a predetermined website and stores the guideline as a guideline.

The predetermined website may be, for example, the Korean Society of Thyroid Imaging Medicine (http://thyroidimaging.kr/), and the system according to the present invention sets the index of "reference" among the data included in the website. The "KSThR guideline" can be extracted automatically. The extracted thyroid nodule (KSThR guideline) is, for example, the basis for the classification of benign nodules, indeterminate nodules, suspicious malignants, and normal nodules with nominal diameters of 2 cm, 1 cm, and 5 mm. In this case, since the diameter of the nodule is one of the values analyzed by the above-described image analyzing means 120, it is automatically determined where it corresponds.

The above-mentioned thyroid nodule (KSThR guideline) of the Korean Society of Thyroid Imaging Medicine is just one example, and as described above, it can be applied to any website that can extract the thyroid nodule by setting the "reference" as an index. to be.

Description of automatic thyroid cancer detection method

Hereinafter, a method for automatically determining thyroid cancer will be described with reference to FIGS. 3 and 5.

Similarly to the prior art, the attending physician treats the patient and recommends taking an ultrasound image when the thyroid cancer is suspected, and the patient takes an ultrasound image in the ultrasound machine 100. The captured ultrasound image is extracted by the image extracting means 110 (S100).

Here, the ultrasound image may be an image in which a doctor directly takes an ultrasound image of a patient using a probe, or may be an image automatically scanned using a gel-pad probe.

Next, the image analysis means 120 of the present invention detects the nodules in the extracted ultrasound image, determines the priority, and calculates values of a plurality of factors for each nodule (S200). Many of the factors computed here are seven: Diameter, Internal Content, Spongiform appearance, Shape, Margin, Echogenicity, and Calcification. Dog.

Meanwhile, in another embodiment, the value of the plurality of factors may be manually input by using a separate input unit 315.

Next, the risk calculating unit 310 calculates the risk in a predetermined manner by combining the values of the plurality of factors (S300).

For example, after assigning 0 to 10 points to each of the seven factors (10 points when the risk of thyroid cancer is high), a method of adding them up may be used. As another example, the diameter factor may be excluded as it is used as another criterion to be described later. As another example, each factor may be weighted. For another example, if any one factor approaches 10, it may be calculated in a manner that increases the risk. As such, if seven factors are automatically or manually extracted, they may be combined in any way, and the scope of the present invention is to determine whether thyroid cancer is used using a plurality of factors, and calculates the value of the factor. Note that it is not a specific formula to.

On the other hand, when there is an existing biopsy record, the medical record stored in the EMR database 220, that is, the previous biopsy result (Previous Biopsy Result) is further used. Previous Biopsy Results are stored in a six-step diagnostic system (Bethesda System for Reporting Thyroid Cytopathology), as described above, to calculate each value, which can be added to the seven factors described above. have.

When the risk is calculated by the system according to the invention in this way, another diagnosis is made based on whether the risk is below a preset range. As shown in the figure, if the risk is less than 5% (S400) if the diameter is 2cm or less (S500), a biopsy is not necessary and can be diagnosed as "overview". That is, "pass observation" is output through the output unit 400 (S600). If the diameter is more than 2cm, because a biopsy is necessary, the "history examination" is output through the output unit 400 (S700).

Such output is output along with the ultrasound image through the monitor provided in the ultrasound machine 100, through the monitor associated with the PACS or EMR in the hospital, or of course there is no limitation.

On the other hand, if the risk is calculated to be 5% or more by the system according to the present invention, it is preferable to apply the guidelines once again from the guidelines database 500 rather than unconditional biopsy. For example, if the guideline is the above-mentioned thyroid nodule (KSThR guideline) of the Korean Society of Thyroid Imaging Medicine (S410), if the nodule (indeterminate) (S410) If the diameter is less than 1cm (S510), a biopsy is not necessary. It can be diagnosed as follow-up observation. That is, "pass observation" is output through the output unit 400 (S600). Lowering from 2 cm to 1 cm is a higher risk and a stricter standard. If the diameter is more than 1 cm, since a biopsy is necessary, the "history examination" is output through the output unit 400 (S700).

Here, if the diameter is not 0.5cm or less if the indeterminate nodule (S520) does not require a biopsy and can be diagnosed as a follow-up observation. That is, "pass observation" is output through the output unit 400 (S600). The decrease from 1 cm to 0.5 cm reflects a higher risk because it is not an indeterminate nodule. If the diameter is more than 0.5cm, because a biopsy is necessary, the "history examination" is output through the output unit 400 (S700).

If the biopsy is performed through such a procedure, the result is stored in the EMR database 220 again (S810). The form to be stored is as described above.

In addition, the second ultrasonic inspection is performed through the ultrasonic machine 100, and the corresponding ultrasonic image is extracted by the image extracting means as the second ultrasonic image (S820), and each value of the plurality of factors is determined by the image analyzing means. It is recalculated (S830). The method is as described above.

Now, the risk calculating means 310 calculates the second risk once again by combining the values of the plurality of factors recomputed and the medical records updated by uploading the biopsy to the EMR database 220 in step S720 (S840). ). If the secondary risk is still 5% or more as a result of the calculation (S850), the output unit 400 outputs a "history examination" again (S860), and still outputs an "observation".

This method has the following advantages.

While the prior art relies heavily on the subjective point of view of the reading, the present invention can improve the universality and accuracy by presenting an objective criterion based on a total of seven or more factors.

In addition, in the prior art, if it was determined whether the observation and histological examination were based on only the primary risk determined according to the subjective viewpoint of the reading (that is, the procedure is terminated at the step S400 based on the subjective viewpoint of the judgment). The invention excludes the subjective point of view of judgment and utilizes previous biopsy results, guideline criteria, and more than two times of ultrasound image analysis results. Or cancer removal surgery is not recommended.

In addition, all histological records, as well as values of a number of factors extracted from all ultrasound images, are accumulated, and whether or not it is finally diagnosed as cancer. As a result, the accumulated data form big data, thereby increasing the accuracy of the discriminating method according to the present invention.

Moreover, even if the guidelines change in the future or the Six-stage System for Reporting Thyroid Cytopathology changes, the criteria are automatically applied so that the latest standards can be utilized.

Although described above with reference to a preferred embodiment of the present invention, those of ordinary skill in the art various modifications and variations of the present invention within the scope and spirit of the present invention described in the claims below It will be appreciated that it can be changed.

Claims (18)

1. (a) the image extracting means 110 extracting the ultrasound image;

   (b) the image analyzing means 120 detecting a nodule in the extracted ultrasound image and calculating values of a plurality of factors of the nodule;

   (c) the risk calculating means (310) retrieving medical records from a database, and calculating the risk by a predetermined method using the values of the plurality of factors and the medical records of step (b);

   (d) determining, by the risk calculating means 310, whether the detected nodule is a nodule of a predetermined shape according to a guideline criterion when the calculated risk is greater than or equal to a predetermined criterion;

   (e) in the case of a nodule of a predetermined shape according to the guidelines, the diagnostic information calculating means 320 determining whether the diameter of the nodule is greater than 1 cm;

   (f) the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400 when the diameter of the nodule is greater than 1 cm, and outputs 400 when the diameter of the nodule is 1 cm or less. Outputting "observation" through),

   The medical record of step (c) includes a previous biopsy result (Previous Biopsy Result),

   Automated determination of thyroid cancer.
2. The method of claim 1,

   After the step (d),

   (d1) when the calculated risk is less than the predetermined criterion, the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400 when the diameter of the nodule is greater than 2 cm, and the If the diameter of the nodules 2cm or less, further comprising the step of outputting "transspective observation" through the output unit 400,

   Automated determination of thyroid cancer.
3. The method of claim 1,

   In step (d), the nodule of the predetermined shape is an indeterminate nodule,

   Automated determination of thyroid cancer.
4. The method of claim 1,

   After the step (d),

   (e1) determining that the diameter of the nodule is greater than 0.5 cm by the diagnostic information calculating means (320) when it is not the first nodule according to the guidelines;

   (f1) the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400 when the diameter of the nodule is greater than 0.5 cm, and the output unit when the diameter of the nodule is 0.5 cm or less. And further outputting "observation" through 400;

   Automated determination of thyroid cancer.
5. The method of claim 1,

   After the step (f),

   (g) inputting a biopsy result into the database to update a medical record;

   (h) extracting the second ultrasound image by the image extracting means (110);

   (i) the image analyzing means (120) detecting a nodule in the extracted second ultrasound image and recalculating a value of a plurality of factors of the nodule;

   (j) the risk calculating means 310 uses the values of the plurality of factors recalculated in the step (i) and the medical record updated in the step (g), by the predetermined method of the step (c). Calculating a risk;

   (k) If the calculated risk is greater than or equal to the predetermined criterion, the diagnostic information calculating means 320 outputs a "tissue examination" through the output unit 400, otherwise, through the output unit 400. And further outputting "observation",

   Automated determination of thyroid cancer.
6. The method of claim 5, wherein

   The previous biopsy result of step (c) and the biopsy result input in step (g) are divided into the procedure and the biopsy result,

   The biopsy result is a result recorded as a value by a six-stage diagnostic classification method (Bethesda System for Reporting Thyroid Cytopathology),

   Automated determination of thyroid cancer.
7. The method according to any one of claims 1 to 6,

   The plurality of factors include diameter, internal content, sponge appearance, shape, margin, echogenicity, and calcification,

   Automated determination of thyroid cancer.
8. The method of claim 7, wherein

   The image analyzing means 120 calculates the diameter of the nodule by using the image extracted in the step (a), and the internal components of the nodule are solid, almost solid, Calculate almost any cystic or cystic, calculate the presence or absence of spongy presence of the nodule, and shape the nodule in the shape of oval / round, irregular, or taler than wide. Compute the boundary of the nodule as one of smooth, ill-defined, and spiculated, and calculate the echogenicity of the nodule. Calculate any of the marked hypoechogenicity, hypoechogenicity, isoechogenicity, hyperechogenicity, and calculate the calcification of the nodule in microcalcification, rim calcification, large Operating with any one of calcification,

   Automated determination of thyroid cancer.
9. The method of claim 7, wherein

   In the step (b), the input unit 315 is a step of inputting values of a plurality of factors of the nodule detected in the extracted ultrasound image.

   Automated determination of thyroid cancer.
10. The method of claim 1,

    In the step (b), the image analyzing means 120 includes detecting a plurality of nodules in the extracted ultrasound image and setting the priority of the plurality of nodules based on their size.

    Automated determination of thyroid cancer.
11. Image extracting means (110) for extracting an ultrasound image;

    Image analysis means (120) for detecting a nodule in the ultrasound image extracted by the image extraction means and calculating values of a plurality of factors of the nodule;

    Retrieve a medical record from a database, calculate the risk by a predetermined method using the values of the plurality of factors and the medical record, and if the calculated risk is above a predetermined criterion, the detected nodule is based on guidelines. A risk calculating means 310 for determining whether a nodule of a predetermined shape according to;

    Diagnostic information calculating means (320) for determining whether the diameter of the nodule is greater than 1 cm in the case of a nodule of a predetermined shape according to the guidelines; And

    As a result of the diagnosis information calculation means 320, the diameter of the nodule is more than 1cm "history examination" is output, and if the diameter of the nodules includes an output unit 400 for outputting "transspective observation" ,

    The medical record includes a previous biopsy result,

    Thyroid cancer automatic identification system.
12. The method of claim 11,

    The database is an electronic medical record (EMR) database 220 of the hospital database 200,

    Thyroid cancer automatic identification system.
13. The method of claim 12,

    The hospital database 200 may include a picture archiving and communication system (PACS) database in which ultrasound images extracted by the image extracting means 100 and values of a plurality of factors calculated by the image analyzing means 120 are uploaded. Further comprising 210),

    Thyroid cancer automatic identification system.
14. The method of claim 12,

    The EMR database 220 may be further input biopsy results,

    Thyroid cancer automatic identification system.
15. The method of claim 14,

    The previous biopsy result (Previous Biopsy Result) and the biopsy results are divided into the procedure and the biopsy results, the biopsy results are recorded as a value by the six-step diagnostic system (Bethesda System for Reporting Thyroid Cytopathology) Result,

    The automatic thyroid cancer determination system further includes a biopsy record input unit 225 capable of inputting the previous biopsy result and the biopsy result.

    Thyroid cancer automatic identification system.
16. The method of claim 11,

    Further comprising a guideline database 500 for storing the guideline criteria,

    Thyroid cancer automatic identification system.
17. The method of claim 16,

    The guideline database 500 is to extract information about the guideline according to a predetermined method from a predetermined website and to store the guideline reference,

    Thyroid cancer automatic identification system.
18. The method of claim 11,

    The plurality of factors include diameter, internal components, sea surface, shape, boundary, echo, and calcification,

    The automatic thyroid cancer determination system further includes an input unit 315 capable of inputting values of the plurality of factors.

    Thyroid cancer automatic identification system.

Images