Method for quantitative evaluation of benign and malignant thyroid nodules by elastography

Abstract

The present application relates to the technical field of thyroid ultrasound image quantitative diagnosis, and particularly discloses a method for evaluating the elasticity of thyroid benign and malignant nodules, which comprises the following steps: S1, collecting original data of elasticity ultrasound images corresponding to the thyroid nodules to be detected; S2, carrying out data cleaning on all the original data of elasticity ultrasound images to eliminate useless data and fuzzy imaging data. In the present application, a peripheral normal thyroid parenchyma reference area is set to simultaneously measure and calculate and correct parameters, thereby solving the problem of individual gland hardness interference caused by the lack of baseline correction in the prior art; and the prior scheme directly uses original elasticity quantitative parameters of nodules for grading, which causes confusion in the risk determination of the same values of soft glands in young people and fibrotic glands in the middle-aged and the elderly.

Description

Technical Field

[0001] This invention relates to the field of quantitative diagnostic technology for thyroid ultrasound imaging, and more particularly to a method for quantitative elastography assessment of benign and malignant thyroid nodules. Background Technology

[0002] Thyroid nodules are a common endocrine disorder. Clinically, differentiating between benign and malignant nodules is the core basis for formulating surgical, follow-up, and drug intervention plans. Elastography relies on the difference in tissue hardness for imaging, and compared with traditional two-dimensional ultrasound, it is more sensitive in identifying malignant nodules. It has now become a routine auxiliary examination method for thyroid nodules.

[0003] Current clinically accepted elastography assessment methods rely solely on physicians' visual interpretation of the color distribution of elastography images for qualitative grading, lacking a standardized quantitative correction process. Existing technologies directly extract elastography values ​​from thyroid nodule regions as diagnostic criteria. However, individual differences exist in the baseline hardness of normal thyroid glands among different patients. Adolescent thyroid glands are softer, while middle-aged and elderly glands tend to have higher overall hardness due to fibrosis. The same hardness value corresponds to completely different lesion risks in different populations. Directly using the original quantitative elastography parameters of nodules can lead to a large number of false positive and false negative diagnostic results. Therefore, to address these issues, a quantitative elastography assessment method for benign and malignant thyroid nodules has been proposed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a quantitative elastography method for benign and malignant thyroid nodules.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A quantitative elastography method for assessing benign and malignant thyroid nodules, comprising the following steps:

[0007] S1. Collect raw data of elastography images corresponding to the thyroid nodules to be detected;

[0008] S2. Perform data cleaning on all raw data of elastic ultrasound images to remove useless data and blurry imaging data;

[0009] S3. Collect and store the raw data of elastic ultrasound images after the cleaning operation is completed for later use.

[0010] S4. Delineate the region of interest and the surrounding normal thyroid parenchyma reference area within the elastography image;

[0011] S5. Based on the raw data of the elastography image, calculate and output the corresponding quantitative elastography parameters of the region of interest and the surrounding normal thyroid parenchyma reference area.

[0012] S6. Use the elasticity quantitative parameters of the surrounding normal thyroid parenchyma reference area to correct the elasticity quantitative parameters of the region of interest, and calculate to generate standardized elasticity assessment parameters.

[0013] S7. Match the standardized elasticity assessment parameters with the preset benign and malignant risk grading standards, and combine the two-dimensional ultrasound signs of thyroid nodules to comprehensively determine the benign and malignant risk of nodules, and complete the quantitative assessment of thyroid nodules by elastic ultrasound.

[0014] Preferably, S2 data cleaning includes a three-level hierarchical processing flow:

[0015] Primary cleaning removes useless raw elastic ultrasound image data due to imaging truncation and signal loss.

[0016] Secondary cleaning removes blurry, raw elastic ultrasound images with grayscale differences below a preset threshold and indistinguishable nodule boundaries.

[0017] The third-level cleaning process removes redundant raw elastic ultrasound images of the same nodule that are repeatedly acquired or whose body position deviation exceeds the standard range.

[0018] After the three-stage cleaning is completed, a standardized raw image dataset with a unified format and pixel size is generated and directly transferred to S3 for collection and storage. The three-stage cleaning synchronously records the removal type, removal time, and original image number of each removed data, forming a traceable cleaning log stored in the local storage module of the ultrasound equipment.

[0019] Preferably, the S4 delineation of the region of interest and the surrounding normal thyroid parenchyma reference area follow a unified delineation rule:

[0020] The region of interest completely encompasses the entire solid area of ​​the thyroid nodule, with a 1mm buffer zone extending outward from the boundary. It does not include the calcified or cystic liquefaction areas around the nodule.

[0021] The reference area for the surrounding normal thyroid parenchyma is selected as a homogeneous thyroid gland region on the same side as the nodule, within 3mm to 8mm from the edge of the nodule, without calcification, inflammatory infiltration, or microproliferative foci. The area of ​​the reference area is not less than 1.2 times the area of ​​the region of interest. Only a single set of matched regions of interest and the surrounding normal thyroid parenchyma reference area are delineated in the same frame of elastography image. The delineation operation supports both manual fine-tuning and automatic delineation by the device's AI. After the delineation is completed, the coordinate parameters of the two areas are locked and synchronously bound to the corresponding original elastography image data.

[0022] Preferably, the S5 elastic quantitative parameters include four fixed parameters: mean Young's modulus, maximum Young's modulus, elastic strain ratio, and tissue stiffness dispersion coefficient. All parameters are calculated based on the pixel elastic strain gray value matrix inside the original data of the elastic ultrasound image. The four elastic quantitative parameters are output synchronously, stored independently, and bound one by one with the coordinates of the corresponding region of interest and the surrounding normal thyroid parenchyma reference area.

[0023] Preferably, the S6 correction calculation formula is: Standardized elasticity assessment parameter = single type of elasticity quantitative parameter of region of interest ÷ same type of elasticity quantitative parameter of reference area of ​​surrounding normal thyroid parenchyma. The four types of parameters, namely the mean Young's modulus, the maximum Young's modulus, the elastic strain ratio, and the tissue stiffness dispersion coefficient, are independently corrected to generate four sets of independent standardized elasticity assessment parameters. The four sets of parameters together constitute a complete quantitative assessment index set for nodules.

[0024] Preferably, the S7 preset benign and malignant risk grading standard is divided into five levels, and the grading threshold is determined based on a large sample of thyroid nodule pathological control data.

[0025] Level 1 corresponds to very low risk of benign disease; Level 2 corresponds to low risk of benign disease; Level 3 corresponds to risk at the borderline between benign and malignant disease; Level 4 corresponds to high risk of malignant disease; and Level 5 corresponds to very high risk of malignant disease.

[0026] During the tiered matching process, four sets of standardized resilience assessment parameters are simultaneously used for threshold comparison. If any two or more parameters exceed the malignancy threshold, the risk level is upgraded by one level.

[0027] Preferably, the two-dimensional ultrasound features of S7 thyroid nodules include six core features: nodule morphology, boundary clarity, internal echo type, presence of microcalcifications, aspect ratio, and capsule invasion. Each two-dimensional ultrasound feature is pre-classified into three categories: benign, suspicious, and malignant. All feature judgment labels are simultaneously entered into the evaluation system and linked with standardized elasticity assessment parameters for analysis.

[0028] The preferred comprehensive evaluation logic is as follows: First, output the basic risk level based on four sets of standardized elastic assessment parameters. Then, adjust the basic risk level upward based on the number of malignant labels in the six two-dimensional ultrasound features. When there are ≥3 malignant labels, the basic risk level is directly increased by one level. When there are ≥5 malignant labels, the basic risk level is directly increased by two levels. Finally, the adjusted comprehensive risk result of benign and malignant nodules is output.

[0029] Preferably, the S3 collection and storage adopts a dual storage mode of encrypted local storage + cloud backup. The collected elastic ultrasound image raw data is bound to a four-dimensional index label with patient number, acquisition time, thyroid left and right lobe location, and nodule number. When retrieving data, the corresponding image is quickly located by relying on the four-dimensional index.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] 1. This invention establishes a synchronous measurement and parameter correction calculation process for the surrounding normal thyroid parenchyma reference area, solving the problem of individual glandular hardness interference caused by the lack of baseline correction in existing technologies. Existing solutions directly use the original quantitative elasticity parameters of nodules for grading, which can lead to confusion in risk assessment between soft glands in adolescents and fibrotic glands in middle-aged and elderly individuals with the same values. This invention uses the patient's own normal glands within the same frame of image as a reference, and generates standardized elasticity assessment parameters through ratio calculation to offset the data deviation caused by the basic thyroid hardness of different populations. The four-level fixed quantitative elasticity parameters are synchronously corrected to avoid misjudgment caused by fluctuations in a single parameter.

[0032] 2. This invention introduces a three-level standardized data cleaning process to unify the pre-processing specifications for raw elastic ultrasound image data, addressing the issue of inconsistent image quality in clinically acquired data interfering with computational accuracy. Routine clinical examinations generate a large amount of truncated, blurred, and repeatedly acquired invalid raw elastic ultrasound image data. Existing equipment lacks a unified cleaning mechanism, and directly using substandard images in calculations results in distorted elasticity quantitative parameters. This invention sequentially removes useless, blurred, and redundant images in three levels, simultaneously generating a cleaning log to record the reason and number for each data removal. Filtered images are uniformly formatted and stored according to pixel and size specifications, ensuring the accuracy of subsequent elasticity quantitative parameter calculations from the data source level.

[0033] 3. This invention integrates standardized elasticity assessment parameters with six two-dimensional ultrasound feature labels for coordinated evaluation, overcoming the limitations of existing technologies that rely solely on a single dimension of elastography for diagnosis. Elastography only reflects tissue stiffness and cannot identify malignant features such as nodule capsule invasion, microcalcification, and aspect ratio. This invention pre-classifies nodule morphology, boundary, echogenicity, microcalcification, aspect ratio, and capsule features into three categories of labels. The basic risk level of the elasticity parameter output is dynamically adjusted based on the number of malignant labels, establishing a two-dimensional complementary evaluation logic. Attached Figure Description

[0034] Figure 1 This is a flowchart of the quantitative elastography method for evaluating benign and malignant thyroid nodules proposed in this invention;

[0035] Figure 2 This is a flowchart of the data cleaning process for the quantitative elastography method for benign and malignant thyroid nodules proposed in this invention.

[0036] Figure 3 This is a diagram showing the risk grading standards for benign and malignant thyroid nodules according to the quantitative elastography assessment method for benign and malignant thyroid nodules proposed in this invention. Detailed Implementation

[0037] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0038] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0039] Reference Figure 1-3 A quantitative elastography method for evaluating benign and malignant thyroid nodules, comprising the following steps:

[0040] S1. Collect raw data of elastography images corresponding to the thyroid nodules to be detected;

[0041] S2. Perform data cleaning on all raw data of elastic ultrasound images to remove useless data and blurry imaging data;

[0042] S3. Collect and store the raw data of elastic ultrasound images after the cleaning operation is completed for later use.

[0043] S4. Delineate the region of interest and the surrounding normal thyroid parenchyma reference area within the elastography image;

[0044] S5. Based on the raw data of the elastography image, calculate and output the corresponding quantitative elastography parameters of the region of interest and the surrounding normal thyroid parenchyma reference area.

[0045] S6. Use the elasticity quantitative parameters of the surrounding normal thyroid parenchyma reference area to correct the elasticity quantitative parameters of the region of interest, and calculate to generate standardized elasticity assessment parameters.

[0046] S7. Match the standardized elasticity assessment parameters with the preset benign and malignant risk grading standards, and combine the two-dimensional ultrasound signs of thyroid nodules to comprehensively determine the benign and malignant risk of nodules, and complete the quantitative assessment of thyroid nodules by elastic ultrasound.

[0047] As a technical optimization of the present invention, S2 data cleaning includes a three-level hierarchical processing flow:

[0048] Primary cleaning removes useless raw elastic ultrasound image data due to imaging truncation and signal loss.

[0049] Secondary cleaning removes blurry, raw elastic ultrasound images with grayscale differences below a preset threshold and indistinguishable nodule boundaries.

[0050] The third-level cleaning process removes redundant raw elastic ultrasound images of the same nodule that are repeatedly acquired or whose body position deviation exceeds the standard range.

[0051] After the three-stage cleaning is completed, a standardized raw image dataset with a unified format and pixel size is generated and directly transferred to S3 for collection and storage. The three-stage cleaning synchronously records the removal type, removal time, and original image number of each removed data, forming a traceable cleaning log stored in the local storage module of the ultrasound equipment.

[0052] As a technical optimization of the present invention, S4 delineates the region of interest and the surrounding normal thyroid parenchyma reference area according to a unified delineation rule:

[0053] The region of interest completely encompasses the entire solid area of ​​the thyroid nodule, with a 1mm buffer zone extending outward from the boundary. It does not include the calcified or cystic liquefaction areas around the nodule.

[0054] The reference area for the surrounding normal thyroid parenchyma is selected as a homogeneous thyroid gland region on the same side as the nodule, within 3mm to 8mm from the edge of the nodule, without calcification, inflammatory infiltration, or microproliferative foci. The area of ​​the reference area is not less than 1.2 times the area of ​​the region of interest. Only a single set of matched regions of interest and the surrounding normal thyroid parenchyma reference area are delineated in the same frame of elastography image. The delineation operation supports both manual fine-tuning and automatic delineation by the device's AI. After the delineation is completed, the coordinate parameters of the two areas are locked and synchronously bound to the corresponding original elastography image data.

[0055] As a technical optimization scheme of the present invention, the S5 elastic quantitative parameters include four fixed parameters: mean Young's modulus, maximum Young's modulus, elastic strain ratio, and tissue stiffness dispersion coefficient. All parameters are obtained by calculation based on the pixel elastic strain gray value matrix inside the original data of elastic ultrasound image. The four elastic quantitative parameters are output synchronously, stored independently, and bound one by one with the coordinates of the corresponding region of interest and the surrounding normal thyroid parenchyma reference area.

[0056] As a technical optimization scheme of the present invention, the S6 correction calculation formula is: Standardized elasticity assessment parameter = single type of elasticity quantitative parameter of region of interest ÷ same type of elasticity quantitative parameter of surrounding normal thyroid parenchyma reference area. The four types of parameters, namely the mean Young's modulus, the maximum Young's modulus, the elastic strain ratio, and the tissue stiffness dispersion coefficient, are independently corrected to generate four sets of independent standardized elasticity assessment parameters. The four sets of parameters together constitute a complete quantitative assessment index set for nodules.

[0057] As a technical optimization of the present invention, S7 presets a benign and malignant risk grading standard into five levels, and the grading threshold is determined based on a large sample of thyroid nodule pathological control data.

[0058] Level 1 corresponds to very low risk of benign disease; Level 2 corresponds to low risk of benign disease; Level 3 corresponds to risk at the borderline between benign and malignant disease; Level 4 corresponds to high risk of malignant disease; and Level 5 corresponds to very high risk of malignant disease.

[0059] During the tiered matching process, four sets of standardized resilience assessment parameters are simultaneously used for threshold comparison. If any two or more parameters exceed the malignancy threshold, the risk level is upgraded by one level.

[0060] As a technical optimization of the present invention, the two-dimensional ultrasound features of S7 thyroid nodules include six core features: nodule morphology, boundary clarity, internal echo type, presence of microcalcifications, aspect ratio, and capsule invasion. Each two-dimensional ultrasound feature is pre-classified into three categories: benign, suspicious, and malignant. All feature judgment labels are simultaneously entered into the evaluation system and linked with standardized elastic evaluation parameters for analysis.

[0061] As a technical optimization scheme of the present invention, the comprehensive evaluation logic is as follows: First, output the basic risk level based on four sets of standardized elastic assessment parameters, and then adjust the basic risk level upward based on the number of malignant labels in the six two-dimensional ultrasound signs. When there are ≥3 malignant labels, the basic risk level is directly increased by one level, and when there are ≥5 malignant labels, the basic risk level is directly increased by two levels. Finally, the corrected comprehensive risk result of benign and malignant nodules is output.

[0062] As a technical optimization of the present invention, the S3 collection and storage adopts a dual storage mode of encrypted local storage + cloud backup. The collected elastic ultrasound image raw data is bound to a four-dimensional index label with patient number, acquisition time, thyroid left and right lobe location, and nodule number. When retrieving data, the corresponding image is quickly located by relying on the four-dimensional index.

[0063] This embodiment uses a conventional color Doppler elastography diagnostic instrument. The device has a built-in custom-designed calculation software module, eliminating the need for additional hardware. The software includes pre-set standardized benign / malignant risk grading criteria, two-dimensional ultrasound sign identification label templates, data cleaning threshold parameters, and standardized elastography parameter correction calculation formulas. The subjects in this study were female patients with a single solid thyroid nodule. Preoperative elastography and fine-needle aspiration biopsy were performed simultaneously, and the pathology ultimately diagnosed papillary thyroid carcinoma.

[0064] The first step involves acquiring raw elastography data corresponding to the thyroid nodule to be detected. The clinician adjusts the ultrasound probe frequency to the thyroid-specific 5-12MHz band, instructs the patient to lie supine with their neck fully extended to expose the anterior thyroid region, applies coupling gel to the probe to ensure it adheres tightly to the skin, avoiding areas obstructed by neck blood vessels and the trachea, and continuously and stably applies pressure to the single nodule in the right lobe of the thyroid gland to acquire an elastography sequence. A single examination continuously acquires 25 frames of images. The raw pixel matrix, grayscale strain data, imaging timestamps, and probe coordinate information of all images are simultaneously packaged to generate raw elastography data. The device automatically binds the patient's unique medical record number, the acquisition location in the right lobe of the thyroid gland, and the nodule's index label (No. 1). After acquisition, all raw elastography data is transmitted in real-time to the device's built-in processing module for processing. Of the 25 frames of raw elastography data acquired in this study, 2 frames showed probe misalignment and blurred nodule boundaries, 3 frames were redundant with no change, and the remaining 20 frames were complete and clear. This portion of raw data proceeds to the next step, the S2 data cleaning process.

[0065] The second step involves data cleaning of all raw elastography images, strictly following a three-tiered cleaning process (Level 1, Level 2, and Level 3). The entire process is automated, generating cleaning logs stored on the local hard drive. Level 1 cleaning first screens all 25 sets of raw elastography images, identifying images with truncated images or signal loss. Since no images with signal loss were found in this sample, no data was removed. Level 2 cleaning automatically calculates the overall grayscale difference for each frame, with a preset grayscale difference threshold of 45. Images below this value are considered blurry elastography images. This screening identified two blurry images with grayscale differences of only 28 and 32, which were automatically marked and removed. The cleaning log records the removal type: Level 2 blurry data, image number, and removal time. Level 3 cleaning compares the pixel matrix similarity of the 25 images. Images with a similarity higher than 92% are considered redundant elastography images. This screening identified three highly repetitive images (numbers 009, 013, and 021), which were simultaneously marked and removed, and written into the cleaning log. After the three-stage cleaning is completed, the remaining 20 frames of qualified elastic ultrasound image raw data are automatically unified by the software to a pixel size of 800×600 and a grayscale storage format, forming a standardized raw image dataset, which is then automatically pushed to the collection and storage module.

[0066] The third step involves collecting and storing the raw elastic ultrasound images after the cleaning process. A dual storage mode of encrypted local storage and cloud backup is used, and the data is archived according to four-dimensional index tags. The four-dimensional index tags are patient visit number, acquisition time, location of the left and right thyroid lobes, and nodule number. In this dataset, the four-dimensional tags are fixed as: visit number, acquisition time, right thyroid lobe, and nodule number 1. Local storage uses the device's built-in encrypted solid-state drive, while cloud backup is synchronously transmitted to a dedicated encrypted server in the hospital's radiology department. Cloud data transmission is encrypted with SSL throughout to prevent leakage of patient image data. After collection, 20 qualified raw elastic ultrasound images are generated into a separate data folder. The software is set to retrieve the channel with one click, allowing for quick access to any clear image frame for subsequent area selection. Frame 003, with the nodule centered, gland unobstructed, and uniform grayscale, is prioritized as the core calculation frame.

[0067] The fourth step involves delineating the region of interest (ROI) and the surrounding normal thyroid parenchyma reference area within the elastography image, supporting both AI-automated delineation and physician-manual fine-tuning modes. The software first uses image recognition algorithms to automatically identify the thyroid nodule entity in frame 003, initially generating the ROI outline. This outline automatically and completely covers the entire nodule entity area, with a 1mm buffer zone extending outwards. The algorithm automatically avoids small cystic liquefaction and punctate calcification areas within the nodule, ensuring that hard calcifications are not included in the ROI. When automatically delineating the surrounding normal thyroid parenchyma reference area, the right lobe of the thyroid gland on the same side as the nodule is selected, along with a homogeneous, non-proliferative, and non-inflammatory normal glandular area 4mm from the nodule edge. The calculated reference area area is 1.35 times the ROI area, meeting the minimum requirement of 1.2 times. Only one matching set of two-area outlines is generated for each frame.

[0068] The fifth step involves calculating and outputting the quantitative elastic parameters corresponding to the region of interest (ROI) and the surrounding normal thyroid parenchyma reference area based on the raw data of the elastography image: mean Young's modulus, maximum Young's modulus, elastic strain ratio, and tissue stiffness dispersion coefficient. The software retrieves the grayscale value matrix of all pixels within the raw elastography image of frame 003, extracts the pixel matrix of the ROI and the pixel matrix of the surrounding normal thyroid parenchyma reference area, calculates them independently, and outputs the values ​​of the four quantitative elastic parameters for the two regions separately.

[0069] The sixth step involves correcting the region of interest's elasticity quantitative parameters using the elasticity quantitative parameters of the surrounding normal thyroid parenchyma reference area. The formula is standardized as follows: Standardized elasticity assessment parameter = Single type of elasticity quantitative parameter of region of interest ÷ Same type of elasticity quantitative parameter of the surrounding normal thyroid parenchyma reference area. The correction results for the four types of parameters are calculated sequentially: Standardized parameter for the mean Young's modulus = 68.2 ÷ 16.4 = 4.16; Standardized parameter for the maximum Young's modulus = 116.5 ÷ 24.8 = 4.70; Standardized parameter for the elastic strain ratio = 7.3 ÷ 1.2 = 6.08; Standardized parameter for the tissue stiffness dispersion coefficient = 0.42 ÷ 0.08 = 5.25. These four sets of standardized elasticity assessment parameters are output simultaneously, forming a complete quantitative assessment index set for this nodule. The software automatically stores the four sets of values.

[0070] The seventh step involves matching standardized elastic assessment parameters with preset benign and malignant risk grading standards, simultaneously combining two-dimensional ultrasound findings of the thyroid nodule to comprehensively determine the benign or malignant risk. This process is divided into three steps: basic grading, finding the correct risk level, and outputting the final risk. The first step, basic grading matching, uses four preset standardized elastic assessment parameters with malignant thresholds: mean Young's modulus ≥ 3.0, maximum Young's modulus ≥ 3.5, elastic strain ratio ≥ 4.0, and tissue stiffness dispersion coefficient ≥ 3.0. In this case, all four standardized elastic assessment parameters exceeded the malignant threshold, satisfying the rule of increasing the risk level by one level for any two parameters exceeding the threshold. The basic risk level is directly determined as Grade IV, high risk of malignancy. The second step involves inputting six two-dimensional ultrasound findings and labeling them. A total of five malignant labels are counted. When there are ≥ 5 malignant labels, the basic risk level is increased by two levels. After increasing the basic risk level by two levels from Grade IV, it becomes Grade V, extremely high risk of malignancy. The third step integrates the four sets of standardized elastic assessment parameter values, the grading matching process, the six two-dimensional ultrasound findings labels, and the risk correction logic to output the final comprehensive risk result.

[0071] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A quantitative elastography method for evaluating benign and malignant thyroid nodules, characterized in that, The evaluation method includes the following steps: S1. Collect raw data of elastography images corresponding to the thyroid nodules to be detected; S2. Perform data cleaning on all raw data of elastic ultrasound images to remove useless data and blurry imaging data; S3. Collect and store the raw data of elastic ultrasound images after the cleaning operation is completed for later use. S4. Delineate the region of interest and the surrounding normal thyroid parenchyma reference area within the elastography image; S5. Based on the raw data of the elastography image, calculate and output the corresponding quantitative elastography parameters of the region of interest and the surrounding normal thyroid parenchyma reference area. S6. Use the elasticity quantitative parameters of the surrounding normal thyroid parenchyma reference area to correct the elasticity quantitative parameters of the region of interest, and calculate to generate standardized elasticity assessment parameters. S7. Match the standardized elasticity assessment parameters with the preset benign and malignant risk grading standards, and combine the two-dimensional ultrasound signs of thyroid nodules to comprehensively determine the benign and malignant risk of nodules, and complete the quantitative assessment of thyroid nodules by elastic ultrasound.

2. The method for quantitative evaluation of benign and malignant thyroid nodules using elastography according to claim 1, characterized in that, S2 data cleaning includes a three-level, layered processing flow: Primary cleaning removes useless raw elastic ultrasound image data due to imaging truncation and signal loss. Secondary cleaning removes blurry, raw elastic ultrasound images with grayscale differences below a preset threshold and indistinguishable nodule boundaries. The third-level cleaning process removes redundant raw elastic ultrasound images of the same nodule that are repeatedly acquired or whose body position deviation exceeds the standard range. After the three-stage cleaning is completed, a standardized raw image dataset with a unified format and pixel size is generated and directly transferred to S3 for collection and storage. The three-stage cleaning synchronously records the removal type, removal time, and original image number of each removed data, forming a traceable cleaning log stored in the local storage module of the ultrasound equipment.

3. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 1, characterized in that, The S4 delineation of the region of interest and the surrounding normal thyroid parenchyma reference area follows a unified delineation rule: The region of interest completely encompasses the entire solid area of ​​the thyroid nodule, with a 1mm buffer zone extending outward from the boundary. It does not include the calcified or cystic liquefaction areas around the nodule. The reference area for the surrounding normal thyroid parenchyma is selected as a homogeneous thyroid gland region on the same side as the nodule, within 3mm to 8mm from the edge of the nodule, without calcification, inflammatory infiltration, or microproliferative foci. The area of ​​the reference area is not less than 1.2 times the area of ​​the region of interest. Only a single set of matched regions of interest and the surrounding normal thyroid parenchyma reference area are delineated in the same frame of elastography image. The delineation operation supports both manual fine-tuning and automatic delineation by the device's AI. After the delineation is completed, the coordinate parameters of the two areas are locked and synchronously bound to the corresponding original elastography image data.

4. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 1, characterized in that, The S5 quantitative elasticity parameters include four fixed parameters: mean Young's modulus, maximum Young's modulus, elastic strain ratio, and tissue stiffness dispersion coefficient. All parameters are calculated based on the pixel elastic strain grayscale value matrix within the original data of the elastic ultrasound image. The four quantitative elasticity parameters are output synchronously, stored independently, and bound one-to-one with the coordinates of the corresponding region of interest and the surrounding normal thyroid parenchyma reference area.

5. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 4, characterized in that, The S6 correction formula is: Standardized elasticity assessment parameter = single-type elasticity quantitative parameter of region of interest ÷ same-type elasticity quantitative parameter of surrounding normal thyroid parenchyma reference area. The four types of parameters, namely the mean Young's modulus, the maximum Young's modulus, the elastic strain ratio, and the tissue stiffness dispersion coefficient, are independently corrected to generate four sets of independent standardized elasticity assessment parameters. The four sets of parameters together constitute a complete quantitative assessment index set for nodules.

6. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 1, characterized in that, The S7 pre-defined benign and malignant risk grading standard is divided into five levels, and the grading threshold is based on a large sample of thyroid nodule pathological control data. Level 1 corresponds to very low risk of benign disease; Level 2 corresponds to low risk of benign disease; Level 3 corresponds to risk at the borderline between benign and malignant disease; Level 4 corresponds to high risk of malignant disease; and Level 5 corresponds to very high risk of malignant disease. During the tiered matching process, four sets of standardized resilience assessment parameters are simultaneously used for threshold comparison. If any two or more parameters exceed the malignancy threshold, the risk level is upgraded by one level.

7. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 1, characterized in that, The two-dimensional ultrasound features of S7 thyroid nodules include six core features: nodule morphology, border clarity, internal echo type, presence of microcalcifications, aspect ratio, and capsule invasion. Each feature is pre-classified into three categories: benign, suspicious, and malignant. All feature labels are simultaneously entered into the evaluation system and linked with standardized elasticity assessment parameters for analysis.

8. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 7, characterized in that, The comprehensive evaluation logic is as follows: First, the basic risk level is output based on four sets of standardized elastic assessment parameters. Then, the basic risk level is adjusted upward based on the number of malignant labels in the six two-dimensional ultrasound features. When there are ≥3 malignant labels, the basic risk level is directly increased by one level. When there are ≥5 malignant labels, the basic risk level is directly increased by two levels. Finally, the adjusted comprehensive risk result of benign and malignant nodules is output.

9. The method for quantitative evaluation of benign and malignant thyroid nodules by elastography according to claim 1, characterized in that, The S3 collection and storage adopts a dual storage mode of encrypted local storage + cloud backup. The collected elastic ultrasound image raw data is bound to a four-dimensional index label with patient number, acquisition time, thyroid left and right lobe location, and nodule number. When retrieving data, the corresponding image is quickly located by relying on the four-dimensional index.

Images