Biomarker-based thyroid nodule benign and malignant identification method and system

Abstract

The invention relates to the field of medical informatics, and provides a biomarker-based thyroid nodule benign and malignant identification method and system. The method comprises the following steps: collecting a serum sample; performing quantitative detection on the multiple markers in the serum sample to obtain concentration data of the multiple markers; inputting the multi-marker concentration data into a pre-constructed probability calculation model to obtain a prediction probability value; and according to a preset diagnosis threshold value and the prediction probability value, judging whether the thyroid nodule is benign or malignant to obtain a thyroid nodule benign or malignant identification result. By integrating multi-marker information, the biological difference between benign and malignant nodules is effectively captured, and the overall efficiency of differential diagnosis is improved.

Description

Technical Field

[0001] This invention relates to the field of medical informatics technology, and in particular to a method and system for differentiating benign and malignant thyroid nodules based on biomarkers. Background Technology

[0002] Thyroid nodules are a common endocrine disorder in clinical practice. Differentiated thyroid cancer, the most common subtype of thyroid cancer, is experiencing a gradual increase in incidence. Currently, thyroid color Doppler ultrasound is the preferred diagnostic method for thyroid nodules, and ultrasound-guided fine-needle aspiration cytology (FNA) is considered the gold standard for differentiating between benign and malignant thyroid nodules. Simultaneously, serum tumor marker testing is widely used in the auxiliary diagnosis of thyroid nodules, with thyroglobulin (TG) and carcinoembryonic antigen (CEA) being the most commonly used serum markers.

[0003] However, existing technologies have significant shortcomings in differentiating between benign and malignant thyroid nodules: the accuracy of ultrasound examination depends on the operator's experience level, and there is a significant risk of misdiagnosis for small nodules or nodules with complex characteristics. While FNA examination has high diagnostic accuracy, it carries the risk of complications such as bleeding and infection, as well as the possibility of false negative results. Although the detection of single serum markers such as TG or CEA has some diagnostic value, its sensitivity and specificity are not ideal, and it is easily affected by benign diseases such as thyroiditis. Even the combined detection of TG and CEA can improve diagnostic specificity to some extent, but the overall sensitivity is still insufficient, and the effect on the early detection of certain subtypes of malignant thyroid nodules is not significant, failing to meet the urgent clinical need for high-precision, non-invasive diagnostic methods. Summary of the Invention

[0004] This invention provides a method and system for differentiating benign and malignant thyroid nodules based on biomarkers, in order to overcome the shortcomings of existing technologies.

[0005] This invention provides a method for differentiating benign and malignant thyroid nodules based on biomarkers, comprising: S1: Collect serum samples; S2: Quantitatively detect multiple biomarkers in the serum sample to obtain concentration data of multiple biomarkers; S3: Input the multi-marker concentration data into the pre-constructed probability calculation model to obtain the predicted probability value; S4: Based on the preset diagnostic threshold and the predicted probability value, the benign or malignant thyroid nodules are determined, and the results of the benign or malignant thyroid nodule identification are obtained.

[0006] According to the present invention, a method for differentiating benign and malignant thyroid nodules based on biomarkers, in step S2, the concentration data of the multiple biomarkers includes: Concentration values ​​of growth differentiation factor 15, thyroglobulin, and carcinoembryonic antigen.

[0007] According to the present invention, a method for differentiating benign and malignant thyroid nodules based on biomarkers, step S2 specifically includes: The serum sample was quantitatively analyzed by enzyme-linked immunosorbent assay (ELISA) to obtain the concentration value of growth differentiation factor 15. The serum sample was subjected to combined detection of thyroglobulin and carcinoembryonic antigen using chemiluminescent immunoassay to obtain the concentration values ​​of thyroglobulin and carcinoembryonic antigen.

[0008] According to the present invention, in a method for differentiating benign and malignant thyroid nodules based on biomarkers, in step S3, the probability calculation model is a binary logistic regression model.

[0009] According to the present invention, in a method for differentiating benign and malignant thyroid nodules based on biomarkers, in step S3, the expression for the predicted probability value is:

[0010] in, This refers to the predicted probability value output by the probability calculation model. For model constants, The first regression coefficient, The second regression coefficient, The third regression coefficient, For conversion functions, This represents the concentration of growth differentiation factor 15. This represents the concentration of thyroglobulin. This represents the concentration value of carcinoembryonic antigen.

[0011] According to the present invention, a method for differentiating benign and malignant thyroid nodules based on biomarkers, in step S4, the results of the benign and malignant thyroid nodule identification include: When the predicted probability value is less than or equal to the preset diagnostic threshold, the benign or malignant differentiation result of the thyroid nodule is benign. When the predicted probability value is greater than the preset diagnostic threshold, the thyroid nodule is classified as malignant.

[0012] According to the present invention, in a method for differentiating benign and malignant thyroid nodules based on biomarkers, in step S4, the preset diagnostic threshold is the probability value corresponding to the maximum Youden index of the probability calculation model.

[0013] A second aspect of the present invention provides a system for differentiating benign and malignant thyroid nodules based on biomarkers, comprising: Receiving module: Used to receive the concentration data of multiple biomarkers obtained from the quantitative detection of multiple biomarkers in patient serum samples; A calculation module, configured as a pre-built probability calculation model, is used to output predicted probability values ​​based on the multi-marker concentration data; The discrimination module is used to distinguish between benign and malignant thyroid nodules based on a preset diagnostic threshold and the predicted probability value, and to obtain the benign and malignant identification results of thyroid nodules.

[0014] A third aspect of the present invention provides a device for differentiating benign and malignant thyroid nodules based on biomarkers, comprising: A memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause a biomarker-based thyroid nodule benign / malignant identification device to perform a biomarker-based thyroid nodule benign / malignant identification method as described in any of the preceding claims.

[0015] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement a method for differentiating benign and malignant thyroid nodules based on biomarkers as described in any of the preceding claims.

[0016] This invention provides a method and system for differentiating benign and malignant thyroid nodules based on biomarkers. Firstly, it uses serum samples to provide data support, avoiding the dependence on operator experience in traditional ultrasound examinations and the risks of complications such as bleeding and infection associated with fine-needle aspiration cytology. This provides patients with a safe, non-invasive, and repeatable testing method, significantly reducing patient suffering and medical risks. Secondly, it uses immunoassay technology to quantitatively detect multiple biomarkers in serum samples, simultaneously obtaining the concentration values ​​of three biomarkers: growth differentiation factor 15, thyroglobulin, and carcinoembryonic antigen. Compared to single biomarker detection, this method can capture characteristic information of thyroid nodules from multiple biological dimensions, significantly reducing the problem of single biomarker interference from benign diseases or insufficient sensitivity to specific cancer types, providing a more comprehensive and reliable data foundation for subsequent risk assessment. In data processing, the log1p function is used to analyze the growth... Logarithmic transformation of differentiation factor 15 and carcinoembryonic antigen concentration values ​​resolved common issues of skewed distribution and zero values ​​in biomarker data, making the data more consistent with the assumptions of the statistical model and thus improving the model's stability and prediction accuracy. Subsequently, a pre-constructed binary logistic regression model was used to calculate the risk probability of multi-marker data, scientifically integrating information from various markers and quantifying malignancy risk through mathematical modeling. This avoided biases in subjective judgment, making diagnostic results more objective and standardized. Furthermore, the selection of the optimal diagnostic threshold ensured an optimal balance between sensitivity and specificity, minimizing false positives and false negatives. This provides clinicians with more reliable diagnostic evidence, helps optimize the allocation of medical resources, reduces unnecessary overtreatment, and improves the early detection rate of malignant nodules, thus securing better treatment opportunities and prognoses for patients. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 A schematic diagram of a method for differentiating benign and malignant thyroid nodules based on biomarkers provided by this invention; Figure 2 A schematic diagram of a biomarker-based system for differentiating benign and malignant thyroid nodules is provided for this invention. Figure 3 This is a schematic diagram comparing serum marker levels in patients with benign and malignant thyroid nodules with those with malignant thyroid nodules, as provided by the present invention. Figure 4This is a schematic diagram comparing the diagnostic and differential diagnostic efficacy of different indicators and combinations provided by the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0020] To better understand this invention, the research background of this invention will be explained in detail below.

[0021] Thyroid nodules are quite common in clinical practice. The risk of developing palpable thyroid nodules increases with age, with a lifetime risk estimated at 5-10%. Differentiated thyroid cancer is the most common subtype of thyroid cancer, and its incidence is gradually increasing. Despite the high prevalence of thyroid nodules, only 7% to 15% are diagnosed as malignant. The current clinical challenge in treating thyroid nodules is finding the most appropriate and systematic approach to reliably and cost-effectively differentiate between benign and malignant nodules.

[0022] However, traditional diagnostic methods have the following limitations: Thyroid ultrasound is the most common examination method for thyroid diseases, but its accuracy is lower than that of surgical pathology results. Ultrasound-guided fine-needle aspiration cytology (FNA) carries the risk of bleeding and false negatives; single marker detection (such as TG) has certain diagnostic value, but its sensitivity and specificity are insufficient and it is affected by disease states such as thyroiditis; although the combination of TG and CEA can improve specificity, the sensitivity is still low.

[0023] However, existing technologies have the following problems and shortcomings in the differential diagnosis of benign and malignant thyroid nodules: Limitations of ultrasound examination: As the most commonly used examination method for thyroid diseases, ultrasound has relatively low accuracy. Its diagnostic ability depends on the operator's experience, and there is a large margin of error in small nodules or nodules with complex characteristics. In addition, although ultrasound-guided fine-needle aspiration cytology (FNA) improves diagnostic accuracy, it carries certain risks (such as bleeding or false negative results).

[0024] Limitations of single biomarker detection: For example, detecting serum thyroglobulin (TG) or carcinoembryonic antigen (CEA) alone, although valuable in certain situations, has low sensitivity and specificity.

[0025] Limitations of classic biomarker combinations: Although combined detection of thyroglobulin (TG) and carcinoembryonic antigen (CEA) can improve diagnostic specificity to some extent, the sensitivity is still insufficient, and the effect on early detection of certain subtypes of malignant thyroid nodules is not significant.

[0026] Growth differentiation factor 15 (GDF15) is an important member of the transforming growth factor β (TGF-β) family. Its circulating levels are significantly correlated with various malignant tumors, showing a sharp increase during tumor tissue stress or malignant transformation. Studies in papillary thyroid carcinoma (PTC) have found that GDF15 circulating levels are significantly higher than in healthy individuals, which may be closely related to tumor development and progression. However, no studies have yet explored the application of GDF15 in combination with classic tumor markers (TG and CEA) in differentiating benign from malignant thyroid nodules.

[0027] This invention addresses this technological gap by proposing an optimized diagnostic model based on combined detection. The technical objective of this invention is to establish a highly sensitive and specific model for differentiating benign from malignant thyroid nodules through the combined detection of GDF15 and classic thyroid tumor markers (TG, CEA), providing a non-invasive, economical, and efficient diagnostic method for clinical use, overcoming the limitations of single-technology approaches.

[0028] The embodiments of the present invention are described below with reference to the figures.

[0029] like Figure 1 As shown, this invention provides a method for differentiating benign and malignant thyroid nodules based on biomarkers, comprising: S1: Collect serum samples.

[0030] Specifically, in step S1, the present invention first collects peripheral venous blood from the subject who needs to determine the nature of thyroid nodules, and then processes the blood sample according to the clinical standard operating procedure, separating and collecting serum.

[0031] S2: Quantitatively detect multiple biomarkers in the serum sample to obtain concentration data of multiple biomarkers.

[0032] In step S2, the concentration data of the multiple biomarkers include: the concentration values ​​of growth differentiation factor 15, thyroglobulin, and carcinoembryonic antigen.

[0033] Furthermore, the quantitative detection of multiple biomarkers in step S2 is a biochemical detection process based on immunoassay technology. It uses specific antibodies to bind to target proteins to generate detectable signals to quantitatively determine the concentration of specific biomarkers in serum. The concentration data of multiple biomarkers include three numerical parameters: growth differentiation factor 15 concentration, thyroglobulin concentration, and carcinoembryonic antigen concentration. Each parameter is obtained through different immunoassay methods described below.

[0034] Specifically, step S2 includes: The serum sample was quantitatively analyzed by enzyme-linked immunosorbent assay (ELISA) to obtain the concentration value of growth differentiation factor 15; the serum sample was analyzed by chemiluminescent immunoassay to detect the concentration values ​​of thyroglobulin and carcinoembryonic antigen (CEA) to obtain the concentration values ​​of thyroglobulin and CEA.

[0035] Furthermore, when identifying multiple biomarkers using the methods described above, the concentration value of each biomarker must undergo quality control verification, including the setting of blank controls, negative controls, and positive controls, as well as the evaluation of intra-batch and inter-batch precision. The final quantitative values ​​of growth differentiation factor 15 concentration, thyroglobulin concentration, and carcinoembryonic antigen concentration are obtained in pg / mL, ng / mL, and ng / mL, respectively. The obtained concentration values ​​constitute multi-biomarker concentration data, which are used as input parameters for subsequent Logistic regression model calculations.

[0036] S3: Input the multi-marker concentration data into the pre-constructed probability calculation model to obtain the predicted probability value.

[0037] Furthermore, the pre-construction process involves model training. First, a training dataset is input, which contains multi-marker concentration data of thyroid nodules patients with known benign or malignant diagnoses. Benign nodules are coded as 0, and malignant nodules are coded as 1. Then, a likelihood function is constructed through maximum likelihood estimation to calculate regression coefficients. Subsequently, by solving a system of equations in which the first derivative of the likelihood function equals zero, the optimal regression coefficients are obtained using the Newton-Raphson iterative algorithm or the quasi-Newton method. The iteration continues until the parameters converge or the preset maximum number of iterations is reached, thus obtaining the optimal parameters and the corresponding model.

[0038] In step S3, the probability calculation model is a binary logistic regression model.

[0039] Furthermore, the probability calculation model of the present invention is a mathematical model constructed based on the binary logistic regression algorithm. The model outputs the predicted probability value of the malignant risk of thyroid nodules by performing linear combination and nonlinear transformation on the concentration data of multiple biomarkers. The binary logistic regression model is a generalized linear model, which is specifically used to handle regression analysis problems where the dependent variable is a binary variable. The output of the linear regression is mapped to the probability space between 0 and 1 through the Logit function.

[0040] In step S3, the expression for the predicted probability value is:

[0041] in, This refers to the predicted probability value output by the probability calculation model. For model constants, The first regression coefficient, The second regression coefficient, The third regression coefficient, For conversion functions, This represents the concentration of growth differentiation factor 15. This represents the concentration of thyroglobulin. This represents the concentration value of carcinoembryonic antigen.

[0042] Furthermore, in the prediction process, step S3 first preprocesses the multi-marker concentration data to transform the original concentration data from a skewed distribution to a form closer to a normal distribution, while avoiding the occurrence of negative infinity when the concentration value is zero. Specifically, the concentration values ​​of growth differentiation factor 15 and carcinoembryonic antigen are transformed using log1p, while the concentration value of thyroglobulin is directly used as the model input without transformation.

[0043] The model primarily calculates linear predicted values ​​during computation. Then, an inverse Logit transformation converts these linear predicted values ​​into probability values, ensuring that the output predicted probability value P remains between 0 and 1, conforming to the mathematical definition of probability. The input data, after model computation, outputs a predicted probability value, representing the probability that the thyroid nodule is malignant based on the current multi-marker concentration data. The predicted probability value ranges from 0 to 1; a value closer to 1 indicates a higher risk of malignancy, while a value closer to 0 indicates a higher risk of benignity.

[0044] S4: Based on the preset diagnostic threshold and the predicted probability value, the benign or malignant thyroid nodules are determined, and the results of the benign or malignant thyroid nodule identification are obtained.

[0045] In step S4, the present invention performs benign or malignant thyroid nodules based on the probability values ​​output in step S3. This is a binary classification decision process based on numerical comparison. By comparing the predicted probability values ​​with a preset diagnostic threshold, a clear benign or malignant classification result is output.

[0046] In step S4, the preset diagnostic threshold is the probability value corresponding to the maximum Yoden index of the probability calculation model.

[0047] Furthermore, the Youden Index is a statistical metric for evaluating the performance of a binary classifier. The Youden Index ranges from -1 to 1, with a higher value indicating better overall classifier performance. Specifically, when determining the threshold, a validation dataset is first constructed, containing thyroid nodule patient samples with known benign or malignant diagnoses and their corresponding predicted probability values. Then, the predicted probability values ​​in the validation dataset are arranged in ascending order to obtain a probability sequence. Subsequently, each probability value is used as a candidate threshold, and the corresponding confusion matrix element is calculated. When the threshold is set to 'a', samples with predicted probability values ​​greater than 'a' are classified as malignant, and samples with predicted probability values ​​less than or equal to 'a' are classified as benign. By comparing the predicted classification results with the true labels, the number of TP, FP, TN, and FN is counted. For each candidate threshold 'a', the corresponding sensitivity and specificity are calculated, and then the Youden Index is calculated. After iterating through all candidate thresholds, the threshold with the largest Youden Index is selected as the preset diagnostic threshold.

[0048] In step S4, the results of the benign or malignant differentiation of the thyroid nodules include: When the predicted probability value is less than or equal to the preset diagnostic threshold, the thyroid nodule is benign; when the predicted probability value is greater than the preset diagnostic threshold, the thyroid nodule is malignant.

[0049] In addition, the data structure for the benign / malignant differentiation results includes classification labels and related numerical information. The classification labels are in string or numerical encoding form, with benign nodules encoded as 0 or "benign" and malignant nodules encoded as 1 or "malignant". The related numerical information includes predicted probability values, preset diagnostic thresholds, and probability differences. The absolute value of the probability difference reflects the confidence level of the classification decision; the larger the value, the more reliable the classification result.

[0050] like Figure 2 As shown, the present invention also provides a system for differentiating benign and malignant thyroid nodules based on biomarkers, comprising: Receiving module: Used to receive the concentration data of multiple biomarkers obtained from the quantitative detection of multiple biomarkers in patient serum samples; A calculation module, configured as a pre-built probability calculation model, is used to output predicted probability values ​​based on the multi-marker concentration data; The discrimination module is used to distinguish between benign and malignant thyroid nodules based on a preset diagnostic threshold and the predicted probability value, and to obtain the benign and malignant identification results of thyroid nodules.

[0051] This invention also provides a device for differentiating benign and malignant thyroid nodules based on biomarkers, comprising: A memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause a biomarker-based thyroid nodule benign / malignant identification device to perform a biomarker-based thyroid nodule benign / malignant identification method as described in any of the preceding claims.

[0052] The present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement a method for differentiating benign and malignant thyroid nodules based on biomarkers as described in any of the preceding claims.

[0053] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0054] The following describes a specific embodiment of the present invention, which provides a method and system for differentiating benign and malignant thyroid nodules based on biomarkers.

[0055] First, 96 patients with thyroid nodules who underwent surgical treatment and obtained a clear pathological diagnosis were collected. Inclusion criteria: (1) age ≥18 years; (2) preoperative ultrasound examination showed the presence of thyroid nodules; (3) first thyroid surgery; (4) fasting serum sample collected before surgery; (5) complete clinical and pathological data. Exclusion criteria: (1) history of thyroid surgery, 131I treatment or neck radiotherapy; (2) coexisting malignant tumors in other sites; (3) other diseases that can seriously affect GDF15, TG, and CEA levels (such as severe liver and kidney dysfunction, acute infection, etc.); (4) pregnant or lactating women; (5) currently taking drugs that affect thyroid function. Based on the postoperative pathological results, the patients were divided into a benign group (n=62, including nodular goiter, thyroid adenoma, etc.) and a malignant group (n=34, mainly papillary thyroid carcinoma).

[0056] Subsequently, serum samples from the aforementioned patients, which were frozen at -80°C before surgery, were thawed at room temperature, and the concentration of GDF15 (pg / mL) was measured using a human GDF15 ELISA kit according to the instructions. The concentrations of TG (ng / mL) and CEA (ng / mL) were measured using a fully automated chemiluminescence immunoassay analyzer and its matching kit. All tests were strictly quality controlled.

[0057] In the data analysis and construction and validation of the logistic regression model, R software was used for data analysis. First, biomarker levels were compared: the levels of GDF15, TG, and CEA in patients in the benign and malignant groups were described (median and interquartile range), and the Mann-Whitney U test was used to compare differences between groups. The results showed... Figure 3 As shown. Figure 3 In the study, the median GDF15 concentration in the malignant group was 610 (IQR: 350-1100) pg / mL, significantly higher than the 105 (IQR: 65-160) pg / mL in the benign group (P<0.001); the median TG concentration in the malignant group was 55 (IQR: 25-120) ng / mL, higher than the 38 (IQR: 20-70) ng / mL in the benign group (P = 0.04); and the median CEA concentration in the malignant group was 5.2 (IQR: 2.5-15) ng / mL, significantly higher than the 1.5 (IQR: 1.1-2.0) ng / mL in the benign group (P<0.001).

[0058] Subsequently, a binary logistic regression model was constructed with pathological results (benign = 0, malignant = 1) as the dependent variable and log1p(GDF15), TG, and log1p(CEA) as independent variables. The resulting model equation is:

[0059] After obtaining the model equation, diagnostic efficacy is evaluated. In this embodiment, the pROC package is used for ROC analysis to calculate the AUC values ​​of each individual biomarker and the combined model. The results are as follows: Figure 4 As shown.

[0060] Figure 4In the figure, the horizontal axis represents "1 - Specificity (False Positive Rate)," which has a value range of 0-1. The false positive rate represents the proportion of all real benign nodules that are incorrectly classified as malignant. The larger the value on the horizontal axis, the higher the probability that the model will misclassify benign nodules as malignant. The vertical axis represents "Sensitivity (TruePositive Rate)," or sensitivity, with a value ranging from 0 to 1. The true positive rate represents the proportion of all truly malignant nodules that are correctly identified as malignant. The larger the value on the vertical axis, the stronger the model's ability to correctly identify malignant nodules. In the curves, GDF15: AUC = 0.974 (95% CI: 0.950-0.998), TG: AUC = 0.659 (95% CI: 0.537-0.781), CEA: AUC = 0.853 (95% CI: 0.740-0.965), and the combined model (GDF15+TG+CEA): AUC = 0.987 (95% CI: 0.969-1.000). The results showed that the AUC of the combined model was significantly higher than that of GDF15, TG, and CEA diagnosed individually (DeLong test, P<0.01 for all comparisons with combined model), demonstrating superior comprehensive diagnostic performance.

[0061] When determining the optimal diagnostic threshold, based on the ROC curve of the joint model, the predicted probability threshold when the Youden index is maximized is calculated to be 0.35. Under this threshold, the diagnostic performance of the joint model is: sensitivity = 88.2% (30 / 34), specificity = 85.5% (53 / 62), and accuracy = (30+53) / 96 = 86.5%.

[0062] The final results demonstrate that this invention successfully constructed a logistic regression diagnostic model based on the combined detection of serum GDF15, TG, and CEA. This model exhibited high diagnostic efficacy (AUC=0.987) in differentiating between benign and malignant thyroid nodules, significantly outperforming single-marker detection. At the optimal diagnostic threshold, the model demonstrated good sensitivity (88.2%) and specificity (85.5%). This indicates that the method proposed in this invention is a promising, accurate, and non-invasive auxiliary diagnostic tool for benign and malignant thyroid nodules, effectively addressing the shortcomings of existing technologies and providing crucial decision support for clinical practice.

[0063] This invention integrates information from the novel biomarker GDF15 with traditional biomarkers TG and CEA, and optimizes this combination using a logistic regression model. Compared to single biomarker detection, this method more effectively captures the biological differences between benign and malignant nodules, thus significantly improving the overall efficacy of differential diagnosis. Furthermore, this invention only requires peripheral blood sampling, avoiding the pain, complication risks, and resource requirements of invasive examinations such as femoral head irradiation (FNA), making it patient-friendly, easy to accept, and repeatable. The test results provided by this invention can offer clinicians a more reliable basis for risk assessment. For nodules predicted as low-risk by the model, it may reduce unnecessary FNA, save medical costs, and alleviate patient anxiety; for high-risk results, it can suggest prioritizing FNA or surgical intervention, contributing to the early detection and treatment of malignant tumors. The biomarker detection technologies involved in this invention (such as ELISA and CLIA) are mature clinical laboratory techniques, and the logistic regression model is a classic and easily implemented statistical model, easy to standardize and widely apply.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for differentiating benign and malignant thyroid nodules based on biomarkers, characterized in that, include: S1: Collect serum samples; S2: Quantitatively detect multiple biomarkers in the serum sample to obtain concentration data of multiple biomarkers; S3: Input the multi-marker concentration data into the pre-constructed probability calculation model to obtain the predicted probability value; S4: Based on the preset diagnostic threshold and the predicted probability value, the benign or malignant thyroid nodules are determined, and the results of the benign or malignant thyroid nodule identification are obtained.

2. The method for differentiating benign and malignant thyroid nodules based on biomarkers according to claim 1, characterized in that, In step S2, the multi-marker concentration data includes: Concentration values ​​of growth differentiation factor 15, thyroglobulin, and carcinoembryonic antigen.

3. The method for differentiating benign and malignant thyroid nodules based on biomarkers according to claim 2, characterized in that, Step S2 specifically includes: The serum sample was quantitatively analyzed by enzyme-linked immunosorbent assay (ELISA) to obtain the concentration value of growth differentiation factor 15. The serum sample was subjected to combined detection of thyroglobulin and carcinoembryonic antigen using chemiluminescent immunoassay to obtain the concentration values ​​of thyroglobulin and carcinoembryonic antigen.

4. The method for differentiating benign and malignant thyroid nodules based on biomarkers according to claim 1, characterized in that, In step S3, the probability calculation model is a binary logistic regression model.

5. The method for differentiating benign and malignant thyroid nodules based on biomarkers according to claim 1, characterized in that, In step S3, the expression for the predicted probability value is: in, This refers to the predicted probability value output by the probability calculation model. For model constants, The first regression coefficient, The second regression coefficient, The third regression coefficient, For conversion functions, This represents the concentration of growth differentiation factor 15. This represents the concentration of thyroglobulin. This represents the concentration value of carcinoembryonic antigen.

6. The method for differentiating benign and malignant thyroid nodules based on biomarkers according to claim 1, characterized in that, In step S4, the results of the benign or malignant differentiation of the thyroid nodules include: When the predicted probability value is less than or equal to the preset diagnostic threshold, the benign or malignant differentiation result of the thyroid nodule is benign. When the predicted probability value is greater than the preset diagnostic threshold, the thyroid nodule is classified as malignant.

7. The method for differentiating benign and malignant thyroid nodules based on biomarkers according to claim 1, characterized in that, In step S4, the preset diagnostic threshold is the probability value corresponding to the maximum Yoden index of the probability calculation model.

8. A system for differentiating benign and malignant thyroid nodules based on biomarkers, characterized in that, include: Receiving module: Used to receive the concentration data of multiple biomarkers obtained from the quantitative detection of multiple biomarkers in patient serum samples; A calculation module, configured as a pre-built probability calculation model, is used to output predicted probability values ​​based on the multi-marker concentration data; The discrimination module is used to distinguish between benign and malignant thyroid nodules based on a preset diagnostic threshold and the predicted probability value, and to obtain the benign and malignant identification results of thyroid nodules.

9. A device for differentiating benign and malignant thyroid nodules based on biomarkers, characterized in that, include: A memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause a biomarker-based thyroid nodule benign / malignant identification device to perform a biomarker-based thyroid nodule benign / malignant identification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a processor, implement a method for differentiating benign and malignant thyroid nodules based on biomarkers as described in any one of claims 1 to 7.

Images