Thyroid cancer patient recurrent laryngeal nerve invasion evaluation system

Abstract

The invention belongs to the technical field of thyroid cancer preoperative evaluation, and particularly relates to a thyroid cancer patient recurrent laryngeal nerve invasion evaluation system. The system comprises an input module used for inputting ultrasonic detection data and clinical diagnosis data of a patient; the prediction module is used for extracting features from the ultrasonic detection data and the clinical diagnosis data and calculating the features by utilizing a prediction model to obtain a recurrent laryngeal nerve invasion evaluation result; and the output module is used for outputting the recurrent laryngeal nerve invasion evaluation result. According to the method, the ultrasonic diagnosis features and the clinical diagnosis features are optimized, and the prediction model is constructed, so that accurate evaluation of the invasion condition of the recurrent laryngeal nerves of the thyroid cancer patient before the operation is realized, and the method has a good application prospect.

Description

Technical Field

[0001] The present invention belongs to the technical field of preoperative evaluation of thyroid cancer, and in particular relates to a recurrent laryngeal nerve invasion evaluation system for thyroid cancer patients. Background Art

[0002] The global incidence of thyroid cancer is gradually increasing. Although most patients have a good prognosis, some thyroid cancers are biologically aggressive, which is the main factor affecting the mortality rate of thyroid cancer. Extrathyroidal invasion of the thyroid gland is one of the main reasons for the poor prognosis of patients, and the ten-year survival rate of patients can be reduced by 55%. Among them, the recurrent laryngeal nerve is the structure that is most prone to extrathyroidal invasion after the strap muscle.

[0003] The recurrent laryngeal nerve is an important nerve that controls the movement of the vocal cords. Invasion of the recurrent laryngeal nerve can cause damage to the recurrent laryngeal nerve, which can lead to many complications, such as hoarseness, loss of voice, coughing when drinking water, difficulty breathing or suffocation. For some professions that require a good voice, such as singers and teachers, severe recurrent laryngeal nerve damage can even change their original careers. In addition, invasion of the recurrent laryngeal nerve can also lead to a significantly increased tumor recurrence rate. Therefore, in the surgical treatment of thyroid cancer, it is necessary to evaluate the invasion of the recurrent laryngeal nerve and adopt appropriate surgical strategies and auxiliary treatment plans to reduce the risk of postoperative recurrence and preserve the patient's nerve function to the greatest extent.

[0004] For resectable thyroid cancer, the American Thyroid Association (ATA) requires preoperative imaging examinations for tumor staging, because the degree of local infiltration can affect surgical decisions. However, the current preoperative evaluation content and methods are still imperfect. The preoperative evaluation content is limited to thyroid lesions. There is currently no relevant evaluation system for the infiltration of surrounding tissues such as blood vessels and nerves, especially for the invasion of the recurrent laryngeal nerve. In the prior art, the only evaluation of the recurrent laryngeal nerve before thyroid cancer surgery is laryngoscopy, and its principle is to indirectly judge the invasion of the recurrent laryngeal nerve by observing the mobility of the vocal cords. In addition, some literature records the idea of ​​using ultrasound results to evaluate the invasion of the recurrent laryngeal nerve. However, these documents do not specifically record how to use ultrasound results to conclude on the invasion of the recurrent laryngeal nerve.

[0005] Therefore, this field urgently needs a new method to establish a prediction model based on ultrasound features and clinical characteristics to identify high-risk patients with recurrent laryngeal nerve invasion before surgery and help surgeons make preoperative decisions. Summary of the invention

[0006] In view of the problems of the prior art, the present invention provides a recurrent laryngeal nerve invasion assessment system for thyroid cancer patients.

[0007] A system for evaluating recurrent laryngeal nerve invasion in patients with thyroid cancer, comprising:

[0008] An input module, used for inputting ultrasound detection data and clinical diagnosis data of patients;

[0009] A prediction module, used to extract features from the ultrasound detection data and clinical diagnosis data, calculate the features using a prediction model, and obtain a recurrent laryngeal nerve invasion assessment result;

[0010] An output module, used for outputting the recurrent laryngeal nerve invasion assessment result;

[0011] Among them, the characteristics include at least one of the following information: age, body mass index, maximum diameter of the tumor, whether the tumor is close to the lateral capsule of the thyroid gland, whether the tumor is close to the medial capsule of the thyroid gland, whether the tumor is close to the deep capsule of the thyroid gland, the distance between the tumor and the recurrent laryngeal nerve, and whether the typical nerve structure of the recurrent laryngeal nerve in the two-dimensional longitudinal section of ultrasound exists.

[0012] Preferably, the prediction model assigns points according to the features, and adds up the scores of all the features to obtain a final total score, and the recurrent laryngeal nerve invasion assessment result is obtained based on the total score.

[0013] Preferably, the prediction model is selected from model 1, and the features and scoring methods included in model 1 are:

[0014] 1) Age: The scoring method is as follows: <30 years old, 0 points; 30-40 years old, excluding 40 years old, 17 points; 40-50 years old, excluding 50 years old, 45 points; 50-60 years old, excluding 60 years old, 29 points; ≥60 years old, 58 points;

[0015] 2) BMI, which is scored as follows: <18.5, 22 points; 18.5-24, excluding 24, 0 points; ≥24, 29 points;

[0016] 3) Maximum diameter of the tumor, which was scored as follows: ≤10 mm, 0 points; 10-20 mm, excluding 10 mm, 23 points; 20-40 mm, excluding 20 mm, 42 points; >40 mm, 97 points;

[0017] 4) The distance between the tumor and the recurrent laryngeal nerve was scored as follows: ≥1 mm, 0 points; <1 mm, 100 points;

[0018] 5) Whether the tumor is closely attached to the lateral capsule of the thyroid gland, the scoring method is: yes, 0 points; no, 52 points;

[0019] 6) Whether the tumor is closely attached to the inner capsule of the thyroid gland, the scoring method is: yes, 43 points; no, 0 points;

[0020] 7) Whether the tumor is closely attached to the deep capsule of the thyroid gland, the scoring method is: yes, 92 points; no, 0 points.

[0021] Preferably, the total score is ≥ 260 points, indicating a high risk of recurrent laryngeal nerve invasion.

[0022] Preferably, the model 1 is suitable for patients whose ultrasound detection results cannot clearly show the two-dimensional longitudinal section structure of the recurrent laryngeal nerve.

[0023] Preferably, the prediction model is selected from model 2, and the features and scoring methods included in model 2 are:

[0024] 1) The distance between the tumor and the recurrent laryngeal nerve was scored as follows: ≥1 mm, 0 points; <1 mm, 59 points;

[0025] 2) Whether the tumor is closely attached to the deep capsule of the thyroid gland, the scoring method is: yes, 66 points; no, 0 points;

[0026] 3) Whether the typical nerve structure of the recurrent laryngeal nerve in the two-dimensional longitudinal section of ultrasound is present or not, the scoring method is: yes, 0 points; no, 100 points.

[0027] Preferably, the total score is ≥135 points, indicating a high risk of recurrent laryngeal nerve invasion.

[0028] Preferably, the model 2 is suitable for patients whose ultrasound detection results can clearly show the two-dimensional longitudinal section structure of the recurrent laryngeal nerve.

[0029] Preferably, the prediction model is established by the statistical methods of logistics regression and nomogram.

[0030] The present invention also provides a computer-readable storage medium, on which is stored: a computer program for implementing the above-mentioned recurrent laryngeal nerve invasion assessment system for thyroid cancer patients.

[0031] The present invention selects ultrasound diagnostic features and clinical diagnostic features suitable for evaluating recurrent laryngeal nerve invasion, and constructs a prediction model for evaluating the recurrent laryngeal nerve invasion using these features.

[0032] The present invention provides two prediction models according to whether the ultrasonic diagnostic features can clearly display the two-dimensional longitudinal section structure of the recurrent laryngeal nerve. When the ultrasonic detection results cannot clearly display the two-dimensional longitudinal section structure of the recurrent laryngeal nerve, using model 1 can effectively improve the sensitivity of diagnosis; when the ultrasonic detection results can clearly display the two-dimensional longitudinal section structure of the recurrent laryngeal nerve, using model 2 can effectively improve the specificity and accuracy of diagnosis.

[0033] Based on the above two prediction models, the present invention can evaluate the recurrent laryngeal nerve invasion of thyroid cancer patients before surgery in an intuitive, accurate and non-invasive manner through ultrasound visualization, predict the degree of recurrent laryngeal nerve preservation, and assist doctors in formulating surgical treatment methods, and has good application prospects.

[0034] Obviously, according to the above contents of the present invention, in accordance with common technical knowledge and customary means in the art, without departing from the above basic technical ideas of the present invention, other various forms of modification, replacement or change may be made.

[0035] The above contents of the present invention are further described in detail below through specific implementation methods in the form of embodiments. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following examples. All technologies realized based on the above contents of the present invention belong to the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a schematic diagram of the system flow of Example 1;

[0037] Figure 2 This is a schematic diagram of the system flow of Example 2;

[0038] Figure 3 These are ultrasound images related to the recurrent laryngeal nerve, among which A is a short-axis schematic diagram of the recurrent laryngeal nerve; B is a long-axis schematic diagram of the recurrent laryngeal nerve; C is a schematic diagram of the short-axis invasion of the recurrent laryngeal nerve; D is a schematic diagram of the long-axis invasion of the recurrent laryngeal nerve; E is a short-axis section in which the tumor is close to the recurrent laryngeal nerve, but the nerve is not involved; F is a long-axis section in which the tumor is close to the recurrent laryngeal nerve, but the nerve is not involved; in the figure, the hexagon represents the trachea, the triangle represents the esophagus, the quadrilateral represents the cricoid cartilage, the arrow represents the recurrent laryngeal nerve, and MASS represents thyroid tumor. DETAILED DESCRIPTION

[0039] It should be noted that the algorithms of data collection, transmission, storage and processing steps not specifically described in the embodiments, as well as the hardware structure, circuit connection, etc. not specifically described can all be implemented through the contents disclosed in the prior art.

[0040] Example 1: Assessment system for recurrent laryngeal nerve invasion in thyroid cancer patients

[0041] The system flow chart of this embodiment is as follows: Figure 1 As shown, specifically including:

[0042] An input module, used for inputting ultrasound detection data and clinical diagnosis data of patients;

[0043] A prediction module, used to extract features from the ultrasound detection data and clinical diagnosis data, calculate the features using a prediction model, and obtain a recurrent laryngeal nerve invasion assessment result;

[0044] An output module is used to output the recurrent laryngeal nerve invasion assessment result.

[0045] The prediction model assigns points according to the features, and adds up the scores of all the features to obtain a final total score, and the recurrent laryngeal nerve invasion assessment result is obtained based on the total score.

[0046] The features and scoring methods of the prediction model (model 1) are as follows:

[0047] 1) Age: The scoring method is as follows: <30 years old, 0 points; 30-40 years old, excluding 40 years old, 17 points; 40-50 years old, excluding 50 years old, 45 points; 50-60 years old, excluding 60 years old, 29 points; ≥60 years old, 58 points;

[0048] 2) BMI, which is scored as follows: <18.5, 22 points; 18.5-24, excluding 24, 0 points; ≥24, 29 points;

[0049] 3) Maximum diameter of the tumor, which was scored as follows: ≤10 mm, 0 points; 10-20 mm, excluding 10 mm, 23 points; 20-40 mm, excluding 20 mm, 42 points; >40 mm, 97 points;

[0050] 4) The distance between the tumor and the recurrent laryngeal nerve was scored as follows: ≥1 mm, 0 points; <1 mm, 100 points;

[0051] 5) Whether the tumor is closely attached to the lateral capsule of the thyroid gland, the scoring method is: yes, 0 points; no, 52 points;

[0052] 6) Whether the tumor is closely attached to the inner capsule of the thyroid gland, the scoring method is: yes, 43 points; no, 0 points;

[0053] 7) Whether the tumor is closely attached to the deep capsule of the thyroid gland, the scoring method is: yes, 92 points; no, 0 points.

[0054] If the total score is ≥260 points, it is judged that there is a high risk of recurrent laryngeal nerve invasion.

[0055] by Figure 3 For example, Figure 3 A and Figure 3 B shows the short-axis and long-axis diagrams of the normal recurrent laryngeal nerve structure. Figure 3 C and Figure 3 The maximum diameter of the thyroid tumor in D is about 28 mm. In the short-axis section of ultrasound, the tumor is seen to be closely attached to the inner capsule of the thyroid gland ( Figure 3 C), in the long axis section of ultrasound, close to the deep surface of the thyroid capsule ( Figure 3D), the distance between the tumor and the recurrent laryngeal nerve was less than 1 mm, and the neural structure of the recurrent laryngeal nerve was unclear. Combined with the existing ultrasound features, the total model score was >260 points, so it was judged that this case was at high risk of recurrent laryngeal nerve invasion. The final intraoperative results confirmed that the tumor invaded the recurrent laryngeal nerve.

[0056] Model 1 is suitable for patients whose ultrasound examination results cannot clearly show the two-dimensional longitudinal section structure of the recurrent laryngeal nerve. For such patients, the use of Model 1 for judgment can effectively improve the sensitivity of diagnosis.

[0057] In order to verify the diagnostic performance of Model 1 of this embodiment, 104 patients (19 of whom had recurrent laryngeal nerve invasion) were collected and the recurrent laryngeal nerve invasion assessment system for thyroid cancer patients was used to assess the recurrent laryngeal nerve invasion. The system had a diagnostic sensitivity of 94.74%, a specificity of 74.12%, a positive predictive value of 45%, a negative predictive value of 98.43%, and a diagnostic accuracy of 76.42%. It can be seen that Model 1 of this embodiment can sensitively identify patients at high risk of recurrent laryngeal nerve invasion.

[0058] Example 2: Assessment system for recurrent laryngeal nerve invasion in thyroid cancer patients

[0059] The system flow chart of this embodiment is as follows: Figure 2 As shown, specifically including:

[0060] An input module, used for inputting ultrasound detection data of the patient;

[0061] A prediction module, used to extract features from the ultrasound detection data, calculate the features using a prediction model, and obtain a recurrent laryngeal nerve invasion assessment result;

[0062] An output module is used to output the recurrent laryngeal nerve invasion assessment result.

[0063] The prediction model assigns points according to the features, and adds up the scores of all the features to obtain a final total score, and the recurrent laryngeal nerve invasion assessment result is obtained based on the total score.

[0064] The features and scoring methods of the prediction model (model 2) are as follows:

[0065] 1) The distance between the tumor and the recurrent laryngeal nerve was scored as follows: ≥1 mm, 0 points; <1 mm, 59 points;

[0066] 2) Whether the tumor is closely attached to the deep capsule of the thyroid gland, the scoring method is: yes, 66 points; no, 0 points;

[0067] 3) Whether the typical nerve structure of the recurrent laryngeal nerve in the two-dimensional longitudinal section of ultrasound is present or not, the scoring method is: yes, 0 points; no, 100 points.

[0068] If the total score is ≥135 points, it is judged that there is a high risk of recurrent laryngeal nerve invasion.

[0069] by Figure 3 E and Figure 3 For example, the maximum diameter of the thyroid tumor is about 12 mm, and it is close to the inner capsule of the thyroid gland in the cross section ( Figure 3 E), in longitudinal section, close to the deep surface of the thyroid capsule ( Figure 3 F), the distance between the tumor and the recurrent laryngeal nerve was less than 1 mm, but the structure of the recurrent laryngeal nerve was clear and complete. Combined with the ultrasound characteristics, the total model score was less than 135 points. Therefore, it was judged that the possibility of recurrent laryngeal nerve invasion in this case was low. The final intraoperative results confirmed that the tumor did not invade the recurrent laryngeal nerve.

[0070] Model 2 is suitable for patients whose ultrasound examination results can clearly show the two-dimensional longitudinal section structure of the recurrent laryngeal nerve. For such patients, the use of Model 2 for judgment can effectively improve the specificity of diagnosis.

[0071] In order to verify the diagnostic performance of Model 2 of this embodiment, 104 patients (19 of whom had recurrent laryngeal nerve invasion) were collected and the recurrent laryngeal nerve invasion assessment system for thyroid cancer patients was used to assess the recurrent laryngeal nerve invasion. The system diagnostic sensitivity was 57.89%, the specificity was 95.29%, the positive predictive value was 73.26%, the negative predictive value was 91.03%, and the diagnostic accuracy was 88.46%. It can be seen that Model 2 of this embodiment can reduce false positives and increase the specificity and accuracy of the diagnosis of recurrent laryngeal nerve invasion.

[0072] It can be seen from the above embodiments that the present invention achieves accurate assessment of recurrent laryngeal nerve invasion in patients with thyroid cancer before surgery by optimizing ultrasound diagnostic features and clinical diagnostic features and constructing a prediction model, and has a good application prospect.

Claims

1. A system for assessing recurrent laryngeal nerve invasion in patients with thyroid cancer, characterized in that: include: An input module, used for inputting ultrasound detection data and clinical diagnosis data of patients; A prediction module, used to extract features from the ultrasound detection data and clinical diagnosis data, calculate the features using a prediction model, and obtain a recurrent laryngeal nerve invasion assessment result; An output module, used for outputting the recurrent laryngeal nerve invasion assessment result; Among them, the characteristics include at least one of the following information: age, body mass index, maximum diameter of the tumor, whether the tumor is close to the lateral capsule of the thyroid gland, whether the tumor is close to the medial capsule of the thyroid gland, whether the tumor is close to the deep capsule of the thyroid gland, the distance between the tumor and the recurrent laryngeal nerve, and whether the typical nerve structure of the recurrent laryngeal nerve in the two-dimensional longitudinal section of ultrasound exists.

2. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 1, characterized in that: The prediction model assigns points according to the features, and adds up the scores of all the features to obtain a final total score, and the recurrent laryngeal nerve invasion assessment result is obtained based on the total score.

3. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 2, characterized in that: The prediction model is selected from model 1, and the features and scoring methods of model 1 are as follows: 1) Age: The scoring method is as follows: <30 years old, 0 points; 30-40 years old, excluding 40 years old, 17 points; 40-50 years old, excluding 50 years old, 45 points; 50-60 years old, excluding 60 years old, 29 points; ≥60 years old, 58 points; 2) BMI, which is scored as follows: <18.5, 22 points; 18.5-24, excluding 24, 0 points; ≥24, 29 points; 3) Maximum diameter of the tumor, which was scored as follows: ≤10 mm, 0 points; 10-20 mm, excluding 10 mm, 23 points; 20-40 mm, excluding 20 mm, 42 points; >40 mm, 97 points; 4) The distance between the tumor and the recurrent laryngeal nerve was scored as follows: ≥1 mm, 0 points; <1 mm, 100 points; 5) Whether the tumor is closely attached to the lateral capsule of the thyroid gland, the scoring method is: yes, 0 points; no, 52 points; 6) Whether the tumor is closely attached to the inner capsule of the thyroid gland, the scoring method is: yes, 43 points; no, 0 points; 7) Whether the tumor is closely attached to the deep capsule of the thyroid gland, the scoring method is: yes, 92 points; No, 0 points.

4. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 3, characterized in that: If the total score is ≥260 points, it is judged that there is a high risk of recurrent laryngeal nerve invasion.

5. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 3 or 4, characterized in that: The model 1 is suitable for patients whose ultrasound detection results cannot clearly show the two-dimensional longitudinal section structure of the recurrent laryngeal nerve.

6. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 2, characterized in that: The prediction model is selected from model 2, and the features and scoring methods included in model 2 are: 1) The distance between the tumor and the recurrent laryngeal nerve was scored as follows: ≥1 mm, 0 points; <1 mm, 59 points; 2) Whether the tumor is closely attached to the deep capsule of the thyroid gland, the scoring method is: yes, 66 points; no, 0 points; 3) Whether the typical nerve structure of the recurrent laryngeal nerve in the two-dimensional longitudinal section of ultrasound is present or not, the scoring method is: yes, 0 points; no, 100 points.

7. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 6, characterized in that: If the total score is ≥135 points, it is judged that there is a high risk of recurrent laryngeal nerve invasion.

8. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 6 or 7, characterized in that: Model 2 is suitable for patients whose ultrasound test results can clearly show the two-dimensional longitudinal section structure of the recurrent laryngeal nerve.

9. The recurrent laryngeal nerve invasion assessment system for thyroid cancer patients according to claim 1, characterized in that: The prediction model is established by the statistical methods of logistics regression and nomogram.

10. A computer-readable storage medium, characterized in that: Stored thereon is: a computer program for implementing the recurrent laryngeal nerve invasion assessment system for thyroid cancer patients as described in any one of claims 1-9.

Images