Tumor endoscope resection difficulty prediction system based on detection data

Abstract

The invention provides a tumor endoscopic resection difficulty prediction system based on detection data, and belongs to the field of tumor detection. The problem of inaccurate prediction of the tumor excision difficulty is solved; specifically, the medical record acquisition module acquires tumor reference data; the medical record analysis module is used for calculating standard treatment time and a treatment time equation of each tumor; the difficulty analysis module is used for calculating a judgment equation of each tumor; obtaining a medical examination report of the patient, determining the treatment difficulty of the tumor type of the patient, and obtaining the medical examination report; the doctor interaction module is used for feeding back the medical examination report to a doctor; continuously acquiring a medical examination report of the patient, and generating the medical examination report; by acquiring, processing and analyzing the related data of the tumor patient, the tumor type, the treatment difficulty and the treatment sequence of the patient are determined, and a doctor can conveniently perform a tumor endoscopic resection operation.

Description

Technical Field

[0001] The present invention discloses a tumor endoscopic resection difficulty prediction system based on detection data, and relates to the technical field of tumor analysis. Background Art

[0002] The existing systems for predicting the difficulty of tumor resection have the following shortcomings: Data dependency: The performance of existing systems is highly dependent on the accuracy and completeness of input data. If there are errors or missing data, the prediction results will be inaccurate. At the same time, a large amount of redundant data will also reduce the computing power of the system.

[0003] High technical threshold: The tumor endoscopic resection difficulty prediction system involves complex data processing and analysis technologies, and requires a professional technical team to develop and maintain it. This places high demands on the hospital's technical strength and human resources, limiting the popularization and application of the system.

[0004] Limited scope of application: Most existing systems use machine learning to process data on the difficulty of endoscopic tumor resection. This method requires a large amount of image data or physiological data for large-scale data fitting and model training. In addition, the existing system cannot accurately identify certain tumors with similar symptoms, resulting in a narrow scope of application. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a tumor endoscopic resection difficulty prediction system based on detection data, aiming to solve the problem of inaccurate prediction of tumor resection difficulty.

[0006] In order to achieve the above object, the present invention is implemented through the following technical solutions: A tumor endoscopic resection difficulty prediction system based on detection data includes: Medical record acquisition module: used to obtain all tumor medical record information in the target hospital as reference medical records; extract the tumor type, age, gender, treatment time and lesion index content of all tumor patients in the reference medical records to obtain tumor reference data; The content of pathological indicators includes: tumor marker content, immune cell content, non-immune cell content and liver and kidney secretion content; Medical record analysis module: used to analyze tumor reference data based on treatment time and calculate the standard treatment time for each tumor; The influence coefficients of tumor patients' age, gender and lesion index content on treatment time were calculated, and the tumor reference data were secondary analyzed to construct the treatment time equation for each tumor. Difficulty analysis module: used to perform Logistic regression analysis on tumor reference data according to the lesion index content of different tumor patients, calculate the incidence weight of each tumor, and obtain the tumor judgment equation; Obtain the patient's medical examination report, and determine the treatment difficulty of the patient's tumor type by combining the tumor judgment equation and treatment time equation for each tumor, and obtain a medical examination report; Physician interaction module: used to feed back medical examination reports to physicians; continuously obtain patients' medical examination reports and generate medical examination reports.

[0007] Furthermore, the workflow of the medical record analysis module is as follows: Process A1: Count the number of reference medical records, recorded as bl; Count the number of all tumor types in the reference medical records, recorded as zl; Count the number of patients with the first type of tumor in the reference medical records, denoted as bl （1） ; The number of patients with the second tumor, denoted as bl （2） ; Similarly, the number of patients with the zlth tumor is denoted as bl （zl） ; Among them, bl （1） +bl （2） +……+bl （zl） =bl; Calculate the incidence of the first tumor (among tumor patients), denoted as ir （1） ;ir （1） The calculation formula is: （1） =bl （1） / bl; The incidence of the second tumor is denoted as ir （2） ;ir （2） The calculation formula is: （2） =bl （2） / bl; By analogy, the incidence of the zlth tumor is denoted as ir （zl） ;ir （zl） The calculation formula is: （zl） =bl （zl） / bl; Process A2: Group the reference medical records by tumor type to obtain group t 1 , Group t 2 ~ Group t zl ; Among them, group t 1 represents all reference medical records of patients with the first type of tumor; group t 2 represents all the reference medical records of patients with the second type of tumor; and so on, group t zl Represents all reference medical records of patients with the zlth type of tumor; Group t 1 ~ Group t zl In the above table, the treatment time of different tumor patients is recorded as to(1,1), to(1,2)~to(1,bl（1） )~to(zl,bl （zl） ); Among them, to (1,1) represents group t 1 The treatment time of the first tumor patient in the group t 1 The treatment time of the second tumor patient in the test; and so on, to (1, bl （1） ) represents group t 1 bl （1） Treatment time of each tumor patient; to (zl, bl （zl） ) represents group t zl bl （zl） Treatment time for each cancer patient; Process A3: According to to (1,1) ~ to (zl, bl （zl） ), calculate the standard treatment time for each tumor in turn, and get zt （1） 、zt （2） ~zt （zl） ; Among them, zt （1） represents the standard treatment time for the first type of tumor; zt （2） represents the standard treatment time for the second type of tumor; and so on, zt （zl） represents the standard treatment time for the zlth tumor; Process A4: According to group t 1 ~ Group t zl The age, gender and lesion index content of different tumor patients are calculated, the influence coefficient of the age, gender and lesion index content of tumor patients on the treatment time is calculated, the treatment time equation for each tumor is constructed, and the difficulty analysis module is entered.

[0008] Furthermore, the specific process of process A3 is as follows: Process A31: According to to (1,1), to (1,2) ~ to (1,bl （1） ) Calculate the standard treatment time for the first type of tumor and get zt （1） ; Calculate to(1,1)~to(1,bl （1） ), denoted as ato(1,-); Process A32: Extract to (1, 1) ~ to (1, bl （1） ) is denoted as mto(1,-), and the minimum value is denoted as lto(1,-); Taking [0,1] as the error range, extract to (1,1) ~ to (1,bl （1） ) is denoted as zto(1,-); Process A33: Take [0,1] as the error range, between to(1,1)~to(1,bl （1） ) is used to count the number of tumor patients whose treatment time is equal to zto (1, -), denoted as tp; In to (1,1) ~ to (1,bl （1） ), the number of tumor patients whose treatment time is between [lto (1, -), ato (1, -)) is counted, recorded as tp1; the number of tumor patients whose treatment time is between [ato (1, -), mto (1, -)] is counted, recorded as tp2; Process A34: Comparison (tp / bl （1） ) and 0.5, determine zt （1） The value of If (tp / bl （1） )≥0.5, then zt （1） The value is: zto (1, -); If (tp / bl （1） )<0.5, then zt （1） The calculation formula is: ; Process A35: Repeat the same process of calculating the standard treatment time for the first tumor, and calculate the standard treatment time for the second to zl tumors to obtain zt （2） ~zt （zl） .

[0009] Furthermore, the specific process of process A4 is as follows: Process A41: Calculate group t 1 The effect coefficients of age, gender and lesion index content of tumor patients on treatment time were calculated to construct the treatment time equation for the first type of tumor. Group t 1 The age of the first patient is recorded as y (1,1), the tumor marker content is recorded as tu (1,1), the immune cell content is recorded as im (1,1), the non-immune cell content is recorded as ui (1,1), and the liver and kidney secretion content is recorded as lk (1,1); The age of the second patient is recorded as y (1, 2), the tumor marker content is recorded as tu (1, 2), the immune cell content is recorded as im (1, 2), the non-immune cell content is recorded as ui (1, 2), and the liver and kidney secretion content is recorded as lk (1, 2); By analogy, the bl （1） The age of each patient is denoted as y(1,bl （1） ), the tumor marker content is recorded as tu (1,bl （1） ), the immune cell content is recorded as im (1,bl （1） ), the non-immune cell content is recorded as ui (1,bl（1） ), the liver and kidney secretion content is recorded as lk (1,bl （1） ); Process A42: Obtain the standard treatment time zt for the first tumor （1） ; Calculation group t 1 The influence coefficient of the age of the first patient on the treatment time is denoted as α(1,1). The calculation formula of α(1,1) is: α(1,1) = to(1,1) / zt （1） ; The influence coefficient of the second patient's age on the treatment time is denoted as α(1,2). The calculation formula of α(1,2) is: α(1,2) = to(1,2) / zt （1） ; By analogy, the bl （1） The influence coefficient of the patient's age on the treatment time is denoted as α (1,bl （1） ), α(1,bl （1） ) is calculated as: α(1,bl （1） )=to(1,bl （1） ) / zt （1） ; Calculate α(1,1)~α(1,bl （1） ) is recorded as α(1,-); α(1,-) is taken as the age influence coefficient of the first type of tumor; Process A43: Statistical group t 1 The number of male patients in the test is recorded as bem, and the number of female patients is recorded as bew; where bem + bew = bl （1） ; Calculate the influence coefficient of male patients on treatment time, recorded as β(1, m), the calculation formula of β(1, m) is: β(1, m) = (bem / bl （1） )×zt （1） ; Calculate the effect coefficient of female patients on treatment time β(1, w), the calculation formula of β(1, w) is: β(1, w) = (bew / bl （1） )×zt （1） .

[0010] Furthermore, the subsequent process of the process A43 is as follows: Process A44: Calculate group t 1 The influence coefficient of the lesion index content of the first patient on the treatment time is denoted as λ(1,1); If the first patient is male, the calculation formula for λ(1,1) is: ; If the first patient is a female, the calculation formula for λ(1,1) is: ; Calculation group t 1 The influence coefficient of the lesion index content of the second patient on the treatment time is denoted as λ(1,2); If the second patient is male, the calculation formula for λ(1,2) is: ; If the second patient is a female, the calculation formula for λ(1,2) is: ; Similarly, calculate group t 1 bl （1） The influence coefficient of the lesion index content of each patient on the treatment time is denoted as λ (1,bl （1） ); If the first patient is male, then λ(1,bl （1） ) is calculated as: ; If the first patient is a female, then λ(1,bl （1） ) is calculated as: ; Process A45: Calculate λ(1,1)~λ(1,bl （1） ) is recorded as λ(1,-), and λ(1,-) is used as the influence coefficient of the pathological index of the first tumor; Summarize the data in process A41 to process A45, construct the treatment time equation for the first tumor, and obtain: Patients are male: Ttm(1) = α(1, -) × xn(1, -) + β(1, m) + λ(1, -) × xz(1, -); Patients are female: Ttw(1) = α(1, -) × xn(1, -) + β(1, w) + λ(1, -) × xz(1, -); Wherein, Ttm(1) represents the treatment time of male patients with the first tumor; Ttw(1) represents the treatment time of female patients with the first tumor; xn(1, -) represents the age of patients with the first tumor; xz(1, -) represents the geometric mean of the lesion index of patients with the first tumor; Process A46: Repeat the same process of constructing the first tumor treatment time equation, according to the group t 2 ~ Group t zl Construct the treatment time equation for the 2nd to zlth tumors.

[0011] Furthermore, the workflow of the difficulty analysis module is as follows: Process B1: Extract the incidence of tumors from type 1 to type zl to obtain ir （1） ~ir （zl） ; Get Group 1 The lesion index of all patients: Tumor marker content: tu (1,1) ~ tu (1,bl （1） ); Immune cell content: im(1,1)~im(1,bl （1） ); Non-immune cell content: ui (1,1) ~ ui (1,bl （1） ); Liver and kidney secretion content: lk (1,1) ~ lk (1,bl （1） ); Process B2: Get group t 2 The lesion index of all patients: Tumor marker levels are recorded as: tu (2,1) ~ tu (2,bl （2） ); The immune cell content is recorded as: im(2,1)~im(2,bl （2） ); The non-immune cell content is recorded as: ui (2,1) ~ ui (2,bl （2） ); The content of liver and kidney secretion is recorded as: lk (2,1) ~ lk (2,bl （2） ); And so on, get group t zl The lesion index of all patients: Tumor marker content: tu (zl, 1) ~ tu (zl, bl （zl） ); Immune cell content: im(zl,1)~im(zl,bl （zl） ); Non-immune cell content: ui (zl, 1) ~ ui (zl, bl （zl） ); Liver and kidney secretion content: lk (zl, 1) ~ lk (zl, bl （zl） ); Process B3: Obtain the normal tumor marker content range in the human body, recorded as [tuj, tuk]; The normal range of immune cell content is denoted as [imj, imk]; The content range of normal non-immune cells is recorded as [uij, uik]; The content range of normal liver and kidney secretions is recorded as [lkj, lkk]; Summarize the data in process B1 to process B3, calculate the lesion weight of the first tumor by using logistic regression, and construct the judgment equation of the first tumor; Process B4: Repeat the same process of constructing the judgment equation for the first tumor to construct the judgment equations for the second to zlth tumors; Obtain the patient's medical examination report, extract the patient's lesion index, and substitute it into the judgment equation of the first to zlth tumors to determine the patient's tumor type as the diseased tumor; According to the patient's medical examination report, extract the patient's age and gender, and extract the treatment time equation of the corresponding tumor, and calculate the patient's estimated treatment time, recorded as ht; According to the diseased tumor, the standard treatment time of the corresponding tumor is extracted and recorded as zht; zht∈{zt （1） 、zt （2） ~zt （zl）}; Taking [0, 1] as the error range, determine whether ht and zht are equal; If ht = zht, it means that the patient's tumor resection is moderately difficult; If ht<zht, it means that the patient's tumor is easier to remove; If ht>zht, it means that the patient's tumor is difficult to remove.

[0012] Furthermore, the specific process of process B3 is as follows: Process B31: Let the content of tumor markers in the blood of tumor patients be Xtu, the content of immune cells be Xim, the content of non-immune cells be Xui, and the content of liver and kidney secretions be Xlk; Assume that the mathematical expression of the judgment equation of the first type of tumor is equation Inr (1), and the mathematical expression of equation Inr (1) is: Inr(1)=1 / [1+e -（z（1）） ]; Among them, Inr(1) represents the probability that the patient has the first tumor; z(1) represents the linear characteristic sum of the patient suffering from the first tumor; the mathematical expression of z(1) is: z(1)=wt （1） ×Xtu+wi （1） ×Xim+wu （1） ×Xui+wl （1） ×Xlk+b(1) Among them, b(1) represents the bias that the patient has the first tumor, and the initial value of b(1) is 0; wt （1）, wi （1） 、wu （1） and wl （1） , respectively represent the tumor marker lesion weight, immune cell lesion weight, non-immune cell lesion weight and liver and kidney secretion lesion weight of the patient suffering from the first tumor; wt （1） , wi （1） 、wu （1） and wl （1） The initial value of is 1; Process B32: Let the value of Inr(1) be ir （1） , tu(1,1)~tu(1,bl （1） )、im(1,1)~im(1,bl （1） )、ui(1,1)~ui(1,bl （1） ) and lk(1,1)~lk(1,bl （1） ), as Xtu, Xim, Xui and Xlk respectively and substituted into z(1), we get equation group B-1: ; Use MATLAB software to calculate process group B-1 and get wt （1） , wi （1） 、wu （1） wl （1） and a first estimate of b(1), denoted by wt1 （1） ,wi1 （1） 、wu1 （1） wl1 （1） and b1(1); Substituting into the mathematical expression of z(1), we get the mathematical expression z(1)-1; z(1)-1=wt1 （1） ×Xtu+wi1 （1） ×Xim+wu1 （1） ×Xui+wl1 （1） ×Xlk+b1(1).

[0013] Furthermore, the subsequent process of the process B32 is as follows: Process B33: tu(1,1)~tu(1,bl （1） )、im(1,1)~im(1,bl （1） )、ui(1,1)~ui(1,bl （1） ) and lk(1,1)~lk(1,bl （1） ), as Xtu, Xim, Xui and Xlk respectively and substitute into z(1)-1, recalculate the group t 1The estimated incidence rate of all patients in the study is: Group 1 The estimated incidence rate of the first patient in the study is denoted as irr(1). The calculation formula for irr(1) is: ; Among them, zz(1) represents (group t 1 The new linear feature sum of the first patient is calculated as: zz(1)=wt1 （1） ×tu(1,1)+wi1 （1） ×im(1,1)+wu1 （1） ×ui(1,1)+wl1 （1） ×lk(1,1)+b1(1); Group 1 The estimated incidence rate of the second patient is recorded as irr(2). The calculation formula of irr(2) is: ; Among them, zz(2) represents (group t 1 The new linear feature sum of the second patient, zz(2) is calculated as: zz(2)=wt1 （1） ×tu(1,2)+wi1 （1） ×im(1,2)+wu1 （1） ×ui(1,2)+wl1 （1） ×lk(1,2)+b1(1); Similarly, group t 1 bl （1） The estimated incidence rate per patient is denoted as irr (bl （1） );irr(bl （1） ) is calculated as: ; Among them, zz (bl （1） ) means (group t 1 Middle) No. bl （1） The new linear features and zz(bl （1） ) is calculated as: zz(bl （1） )=wt1 （1） ×tu(1,bl （1） )+wi1 （1） ×im(1,bl （1） )+wu1 （1） ×ui(1,bl （1） )+wl1 （1） ×lk(1,bl （1）)+b1(1); Process B34: Calculate the logarithmic loss of the first tumor judgment equation, denoted as LL （1） LL （1） The calculation formula is as follows: ; Where irr(i) represents group t 1 The estimated incidence rate of the i-th patient in the study, the value range of i is: 1~bl （1） ; Calculate wt1 （1） The gradient of is denoted as δ(L-tu); the calculation formula of δ(L-tu) is: ; Calculate wi1 （1） The gradient of is denoted as δ(L-im); the calculation formula of δ(L-im) is: ; Calculate wu1 （1） The gradient of is denoted as δ(L-ui); the calculation formula of δ(L-ui) is: ; Calculate wl1 （1） The gradient of is denoted as δ(L-lk); the calculation formula of δ(L-lk) is: ; Calculate the gradient of b(1), denoted as δ(Lb); the calculation formula of δ(Lb) is: ; Where irr(j) represents the group t 1 The estimated incidence rate of the jth patient in the study, the value range of j is: 1~bl （1） ; tu(1,j),im(1,j),ui(1,j) and lk(1,j), respectively, represent the 1 The levels of tumor markers, immune cells, non-immune cells, and liver and kidney secretions in the blood of the jth patient.

[0014] Furthermore, the subsequent process of process B34 is as follows: Process B35: Assume that the learning rate is ε, the initial value of ε is: 0.001, and the value range of ε is: [0.001, 0.1]; Let wt1 （1） After the change, the weight is wt2 （1） , wt2 （1） The calculation formula is: wt2 （1） =wt1（1） ±(ε×δ(L-tu)); Set wi1 （1） After the change, the authority is wi2 （1） ,wi2 （1） The calculation formula is: wi2 （1） =wi1 （1） ±(ε×δ(L-im)); Set wu1 （1） After the change, the authority is wu2 （1） , wu2 （1） The calculation formula is: wu2 （1） =wu1 （1） ±(ε×δ(L-ui)); Assume wl1 （1） After the change, the authority is wl2 （1） ,wl2 （1） The calculation formula is: wl2 （1） =wl1 （1） ±(ε×δ(L-lk)); Assume that the weight of b1(1) after change is b2(1), and the calculation formula of b2(1) is: b2(1)=b1(1)±(ε×δ(Lb)); Process B36: wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1), and then substitute them into the mathematical expression of z(1) in reverse order, and re-execute process B32 to process B34 to obtain the new logarithmic loss of the first tumor judgment equation, denoted as LL （1） 、 ; Compare LL （1） 、 With LL （1） size; If LL （1） 、 <LL （1） , then wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1) as the weights and biases of the first tumor judgment equation, and wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） Substituting b2(1) into the mathematical expression of z(1), we obtain the primary first type tumor judgment equation; If LL （1） 、 ≥LL （1） , then reset the learning rate ε in process B35, and repeat process B32 to process B36 until LL （1） 、 <LL （1） , and LL （1） 、 The value of is the smallest, and the primary first type of tumor judgment equation is obtained; Process B37: Substitute tuj, imj, uij and lkj as well as tuk, imk, uik and lkk into the primary first type tumor judgment equation respectively, and determine whether the results of the two outputs of the primary first type tumor judgment equation are both 0; If yes, it means that the primary first tumor judgment equation is accurate, and the primary first tumor judgment equation is used as the first tumor judgment equation; If not, it means that the primary first type tumor judgment equation is inaccurate. Reset the value of the learning rate ε in process B35, and repeat processes B32 to B37 until the two output results of the primary first type tumor judgment equation are both 0, and the first type tumor judgment equation is obtained.

[0015] Compared with the prior art, the present invention has the following beneficial effects: Improve diagnosis and treatment efficiency: The present invention can automatically process and analyze patient test data and generate a surgical difficulty assessment report, thereby reducing the workload of doctors and improving diagnosis and treatment efficiency. At the same time, the present invention uses the fusion of different physiological data of tumor patients as a way to diagnose tumor types. Compared with traditional tumor image analysis methods, the present invention can achieve rapid response and accurate judgment.

[0016] Reduce medical costs: With the assistance of the prediction system, doctors can more accurately assess the risks and difficulty of surgery, avoiding unnecessary overtreatment and waste of resources.

[0017] Improve patient satisfaction: The present invention can sort patients according to the difficulty of tumor resection, thereby improving the treatment effect and quality of life of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 It is a schematic diagram of the system of the present invention; Figure 2 This is a schematic diagram of data processing of the present invention; Figure 3 Schematic diagram of the (fibroglandular) tumor of the present invention. DETAILED DESCRIPTION

[0019] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] See also Figure 1 and Figure 3 , a tumor endoscopic resection difficulty prediction system based on detection data includes: a (tumor) medical record acquisition module, a medical record analysis module, a (resection) difficulty analysis module, a physician interaction module, a database and a server; wherein the (tumor) medical record acquisition module, the medical record analysis module, the (resection) difficulty analysis module and the physician interaction module are connected to the database and the server respectively; Database and server: used for data transmission and data calculation between tumor medical record acquisition module, medical record analysis module, (resection) difficulty analysis module and physician interaction module; (Tumor) medical record acquisition module: used to obtain all tumor medical record information in the target hospital as reference medical records; extract the tumor type, age, gender, treatment time and lesion index content of all tumor patients in the reference medical records to obtain tumor reference data; The content of pathological indicators includes: the content of tumor markers (in the blood), the content of immune cells (in the blood), the content of non-immune cells (in the blood) and the content of liver and kidney secretions (in the blood); Tumor markers include: alpha-fetoprotein, carcinoembryonic antigen, prostate-specific antigen, etc. Liver and kidney secretions include: alanine aminotransferase, aspartate aminotransferase, blood creatinine, etc.

[0021] It should be noted that the "target hospital" in the present invention means: a city or county-level hospital that uses the present invention (a tumor endoscopic resection difficulty prediction system based on detection data) to predict the difficulty of tumor endoscopic resection; The "entire tumor medical record information" in the present invention means: the medical records of patients who have undergone endoscopic tumor resection surgery.

[0022] Medical record analysis module: used to analyze tumor reference data based on treatment time and calculate the standard treatment time for each tumor; The influence coefficients of tumor patients' age, gender and lesion index content on treatment time were calculated, and the tumor reference data were secondary analyzed to construct the treatment time equation for each tumor. Process A: Please refer to Figure 2 ,The workflow of the medical record analysis module is as follows: Process A1: Count the number of reference medical records, recorded as bl; Count the number of all tumor types in the reference medical records, recorded as zl; Count the number of patients with the first type of tumor in the reference medical records, denoted as bl（1） ; The number of patients with the second tumor, denoted as bl （2） ; Similarly, the number of patients with the zlth tumor is denoted as bl （zl） ; Among them, bl （1） +bl （2） +……+bl （zl） =bl; Calculate the incidence of the first tumor (among tumor patients), denoted as ir （1） ;ir （1） The calculation formula is: （1） =bl （1） / bl; The incidence of the second tumor is denoted as ir （2） ;ir （2） The calculation formula is: （2） =bl （2） / bl; By analogy, the incidence of the zlth tumor is denoted as ir （zl） ;ir （zl） The calculation formula is: （zl） =bl （zl） / bl; Process A2: Group the reference medical records by tumor type to obtain group t 1 , Group t 2 ~ Group t zl ; Among them, group t 1 represents all reference medical records of patients with the first type of tumor; group t 2 represents all the reference medical records of patients with the second type of tumor; and so on, group t zl Represents all reference medical records of patients with the zlth type of tumor; Group t 1 ~ Group t zl In the above table, the treatment time of different tumor patients is recorded as to(1,1), to(1,2)~to(1,bl （1） )~to(zl,bl （zl） ); (Unit: day) Among them, to (1,1) represents group t 1 The treatment time of the first tumor patient in the group t 1 The treatment time of the second tumor patient in the test; and so on, to (1, bl （1） ) represents group t 1 bl （1） Treatment time of tumor patients; to (zl, bl （zl） ) represents group t zl bl （zl）Treatment time for each cancer patient; Process A3: According to to (1,1) ~ to (zl, bl （zl） ), calculate the standard treatment time for each tumor in turn, and get zt （1） 、zt （2） ~zt （zl） ; Among them, zt （1） represents the standard treatment time for the first type of tumor; zt （2） represents the standard treatment time for the second type of tumor; and so on, zt （zl） represents the standard treatment time for the zlth tumor; Process A31: According to to (1,1), to (1,2) ~ to (1,bl （1） ) Calculate the standard treatment time for the first type of tumor and get zt （1） ; Calculate to(1,1)~to(1,bl （1） ), denoted as ato(1,-); Process A32: Extract to (1, 1) ~ to (1, bl （1） ) is denoted as mto(1,-), and the minimum value is denoted as lto(1,-); Taking [0,1] as the error range, extract to (1,1) ~ to (1,bl （1） ) is denoted as zto(1,-); Process A33: Take [0,1] as the error range, between to(1,1)~to(1,bl （1） ) is used to count the number of tumor patients whose treatment time is equal to zto (1, -), denoted as tp; In to (1,1) ~ to (1,bl （1） ), the number of tumor patients whose treatment time is between [lto (1, -), ato (1, -)) is counted, recorded as tp1; the number of tumor patients whose treatment time is between [ato (1, -), mto (1, -)] is counted, recorded as tp2; Process A34: Comparison (tp / bl （1） ) and 0.5, determine zt （1） The value of If (tp / bl （1） )≥0.5, then zt （1） The value is: zto (1, -); If (tp / bl （1） )<0.5, then zt （1） The calculation formula is: ; Process A35: Repeat the same process of calculating the standard treatment time for the first tumor (i.e., process A31 to process A34), calculate the standard treatment time for the second to zlth tumors, and obtain zt （2） ~zt （zl） ; Process A4: According to group t 1 ~ Group t zl The age, gender and lesion index content of different tumor patients are calculated, the influence coefficient of the age, gender and lesion index content of tumor patients on the treatment time is calculated, the treatment time equation of each tumor is constructed, and the difficulty analysis module is entered; Process A41: Calculate group t 1 The effect coefficients of age, gender and lesion index content of tumor patients on treatment time were calculated to construct the treatment time equation for the first type of tumor. Group t 1 The age of the first patient is recorded as y (1,1), the tumor marker content (in the blood) is recorded as tu (1,1), the immune cell content (in the blood) is recorded as im (1,1), and the non-immune cell content (in the blood) is recorded as ui (1,1), and the liver and kidney secretion content (in the blood) is recorded as lk (1,1); The age of the second patient is recorded as y(1,2), the tumor marker content (in the blood) is recorded as tu(1,2), the immune cell content (in the blood) is recorded as im(1,2), and the non-immune cell content (in the blood) is recorded as ui(1,2), and the liver and kidney secretion content (in the blood) is recorded as lk(1,2); By analogy, the bl （1） The age of each patient is denoted as y(1,bl （1） The tumor marker content in the blood is recorded as tu(1,bl （1） , (The content of immune cells in the blood) is recorded as im(1,bl （1） , (the content of non-immune cells in the blood) is recorded as ui (1,bl （1） The content of liver and kidney secretion (in blood) is recorded as lk (1,bl （1） ); Process A42: Obtain the standard treatment time zt for the first tumor （1） ; Calculation group t 1 The influence coefficient of the age of the first patient on the treatment time is denoted as α(1,1). The calculation formula of α(1,1) is: α(1,1) = to(1,1) / zt （1） ; The influence coefficient of the second patient's age on the treatment time is denoted as α(1,2). The calculation formula of α(1,2) is: α(1,2) = to(1,2) / zt （1） ; By analogy, the bl （1） The influence coefficient of the patient's age on the treatment time is denoted as α (1,bl （1） ), α(1,bl （1） ) is calculated as: α(1,bl （1） )=to(1,bl （1） ) / zt （1） ; Calculate α(1,1)~α(1,bl （1） ) is recorded as α(1,-); α(1,-) is taken as the age influence coefficient of the first type of tumor; Process A43: Statistical group t 1 The number of male patients in the test is recorded as bem, and the number of female patients is recorded as bew; where bem + bew = bl （1） ; Calculate the effect coefficient of male patients on the treatment time of (type 1 tumor), denoted as β(1, m), and the calculation formula of β(1, m) is: β(1, m) = (bem / bl （1） )×zt （1） ; Calculate the effect coefficient of female patients on the treatment time of (type 1 tumor) β(1, w), the calculation formula of β(1, w) is: β(1, w) = (bew / bl （1） )×zt （1） ; Process A44: Calculate group t 1 The influence coefficient of the lesion index content of the first patient on the treatment time is denoted as λ(1,1); If (group t 1 If the first patient in the equation is male, the formula for calculating λ(1,1) is: If the first patient is male, the calculation formula for λ(1,1) is: ; If (group t 1 If the first patient in the equation is female, the formula for calculating λ(1,1) is: ; Calculation group t 1 The influence coefficient of the lesion index content of the second patient on the treatment time is denoted as λ(1,2); If (group t 1 If the second patient in the test is male, the formula for calculating λ(1,2) is: ; If (group t 1 If the second patient in the above equation is female, the calculation formula for λ(1,2) is: ; Similarly, calculate group t 1 bl （1） The influence coefficient of the lesion index content of each patient on the treatment time is denoted as λ (1,bl （1） ); If (group t 1 The first patient in the test is male, so λ(1,bl （1） ) is calculated as: ; If (group t 1 The first patient in the test is a female, so λ(1,bl （1） ) is calculated as: ; Process A45: Calculate λ(1,1)~λ(1,bl （1） ) is recorded as λ(1,-), and λ(1,-) is used as the influence coefficient of the pathological index of the first tumor; Summarize the data in process A41 to process A45, construct the treatment time equation for the first tumor, and obtain: Patients are male: Ttm(1) = α(1, -) × xn(1, -) + β(1, m) + λ(1, -) × xz(1, -); Patients are female: Ttw(1) = α(1, -) × xn(1, -) + β(1, w) + λ(1, -) × xz(1, -); Wherein, Ttm(1) represents the treatment time of male patients with the first tumor; Ttw(1) represents the treatment time of female patients with the first tumor; xn(1, -) represents the age of patients with the first tumor; xz(1, -) represents the geometric mean of the lesion index of patients with the first tumor; Process A46: Repeat the same process of constructing the first tumor treatment time equation (i.e., process A41 to process A45), according to the group t 2 ~ Group t zl Construct the treatment time equation for the 2nd to zlth tumors.

[0023] (Resection) Difficulty Analysis Module: It is used to perform Logistic regression analysis on tumor reference data according to the lesion index content of different tumor patients, calculate the incidence weight of each tumor, and obtain the tumor judgment equation; Obtain the patient's medical examination report, and determine the treatment difficulty of the patient's tumor type by combining the tumor judgment equation and treatment time equation for each tumor, and obtain a medical examination report; Process B: The workflow of the difficulty analysis module is as follows: Process B1: Extract the incidence of the first to zlth tumors (in tumor patients) to obtain ir （1） ~ir （zl） ; Get Group 1 The lesion index of all patients: Tumor marker levels (in blood): tu (1,1) ~ tu (1,bl （1） ); Immune cell content (in blood): im(1,1)~im(1,bl （1） ); Non-immune cell content (in blood): ui (1,1) ~ ui (1,bl （1） ); (in blood) liver and kidney secretion content: lk (1,1) ~ lk (1,bl （1） ); Process B2: Get group t 2 The lesion index of all patients: The tumor marker content (in the blood) is recorded as: tu (2,1) ~ tu (2,bl （2） ); The immune cell content in the blood is recorded as: im(2,1)~im(2,bl （2） ); The non-immune cell content (in the blood) is recorded as: ui (2,1) ~ ui (2,bl （2） ); The content of liver and kidney secretion (in blood) is recorded as: lk (2,1) ~ lk (2,bl （2） ); And so on, get group t zl The lesion index of all patients: (in blood) tumor marker content: tu (zl,1) ~ tu (zl,bl （zl） ); Immune cell content (in blood): im (zl, 1) ~ im (zl, bl （zl） ); Non-immune cell content (in blood): ui (zl, 1) ~ ui (zl, bl （zl） ); (in blood) liver and kidney secretion content: lk (zl,1) ~ lk (zl,bl （zl） ); Process B3: Obtain the normal (blood) tumor marker content range in a (healthy) human body, recorded as [tuj, tuk]; The normal range of immune cell content (in the blood) is denoted as [imj, imk]; The normal range of non-immune cells (in the blood) is denoted as [uij, uik]; Normal (in blood) range of liver and kidney secretions, denoted as [lkj, lkk]; Summarize the data in process B1 to process B3, calculate the lesion weight of the first tumor by using logistic regression, and construct the judgment equation of the first tumor; Process B31: Let the content of tumor markers in the blood of tumor patients be Xtu, the content of immune cells be Xim, the content of non-immune cells be Xui, and the content of liver and kidney secretions be Xlk; Assume that the mathematical expression of the judgment equation of the first type of tumor is equation Inr (1), and the mathematical expression of equation Inr (1) is: Inr(1)=1 / [1+e -（z（1）） ]; Among them, Inr(1) represents the probability that the patient has the first tumor; z(1) represents the linear characteristic sum of the patient suffering from the first tumor; the mathematical expression of z(1) is: z(1)=wt （1） ×Xtu+wi （1） ×Xim+wu （1） ×Xui+wl （1） ×Xlk+b(1) Among them, b(1) represents the bias that the patient has the first tumor, and the initial value of b(1) is 0; wt （1） , wi （1） 、wu （1） and wl （1） , respectively represent the tumor marker lesion weight, immune cell lesion weight, non-immune cell lesion weight and liver and kidney secretion lesion weight of the patient suffering from the first tumor; wt （1） , wi （1） 、wu （1） and wl （1） The initial value of is 1; Process B32: Let the value of Inr(1) be ir （1） , tu(1,1)~tu(1,bl （1） )、im(1,1)~im(1,bl （1） )、ui(1,1)~ui(1,bl （1） ) and lk(1,1)~lk(1,bl （1） ), as Xtu, Xim, Xui and Xlk respectively and substituted into z(1), we get equation group B-1: ; Use MATLAB software to calculate process group B-1 and get wt （1） , wi （1） 、wu （1） wl （1） and a first estimate of b(1), denoted by wt1 （1） ,wi1 （1） 、wu1 （1） wl1 （1） and b1(1); Substituting into the mathematical expression of z(1), we get the mathematical expression z(1)-1; z(1)-1=wt1 （1） ×Xtu+wi1 （1） ×Xim+wu1 （1） ×Xui+wl1 （1） ×Xlk+b1(1); Process B33: tu(1,1)~tu(1,bl （1） )、im(1,1)~im(1,bl （1） )、ui(1,1)~ui(1,bl （1） ) and lk(1,1)~lk(1,bl （1） ), as Xtu, Xim, Xui and Xlk respectively and substitute into z(1)-1, recalculate the group t 1 The estimated incidence rate of all patients in the study is: Group 1 The estimated incidence rate of the first patient in the study is denoted as irr(1). The calculation formula for irr(1) is: ; Among them, zz(1) represents (group t 1 The new linear feature sum of the first patient is calculated as: zz(1)=wt1 （1） ×tu(1,1)+wi1 （1） ×im(1,1)+wu1 （1） ×ui(1,1)+wl1 （1） ×lk(1,1)+b1(1); Group 1 The estimated incidence rate of the second patient is recorded as irr(2). The calculation formula of irr(2) is: ; Among them, zz(2) represents (group t 1 The new linear feature sum of the second patient, zz(2) is calculated as: zz(2)=wt1（1） ×tu(1,2)+wi1 （1） ×im(1,2)+wu1 （1） ×ui(1,2)+wl1 （1） ×lk(1,2)+b1(1); Similarly, group t 1 bl （1） The estimated incidence rate per patient is denoted as irr (bl （1） );irr(bl （1） ) is calculated as: ; Among them, zz (bl （1） ) means (group t 1 Middle) No. bl （1） The new linear features and zz(bl （1） ) is calculated as: zz(bl （1） )=wt1 （1） ×tu(1,bl （1） )+wi1 （1） ×im(1,bl （1） )+wu1 （1） ×ui(1,bl （1） )+wl1 （1） ×lk(1,bl （1） )+b1(1); Process B34: Calculate the logarithmic loss of the first tumor judgment equation, denoted as LL （1） LL （1） The calculation formula is as follows: ; Where irr(i) represents group t 1 The estimated incidence rate of the i-th patient in the study, the value range of i is: 1~bl （1） ; Calculate wt1 （1） The gradient of LL （1） About wt1 （1） The partial derivative of ), denoted as δ(L-tu); the calculation formula of δ(L-tu) is: ; Calculate wi1 （1） The gradient of LL （1） About wi1 （1） The partial derivative of L-im is denoted as δ(L-im). The calculation formula of δ(L-im) is: ; Calculate wu1 （1）The gradient of LL （1） About wu1 （1） The partial derivative of ), denoted as δ(L-ui); the calculation formula of δ(L-ui) is: ; Calculate wl1 （1） The gradient of LL （1） About wl1 （1） The partial derivative of L-lk is denoted as δ(L-lk). The calculation formula of δ(L-lk) is: ; Calculate the gradient of b(1) (i.e. LL （1） The partial derivative of b(1) is denoted as δ(Lb); the calculation formula of δ(Lb) is: ; Where irr(j) represents the group t 1 The estimated incidence rate of the jth patient in the study, the value range of j is: 1~bl （1） ; tu(1,j),im(1,j),ui(1,j) and lk(1,j), respectively, represent the 1 The levels of tumor markers, immune cells, non-immune cells, and liver and kidney secretions in the blood of the jth patient; Process B35: Assume that the learning rate is ε, the initial value of ε is: 0.001, and the value range of ε is: [0.001, 0.1]; Assume wt1 （1） After the change, the weight is wt2 （1） ,wt2 （1） The calculation formula is: wt2 （1） =wt1 （1） ±(ε×δ(L-tu)); Set wi1 （1） After the change, the authority is wi2 （1） ,wi2 （1） The calculation formula is: wi2 （1） =wi1 （1） ±(ε×δ(L-im)); Set wu1 （1） After the change, the authority is wu2 （1） , wu2 （1） The calculation formula is: wu2 （1） =wu1 （1） ±(ε×δ(L-ui)); Assume wl1 （1） After the change, the authority is wl2（1） ,wl2 （1） The calculation formula is: wl2 （1） =wl1 （1） ±(ε×δ(L-lk)); Assume that the weight of b1(1) after change is b2(1), and the calculation formula of b2(1) is: b2(1)=b1(1)±(ε×δ(Lb)); It should be noted that the initial value and value range of the "learning rate ε" in the present invention are only examples, and users or relevant technical personnel can adjust the initial value and value range of the learning rate ε according to actual needs; Process B36: wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1), and then substitute it back into the mathematical expression of z(1) (i.e., use “wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1)” replaces “wt” in the mathematical expression of z(1) （1） , wi （1） 、wu （1） wl （1） and b(1)”), re-execute process B32 to process B34 to obtain the new logarithmic loss of the first tumor judgment equation, denoted as LL （1） 、 ; Compare LL （1） 、 With LL （1） size; If LL （1） 、 <LL （1） , then wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1) as the weights and biases of the first tumor judgment equation, and wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） Substituting b2(1) into the mathematical expression of z(1), we obtain the (definite) primary first type tumor judgment equation; If LL （1） 、 ≥LL （1）, then reset the learning rate ε in process B35, and repeat process B32 to process B36 until LL （1） 、 <LL （1） , and LL （1） 、 The value of is the smallest, and the (definite) primary first type tumor judgment equation is obtained; Process B37: Substitute "tuj, imj, uij and lkj" and "tuk, imk, uik and lkk" into the (determined) primary first type tumor judgment equation, and determine whether the results of the two outputs of the (determined) primary first type tumor judgment equation are both 0; If yes, it means that the (determined) primary first type tumor judgment equation is accurate, and the (determined) primary first type tumor judgment equation is used as the (determined) first type tumor judgment equation; If not, it means that the (determined) primary first type tumor judgment equation is inaccurate, and the value of the learning rate ε in process B35 is reset, and processes B32 to B37 are repeatedly executed until the two output results of the (determined) primary first type tumor judgment equation are both 0, and the (determined) first type tumor judgment equation is obtained; Process B4: Repeat the same process of constructing the judgment equation for the first tumor (i.e., process B3) to construct judgment equations for the second to zlth tumors; Obtain the patient's medical examination report, extract the patient's lesion index, and substitute it into the judgment equations of the 1st to zlth tumors to determine the patient's tumor type as the diseased tumor; (that is, among the judgment equations corresponding to the 1st to zlth tumors, the judgment equation output is not 0 and the tumor type corresponding to the judgment equation with the highest incidence rate) According to the patient's medical examination report, extract the patient's age and gender, and extract the treatment time equation of the corresponding tumor, and calculate the patient's estimated treatment time, recorded as ht; According to the diseased tumor, the standard treatment time of the corresponding tumor is extracted and recorded as zht; zht∈{zt （1） 、zt （2） ~zt （zl）}; Taking [0, 1] as the error range, determine whether ht and zht are equal; If ht = zht, the patient's tumor is of moderate difficulty (endoscopic resection); If ht < zht, it means that the patient's tumor is easier to remove (endoscopically); If ht>zht, it means that the patient’s tumor is difficult to remove (endoscopically); Calculate the patient's difficulty deviation coefficient, denoted as ny; the calculation formula of ny is: ny = (ht-zht) / zht; If the difficulty of tumor (endoscopic resection) of multiple patients is the same, and the difficulty deviation coefficients of the corresponding patients are all less than 0, the patients will be sorted and treated in ascending order of their difficulty deviation coefficients; If the difficulty of tumor (endoscopic resection) for multiple patients is the same, and the difficulty deviation coefficients of the corresponding patients are all greater than 0, the patients will be sorted and treated in descending order of their difficulty deviation coefficients, and the patients will be transferred to the emergency department.

[0024] It should be noted that, due to the excessive number of physiological parameters in the human body, the “prediction of difficulty of endoscopic tumor resection” of the present invention focuses on logical design; The labels such as "lesion index content" shown in the present invention are all fuzzy definitions, and the specific definition of the lesion index content needs to be defined by the user or relevant technical personnel in actual use; When using the present invention, users or relevant technical personnel can define or design more labels or indicators to improve the accuracy of the present invention in predicting the difficulty of endoscopic resection of different tumors.

[0025] Physician interaction module: used to feed back medical examination reports to physicians; continuously obtain patients' medical examination reports and generate medical examination reports.

[0026] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technical personnel in this field according to actual conditions. For example, if there are weight coefficients and proportional coefficients, their set sizes are to quantify each parameter to obtain a specific value, which is convenient for subsequent comparison. Regarding the size of the weight coefficient and the proportional coefficient, it is sufficient as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0027] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A tumor endoscopic resection difficulty prediction system based on detection data, characterized in that: The system comprises: Medical record acquisition module: used to obtain all tumor medical record information in the target hospital as reference medical records; extract the tumor type, age, gender, treatment time and lesion index content of all tumor patients in the reference medical records to obtain tumor reference data; The content of pathological indicators includes: tumor marker content, immune cell content, non-immune cell content and liver and kidney secretion content; Medical record analysis module: used to analyze tumor reference data based on treatment time and calculate the standard treatment time for each tumor; The influence coefficients of tumor patients' age, gender and lesion index content on treatment time were calculated, and the tumor reference data were secondary analyzed to construct the treatment time equation for each tumor. Difficulty analysis module: used to perform Logistic regression analysis on tumor reference data according to the lesion index content of different tumor patients, calculate the incidence weight of each tumor, and obtain the tumor judgment equation; Obtain the patient's medical examination report, and determine the treatment difficulty of the patient's tumor type by combining the tumor judgment equation and treatment time equation for each tumor, and obtain a medical examination report; Physician interaction module: used to feed back medical examination reports to physicians; continuously obtain patients' medical examination reports and generate medical examination reports.

2. The tumor endoscopic resection difficulty prediction system based on detection data according to claim 1, characterized in that: The workflow of the medical record analysis module is as follows: Process A1: Count the number of reference medical records, recorded as bl; Count the number of all tumor types in the reference medical records, recorded as zl; Count the number of patients with the first type of tumor in the reference medical records, denoted as bl （1） ; The number of patients with the second tumor, denoted as bl （2） ; Similarly, the number of patients with the zlth tumor is denoted as bl （zl） ; Among them, bl （1） +bl （2） +……+bl （zl） =bl; Calculate the incidence of the first tumor, denoted as ir （1） ;ir （1） The calculation formula is: （1） =bl （1） / bl; By analogy, the incidence of the zlth tumor is denoted as ir （zl） ;ir （zl） The calculation formula is: （zl） =bl （zl） / bl; Process A2: Group the reference medical records by tumor type to obtain groups t1, t2, t3, t4, t5, t6, t7, t8, t9, t10, t11, t22, t13, t14, t15, t16, t17, t18, t19, t20, t21, t22, t30, t31, zl ; Among them, group t1 represents all the reference medical records of patients with the first type of tumor; and so on, group t zl Represents all reference medical records of patients with the zlth type of tumor; Group t1~group t zl In the above table, the treatment time of different tumor patients is recorded as to(1,1), to(1,2)~to(1,bl （1） )~to(zl,bl （zl） ); Among them, to (1, 1) represents the treatment time of the first tumor patient in group t1; and so on, to (1, bl （1） ) indicates the bl-th （1） Treatment time of tumor patients; to (zl, bl （zl） ) represents group t zl bl （zl） Treatment time for each cancer patient; Process A3: According to to (1,1) ~ to (zl, bl （zl） ), calculate the standard treatment time for each tumor in turn, and get zt （1） ~zt （zl） ; Among them, zt （1） represents the standard treatment time for the first type of tumor; and so on, zt （zl） represents the standard treatment time for the zlth tumor; Process A4: According to group t1 to group t zl The age, gender and lesion index content of different tumor patients are calculated, the influence coefficient of the age, gender and lesion index content of tumor patients on the treatment time is calculated, the treatment time equation for each tumor is constructed, and the difficulty analysis module is entered.

3. The tumor endoscopic resection difficulty prediction system based on detection data according to claim 2, characterized in that: The specific process of process A3 is as follows: Process A31: According to to (1, 1) ~ to (1, bl （1） ) Calculate the standard treatment time for the first type of tumor and get zt （1） ; Calculate to(1,1)~to(1,bl （1） ), denoted as ato(1,-); Process A32: Extract to (1, 1) ~ to (1, bl （1） ) is denoted as mto(1,-), and the minimum value is denoted as lto(1,-); Taking [0,1] as the error range, extract to (1,1) ~ to (1,bl （1） ) is denoted as zto(1,-); Process A33: Take [0,1] as the error range, between to(1,1)~to(1,bl （1） ) is used to count the number of tumor patients whose treatment time is equal to zto (1, -), denoted as tp; In to (1,1) ~ to (1,bl （1） ), the number of tumor patients whose treatment time is between [lto (1, -), ato (1, -)) is counted, recorded as tp1; the number of tumor patients whose treatment time is between [ato (1, -), mto (1, -)] is counted, recorded as tp2; Process A34: Comparison (tp / bl （1） ) and 0.5, determine zt （1） The value of If (tp / bl （1） )≥0.5, then zt （1） The value is: zto (1, -); If (tp / bl （1） )<0.5, then zt （1） The calculation formula is: ； Process A35: Repeat the same process of calculating the standard treatment time for the first tumor, and calculate the standard treatment time for the second to zl tumors to obtain zt （2） ~zt （zl） .

4. The system for predicting the difficulty of endoscopic tumor resection based on detection data according to claim 2, characterized in that: The specific process of process A4 is as follows: Process A41: Calculate the influence coefficients of age, gender and lesion index content of tumor patients in group t1 on treatment time, and construct the treatment time equation for the first type of tumor; The age of the first patient in group t1 is recorded as y (1,1), the tumor marker content is recorded as tu (1,1), the immune cell content is recorded as im (1,1), the non-immune cell content is recorded as ui (1,1), and the liver and kidney secretion content is recorded as lk (1,1); By analogy, the bl （1） The age of each patient is denoted as y(1,bl （1） ), the tumor marker content is recorded as tu (1,bl （1） ), the immune cell content is recorded as im (1,bl （1） ), the non-immune cell content is recorded as ui (1,bl （1） ), the liver and kidney secretion content is recorded as lk (1,bl （1） ); Process A42: Obtain the standard treatment time zt for the first tumor （1） ; Calculate the effect of the age of the first patient in group t1 on the treatment time, denoted as α(1,1). The calculation formula of α(1,1) is: α(1,1) = to(1,1) / zt （1） ; By analogy, the bl （1） The influence coefficient of the patient's age on the treatment time is denoted as α (1,bl （1） ), α(1,bl （1） ) is calculated as: α(1,bl （1） )=to(1,bl （1） ) / zt （1） ; Calculate α(1,1)~α(1,bl （1） ) is recorded as α(1,-); α(1,-) is taken as the age influence coefficient of the first type of tumor; Process A43: The number of male patients in the statistical group t1 is recorded as bem, and the number of female patients is recorded as bew; where bem + bew = bl （1） ; Calculate the influence coefficient of male patients on treatment time, recorded as β(1, m), the calculation formula of β(1, m) is: β(1, m) = (bem / bl （1） )×zt （1） ; Calculate the effect coefficient of female patients on treatment time β(1, w), the calculation formula of β(1, w) is: β(1, w) = (bew / bl （1） )×zt （1） .

5. The system for predicting the difficulty of endoscopic tumor resection based on detection data according to claim 4, characterized in that: The subsequent process of the process A43 is as follows: Process A44: Calculate the influence coefficient of the lesion index content of the first patient in group t1 on the treatment time, denoted as λ(1,1); If the first patient is male, the calculation formula for λ(1,1) is: ； If the first patient is a female, the calculation formula for λ(1,1) is: ； Similarly, calculate the bl-th （1） The influence coefficient of the lesion index content of each patient on the treatment time is denoted as λ (1,bl （1） ); If the first patient is male, then λ(1,bl （1） ) is calculated as: ； If the first patient is a female, then λ(1,bl （1） ) is calculated as: ； Process A45: Calculate λ(1,1)~λ(1,bl （1） ) is recorded as λ(1,-), and λ(1,-) is used as the influence coefficient of the pathological index of the first tumor; Summarize the data in process A41 to process A45, construct the treatment time equation for the first tumor, and obtain: Patients are male: Ttm(1) = α(1, -) × xn(1, -) + β(1, m) + λ(1, -) × xz(1, -); Patients are female: Ttw(1) = α(1, -) × xn(1, -) + β(1, w) + λ(1, -) × xz(1, -); Wherein, Ttm(1) represents the treatment time of male patients with the first tumor; Ttw(1) represents the treatment time of female patients with the first tumor; xn(1, -) represents the age of patients with the first tumor; xz(1, -) represents the geometric mean of the lesion index of patients with the first tumor; Process A46: Repeat the same process of constructing the first tumor treatment time equation, according to group t2~group t zl Construct the treatment time equation for the 2nd to zlth tumors.

6. The system for predicting the difficulty of endoscopic tumor resection based on detection data according to claim 2, characterized in that: The workflow of the difficulty analysis module is as follows: Process B1: Extract the incidence of tumors from type 1 to type zl to obtain ir （1） ~ir （zl） ; Get the lesion index of all patients in group t1: Tumor marker content: tu (1,1) ~ tu (1,bl （1） ); Immune cell content: im(1,1)~im(1,bl （1） ); Non-immune cell content: ui (1,1) ~ ui (1,bl （1） ); Liver and kidney secretion content: lk (1,1) ~ lk (1,bl （1） ); Process B2: Get group t zl The lesion index of all patients: Tumor marker content: tu (zl, 1) ~ tu (zl, bl （zl） ); Immune cell content: im(zl,1)~im(zl,bl （zl） ); Non-immune cell content: ui (zl, 1) ~ ui (zl, bl （zl） ); Liver and kidney secretion content: lk (zl, 1) ~ lk (zl, bl （zl） ); Process B3: Obtain the normal tumor marker content range in the human body, recorded as [tuj, tuk]; The normal range of immune cell content is denoted as [imj, imk]; The content range of normal non-immune cells is recorded as [uij, uik]; The content range of normal liver and kidney secretions is recorded as [lkj, lkk]; Summarize the data in process B1 to process B3, calculate the lesion weight of the first tumor by using logistic regression, and construct the judgment equation of the first tumor; Process B4: Repeat the same process of constructing the judgment equation for the first tumor to construct the judgment equations for the second to zlth tumors; Obtain the patient's medical examination report, extract the patient's lesion index, and substitute it into the judgment equation of the first to zlth tumors to determine the patient's tumor type as the diseased tumor; According to the patient's medical examination report, extract the patient's age and gender, and extract the treatment time equation of the corresponding tumor, and calculate the patient's estimated treatment time, recorded as ht; According to the diseased tumor, the standard treatment time of the corresponding tumor is extracted and recorded as zht; zht∈{zt （1） 、zt （2） ~zt （zl） }; Taking [0, 1] as the error range, determine whether ht and zht are equal; If ht = zht, it means that the patient's tumor resection is moderately difficult; If ht<zht, it means that the patient's tumor is easier to remove; If ht>zht, it means that the patient's tumor is difficult to remove.

7. The system for predicting the difficulty of endoscopic tumor resection based on detection data according to claim 6, characterized in that: The specific process of process B3 is as follows: Process B31: Let the content of tumor markers in the blood of tumor patients be Xtu, the content of immune cells be Xim, the content of non-immune cells be Xui, and the content of liver and kidney secretions be Xlk; Assume that the mathematical expression of the judgment equation of the first type of tumor is equation Inr (1), and the mathematical expression of equation Inr (1) is: Inr(1)=1 / [1+e -（z（1）） ]; Among them, Inr(1) represents the probability that the patient has the first tumor; z(1) represents the linear characteristic sum of the patient suffering from the first tumor; the mathematical expression of z(1) is: z(1)=wt （1） ×Xtu+wi （1） ×Xim+wu （1） ×Xui+wl （1） ×Xlk+b(1); Among them, b(1) represents the bias that the patient has the first tumor, and the initial value of b(1) is 0; wt （1） , wi （1） 、wu （1） and wl （1） , respectively represent the tumor marker lesion weight, immune cell lesion weight, non-immune cell lesion weight and liver and kidney secretion lesion weight of the patient suffering from the first tumor; wt （1） , wi （1） 、wu （1） and wl （1） The initial value of is 1; Process B32: Let the value of Inr(1) be ir （1） , tu(1,1)~tu(1,bl （1） )、im(1,1)~im(1,bl （1） )、ui(1,1)~ui(1,bl （1） ) and lk(1,1)~lk(1,bl （1） ), as Xtu, Xim, Xui and Xlk respectively and substituted into z(1), we get equation group B-1: ； Use MATLAB software to calculate process group B-1 and get wt （1） , wi （1） 、wu （1） wl （1） and a first estimate of b(1), denoted by wt1 （1） ,wi1 （1） 、wu1 （1） wl1 （1） and b1(1); Substituting into the mathematical expression of z(1), we get the mathematical expression z(1)-1; from(1)-1=t1 （1） ×Xtu+wi1 （1） ×Xim+wu1 （1） ×Xui+wl1 （1） ×Xlk+b1(1)。 8. The system for predicting the difficulty of endoscopic tumor resection based on detection data according to claim 7, characterized in that: The subsequent process of process B32 is as follows: Process B33: tu(1,1)~tu(1,bl （1） )、im(1,1)~im(1,bl （1） )、ui(1,1)~ui(1,bl （1） ) and lk(1,1)~lk(1,bl （1） ), as Xtu, Xim, Xui and Xlk respectively and substituted into z(1)-1, recalculate the estimated incidence rate of all patients in group t1, and obtain: The estimated incidence rate of the first patient in group t1 is denoted as irr(1); the calculation formula for irr(1) is: ； Among them, zz(1) represents the new linear feature sum of the first patient, and the calculation formula of zz(1) is: zz(1)=wt1 （1） ×tu(1,1)+wi1 （1） ×im(1,1)+wu1 （1） ×ui(1,1)+wl1 （1） ×lk(1,1)+b1(1); Similarly, the blth （1） The estimated incidence rate per patient is denoted as irr (bl （1） );irr(bl （1） ) is calculated as: ； Among them, zz (bl （1） ) indicates the bl （1） The new linear features and zz(bl （1） ) is calculated as: zz(bl （1） )=wt1 （1） ×tu(1,bl （1） )+wi1 （1） ×im(1,bl （1） )+wu1 （1） ×ui(1,bl （1） )+wl1 （1） ×lk(1,bl （1） )+b1(1); Process B34: Calculate the logarithmic loss of the first tumor judgment equation, denoted as LL （1） LL （1） The calculation formula is as follows: ； Where irr(i) represents the estimated incidence rate of the i-th patient in group t1, and the value range of i is: 1~bl （1） ; Calculate wt1 （1） The gradient of is denoted as δ(L-tu); the calculation formula of δ(L-tu) is: ； Calculate wi1 （1） The gradient of is denoted as δ(L-im); the calculation formula of δ(L-im) is: ； Calculate wu1 （1） The gradient of is denoted as δ(L-ui); the calculation formula of δ(L-ui) is: ； Calculate wl1 （1） The gradient of is denoted as δ(L-lk); the calculation formula of δ(L-lk) is: ； Calculate the gradient of b(1), denoted as δ(Lb); the calculation formula of δ(Lb) is: ； Where irr(j) represents the estimated incidence rate of the jth patient in group t1, and the value range of j is: 1~bl （1） ; tu(1,j),im(1,j),ui(1,j) and lk(1,j) respectively represent the tumor marker content, immune cell content, non-immune cell content and liver and kidney secretion content in the blood of the jth patient in group t1.

9. The system for predicting the difficulty of endoscopic tumor resection based on detection data according to claim 8, characterized in that: The subsequent process of process B34 is as follows: Process B35: Assume that the learning rate is ε, the initial value of ε is: 0.001, and the value range of ε is: [0.001, 0.1]; Assume wt1 （1） After the change, the weight is wt2 （1） ,wt2 （1） The calculation formula is: wt2 （1） =wt1 （1） ±(ε×δ(L-tu)); Set wi1 （1） After the change, the authority is wi2 （1） ,wi2 （1） The calculation formula is: wi2 （1） =wi1 （1） ±(ε×δ(L-im)); Set wu1 （1） After the change, the authority is wu2 （1） , wu2 （1） The calculation formula is: wu2 （1） =wu1 （1） ±(ε×δ(L-ui)); Assume wl1 （1） After the change, the authority is wl2 （1） ,wl2 （1） The calculation formula is: <h2 style=";text-align:left;direction:ltr">wl2<h2 style=";text-align:left;direction:ltr"> （1） <h2 style=";text-align:left;direction:ltr"> (wl1)<h2 style=";text-align:left;direction:ltr"> （1） <h2 style=";text-align:left;direction:ltr"> ±(ε×δ(L-lk)); Assume that the weight of b1(1) after change is b2(1), and the calculation formula of b2(1) is: b2(1)=b1(1)±(ε×δ(Lb)); Process B36: wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1), and then substitute them into the mathematical expression of z(1) in reverse order, and re-execute process B32 to process B34 to obtain the new logarithmic loss of the first tumor judgment equation, denoted as LL （1） 、 ; Compare LL （1） 、 With LL （1） size; If LL （1） 、 <LL （1） , then wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） and b2(1) as the weights and biases of the first tumor judgment equation, and wt2 （1） ,wi2 （1） 、wu2 （1） wl2 （1） Substituting b2(1) into the mathematical expression of z(1), we obtain the primary first type tumor judgment equation; If LL （1） 、 ≥LL （1） , then reset the learning rate ε in process B35, and repeat process B32 to process B36 until LL （1） 、 <LL （1） , and LL （1） 、 The value of is the smallest, and the primary first type of tumor judgment equation is obtained; Process B37: Substitute tuj, imj, uij and lkj as well as tuk, imk, uik and lkk into the primary first type tumor judgment equation respectively, and determine whether the results of the two outputs of the primary first type tumor judgment equation are both 0; If yes, it means that the primary first tumor judgment equation is accurate, and the primary first tumor judgment equation is used as the first tumor judgment equation; If not, it means that the primary first type tumor judgment equation is inaccurate. Reset the value of the learning rate ε in process B35, and repeat processes B32 to B37 until the two output results of the primary first type tumor judgment equation are both 0, and the first type tumor judgment equation is obtained.

Images