Three-level lymph structure-based prognosis prediction method and system for new-assisted triple negative breast cancer

Abstract

The invention discloses a prognosis prediction method and system for new-assisted triple negative breast cancer based on a three-level lymph structure, and the method comprises the steps: obtaining breast cancer tissue paraffin block data corresponding to the new-assisted triple negative breast cancer, and carrying out the slicing processing of a corresponding breast cancer tissue paraffin block; based on the three-level lymph structure, in combination with breast cancer tissue paraffin block data, grade data corresponding to the TLS density and the TLS maturity degree are generated respectively, and corresponding estimation data are generated; constructing a prediction model corresponding to the new-assisted triple negative breast cancer, and combining the estimation data to generate prediction data corresponding to the prognosis of the new-assisted triple negative breast cancer in real time; the application value of TLS maturity and spatial positioning characteristics in triple negative breast cancer prognosis evaluation can be accurately evaluated, and meanwhile, a more accurate prognosis prediction model is constructed, so that a powerful basis is provided for formulating a personalized treatment scheme, and the survival rate and life quality of triple negative breast cancer patients are promoted to be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of data prediction and processing, and in particular relates to a method and system for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure. Background Art

[0002] In many types of solid malignant tumors, the localization differences of tertiary lymphoid structures (TLS) are closely related to changes in their maturity and have distinct prognostic values. Despite this, in the field of breast cancer, especially in triple-negative breast cancer, the prognostic effects of TLS with different localization differences are still poorly understood. At a time when neoadjuvant therapy has become the mainstream treatment strategy for patients with triple-negative breast cancer, TLS evaluation that relies on surgical specimens has to consider the impact of neoadjuvant therapy, which may make the prognostic value of TLS in triple-negative breast cancer after neoadjuvant therapy different from the conclusions drawn in other tumors. Most previous studies on the prognostic value of TLS in triple-negative breast cancer have ignored the differences in TLS with different spatial locations, or did not consider the key factor of neoadjuvant therapy.

[0003] In addition, there is currently no large-sample clinical study on the prognostic value of different heterogeneous TLS in triple-negative breast cancer after neoadjuvant therapy, and it is impossible to accurately evaluate the application value of TLS maturity and spatial positioning characteristics in the prognostic evaluation of triple-negative breast cancer, and it is impossible to construct an accurate prognostic prediction model to provide a strong basis for the formulation of personalized treatment plans, thereby promoting the improvement of survival rate and quality of life of patients with triple-negative breast cancer.

[0004] Therefore, in view of the above technical problems that most previous studies on the prognostic value of TLS in triple-negative breast cancer have ignored the differences in TLS with different spatial positioning, or have not considered the key factor of neoadjuvant therapy, and have been unable to accurately evaluate the maturity and spatial positioning characteristics of TLS in the prognostic evaluation of triple-negative breast cancer, as well as the lack of large-sample clinical studies on the prognostic value of different heterogeneous TLS in neoadjuvant triple-negative breast cancer, it is urgent to design and develop a prognosis prediction method and system for neoadjuvant triple-negative breast cancer based on the tertiary lymphatic structure. Summary of the invention

[0005] In order to overcome the shortcomings and difficulties of the above-mentioned prior art, the purpose of the present invention is to provide a method, system and platform for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure, which can accurately evaluate the application value of the maturity and spatial positioning characteristics of TLS in the prognosis evaluation of triple-negative breast cancer, and at the same time construct a more accurate prognosis prediction model, in order to provide a strong basis for the formulation of personalized treatment plans, thereby promoting the improvement of the survival rate and quality of life of patients with triple-negative breast cancer.

[0006] The first purpose of the present invention is to provide a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure; the second purpose of the present invention is to provide a system for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure; the third purpose of the present invention is to provide a platform for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure.

[0007] The first object of the present invention is achieved in that the method comprises the following steps:

[0008] Acquire breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slice and process corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data;

[0009] Based on the three-level lymphatic structure, combined with the breast cancer tissue paraffin block data, first-level data corresponding to TLS density and second-level data corresponding to TLS maturity are generated respectively, and estimated data corresponding to the level data are generated; wherein the first-level data is a semi-quantitative abundance level; the second-level data is a maturity level;

[0010] A prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy is constructed, and the estimated data is combined to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time.

[0011] Further, the method of acquiring breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slicing and processing corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data, further includes:

[0012] Perform H&E staining and Ki67 immunohistochemical staining on the paraffin blocks of breast cancer tissues, and generate corresponding KFB format files;

[0013] According to the format file of the KFB, convert and generate corresponding pathological image data in the SVS format;

[0014] The WSI file in the SVS format is subjected to region delineation processing, area measurement processing and ROI region cell number calculation processing respectively.

[0015] Furthermore, the construction of a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes:

[0016] Obtain sample data corresponding to triple-negative breast cancer after neoadjuvant therapy, and determine and estimate whether there are significant prognostic differences between different groups of samples;

[0017] Cox proportional hazards regression analysis was performed on each covariate separately, and covariate data with significant association with survival time were generated;

[0018] A multi-factor Cox model was constructed, and continuous variables were converted into restricted cubic spline form and added into the multi-factor Cox model to generate corresponding node quantity data.

[0019] Furthermore, the construction of a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes:

[0020] Based on the prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, the relationship data between the three-level lymphatic structure and disease-free survival and overall survival were generated respectively;

[0021] Acquire clinical pathological characteristic data corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate correlation data between the clinical pathological characteristic data and the characteristic data of tertiary lymphatic structure;

[0022] Linear correlation data between the tertiary lymphatic structure characteristic data and the prognosis data are created, and corresponding correction data are generated according to the linear correlation data; wherein the linear correlation data includes linear correlation data and nonlinear correlation data.

[0023] Furthermore, the construction of a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes:

[0024] Based on the three-level lymphatic structure characteristic data, a DFS prediction model for triple-negative breast cancer was constructed, and the performance of each prediction model was compared to generate model performance evaluation index data corresponding to the performance.

[0025] The second object of the present invention is achieved as follows: the system is applied to a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure, and the system comprises:

[0026] A data acquisition processing unit, used to acquire breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slice and process the corresponding breast cancer tissue paraffin block according to the breast cancer tissue paraffin block data;

[0027] A first data generating unit is used to generate first-level data corresponding to TLS density and second-level data corresponding to TLS maturity based on the three-level lymphatic structure and in combination with the breast cancer tissue paraffin block data, and generate estimated data corresponding to the level data; wherein the first-level data is a semi-quantitative abundance level; and the second-level data is a maturity level;

[0028] The second data generating unit is used to construct a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, and to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time in combination with the estimated data.

[0029] Furthermore, the data acquisition processing unit further includes:

[0030] The first processing module is used to perform H&E staining and Ki67 immunohistochemical staining on the paraffin block of breast cancer tissue respectively, and generate corresponding KFB format files;

[0031] A second processing module is used to convert and generate corresponding pathological image data in SVS format according to the format file of KFB;

[0032] The third processing module is used to perform region delineation processing, area measurement processing and ROI region cell number calculation processing on the WSI file in the SVS format.

[0033] Furthermore, the second data generating unit further includes:

[0034] The first determination module is used to obtain sample data corresponding to triple-negative breast cancer after neoadjuvant therapy, and to determine and estimate whether there is a significant prognostic difference between different groups of samples;

[0035] The first data generation module is used to perform Cox proportional hazard regression analysis on each covariate separately and generate covariate data that are significantly associated with survival time;

[0036] The second data generation module is used to construct a multi-factor Cox model, and at the same time convert the continuous variables into a restricted cubic spline form and add them to the multi-factor Cox model to generate corresponding node quantity data.

[0037] Furthermore, the second data generating unit further includes:

[0038] The third data generation module is used to generate the relationship data between the three-level lymphatic structure and the disease-free survival and overall survival based on the prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy;

[0039] The fourth data generation module is used to obtain clinical pathological characteristic data corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate correlation data between the clinical pathological characteristic data and the tertiary lymphatic structure characteristic data;

[0040] A fifth data generation module, used to create linear correlation data between the three-level lymphatic structure characteristic data and the prognosis data, and generate corresponding correction data according to the linear correlation data; wherein the linear correlation data includes linear correlation data and nonlinear correlation data;

[0041] The second data generating unit further includes:

[0042] The sixth data generation module is used to construct a DFS prediction model for triple-negative breast cancer based on the tertiary lymphatic structure feature data, compare the performance of each prediction model, and generate model performance evaluation indicator data corresponding to the performance.

[0043] The third object of the present invention is achieved as follows: it includes a processor, a memory and a neoadjuvant triple-negative breast cancer prognosis prediction platform control program based on a three-level lymphatic structure; wherein the neoadjuvant triple-negative breast cancer prognosis prediction platform control program based on a three-level lymphatic structure is executed by the processor, the neoadjuvant triple-negative breast cancer prognosis prediction platform control program based on a three-level lymphatic structure is stored in the memory, and the neoadjuvant triple-negative breast cancer prognosis prediction platform control program based on a three-level lymphatic structure implements the neoadjuvant triple-negative breast cancer prognosis prediction method based on a three-level lymphatic structure.

[0044] The scheme of the present invention obtains breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slices and processes corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data;

[0045] Based on the three-level lymphatic structure, combined with the breast cancer tissue paraffin block data, first-level data corresponding to TLS density and second-level data corresponding to TLS maturity are generated respectively, and estimated data corresponding to the level data are generated; wherein the first-level data is a semi-quantitative abundance level; the second-level data is a maturity level;

[0046] A prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy is constructed, and combined with the estimated data, prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy is generated in real time; and a system and platform corresponding to the method can accurately evaluate the application value of the maturity and spatial positioning characteristics of TLS in the prognosis evaluation of triple-negative breast cancer, and at the same time construct a more accurate prognosis prediction model, in order to provide a strong basis for the formulation of personalized treatment plans, thereby promoting the improvement of the survival rate and quality of life of patients with triple-negative breast cancer.

[0047] That is to say, this scheme is a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure. By constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data, high-precision prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy is generated in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 This is a schematic diagram of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0050] Figure 2 This is a schematic diagram of an example of pathological section evaluation of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0051] Figure 3 This is a schematic diagram of the comparison of different maturation stages of TLS and SLO (tumor-infiltrated axillary lymph nodes) in an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure of the present invention;

[0052] Figure 4 This is a schematic diagram of the research process of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0053] Figure 5 This is a schematic diagram of mulberries showing the TLS area and maturity in the SYSU cohort of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0054] Figure 6 Schematic diagram of the estimated disease-free survival rate and overall survival rate (Kaplan-Meier) of the SYSU cohort and the TCGA cohort in an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0055] Figure 7 This is a heat map showing the correlation between clinical pathological characteristics and tertiary lymphatic structure characteristics of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0056] Figure 8This is a forest diagram of single factor Cox regression analysis of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0057] Fig. 9 This is a forest diagram of multi-factor Cox regression analysis of DFS and OS in Model 1 of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0058] Fig.10 This is a forest diagram of multi-factor Cox regression analysis of DFS and OS in Model 2 of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0059] Fig.11 This is a forest diagram of multi-factor Cox regression analysis of DFS and OS in Model 3 of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0060] Fig.12 This is a forest diagram of multi-factor Cox regression analysis of DFS and OS in Model 4 of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0061] Fig.13 It is a schematic diagram of the nonlinear association between the adjusted TLS features and the DFS hazard ratio (HR) of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0062] Fig.14 The ROC curve and DCA decision curve of the Model 1-Model of the embodiment of the method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure of the present invention;

[0063] Fig.15 The three-year and five-year calibration curves of the Model 1-Model 5 of the embodiment of the method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure of the present invention;

[0064] Fig.16 This is a schematic diagram of a DFS prognostic model constructed based on TLS features in an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0065] Fig.17 This is a schematic diagram of DFS and OS stratification (Kaplan-Meier) of the SYSU cohort and TCGA cohort based on TLS immune risk score in an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure of the present invention;

[0066] Fig.18 This is a schematic diagram of spatial classification based on TLS in an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0067] Fig.19 This is a schematic diagram of DFS and OS stratification (Kaplan-Meier) of SYSU cohort and TCGA cohort based on TLS spatial typing in an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0068] Fig. 20 This is a schematic diagram of sensitivity analysis of the TCGA cohort of an embodiment of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure of the present invention;

[0069] Fig.21 This is a schematic diagram of the architecture of a prognosis prediction system for triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure of the present invention;

[0070] Fig. 22 This is a schematic diagram of the architecture of a prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure of the present invention;

[0071] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0072] In order to better understand the purpose, technical solutions and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0073] The present invention may also be implemented or applied through other different specific examples, and the details in this specification may also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0074] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0075] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. Secondly, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0076] Preferably, the method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure of the present invention is applied in one or more terminals or servers. The terminal is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASIC), programmable gate arrays (FPGA), digital signal processors (DSP), embedded devices, etc.

[0077] The terminal can be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal can interact with the client through a keyboard, a mouse, a remote control, a touch pad, or a voice control device.

[0078] The present invention aims to realize a method, system and platform for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure.

[0079] like Figure 1 , which is a flow chart of a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure provided by an embodiment of the present invention.

[0080] In this embodiment, the neoadjuvant triple-negative breast cancer prognosis prediction method based on the tertiary lymphatic structure can be applied to a terminal or a fixed terminal with a display function, and the terminal is not limited to a personal computer, a smart phone, a tablet computer, a desktop computer or an all-in-one computer equipped with a camera, etc.

[0081] The method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure can also be applied to a hardware environment consisting of a terminal and a server connected to the terminal through a network. The network includes but is not limited to: a wide area network, a metropolitan area network or a local area network. The method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure of the embodiment of the present invention can be executed by a server, can be executed by a terminal, or can be executed by both a server and a terminal.

[0082] For example, for a terminal that needs to predict the prognosis of triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure, the function of predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure provided by the method of the present invention can be directly integrated on the terminal, or a client for implementing the method of the present invention can be installed. For another example, the method provided by the present invention can also be run on a server or other device in the form of a software development kit (SDK), and an interface for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure is provided in the form of SDK. The terminal or other device can realize the function of predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure through the provided interface. The present invention is further described below in conjunction with the accompanying drawings.

[0083] like Figure 1 As shown, the present invention provides a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure, wherein the method comprises the following steps:

[0084] S01, obtaining breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slicing and processing corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data;

[0085] S02, based on the three-level lymphatic structure, combined with the breast cancer tissue paraffin block data, respectively generate first-level data corresponding to TLS density and second-level data corresponding to TLS maturity, and generate estimated data corresponding to the level data; wherein the first-level data is a semi-quantitative abundance level; the second-level data is a maturity level;

[0086] S03. Construct a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time in combination with the estimated data.

[0087] The method of obtaining breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slicing and processing corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data, further includes:

[0088] S011. Perform H&E staining and Ki67 immunohistochemical staining on the paraffin blocks of breast cancer tissue respectively, and generate corresponding KFB format files;

[0089] S012. According to the format file of the KFB, convert and generate corresponding pathological image data in the SVS format;

[0090] S013, performing region delineation processing, area measurement processing and ROI region cell number calculation processing on the WSI file in the SVS format respectively.

[0091] The method of constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes:

[0092] S031. Obtain sample data corresponding to triple-negative breast cancer after neoadjuvant therapy, and determine and estimate whether there are significant prognostic differences between different groups of samples;

[0093] S032. Perform Cox proportional hazard regression analysis on each covariate separately and generate covariate data that are significantly associated with survival time;

[0094] S033. Construct a multi-factor Cox model, convert continuous variables into restricted cubic spline form, and add them to the multi-factor Cox model to generate corresponding node quantity data.

[0095] The method of constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes:

[0096] S034. Based on the prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, the relationship data between the three-level lymphatic structure and disease-free survival and overall survival were generated respectively;

[0097] S035. Acquire clinical pathological characteristic data corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate correlation data between the clinical pathological characteristic data and the tertiary lymphatic structure characteristic data;

[0098] S036. Create linear correlation data between the tertiary lymphatic structure characteristic data and the prognosis data, and generate corresponding correction data based on the linear correlation data; wherein the linear correlation data includes linear correlation data and nonlinear correlation data.

[0099] The method of constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes:

[0100] S037. Based on the tertiary lymphatic structure characteristic data, a DFS prediction model for triple-negative breast cancer is constructed, and the performance of each prediction model is compared to generate model performance evaluation indicator data corresponding to the performance.

[0101] Specifically, in the embodiment of the present invention, mature TLS and TLS in the proximal (T) region indicate a better prognosis for patients with triple-negative breast cancer, while TLS in the distal (P) region is associated with a worse prognosis for patients with triple-negative breast cancer.

[0102] SYSU cohort, this protocol retrospectively analyzed an independent long-term follow-up cohort (SYSU cohort) from Sun Yat-sen Memorial Hospital, Sun Yat-sen University. The inclusion criteria of the study population are as follows:

[0103] (1) Aged ≥20 years and <75 years at the time of first diagnosis, female.

[0104] (2) Breast cancer patients who received standardized neoadjuvant treatment and then underwent surgery at Sun Yat-sen Memorial Hospital of Sun Yat-sen University between 2011 and 2021.

[0105] (3) Pathological molecular classification is triple-negative breast cancer, that is, both ER and PR are negative (IHC ER and PR positive percentage <1%), and Her2 is negative or low expression (IHC Her2- / + or IHC Her2++ but FISH-).

[0106] (4) The pathological type is invasive carcinoma.

[0107] (5) Sign the informed consent form for the reuse of discarded specimens.

[0108] Exclusion criteria were as follows:

[0109] (1) The patient's electronic medical record is incomplete, the neoadjuvant therapy course has not been completed, or follow-up data are difficult to obtain.

[0110] (2) The pathological sections of the patient's surgical resection are missing or difficult to obtain.

[0111] (3) The patient already has recurrence or distant metastasis at the time of diagnosis.

[0112] (4) Pathological diagnosis was inflammatory breast cancer.

[0113] (5) History of autoimmune disease or any active autoimmune disease.

[0114] The study cohort ultimately included a total of 235 patients with early- and middle-stage triple-negative breast cancer who underwent surgery at the Breast Cancer Center of Sun Yat-sen Memorial Hospital, Sun Yat-sen University from August 2011 to September 2021.

[0115] The main outcome measure of the study was disease-free survival (DFS), which was defined as the time from the date of completion of surgical treatment to cancer recurrence, metastasis, or all-cause death. The main outcome events were disease progression and death, that is, any one of cancer recurrence, metastasis, or all-cause death. The evaluation method was mainly clinical follow-up, including imaging examinations (such as breast ultrasound, mammography, CT, MRI, etc.) and blood test results, outpatient treatment records, hospitalization records, etc. performed in our hospital. For patients who lacked routine follow-up records after surgery, telephone follow-up was used to contact the patients or contacts.

[0116] The secondary outcome measure was overall survival (OS), which was defined as the time from enrollment to death due to any cause. The same clinical follow-up and data recording methods as DFS were used. The secondary outcome event was death from any cause.

[0117] The clinical preoperative baseline data of all patients were collected, including age, gender, height, weight at the time of initial diagnosis, preoperative ultrasound tumor size (to determine clinical T stage), preoperative axillary lymph node examination and ultrasound results (clinical N stage), clinical TNM stage according to the American Joint Committee on Cancer (AJCC) standard, and operation time.

[0118] The postoperative pathological baseline data of all patients were collected, including pathological number, number of intraoperative positive lymph nodes, Ki67 level, and intravascular tumor thrombus infiltration (Lymph-Vascular Invasion, LVI).

[0119] The latest update of follow-up data was in January 2023. If the outcome event has occurred, the final follow-up time is recorded as the time of the outcome event; if the outcome event has not occurred, the final follow-up time is recorded as the time of the last follow-up. The median follow-up time is 26.9 months.

[0120] TCGA cohort: TCGA (The Cancer Genome Atlas) is a large-scale cancer genome project jointly launched by the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI) of the National Institutes of Health (NIH). TCGA-BRCA refers to a subset of the TCGA project specifically for breast cancer, which includes clinical information, follow-up information, transcriptome analysis, epigenetics, pathological section data, etc. of most subjects. In terms of pathological sections, TCGA provides two types of data: rapid frozen sections and formalin-fixed paraffin-embedded (FFEP) sections. Files with "TS#" or "BS#" logos are rapid frozen sections, and files with "DX#" logos are formalin-fixed paraffin-embedded sections. Rapid frozen sections have a small area, and the TLS on them is difficult to identify. Therefore, the present invention only selects formalin-fixed paraffin-embedded sections marked with "DX#".

[0121] According to the clinical baseline data downloaded from the TCGA official website (https: / / portal.gdc.com) and the definition of triple-negative breast cancer, patients in the cohort with negative ER, PR, and Her2 were considered to be triple-negative breast cancer patients. The inclusion criteria are as follows: (1) ER negative, that is, "breast carcinoma progesterone receptor status" is "Negative". (2) PR negative, that is, "breast carcinoma estrogen receptor status" is "Negative". (3) Her2 negative, that is, "lab proc her2 neu immunohistochemistry receptor status" or "lab procedure her2 neu in situ hybrid outcome type" is "Negative". (4) The pathological type is invasive ductal carcinoma, that is, the "histological type" is "Infiltrating Ductal Carcinoma".

[0122] The exclusion criteria were as follows: (1) distant metastasis was present at the time of diagnosis, i.e., TNM stage M1. (2) lack of paraffin pathology section data or the area of ​​the pathology section was too small. (3) paraffin pathology specimens lacked the peritumoral area and could not be evaluated.

[0123] A total of 147 triple-negative breast cancer patients were included in the study cohort. The RNAseq data (level 3), corresponding clinical baseline information, and SVS format files with complete paraffin pathology section images of these 147 triple-negative breast cancer patients were downloaded from the TCGA official website (https: / / portal.gdc.com).

[0124] The primary evaluation index was DFS. The secondary evaluation index was OS. The definition of DFS and OS was the same as that in 2.1.1. The survival follow-up data of these 147 patients with triple-negative breast cancer were downloaded from the officially recognized proxy website UCSC Xena (https: / / xena.ucsc.edu / ). The median follow-up time was 26.3 months. The evaluation of the primary and secondary outcome indicators was the same as that of the SYSU cohort.

[0125] Experimental materials: The breast cancer tissue specimens used in this study were collected from patients who met the inclusion criteria of the SYSU cohort and were collected by the pathology department after surgical resection. They were fixed, dehydrated, transparent, and waxed and preserved in the form of tissue wax blocks. All samples were obtained from the pathology department of Sun Yat-sen Memorial Hospital to ensure the consistency of sample processing.

[0126] Submit a list of patients including pathology numbers to the Pathology Department and request the available H&E stained sections and Ki67 immunohistochemical stained sections for patients who meet the inclusion criteria.

[0127] For patients with the following conditions, additional corresponding tissue wax blocks should be requested:

[0128] (1) H&E staining and Ki67 immunohistochemical staining sections are missing, stained, or faded.

[0129] (2) The peritumoral area of ​​H&E staining and Ki67 immunohistochemical staining sections accounted for <30%, making it impossible to evaluate the peritumoral area.

[0130] The tumor border area of ​​the tissue wax block was supplemented with stain to ensure that the evaluation of TLS on the slices could cover the tumor area and the peritumoral area.

[0131] This study was approved by the hospital ethics committee (approval number: [SYSKY-2024-058-01]), and all patients signed informed consent for the reuse of discarded specimens.

[0132] The experimental reagents are shown in the following table:

[0133] Table 1: Experimental reagents

[0134] Reagent name Manufacturer Eosin Hematoxylin Stain (G1005-500ML) Wuhan Seville Company Hematoxylin blueing solution (G1040-500ML) Wuhan Seville Company Hematoxylin differentiation solution (G1039-500ML) Wuhan Seville Company Ki67 antibody (monoclonal, clone ID: D2H10) CST (Cell Signaling Technology) DAB colorimetric reagent Wuhan Seville Company PBS Phosphate Buffered Saline Powder Beijing Solarbio Company Xylene Sigma 100× Sodium Citrate Buffer Shanghai Maclin Company Anhydrous ethanol Guangzhou Reagent Factory 3% Hydrogen Peroxide Guangzhou Reagent Factory 5% BSA Gibco Corporation 4% paraformaldehyde tissue fixative Wuhan Seville Company

[0135] Experimental methods: tissue processing

[0136] All additional tissue specimens requested were paraffin blocks of breast cancer tissue obtained from the Department of Pathology after surgery and from patients who met the inclusion criteria. After paraffin fixation, the tissues were serially sliced ​​to a thickness of approximately 4 μm.

[0137] H&E staining: (1) Preheat the paraffin sections in a 58°C oven and dewax after 60 minutes of baking. (2) Dewax with xylene and then dehydrate with 100%, 95%, and 70% ethanol. (3) Add hematoxylin stain for 5 minutes, wash quickly with water, and use aluminum acetate solution for identification and staining for 1 minute. (4) Add eosin stain for 2 minutes, dehydrate with 95% and 100% ethanol and xylene to make it transparent. (5) Wax sealing to complete the slide preparation process.

[0138] Ki67 immunohistochemical staining: (1) The dewaxing and hydration steps are consistent with those of H&E staining. (2) Add pH 6.0 citrate buffer for antigen retrieval, heat to 121°C in an autoclave, and maintain at this temperature for 2 minutes. (3) Cool the specimen to ambient temperature and rinse with PBS buffer three times, each rinse lasting about 5 minutes. (4) To prevent interference from endogenous peroxidase, incubate the specimen in a solution containing 3% hydrogen peroxide for 20 minutes. (5) Block nonspecific binding sites with 5% bovine serum albumin for 30 minutes. (6) Add the primary antibody Ki67 (1:150 dilution) and incubate overnight at 4°C. (7) Incubate the specimen with the secondary antibody at ambient temperature for 1 hour. (8) The DAB color development lasts for 1-2 minutes, and the reaction should be closely monitored to prevent over-staining. (9) After a slight counterstaining with eosin, dehydrate and clear, and then mount the slides with paraffin.

[0139] Microscope image acquisition: All H&E stained specimens and Ki67 immunohistochemical stained specimens were scanned by a digital scanner (model KF-PRO-040-HI, Jiangfeng Bio) for whole slide imaging (WSI) and stored as KFB format files, which were converted into pathological image data in SVS format by format conversion software. The publicly available Qupath software (version 0.3.2) was used to delineate the region of interest (ROI), measure the area, and calculate the number of cells in the ROI area for the WSI files in SVS format.

[0140] The format conversion software is open source (https: / / github.com / wilmerwang / SLFCD / releases).

[0141] Interpretation and evaluation of pathological sections: Overall evaluation principle: The SVS format digital pathological sections of H&E-stained breast cancer tissues or adjacent normal breast tissues of the SYSU cohort and TCGA cohort were independently evaluated by the author and another breast surgeon trained in pathology. When there was a disagreement in the evaluation, the pathologist participated in the evaluation process and discussed the final evaluation results. The interpretation and evaluation of these pathological sections followed a clearly defined method. Throughout the process, another breast surgeon and pathologist were completely isolated from any clinical information or follow-up results to ensure an unbiased evaluation. The main contents of the evaluation included tertiary lymphoid structures (TLS), tumor-infiltrating lymphocytes (TILs), and vascular-infiltrating tumor emboli (LVI). Ki67-stained sections of the SYSU cohort were used to assist in the evaluation of H&E-stained sections. Due to the lack of Ki67 staining data in the TCGA cohort, the interpretation of the H&E-stained sections in this cohort relied on the rich evaluation experience gained from the analysis of H&E-stained sections in the SYSU cohort.

[0142] TLS assessment method: TLS, or tertiary lymphoid structures, can be broadly defined as organized lymphocyte aggregates

[16] , which can be observed and analyzed on H&E stained sections. However, in a specific context, in order to minimize the potential confusion between the presence of small TLS and TILs, this protocol stipulates that structures formed by 100 or more organized lymphocyte aggregates are TLS

[47] .

[0143] Therefore, according to the purpose of the study, the samples were divided into two different groups according to the presence or absence of TLS (e.g. Figure 2 ):(in, Figure 2 Figure A is a comparison between the TLS(-) group and the TLS(+) group. The solid yellow line outlines the TLS in the TLS(+) group as the ROI. Figure B is an example of the near-tumor (T) area and far-tumor (P) area of ​​TLS distribution. The area within the blue solid line is the T area, and the area outside the area is the P area. Invasive cancer areas, areas with residual degenerative cancer cells, areas where cancer cells are necrotic or completely disappeared and replaced by granulation tissue or fibrosis, and areas within 0.5 mm of their edges are all considered near-tumor (T) areas. Far-tumor (P) areas are areas on the sections that do not meet the above criteria. Figure C is an example of intravascular invasion (LVI). The red arrow points to the intravascular tumor thrombus. Figure D is an example of TIL grouping based on the TIL guidelines: low-infiltration group (0-10%), medium-infiltration group (11%-40%), and high-infiltration group (41%-90%)).

[0144] (1) The TLS(+) group was characterized by the presence of at least one aggregate containing at least 100 lymphocytes;

[0145] (2) The TLS(-) group included cases with no TLS at all or only small aggregates of less than 100 cells.

[0146] Qupath software (version 0.3.2) was used to count the number of cells, mark the region of interest (ROI) in the TLS(+) group, measure the area, and save the ROI marking data as a geojson format file.

[0147] To investigate TLS heterogeneity in more depth, various factors including the spatial distribution, abundance, and maturation stage of TLS in the TLS(+) group were comprehensively analyzed.

[0148] In order to determine the spatial distribution of TLS, this protocol refers to the previously published TLS research standards in different types of tumors: the distribution area of ​​TLS is subdivided into two sub-areas, the near-tumor (T) area and the far-tumor (P) area. The T area is the marginal area within 0.5 mm from the invasive edge of the tumor plus the invasive cancer area, and the P area is the interstitial area 0.5 mm and above from the invasive edge of the tumor. It is worth noting that, unlike other types of tumors, most patients with triple-negative breast cancer receive neoadjuvant chemotherapy before surgery, and then surgically remove tissue samples for pathological examination and evaluation of TLS. According to the hypothesis of this protocol study, after the cancer tissue is necrotic and replaced by granulation tissue or fibrosis, TLS located near the tumor area before neoadjuvant treatment may not be directly adjacent to the periphery of invasive cancer. Nevertheless, these TLS must still be classified as part of the T area. Therefore, it is necessary to adjust the standards commonly used in previous studies when distinguishing the localization sub-areas of TLS.

[0149] Therefore, in this study, with reference to the pathological evaluation criteria for breast cancer after neoadjuvant therapy

[61] , a new TLS distribution area standard was developed: invasive cancer areas, areas with residual degenerative cancer cells, areas where cancer cells are necrotic or completely disappeared and replaced by granulation tissue or fibrosis, and areas within 0.5 mm of their edges should all be considered as tumor-proximal (T) areas. Tumor-distal (P) areas are areas on the sections that do not meet the above criteria.

[0150] Three methods were used to evaluate the abundance of TLS: rating level, area ratio per (‰), and count. Referring to the standards set by published studies

[47] and the research practice of this scheme, the number of TLS was divided into four levels:

[0151] (1) None (grade 0) - no TLS were observed or only small aggregates of less than 100 cells were observed; (2) Low density (grade 1) - 1 to 3 TLS were observed; (3) Medium density (grade 2) - 4 or more TLS were observed, but the high density criteria were not met; (4) High density (grade 3) - some TLS were connected into sheets.

[0152] The TLS density was evaluated in the T and P regions to obtain the corresponding semi-quantitative abundance levels: T_level and P_level. In addition to the semi-quantitative quantity rating, a quantitative method was also used to evaluate the abundance of TLS. The TLS area (S-TLS) on each H&E-stained slide of the patient was measured, including the TLS area in the T and P regions (TS-TLS & PS-TLS) and the total area of ​​the tissue section WSI (S-WSI). For cases with multiple sections, the measured areas were averaged. T_per (‰) = TS-TLS / S-WSI * 1000 and P_per (‰) = PS-TLS / S-WSI * 1000 were used as one of the indicators for evaluating the abundance of TLS. The number of ROIs marked in different regions was recorded as T_count and P_count.

[0153] In the assessment of TLS maturity, the currently widely recognized assessment criteria [30-32,47] are based on lymphoid follicles (Fol) and germinal centers (GC): Early TLS (E-TLS) is a dense aggregate of mixed immature B cells and T cells, lacking lymphoid follicles and germinal centers (Fol-&GC-); Primary follicle-like TLS (PFL-TLS) is a lymphocyte aggregate with lymphoid follicles but lacking germinal centers (Fol+&GC-); Secondary follicle-like TLS (SFL-TLS) is characterized by both lymphoid follicles and germinal centers and is the most mature TLS (Fol+&GC+). Lymphoid follicles are mainly composed of rapidly proliferating B lymphocytes and FDCs

[62] . Previous studies [30-32, 47] usually use CD21 and CD23 to mark FDC to assist in the identification of lymph follicles and thus evaluate the maturity of TLS. Ki67 is a marker of cell proliferation and can be used to mark highly proliferating B cells in TLS

[21] . Previous studies have also demonstrated the feasibility of Ki67 marking lymph follicles

[63] . At the same time, Ki67 is also one of the most commonly used immunohistochemical staining for breast cancer. Therefore, Ki67 immunohistochemical staining was chosen to assist in the identification of lymph follicles and germinal centers by analogy with SLO, i.e., axillary lymph nodes, to evaluate the maturity of TLS ( Figure 3 ).in, Figure 3A, three maturation stages of TLS: E-TLS (Early TLS): no lymphoid follicles, no Ki67 staining; PFL-TLS (Primary follicle-like TLS): lymphoid follicles, but no germinal centers, Ki67 staining but not aggregated; SFL-TLS (Secondary follicle-like TLS): germinal centers. Ki67 aggregates in the germinal centers. B. Three maturation stages of SLO. H&E staining and Ki67 staining in SLO are similar to those in TLS, proving that the evaluation method of TLS maturity by Ki67 is reliable.

[0154] Therefore, similar to lymph nodes, TLS with no or very little Ki67 positivity indicates almost no B cell proliferation, which is considered to be E-TLS; scattered Ki67 positivity indicates that proliferating B cells are not aggregated and are distributed in primary lymphoid follicles, which is considered to be PFL-TLS; aggregated Ki67 positivity indicates the presence of germinal centers, which is considered to be SFL-TLS. Attention should be paid to the differentiation of Ki67 positivity in tumor cells and Ki67 positivity in B cells, and the differentiation method is mainly based on the comparison with the nuclei of tumor cells in the tumor area.

[0155] The maturity of TLS was evaluated in the T and P regions respectively, and the cases were divided into 4 levels according to the maturity of TLS: (1) cases where no TLS was observed were classified as level 0; (2) cases where only E-TLS was observed in all sections were classified as level 1; (3) cases where PFL-TLS was observed in some sections but SFL-TLS was not found in all sections were classified as level 2; (4) cases where SFL-TLS was observed were classified as level 3.

[0156] LVI assessment method: Lymphovascular invasion (LVI) refers to the infiltration of cancer cells into blood vessels or lymphatic vessels during pathological examination. Traditionally, LVI is one of the most important pathological features of malignant tumors and is associated with a poor prognosis in different cancers, including breast cancer

[64] . Therefore, LVI is also included in the pathological variables evaluated. In the peritumoral area of ​​the H&E-stained section, look for lymphatic vessels and capillary lumens composed of endothelial cells. Evaluate whether there are tumor thrombi composed of cancer cells in the lumen and record them.

[0157] TIL assessment method: TILs in SVS digital pathology sections were assessed according to the criteria set out in the 2017 TIL assessment guidelines of the International Breast Cancer Immuno-Oncology Biomarker Working Group

[38] . Specific precautions are as follows: (1) A low-power macroscopic scan was performed to determine the area for TIL assessment, including the central tumor and the invasive margin. The invasive margin was defined as a 1 mm wide area centered on the boundary between the cancer nest and normal tissue; (2) the average TIL level in the tumor area was assessed, rather than being limited to the hot spot area; (3) TILs should be reported as a percentage of a continuous variable during the assessment. According to the guidelines, TILs were divided into three groups: low invasive group (0-10%), medium invasive group (11%-40%), and high invasive group (41%-90%).

[0158] Ki67 evaluation method: The Ki67 expression level in the SYSU cohort was interpreted using a standardized “typewriter” visual evaluation method

[65] , and Ki67 ≤ 30% was classified as a low expression group, and Ki67 > 30% was classified as a high expression group. However, the TCGA cohort lacked Ki67 immunohistochemistry results, so this protocol used RNAseq of the MIB1 gene (which regulates Ki67 protein expression) instead, and used the median expression level as the cutoff point to divide the TCGA cohort into a low expression group and a high expression group.

[0159] Statistical analysis: All statistical analyses were performed using R 4.3.2 software.

[0160] Baseline statistics: The quantitative data that conformed to the normal distribution were expressed as Mean (SD), and those that did not conform to the normal distribution were expressed as Median [IQR]. For the data that conformed to the normal distribution, the independent sample t test was used for the comparison between the two groups, and the Mann-Whitney U rank sum test was used for the comparison between the two groups that did not conform to the normal distribution. The count data were described by n (%), and the Spearson chi-square test or Fisher's exact test was used to compare the differences between the groups. The difference was statistically significant when the two-sided test p < 0.05. For the grade data in this protocol, such as clinical TNM stage, TLS Mature, T level, P level and TIL infiltration level, they were all included in the analysis as quantitative variables.

[0161] Mulberry graph drawing: drawn using the networkD3 package.

[0162] KM analysis: The survfit function of the survival package was used to analyze the Kaplan-Meier survival function curves of disease-free survival and overall survival in different groups in the SYSU cohort and TCGA cohort. The log-rank test was used to assess whether there were significant prognostic differences between samples in different groups.

[0163] Correlation analysis: The pheatmap package was used to describe the correlation between quantitative variables that were not normally distributed. A p value of less than 0.05 was considered statistically significant.

[0164] Univariate and multivariate Cox regression analysis: Cox proportional hazard regression analysis was performed on each covariate using the coxph function in the survival package. Covariates significantly associated with survival time were identified. On the basis of univariate analysis, statistically significant variables were selected for multivariate Cox regression, and a multivariate Cox regression model was established using the coxph function.

[0165] Nonlinear analysis: The continuous variables were converted into restricted cubic spline form using the rcs function in the rms package and added to the multifactor Cox model. The number of RCS knots (k value) was selected based on the AIC value. The AIC values ​​of all models from k = 3 to k = 8 were compared, and the k with the smallest AIC value was selected as the optimal number of knots for restricted cubic spline.

[0166] Prediction model construction: A prediction model was constructed based on the results of multivariate Cox regression analysis. The predict function was used to predict the probability of survival. The survivalROC package was used to evaluate the prediction accuracy of the model and to draw the ROC curve. The timeROC package was used to calculate the area under the curve (AUC) value based on the time-dependent ROC curve. The stdca package was used to perform decision curve analysis (DCA) to evaluate the clinical decision utility. The calibration of the nomogram was evaluated by multiple resampling using the calibrate function and the bootstrap method, and the calibration curve was drawn using the plot function. The nomogram function in the rms package was used to draw the nomogram based on the Model5 multivariate Cox model.

[0167] Online nomogram establishment: Based on the Model 5 established after screening, a dynamic nomogram is established based on DynNom. Based on the rsconnect package, the model is uploaded and deployed through shinyapp. The dynamic nomogram website after deployment is open source: https: / / 4ravewaving.shinyapps.io / TLS_Nomogram / .

[0168] like Figure 4 As shown, it is a flow chart of the study of an embodiment of the scheme of the present invention.

[0169] Results and analysis: Clinical characteristics of the study population: There were 235 patients in the SYSU cohort, all of whom were Asians, including 177 patients in the TLS(+) group, accounting for about 75%, and 58 patients in the TLS(-) group, accounting for about 25%. The differences in baseline characteristics between the TLS(+) group and the TLS(-) group are shown in Table 2.1. The results show that there were significant statistical differences between the two groups of patients in terms of clinical T stage, clinical TNM stage, pathological complete remission (pCR) rate, Ki67 expression, and disease-free survival (DFS) (p<0.05). Among them, compared with the TLS(-) group, the TLS(+) group had lower clinical T stage, lower clinical TNM stage, lower Ki67 expression, and lower disease-free survival. Surprisingly, the TLS(+) group had a lower pCR rate, which may be related to the lack of corresponding inflammatory factors after the tumor completely regressed after neoadjuvant therapy, but TLS can still be found in the sections of many pCR cases. (as shown in Table 2)

[0170] There were 147 patients in the TCGA cohort, mainly Europeans and Africans, 89 and 47 respectively, and only 8 Asians. Due to differences in age and the development of treatment, compared with the SYSU cohort, none of the TCGA cohort received neoadjuvant therapy. Since the breast cancer cohort in the TCGA database did not disclose data such as height and weight, the height and weight data of the TCGA cohort could not be obtained. When the TCGA database was established, immunohistochemistry of Ki67 in breast cancer had not been widely used, so there was a lack of Ki67-related data in the database. Therefore, the expression level of MIB1, a gene that regulates Ki67 expression, was used instead, which may cause a certain bias. As can be seen from the table, the clinical T stage of the TLS(+) group was lower than that of the TLS(-) group. (As shown in Table 2)

[0171] Since the TCGA cohort lacks many clinical indicators and all of them are people who have not received neoadjuvant chemotherapy, the data structure is not suitable as an external validation set to verify the constructed prediction model. Therefore, the TCGA cohort was selected as a sensitivity analysis cohort to verify the conclusions and risk stratification in people who have not received neoadjuvant chemotherapy to evaluate the robustness of the conclusions in the face of people who have not received neoadjuvant triple-negative breast cancer.

[0172] Table 2 Baseline clinical characteristics of SYSU cohort and TCGA cohort

[0173]

[0174]

[0175]

[0176] 1. Mean ± SD; Median [IQR]; n (%).

[0177] 2. Independent sample t test, Mann-Whitney U rank sum test, Spearson chi-square test, Fisher exact test; p < 0.05 is statistically significant and is displayed in bold.

[0178] Pathological and tertiary lymphoid structural characteristics of the study population: Based on previous studies [31,32], in addition to TLS-related spatial and maturation characteristics, TILs and LVIs were additionally evaluated for prognosis. All indicators of the SYSU cohort and some indicators of the TCGA cohort were evaluated according to the criteria in Methods 2.3.6.

[0179] In the SYSU cohort, a total of 235 patients underwent pathological section evaluation. Among them, 58 cases (24.68%) were in the TLS (-) group and 177 cases (75.32%) were in the TLS (+) group. Among them, the abundance rating of TLS in the near-tumor (T) area was grade 0, that is, no TLS, with 78 cases (33.19%), grade 1 with 43 cases (18.30%), grade 2 with 90 cases (38.30%), and grade 3 with 24 cases (10.21%); the average TLS count in the T area was 5.12; the average ratio of TLS in the T area to the total area of ​​the section was 4.73‰. The abundance rating of TLS in the far tumor (P) region was as follows: there were 160 cases (68.09%) at grade 0, 36 cases (15.32%) at grade 1, 37 cases (15.74%) at grade 2, and 2 cases (0.85%) at grade 3; the average TLS count in the P region was 1.52; the average ratio of TLS in the P region to the total area of ​​the slice was 0.78‰.

[0180] (Table 3)

[0181] In the TCGA cohort, a total of 147 patients underwent pathological section evaluation. Among them, 48 cases (32.65%) were in the TLS (-) group and 99 cases (67.35%) were in the TLS (+) group. The TLS zoning abundance rating, TLS number and vascular invasion (LVI) of this cohort were mainly evaluated. Among them, the abundance rating of TLS in the near-tumor (T) region was 51 cases (34.69%) at level 0, 24 cases (16.33%) at level 1, 58 cases (39.46%) at level 2, and 14 cases (9.52%) at level 3; the average TLS count in the T region was 5.54. The abundance rating of TLS in the far-tumor (P) region was 85 cases (57.82%) at level 0, 22 cases (14.97%) at level 1, and 40 cases (27.21%) at level 2. There was no P region TLS connected in sheets, that is, there was 0 case at level 3; the average TLS count in the P region was 2.80. In addition. Due to lack of validation of Ki67 staining, no maturity score was performed for the TCGA cohort. (Table 3)

[0182] In both cohorts, the abundance of TLS in the T region was mainly grade 0 and grade 2. There was almost no TLS with an abundance of grade 3 in the P region. There was no significant difference in TLS grouping and TLS abundance in the T region between the two cohorts. However, there were significant differences in TLS abundance and vascular infiltration in the P region. In the TCGA cohort, TLS abundance in the P region was higher and vascular infiltration was lower, which may be related to the lack of immunohistochemical verification.

[0183] In the SYSU cohort, the relationship between the spatial distribution of TLS and maturity was preliminarily explored.

[0184] At the TLS level, a total of 1570 TLS were detected in the slices of 235 patients in the SYSU cohort. Among them, there were 1213 (77.26%) in the T area and 357 (22.74%) in the P area. Among the TLS with different maturity levels, there were 589 (27.52%) E-TLS, 944 (60.13%) PFL-TLS, and 37 (2.36%) SFL-TLS. The spatial distribution of TLS was significantly correlated with the maturity level (r = 0.357, p < 0.001, as shown in Figure 2). Figure 5 Among them, SFL-TLS is relatively rare, T area TLS is mainly PFL-TLS, and P area TLS is mainly E-TLS.

[0185] At the patient and section level, among the 235 patients in the SYSU cohort, 58 (24.68%) had a TLS maturity score of 0, 65 (27.66%) had a grade 1, 98 (41.70%) had a grade 2, and 14 (5.96%) had a grade 3 (Table 2.2). The TLS maturity score also correlated with the TLS abundance level in the T and P regions (r = 0.796, p < 0.001; r = 0.279, p < 0.001; e.g. Figure 5 ). Among them, Figure 5 A. TLS level rating. The left square is the distribution area of ​​TLS, and the right square is the maturity stage of TLS. The white numbers in the square are the number of TLS. B. WSI level rating. The middle is the maturity rating, and the two sides are the TLS abundance ratings in the T and P regions. For specific rating criteria, see Method 2.4.5. r is the Spearson correlation coefficient, and p < 0.05 is statistically significant.

[0186] Table 3 Baseline characteristics of pathological evaluation of SYSU cohort and TCGA cohort

[0187]

[0188]

[0189] 1. Mean ± SD; Median [IQR]; n (%).

[0190] 2. Independent sample t test, Mann-Whitney U rank sum test, Spearson chi-square test, Fisher exact test; p < 0.05 is statistically significant and is displayed in bold.

[0191] Relationship between tertiary lymphatic structure and disease-free survival and overall survival: During the median follow-up period of 26.9 months, 76 (32.3%) of the 235 patients in the SYSU cohort who received neoadjuvant chemotherapy experienced a primary outcome event, i.e., disease progression or death, and 41 (17.4%) patients experienced a secondary outcome event, i.e., death from all causes. As shown in the KM survival curve, the TLS(+) group had a significant prognostic advantage in DFS compared with the TLS(-) group (p<0.001, as shown in the KM survival curve). Figure 6 Although there was no significant difference in survival between the two groups in terms of the secondary outcome of all-cause mortality (p = 0.11, as Figure 6 The TLS(+) group showed a better trend than the TLS(-) group (as shown in C), but the TLS(+) group still showed a better trend. However, compared with the TLS(-) group, the TLS(+) group had a better clinical TNM stage (p=0.03, Table 2.1), which may affect the prognosis to a certain extent. Therefore, multivariate analysis should be used to correct the impact of clinical TNM stage in the follow-up.

[0192] In the 147 patients in the TCGA cohort who did not receive neoadjuvant chemotherapy, during a median follow-up period of 26.3 months, 31 (21.0%) patients experienced disease progression or death, and 23 (15.6%) all-cause deaths. By comparison, it can be found that the overall survival rates of the two cohorts are roughly the same, while the disease-free survival rates are very different (Chi-square test p = 0.001). The reason is speculated to be that the TCGA-BRCA project started in 2006, when the detection technology and follow-up conditions were relatively backward, resulting in some recurrence and metastasis events in the TCGA cohort not being discovered in time. On the KM survival curve, the TLS(+) group also failed to show better disease-free survival and overall survival (such as Figure 6 Middle B, Figure 6 This may be related to the heterogeneity of TLS. Figure 6 Figure 3. Distribution of KM survival curves of disease-free survival (DFS) of TLS(+) and TLS(-) groups in the AB.SYSU cohort and TCGA cohort, and the differences between the groups were tested by log rank. Distribution of KM survival curves of overall survival (OS) of TLS(+) and TLS(-) groups in the CD.SYSU cohort and TCGA cohort, and the differences between the groups were tested by log rank.

[0193] Correlation between clinical pathological characteristics and tertiary lymphatic structural characteristics: When constructing a Cox regression model, if there is a high correlation or exact correlation between the included explanatory variables, the model estimation will be distorted. Therefore, it is necessary to conduct a correlation analysis before incorporating explanatory variables. For collinear variables, selective inclusion or merging into one variable should be adopted.

[0194] In the Spearson correlation analysis, p < 0.001 was set as a sufficiently significant collinear relationship between two variables.

[0195] Among the clinical characteristics, age was significantly correlated with menopausal status, Her2, and height; clinical T stage, clinical N stage, clinical TNM stage, and the number of positive axillary lymph nodes were significantly correlated; height was significantly correlated with weight ( Figure 7 ). By screening and merging relevant variables, age, clinical TNM stage, Her2, and Ki67 were selected in the clinical characteristics for univariate regression analysis, and height and weight data were merged into body mass index (BMI) for analysis.

[0196] In the pathological characteristics, whether it is T-region TLS or P-region TLS, the level, area ratio and count [level, per(‰), count] as the three indicators for evaluating TLS abundance all show significant correlation, which also indirectly proves that the evaluation method is reliable and repeatable. The maturity of TLS is highly correlated with the TLS abundance index of the T region, especially T level, which suggests that the TLS abundance index of the T region may be highly collinear with the maturity of TLS. In addition, the level of TIL infiltration is also highly positively correlated with the maturity of TLS and the abundance index of TLS in the T region ( Figure 7 ).exist Figure 7 Heat map of the correlation between multiple clinical pathological features and tertiary lymphatic structural features. The horizontal and vertical axes represent clinical pathological features, and the colors and circle sizes represent correlation coefficients. Red represents positive correlation, blue represents negative correlation, and the darker the color and the larger the circle, the stronger the correlation between the two. ***p<0.001, only p<0.001 is sufficiently significant and marked in the figure. TLS Mature, the maturity of TLS. T_per(‰) and P_per(‰), the percentage of T-zone TLS and P-zone TLS in the total area of ​​tissue sections. T_per(‰)=TS-TLS / S-WSI*1000; P_per(‰)=PS-TLS / S-WSI*1000. T count and P count, the number of T-zone TLS and P-zone TLS. T level and P level, the level of T-zone TLS and P-zone TLS, see Methods for details.

[0197] Linear correlation between tertiary lymphatic structural characteristics and prognosis

[0198] First, the relationship between each variable and DFS was analyzed one by one by univariate Cox regression, and the indicators related to the primary outcome DFS were preliminarily determined and included in multivariate correction.

[0199] In univariate Cox regression analysis, clinical TNM stage, Ki67 level, P-region TLS abundance index [P_per(‰) and P count] and vascular invasion LVI were significant adverse prognostic factors, while TLS maturity, T-region TLS abundance index [Tlevel, T_per(‰) and T count], pathological complete remission pCR and TIL infiltration degree were significant favorable prognostic factors ( Figure 8 ). It can be found that P level is not as good as P_per(‰) and P count in prognostic performance. This is probably because P level, as a semi-quantitative indicator, loses more information than the other two indicators during evaluation, which greatly affects its prognostic value. In the next step of analysis, in order to avoid ignoring important variables, p<0.10 was set as the entry threshold, so BMI was also included.

[0200] Subsequently, multivariate Cox regression analysis was performed to correct for confounding factors and further explore the role of TLS characteristics in prognosis. Based on the correlation analysis in 3.4 and the univariate regression analysis in the previous article, TLS maturity was selected as the main analysis variable for Model 1, the TLS area ratio in the T region and the TLS area ratio in the P region were selected as the main analysis variables for Model 2, the number of TLS in the T region and the number of TLS in the P region were selected as the main analysis variables for Model 3, and the TLS level in the T region and the number of TLS in the P region were selected as the main analysis variables for Model 4.

[0201] Five indicators, including clinical TNM stage, BMI, pCR, Ki67, and LVI, were included as covariates for correction, and the relationship between TLS characteristics and disease-free survival (DFS) and overall survival (OS) was evaluated.

[0202] In Model 1, it was found that TLS maturity was a significant favorable prognostic factor for DFS (HR 0.575, 95% CI: 0.443-0.747, p < 0.001, Fig. 9 ), but the relationship with OS was not significant (HR 0.758, 95% CI: 0.540-0.747, p = 0.109, Fig. 9 ).

[0203] In Model 2, it was determined that the T-region TLS area ratio [T_per(‰)] was not only a significant favorable prognostic factor for DFS (HR 0.851, 95% CI: 0.775-0.934, p=0.001, Fig.10 ), and was also a significant protective factor for OS (HR 0.905, 95% CI: 0.829-0.987, p = 0.024, Fig.10 ); while the P region TLS area ratio [P_per(‰)] was a significant unfavorable factor for DFS (HR 1.183, 95%CI: 1.044-1.34, p = 0.008, Fig.10 ), which may also indicate an unfavorable trend in OS, although this trend was not significant (HR 1.168, 95% CI: 0.998-1.367, p = 0.052, Fig.10 ).

[0204] In Model 3, the number of TLS in the T region (T count) was a significant protective factor for DFS (HR 0.903, 95% CI: 0.847-0.962, p = 0.002, Fig.11 ), but the relationship with OS was not significant (HR 1.041, 95% CI: 0.978-1.108, p = 0.209, Fig.11 ). The number of TLS in the P region (P count) has no significant relationship with DFS and OS.

[0205] In Model 4, for the convenience of model application, the feature combination of T region TLS level (T level) and P region TLS number (P count) was selected. T region TLS level (T level) is not only a significant favorable prognostic factor for DFS (HR0.52, 95%CI: 0.402-0.671, p<0.001, Fig.12 ), and was also a significant protective factor for OS (HR 0.681, 95% CI: 0.496-0.934, p = 0.017, Fig.12 ); the number of TLS in the P region (P count) was a significant unfavorable factor for DFS after correction (HR 1.09, 95% CI: 1.022-1.163, p = 0.009, Fig.12 ).

[0206] In summary, Model 1 and Model 2 indicate that the maturity of TLS and TLS in the proximal (T) region can indicate a more positive DFS in triple-negative breast cancer, while TLS in the distal (P) region is associated with a poor DFS prognosis in triple-negative breast cancer to a certain extent, which is consistent with the hypothesis. In order to further verify the effect of TLS regional positioning in Model 2, TLS abundance was evaluated by multiple evaluation methods, and Model 3 for counting and quantification of TLS and Model 4 for graded quantification of TLS abundance were constructed for sensitivity analysis, proving that the conclusions are still robust in different evaluation systems.

[0207] Among them, Figure 8 In the figure, the univariate Cox regression forest plot shows the univariate Cox regression results and p-values ​​of clinical characteristics, pathological characteristics, and TLS characteristics on DFS. Colors represent different hazard ratios (HR). Red represents favorable factors (HR>1), blue represents unfavorable factors (HR<1), and bold represents factors included in subsequent analysis.

[0208] exist Fig. 9 Five covariates, clinical TNM stage, BMI, pCR, Ki67, and LVI, were included for adjustment, with clinical TNM stage as a continuous variable. The relationship between TLS maturity and DFS and OS was analyzed. The forest plot shows the multivariate Cox regression results and p-values ​​for DFS and OS. The colors of TLS features represent different hazard ratios (HRs), with red representing favorable TLS features (HR>1), blue representing unfavorable TLS features (HR<1), and gray representing insignificant.

[0209] exist Fig.10 In the analysis, five covariates including clinical TNM stage, BMI, pCR, Ki67, and LVI were included for adjustment, among which clinical TNM stage was used as a continuous variable. The relationship between T-region TLS area ratio, P-region TLS area ratio and DFS and OS was analyzed. The forest plot shows the multivariate Cox regression results and p-values ​​for DFS and OS. The color of TLS features represents different hazard ratios (HR), red represents favorable TLS features (HR>1), and blue represents unfavorable TLS features (HR<1).

[0210] exist Fig.11 In the analysis, five covariates including clinical TNM stage, BMI, pCR, Ki67, and LVI were included for adjustment, among which clinical TNM stage was used as a continuous variable. The relationship between T-region TLS count, P-region TLS count and DFS and OS was analyzed. The forest plot shows the multivariate Cox regression results and p-values ​​for DFS and OS. The color of TLS features represents different hazard ratios (HR), red represents favorable TLS features (HR>1), blue represents unfavorable TLS features (HR<1), and gray represents insignificant.

[0211] exist Fig.12 Five covariates, clinical TNM stage, BMI, pCR, Ki67, and LVI, were included for adjustment, with clinical TNM stage as a continuous variable. The relationship between T-region TLS grade and P-region TLS count and DFS and OS was analyzed. The forest plot shows the multivariate Cox regression results and p-values ​​for DFS and OS. The color of TLS features represents different hazard ratios (HRs), red represents favorable TLS features (HR>1), blue represents unfavorable TLS features (HR<1), and gray represents insignificant.

[0212] Nonlinear association between tertiary lymphatic structure characteristics and prognosis: Among the characteristics of tertiary lymphatic structure (TLS), area ratio [per(‰)] and count (count) are continuous variables, and an important assumption of Cox regression analysis is that the independent variable and survival are linearly associated. We want to further study whether the association between T_per(‰) and P_per(‰), T count and P count and prognosis is nonlinear, which may help to further understand and analyze the prognostic impact of TLS localization heterogeneity.

[0213] Restricted Cubic Splines (RCS) is a regression analysis method for fitting nonlinear data. It mainly divides the independent variables into segments according to nodes in the test, transforms them into several linear segments, and then uses a cubic function to build a model in each linear segment. RCS was used to conduct a nonlinear study on the continuous characteristics of TLS and disease-free survival (DFS), including the area ratio and count of TLS. In terms of the adjustment of confounding factors, the model used in the multivariate Cox regression analysis above was used. The selection of the number of RCS knots (k value) is closely related to the results of RCS. The optimal k value was found using the principle of minimum AIC (Akaikeinformation criterion). The AIC values ​​of all models from k=3 to k=8 were compared, and the k with the smallest AIC value was selected as the optimal number of knots for restricted cubic splines.

[0214] The results of RCS analysis showed that, regardless of area ratio or count, after RCS conversion, T-region TLS still indicated a more positive DFS in triple-negative breast cancer, while P-region TLS was still associated with a poor DFS prognosis in triple-negative breast cancer (e.g. Fig.13 AD, Total p<0.05). This is consistent with the conclusion of multivariate Cox regression analysis. However, T-region TLS, including area ratio and count, has a very significant nonlinear association with DFS (such as Fig.13 As shown in A: k = 3, such as Fig.13As shown in C: k = 3, nonlinear p < 0.001). The threshold line is x = 1.5 (as shown in Fig.13 A in the figure) and x = 3.0 (as shown in Fig.13 This indicates that when the TLS area ratio in the T region is less than 1.5‰ or the count is less than 3, the increase in TLS in the T region will increase the negative prognostic effect (HR>1); only when the TLS area ratio in the T region is greater than 1.5‰ or the count is greater than 3, the increase in TLS in the T region is a prognostic protective factor (HR<1). In contrast, TLS in the P region is linearly correlated with survival ( Fig.13 B: k = 5, Fig.13 D: k=3, nonlinear p>0.05).

[0215] An interesting conclusion was reached: In this study, when there are fewer TLS in the tumor-proximal (T) area, it will become an important risk factor for prognosis. Only when there are more TLS in the tumor-proximal (T) area can it actually play a protective role. TLS in the distant (P) area has always been a risk factor. In addition, there is an interesting correlation between the threshold line T count = 3.0 and the T level, and this threshold line is completely consistent with the established definition of T level. According to the designed definition of T level, when the number of TLS in the tumor-proximal (T) area is 1-3, it is level 1, that is, when T level = 1, it is high risk, and when T level> 1, it is low risk. Among them, in Fig.13 In AD. Different models correspond to the confounding factor adjustment strategies in the previous multivariate Cox. The variables selected by different models are as follows: Based on clinical TNM stage, BMI, pCR, Ki67, and LVI, Model 2: T_per(‰) and P_per(‰); Model 3: T count and P count; Model 4: T level and P count. The y-axis is the hazard ratio (HR) of DFS, and the x-axis is the continuous feature of TLS. The yellow solid line represents the hazard ratio at different values ​​of the x-axis, and the blue dots on the yellow solid line represent the number and value of nodes (k). The light blue area represents the 95% confidence interval. The black dotted line is the reference line for HR=1. The red solid line is the threshold line when HR=1, Fig.13 The threshold line of A is x=1.5, Fig.13 The middle C threshold line is x=3.0.

[0216] Construction of a DFS prediction model for triple-negative breast cancer based on the characteristics of the three-level lymphatic structure and comparison of its performance

[0217] According to the analysis of multivariate Cox regression in the previous article, four prediction models, Model 1-Model 4, were constructed based on TLS characteristics. According to the conclusion of the previous section, T count = 3.0 may be a cutoff point with important significance for DFS prognosis. Therefore, the original four-category T level of TLS abundance rating system used in other literature was modified to obtain a two-category new T level: (1) No or low abundance (level 0) - 0 to 3 TLS were observed in the T area; (2) Medium-high abundance (level 1) - more than 3 TLS were observed in the T area. A new model, Model 5, was constructed based on the new T level replacing the T level in Model 4. Fig.14 and Fig.15 The specific variables included in each model are given in the figure legend.

[0218] In order to screen a better model, the ROC curves, DCA curves, one-year disease-free survival calibration curves, three-year disease-free survival calibration curves and C-index (Table 4) of Model 1-Model 5 were compared.

[0219] For the discrimination of the model, the ability of the model to accurately distinguish different survival states in one year and three years was quantified by time-dependent ROC curve analysis. A high AUC value indicates that the model has good discrimination for predicting survival outcomes. Comparing the AUC values ​​of different models, after using the new T level binary variable to reduce the quantification of TLS abundance in the near-tumor area, Model 5 (AUC1-Year = 0.796, AUC3-Year = 0.848) has better discrimination than Model 3 (AUC1-Year = 0.778, AUC3-Year = 0.816) using the T count variable and Model 1 (AUC1-Year = 0.791, AUC3-Year = 0.841) using the TLS maturity variable, and has little difference in discrimination compared with Model 4 (AUC1-Year = 0.798, AUC3-Year = 0.848) using the T level four-category variable.

[0220] The DCA curve evaluates the clinical utility of the model by calculating the net benefit of using the corresponding model compared to not using any model (None) or treating all patients (All) under a series of potential threshold probabilities. The DCA curves of Model 1-Model 3 intersect with the blue line at low predicted risks (approximately <0.1), indicating that the net benefit of Model 1-Model 3 is not high at low risk thresholds, and the prediction model should be avoided. Model 4-Model 5 has a high net benefit in almost the entire threshold probability range, indicating that the prediction model has a high benefit in clinical use.

[0221] The calibration curve evaluates the consistency between the survival probability predicted by the model and the survival probability actually observed in the patient. The dotted reference line is the most ideal model, that is, the predicted and actual incidence rates are exactly the same. The closer the fitting curve is to the diagonal line, the closer the predicted and actual incidence rates are, and the more consistent the model is. The bootstrap resampling method simulates the process of extracting multiple samples from the population, thereby providing a more robust estimate of the model calibration performance. Comparing the one-year survival rate and the three-year predicted rate, it can be found that Model 5 is better than Model 1-Model 4 in predictive consistency.

[0222] The consistency index (C-index) mainly measures the consistency between the risk score predicted by the model and the actual survival results. It takes into account the information of pairing, time and events between individuals and is more effective than traditional ROC analysis in evaluating the discrimination of survival prediction models. Models 1-5 all have similar C-index (0.77-0.80), indicating that the model has good discrimination (Table 4). Fig.14 A. ROC curves of Model 1-Model 5. The three-year survival prediction is marked with a solid red line, and the one-year survival prediction is marked with a solid blue line. The horizontal axis is the false positive probability (FPR, 1-Specificity), and the vertical axis is the true positive probability (TPR, Sensitivity). The AUC area of ​​​​the ROC has been marked in the figure, that is, the area under the curve and the coordinate axis. The closer the AUC is to 1, the higher the model prediction discrimination and the higher the accuracy. When AUC is less than 0.5, it means that the model's predictive ability is worse than random prediction.

[0223] B. Multivariate Cox regression decision curves for Model 1-Model 5. The horizontal axis is the threshold probability, and the vertical axis is the net benefit. The None green line means that no model is used for treatment, and the net benefit is 0. The All blue line means that all patients are treated. Fig.15This is the calibration graph of the one-year survival rate (blue line) and three-year survival rate (red line) of the multivariate Cox regression model. The calibration curve can show the relationship between the actual incidence and the predicted incidence. The horizontal axis is the predicted survival rate, the vertical axis is the actual survival rate, and the dotted diagonal line is used as the reference line, representing that the predicted probability is equal to the actual probability. The closer the fitting curve is to the diagonal line, the stronger the correspondence between the predicted incidence and the actual incidence, indicating a higher quality model.

[0224] Model 1: cTNM+BMI+Ki67+pCR+LVI+TLS Mature.

[0225] Model 2: cTNM+BMI+Ki67+pCR+LVI+T_per(‰)+P_per(‰).

[0226] Model 3: cTNM+BMI+Ki67+pCR+LVI+T count+P count.

[0227] Model 4: cTNM+BMI+Ki67+pCR+LVI+T level+P count.

[0228] Model 5: cTNM+BMI+Ki67+pCR+LVI+new T level+P count.

[0229] Table 4 C-index and standard error of each model

[0230] Model C-index SE Model1 0.78 0.03 Model2 0.79 0.02 Model 3 0.77 0.03 Model 4 0.80 0.02 Model5 0.79 0.02

[0231] Final model construction and clinical application: Considering the accuracy comparison results of the model and the convenience of model application, Model 5 was finally selected as the final prediction model. The following table shows the specific modeling coefficients of Model 5 (Table 5). Based on the 7 parameters of Model 5, with disease progression and death as target events, a Nomogram prediction model for the prognosis of triple-negative breast cancer disease-free survival (DFS) was constructed ( Fig.16 ). In the Nomogram prediction model, the corresponding score can be obtained according to each predictive factor of the case. The predicted probability corresponding to the total score obtained by adding up these scores is the probability of disease-free survival of the case in one year and three years. Fig.16A. A nomogram based on the Model 5 prediction model. Each predictive factor has a corresponding score, and the total score obtained by adding up corresponds to the one-year and three-year disease-free survival rates. B. An interactive operation interface of an online web dynamic nomogram based on a prediction model. The image shows the input content of two triple-negative breast cancer cases. C. The display interface of an online web dynamic nomogram based on a prediction model, including numerical summaries, visualization of survival rate predictions, and a graph of survival rate changes as follow-up time increases. Serial numbers 1-2 are Fig.16 One-year and three-year disease-free survival results for case 1 (black and blue segments, black survival curve); No. 3-4 is Fig.16 The one-year and three-year disease-free survival results of Case 2 (red and green line segments, blue survival curve).

[0232] Table 5 Regression coefficients of variables in Model 5 1

[0233] Model5 coef exp(coef) SE pvalue TnM 0.420 1.522 0.100 <0.001 BMI 0.079 0.079 0.030 0.009 Ki67 0.122 1.130 0.292 0.676 pCR -1.027 -1.027 0.612 0.094 LVI 1.144 3.138 0.248 <0.001 Pcount 0.082 1.085 0.033 0.013 newTlevel -1.328 0.265 0.286 <0.001

[0234] 1 coef, regression coefficient. exp(coef), risk ratio, i.e. HR value. SE, standard error.

[0235] A nomogram is a traditional visual statistical prediction model that integrates multiple predictive variables to estimate a patient's disease prognosis. However, traditional nomograms are usually static in nature, so they cannot adapt well to new data or changes. In addition, these models cannot be updated in real time, limiting their potential practical use. It is worth noting that traditional nomograms are usually in paper form or require manual reading, which may lead to user misreading or calculation errors.

[0236] The dynamic nomogram, or interactive nomogram, is a contemporary enhanced iteration of the traditional model, usually built on a web or software application. Users can enter data through simple input and click operations in an interactive interface, and the model will calculate and update the results in real time, reducing calculation errors. Therefore, the model can calculate and update results instantly, significantly reducing the possibility of calculation errors. In addition, due to its web-based or application-based nature, these dynamic nomograms can be easily accessed from any device with an internet connection, making them easy to apply in clinical settings.

[0237] The prediction model was deployed on the web using the R language DynNom package and a shinyapps.io cloud account linked to Rstudio. The model is open source and can be accessed directly from the web.

[0238] The access address is as follows: https: / / 4ravewaving.shinyapps.io / TLS_Nomogram / .

[0239] Two triple-negative breast cancer cases were included for comparison. In the traditional nomogram: Case 1 (cTNM: IIIA, 90 points; BMI: 24, 10 points; Ki67: Low, 0 points; pCR: No, 15 points; LVI: Yes, 15 points; new T level: No to Low, 20 points; P count: 2, 5 points), with a total score of 155 points, a one-year disease-free survival rate of approximately 0.65, and a three-year disease-free survival rate of approximately 0.25. Case 2 (cTNM: IIIA, 90 points; BMI: 24, 10 points; Ki67: Low, 0 points; pCR: No, 15 points; LVI: No, 0 points; new Tlevel: Medium to High, 0 points; P count: 2, 5 points), with a total score of 120 points, a one-year disease-free survival rate of approximately 0.97, and a three-year disease-free survival rate of approximately 0.8. In the dynamic nomogram, the same two case baseline characteristics are entered in the interactive interface. Serial number 1-2 indicates that the one-year disease-free survival rate of case 1 is 0.680, and the three-year disease-free survival rate is 0.226; serial number 3-4 indicates that the one-year survival rate of case 2 is 0.970, and the three-year disease-free survival rate is 0.880. In addition to being more convenient and quick, the dynamic nomogram also has a more accurate survival rate prediction value and can provide a visualization of the 95% confidence interval.

[0240] Prognostic stratification based on immune risk score and spatial classification of three-level lymphatic structure: Based on the model regression coefficient of Model 5, the immune risk score calculation formula based on TLS was obtained:

[0241] TLSRisk Score = 0.082 * P count - 1.328 * new T level (1)

[0242] Among them, new T level: none or low abundance = 0, medium and high abundance = 1.

[0243] The TLS immune risk score was calculated in the SYSU cohort and divided into a low-risk group and a high-risk group using the median as the cut-off value. In the KM survival analysis, it can be seen that significant risk stratification was achieved in the primary outcome DFS (log-rank p < 0.0001, such as Fig.17 In the secondary outcome OS, the survival difference was not significant (log-rank p = 0.091, as Fig.17 Middle C), but the low-risk group had a significantly better prognostic trend.

[0244] The TLS immune risk score was applied in the TCGA cohort that did not receive neoadjuvant therapy, and the patients were divided into low-risk group and high-risk group with the median as the cutoff value. Compared with the TLS immune risk score that was directly grouped based on the presence or absence of TLS in Section 3.3, DFS (log-rank p = 0.0009, as shown in Figure 3.3) was achieved. Fig.17 (B) and OS (log-rank p = 0.0010, as Fig.17 This demonstrates that the TLS immune risk score is robust in other heterogeneous cohorts.

[0245] In order to further verify the conclusion, TLS was classified according to its spatial and quantitative distribution. Type I: no TLS or less TLS in the T region, no TLS in the P region; Type II: more TLS in the T region, no TLS in the P region; Type III: more TLS in the T region, TLS in the P region; Type IV: no TLS or less TLS in the T region, TLS in the P region. The schematic diagram of the four TLS spatial classifications is shown below. Fig.17 shown.

[0246] KM survival analysis was performed on the four TLS spatial types based on DFS and OS in the SYSU cohort and TCGA cohort. It was found that type II and type III had better prognosis than type I and type IV. Type IV with TLS in the P region had a worse prognosis than type I without TLS ( Fig.19 ). This further verifies the conclusion.

[0247] Among them, Fig.17 In AD, the TLS immune risk scores calculated by the Model 5 regression coefficient were used for stratification. The patients were divided into low-risk group and high-risk group based on the median risk score. KM survival analysis of DFS and OS was performed in the SYSU cohort and TCGA cohort, respectively. The differences between the groups were tested by log rank.

[0248] exist Fig.18 In the spatial classification based on TLS, there are: Type I: no TLS or less TLS in the T area, and no TLS in the P area; Type II: more TLS in the T area, and no TLS in the P area; Type III: more TLS in the T area, and TLS in the P area; Type IV: no TLS or less TLS in the T area, and TLS in the P area.

[0249] According to the research, most TLS in the T region are PFL-TLS containing Ki67+B cells, and most TLS in the P region are E-TLS containing Ki67-B cells.

[0250] exist Fig.19AD was stratified according to TLS spatial typing. KM survival analysis of DFS and OS was performed in the SYSU cohort and TCGA cohort, respectively. Differences between the groups were tested by log rank.

[0251] Sensitivity analysis: In the TCGA cohort, the same nonlinear restricted cubic spline analysis based on DFS was performed on the abundance of TLS in the T region (e.g. Fig. 20 A in the figure). It was found that after incomplete adjustment according to the variables of Model 5, the TLS count in the T region had an inflection point of the HR value at x = 3.0, that is, when x < 3.0, the TLS count in the T region existed as a risk factor for DFS (HR>1). When x> 3.0, the TLS count in the T region was a protective factor for DFS, although this protective effect became insignificant after the number of TLS in the T region increased. This finding suggests that the nonlinear association between TLS in the T region and disease-free survival is robust, and the number of TLS in the T region may be a statistically significant cutoff when it is equal to 3. It further shows that TLS is associated with a more positive DFS prognosis only when there are enough TLS in the T region.

[0252] Multivariate Cox regression analysis was also performed on the variables in Model 5 (except BMI and pCR) based on DFS in the TCGA cohort ( Fig. 20 Middle B). In the TCGA cohort, the TLS abundance rating in the T region was significantly associated with a good DFS prognosis (HR=0.335, p=0.022), and the clinical TNM stage (HR=1.835, p<0.001) and vascular invasion (HR=9.415, p<0.001) were significantly associated with a poor prognosis. However, no significant association was found between the TLS count in the P region and the prognosis.

[0253] exist Fig. 20 A. Nonlinear analysis of adjusted T counts in the TCGA cohort. Due to the absence of BMI and pCR, the adjustment variables were cTNM, Ki67, and LVI. The y-axis is the hazard ratio (HR) for DFS, and the x-axis is the count of TLS in the tumor-proximal (T) area. The yellow solid line represents the hazard ratio at different values ​​of the x-axis, and the blue dots on the yellow solid line represent the number and value of nodes (k). The light blue area represents the 95% confidence interval. The black dashed line is the reference line for HR=1. The red solid line is the threshold line when HR=1, and the threshold line is x=3.0.

[0254] B. Multivariate Cox regression forest plot for disease-free survival (DFS) in the TCGA cohort. Due to the absence of BMI and pCR, the variables included were cTNM, Ki67, LVI, new T level, and P count in Model 5.

[0255] Discussion and conclusion: This study revealed the complex role of TLS in triple-negative breast cancer patients who received neoadjuvant chemotherapy based on the number and spatial location. The spatial distribution of TLS was closely related to the degree of maturity. The proximal TLS mainly showed the characteristics of PFL-TLS, while the distal TLS mainly showed the characteristics of E-TLS. This is inconsistent with the conclusion of Ding et al. in intrahepatic cholangiocarcinoma

[32] . The distal TLS also showed the characteristics of E-TLS, while the proximal TLS mainly showed the characteristics of SFL-TLS. In the study of Ding et al.

[32] , a total of 87 SFL-TLS were found, accounting for 15.2% of the total TLS (571). In the triple-negative breast cancer cohort, only 37 SFL-TLS were found, accounting for 2.4% of the total TLS (1570). In cross-cancer TLS studies

[57] , it has been found that the number and quality of TLS vary in different cancer types. Therefore, it can be speculated that compared with intrahepatic cholangiocarcinoma, the proportion of TLS that develop to full maturity in triple-negative breast cancer is relatively small, which may be related to the differences in factors that promote TLS maturation in different tumor immune microenvironments. This may suggest that promoting TLS maturation in triple-negative breast cancer could be a new immunotherapy strategy.

[0256] The abundance of mature TLS and TLS in the tumor-proximal region is an effective predictor of good DFS prognosis in triple-negative breast cancer, while the presence of TLS in the tumor-distal region is significantly associated with poor DFS prognosis. In previous studies, the dual functions of spatially different TLS have been recognized in intrahepatic cholangiocarcinoma

[32] , liver metastases of colon cancer

[31] , and renal clear cell carcinoma

[30] , but the underlying mechanism remains uncertain. Zhang et al.

[31] and Ding et al.

[32] found that TLS close to the tumor contained more CD4+Bcl6+Tfh cells than TLS far from the tumor. Tfh cells are important components for activating B cell proliferation, building germinal centers, and forming mature TLS, which help shape the anti-tumor immune microenvironment. The assessment of TLS maturity by labeling proliferating B cells with Ki67 also demonstrated that TLS close to the tumor region are mostly PFL-TLS containing more Ki67+ proliferating B cells and are associated with better prognosis. Wolf et al. reported

[66] that B cells and their functions are highly dependent on the presence and maturation of TLS, and that highly proliferative and fully mature B cells grow into plasma cells and produce high-affinity anti-tumor IgG and IgA antibodies, thereby forming an anti-cancer immune microenvironment. This may explain one of the mechanisms by which the number and spatial location of TLS near the tumor area and prognosis in the study are complex.

[0257] The relationship between TLS in distant tumor areas and adverse outcomes is consistent with previous studies in other subtypes of breast cancer

[47] . Previous studies in mouse liver cancer models

[67] reported that TLS around tumors reflect an inflammatory environment that supports tumor growth and may serve as a site for tumor progenitor cell migration. Sofopoulos et al.'s study of peritumoral TLS in breast cancer

[47] also found that many TLS far from the tumor were infiltrated by cancer cell clusters. It can be speculated that TLS far from the tumor may reflect the process of tumor cells gradually migrating from the tumor area to the external area of ​​the tumor associated with invasion.

[0258] Although the results in the SYSU cohort can indirectly verify this hypothesis, the connection between TLS near the tumor area and TLS far from the tumor area may be far more complicated than imagined. Studies in intrahepatic cholangiocarcinoma

[32] reported that as the abundance of peritumoral TLS increased, the frequency of Treg cells in intratumoral TLS increased significantly, which means that there is a potential connection between peritumoral TLS and intratumoral TLS. In addition to being immature and dysfunctional, peritumoral TLS may also have the function of destroying the anti-tumor immunity of intratumoral TLS. However, the study did not further analyze the specific cell phenotypes of TLS in different regions of triple-negative breast cancer through multiple immunofluorescence analysis, so the relevant conclusions could not be verified.

[0259] Compared with previous similar studies, a nonlinear analysis of the abundance of TLS in the proximal and distal regions was performed, and a new conclusion was drawn: in the SYSU cohort, patients with triple-negative breast cancer who received neoadjuvant chemotherapy and had a proximal region TLS abundance count threshold of more than 3.0 indicated a good prognosis. According to the commonly used TLS abundance rating criteria in the literature [30-32], low-abundance proximal region TLS may have a worse prognosis than no proximal region TLS, which also indirectly corroborates and supports the research results. A sensitivity analysis was also performed in the TCGA triple-negative cohort, confirming this conclusion, and the threshold for counting in the TCGA cohort was also 3.0. The mechanism behind this may be related to the interaction between TLS. It is speculated that there may be a positive feedback mechanism for mutually promoting maturation of proximal region TLS, and low-abundance TLS is not enough to reach the positive feedback promotion threshold of TLS.

[0260] According to the conclusions of this study, in the construction of the prediction model, the commonly used rating method in the literature was revised, and a new T level was proposed, which merged the non-near-tumor area TLS and the low-abundance near-tumor area TLS into high-risk items, and the medium-high abundance near-tumor area TLS was a low-risk item. Multiple prediction models were constructed and compared through prediction performance, and the final model (AUC1-Year = 0.796, AUC3-Year = 0.848, C-index = 0.79) was screened. The risk score of the model can also achieve risk stratification in the TCGA cohort.

[0261] Conclusions: (1) In the triple-negative breast cancer cohort receiving neoadjuvant chemotherapy, the spatial distribution of TLS was closely related to maturity. The proximal TLS mainly showed the characteristics of PFL-TLS, while the distal TLS mainly showed the characteristics of E-TLS. (2) The analysis found that the abundance of mature TLS and proximal TLS was associated with a more positive disease-free survival, while the abundance of distal TLS was associated with a poor disease-free survival prognosis. (3) The abundance of proximal TLS was significantly nonlinearly correlated with the prognosis of disease-free survival. When the proximal TLS was less (area ratio less than 1.5‰ or count less than 3), the proximal TLS was associated with a poor prognosis. Only when the proximal TLS was more, it was associated with a more positive disease-free survival. (4) Four models were constructed according to the commonly used standards in the past, and the commonly used standards were adjusted according to the conclusions of the study to construct the fifth model. After a comprehensive comparison of the performance of the five models, the best model Model 5 (AUC1-Year=0.796, AUC3-Year=0.848, C-index=0.79) was obtained. (5) Based on the optimal model, a simple and convenient online dynamic nomogram analysis tool was constructed and made open source on the Internet for use by a wide range of users.

[0262] To achieve the above object, the present invention also provides a system for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure, wherein the system is applied to the method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure, such as Fig.21 As shown, the system comprises:

[0263] A data acquisition processing unit, used to acquire breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slice and process the corresponding breast cancer tissue paraffin block according to the breast cancer tissue paraffin block data;

[0264] A first data generating unit is used to generate first-level data corresponding to TLS density and second-level data corresponding to TLS maturity based on the three-level lymphatic structure and in combination with the breast cancer tissue paraffin block data, and generate estimated data corresponding to the level data; wherein the first-level data is a semi-quantitative abundance level; and the second-level data is a maturity level;

[0265] The second data generating unit is used to construct a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, and to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time in combination with the estimated data.

[0266] The data acquisition processing unit further includes:

[0267] The first processing module is used to perform H&E staining and Ki67 immunohistochemical staining on the paraffin block of breast cancer tissue respectively, and generate corresponding KFB format files;

[0268] A second processing module is used to convert and generate corresponding pathological image data in SVS format according to the format file of KFB;

[0269] The third processing module is used to perform region delineation processing, area measurement processing and ROI region cell number calculation processing on the WSI file in the SVS format.

[0270] The second data generating unit further includes:

[0271] The first determination module is used to obtain sample data corresponding to triple-negative breast cancer after neoadjuvant therapy, and to determine and estimate whether there is a significant prognostic difference between different groups of samples;

[0272] The first data generation module is used to perform Cox proportional hazard regression analysis on each covariate separately and generate covariate data that are significantly associated with survival time;

[0273] The second data generation module is used to construct a multi-factor Cox model, and at the same time convert the continuous variables into a restricted cubic spline form and add them to the multi-factor Cox model to generate corresponding node quantity data.

[0274] The second data generating unit further includes:

[0275] The third data generation module is used to generate the relationship data between the three-level lymphatic structure and the disease-free survival and overall survival based on the prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy;

[0276] The fourth data generation module is used to obtain clinical pathological characteristic data corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate correlation data between the clinical pathological characteristic data and the tertiary lymphatic structure characteristic data;

[0277] A fifth data generation module, used to create linear correlation data between the three-level lymphatic structure characteristic data and the prognosis data, and generate corresponding correction data according to the linear correlation data; wherein the linear correlation data includes linear correlation data and nonlinear correlation data;

[0278] The second data generating unit further includes:

[0279] The sixth data generation module is used to construct a DFS prediction model for triple-negative breast cancer based on the tertiary lymphatic structure feature data, compare the performance of each prediction model, and generate model performance evaluation indicator data corresponding to the performance.

[0280] In the system solution embodiment of the present invention, the method steps involved in the prognosis prediction of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure have been described above in detail, that is, the functional modules in the system are used to implement the steps or sub-steps in the above method embodiment, which will not be repeated here.

[0281] To achieve the above objectives, the present invention also provides a prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure. Fig. 22 As shown, it includes a processor, a memory, and a control program for the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure; wherein the control program for the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure is executed by the processor, the control program for the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure is stored in the memory, and the control program for the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure implements the steps of the prognosis prediction method for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure. For example:

[0282] S01, obtaining breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slicing and processing corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data;

[0283] S02, based on the three-level lymphatic structure, combined with the breast cancer tissue paraffin block data, respectively generate first-level data corresponding to TLS density and second-level data corresponding to TLS maturity, and generate estimated data corresponding to the level data; wherein the first-level data is a semi-quantitative abundance level; the second-level data is a maturity level;

[0284] S03. Construct a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time in combination with the estimated data.

[0285] The specific details of the steps have been explained above and will not be repeated here.

[0286] In the embodiment of the present invention, the built-in processor of the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure can be composed of an integrated circuit, for example, it can be composed of a single packaged integrated circuit, or it can be composed of multiple integrated circuits with the same or different functions, including one or more central processing units (CPU), microprocessors, digital processing chips, graphics processors, and various control chips. The processor uses various interfaces and lines to connect various components, and executes or executes programs or units stored in the memory, and calls data stored in the memory to perform various functions and process data for the prognosis prediction of triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure;

[0287] The memory is used to store program codes and various data, and is installed in the neoadjuvant triple-negative breast cancer prognosis prediction platform based on the three-level lymphatic structure, and realizes high-speed and automatic access to programs or data during operation.

[0288] The memory includes read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electronically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0289] The scheme of the present invention obtains breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slices and processes corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data;

[0290] Based on the three-level lymphatic structure, combined with the breast cancer tissue paraffin block data, first-level data corresponding to TLS density and second-level data corresponding to TLS maturity are generated respectively, and estimated data corresponding to the level data are generated; wherein the first-level data is a semi-quantitative abundance level; the second-level data is a maturity level;

[0291] A prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy is constructed, and combined with the estimated data, prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy is generated in real time; and a system and platform corresponding to the method can accurately evaluate the application value of the maturity and spatial positioning characteristics of TLS in the prognosis evaluation of triple-negative breast cancer, and at the same time construct a more accurate prognosis prediction model, in order to provide a strong basis for the formulation of personalized treatment plans, thereby promoting the improvement of the survival rate and quality of life of patients with triple-negative breast cancer.

[0292] That is to say, this scheme is a method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure. By constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data, high-precision prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy is generated in real time.

[0293] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure, characterized in that: The method comprises the steps of: Acquire breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slice and process corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data; Based on the three-level lymphatic structure, combined with the breast cancer tissue paraffin block data, first-level data corresponding to TLS density and second-level data corresponding to TLS maturity are generated respectively, and estimated data corresponding to the level data are generated; wherein the first-level data is a semi-quantitative abundance level; the second-level data is a maturity level; A prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy is constructed, and the estimated data is combined to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time.

2. The method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure according to claim 1, characterized in that: The method of obtaining breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slicing and processing corresponding breast cancer tissue paraffin blocks according to the breast cancer tissue paraffin block data, further includes: Perform H&E staining and Ki67 immunohistochemical staining on the paraffin blocks of breast cancer tissues, and generate corresponding KFB format files; According to the format file of the KFB, convert and generate corresponding pathological image data in the SVS format; The WSI file in the SVS format is subjected to region delineation processing, area measurement processing and ROI region cell number calculation processing respectively.

3. The method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure according to claim 1, characterized in that: The method of constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes: Obtain sample data corresponding to triple-negative breast cancer after neoadjuvant therapy, and determine and estimate whether there are significant prognostic differences between different groups of samples; Cox proportional hazards regression analysis was performed on each covariate separately, and covariate data with significant association with survival time were generated; A multi-factor Cox model was constructed, and continuous variables were converted into restricted cubic spline form and added into the multi-factor Cox model to generate corresponding node quantity data.

4. A method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure according to claim 1 or 3, characterized in that: The method of constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes: Based on the prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, the relationship data between the three-level lymphatic structure and disease-free survival and overall survival were generated respectively; Acquire clinical pathological characteristic data corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate correlation data between the clinical pathological characteristic data and the characteristic data of tertiary lymphatic structure; Linear correlation data between the tertiary lymphatic structure characteristic data and the prognosis data are created, and corresponding correction data are generated according to the linear correlation data; wherein the linear correlation data includes linear correlation data and nonlinear correlation data.

5. The method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy based on the tertiary lymphatic structure according to claim 4, characterized in that: The method of constructing a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy and combining the estimated data to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time also includes: Based on the three-level lymphatic structure characteristic data, a DFS prediction model for triple-negative breast cancer was constructed, and the performance of each prediction model was compared to generate model performance evaluation index data corresponding to the performance.

6. A prognosis prediction system for triple-negative breast cancer after neoadjuvant therapy based on the three-level lymphatic structure, characterized in that: The system is applied to the method for predicting the prognosis of triple-negative breast cancer after neoadjuvant therapy according to any one of claims 1 to 5, and the system comprises: A data acquisition processing unit, used to acquire breast cancer tissue paraffin block data corresponding to triple-negative breast cancer after neoadjuvant therapy, and slice and process the corresponding breast cancer tissue paraffin block according to the breast cancer tissue paraffin block data; A first data generating unit is used to generate first-level data corresponding to TLS density and second-level data corresponding to TLS maturity based on the three-level lymphatic structure and in combination with the breast cancer tissue paraffin block data, and generate estimated data corresponding to the level data; wherein the first-level data is a semi-quantitative abundance level; and the second-level data is a maturity level; The second data generating unit is used to construct a prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy, and to generate prediction data corresponding to the prognosis of triple-negative breast cancer after neoadjuvant therapy in real time in combination with the estimated data.

7. The prognosis prediction system for triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure according to claim 6, characterized in that: The data acquisition processing unit further includes: The first processing module is used to perform H&E staining and Ki67 immunohistochemical staining on the paraffin block of breast cancer tissue respectively, and generate corresponding KFB format files; A second processing module is used to convert and generate corresponding pathological image data in SVS format according to the format file of KFB; The third processing module is used to perform region delineation processing, area measurement processing and ROI region cell number calculation processing on the WSI file in the SVS format.

8. The neoadjuvant triple-negative breast cancer prognosis prediction system based on the tertiary lymphatic structure according to claim 6, characterized in that: The second data generating unit further includes: The first determination module is used to obtain sample data corresponding to triple-negative breast cancer after neoadjuvant therapy, and to determine and estimate whether there is a significant prognostic difference between different groups of samples; The first data generation module is used to perform Cox proportional hazard regression analysis on each covariate separately and generate covariate data that are significantly associated with survival time; The second data generation module is used to construct a multi-factor Cox model, and at the same time convert the continuous variables into a restricted cubic spline form and add them to the multi-factor Cox model to generate corresponding node quantity data.

9. A neoadjuvant triple-negative breast cancer prognosis prediction system based on tertiary lymphatic structure according to claim 6 or 8, characterized in that: The second data generating unit further includes: The third data generation module is used to generate the relationship data between the three-level lymphatic structure and the disease-free survival and overall survival based on the prediction model corresponding to triple-negative breast cancer after neoadjuvant therapy; The fourth data generation module is used to obtain clinical pathological characteristic data corresponding to triple-negative breast cancer after neoadjuvant therapy, and generate correlation data between the clinical pathological characteristic data and the tertiary lymphatic structure characteristic data; A fifth data generation module, used to create linear correlation data between the three-level lymphatic structure characteristic data and the prognosis data, and generate corresponding correction data according to the linear correlation data; wherein the linear correlation data includes linear correlation data and nonlinear correlation data; The second data generating unit further includes: The sixth data generation module is used to construct a DFS prediction model for triple-negative breast cancer based on the tertiary lymphatic structure feature data, compare the performance of each prediction model, and generate model performance evaluation indicator data corresponding to the performance.

10. A prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on tertiary lymphatic structure, characterized in that: The invention comprises a processor, a memory and a control program of a prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure; wherein the control program of the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure is executed by the processor, the control program of the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure is stored in the memory, and the control program of the prognosis prediction platform for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure realizes the prognosis prediction method for triple-negative breast cancer after neoadjuvant therapy based on a three-level lymphatic structure as described in any one of claims 1 to 5.

Images