Auxiliary decision method and device for optimal wax block under molecular pathological examination doctor's order

Abstract

The application provides an auxiliary decision-making method and device for optimal wax blocks under molecular pathology special examination doctor's orders, and relates to the technical field of molecular pathology detection and medical artificial intelligence. The method comprises the following steps: in response to the order input operation of the user interface, obtaining the identity in the molecular pathology special examination doctor's order, and based on the identity, calling the preset information of a plurality of candidate wax blocks. For each candidate wax block, the missing data in the preset information is compensated and the wax block confidence is generated; based on the compensated data, the detection of multiple indexes is carried out, and the score of each index is generated. The compensated preset information of each candidate wax block is automatically compared, and the conflict information between each candidate wax block is generated; according to the conflict information, the wax block confidence is updated. Based on the score of each candidate wax block and the updated wax block confidence, the multiple candidate wax blocks are graded, and a detection strategy containing the main examination wax block recommendation is generated. The method is used to improve the existing wax block detection accuracy and detection efficiency.

Description

Technical Field

[0001] This application relates to the fields of molecular pathology testing and medical artificial intelligence technology, and more specifically, to an auxiliary decision-making method and device for determining the optimal paraffin block under a specific molecular pathology examination order. Background Technology

[0002] Currently, in the entire process of molecular pathology testing, it is necessary to select a suitable paraffin block from multiple paraffin-embedded blocks to be tested, and then complete subsequent analysis and other operations.

[0003] In existing technologies, wax blocks are usually selected with the assistance of AI technology to obtain wax blocks suitable for testing.

[0004] However, while existing technologies attempt to apply AI to paraffin block screening, they all remain at the level of a simple logic of "single paraffin block index scoring - fixed weight fusion - ranking recommendation". Essentially, they are just scoring and ranking existing paraffin blocks, without providing solutions to the complex pain points of real-world molecular pathology clinical scenarios, resulting in poor accuracy in paraffin block screening. Summary of the Invention

[0005] The purpose of this application is to provide an auxiliary decision-making method and device for optimal paraffin blocks under a specific medical order for molecular pathology examination, so as to solve the above-mentioned problems existing in the prior art, ensure the reliability of molecular pathology test results, realize the optimal allocation of paraffin block resources and standardized management of the testing process, and solve the problems of low main accuracy and poor testing efficiency of existing paraffin block detection.

[0006] Firstly, a method for auxiliary decision-making regarding the optimal paraffin block under a specific molecular pathology examination order is provided, applied to a pathology information system, wherein the pathology information system includes a user graphical interface; the method may include: In response to the medical order input operation of the user graphical interface, the identity identifier in the molecular pathology special examination medical order is obtained, and the preset information of multiple candidate paraffin blocks is retrieved based on the identity identifier; For each candidate wax block, the missing data in the preset information is compensated and a wax block confidence score is generated; based on the compensated data, multiple indicators are detected and scores for each indicator are generated. Automatically compare the compensated preset information of each candidate wax block to generate conflict information between each candidate wax block; update the confidence level of the wax block based on the conflict information; Based on the score of each candidate wax block and the updated wax block confidence, the multiple candidate wax blocks are classified, a detection strategy including the main inspection wax block recommendation is generated, and the detection strategy is pushed to the user graphical interface for display.

[0007] Secondly, a decision-making auxiliary device for optimizing paraffin blocks under specific molecular pathology examination orders is provided, applied to a pathology information system, wherein the pathology information system includes a user graphical interface; the device may include: The acquisition module is used to respond to the medical order input operation of the user's graphical interface, acquire the identity identifier in the molecular pathology special examination medical order, and retrieve the preset information of multiple candidate wax blocks based on the identity identifier; The detection module is used to compensate for missing data in the preset information for each candidate wax block and generate a wax block confidence score; based on the compensated data, it performs detection on multiple indicators and generates scores for each indicator. The update module is used to automatically compare the compensated preset information of each candidate wax block and generate conflict information between each candidate wax block; and update the confidence level of the wax block according to the conflict information. The generation module is used to classify multiple candidate wax blocks based on the score of each candidate wax block and the updated wax block confidence, generate a detection strategy including the main inspection wax block recommendation, and push the detection strategy to the user graphical interface for display.

[0008] This application provides an auxiliary decision-making method and apparatus for optimal paraffin blocks under a molecular pathology examination order. Responding to a user's graphical interface input of the examination order, the method obtains an identifier from the order and retrieves preset information for multiple candidate paraffin blocks based on this identifier. For each candidate paraffin block, missing data in the preset information is compensated, and a block confidence score is generated. Multiple indicators are detected based on the compensated data, and scores for each indicator are generated. The compensated preset information for each candidate paraffin block is automatically compared to generate conflict information between them. The block confidence score is updated based on the conflict information. Based on the score of each candidate paraffin block and the updated block confidence score, the multiple candidate paraffin blocks are graded, generating a detection strategy including recommendations for the primary examination block, and the detection strategy is pushed to the user's graphical interface for display. This solution breaks through the existing simple technical framework of "scoring-ranking-recommendation" and constructs a complete technical system of "full-dimensional information collection-intelligent compensation for missing data-multi-dimensional hierarchical analysis-multi-block collaborative conflict decision-making-intelligent detection strategy generation". It not only solves the inherent drawbacks of manual screening, but also addresses the core pain points in clinical scenarios such as missing wax block information, conflicting results from multiple wax blocks, and rigid detection processes. It significantly improves the standardization, accuracy, detection efficiency, and utilization rate of wax block resources in molecular pathology testing, ensures the reliability of molecular pathology test results, achieves optimal allocation of wax block resources and standardized management of detection processes, and solves the problems of low main accuracy and poor detection efficiency in existing wax block detection. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart illustrating an auxiliary decision-making method for determining the optimal paraffin block under a specific medical order for molecular pathology examination, provided in this application embodiment; Figure 2 A flowchart illustrating another auxiliary decision-making method for determining the optimal paraffin block under a specific molecular pathology examination order, provided for an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an auxiliary decision-making device for determining the optimal paraffin block under a specific medical order for molecular pathology examination, provided in an embodiment of this application. Detailed Implementation

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by those skilled in the art. The words "first," "second," and similar terms used in this application do not indicate any order, quantity, or importance, but are only used to distinguish different components. The words "comprising" or "including," etc., mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but do not exclude other elements or objects. The words "connected," "coupled," or "connected," etc., are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up," "down," "left," "right," etc., are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0012] Currently, in the entire process of molecular pathology testing, it is necessary to select a suitable paraffin block from multiple paraffin-embedded blocks to be tested, and then complete subsequent analysis and other operations.

[0013] In one example, AI technology is typically used to assist in the screening of paraffin blocks to obtain blocks suitable for testing. However, while existing technologies attempt to apply AI to paraffin block screening, they all remain at the level of a simple logic of "single paraffin block index scoring - fixed weight fusion - ranking recommendation". Essentially, this involves scoring and ranking existing paraffin blocks, without providing solutions to the complex pain points of real-world molecular pathology clinical scenarios, resulting in poor accuracy in paraffin block screening.

[0014] In one example, suitable paraffin blocks for testing are typically obtained through manual screening. The inherent drawbacks of manual screening include: high dependence on individual clinical experience, low screening efficiency, the need to verify information for each block in multi-block scenarios (time-consuming and susceptible to differences in experience and fatigue); poor screening accuracy, unable to precisely quantify core indicators such as block quality and cancer cell content, easily leading to test failure and distorted results due to inappropriate block selection; significant waste of paraffin block resources, difficulty in accurately controlling remaining usable quantities, easily resulting in duplicate sampling or excessive consumption of valuable specimens; and insufficient process standardization, with the arbitrariness of manual operation leading to inconsistent testing procedures and affecting the homogeneity of test results.

[0015] For ease of understanding, the terms used in the embodiments of this application are explained below: The auxiliary decision-making method for determining the optimal paraffin block under a specific molecular pathology examination order provided in this application can be applied to a pathology information system.

[0016] The preferred embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit this application. Furthermore, the embodiments and features in the embodiments of this application can be combined with each other without conflict.

[0017] Figure 1 This is a flowchart illustrating an auxiliary decision-making method for optimal paraffin block selection under a specific molecular pathology examination order, provided as an embodiment of this application. It is applied to a pathology information system, which includes a user graphical interface; such as... Figure 1 As shown, the method may include: Step S101: In response to the user's graphical interface medical order input operation, obtain the identity identifier in the molecular pathology special examination medical order, and retrieve the preset information of multiple candidate paraffin blocks based on the identity identifier.

[0018] For example, a molecular pathology special test order refers to an instruction issued by a user through the user graphical interface of a pathology information system (LIS) for a specific molecular pathology test. The instruction includes what test to do, what platform to use, and quality control standards.

[0019] In this step, the user issues a molecular pathology special examination order through the Pathology Information System (LIS). The system automatically extracts the core information of the order, including a unique identifier, the type of special examination item, the testing platform, the nucleic acid type requirements, the quality control thresholds for the special examination item (including the minimum percentage of tumor cells, the minimum amount of wax roll used, and the sample compliance requirements), and clinical diagnostic correlation information.

[0020] Based on the unique identifier, the system integrates the Pathology Information System (LIS), the paraffin block management system, and the pathology slide digitization system. It automatically retrieves pre-defined information for all candidate paraffin blocks corresponding to that identifier. This pre-defined information includes multiple dimensions, such as: basic paraffin block identification information (including unique identifier, storage location, sampling time and site), full lifecycle quality control information (including original total volume, used volume, usage records, cold ischemia time, fixation time, fixative type, decalcification treatment records, and slide preparation records), associated pathological information (including corresponding HE slide digital images, pathological diagnosis results, historical immunohistochemical test results, and previous molecular testing records), and historical quality control and testing data (historical test success rate and test result quality control status). Therefore, by automatically retrieving comprehensive information from candidate paraffin blocks based on a unique identifier, unified collection, standardized processing, and centralized management of multi-source data are achieved.

[0021] Step S102: For each candidate wax block, perform data compensation on the missing data in the preset information and generate the wax block confidence score; based on the compensated data, perform multi-indicator detection and generate the score for each indicator.

[0022] For example, for each candidate paraffin block, missing data in the preset information is identified, and the missing data in the preset information is compensated for, and the confidence level of the paraffin block is calibrated; the confidence level of the paraffin block refers to the confidence level obtained after initialization based on the compensation of missing data. Based on the compensated data, multiple indicators are detected to obtain the score of each indicator; the detection of multiple indicators includes the detection of slide quality, the detection of tumor cell content, the detection of resource suitability, and the detection of quality control compliance, etc.

[0023] Optionally, this step can also perform hard rule screening before data compensation, that is, complete the admission judgment of wax blocks through preset decision tree rules. Specifically, based on the hard requirements of the special inspection project type, the first round of elimination is completed through decision tree rules, including but not limited to: if the remaining usable capacity of the wax block (remaining usable capacity of the wax block = original total capacity - used capacity) < the preset minimum wax roll usage threshold corresponding to the special inspection project type, it is directly judged as unusable and excluded from the candidate wax block set; if the wax block has a clear quality control failure record, such as fixation failure, severely damaged sections that cannot be repaired, and decalcification treatment affecting nucleic acid detection, it is directly judged as unusable and excluded from the candidate wax block set; if the pathological diagnosis corresponding to the wax block clearly shows no tumor cells, it is directly judged as unusable and excluded from the candidate wax block set; finally, the admission candidate set is obtained, which includes multiple candidate wax blocks corresponding to the identity identifier. Here, the confidence level in this application refers to a value between 0 and 1, used to represent the degree of credibility, and is used as a multiplier when adjusting the weights, that is, the confidence coefficient.

[0024] Step S103: Automatically compare the compensated preset information of each candidate wax block to generate conflict information between each candidate wax block; update the confidence level of the wax block based on the conflict information.

[0025] For example, the system automatically compares the fields in the compensated preset information of each candidate paraffin block to identify conflict information between them, and updates the confidence level of the paraffin blocks based on the conflict information. Conflict information refers to the conflict between different candidate paraffin blocks on preset key indicators. For example, preset key indicators include PD-L1 expression, gene mutation results, etc. Conflict information refers to the conflict between pathological diagnosis results and historical test results (such as PD-L1 expression, gene mutation results).

[0026] Step S104: Based on the score of each candidate wax block and the updated wax block confidence, classify multiple candidate wax blocks, generate a detection strategy that includes the main inspection wax block recommendation, and push the detection strategy to the user's graphical interface for display.

[0027] For example, based on the score of each candidate wax block and the updated wax block confidence, a comprehensive fit score is calculated for each candidate wax block. According to the comprehensive fit score, multiple candidate wax blocks are classified into three levels: primary inspection wax block, verification wax block, and backup wax block. Based on these three levels, a detection strategy including recommendations for the primary inspection wax block is generated. The detection strategy includes: primary inspection wax block identification and cutting amount; preset verification trigger conditions and corresponding verification wax blocks; and generating a minimum cutting scheme for candidate wax blocks with remaining amounts below a threshold.

[0028] The method provided in this application, in response to a user's graphical interface input of a medical order, obtains an identity identifier from a molecular pathology special examination medical order and retrieves preset information for multiple candidate paraffin blocks based on this identity identifier. For each candidate paraffin block, data compensation is performed on missing data in the preset information to generate a paraffin block confidence score; multiple indicators are detected based on the compensated data to generate scores for each indicator. The compensated preset information for each candidate paraffin block is automatically compared to generate conflict information between the candidate paraffin blocks; the paraffin block confidence score is updated based on the conflict information. Based on the score of each candidate paraffin block and the updated paraffin block confidence score, the multiple candidate paraffin blocks are graded, a detection strategy including the main examination paraffin block recommendation is generated, and the detection strategy is pushed to the user's graphical interface for display. This solution breaks through the existing simple technical framework of "scoring-ranking-recommendation" and constructs a complete technical system of "full-dimensional information collection-intelligent compensation for missing data-multi-dimensional hierarchical analysis-multi-block collaborative conflict decision-making-intelligent detection strategy generation". It not only solves the inherent drawbacks of manual screening, but also addresses the core pain points in clinical scenarios such as missing wax block information, conflicting results from multiple wax blocks, and rigid detection processes. It significantly improves the standardization, accuracy, detection efficiency, and utilization rate of wax block resources in molecular pathology testing, ensures the reliability of molecular pathology test results, achieves optimal allocation of wax block resources and standardized management of detection processes, and solves the problems of low main accuracy and poor detection efficiency in existing wax block detection.

[0029] Figure 2 A flowchart illustrating another auxiliary decision-making method for optimal paraffin block selection under specific molecular pathology examination orders provided in this application is shown below. Figure 2 As shown, in this embodiment... Figure 1 Based on the embodiments, the method is described in detail below, and the method includes: Step S201: In response to the user's graphical interface medical order input operation, obtain the identity identifier in the molecular pathology special examination medical order, and retrieve the preset information of multiple candidate paraffin blocks based on the identity identifier.

[0030] In one example, the preset information includes at least: basic identification information of the paraffin block, quality control information throughout its entire life cycle, related pathological information, and historical testing data.

[0031] For example, this step is described in step S101, and will not be repeated here.

[0032] Step S202: For each candidate wax block, perform data compensation on the missing data in the preset information and generate the wax block confidence score.

[0033] In one example, S202 includes: for each candidate wax block, identifying the missing indicator type of missing data in the preset information according to preset key indicators; the missing indicator type includes supplementable missing indicators and non-supplementable missing indicators; for non-supplementable missing indicators, reducing the wax block confidence of the candidate wax block to a preset confidence threshold; for supplementable missing indicators, determining reference data according to a hierarchical supplementation strategy, performing data compensation based on the reference data, generating compensated data; and calibrating the wax block confidence based on the compensated data.

[0034] In one example, for supplementable missing indicators, reference data is determined according to a stratified supplementation strategy. Data compensation is then performed based on the reference data to generate compensated data. This includes: for supplementable missing indicators, according to the stratified supplementation strategy, the measured data of the same indicator from the same paraffin block of the same specimen is retrieved as reference data, and compensated data is generated based on the reference data; if there is no measured data of the same indicator from the same paraffin block, then a case-level statistical model of the same pathological type and the same sampling site is retrieved to generate compensated data.

[0035] For example, indicator confidence refers to the degree of credibility that the system assigns to each indicator (such as cold ischemia time, fixation time, etc.) after compensation for missing data, based on the homology of the reference data. The confidence level inferred from measured data of homologous paraffin blocks is relatively high (e.g., 0.8), while the confidence level inferred from case-level statistical models is relatively low (e.g., 0.6). Indicator confidence is used to dynamically adjust the fusion weights of each indicator in subsequent analyses.

[0036] Wax block confidence level: This represents the overall credibility of the candidate wax block and is a comprehensive confidence value. In this application, the initial value of the wax block confidence level is determined based on the comprehensive confidence levels of each indicator (e.g., taking the minimum value of all indicator confidence levels, or a weighted average value), and serves as the overall credibility of the wax block before conflict information decision-making. In the conflict information decision-making step (i.e., step S205), the wax block confidence level is updated according to the obtained benchmark information to obtain the final wax block confidence level. The final wax block confidence level is used for subsequent calculation of the comprehensive fit score.

[0037] For each candidate paraffin block, based on preset key indicators, the missing indicator type in the preset information is identified; the missing indicator types include supplementable missing indicators and non-supplementable missing indicators. Supplementable missing indicators include cold ischemia time, fixation time, and quality control parameters related to the sampling site, while non-supplementable missing indicators are core analytical carriers such as the HE slice digital image corresponding to the candidate paraffin block.

[0038] For missing indicators that cannot be supplemented, the overall confidence level of the candidate paraffin block is directly reduced to a pre-set confidence threshold to obtain the block confidence level. This candidate block is then marked as a high-risk candidate block, and a weighted penalty coefficient is applied in subsequent scoring. For missing indicators that can be supplemented, a stratified supplementation strategy is implemented. Reference data is determined, and data compensation is performed based on this reference data to generate compensated data. Specifically, measured data of the same indicator from paraffin blocks of the same specimen are prioritized as reference data. Compensated data is generated by performing statistical distribution on the reference data. If no measured data of the same indicator from paraffin blocks of the same origin is available, a case-level statistical model of the same pathological type and sampling site is used to estimate the indicator and generate compensated data.

[0039] After data compensation is completed, the confidence levels of each indicator (missing indicators can be supplemented) are simultaneously calibrated. The indicator confidence level is positively correlated with the homology of the reference data, and the corresponding confidence level is matched in subsequent scoring. Finally, an admission candidate set with complete indicator and confidence level calibration is output, which includes multiple candidate wax blocks corresponding to the identity identifier.

[0040] Step S203: Based on the compensated data, perform multiple indicator detections and generate scores for each indicator.

[0041] In one example, the detection of multiple indicators includes slide quality detection, tumor cell content detection, resource suitability detection, and quality control compliance detection. Slide quality detection includes: using a CNN model to perform image recognition on the HE slide digital image corresponding to the candidate paraffin block, outputting sub-scores for three sub-items: integrity, staining uniformity, and cell morphology clarity, and weighting the sub-scores to generate the score corresponding to slide quality. Tumor cell content detection includes: using a U-Net model to segment the regions where different types of cells are located in the HE slide digital image to generate tumor cell regions; determining the effective tumor cell percentage in the tumor cell regions and mapping it to obtain the score corresponding to the effective tumor cell percentage. Resource suitability detection includes: generating the score corresponding to resource suitability based on the remaining available capacity of the candidate paraffin block, the preset minimum detection usage threshold, and the preset reserved verification usage. Quality control compliance detection includes: generating the score corresponding to quality control compliance based on the cold ischemia time, fixation time, and preset special test item thresholds of the candidate paraffin block.

[0042] For example, for each candidate paraffin block in the admission candidate set, based on the requirements of the special inspection project type and the confidence level of the paraffin block, a hierarchical quantitative analysis of multi-dimensional indicators is completed, generating standardized scores and dynamic adaptation weights for each dimension; the detection of multiple indicators includes the detection of slide quality, tumor cell content, resource suitability, and quality control compliance; specific analysis dimensions include: 1. Intelligent Analysis of Slide Quality: Based on an improved CNN deep learning model, the digital images of HE slides corresponding to candidate paraffin blocks are analyzed to quantitatively evaluate three core sub-items: slide integrity, staining uniformity, and cell morphology clarity. Each sub-item is scored from 0 to 10, resulting in sub-scores for each of the three sub-items. A base weight is set based on the clinical importance of each sub-item. The base total score for slide quality = (sub-score for integrity × 0.4) + (sub-score for staining uniformity × 0.3) + (sub-score for cell morphology clarity × 0.3). This base total score is the score corresponding to the slide quality. Simultaneously, if local defects exist in the HE slide digital image, the impact of the defective region on the detection is marked, generating a confidence level for the slide quality index.

[0043] 2. Precise Analysis of Tumor Cell Content: Based on the U-Net image segmentation algorithm, precise segmentation of tumor cell regions, normal cell regions, and necrotic regions in the HE slice digital image is achieved. The effective tumor cell percentage is calculated as: Effective tumor cell percentage = Area of ​​effective tumor cell region / Total area of ​​effective cell region in slice × 100%. Combining the specific test item type and differentiated scoring rules, a score corresponding to the effective tumor cell percentage is obtained. The differentiated scoring rules include: For specific test items with high tumor cell content requirements (such as targeted gene detection, fusion gene detection, etc.), a tumor cell percentage ≥50% is optimal (corresponding to 10 points), 30%-50% is good (corresponding to 6-9 points), and <30% is unqualified (corresponding to 0-5 points); For specific test items with low tumor cell content requirements (such as pan-cancer screening, methylation detection), a tumor cell percentage ≥30% is qualified (corresponding to 6-10 points), and <30% is unqualified (corresponding to 0-5 points); Simultaneously, the confidence level of the tumor cell content index is adjusted by combining the previous immunohistochemical and molecular detection tumor cell content validation data of the candidate paraffin block.

[0044] 3. Analysis of Remaining Available Wax Block Quantity and Resource Suitability: Based on the remaining available capacity of candidate wax blocks, combined with the preset minimum testing threshold and preset reserved verification quantity for the special inspection project, the resource suitability score is calculated. The calculation process includes: if the remaining available capacity ≥ the total quantity required for the entire special inspection project (including the main inspection quantity + reserved verification quantity), the resource suitability score = 10 points; if the minimum testing threshold ≤ the remaining available capacity < the total quantity required for the entire process, the resource suitability score = (remaining available capacity / minimum testing threshold) × 10, with a maximum of 10 points, and is simultaneously marked as a wax block with no verification remaining quantity; if the remaining available capacity < the minimum testing threshold, the candidate wax block is directly excluded. This exclusion and screening of wax blocks is linked to the hard rule screening in step S101 above.

[0045] 4. Analysis of Quality Control Compliance of Paraffin Blocks Throughout Their Lifecycle: Based on quality control indicators such as cold ischemia time, fixation time, fixative type, and decalcification treatment of candidate paraffin blocks, and combined with the requirements for sample pretreatment for special testing projects, the quality control compliance score (0-10 points) is quantitatively evaluated; among them, for nucleic acid testing projects, deduction rules and confidence penalties are set for paraffin blocks that have exceeded the decalcification treatment or fixation time.

[0046] Optionally, the confidence level for slice quality is determined based on factors such as the presence of local defects in the image and the model's output probability. The confidence level for tumor cell content is determined based on the consistency between previous validation data and the current segmentation results. The confidence level for resource suitability is calculated based on objective usage, with a default confidence level of 1.0. The confidence level for quality control compliance is determined based on rules such as the completeness of quality control records and whether decalcification treatment was performed (e.g., 1.0 for complete records, 0.6 for decalcification treatment).

[0047] Optionally, based on the type of special inspection project and the confidence level of each dimension's indicators, the default base weights of each dimension's scores are dynamically adjusted to obtain the fusion weights. The default base weights are: tumor cell content score × 0.45, slide quality score × 0.25, quality control compliance score × 0.15, and resource suitability score × 0.15. If a certain dimension's indicator has insufficient data inference or confidence level, the weight of that dimension is reduced proportionally according to the confidence level, and the weight difference is allocated to the high-confidence dimension to ensure the reliability of the scoring results. Then, the scores of each indicator and the preset base weights (or dynamically adjusted fusion weights) can be weighted to calculate the provisional comprehensive score of each candidate wax block.

[0048] Optionally, for candidate wax blocks marked as "high-risk candidate wax blocks," the system, in addition to dynamic weight adjustment, multiplies all dimension weights of the candidate wax block by a weight penalty coefficient (e.g., preset to 0.7) to reduce the credibility of the candidate wax block in the final score. This weight penalty coefficient is multiplied by the confidence level to obtain the final actual weight coefficient.

[0049] Step S204: Automatically compare the compensated preset information of each candidate wax block to generate conflict information between each candidate wax block.

[0050] For example, the fields of the compensated preset information of each candidate paraffin block are automatically compared to generate conflict information between the candidate paraffin blocks. Optionally, the preset information of all candidate paraffin blocks is aligned, for example, the key indicator information in the preset information of all candidate paraffin blocks is aligned. The key indicator information includes pathological diagnosis results, tumor type determination, previous immunohistochemical / molecular detection results, and sampling site information. Based on the aligned information, the fields of the preset information of each candidate paraffin block are automatically compared to identify conflict information. Conflict information includes, but is not limited to: conflict in tumor pathological type diagnosis of different paraffin blocks, conflict in positive / negative results of tumor cells, conflict in expression results of previously detected biomarkers, and conflict in detection results between paraffin blocks of primary lesions and metastatic lesions.

[0051] Step S205: Update the confidence level of the wax block based on the conflict information.

[0052] In one example, S205 includes: determining the baseline information in the conflict information; and updating the wax block confidence based on the baseline information.

[0053] In one example, determining the baseline information in the conflict information includes several implementation methods: The first implementation method includes: quantifying the attribute values ​​of each candidate wax block on multiple priority criteria to obtain quantified attribute values; the multiple priority criteria include: sample representativeness criterion, detection result credibility criterion, and tumor cell content criterion; inputting the quantified attribute values ​​into a multi-criteria decision algorithm, outputting the comprehensive priority score of each candidate wax block with conflicting information, and determining the information corresponding to the candidate wax block with the highest comprehensive priority score as the benchmark information in the conflicting information.

[0054] The second implementation method includes: judging the identified conflict information step by step according to a preset priority adjustment strategy; the preset priority adjustment strategy includes multiple levels of priority rules; if a conflict information with higher priority is obtained at any level, the judgment is terminated, and the conflict information with higher priority is used as the reference information of the conflict information.

[0055] For example, for all candidate paraffin blocks in the admission candidate set, a collaborative analysis of pathological information and historical testing data of multiple paraffin blocks is performed to identify conflicting information. A preset conflict resolution strategy is then executed to determine the baseline information in the conflicting information. Based on the baseline information, the confidence level of the paraffin blocks is updated, thereby completing the preliminary adjustment of the paraffin block priority. Specifically, this includes the following multiple implementation methods: In the first implementation, the attribute values ​​of each candidate wax block are quantified according to multiple priority criteria, resulting in a quantified attribute value for each priority criterion. These multiple priority criteria include: sample representativeness criterion, detection result reliability criterion, and tumor cell content criterion. The quantified attribute values ​​are input into a multi-criteria decision algorithm, which outputs a comprehensive priority score for each candidate wax block with conflicting information. The information corresponding to the candidate wax block with the highest comprehensive priority score is determined as the baseline information in the conflict information. The multi-criteria decision algorithm can be the analytic hierarchy process (AHP) or a method for ranking approximate ideal solutions, etc., without limitation. The priority criteria also include a temporary comprehensive quality criterion, the quantified value of which is calculated based on the scores of each evaluation indicator and a preset basic weight.

[0056] In the second implementation, the identified conflicting information is graded according to a preset priority adjustment strategy. The preset priority adjustment strategy, from high to low, includes sample representativeness priority, detection result reliability priority, overall quality priority, and tumor cell content priority. The comparison strategy within each priority level is as follows: Sample representativeness priority: paraffin blocks from primary lesions have higher priority than those from metastatic lesions; paraffin blocks from surgical specimens have higher priority than those from biopsy specimens; paraffin blocks from the main tumor tissue have higher priority than those from the tumor margin. Test result reliability priority: paraffin blocks with historical validation data have higher priority than those without historical data; paraffin blocks with a history of successful testing on the same platform have higher priority than those without testing records. Overall quality priority: paraffin blocks with higher multi-dimensional comprehensive scores (i.e., provisional comprehensive scores) have higher priority than those with lower scores. Tumor cell content priority: paraffin blocks with a higher percentage of effective tumor cells have higher priority than those with a lower percentage. Based on the aforementioned priority adjustment strategy, the representativeness of the samples is compared first: if the conflicting information of two candidate wax blocks can be distinguished under this criterion (e.g., one is a primary lesion and the other is a metastatic lesion), the judgment is terminated, and the information with higher sample representativeness is used as the benchmark information. If the representativeness of the samples cannot be distinguished (e.g., both are primary lesions), the next priority criterion (i.e., the reliability of the detection result) is applied, and so on. This continues until a certain criterion can distinguish between them, or if all criteria have been compared and still cannot be distinguished (in which case the highest comprehensive score can be selected by default or a random selection can be made).

[0057] After determining the baseline information, the wax block confidence level is updated. Updating the wax block confidence level includes: increasing the confidence level of candidate wax blocks to which the baseline information belongs, and decreasing the confidence level of other candidate wax blocks with conflicting information. For example, the first method of updating the confidence level includes: for candidate wax blocks decided to belong to the baseline information: new wax block confidence level = min(original confidence level × gain coefficient, 1.0), where the gain coefficient is preset to 1.1; for other candidate wax blocks with conflicting information: new wax block confidence level = original confidence level × attenuation coefficient, where the attenuation coefficient is preset to 0.9; for candidate wax blocks without conflicting information: the wax block confidence level remains unchanged. The confidence level ranges from 0 to 1, with the initial value set based on the inferred homology, and the updated value not exceeding 1.

[0058] The second method for updating confidence levels involves adjusting the confidence level of candidate wax blocks to which the baseline information belongs based on their priority level during the decision-making process: higher priority levels result in smaller increases (because higher priority levels inherently possess high confidence); lower priority levels result in larger increases (because lower priority levels indicate significant advantages in other dimensions). For example, if the sample representativeness level wins, the wax block confidence level increases by 5%; if the detection confidence level wins, the wax block confidence level increases by 10%; if the overall quality level wins, the wax block confidence level increases by 15%; if the tumor cell content level wins, the wax block confidence level increases by 20%; and the confidence levels of rejected wax blocks are uniformly multiplied by 0.8.

[0059] Therefore, based on the above priority adjustment strategy, the confidence level of candidate wax blocks with conflicting information is recalibrated and the priority is adjusted to obtain the updated confidence level of the wax blocks. The priority adjustment is as follows: the priority of the candidate wax block to which the benchmark information belongs is adjusted to the highest; the remaining candidate wax blocks with conflicting information are sorted from high to low according to the temporary comprehensive score calculated in step S203; the candidate wax blocks without conflicting information are kept in the original order; and conflict explanations and decision-making basis are generated at the same time.

[0060] Optionally, for special scenarios where the same patient has multiple types of paraffin blocks, such as surgical specimen paraffin blocks, puncture biopsy paraffin blocks, and paraffin blocks of recurrent / metastatic lesions, the paraffin blocks can be collaboratively classified in conjunction with the detection objectives of special examination items. Paraffin blocks that are suitable for the main examination and paraffin blocks that are suitable for disease progression comparison can be marked, providing a basis for the generation of subsequent detection strategies.

[0061] Step S206: Based on the score of each candidate wax block and the updated wax block confidence, classify multiple candidate wax blocks and generate a detection strategy that includes recommendations for the main inspection wax block.

[0062] In one example, S206 includes: determining the comprehensive fit score of each candidate wax block based on the score of each candidate wax block and the updated wax block confidence; classifying multiple candidate wax blocks according to the comprehensive fit score and generating a detection strategy that includes the main inspection wax block recommendation; the detection strategy includes: the main inspection wax block identifier and the cutting amount; preset review trigger conditions and the corresponding review wax blocks used; and generating a minimum cutting scheme for candidate wax blocks with remaining amounts below a threshold.

[0063] For example, by combining the multi-dimensional analysis results of a single paraffin block in step S203 (including scores for slice quality, tumor cell content, resource suitability, and quality control compliance) and the collaborative decision-making results of multiple paraffin blocks (including baseline information, updated paraffin block confidence, and priority ranking of candidate paraffin blocks), a comprehensive score and ranking of the suitability of candidate paraffin blocks is completed. This overcomes the limitations of single paraffin block recommendations and generates intelligent detection strategies adapted to specific testing projects and actual testing scenarios, specifically including: 1. Fit score calculation: For each candidate wax block, the scores of each dimension are weighted and fused based on the dynamic fit fusion weight. Combined with the updated wax block confidence, the final fit score (0-10 points) is calculated. The higher the score, the stronger the fit between the candidate wax block and this special inspection. 2. Wax block stratification and grading: The final comprehensive suitability scores are sorted from high to low, and the candidate wax blocks are divided into three levels: (1) Main inspection wax block: The candidate wax block with the highest comprehensive score is the priority recommended wax block for testing; (2) Verification wax block: The candidate wax block with the second highest comprehensive score, which has sufficient remaining available quantity and has information complementary or conflicting verification value with the main inspection wax block, is used for intelligent trigger verification testing; (3) Backup wax block: The candidate wax block with a comprehensive score that meets the standard, but whose suitability is weaker than the main inspection and verification wax blocks, is used as a backup option for abnormal scenarios; 3. Intelligent Detection Strategy Generation: Combining special examination requirements, paraffin block stratification results, and clinical diagnostic information, a differentiated full-process detection strategy is generated, including: Basic Detection Strategy: Clarifying the requirements for the primary paraffin block extraction, quality control points, and detection process nodes; Intelligent Review Trigger Rules: Presetting review trigger conditions, including but not limited to: conflict between the primary paraffin block test results and the preliminary clinical diagnosis, unqualified quality control of the primary paraffin block test, and the actual value of the tumor cell percentage in the primary paraffin block being lower than the estimated value; Clarifying the priority paraffin blocks and testing requirements for triggering review; Special Scenario Adaptation Strategy: For cases with conflicting results, generating a joint detection strategy of "primary examination + parallel review"; For special examinations related to disease progression monitoring, generating a comparative detection strategy of "primary lesion + metastatic lesion"; Paraffin Block Resource Protection Strategy: For precious specimens with limited remaining paraffin blocks, generating a minimum usage extraction plan and paraffin block retention suggestions to avoid excessive consumption of paraffin block resources. It should be noted that the values ​​in this embodiment are for illustrative purposes only and are not intended to be limiting.

[0064] Step S207: Push the detection strategy to the user's graphical interface for display.

[0065] For example, the detection strategy is pushed to the user's graphical interface for display. The system will simultaneously push the generated optimal wax block information, layered wax block details, and intelligent detection strategy to the corresponding terminal devices or user graphical interfaces throughout the entire process, realizing intelligent prompts and execution control throughout the entire process, specifically including: 1. Intelligent prompts on the operating end: The terminal devices (slicer display screen and mobile terminal) in the wax block storage area and the slicing operation area automatically pop up prompt information, including the unique identifier of the main inspection wax block, storage location, comprehensive compatibility score, detailed analysis of each dimension, recommendation reasons, and cutting dosage requirements; the core information and suitable scenarios of the review wax block and the spare wax block are displayed simultaneously for operators to view and select; 2. Synchronization of detection process strategy: The verification trigger rules and quality control requirements in the intelligent detection strategy are synchronized to the molecular detection laboratory system and LIS system. Automatic prompts and verifications are triggered at the corresponding nodes of the detection process. For example, after nucleic acid extraction is completed, the measured value of the tumor cell percentage is automatically verified. If it is lower than the threshold, the detection prompt of the paraffin block is automatically triggered. 3. Operation permissions and feedback entry: The terminal device provides an operation entry point. Operators can select the wax block and detection strategy recommended by the system, or manually adjust the selection. At the same time, they need to fill in the reason for the manual adjustment to realize the implementation of the strategy and the real-time collection of operation data, ensuring the traceability of the operation.

[0066] Optionally, full-link feedback and closed-loop iterative optimization are performed. The system automatically collects operational data, detection result data, and feedback data throughout the entire process to construct a closed-loop optimization dataset, continuously optimizing the system's decision rules and AI model accuracy, specifically including: (1) Full-link data collection: The collected data includes, but is not limited to: wax block AI analysis data, wax blocks and detection strategies recommended by the system, operator selection results and reasons for adjustment, measured indicators after wax block cutting, quality control data of the detection process, final detection results, detection success rate, and clinical diagnosis verification of the detection results; (2). AI model iterative optimization: The labeled slice images, tumor cell segmentation data, and detection result verification data are used as training samples to continuously optimize the accuracy and generalization ability of the slice quality analysis model and the tumor cell content segmentation model; (3). Optimization of decision rules and weight system: Based on the manual adjustment data of operators and the attribution data of success / failure of detection, the pre-admission decision tree rules, hierarchical inference strategy, preset priority adjustment strategy of conflict information, and dynamic weight adaptation model are continuously optimized to make the decision results of the system more in line with the actual clinical needs. (4) Detection strategy optimization: Based on the verification trigger data of the detection process and the consistency data between the detection results and the clinical diagnosis, continuously optimize the generation rules and verification trigger threshold of the intelligent detection strategy to improve the intelligence and reliability of the entire detection process.

[0067] (5) Different users are assigned hierarchical operation permissions for different roles in the system. Among them, pathologists only have the permission to issue medical orders and view the basic information and decision results of paraffin blocks; molecular physicians and testing operators have the permission to select paraffin blocks, view testing strategies, and provide operation feedback; system administrators have the permission to configure system parameters, adjust rules and weights, manage data, and allocate permissions. They also complete data encryption, operation traceability, and audit log management to ensure system data security and operational compliance.

[0068] The method provided in this application, in response to a user's graphical interface input of a medical order, obtains an identity identifier from a molecular pathology special examination medical order and retrieves preset information for multiple candidate paraffin blocks based on this identity identifier. For each candidate paraffin block, missing data in the preset information is compensated, and a block confidence score is generated. Multiple indicators are detected based on the compensated data, and scores for each indicator are generated. The compensated preset information for each candidate paraffin block is automatically compared to generate conflict information between the candidate blocks. The block confidence score is updated based on the conflict information. Based on the score of each candidate paraffin block and the updated block confidence score, the multiple candidate paraffin blocks are graded, and a detection strategy including recommendations for the primary examination block is generated. The detection strategy is then pushed to the user's graphical interface for display. This solution breaks through the existing simple technical framework of "scoring-ranking-recommendation" and constructs a complete technical system of "full-dimensional information collection-intelligent compensation for missing data-multi-dimensional hierarchical analysis-multi-block collaborative conflict decision-making-intelligent detection strategy generation". It not only solves the inherent drawbacks of manual screening, but also addresses the core pain points in clinical scenarios such as missing wax block information, conflicting results from multiple wax blocks, and rigid detection processes. It significantly improves the standardization, accuracy, detection efficiency, and utilization rate of wax block resources in molecular pathology testing, ensures the reliability of molecular pathology test results, achieves optimal allocation of wax block resources and standardized management of detection processes, and solves the problems of low main accuracy and poor detection efficiency in existing wax block detection.

[0069] In one embodiment, the present application is further described in detail.

[0070] This embodiment takes the targeted gene detection (NGS detection with high tumor cell content) of lung adenocarcinoma patients after surgery as an example to explain in detail the specific implementation process of this application. This embodiment covers common and complex clinical scenarios such as missing paraffin block data and differences in pathological results of multiple paraffin blocks, which can fully demonstrate the technical advantages of this application.

[0071] 1. Medical order issuance and comprehensive information collection The pathologist issued a lung cancer targeted gene testing order for patient XXX (patient ID: ID2026001) through the LIS system. The system automatically extracted the core information of the order: the special test item is NGS testing related to lung cancer targeted drugs, requiring a minimum tumor cell percentage of ≥30%, and an optimal percentage of ≥50%. The minimum wax roll dosage for a single sample test is 12mg, and it is recommended to reserve 10mg for verification. The testing platform is the Illumina sequencing platform. The initial clinical diagnosis is postoperative lung adenocarcinoma, and it is necessary to clarify the targeted gene mutation status.

[0072] Based on a unique identifier, the system integrates the LIS system, paraffin block management system, and pathology slide digitization system to automatically retrieve five candidate paraffin blocks for the patient (unique identifiers for each block: L01, L02, L03, L04, and L05). It collects comprehensive information (i.e., preset information) for each candidate block and simultaneously retrieves the patient's clinical diagnosis and previous testing information. The core information is as follows: (1) L01: The primary lesion paraffin block of the surgical specimen, with an original total volume of 50mg, 10mg used, and 40mg remaining; the specimen was taken from the main body of the tumor, with a cold ischemia time of 45min and fixation in 10% neutral formalin for 12h, without decalcification treatment; the corresponding HE section was intact, and the pathological diagnosis was invasive lung adenocarcinoma, mainly of the glandular type; there were no previous test records. (2) L02: Paraffin block of the primary lesion of the surgical specimen, original total volume 50mg, 8mg used, 42mg remaining; sample taken from the edge of the tumor, cold ischemia time and fixation time records are missing; corresponding HE section is complete, pathological diagnosis is lung adenocarcinoma with a small amount of necrosis; no previous test records; (3) L03: surgical specimen hilar lymph node paraffin block, original total volume 50mg, 35mg used, 15mg remaining; cold ischemia time 50min, fixation 14h; corresponding HE section intact, pathological diagnosis of adenocarcinoma metastasis in lymph nodes; no previous test records; (4) L04: Preoperative puncture biopsy paraffin block, original total volume 30mg, 25mg used, 5mg remaining; cold ischemia time 30min, fixation 8h; pathological diagnosis: adenocarcinoma component found in puncture tissue; (5) L05: Surgical specimen: normal lung tissue paraffin block, original total volume 50mg, 5mg used, 45mg remaining; pathological diagnosis: normal lung tissue, no tumor cells found.

[0073] 2. Pre-admission decision-making for candidate wax blocks and intelligent compensation for missing data The system performs pre-access decision-making and data preprocessing, the specific process of which is as follows: (1) Hard rule admission exclusion: L04 has a remaining available capacity of 5mg, which is less than the minimum dosage of 12mg for this test, so it is directly excluded; L05 has no tumor cells in the pathological diagnosis, so it is directly excluded; the final admission candidate set is L01, L02, and L03; (2) Identification and compensation of missing data: The key quality control indicators of cold ischemia time and fixation time of L02 were identified as missing, which are inferred missing indicators; the cold ischemia time and fixation time data of the same paraffin block L01 of the same surgical specimen were retrieved first, and combined with the overall quality control record of the surgical specimen, it was inferred that the cold ischemia time of L02 was 40-50 min and the fixation time was 10-14 h, which met the quality control requirements of NGS detection. At the same time, the confidence coefficient of the inferred indicator was calibrated to be 0.8. (3) Confidence rating: L01 and L03 have all core indicators complete, with a confidence coefficient of 1.0; L02 has inferred indicators, with an overall confidence coefficient of 0.9.

[0074] 3. Multi-dimensional hierarchical quantitative analysis and dynamic weight adaptation For the three admission paraffin blocks L01, L02, and L03, the system completed multi-dimensional quantitative analysis and dynamic weight adaptation. This special test was a targeted gene detection, with the basic weights set as follows: tumor cell content 0.45, slide quality 0.25, quality control compliance 0.15, and resource suitability 0.15. The specific analysis results are as follows: (1) Analysis of slice quality: L01: Integrity 10 points, staining uniformity 9 points, cell morphology clarity 9 points, total score = 10×0.4+9×0.3+9×0.3=9.4 points, confidence level 1.0; L02: Integrity 9 points, staining uniformity 9 points, cell morphology clarity 8 points, total score = 9×0.4+9×0.3+8×0.3=8.7 points, confidence level 1.0; L03: Integrity 10 points, staining uniformity 9 points, cell morphology clarity 9 points, total score 9.4 points, confidence level 1.0; (2) Analysis of tumor cell content: AI image segmentation calculation shows that L01 has an effective tumor cell ratio of 68%, which meets the optimal standard for high-requirement special examination, scoring 10 points with a confidence level of 1.0. L02 has an effective tumor cell rate of 42%, which is in the good range, scoring 8 points with a confidence level of 1.0; L03 has an effective tumor cell ratio of 55%, which meets the optimal standard, and scores 9.5 points with a confidence level of 1.0. (3) Resource adaptability analysis: The total dosage used in this test (main inspection + verification) is 22mg. L01 has 40mg remaining, which is ≥22mg, so it gets 10 points; L02 has 42mg remaining, which is ≥22mg, so it gets 10 points; L03 has 15mg remaining, which is ≥12mg, the minimum dosage, but <22mg of the total dosage. The score is (15 / 12)×10=10 points, and it is marked as a wax block with no remaining amount for verification. (4) Quality control compliance analysis: L01 cold ischemia and fixation time both met the NGS quality control requirements, with no violations handled, scoring 10 points with a confidence level of 1.0; The quality control indicators inferred by L02 meet the requirements. Based on data inference, the confidence level is 0.8, and the final score is 10 × 0.8 = 8 points. L03 meets the quality control requirements, scoring 10 points with a confidence level of 1.0; (5) Dynamic weight adaptation: Among the three wax blocks, only the quality control compliance indicator of L02 has insufficient confidence. The weight of this dimension is reduced from 0.15 to 0.12. The reduced weight of 0.03 is proportionally allocated to the tumor cell content (0.015) and slice quality (0.015) dimensions to ensure the weight ratio of high confidence indicators.

[0075] 4. Collaborative analysis of multiple wax blocks and intelligent decision-making based on conflict outcomes The system performs collaborative analysis on the information from the three paraffin blocks. The pathological diagnosis of all three blocks was adenocarcinoma, with no conflicting results. Based on the representativeness priority of the samples, L01 and L02 are paraffin blocks from the primary lesion, with higher priority than L03, the paraffin block from the lymph node metastasis. Among them, L01 is a sample taken from the main tumor, with higher priority than L02, which is a sample taken from the tumor margin. The pre-calibration of the paraffin block priority is completed, with no conflicting information, and no need to adjust the confidence level.

[0076] 5. Generation of comprehensive decision-making and intelligent detection strategies for wax block compatibility. (1) Calculation of overall fit score: L01 (weights 0.45 / 0.25 / 0.15 / 0.15): 9.4×0.25 + 10×0.45 + 10×0.15 + 10×0.15 = 2.35 + 4.5 + 1.5 + 1.5 = 9.85 points; L02 (weights 0.465 / 0.265 / 0.12 / 0.15): 8.7×0.265 + 8×0.465 + 8×0.12 + 10×0.15 ≈ 2.31 + 3.72 + 0.96 + 1.5 = 8.49 points; L03 (weights 0.45 / 0.25 / 0.15 / 0.15): 9.4×0.25 + 9.5×0.45 + 10×0.15 + 10×0.15 = 2.35 + 4.275 + 1.5 + 1.5 = 9.625 points; (2) Wax blocks are classified and graded: according to the comprehensive score, L01 (9.85 points) is the main inspection wax block, L03 (9.625 points) is the verification wax block, and L02 (8.49 points) is the backup wax block; (3) Intelligent detection strategy generation: Basic testing strategy: We recommend using paraffin block L01 as the main test block, with a paraffin roll volume of 12mg, leaving room for verification. This paraffin block has excellent slice quality, sufficient tumor cell content, compliant quality control, and sufficient remaining volume, which is fully suitable for the needs of this targeted gene testing. Intelligent review triggering rules: The review test will be automatically triggered when the following situations occur, and L03 paraffin blocks will be used first: ① The actual proportion of tumor cells after the main inspection paraffin block is cut is <30%; ② The quality control of nucleic acid extraction is unqualified, and the concentration / purity does not meet the standards; ③ The test results conflict with the clinical pathological diagnosis, and there are no clear target-related variants. Wax block resource protection strategy: L03 has only 15mg remaining and is only used for verification testing. It is not recommended to use it as the primary test wax block to avoid test failure due to insufficient remaining amount; L02 is used as a backup wax block and is only used when L01 and L03 cannot complete the test. Quality control tips: Pay special attention to the tumor cell area of ​​the L01 paraffin block, avoiding necrotic and normal cell areas to ensure the quality of the test sample.

[0077] 6. Intelligent prompts and policy execution control throughout the entire process When a molecular doctor retrieves a paraffin block from the paraffin storage area, a prompt message automatically appears on the terminal display screen in the slide operation area. The core content includes: (1) [Chief Inspector Recommended Paraffin Block] L01, storage location: Pathology Library A, Shelf 05, Floor 03, No. 12, overall fit score 9.85.

[0078] Key indicators: slice quality 9.4 points, tumor cell percentage 68%, remaining usable volume 40mg, quality control compliance 10 points.

[0079] Recommendation reason: The paraffin block is taken from the main tumor site, the slice quality is excellent, the tumor cell content far exceeds the requirements of this test, the paraffin block quality control process is compliant, the remaining volume can meet the needs of the main examiner and the reviewer, and it has the highest compatibility with this targeted gene test.

[0080] Cutting requirements: It is recommended to cut 12mg of wax roll, and prioritize areas with dense tumor cells.

[0081] (2) [Review Paraffin Block] L03, with an overall fit score of 9.625, is a paraffin block for lymph node metastasis, suitable for review testing scenarios.

[0082] (3) [Spare wax block] L02, with an overall compatibility score of 8.49, is a backup option.

[0083] (4) [Intelligent detection strategy] This test is set with a pre-defined review trigger rule. When the proportion of tumor cells is insufficient, the nucleic acid quality control is unqualified, or the test results conflict with the clinical diagnosis, the system will automatically prompt to activate the review wax block L03.

[0084] At the same time, the system will synchronize the verification trigger rules to the LIS system and the molecular detection laboratory system, and automatically complete the verification and prompts at the corresponding nodes of the detection process.

[0085] 7. End-to-end feedback and closed-loop iterative optimization In this test, the molecular physician selected the L01 paraffin block recommended by the system for extraction and testing. Nucleic acid extraction quality control was passed, and sequencing was successful, detecting an EGFR exon 19 deletion mutation, consistent with the clinical pathological diagnosis. The test results were adopted by clinical practice. The system automatically collected data from the entire process, including AI analysis results, paraffin block selection results, testing quality control data, and clinical validation results, incorporating them into the training dataset for subsequent optimization and iteration of the AI ​​model and decision rules.

[0086] Therefore, by constructing a complete technical system of "full-dimensional information collection - intelligent compensation for missing data - multi-dimensional hierarchical analysis - multi-block collaborative conflict decision-making - intelligent detection strategy generation - full-process dynamic management and control - full-link closed-loop optimization", we can achieve full-process intelligentization of wax block screening, detection strategy generation, and result verification, and upgrade from "wax block recommendation" to "full-process detection decision-making". This not only solves the inherent drawbacks of manual screening, but also fundamentally addresses the problems of existing AI solutions lacking scenario-based decision-making capabilities and being unable to cope with complex clinical pain points. It significantly improves the creativity and clinical applicability of the solution, ensures the reliability of molecular pathology test results, and achieves optimal allocation of wax block resources and standardized management and control of the detection process.

[0087] Corresponding to the above method, this application also provides an auxiliary decision-making device for determining the optimal paraffin block under a molecular pathology examination order, applied to a pathology information system, which includes a user graphical interface; such as Figure 3 As shown, the device includes: The acquisition module 41 is used to respond to the user's graphical interface medical order input operation, acquire the identity identifier in the molecular pathology special examination medical order, and retrieve the preset information of multiple candidate wax blocks based on the identity identifier; The detection module 42 is used to perform data compensation on the missing data in the preset information for each candidate wax block and generate the wax block confidence score; based on the compensated data, it performs detection on multiple indicators and generates scores for each indicator. The update module 43 is used to automatically compare the compensated preset information of each candidate wax block and generate conflict information between each candidate wax block; and update the confidence of the wax block based on the conflict information. The generation module 44 is used to classify multiple candidate wax blocks based on the score of each candidate wax block and the updated wax block confidence, generate a detection strategy including the main inspection wax block recommendation, and push the detection strategy to the user's graphical interface for display.

[0088] The functions of each unit in the auxiliary decision-making device for optimal paraffin block under specific medical orders for molecular pathology examinations provided in the above embodiments of this application can be realized through the above-described methods and steps. Therefore, the specific working process and beneficial effects of each unit in the auxiliary decision-making device for optimal paraffin block under specific medical orders for molecular pathology examinations provided in the embodiments of this application will not be repeated here.

[0089] Those skilled in the art will understand that the embodiments in this application can be provided as methods, systems, or computer program products. Therefore, the embodiments in this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments in this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This application describes embodiments of methods, apparatus (systems), and computer program products according to embodiments of this application with reference to flowchart illustrations and / or block diagrams. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] Although preferred embodiments have been described in this application, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of this application.

[0094] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims in this application and their equivalents, then this application also intends to include these modifications and variations.

Claims

1. A method for auxiliary decision-making regarding the optimal paraffin block under a specific molecular pathology examination order, characterized in that, The method is applied to a pathology information system, wherein the pathology information system includes a user graphical interface; the method includes: In response to the medical order input operation of the user graphical interface, the identity identifier in the molecular pathology special examination medical order is obtained, and the preset information of multiple candidate paraffin blocks is retrieved based on the identity identifier; For each candidate wax block, the missing data in the preset information is compensated and a wax block confidence score is generated; based on the compensated data, multiple indicators are detected and scores for each indicator are generated. Automatically compare the compensated preset information of each candidate wax block to generate conflict information between each candidate wax block; update the confidence level of the wax block based on the conflict information; Based on the score of each candidate wax block and the updated wax block confidence, the multiple candidate wax blocks are classified, a detection strategy including the main inspection wax block recommendation is generated, and the detection strategy is pushed to the user graphical interface for display.

2. The method as described in claim 1, characterized in that, Based on the conflict information, update the confidence level of the wax block, including: Determine the baseline information in the conflict information; The confidence level of the wax block is updated based on the baseline information.

3. The method as described in claim 2, characterized in that, Determining the baseline information in the conflict information includes: The attribute values ​​of each candidate wax block are quantified according to multiple priority criteria to obtain quantified attribute values; the multiple priority criteria include: sample representativeness criterion, test result reliability criterion, and tumor cell content criterion. The quantified attribute values ​​are input into a multi-criteria decision algorithm, which outputs the comprehensive priority score of each candidate wax block with conflict information. The information corresponding to the candidate wax block with the highest comprehensive priority score is determined as the benchmark information in the conflict information.

4. The method as described in claim 2, characterized in that, Determining the baseline information in the conflict information includes: According to the preset priority adjustment strategy, the identified conflict information is judged step by step; the preset priority adjustment strategy includes multiple levels of priority rules. If a higher-priority conflict is obtained at any level, the decision is terminated, and the higher-priority conflict is used as the reference information for that conflict.

5. The method as described in claim 1, characterized in that, For each candidate wax block, data compensation is performed on the missing data in the preset information, and a wax block confidence score is generated, including: For each candidate wax block, the missing indicator type of the missing data in the preset information is identified according to preset key indicators; the missing indicator type includes missing indicators that can be supplemented and missing indicators that cannot be supplemented. For the aforementioned unreplaceable missing indicators, the confidence level of the candidate wax blocks is reduced to a preset confidence threshold. For the missing indicators that can be supplemented, reference data is determined according to the hierarchical supplementation strategy, data compensation is performed based on the reference data to generate compensated data, and the confidence level of the wax block is calibrated based on the compensated data.

6. The method as described in claim 5, characterized in that, For the missing indicators that can be supplemented, reference data is determined according to a hierarchical supplementation strategy. Data compensation is then performed based on the reference data to generate compensated data, including: For the missing indicators that can be supplemented, according to the hierarchical supplementation strategy, the measured data of the same indicator of the same paraffin block of the same specimen are retrieved as reference data, and the compensated data is generated based on the reference data. If there is no measured data of the same indicators for homologous paraffin blocks, then retrieve the case-level statistical model of the same pathological type and the same sampling site to generate compensated data.

7. The method as described in claim 1, characterized in that, Based on the score and updated confidence level of each candidate wax block, the candidate wax blocks are graded to generate a detection strategy that includes recommendations for the master wax block, including: Based on the score of each candidate wax block and the updated wax block confidence, a comprehensive fit score is determined for each candidate wax block. Based on the comprehensive fit score, the candidate wax blocks are classified and a detection strategy including the main inspection wax block recommendation is generated. The detection strategy includes: identifying the main inspection wax block and the cutting amount; pre-setting the verification trigger conditions and the corresponding verification wax blocks; and generating a minimum cutting scheme for candidate wax blocks with remaining amounts below a threshold.

8. The method as described in claim 1, characterized in that, The testing of multiple indicators includes the detection of slide quality, tumor cell content, resource suitability, and quality control compliance. The detection of the slice quality includes: using a CNN model to perform image recognition on the digital image of the HE slice corresponding to the candidate paraffin block, outputting sub-scores for three sub-items: integrity, staining uniformity, and cell morphology clarity, and weighting the sub-scores to generate a score corresponding to the slice quality; The detection of tumor cell content includes: segmenting the regions where different types of cells are located in the HE slice digital image using the U-Net model to generate tumor cell regions; determining the effective tumor cell percentage of the tumor cell regions and mapping the score corresponding to the effective tumor cell percentage; The resource adaptability detection includes: generating a score corresponding to the resource adaptability based on the remaining available capacity of the candidate wax block, a preset minimum detection threshold, and a preset reserved verification amount; The quality control compliance detection includes: generating a score corresponding to the quality control compliance based on the cold ischemia time, fixed time, and preset special inspection item thresholds of the candidate wax block.

9. The method according to any one of claims 1-8, characterized in that, The preset information includes at least: basic identification information of the paraffin block, quality control information throughout its entire life cycle, related pathological information, and historical testing data.

10. A decision-making device for determining the optimal paraffin block under a specific molecular pathology examination order, characterized in that, The device is applied to a pathology information system, which includes a user graphical interface; the device includes: The acquisition module is used to respond to the medical order input operation of the user's graphical interface, acquire the identity identifier in the molecular pathology special examination medical order, and retrieve the preset information of multiple candidate wax blocks based on the identity identifier; The detection module is used to compensate for missing data in the preset information for each candidate wax block and generate a wax block confidence score; based on the compensated data, it performs detection on multiple indicators and generates scores for each indicator. The update module is used to automatically compare the compensated preset information of each candidate wax block and generate conflict information between each candidate wax block; and update the confidence level of the wax block according to the conflict information. The generation module is used to classify multiple candidate wax blocks based on the score of each candidate wax block and the updated wax block confidence, generate a detection strategy including the main inspection wax block recommendation, and push the detection strategy to the user graphical interface for display.

Images