A medical report data management analysis method and system for hematopathy

Abstract

This invention discloses a method and system for managing and analyzing medical report data in hematological diseases, belonging to the field of medical information processing technology. The system includes: S1, data acquisition and fingerprint generation; S2, time-series reconstruction and correction; S3, semantic parsing and standardization; S4, feature coupling and secure computation; and S5, report generation and audit updates. This invention utilizes a time-frequency domain dual-domain correction technique, which effectively eliminates periodic errors and fluctuations between different data sources and different acquisition time points, improving the accuracy and consistency of data processing. This correction method goes beyond traditional time alignment; it identifies and removes periodic interference in the signal through frequency domain analysis, ensuring time-series consistency between different sources. This ensures that even with multi-source data, the analysis results maintain high accuracy and reliability, avoiding erroneous diagnoses and recommendations due to data errors, thus providing a more accurate and reliable basis for clinical decision-making.

Description

Technical Field

[0001] This invention relates to the field of medical information processing technology, specifically to a method and system for managing and analyzing medical report data for hematological diseases. Background Technology

[0002] With the continuous development of medical technology, the diagnosis and treatment of hematological diseases increasingly rely on the collection and analysis of large amounts of medical data, including laboratory test data, genetic screening results, and imaging data. This data typically comes from different devices and at different times, and its format and quality are inconsistent, leading to a growing complexity in data processing. Traditional data analysis methods often rely on a single data source or simple time alignment, making them ill-suited for the demands of multi-source data and complex data structures.

[0003] In existing technologies, medical data analysis generally employs simple alignment and standardization methods based on time series data. While these methods can achieve data preprocessing and standardization to some extent, their main limitation lies in the lack of in-depth correction for periodic fluctuations and interference between multi-source data. When periodic fluctuations exist between multiple data sources (e.g., device sampling errors, noise during data transmission), traditional methods cannot effectively eliminate these interferences, resulting in a significant reduction in the accuracy and consistency of data analysis results, which in turn affects subsequent diagnostic and treatment recommendations and decisions.

[0004] In response to this, this application proposes a method and system for managing and analyzing medical report data of hematological diseases to solve the above-mentioned problems. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for managing and analyzing medical report data in hematological diseases, addressing the lack of in-depth correction for periodic fluctuations and interference between multi-source data in existing technologies. When periodic fluctuations exist between multiple data sources (e.g., equipment sampling errors, noise during data transmission), traditional methods cannot effectively eliminate these interferences, leading to a significant reduction in the accuracy and consistency of data analysis results, which in turn affects subsequent diagnostic and treatment recommendations and decisions.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] Firstly, this application provides a method for managing and analyzing medical report data related to hematological diseases, including:

[0008] Collect patients’ original medical reports and original test / genetic data, and generate irreversible time-coded fingerprints for each record to obtain fingerprinted raw data;

[0009] The original fingerprint data is subjected to time-frequency domain dual-domain correction to reconstruct the sampling event time sequence and unify the unit of measurement, resulting in aligned data with dual-domain correction.

[0010] The aligned data of the dual-domain correction is structurally extracted and terminology correction is performed based on the inverse mapping of clinical ontology to obtain structured semantic data.

[0011] The structured semantic data is coupled with event clusters and statistically summarized under homomorphic encryption to extract interpretable coupled aggregated feature data;

[0012] Based on the coupled and aggregated feature data, field-level traceable hierarchical diagnosis and follow-up recommendations are generated, and the corresponding traceability fingerprint summary is written into the permissioned audit chain, outputting a traceability analysis report.

[0013] Furthermore, the original fingerprint data is obtained through irreversible hashing and time encoding based on sample event sequences to maintain the traceability of entries without exposing identity information.

[0014] Furthermore, the dual-domain correction analyzes the time series in parallel in the time and frequency domains to identify and correct intra-batch sampling offsets and periodic interference, thereby improving the accuracy of sampling time sequence reconstruction and maintaining fingerprint mapping consistency.

[0015] Furthermore, the reverse mapping of the clinical ontology is achieved by deriving the minimum semantic equivalence set of non-standard expressions from the target ontology and verifying each item against the clinical validation set, so as to map the aligned data into standard structured semantic data with a chain of evidence.

[0016] Furthermore, the event cluster coupling identifies clinically relevant mutation-trend couplings by aggregating homologous or cross-source events by time window and calculating coupling strength indices.

[0017] The coupling strength serves as an interpretable term for the coupled aggregated feature data.

[0018] Furthermore, the statistical summary under homomorphic encryption is used to synthesize sensitive indicators in the ciphertext domain and generate publicly verifiable digest credentials, so as to verify the correctness of feature statistics without decrypting the original values.

[0019] Furthermore, the audit chain is a permissioned distributed ledger that only writes the source fingerprint digest and evidence index to display a verifiable evidence chain between the structured semantic fields corresponding to each suggestion and the original fingerprint in the report.

[0020] Furthermore, after clinical confirmation, the source-tracing analysis report uses the manually confirmed annotations and follow-up results in the report to write back to the structured semantic data and coupled aggregated feature data according to time windows, so as to drive the periodic rule and correction strategy revision.

[0021] Secondly, this application provides a medical report data management and analysis system for hematological diseases, including:

[0022] The data acquisition and fingerprint generation module is used to collect patients' original medical reports and original test / gene data, and generate irreversible time-coded fingerprints for each record to obtain fingerprinted raw data.

[0023] The timing reconstruction and correction module is used to perform time-frequency dual-domain correction on the fingerprinted raw data to reconstruct the timing of sampling events and unify the units of measurement, so as to obtain aligned data with dual-domain correction.

[0024] The semantic parsing and standardization module is used to extract the structured data from the alignment data of the dual-domain correction and perform term correction based on the inverse mapping of clinical ontology to obtain structured semantic data.

[0025] The feature coupling and secure computation module is used to perform event cluster coupling on the structured semantic data and statistically summarize it under homomorphic encryption to extract interpretable coupled aggregated feature data;

[0026] The report generation and audit update module is used to generate field-level traceable hierarchical diagnosis and follow-up recommendations based on the coupled aggregated feature data, write the corresponding traceability fingerprint summary into the permissioned audit chain, and output a traceability analysis report.

[0027] Compared with existing technologies, the present invention provides a method and system for managing and analyzing medical report data of hematological diseases. Through time-frequency domain dual-domain correction technology, this solution effectively eliminates periodic errors and fluctuations between different data sources and different acquisition time points, improving the accuracy and consistency of data processing. This correction method goes beyond traditional time alignment; it identifies and removes periodic interference in the signal through frequency domain analysis, ensuring the consistency of time series across different sources. This ensures that even with multi-source data, the analysis results maintain high accuracy and reliability, avoiding erroneous diagnoses and recommendations due to data errors, thus providing a more accurate and reliable basis for clinical decision-making. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0029] Figure 1 A flowchart illustrating a method for managing and analyzing medical report data for hematological diseases, provided as an embodiment of the present invention;

[0030] Figure 2 This is a block diagram of a medical report data management and analysis system for hematological diseases provided in an embodiment of the present invention. Detailed Implementation

[0031] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0032] As attached Figure 1 As shown:

[0033] Example 1:

[0034] A method for managing and analyzing medical report data related to hematological diseases, including:

[0035] S1. Collect the patient's original medical reports and original test / gene data, and generate an irreversible time-coded fingerprint for each record to obtain fingerprinted raw data;

[0036] In step S1, the original fingerprint data is obtained by irreversible hashing and time encoding based on sample event sequences to maintain the traceability of entries without exposing identity information;

[0037] Furthermore, medical reports and test results from patients are obtained from hospitals or the healthcare system. This data includes blood tests, genetic screenings, and imaging examinations. The raw report data may contain unstructured or semi-structured information, covering patient basic information, laboratory values, clinical symptoms, etc.

[0038] S2. Perform time-frequency domain dual-domain correction on the fingerprinted raw data to reconstruct the sampling event time sequence and unify the measurement units to obtain aligned data with dual-domain correction;

[0039] In step S2, the dual-domain correction analyzes the time series in parallel in the time and frequency domains to identify and correct intra-batch sampling offsets and periodic interference, thereby improving the accuracy of sampling time series reconstruction and maintaining fingerprint mapping consistency.

[0040] Furthermore, the raw data undergoes preprocessing. First, test results from different units are standardized (e.g., blood glucose concentrations from different units are converted to the same unit). Then, reference intervals and standard values ​​in the data are calibrated to ensure all data meet clinical standards. Finally, all data are synchronized according to their collection time using timestamp alignment to eliminate discrepancies caused by different data collection times and ensure consistency in the analysis.

[0041] S3. The aligned data of the dual-domain correction is structurally extracted and terminology correction is performed based on the inverse mapping of clinical ontology to obtain structured semantic data.

[0042] In step S3, the reverse mapping of the clinical ontology is performed by deriving the minimum semantic equivalence set of non-standard expressions from the target ontology and verifying them item by item against the clinical validation set, so as to map the aligned data into standard structured semantic data with a chain of evidence.

[0043] Furthermore, the aligned and standardized data is transformed into a structured format to facilitate subsequent analysis and processing. Data fields are labeled, and the data is categorized using predefined clinical ontologies or vocabularies, with semantic annotation of key indicators. For example, abnormal blood pressure values ​​can be labeled as "hypertension" for later analysis.

[0044] S4. Perform event cluster coupling on the structured semantic data and statistically summarize it under homomorphic encryption to extract interpretable coupled aggregated feature data;

[0045] In step S4, the event cluster coupling identifies clinically relevant mutation-trend couplings by aggregating homologous or cross-source events by time window and calculating coupling strength indices.

[0046] The coupling strength serves as an interpretable term for the coupled aggregated feature data;

[0047] The statistical summary under homomorphic encryption is used to synthesize sensitive indicators in the ciphertext domain and generate publicly verifiable digest credentials, so as to verify the correctness of feature statistics without decrypting the original values.

[0048] Furthermore, data from different time points are correlated and aggregated into cross-time point features. By extracting abnormal trends from multiple tests, patterns of change related to specific conditions are identified. These patterns may reveal early signs or abnormal fluctuations in certain diseases, helping clinicians to identify potential health risks in advance. The aggregated data provides a rich basis for subsequent recommendations.

[0049] S5. Based on the coupled aggregated feature data, generate field-level traceable hierarchical diagnosis and follow-up suggestions, write the corresponding traceability fingerprint summary into the permissioned audit chain, and output a traceability analysis report;

[0050] In step S5, the audit chain is a permissioned distributed ledger that only writes the source fingerprint digest and evidence index to display the verifiable evidence chain between the structured semantic fields corresponding to each suggestion and the original fingerprint in the report.

[0051] Furthermore, personalized diagnosis, treatment, and follow-up recommendations are generated based on aggregated feature data. A customized treatment or management plan is generated based on the patient's specific condition, historical data, and aggregated features. These recommendations are stratified according to different risk levels, and each recommendation is traceable to its specific data source and processing procedure, generating detailed traceability reports for doctors and patients to refer to.

[0052] As demonstrated above, this scheme effectively eliminates periodic errors and fluctuations between different data sources and acquisition time points through time-frequency domain dual-domain correction technology, improving the accuracy and consistency of data processing. This correction method goes beyond traditional time alignment; it identifies and removes periodic interference in the signal through frequency domain analysis, ensuring time series consistency between different sources (such as laboratory data, genomic data, and self-tested data). This ensures that even with multi-source data, the analysis results maintain high accuracy and reliability, avoiding erroneous diagnoses and recommendations due to data errors, thus providing a more accurate and reliable basis for clinical decision-making.

[0053] Example 2:

[0054] Dynamic surveillance study of minimal residual disease (MRD) in acute myeloid leukemia (AML):

[0055] 1. Data Acquisition and Preprocessing:

[0056] Data source: Treatment data of 100 AML patients in the hematology department of a top-tier hospital over a period of 2 years.

[0057] Original data:

[0058] Medical reports: bone marrow aspiration cytology reports, flow cytometry (FCM) reports, and fusion gene (RT-PCR) quantitative reports, totaling approximately 1200 PDF / Word documents.

[0059] Raw data for testing / genetics: FCS data files exported from flow cytometry, amplification curves and Ct values ​​exported from PCR instruments, and discrete laboratory test results (such as complete blood count) in electronic medical records (EMR).

[0060] Acquisition method: Anonymized batch export through the interface of Hospital Information System (HIS), Laboratory Information System (LIS) and scientific research database.

[0061] 2. Implementation process:

[0062] S1: Generate raw fingerprint data:

[0063] Operation: For each report, each FCS file, and each EMR test record, extract its core content (such as patient de-identified ID, report type, test items, sampling timestamp, and file binary characteristics) as input.

[0064] Irreversible time-coded fingerprint generation: Employs the SHA-256 algorithm (Patient Research ID || Sampling Time (Unix milliseconds) || Report Type Code || File Feature Code). For example, an FCM report dated October 26, 2023 generates fingerprint a1b2...f8. This fingerprint cannot be reversed to retrieve the original information, but uniquely identifies the data entry.

[0065] Output: Create an index table for the fingerprint -> raw data storage path.

[0066] S2: Time-frequency domain dual-domain correction:

[0067] Problem: The "sampling time" for multiple bone marrow aspirations of the same patient may only be the application time or the reporting time in the system, with an offset of several hours to several days; different testing platforms (such as FCM and PCR) have different sample receiving and processing cycles, resulting in time sequence misalignment.

[0068] Time-domain correction: Correct timestamps from other systems based on the most accurate "sample collection nurse operation time" (from the nursing system). Identify and correct significant intra-batch offsets (for example, if samples are submitted in the same batch, the reporting time should approximately follow a specific distribution).

[0069] Frequency domain correction: Perform Fourier transform on time series data to identify and filter out periodic high-frequency noise introduced by the hospital's fixed work cycle (such as weekly centralized testing), thus smoothing the timeline.

[0070] Standardize the units of measurement: Perform standardization preprocessing such as logarithmic transformation on the PCR "copy number" and FCM "positive cell percentage" to prepare for subsequent correlation analysis.

[0071] Output: Each record has a reconstructed accurate sampling time sequence t_corrected, which is mapped to the original fingerprint.

[0072] S3: Structured Extraction and Terminology Correction

[0073] Operation: Use a natural language processing model to extract key information from the report.

[0074] Input: "Flow cytometry detected that approximately 0.1% of cells expressed CD34+, CD117+, CD33+, HLA-DR-, consistent with a leukemia phenotype."

[0075] Original sample: {"Test Item": "Flow Cytometry", "Positive Rate": 0.1%, "Markers": ["CD34+", "CD117+", "CD33+", "HLA-DR-"], "Conclusion": "Leukemia Phenotype"}

[0076] Clinical ontology reverse mapping: The system calls the "Acute Myeloid Leukemia Immunophenotype Ontology".

[0077] Reverse deduction: Based on the biomarker combination CD34+, CD117+, CD33+, HLA-DR-, the minimum semantic equivalence set is deduced to be the leukemia stem cell-like phenotype (LSC-like Phenotype), with an evidence level of "strong correlation".

[0078] Validation: The mapping was validated against an internal clinical validation set (500 reports annotated by experts), confirming that the accuracy of the mapping is >98%.

[0079] Output - Structured semantic data: {"Event Type": "MRD Detection", "Detection Method": "FCM", "Phenological Classification": "LSC-like", "Quantitative Value": 0.001, "Unit": "Proportion", "Time": t_corrected, "Fingerprint": a1b2...f8, "Evidence Chain": [Ontology Concept ID, Validation Set Matching ID]}

[0080] S4: Event Cluster Coupling and Homomorphic Encryption Statistics:

[0081] Event cluster coupling:

[0082] Time window: Detection events set to "within the same treatment cycle (e.g., ±14 days after chemotherapy)" constitute a cluster.

[0083] Coupling analysis: Within a cluster, two cross-source events were identified: “FCM detected LSC-like phenotype” and “PCR detected FLT3-ITD mutation positive”.

[0084] Calculate coupling strength: Based on historical data from 100 patients, calculate the conditional probability and correlation coefficient of the event pair co-occurring within the same time window. Assume the calculated coupling strength index is 0.75 (range 0-1).

[0085] Statistics Summary under Homomorphic Encryption:

[0086] Requirement: To calculate the average MRD value of the sensitive subgroup of all patients with "LSC-like phenotype positivity rate >0.01% and FLT3-ITD positivity" for internal research, but we do not wish to disclose the specific values ​​for each patient.

[0087] Operation: Homomorphically encrypt the MRD value for each patient, ensuring the data remains within the domain. In the encrypted state, the system calculates the sum and count under the encrypted state.

[0088] Generate digest certificate: After decryption, only the final aggregate result (e.g., average MRD=0.25%) is obtained, and a publicly verifiable mathematical certificate is generated to prove that this average value is indeed derived from the correct calculation of the original ciphertext data that meets the conditions and has not been tampered with.

[0089] Output - Coupling Aggregated Feature Data: {"Feature Name": "LSC_FLT3 Coupling Strength", "Value": 0.75}, {"Aggregated Statistics": "Average MRD of High-Risk Subgroup", "Value": 0.0025, "Homomorphic Encryption Certificate": "xyz_verify_token"}

[0090] S5: Generate and update source tracing analysis reports.

[0091] Report generation: The system generates follow-up recommendations for the current patient, such as: "Recommendation 1 (High-risk warning): This test found that the LSC-like phenotype (0.1%) is coupled with FLT3-ITD mutation, which is a high-risk feature. It is recommended to consider more intensive monitoring (such as a follow-up test in 4 weeks) or targeted therapy."

[0092] Write to the permissioned audit chain: Write the fingerprint digest (a1b2...f8, c3d4...e9) and coupled feature fingerprint of the structured semantic data fields on which the recommendation is based (such as the FCM event and PCR event mentioned above) into a blockchain network that can only be accessed by authorized nodes (such as hospitals and drug regulatory departments).

[0093] Output report: Next to "Recommendation 1" in the report, there is a clickable "Trace to Source" button. After clicking, it verifies through the audit chain and displays all the standardized data entries on which this recommendation was based and their original report fingerprints, forming a complete chain of evidence.

[0094] System Update: The doctor adopts the suggestion and follows up in 4 weeks. The doctor confirms the effectiveness of the suggestion in the system and enters the new follow-up results. The system automatically writes the "suggestion effective" annotation and the new results back to the patient's structured semantic data and feature library according to time windows, to optimize future coupling strength calculation rules.

[0095] Example 3:

[0096] Prediction of the risk of transformation from myelodysplastic syndromes (MDS) to AML:

[0097] 1. Data Acquisition:

[0098] Data source: Longitudinal follow-up data of 50 patients with intermediate-to-high-risk MDS over a period of 3 years.

[0099] Original data: a series of blood routine reports, bone marrow pathology reports, chromosome karyotype analysis reports, and next-generation sequencing (NGS) gene mutation reports.

[0100] 2. Flowchart of this embodiment:

[0101] S1 / S2: Fingerprinting and dual-domain correction of scattered blood routine reports to accurately align the detection time points of each indicator (such as absolute neutrophil count) and construct a continuous time series.

[0102] S3: Non-standard descriptions such as "proportion of primitive cells" in pathology reports, "complex karyotype" in chromosome reports, and "TP53 biallelic mutation" in NGS reports are unified into standardized and computable structured data through reverse mapping of clinical ontology.

[0103] S4: Event Cluster Coupling: Couple the events "progressive deterioration of blood routine" (e.g., platelet count slope <0), "TP53 mutation" and "increased proportion of primitive cells" according to the "quarterly" time window, calculate their dynamic coupling strength, and use it as a feature for predicting transformation risk.

[0104] Homomorphic encrypted statistics: In cross-hospital collaborative studies, each hospital encrypts sensitive indicators (such as the frequency of specific mutant alleles) locally. Collaborators can only collaboratively calculate the statistical characteristics of the joint cohort on the encrypted data, thus protecting patient privacy.

[0105] S5: Generates a "high conversion risk" early warning suggestion and records the standardized indicator fingerprints on the blockchain. Subsequent confirmation information from doctors regarding conversion events is fed back to the system to optimize the risk prediction model.

[0106] Comparative Example: Current conventional manual management combined with information technology management methods

[0107] Data management: Researchers manually downloaded PDF reports from various systems and compiled them into Excel spreadsheets. The data was scattered and the versions were inconsistent.

[0108] Time processing: The sampling time is taken as the date the report is printed, ignoring the time difference between the actual sampling and the detection, resulting in a coarse timing.

[0109] Terminology standardization: The report was manually read and keywords were extracted, or simple keyword matching was used. There are many synonyms and variant forms (such as "primitive granulocytes," "blast cells," and "immature myeloid cells"), resulting in low standardization and reliance on manual verification.

[0110] Association analysis: Manually filtering and matching data from different sources in Excel to perform simple cross-tabulation analysis. This method struggles to identify dynamically coupled events across time windows.

[0111] Statistics and Privacy: Directly aggregating patients' plaintext data for statistical analysis, or transmitting raw data after anonymization during collaboration, poses a risk of privacy leakage.

[0112] Reporting and Source Tracing: Analytical conclusions are written in text form in reports, and recommendations are often only noted as "based on a certain date," lacking an immutable chain of evidence pointing to specific entries in the original data. Historical experience is difficult to feed back into the analysis system in a structured manner.

[0113] The overall results are compared in Table 1 below:

[0114] Table 1

[0115]

[0116] As demonstrated above, by employing a traceable report generation and audit chain write-back mechanism, this solution not only preserves the source data for each diagnostic suggestion and feature when generating analytical reports, but also achieves a complete chain of evidence and data traceability. Each analytical step and each generated suggestion can be traced back to the original data and processing flow, and all data updates and modifications are recorded through a permissioned audit chain. This innovative traceability mechanism significantly enhances report transparency, ensuring that clinicians and patients can access the data source and analysis process at any time, thereby improving the credibility and verifiability of decisions. This innovation not only positively impacts patient treatment plans but also prevents risks caused by data misuse or analytical bias in practical applications, thus providing strong protection for medical quality and data security.

[0117] like Figure 2 As shown, in one embodiment, this application also provides a medical report data management and analysis system for hematological diseases, including:

[0118] The data acquisition and fingerprint generation module is used to collect patients' original medical reports and original test / gene data, and generate irreversible time-coded fingerprints for each record to obtain fingerprinted raw data.

[0119] The timing reconstruction and correction module is used to perform time-frequency dual-domain correction on the fingerprinted raw data to reconstruct the timing of sampling events and unify the units of measurement, so as to obtain aligned data with dual-domain correction.

[0120] The semantic parsing and standardization module is used to extract the structured data from the alignment data of the dual-domain correction and perform term correction based on the inverse mapping of clinical ontology to obtain structured semantic data.

[0121] The feature coupling and secure computation module is used to perform event cluster coupling on the structured semantic data and statistically summarize it under homomorphic encryption to extract interpretable coupled aggregated feature data;

[0122] The report generation and audit update module is used to generate field-level traceable hierarchical diagnosis and follow-up recommendations based on the coupled aggregated feature data, write the corresponding traceability fingerprint summary into the permissioned audit chain, and output a traceability analysis report.

[0123] The beneficial effects and technical effects of the same medical report data management and analysis method for hematological diseases will not be elaborated here.

[0124] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for managing and analyzing medical report data of hematological diseases, characterized in that, include: Collect patients’ original medical reports and original test / genetic data, and generate irreversible time-coded fingerprints for each record to obtain fingerprinted raw data; The original fingerprint data is subjected to time-frequency domain dual-domain correction to reconstruct the sampling event time sequence and unify the unit of measurement, resulting in aligned data with dual-domain correction. The aligned data of the dual-domain correction is structurally extracted and terminology correction is performed based on the inverse mapping of clinical ontology to obtain structured semantic data. The structured semantic data is coupled with event clusters and statistically summarized under homomorphic encryption to extract interpretable coupled aggregated feature data; Based on the coupled and aggregated feature data, field-level traceable hierarchical diagnosis and follow-up recommendations are generated, and the corresponding traceability fingerprint summary is written into the permissioned audit chain, outputting a traceability analysis report.

2. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, The fingerprinted raw data is obtained through irreversible hashing and time encoding based on sample event sequences to maintain the traceability of entries without exposing identity information.

3. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, The dual-domain correction analyzes the time series in parallel in the time and frequency domains to identify and correct intra-batch sampling offsets and periodic interference, thereby improving the accuracy of sampling time series reconstruction and maintaining fingerprint mapping consistency.

4. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, The reverse mapping of the clinical ontology is achieved by deriving the minimum semantic equivalence set of non-standard expressions from the target ontology and verifying each item against the clinical validation set, so as to map the aligned data into standard structured semantic data with a chain of evidence.

5. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, The event cluster coupling identifies clinically relevant mutation-trend couplings by aggregating homologous or cross-origin events by time window and calculating coupling strength indices. The coupling strength serves as an interpretable term for the coupled aggregated feature data.

6. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, The statistical summary under homomorphic encryption is used to synthesize sensitive indicators in the ciphertext domain and generate publicly verifiable digest credentials, so as to verify the correctness of feature statistics without decrypting the original values.

7. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, The audit chain is a permissioned distributed ledger that only writes the source fingerprint digest and evidence index to display the verifiable evidence chain between the structured semantic fields corresponding to each suggestion and the original fingerprint in the report.

8. The method for managing and analyzing medical report data of hematological diseases according to claim 1, characterized in that, After clinical confirmation, the source-tracing analysis report uses the manually confirmed annotations and follow-up results in the report to write back to the structured semantic data and coupled aggregated feature data according to time windows, so as to drive the periodic rule and correction strategy revision.

9. A medical report data management and analysis system for hematological diseases, characterized in that, include: The data acquisition and fingerprint generation module is used to collect patients' original medical reports and original test / gene data, and generate irreversible time-coded fingerprints for each record to obtain fingerprinted raw data. The timing reconstruction and correction module is used to perform time-frequency dual-domain correction on the fingerprinted raw data to reconstruct the timing of sampling events and unify the units of measurement, so as to obtain aligned data with dual-domain correction. The semantic parsing and standardization module is used to extract the structured data from the alignment data of the dual-domain correction and perform term correction based on the inverse mapping of clinical ontology to obtain structured semantic data. The feature coupling and secure computation module is used to perform event cluster coupling on the structured semantic data and statistically summarize it under homomorphic encryption to extract interpretable coupled aggregated feature data; The report generation and audit update module is used to generate field-level traceable hierarchical diagnosis and follow-up recommendations based on the coupled aggregated feature data, write the corresponding traceability fingerprint summary into the permissioned audit chain, and output a traceability analysis report.

Images