Intelligent diagnosis and treatment system based on AI technology

Abstract

The invention discloses an intelligent diagnosis and treatment system based on an AI technology. The intelligent diagnosis and treatment system comprises a login unit, a diagnosis unit, a decision-making unit and a data processing unit. And the login unit is used for the patient to log in and input related information. The diagnosis unit is used for receiving the examination data and outputting a diagnosis and treatment scheme and confidence evaluation. In the diagnosis unit, when the confidence coefficient value exceeds a threshold value, a decision-making unit is triggered and used for diagnosing diseases of the patient and relatives or helping a doctor diagnose the patient and providing a scheme for determining final diagnosis and treatment. The data processing unit is used for erasing public network data of the patient after the diagnosis and treatment scheme is completed. A recognition unit is arranged in the login unit and used for recognizing and verifying biological information of a patient, and the patient input information comprises current symptoms, disease history and allergy history. According to the method, the privacy of the patient can be protected, leakage of personal information of the patient is avoided, precise diagnosis and treatment can be realized, an optimal diagnosis and treatment scheme is formulated for the patient, and missed diagnosis, missed diagnosis and excessive treatment are avoided.

Description

Technical Field

[0001] This invention belongs to the field of intelligent diagnosis and treatment technology, specifically relating to an intelligent diagnosis and treatment system based on AI technology. Background Technology

[0002] AI-powered intelligent diagnosis and treatment refers to the use of artificial intelligence technology to assist or replace some medical processes, enabling intelligent services for disease diagnosis, treatment recommendations, and health management. Its core lies in using technologies such as deep learning and natural language processing to intelligently analyze diagnostic data from medical images, clinical tests, and patient symptom descriptions, providing diagnostic or auxiliary diagnostic suggestions. Through the intervention of AI technology, doctors can more accurately assess conditions and provide optimal treatment plans. Simultaneously, AI can rapidly process and analyze large amounts of medical data, reducing the workload of doctors and making the diagnosis and treatment process more efficient. However, existing AI-powered intelligent diagnosis and treatment still have the following problems: ① Data security deficiencies, which cannot guarantee patient privacy, such as the 2024 Stanford Hospital data breach that resulted in the leakage of 100,000 patients' data due to cloud transmission; ② Human decision-making bias, which reduces the occurrence of missed diagnoses and misdiagnoses, as different doctors may have different diagnoses for the same disease, making it difficult to determine the best treatment plan; ③ System response delays, as some AI-powered medical treatments require manual confirmation of all plans, which can easily delay the golden period of treatment in the case of serious illnesses.

[0003] Therefore, in order to address the aforementioned technical issues, it is necessary to provide an intelligent diagnosis and treatment system based on AI technology.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent diagnosis and treatment system based on AI technology, which can solve the problems mentioned above in the background.

[0006] To achieve the above objectives, a specific embodiment of the present invention provides the following technical solution: An AI-based intelligent diagnosis and treatment system includes a login unit, a diagnosis unit, a decision-making unit, and a data processing unit. The login unit is used for patients to log in and input relevant information. The diagnosis unit receives examination data and outputs a treatment plan and confidence level assessment. Within the diagnosis unit, when the confidence level exceeds a threshold, the decision-making unit is triggered to diagnose the patient's and their relatives' illnesses or to assist doctors in diagnosing the patient and providing a final treatment plan. The data processing unit erases the patient's public network data after the treatment plan is finalized.

[0007] In one or more embodiments of the present invention, the login unit is provided with an identification unit, which is used to identify and verify the patient's biometric information. The patient's input information includes current symptoms, medical history, and allergy history.

[0008] In one or more embodiments of the present invention, the diagnostic unit executes a three-tiered treatment pathway, including: S1. Basic Examinations: Based on the patient's symptoms, match clinical treatment guidelines or generate basic examination requirements; S2. Additional checks: When the confidence level of the basic check result exceeds the threshold X, anomaly markers are associated, which may generate support or exclusion of additional check requirements. S3, Treatment Plan: Integrates S1 and S2 to output AI treatment plans.

[0009] In one or more embodiments of the present invention, the diagnostic unit includes a multimodal data fusion unit and a clinical fit index generation unit. The multimodal data fusion unit is used to integrate the results of imaging, pathology, gene testing, clinical trials, and other physiological and physicochemical examinations used to support or exclude diagnosis and treatment from basic examinations and additional examinations. The clinical fit index generation unit outputs a CAI score of 0% to 100% based on the AI ​​diagnosis and treatment plan.

[0010] In one or more embodiments of the present invention, when the CAI score is ≥99%, the diagnostic unit automatically generates a diagnosis and treatment report, and the patient independently carries out subsequent treatment based on the diagnosis and treatment report. When the CAI score is <99%, and the diagnostic unit displays a low confidence dimension, it may continue to generate support or exclude the need for additional examination items and send them to the decision unit.

[0011] In one or more embodiments of the present invention, the triggering condition for the additional check is: S201, physiological and biological test values ​​such as tumor markers and inflammatory markers exceed the threshold; S202, Imaging reports must include descriptions of BI-RADS / TI-RADS 4 or higher. S202, pathological examination revealed malignant cell markers.

[0012] In one or more embodiments of the present invention, the AI ​​diagnosis and treatment plan includes disease diagnosis, drug treatment, and non-drug treatment.

[0013] In one or more embodiments of the present invention, the display terminal of the decision unit presents: case codes, decision tree diagrams, and abnormal result marking areas.

[0014] In one or more embodiments of the present invention, the data processing unit adopts a segmented storage architecture, which divides the patient data into N encrypted fragments for storage. Each fragment does not contain complete personal information, and the fragments are automatically erased M hours after the end of the treatment.

[0015] In one or more embodiments of the present invention, the intelligent diagnosis and treatment system further includes a correction unit for comparing the efficacy differences between AI diagnosis and treatment plans and manually determined diagnosis and treatment plans. When the manually determined diagnosis and treatment plan leads to a decrease in efficacy, the doctor's operating authority is reduced. At the same time, when the AI ​​diagnosis and treatment plan makes n consecutive errors, the correction unit triggers an emergency update of the knowledge base.

[0016] Compared with existing technologies, the AI-based intelligent diagnosis and treatment system of the present invention can not only protect patient privacy and prevent the leakage of patient personal information, but also achieve precise diagnosis and treatment, formulate the best treatment plan for patients, and avoid misdiagnosis, missed diagnosis and overtreatment. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a system block diagram of an AI-based intelligent diagnosis and treatment system according to an embodiment of the present invention; Figure 2 This is a three-level diagnostic and treatment pathway diagram of the diagnostic unit in one embodiment of the present invention; Figure 3 This is a flowchart of an AI-based intelligent diagnosis and treatment system according to one embodiment of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this disclosure.

[0020] like Figures 1 to 3As shown, an AI-based intelligent diagnosis and treatment system according to one embodiment of the present invention includes a login unit, a diagnosis unit, a decision-making unit, and a data processing unit. The login unit is used for patients to log in and input relevant information. The diagnosis unit is used to receive examination data and output a treatment plan and confidence level assessment. In the diagnosis unit, when the confidence level exceeds a threshold, the decision-making unit is triggered to diagnose the patient's and their relatives' illnesses or to assist doctors in diagnosing the patient and providing a final treatment plan. The data processing unit is used to erase the patient's public network data after the treatment plan is completed.

[0021] After verifying their personal information through the login unit, patients complete the login process by inputting information including, but not limited to, current symptoms, medical history, and allergy history. Upon receiving this data, the diagnostic unit automatically generates basic examination or treatment plans, such as ultrasound and / or blood sample testing if needed. This eliminates the need for patients to register, wait, and undergo diagnostic data support and exclusion at medical institutions, significantly reducing waiting time, improving diagnostic efficiency, and lowering treatment costs. For example, for patients with persistent chest pain, the diagnostic unit generates a recommended "ECG + cardiac enzyme test" examination plan based on symptom characteristics and diagnostic needs, instead of directly performing a high-cost CT scan.

[0022] After the patient completes the examination, the diagnostic unit will synchronize the results and generate a confidence score based on the symptoms and examination results. If the confidence score does not exceed the threshold, it indicates that the examination results are good and no further examination is needed. The diagnostic unit will then generate a diagnostic report and / or treatment prescription. If the patient does not require medication, other intervention suggestions will be generated. If the patient requires medication, a prescription will be generated and sent to the client or a nearby pharmacy address, recommended to the client and / or the pharmacy, and reviewed by the pharmacy. Patients can also pick up their medication at the pharmacy after paying through the intelligent treatment system, thus completing the treatment process. For example, for "persistent cough" and its depth and duration, the diagnostic unit generates a basic laboratory diagnostic plan of "blood cell analysis + CRP + CT scan." Based on the laboratory diagnostic results and CT scan showing ground-glass opacities in the lungs ≥50% with a confidence level of 99% and no further suggestions, the output AI treatment plan is "antiviral or antibiotic medication + oxygen therapy."

[0023] If the confidence level exceeds the threshold, additional examinations will be triggered, and the results will be sent to the decision-making unit. The doctor will then manually review the results to determine the final treatment plan. For example, for a patient with hemoptysis, the diagnostic unit generates a basic examination plan of "low-dose CT scan + tuberculin skin test." If the CT scan shows a lesion in the hilum and the tuberculin skin test is negative with a confidence level of 76%, an additional examination plan of "bronchoscopic biopsy + tumor markers" will be generated. The AI ​​treatment plan will be output as "suspected central lung cancer treatment pathway" and sent to the decision-making unit. The doctor will execute the biopsy, obtain the results, and approve them. The patient can then proceed with treatment according to the AI ​​treatment plan.

[0024] Once a patient's diagnosis is completed, the data processing unit will erase the patient's data to prevent the leakage of patient privacy and ensure the security of the patient's personal information.

[0025] like Figure 1 As shown, the login unit includes an identification unit used to identify and verify the patient's biometric information, which includes, but is not limited to, voiceprint, iris, and fingerprint. Patient input information includes, but is not limited to, current symptoms, medical history, and allergy history.

[0026] The login unit uses local closed-loop storage, and patient files are stored only in the institution's or hospital's intranet encrypted server. Storage, computing, and display are physically isolated from the Internet, and medical or private network encryption or quantum encryption channels are used for data transmission.

[0027] In addition, the login unit is equipped with a physical isolation gateway module, which includes a data flow controller. When a network connection request is detected, the hardware circuit breaker is automatically triggered, and an optical character desensitizer is deployed at the data output end to hide sensitive information such as the patient's name and ID number.

[0028] like Figure 2 As shown, the diagnostic unit executes a three-tiered treatment pathway, including: S1. Basic Examinations: Based on the patient's symptoms, match clinical treatment guidelines or generate basic examination requirements; S2. Additional checks: When the confidence level of the basic check result exceeds the threshold X, anomaly markers are associated, which may generate support or exclusion of additional check requirements. S3, Treatment Plan: Integrates S1 and S2 to output AI treatment plans.

[0029] The threshold X is calculated dynamically, i.e., X = 85% - 0.5% × age coefficient + 2%Z × number of critical symptom markers, where the age coefficient = age / 65Y.

[0030] like Figures 1 to 3As shown, the diagnostic unit includes a multimodal data fusion unit and a clinical fit index generation unit. The multimodal data fusion unit integrates results from basic and supplementary examinations, including imaging, pathology, genetic testing, clinical trials, and other physiological and physiochemical examinations used to support or exclude diagnosis. The clinical fit index generation unit outputs a CAI score ranging from 0% to 100% based on the AI-based diagnostic protocol.

[0031] CAI=\frac{sum_{i=1}^{n}(W_i\timesS_i)}{sum_{i=1}^{n}W_i}\times100\% Where Wi is the weight and Si is the accuracy of the indicator.

[0032] In this implementation, when the CAI score is ≥99%, the diagnostic unit automatically generates a treatment report. Patients can then proceed with subsequent treatment independently or under the guidance of a doctor, based on the report. This means the treatment plan can be directly implemented, avoiding delays caused by manual intervention. When the CAI score is <99%, and the diagnostic unit displays a low confidence level, it may continue to generate reports supporting or excluding additional testing requirements, which are then sent to the decision unit.

[0033] In addition, the triggering conditions for additional checks are: S201, Physiological and biological markers such as tumor markers and inflammatory markers exceed the threshold; S202, Imaging reports must include descriptions of BI-RADS / TI-RADS 4 or higher. S202, pathological examination revealed malignant cell markers.

[0034] Specifically, AI-based treatment plans include disease diagnosis, disease risk and risk assessment, health and health status assessment, and drug and non-drug treatment plans. Drug treatment plans include clearly defined drug names, dosages, frequency of use, and precautions. Non-drug treatment plans include, but are not limited to, surgery, radiotherapy, physical rehabilitation, psychological intervention, or lifestyle modifications.

[0035] In this embodiment, the decision unit's display terminal presents: case code, decision tree diagram, and abnormal result marking area. This further protects patient privacy. For example, the decision unit's display terminal presents: Patient ID-JX2025, 58-year-old male, 30-year smoking history, chief complaint "hemoptysis", basic examination plan "low-dose CT + sputum cytology", CT shows a 2cm mass in the hilum, confidence level 76%, additional examination plan "bronchoscopic biopsy + PD-L1 detection", output AI treatment plan "thoracoscopic lobectomy + immunotherapy".

[0036] Meanwhile, the treatment plan of the decision-making unit needs to be activated after the physician's electronic signature, so as to record the physician's actions.

[0037] like Figure 1 As shown, the data processing unit adopts a segmented storage architecture and uses an automatic erasure rule, which divides patient data into N encrypted fragments for storage. Each fragment does not contain complete personal information, and automatic fragment erasure is triggered M hours after the end of treatment. For example, automatic fragment erasure is triggered 24 hours after the end of treatment.

[0038] Of course, patients can also initiate a "data erasure request" at any time after the treatment is completed.

[0039] like Figure 1 As shown, the intelligent diagnosis and treatment system also includes a correction unit, used to compare the efficacy differences between AI-determined and manually determined treatment plans. When a manually determined treatment plan leads to a decrease in efficacy, the doctor's operating privileges are reduced. Simultaneously, if the AI ​​treatment plan makes n consecutive errors, the correction unit triggers an emergency update of the knowledge base. For example, patient ID-JX2025, a 52-year-old female, presents with a history of "thyroid nodules." The basic examination plan is "ultrasound + thyroid function tests." The ultrasound indicates TI-RADS 4 with a confidence level of 82%. Additional examinations are "fine-needle aspiration biopsy + gene testing." The AI-generated treatment plan is "left thyroidectomy." The doctor confirms the gene testing results and modifies it to "surgical resection." If the postoperative pathology confirms a benign result, the correction unit marks the doctor as having an "overtreatment tendency."

[0040] It will be apparent to those skilled in the art that this disclosure is not limited to the details of the exemplary embodiments described above, and that this disclosure can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of this disclosure is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within this disclosure. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0041] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An intelligent diagnosis and treatment system based on AI technology, characterized in that, include: The login unit is used for patients to log in and enter relevant information. The diagnostic unit is used to receive examination data and output treatment plans and confidence assessments. The decision unit is triggered when the confidence level exceeds the threshold. It is used to diagnose the disease of the patient and their relatives or to help doctors diagnose the patient and provide a plan for determining the final treatment. The data processing unit is used to erase the patient's public network data after the treatment plan is completed.

2. The intelligent diagnosis and treatment system based on AI technology according to claim 1, characterized in that, The login unit includes an identification unit, which is used to identify and verify the patient's biometric information. The patient's input information includes current symptoms, medical history, and allergy history.

3. The intelligent diagnosis and treatment system based on AI technology according to claim 1, characterized in that, The diagnostic unit executes a three-tiered treatment pathway, including: S1. Basic Examinations: Based on the patient's symptoms, match clinical treatment guidelines or generate basic examination requirements; S2. Additional checks: When the confidence level of the basic check result exceeds the threshold X, anomaly markers are associated, which may generate support or exclusion of additional check requirements. S3, Treatment Plan: Integrates S1 and S2 to output AI treatment plans.

4. The intelligent diagnosis and treatment system based on AI technology according to claim 3, characterized in that, The diagnostic unit includes a multimodal data fusion unit and a clinical diagnostic and treatment fit index generation unit. The multimodal data fusion unit is used to integrate the results of imaging, pathology, gene testing, clinical trials, and other physiological and physiochemical examinations used to support or exclude diagnosis and treatment from basic examinations and additional examinations. The clinical diagnostic and treatment fit index generation unit outputs a CAI score of 0% to 100% based on the AI ​​diagnosis and treatment plan.

5. The intelligent diagnosis and treatment system based on AI technology according to claim 4, characterized in that, When the CAI score is ≥99%, the diagnostic unit automatically generates a diagnosis and treatment report, and the patient independently carries out subsequent treatment based on the report. When the CAI score is <99%, and the diagnostic unit displays a low confidence dimension, it may continue to generate reports supporting or excluding the need for additional examinations, and send them to the decision unit.

6. The intelligent diagnosis and treatment system based on AI technology according to claim 3, characterized in that, The triggering condition for the additional check is: S201. Tumor marker detection values ​​exceed the threshold; S202, Imaging reports must contain at least four BI-RADS / TI-RADS descriptions; S202, pathological examination revealed malignant cell markers.

7. The intelligent diagnosis and treatment system based on AI technology according to claim 3, characterized in that, The AI-based diagnosis and treatment plan includes disease diagnosis, drug treatment, and non-drug treatment.

8. The intelligent diagnosis and treatment system based on AI technology according to claim 1, characterized in that, The display terminal of the decision unit presents: case code, decision tree diagram and abnormal result marking area.

9. The intelligent diagnosis and treatment system based on AI technology according to claim 1, characterized in that, The data processing unit adopts a segmented storage architecture, which divides patient data into N encrypted fragments for storage. Each fragment does not contain complete personal information, and the fragments are automatically erased M hours after the end of treatment.

10. The intelligent diagnosis and treatment system based on AI technology according to claim 3, characterized in that, The intelligent diagnosis and treatment system also includes a correction unit, which is used to compare the efficacy differences between AI-determined diagnosis and treatment plans and manually determined diagnosis and treatment plans. When the manually determined diagnosis and treatment plan leads to a decrease in efficacy, the doctor's operating authority is reduced. At the same time, when the AI ​​diagnosis and treatment plan makes n consecutive errors, the correction unit triggers an emergency update of the knowledge base.

Images