Clinical auxiliary decision-making system and equipment based on decision-making tree and storable medium

Abstract

The invention belongs to the field of computer aided decision making, and particularly relates to a clinical aided decision making system and device based on a decision-making tree and a storable medium. The system comprises an acquisition unit used for acquiring pain information of a patient, wherein the pain information comprises whether a pain part can be positioned or not and the pain part; the decision-making unit is used for inputting the pain information into a decision-making tree model and then outputting pain causes of the patient; the decision-making unit comprises two sub-modules. The first sub-module is used for obtaining a first-level judgment result based on whether the pain part can be positioned or not. The second submodule is used for obtaining a second-level judgment result based on the first-level judgment result and the pain part; and the decision-making unit outputs the pain cause of the patient based on the second-stage judgment result. According to the method, efficient judgment of clinical pain reasons is realized, and doctors are effectively assisted in decision making.

Description

Technical Field

[0001] This invention is a divisional application of the invention patent with application date of May 8, 2024, application number 202410561464.3, and invention name of an artificial intelligence-based toothache diagnosis method, system and device.

[0002] The present invention relates to the field of computer-assisted decision making, and more specifically, to a clinical decision-assisted system, device, and computer-readable storage medium based on a decision tree. Background Art

[0003] With the development of artificial intelligence, many clinical decision-making support systems and medical expert systems based on artificial intelligence are increasingly used in medical decision-making support, which improves the efficiency and accuracy of clinical diagnosis. Existing medical expert systems mostly use models such as support vector machines and logistic regression to make auxiliary decisions on diseases based on the extracted features. In clinical practice, non-odontogenic toothache is often ignored. The main reason may be that its symptoms are similar to those of odontogenic toothache, but the causes are different, which makes it easy for dentists to misdiagnose it as odontogenic toothache. Therefore, failure to correctly diagnose non-odontogenic toothache may lead to unnecessary dental damage, delay the patient's access to correct treatment, and have a serious impact on the patient's life and health.

[0004] However, the symptoms of toothache are relatively unsuitable for the use of models such as support vector machines, and appropriate models are needed to assist clinical decision-making. Summary of the invention

[0005] In view of the above problems, the present invention provides a clinical decision support system based on a decision tree, which can effectively distinguish the various subcategories of pain through machine learning decision trees, and can more accurately determine the cause of pain, thereby formulating a more appropriate treatment plan.

[0006] A clinical decision support system based on a decision tree, the system comprising: An acquisition unit is used to acquire the patient's pain information, wherein the pain information includes whether the pain site can be located and the pain site; Decision-making unit: used to input the pain information into the decision tree model and then output the cause of the patient's pain; the decision-making unit includes two submodules: The first submodule: used to obtain the first-level judgment result based on whether the pain site can be located; The second submodule is used to obtain the second-level judgment result based on the first-level judgment result and the pain location; The decision unit outputs the cause of the patient's pain based on the second-level judgment result.

[0007] Furthermore, the decision unit also includes a third submodule: for obtaining a third-level judgment result based on the second-level judgment result and the nature and duration of pain, and the decision unit outputs the cause of the patient's pain based on the third-level judgment result.

[0008] A clinical decision-supporting device based on a decision tree comprises: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, any one of the clinical decision-supporting systems based on a decision tree is implemented.

[0009] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements any one of the clinical decision support systems based on a decision tree.

[0010] A clinical decision support system based on ensemble learning, the system comprising: Get scale answer unit: used to obtain the patient's answer to the screening questionnaire. Scale decision unit: used to input the answers into an integrated learning classifier to obtain subclass groupings of pain causes.

[0011] Furthermore, the ensemble learning classifier is an ensemble learning based on the Fisher discriminant equation, and the ensemble learning includes the Fisher discriminant equations of different subclass groupings. After the patient's answers to the screening questionnaire are input into the ensemble learning classifier, the subclass grouping corresponding to the Fisher discriminant equation with the highest Fisher discriminant equation score of the different subgroups is output.

[0012] Furthermore, the Fisher discriminant equations of the different subclass groups are obtained by fitting the Fisher discriminant equations to the different subclass groups based on the clinical data set to obtain the Fisher discriminant equations of the different subclass groups.

[0013] Furthermore, the pain screening questionnaire is a Chinese localized version obtained by cross-cultural adaptation of the original DePaQ, including 13 items; Optionally, the cross-cultural adaptation steps include: Step 1: Different people translated the original DePaQ into Chinese to obtain different translations; Step 2: Combining the different translations to obtain a unified Chinese translation; Step 3: The unified Chinese translation is back-translated into English to obtain a back-translated English version; Step 4: The expert group formulates the final translation by combining the original DePaQ, the different translations, the unified Chinese translation, and the back-translated English version; Step 5: Based on the final translation, the items of the final translation are revised according to the localization data set to obtain the Chinese localized version; Optionally, the revision includes: deleting inappropriate entries of the Chinese localized version, adding localized adapted entries, and modifying options for entries of the Chinese localized version.

[0014] A clinical decision-making support device based on integrated learning comprises: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, the clinical decision-making support system based on integrated learning is generated.

[0015] A computer-readable storage medium stores a computer program, which, when executed by a processor, provides a clinical decision support system based on integrated learning.

[0016] A clinically assisted toothache diagnosis method based on a decision tree, the method comprising: S101: Acquire toothache information of the patient, wherein the toothache information includes toothache location and toothache part; S102: The toothache information is input into a decision tree model, and the cause of the patient's toothache is output. The decision tree model distinguishes the patient into odontogenic toothache and non-odontogenic toothache based on the toothache information. The distinction process of the decision tree includes: S21: if the toothache of the patient cannot be located, it is determined that the patient has non-odontogenic toothache; S22: if the toothache location of the patient is uncertain, determining that the patient has odontogenic toothache; S23: If the toothache of the patient can be located, and the toothache occurs in the maxillary premolars or molars or the left mandibular posterior teeth, the patient is judged to have non-odontogenic toothache. If the toothache of the patient can be located, and the toothache occurs in the third molars or first molars, the patient is judged to have odontogenic toothache.

[0017] Furthermore, the toothache information also includes the nature and duration of the toothache, and the differentiation process of the decision tree also includes: S23': If the patient's toothache is located as being localizable, the toothache occurs in the maxillary canines or premolars, and the nature and duration of the toothache are continuous severe burning pain, the patient is judged to have non-odontogenic toothache; If the patient's toothache is located as being localizable, the toothache occurs in the maxillary canines or premolars, and the nature and duration of the toothache are transient stabbing pain, the patient is judged to have non-odontogenic toothache; If the patient's toothache can be located, the toothache occurs in a specific tooth, and the toothache is of a continuous throbbing nature and duration, the patient is judged to have non-odontogenic toothache; if the patient's toothache can be located, the toothache occurs in a specific tooth, and the toothache is of a transient sensitivity or a continuous pulsating throbbing pain during an acute attack, or a transient pain when chewing during a chronic attack, or a continuous dull pain or severe pain, the patient is judged to have odontogenic toothache; Optionally, the subcategory of non-odontogenic toothache includes temporomandibular joint disease, and the method for determining the temporomandibular joint disease is: when the patient's toothache is located as unlocalizable and the toothache generally occurs in the maxillary, mandibular, and posterior teeth, and the nature and duration of the toothache are persistent dull pain; Optionally, the subtype of non-odontogenic toothache includes trigeminal neuralgia, and the method for determining trigeminal neuralgia is: the patient's toothache is located as unlocalized, and the toothache generally occurs in both the upper and lower jaws, and the nature and duration of the toothache are sudden electric shock-like or needle-like severe pain; Optionally, the subclass of non-odontogenic toothache includes mental factors, and the mental factors are determined as follows: the patient's toothache is located as unlocalized, and the toothache generally occurs in both the upper and lower jaws, and the nature and duration of the toothache are continuous dull pain or throbbing pain; Optionally, the subclass of non-odontogenic toothache includes maxillary sinusitis, and the method for determining maxillary sinusitis is: if the patient's toothache is localizable and the toothache occurs in the maxillary premolars or molars, the nature and duration of the toothache are persistent distending pain; Optionally, the subtype of non-odontogenic toothache includes cluster headache, and the cluster headache is determined as follows: if the patient's toothache is localizable and the toothache occurs in the maxillary canine or premolar, and the nature and duration of the toothache are continuous severe burning pain; Optionally, the subcategories of the non-odontogenic toothache include systemic diseases, and the method for determining the systemic disease is: if the patient's toothache is localizable and the toothache occurs in the left mandibular posterior teeth, the nature and duration of the toothache are referred dull pain; Optionally, the subcategory of the non-odontogenic toothache includes migraine, and the method for determining the migraine is: if the patient's toothache is localizable and the toothache occurs in other specific teeth, the nature and duration of the toothache are continuous throbbing pain; Optionally, the subcategories of odontogenic toothache include dentin hypersensitivity, irreversible pulpitis, wedge-shaped defect, pericoronitis, tooth crack, vertical root fracture, reversible pulpitis, apical periodontitis, periodontal disease, and dental trauma, wherein wedge-shaped defect, tooth crack, vertical root fracture, periodontal disease, and dental trauma are odontogenic toothaches that are easy to diagnose, and dentin hypersensitivity, irreversible pulpitis, pericoronitis, reversible pulpitis, and apical periodontitis are odontogenic toothaches that are difficult to diagnose clinically; Optionally, the method for determining the wedge-shaped defect is: if the patient's toothache is located as being positionable and the toothache occurs in the maxillary canine or premolar, and the nature and duration of the toothache are tooth sensitivity or transient tingling; Optionally, the method for determining the tooth crack is: if the patient's toothache can be located and the toothache occurs in the first molar, the nature and duration of the toothache are spontaneous stabbing pain and obvious pressure pain; Optionally, the method for determining the vertical root fracture is: if the patient's toothache is localizable and the toothache occurs in the first molar, the nature and duration of the toothache are spontaneous stabbing pain and chewing pain is obvious; Optionally, the method for determining periodontal disease is: if the patient's toothache is localizable and the toothache occurs in other specific teeth, the nature and duration of the toothache are continuous dull pain; Optionally, the method for determining the tooth injury is: if the patient's toothache is localizable and the toothache occurs in other specific teeth, the nature and duration of the toothache are severe pain with numbness; Optionally, the toothache information is obtained through a toothache initial diagnosis questionnaire, and the toothache information also includes other symptoms and other associated diseases.

[0018] A clinical auxiliary toothache diagnosis system based on a decision tree, the system comprising: A decision tree acquisition unit 301 is used to acquire toothache information of a patient, wherein the toothache information includes toothache location and toothache part; Decision tree decision unit 302: used to output the cause of the patient's toothache after the toothache information is input into the decision tree model. The decision tree model distinguishes the patient into odontogenic toothache and non-odontogenic toothache based on the toothache information. The distinction process of the decision tree includes: S21: if the toothache of the patient cannot be located, it is determined that the patient has non-odontogenic toothache; S22: if the toothache location of the patient is uncertain, determining that the patient has odontogenic toothache; S23: If the toothache of the patient can be located, and the toothache occurs in the maxillary premolars or molars or the left mandibular posterior teeth, the patient is judged to have non-odontogenic toothache. If the toothache of the patient can be located, and the toothache occurs in the third molars or first molars, the patient is judged to have odontogenic toothache.

[0019] A clinical auxiliary toothache diagnosis device based on a decision tree comprises: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, the clinical auxiliary toothache diagnosis method based on the decision tree is implemented.

[0020] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the clinically assisted toothache diagnosis method based on a decision tree.

[0021] A clinically assisted odontogenic toothache etiology diagnosis method based on ensemble learning, the method comprising: S201: Obtain answers to a dental pain screening questionnaire from patients with odontogenic dental pain. S202: Inputting the answers of the toothache screening questionnaire of the odontogenic toothache patient into an integrated learning classifier to obtain subclass grouping of odontogenic toothache.

[0022] Furthermore, the ensemble learning classifier is an ensemble learning based on the Fisher discriminant equation, the ensemble learning includes the Fisher discriminant equations of different subclass groups, and the answers of the patient's toothache screening questionnaire are input into the ensemble learning to output the subclass grouping corresponding to the Fisher discriminant equation with the highest Fisher discriminant equation score of the different subgroups; Optionally, the Fisher discriminant equations of the different subgroups are obtained by fitting the Fisher discriminant equations of the different subgroups of odontogenic toothache based on the clinical data set to obtain the Fisher discriminant equations of the different subgroups; Optionally, the toothache screening questionnaire is a Chinese localized version obtained by cross-cultural adaptation of the original DePaQ, including 13 items, including the oral pain site, pain duration, pain severity, pain form, pain spread, the effect of eating, swelling degree, tooth looseness, dysphagia, variation of painful teeth, degree of sleep impact, and pain matching words; Optionally, the 13 items are: Where in your mouth do you currently feel pain? How long have you had the pain? How I would describe the intensity of the pain when it was at its worst; Please think about the pain you are experiencing right now, and what form it takes; Please indicate where the pain spreads to the surrounding area; Please describe how much the pain is worse when you chew or eat on the painful side; Please describe the effect on pain after eating / drinking cold food; Please indicate how swollen your gums are now or have been recently; The degree of tooth looseness you feel; You have difficulty swallowing now or recently; Do you feel pain in the tooth? It stretches out a little; Please indicate the extent to which your sleep is currently affected; If pain is present, which of the following words would you use to describe the pain you are experiencing? Optionally, the cross-cultural adaptation steps include: Step 1: Different people translated the original DePaQ into Chinese to obtain different translations; Step 2: Combining the different translations to obtain a unified Chinese translation; Step 3: The unified Chinese translation is back-translated into English to obtain a back-translated English version; Step 4: The expert group formulates the final translation by combining the original DePaQ, the different translations, the unified Chinese translation, and the back-translated English version; Step 5: Based on the final translation, the items of the final translation are revised according to the localization data set to obtain the Chinese localized version; Optionally, the revision includes: deleting inappropriate entries of the Chinese localized version, adding localized adapted entries, and modifying options for entries of the Chinese localized version; Optionally, the revision includes: Revision 1: For item 1, some options were deleted; Revision 2: For the second item, the 5 time periods were consolidated into 2 time periods; Revision 3: Delete the item about the severity of toothache radiating to the surrounding area; Revision 4: Added an item about the area where toothache spreads to the surrounding area; Revision 5: For item 8, the three options were revised to two options; Revision 6: Merge items 13-16 of the original version into one item; Optionally, the sub-groupings of odontogenic toothache include: reversible pulpitis and dentin sensitivity, apical periodontitis and irreversible pulpitis, and pericoronitis.

[0023] A tooth pain etiology diagnosis system based on ensemble learning, the system comprising: The integrated learning acquisition unit 401 is used to obtain answers from a toothache screening questionnaire of a patient with odontogenic toothache; The integrated learning decision unit 402 is used to input the answers of the toothache screening questionnaire of the odontogenic toothache patient into the integrated learning classifier to obtain the subclass grouping of odontogenic toothache.

[0024] A clinically assisted odontogenic toothache diagnosis device based on integrated learning comprises: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, the clinically assisted odontogenic toothache diagnosis method based on integrated learning is implemented.

[0025] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for diagnosing the cause of tooth pain based on ensemble learning.

[0026] Advantages of this application: 1. This application uses a decision tree to make clinical toothache diagnosis simple and efficient, which can help doctors diagnose the cause of toothache or patients make preliminary diagnosis at home, which is of great clinical significance; 2. This application realizes the diagnosis of odontogenic toothache based on the toothache screening questionnaire, which can assist doctors in diagnosing the cause of toothache in clinical practice; 3. This application realizes toothache diagnosis based on artificial intelligence, which can not only reduce the misdiagnosis rate, avoid unnecessary treatment and waste of medical resources, but also improve the treatment effect and quality of life of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 is a schematic flow chart of a clinically assisted toothache diagnosis method based on a decision tree provided by an embodiment of the present invention; Figure 2 is a schematic flow chart of a clinically assisted odontogenic toothache diagnosis method based on ensemble learning provided by an embodiment of the present invention; Figure 3 It is a schematic flow chart of a clinical auxiliary toothache diagnosis system based on a decision tree or a clinical auxiliary odontogenic toothache diagnosis system based on ensemble learning provided by an embodiment of the present invention; Figure 4is a schematic flow chart of a clinically assisted odontogenic toothache diagnosis system based on integrated learning provided by an embodiment of the present invention; Figure 5 It is a schematic block diagram of a clinical auxiliary toothache diagnosis device based on a decision tree or a clinical auxiliary odontogenic toothache diagnosis device based on ensemble learning provided by an embodiment of the present invention; Figure 6 It is a schematic diagram of the toothache location of the toothache initial diagnosis questionnaire or the toothache applet provided in the embodiment of the present invention; Figure 7 is a decision diagram of a decision tree provided by an embodiment of the present invention, wherein the output yellow indicates non-odontogenic toothache, and green indicates odontogenic toothache; Figure 8 It is the overall solution of the toothache diagnosis method provided by the embodiment of the present invention; Fig. 9 A toothache applet and the items included therein are provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0030] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The sequence numbers of the operations, such as S101, S102, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0032] Figure 1 The present invention provides a schematic flow chart of a clinically assisted toothache diagnosis method based on a decision tree. Specifically, the method comprises the following steps: S101: Obtaining toothache information from a patient, wherein the toothache information includes toothache location and toothache part; Step 1: The patient fills in the toothache applet and answers the items in the applet. The applet collects the patient's answers and automatically converts them into features that can be recognized by machine learning. Then, the machine learning decision tree algorithm is used to diagnose some easier-to-identify toothache diseases. Among them, the easier-to-identify toothache diseases include non-odontogenic toothache and easier-to-diagnose subtypes of odontogenic toothache.

[0033] Non-odontogenic toothache refers to pain originating from tissues or organs outside the oral cavity. This pain may be mistaken for toothache because it involves the teeth. The subcategories of non-odontogenic toothache are divided into ① temporomandibular joint disease, ② trigeminal neuralgia, ③ mental factors, ④ maxillary sinusitis, ⑤ cluster headache, ⑥ systemic disease, and ⑦ migraine according to the different causes. Subcategories of odontogenic toothache include: ① dentin hypersensitivity, ② irreversible pulpitis, ③ wedge-shaped defect, ④ pericoronitis or pericoronitis of wisdom teeth, ⑤ cracked tooth, ⑥ vertical root fracture, ⑦ reversible pulpitis, ⑧ apical periodontitis, periodontal disease, and dental trauma; among them, wedge-shaped defect, cracked tooth, vertical root fracture, periodontal disease, and dental trauma are easier to diagnose; dentin hypersensitivity, irreversible pulpitis, reversible pulpitis, apical periodontitis, and pericoronitis of wisdom teeth are more difficult to diagnose clinically.

[0034] In one embodiment, the toothache applet used and the items it contains are as follows: Fig. 9 As shown: First, the patient fills out the toothache applet and diagnoses the easier-to-identify toothache disease based on the machine learning decision tree algorithm.

[0035] S102: Differentiating the patient into odontogenic toothache and non-odontogenic toothache based on the toothache information, wherein the differentiation includes: S21: if the toothache of the patient cannot be located, it is determined that the patient has non-odontogenic toothache; S22: if the toothache location of the patient is uncertain, determining that the patient has odontogenic toothache; S23: If the toothache of the patient can be located, and the toothache occurs in the maxillary premolars or molars or the left mandibular posterior teeth, the patient is judged to have non-odontogenic toothache. If the toothache of the patient can be located, and the toothache occurs in the third molars or first molars, the patient is judged to have odontogenic toothache.

[0036] In one embodiment, a decision tree is used to distinguish between odontogenic and non-odontogenic toothaches. Figure 7As shown, based on the patient's toothache location, nature and duration, the output is whether the patient has odontogenic toothache (green output) or non-odontogenic toothache (yellow output), and the subcategories of non-odontogenic toothache and partial odontogenic toothache are also output. For odontogenic toothache whose subcategories cannot be clearly diagnosed, only odontogenic toothache is output without the subcategories.

[0037] In one embodiment, if the patient's toothache location can be located and the toothache location occurs in any other specific tooth (e.g. Figure 6 The tooth numbered 25 or other numbered teeth in the figure), the toothache nature and duration are continuous throbbing pain, and the patient is judged to have non-odontogenic toothache, the subtype is migraine, wherein the other specific teeth Figure 6 All teeth except the upper and lower mandibular premolars or molars, maxillary canines or premolars, left mandibular posterior teeth, third molars, and first molars.

[0038] In one embodiment, the decision tree is determined as follows: the decision nodes of the decision tree are obtained based on oral clinical research and literature research, a decision tree search space is formed according to the decision tree, and a decision tree to be trained is generated according to the decision tree search space; the clinical data set is divided into a training set and a validation set, and the decision tree to be trained is trained according to the training set to generate multiple decision tree models; the decision tree is selected according to the performance of the multiple decision tree models on the validation set, and the performance on the validation set includes one or more of the following indicators: accuracy, sensitivity, and specificity.

[0039] The formulation of the toothache machine learning decision tree is based on our previous clinical study of 80 toothache patients included in the Department of Stomatology of Xiangya Second Hospital of Central South University. The toothache location decision point placed at the first branch point of the decision tree is 75% more sensitive than the toothache nature and duration decision point placed at the first branch point. For example, 80% of patients who reported toothache in maxillary canines or premolars (toothache location decision point) had wedge-shaped defects and cluster headaches, while patients who reported spontaneous persistent pain (toothache nature and duration decision point) had multiple subtypes of toothache, which were persistent spontaneous pain, mixed and disordered diseases, and a single disease diagnosis rate of less than 45%, making it difficult to locate and treat. In addition, non-odontogenic pain may be accompanied by other physical manifestations and symptoms, which are the characteristics of non-odontogenic pain, so we use it as the third decision point. Therefore, we chose the toothache location as the first decision point, the toothache nature and duration as the second decision point, and other symptoms and other accompanying diseases as the third decision point, and gradually advanced to accurately diagnose the disease.

[0040] In one embodiment, in the determined decision tree algorithm, the primary split feature / node is the pain site of the toothache. Based on this feature, the algorithm will branch based on the location of the toothache, such as uncertain location, maxillary, mandibular, a group of teeth or a specific tooth. Subsequently, the second key split feature / node is the nature and duration of the toothache. For example, whether it is paroxysmal or continuous, and whether it is transiently sensitive, sharp pain, tingling, electric shock-like, dull pain, etc. Finally, the third split feature / node is to consider other symptoms of toothache and possible accompanying other facial or systemic diseases. This may include other symptoms such as bleeding gums, loose teeth, fever, and possible diseases such as temporomandibular joint disorder, angina, etc. Through this step-by-step feature splitting, the decision tree algorithm can systematically analyze the patient's symptoms, thereby diagnosing whether the patient has non-odontogenic toothache or odontogenic toothache, and can output subcategories of non-odontogenic toothache, including ① temporomandibular joint disease, ② trigeminal neuralgia, ③ mental factors, ④ maxillary sinusitis, ⑤ cluster headache, ⑥ systemic disease, ⑦ migraine; and some subcategories of odontogenic toothache: wedge-shaped defect, tooth crack, vertical root fracture, periodontal disease, and dental trauma. However, when the patient is suspected of suffering from ① dentin hypersensitivity, ② irreversible pulpitis, ③ wisdom tooth pericoronitis, ④ reversible pulpitis, and ⑤ apical periodontitis, the decision tree cannot output a clear diagnosis and can only output that the patient has odontogenic toothache.

[0041] In one embodiment, after the patient completes the items of the mini-program, it is output that the patient's toothache is non-odontogenic toothache, and the subtype of non-odontogenic toothache is trigeminal neuralgia.

[0042] In one embodiment, after the patient completes the items of the mini-program, it is output that the patient's toothache is odontogenic toothache, and the subtype of odontogenic toothache is cracked tooth.

[0043] In one embodiment, after the patient answers and completes the items of the applet, it is output that the patient's tooth pain is odontogenic, but the subtype of odontogenic tooth pain is unknown.

[0044] Figure 2 is a schematic flow chart of a clinically assisted odontogenic toothache diagnosis method based on ensemble learning provided by an embodiment of the present invention; S201: Obtain the answers of the dental pain screening questionnaire of patients with odontogenic dental pain; The above decision tree algorithm diagnosed some easier-to-identify toothache diseases, including non-odontogenic toothache and easier-to-diagnose subtypes of odontogenic toothache, while the more difficult-to-diagnose subtypes of odontogenic toothache included: dentin hypersensitivity, irreversible pulpitis, reversible pulpitis, apical periodontitis, and wisdom tooth pericoronitis. The decision tree can only identify odontogenic toothache ( Figure 8The toothache diseases that cannot be clearly diagnosed in the literature cannot be clearly diagnosed into the specific categories of these difficult-to-diagnose subcategories of odontogenic toothache. For these difficult-to-diagnose subcategories of odontogenic toothache, it is necessary to proceed to the next step of diagnosis.

[0045] In one embodiment, after the patient completes the items of the mini-program, the patient's toothache is output as odontogenic toothache, but the subtype of odontogenic toothache is unknown. At this time, the patient's answers to the toothache screening questionnaire are continued to be obtained to further diagnose the cause or subtype of the patient's odontogenic toothache.

[0046] S202: Inputting the answers of the toothache screening questionnaire of the odontogenic toothache patient into an integrated learning classifier to obtain subclass grouping of odontogenic toothache.

[0047] Step 2: Patients with odontogenic toothache continue to fill out the toothache screening questionnaire online. The patient's answers to the 13 items are automatically input into three Fisher discriminant analyses, and the subclass of odontogenic toothache corresponding to the Fisher discriminant equation with the highest score is output, that is, the output diagnosis.

[0048] 1. Odontogenic toothaches include reversible pulpitis, dentin hypersensitivity, irreversible pulpitis, apical periodontitis and pericoronitis. Based on literature review, patient surveys and expert opinions, we divided these five types of toothache diseases into three groups, namely Group 1 (reversible pulpitis and dentin hypersensitivity), Group 2 (apical periodontitis and irreversible pulpitis) and Group 3 (pericoronitis). The reason is that reversible pulpitis and dentin hypersensitivity are generally treated with soothing treatment and do not require root canal treatment, so they are grouped together; apical periodontitis and irreversible pulpitis require root canal treatment and are therefore grouped together; wisdom tooth pericoronitis requires the removal of wisdom teeth and is therefore grouped together separately.

[0049] 2. We used the Chinese version of the toothache screening questionnaire to collect data and used Fisher discriminant analysis to diagnose odontogenic toothache.

[0050] ① Patient inclusion: A total of 290 patients with toothache were recruited from the Department of Stomatology, Xiangya Second Hospital, Central South University. Among them, 104 patients were diagnosed as group 1 (reversible pulpitis and dentin hypersensitivity), 87 patients were diagnosed as group 2 (apical periodontitis and irreversible pulpitis), and 99 patients were diagnosed as group 3 (pericoronitis). According to references and the advice of consulting experts, we randomly divided the 290 patients with toothache into two subsamples. Subsample 1 (training set) had 203 patients with toothache for Fisher discriminant function training, and subsample 2 (validation set) had 87 patients with toothache for Fisher discriminant validity verification.

[0051] Table 1. Training set and validation set division for Fisher discriminant equation training

[0052] ②Obtaining Fisher discriminant function: Based on subsample 1 (203 patients with toothache), we obtained Fisher discriminant equations based on the items of the Chinese version of the toothache screening questionnaire for different groups. The Fisher discriminant equation score explained the relationship between the score of each item on the Chinese version of the toothache screening questionnaire and the three groups of toothache in the form of a linear equation. The scores of each patient's 13 items were substituted into the discriminant equation to obtain the Fisher scores of the three groups of toothache. Participants were divided into the group with the highest Fisher score.

[0053] The discriminant equation obtained based on the training set is as follows: Group 1 = 6.008*Item1 + 5.865*Item2 + 0.943*Item3 + 12.206*Item4 +0.438*Item5 + (-0.481)*Item6 + 3.401*Item7 + (-1.919)*Item8 + 0.288*Item9 +(2.223)*Item10 + 1.605*Item11 + 0.007*Item12 + 0.719*Item13 + (-21.380).

[0054] Group 2 = 10.702*Item1 + 5.779*Item2 + 2.132*Item3 + 12.635*Item4 +2.336*Item5 + 0.587*Item6 + 3.628*Item7 + (-1.829)*Item8 + 1.601*Item9 + (-4.223)*Item10 + 0.404*Item11 + 0.253*Item12 + 0.489*Item13 + (-38.808).

[0055] Group 3 = 11.663*Item1 + 5.874*Item2 + 1.004*Item3 + 12.384*Item4 +0.042*Item5 + 0.279*Item6 + 4.146*Item7 + (-2.091)*Item8 + 0.954*Item9 +(-3.237)*Item10 +1.963*Item11+(-0.050)*Item12 + 0.797*Item13 + (-34.924).

[0056] ③Validity verification: To verify whether the Fisher discriminant equation based on the Chinese version of the toothache screening questionnaire can correctly divide toothache patients into three groups, the Fisher discriminant equation obtained from subsample 1 was applied to subsample 2 (87 participants). Each participant in subsample 2 was classified into the group with the highest Fisher score. The clinical diagnosis was then compared with the diagnosis of the Fisher discriminant equation, and it was finally found that the Fisher discriminant function could correctly distinguish 93.1% of toothache cases into three groups.

[0057] The sensitivity and specificity of the three groups are shown in Table 2: Table 2. Performance of Fisher discriminant equation on the validation set

[0058] 3. Based on the previous verification, we linked the Fisher discriminant equations of three groups of toothache diseases to artificial intelligence. After the patient completed the toothache screening questionnaire (DePaQ), AI automatically input the score of each item into the three discriminant equations. Each patient was classified into the group with the highest score of the Fisher discriminant equation, which was the output diagnosis.

[0059] 4. Output diagnosis: ①Reversible pulpitis and dentin hypersensitivity; ② Apical periodontitis and irreversible pulpitis; ③Pericoronitis of wisdom teeth.

[0060] In one embodiment, the Chinese version of the toothache screening questionnaire used is as follows: Table 3. Chinese version of toothache screening questionnaire

[0061] 3. Clinical Application 1. It can be used to assist dentists in diagnosing toothache; 2. If the patient cannot seek medical treatment in time due to work or other reasons, the patient can make a preliminary diagnosis on his own.

[0062] In one embodiment, a toothache diagnosis method based on artificial intelligence and a toothache diagnosis method based on artificial intelligence are combined, such as Figure 8 As shown, first, a small program written through a toothache initial diagnosis questionnaire is used to automatically collect patients' toothache information, distinguish between some easier-to-diagnose toothache diseases (non-odontogenic toothache and some easier-to-diagnose odontogenic toothache) and other toothache diseases that cannot be clearly diagnosed (more difficult-to-diagnose odontogenic toothache). These odontogenic toothaches that cannot be clearly diagnosed continue to use the diagnostic method of odontogenic toothache based on artificial intelligence.

[0063] In one embodiment, after the patient completes the items of the mini-program, the patient's toothache is output as odontogenic toothache, but the subtype of odontogenic toothache is unknown. At this time, the Chinese version of the patient's toothache screening questionnaire is collected, and according to the patient's answer, the subtype of the patient's odontogenic toothache is obtained as pericoronitis.

[0064] Figure 3 is a schematic flow chart of a toothache diagnosis system based on artificial intelligence provided by an embodiment of the present invention; A clinically assisted toothache diagnosis system based on a decision tree, comprising: Acquisition unit 301: used to acquire the patient's toothache information, wherein the toothache information includes the toothache location and the toothache part; Decision making unit 302: used to distinguish patients into odontogenic toothache and non-odontogenic toothache based on the toothache information, wherein the distinction includes: S21: if the toothache of the patient cannot be located, it is determined that the patient has non-odontogenic toothache; S22: if the toothache location of the patient is uncertain, determining that the patient has odontogenic toothache; S23: If the toothache of the patient can be located, and the toothache occurs in the maxillary premolars or molars or the left mandibular posterior teeth, the patient is judged to have non-odontogenic toothache. If the toothache of the patient can be located, and the toothache occurs in the third molars or first molars, the patient is judged to have odontogenic toothache.

[0065] Figure 4 is a schematic flow chart of a clinically assisted odontogenic toothache diagnosis system based on integrated learning provided by an embodiment of the present invention; A clinically assisted odontogenic toothache diagnosis system based on ensemble learning, the system comprising: An acquisition unit 401 is used to acquire answers of patients with odontogenic toothache to a toothache screening questionnaire; Decision making unit 402 is used to determine the subtype of odontogenic toothache according to the patient's answer to the toothache screening questionnaire.

[0066] Figure 5 It is a schematic block diagram of a clinical auxiliary toothache diagnosis device based on a decision tree or a clinical auxiliary odontogenic toothache diagnosis device based on ensemble learning provided by an embodiment of the present invention; A clinical auxiliary toothache diagnosis device based on a decision tree, comprising: a memory and a processor; The memory is used to store program instructions; The processor is used to call program instructions, and when the program instructions are executed, the steps of the above-mentioned artificial intelligence-based toothache diagnosis method are implemented.

[0067] A clinical auxiliary odontogenic toothache diagnosis device based on integrated learning comprises: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, the odontogenic toothache diagnosis method based on artificial intelligence is implemented.

[0068] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the artificial intelligence-based odontogenic toothache diagnosis method.

[0069] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned artificial intelligence-based toothache diagnosis method.

[0070] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0071] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0072] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0073] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0074] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0075] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0076] The above is a detailed introduction to a computer device provided by the present invention. For a person skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A clinical decision support system based on decision tree, characterized in that: The system comprises: An acquisition unit is used to acquire the patient's pain information, wherein the pain information includes whether the pain site can be located and the pain site; Decision-making unit: used to input the pain information into the decision tree model and then output the cause of the patient's pain; the decision-making unit includes two submodules: The first submodule: used to obtain the first-level judgment result based on whether the pain site can be located; The second submodule is used to obtain the second-level judgment result based on the first-level judgment result and the pain location; The decision unit outputs the cause of the patient's pain based on the second-level judgment result.

2. The clinical decision support system based on decision tree according to claim 1, characterized in that: The decision unit further includes a third submodule: configured to obtain a third-level judgment result based on the second-level judgment result and the nature and duration of pain, and the decision unit outputs the cause of the patient's pain based on the third-level judgment result.

3. A clinical decision-making support device based on a decision tree, characterized in that: include: Memory and processor; The memory is used to store program instructions; The processor is used to call program instructions, and when the program instructions are executed, the clinical decision support system based on the decision tree described in any one of claims 1-2 is implemented.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the decision tree-based clinical decision support system according to any one of claims 1 to 2 is implemented.

5. A clinical decision support system based on ensemble learning, characterized in that: The system comprises: Get scale answer unit: used to obtain the patient's answer to the screening questionnaire, Scale decision unit: used to input the answers into an integrated learning classifier to obtain subclass groupings of pain causes.

6. The clinical decision support system based on integrated learning according to claim 5, characterized in that: The ensemble learning classifier is an ensemble learning based on the Fisher discriminant equation, and the ensemble learning includes the Fisher discriminant equations of different subclass groupings. After the patient's answers to the screening questionnaire are input into the ensemble learning classifier, the subclass grouping corresponding to the Fisher discriminant equation with the highest Fisher discriminant equation score of the different subgroups is output.

7. The clinical decision support system based on integrated learning according to claim 6, characterized in that: The Fisher discriminant equations of the different subclass groups are obtained by fitting the Fisher discriminant equations to the different subclass groups based on the clinical data set to obtain the Fisher discriminant equations of the different subclass groups.

8. The clinical decision support system based on integrated learning according to claim 5, characterized in that: The screening questionnaire is a Chinese localized version obtained by cross-cultural adaptation of the original DePaQ, and includes 13 items; Optionally, the cross-cultural adaptation steps include: Step 1: Different people translated the original DePaQ into Chinese to obtain different translations; Step 2: Combining the different translations to obtain a unified Chinese translation; Step 3: The unified Chinese translation is back-translated into English to obtain a back-translated English version; Step 4: The expert group formulates the final translation by combining the original DePaQ, the different translations, the unified Chinese translation, and the back-translated English version; Step 5: Based on the final translation, the items of the final translation are revised according to the localization data set to obtain the Chinese localized version; Optionally, the revision includes: deleting inappropriate entries of the Chinese localized version, adding localized adapted entries, and modifying options for entries of the Chinese localized version.

9. A clinical decision support device based on ensemble learning, characterized in that: include: Memory and processor; The memory is used to store program instructions; the processor is used to call program instructions, and when the program instructions are executed, the clinical decision support system based on integrated learning as described in any one of claims 5-8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements a clinical decision support system based on integrated learning as described in any one of claims 5 to 8.

Images