CN114613498A - Auxiliary MDT clinical decision method, system and equipment based on machine learning - Google Patents
Auxiliary MDT clinical decision method, system and equipment based on machine learning Download PDFInfo
- Publication number
- CN114613498A CN114613498A CN202210296871.7A CN202210296871A CN114613498A CN 114613498 A CN114613498 A CN 114613498A CN 202210296871 A CN202210296871 A CN 202210296871A CN 114613498 A CN114613498 A CN 114613498A
- Authority
- CN
- China
- Prior art keywords
- treatment
- cancer patient
- mdt
- survival time
- predicted
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/20—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for electronic clinical trials or questionnaires
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Epidemiology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Primary Health Care (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
The invention relates to a machine learning-based MDT clinical decision making assisting method, a machine learning-based MDT clinical decision making system and machine learning-based MDT clinical decision making equipment. The method comprises the following steps: acquiring clinical data of a cancer patient to be detected; inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life time of the cancer patient to be detected; inputting the clinical data of the cancer patient to be detected and the MDT clinical decision treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme; calculating a difference between the second predicted lifetime and the first predicted lifetime; and outputting a result of the auxiliary MDT clinical decision according to the difference. The method has good application value in MDT clinical decision.
Description
Technical Field
The invention relates to the technical field of intelligent medical treatment, in particular to an auxiliary MDT clinical decision method, an auxiliary MDT clinical decision system, auxiliary MDT clinical decision equipment and a computer-readable storage medium based on machine learning.
Background
The multidisciplinary diagnosis and treatment model (Multi disciplinery team MDT) is a working group consisting of department experts from surgery, oncology, radiotherapy, imaging, pathology, etc., proposes an optimal treatment plan suitable for a patient through a regular consultation form for a certain disease, and then performs the treatment plan by related disciplines or multidisciplinary combination. In recent years, the emergence of new methods such as targeted therapy and immunotherapy brings hope to some refractory cancer patients, and the operation time is strived for some non-operable cancer patients, but the selection of the optimal treatment scheme is a challenge for expert groups.
Disclosure of Invention
The method, the system and the equipment for assisting the MDT clinical decision-making based on machine learning are provided based on a clinical scene and by taking the survival time of a patient as an evaluation endpoint, and a computer readable storage medium is provided to assist MDT experts in making the optimal clinical decision.
The application aims to provide an auxiliary MDT clinical decision method based on machine learning, which comprises the following steps:
acquiring clinical data of a cancer patient to be detected;
inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life time of the cancer patient to be detected;
inputting the clinical data of the cancer patient to be detected and the MDT clinical decision treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
calculating a difference between the second predicted lifetime and the first predicted lifetime;
and outputting a result of the auxiliary MDT clinical decision according to the difference.
Further, the treatment plan of the MDT clinical decision includes any one of surgery, local regional treatment, chemotherapy, radiotherapy, surgery-combined chemotherapy, local regional treatment-combined radiotherapy, local regional treatment-combined chemotherapy, immunotherapy or other new treatment plans, and the clinical data of the cancer patient to be tested and any one of the surgery, local regional treatment, chemotherapy, radiotherapy, surgery-combined chemotherapy, local regional treatment-combined radiotherapy, local regional treatment-combined chemotherapy, immunotherapy or other new treatment plans are input into a pre-trained treatment evaluation model, so as to obtain a second predicted survival time of the cancer patient to be tested after the treatment plan.
Further, the clinical data of the cancer patient to be detected and any one treatment scheme of the MDT clinical decision are input into a pre-trained treatment evaluation model, a second predicted survival time of the cancer patient to be detected after the treatment scheme of each MDT clinical decision is obtained, a second predicted survival time corresponding to the treatment scheme with the longest second predicted survival time is reserved, and a difference value between the second predicted survival time and the first predicted survival time is calculated.
Further, the clinical data comprises hospitalization and treatment information data, pathological data, image examination data and laboratory examination data.
Further, the method for constructing the pre-trained life cycle prediction model comprises the following steps: acquiring a clinical data set of a cancer patient as a training set, comparing the predicted survival time of the cancer patient with the actual survival time, generating a loss value, performing back propagation, and optimizing a survival prediction model to obtain the pre-trained survival prediction model;
optionally, an algorithm adopted by the pre-trained lifetime prediction model is selected from multiple machine learning algorithms, and the machine learning algorithm is selected from one or more of the following machine learning algorithms: random forest, logistic regression, linear regression, polynomial regression, stepwise regression, ridge regression, lasso regression, elastic regression.
Further, the method further comprises the steps of obtaining an MDT clinical decision according to the first predicted survival time, wherein the obtaining of the MDT clinical decision according to the first predicted survival time is to compare the first predicted survival time with the median of the cancer survival time of the cancer patient to be detected, when the first predicted survival time is larger than the median of the cancer survival time, the obtained MDT clinical decision is a traditional treatment scheme, and the clinical data of the cancer patient to be detected and the traditional treatment scheme are input into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme; when the first predicted survival time is smaller than the median of the cancer survival period, acquiring an MDT clinical decision as an immunotherapy scheme, and inputting the clinical data of the cancer patient to be detected and the immunotherapy scheme into a pre-trained treatment evaluation model to acquire a second predicted survival time of the cancer patient to be detected after the therapy scheme; optionally, the conventional treatment scheme includes surgery, regional local treatment, chemotherapy, radiotherapy, surgery combined chemotherapy, regional local treatment combined radiotherapy, and regional local treatment combined chemotherapy; such immunotherapeutic regimens include immune checkpoint, etc., immunotherapy, targeted therapy, or other novel therapeutic regimens.
It is an object of the present application to provide a machine learning based assisted MDT clinical decision system, the system comprising:
the acquiring unit is used for acquiring clinical data of a cancer patient to be detected;
the first prediction unit is used for inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life cycle of the cancer patient to be detected;
the second prediction unit is used for inputting the clinical data of the cancer patient to be detected and the treatment scheme of the MDT clinical decision into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
a calculating unit for calculating a difference between the second predicted lifetime and the first predicted lifetime;
and the output unit is used for outputting the result of the auxiliary MDT clinical decision according to the difference.
Further, the system also comprises a judging unit which is positioned behind the first prediction unit, for comparing the first predicted survival time to the median number of cancer survival for the cancer patient to be tested, the second prediction unit comprises a traditional treatment scheme second prediction unit and an immunotherapy scheme second prediction unit, the traditional treatment scheme second prediction unit is used for inputting the clinical data of the cancer patient to be tested and the traditional treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be tested after the treatment scheme, the second prediction unit of the immunotherapy scheme is used for inputting the clinical data of the cancer patient to be tested and the immunotherapy scheme into a pre-trained therapy evaluation model to obtain a second predicted survival time of the cancer patient to be tested after the therapy scheme; when the first predicted survival time is longer than the median of the cancer survival period, acquiring an MDT clinical decision as a traditional treatment scheme, inputting the clinical data of the cancer patient to be detected and the traditional treatment scheme into a pre-trained treatment evaluation model, and acquiring a second predicted survival time of the cancer patient to be detected after the treatment scheme; and when the first predicted survival time is less than the median of the cancer survival period, obtaining an MDT clinical decision as an immunotherapy scheme, inputting the clinical data of the cancer patient to be detected and the immunotherapy scheme into a pre-trained treatment evaluation model, and obtaining a second predicted survival time of the cancer patient to be detected after the therapy scheme.
The application aims to provide an auxiliary MDT clinical decision device based on machine learning, which comprises: a memory and a processor;
the memory is to store program instructions;
the processor is configured to invoke program instructions that, when executed, implement the above-described machine learning-based assisted MDT clinical decision making method steps.
The present application is directed to a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned method steps of a machine learning-based auxiliary MDT clinical decision making.
The advantages of the application are that:
1. the method is based on a clinical scene, clinical data of a patient are input into a pre-trained life cycle prediction model to obtain first predicted life time of the cancer patient to be detected, and whether the patient is suitable for immunotherapy or traditional therapy is judged according to the comparison result of the first predicted life time and median of the life cycle of the type of cancer;
2. inputting clinical data of a cancer patient to be detected and a treatment scheme of any MDT clinical decision into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme, and selecting the treatment scheme with the longest second predicted survival time from a plurality of treatment schemes to provide a predicted result of each treatment scheme and an optimal treatment scheme for clinical experts and patients to select;
3. the application reflects the influence on the survival time of the patient before and after the treatment of each treatment scheme by calculating the difference value between the second predicted survival time and the first predicted survival time, provides the intervention condition of the optimal treatment scheme on the state of an illness of the patient for clinical experts and the patient, and leads the patient to have more intuitive understanding on the treatment scheme by the visualization of the result, thereby being beneficial to relieving the potential medical contradiction.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flow chart of a method for assisting MDT clinical decision based on machine learning according to an embodiment of the present invention;
FIG. 2 is a schematic block diagram of an auxiliary MDT clinical decision making apparatus based on machine learning provided by an embodiment of the present invention;
fig. 3 is a schematic block diagram of an auxiliary MDT clinical decision making system based on machine learning according to an embodiment of the present invention;
fig. 4 is a schematic flow chart of a method for assisting MDT clinical decision based on machine learning according to an embodiment of the present invention.
Fig. 5 is an interface schematic diagram of an auxiliary MDT clinical decision method based on machine learning according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention.
In some of the flows described in the present specification and claims and in the above-described figures, a number of operations are included that occur in a particular order, but it should be clearly understood that these operations may be performed out of order or in parallel as they occur herein, with the order of the operations, e.g., S101, S102, etc., merely being used to distinguish between various operations, and the order of the operations itself does not represent any order of performance. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor limit the types of "first" and "second" to be different.
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of a method for assisting MDT clinical decision based on machine learning according to an embodiment of the present invention, specifically, the method includes the following steps:
s101: acquiring clinical data of a cancer patient to be detected;
in one embodiment, the clinical data includes hospitalization information data, pathology data, imaging examination data, laboratory examination data. Optionally, the clinical data includes data of age, sex, TNM stage, etc. of the cancer patient to be tested. The TNM staging uses the patient's raw data, such as tumor size, number of tumors, peritoneal invasion, vascular invasion, lymph node metastasis, distant metastasis, etc., to determine TNM staging. We stage TNM into five grades, I, II, IIIA, IIIB and IV, mainly according to AJCC eighth edition. When the original data is missing or only the seventh version of AJCC category data, we partition according to the version of the standard.
S102: inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life cycle of the cancer patient to be detected;
in one embodiment, the method further comprises obtaining an MDT clinical decision according to the first predicted survival time, wherein the obtaining of the MDT clinical decision according to the first predicted survival time is to compare the first predicted survival time with a median cancer survival time of the cancer patient to be tested, obtain the MDT clinical decision as a traditional treatment plan when the first predicted survival time is greater than the median cancer survival time, and input clinical data of the cancer patient to be tested and the traditional treatment plan into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be tested after the treatment plan; when the first predicted survival time is smaller than the median of the cancer survival period, acquiring an MDT clinical decision as an immunotherapy scheme, and inputting the clinical data of the cancer patient to be detected and the immunotherapy scheme into a pre-trained treatment evaluation model to acquire a second predicted survival time of the cancer patient to be detected after the therapy scheme; optionally, the conventional treatment scheme includes surgery, regional local treatment, chemotherapy, radiotherapy, surgery combined chemotherapy, regional local treatment combined radiotherapy, and regional local treatment combined chemotherapy; the immunotherapy regimen includes immunotherapy such as immune checkpoints or other novel therapeutic regimens. Optionally, the median cancer life can be adjusted according to the situation, or can be artificially set to a specific value, for example, the median liver cancer life is artificially set to 11 months.
In one embodiment, the method for constructing the pre-trained lifetime prediction model includes: acquiring a clinical data set of a cancer patient as a training set, comparing the predicted survival time of the cancer patient with the actual survival time, generating a loss value, performing back propagation, and optimizing a survival prediction model to obtain the pre-trained survival prediction model; optionally, the algorithm adopted by the pre-trained lifetime prediction model is selected from multiple machine learning algorithms, and the machine learning algorithm is selected from one or more of the following machine learning algorithms: random forest, logistic regression, linear regression, polynomial regression, stepwise regression, ridge regression, lasso regression, elastic regression.
In one embodiment, the pre-trained treatment evaluation model is constructed by: and acquiring a clinical data set (including clinical data, a treatment scheme and the life cycle of the patient) of the cancer patient to be treated as a training set, comparing the predicted life cycle of the cancer patient with the actual life cycle, generating a loss value, performing back propagation, and optimizing a treatment evaluation model to obtain the pre-trained treatment evaluation model. Furthermore, the clinical data set of the cancer patient receiving treatment is dynamically updated, new clinical data of the cancer patient receiving treatment is periodically brought into the training data set, and secondary optimization of the model is performed to obtain an updated pre-trained treatment evaluation model.
S103: inputting the clinical data of the cancer patient to be detected and the MDT clinical decision treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
in one embodiment, the MDT clinical decision-making treatment plan includes any one of a surgical treatment, a regional treatment, chemotherapy, radiotherapy, a surgical combination chemotherapy, a regional treatment combination radiotherapy, a regional treatment combination chemotherapy, immunotherapy or other new treatment plan, and the clinical data of the cancer patient to be tested and any one of the surgical treatment, the regional treatment, chemotherapy, radiotherapy, a surgical combination chemotherapy, a regional treatment combination radiotherapy, a regional treatment combination chemotherapy, a regional treatment combination radiotherapy, immunotherapy or other new treatment plan are input into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be tested after the treatment plan.
In one embodiment, the clinical data of the cancer patient to be tested and any one treatment scheme of the MDT clinical decision are input into a pre-trained treatment evaluation model, so as to obtain a second predicted survival time of the cancer patient to be tested after the treatment scheme of each MDT clinical decision, retain a second predicted survival time corresponding to the treatment scheme with the longest second predicted survival time, and calculate a difference between the second predicted survival time and the first predicted survival time.
In one embodiment, the data used for model training of the present application mainly relates to two parts of data, one part of which is from the fifth medical center of the general hospital of the chinese liberty, and 1240 patients suspected of being diagnosed with ICC were co-registered during the study (7 months to 2 months of 2021, 2007). We excluded patients with imaging ICC compliance but no pathological diagnosis (n 410), patients with ICC pathological diagnosis in the hospital (n 90), patients with biliary cancer at the periphery (n 112) or distal end (n 45), patients with mixed or combined hepatocellular carcinoma-cholangiocarcinoma (n 68), and patients with gallbladder cancer (n 1). In addition, we also excluded patients who died within 1 month after resection (n-5) and patients who were only seen at baseline without long-term follow-up (n-67). Of 519 pathologically confirmed ICC patients, only 504 were recorded with TNM staging. These 504 were included in our study. The other is from a large public database, we included patients with intrahepatic cholangiocarcinoma with histological diagnosis of valid follow-up data in the SEER plus database from 2000 to 2018. The disease ICD-10-CM of the enrolled patients was encoded as C22.1 and ICD-O-3 was encoded as 8160/3. Patients diagnosed with cancer, as evidenced by necropsy or death alone, and patients who were not complete with follow-up data and histologically positive, were excluded. Only patients 15 years and older and with ethnic records were enrolled. Tumor staging was encoded according to the seventh and eighth edition TNM staging system of the united states cancer joint committee. The total number of the admitted patients is 4398. The two sets of data act as a training set and an external validation set with respect to each other. In part of the study, they were mixed together as one data set and further randomly divided into a training set and a test set. The performance of the model is measured in terms of the consistency index of the training set and the test set, and the consistency index of the out-of-bag estimate based on the training set. Time dependent AUC was also calculated to evaluate the model. As part of the model evaluation, the 95% confidence interval for each index was calculated by training 500 times a portion of the samples (60%) of the training and test sets. To evaluate the impact of each feature, we calculated the importance of the feature by measuring the reduction in test score after randomly arranging one feature, and also used SHAP to interpret the model.
S104: calculating a difference between the second predicted lifetime and the first predicted lifetime; .
In one embodiment, the first predicted survival time reflects the survival time of the cancer patient to be tested when not receiving any treatment, the second predicted survival time reflects the survival time of the cancer patient to be tested after receiving the treatment plan of the optimal MDT clinical decision, and the difference between the first predicted survival time and the second predicted survival time reflects the influence on the survival time of the cancer patient to be tested after receiving the corresponding treatment.
S105: and outputting a result of the auxiliary MDT clinical decision according to the difference.
In one embodiment, as shown in fig. 4, clinical data of a cancer patient to be tested is obtained; inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life time of the cancer patient to be detected; comparing the first predicted survival time with the median of the cancer life time of the cancer patient to be detected, obtaining an MDT clinical decision as a traditional treatment scheme when the first predicted survival time is greater than the median of the cancer life time, and respectively inputting the clinical data of the cancer patient to be detected and each traditional treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after each traditional treatment scheme; calculating a difference between the second predicted survival time and the first predicted survival time after each conventional treatment regimen; outputting the treatment scheme with the longest second predicted survival time and the survival time difference thereof as the result of the clinical decision of the auxiliary MDT; optionally, a second predicted survival time of the cancer patient to be tested after each conventional treatment scheme is obtained, the second predicted survival time corresponding to the treatment scheme with the longest second predicted survival time is reserved, a difference between the second predicted survival time and the first predicted survival time is calculated, and the treatment scheme with the longest second predicted survival time and the difference between the second predicted survival time and the first predicted survival time are output as a result of the clinical decision of the auxiliary MDT.
In one embodiment, clinical data of a test cancer patient is obtained; inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life time of the cancer patient to be detected; comparing the first predicted survival time with the median cancer survival time of the cancer patient to be tested, and when the first predicted survival time is less than the median cancer survival time, obtaining an MDT clinical decision as an immunotherapy scheme; inputting the clinical data and the immunotherapy scheme of the cancer patient to be detected into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after each immunotherapy scheme, and calculating a difference value between the second predicted survival time and the first predicted survival time after each immunotherapy scheme; outputting the treatment scheme with the longest second predicted survival time and the survival time difference thereof as the result of the clinical decision of the auxiliary MDT; optionally, a second predicted survival time of the cancer patient to be tested after each immunotherapy scheme is obtained, the second predicted survival time corresponding to the therapy scheme with the longest second predicted survival time is reserved, a difference between the second predicted survival time and the first predicted survival time is calculated, and the therapy scheme with the longest second predicted survival time and the difference between the second predicted survival time and the first predicted survival time are output as a result of the clinical decision of the auxiliary MDT.
Fig. 2 is a device for assisting MDT clinical decision based on machine learning, provided in an embodiment of the present invention, and includes: a memory and a processor;
the memory is to store program instructions;
the processor is configured to invoke program instructions that when executed implement the following machine learning-based assisted MDT clinical decision method steps:
acquiring clinical data of a cancer patient to be detected;
inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life time of the cancer patient to be detected;
inputting the clinical data of the cancer patient to be detected and the MDT clinical decision treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
calculating a difference between the second predicted lifetime and the first predicted lifetime;
and outputting a result of the auxiliary MDT clinical decision according to the difference.
Fig. 3 is a machine learning-based auxiliary MDT clinical decision making system provided by an embodiment of the present invention, including:
the acquiring unit is used for acquiring clinical data of a cancer patient to be detected;
the first prediction unit is used for inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life cycle of the cancer patient to be detected;
the second prediction unit is used for inputting the clinical data of the cancer patient to be detected and the treatment scheme of the MDT clinical decision into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
a calculating unit for calculating a difference between the second predicted lifetime and the first predicted lifetime;
and the output unit is used for outputting the result of the auxiliary MDT clinical decision according to the difference.
In a particular embodiment, a machine learning-based assisted MDT clinical decision system includes:
the acquiring unit is used for acquiring clinical data of a cancer patient to be detected;
the first prediction unit is used for inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life cycle of the cancer patient to be detected;
the judging unit is used for comparing the first predicted survival time with the median of the cancer survival time of the cancer patient to be detected, obtaining an MDT clinical decision as a traditional treatment scheme when the first predicted survival time is greater than the median of the cancer survival time, and obtaining the MDT clinical decision as an immunotherapy scheme when the first predicted survival time is less than the median of the cancer survival time;
the second prediction unit comprises a traditional treatment scheme second prediction unit and an immunotherapy scheme second prediction unit, the traditional treatment scheme second prediction unit is used for inputting the clinical data and the traditional treatment scheme of the cancer patient to be detected into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme, and the immunotherapy scheme second prediction unit is used for inputting the clinical data and the immunotherapy scheme of the cancer patient to be detected into the pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
a calculating unit for calculating a difference between the second predicted lifetime and the first predicted lifetime;
and the output unit is used for outputting the result of the auxiliary MDT clinical decision according to the difference.
In one embodiment, as shown in fig. 5, the clinical data of the intrahepatic cholangiocellular carcinoma patient to be tested and any one of the MDT clinically decided treatment schemes are input into a pre-trained treatment evaluation model, so as to obtain the second predicted survival time of the intrahepatic cholangiocellular carcinoma patient to be tested after each MDT clinically decided treatment scheme.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the above-mentioned machine learning-based aided MDT clinical decision method steps.
The validation results of this validation example show that assigning an intrinsic weight to an indication can improve the performance of the method moderately over the default setting.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: read Only Memory (ROM), Random Access Memory (RAM), magnetic or optical disks, and the like.
It will be understood by those skilled in the art that all or part of the steps in the method for implementing the above embodiments may be implemented by hardware that is instructed to implement by a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
While the invention has been described in detail with reference to specific embodiments thereof, it will be apparent to one skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.
Claims (10)
1. A machine learning-based assisted MDT clinical decision method, comprising:
acquiring clinical data of a cancer patient to be detected;
inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life time of the cancer patient to be detected;
inputting the clinical data of the cancer patient to be detected and the MDT clinical decision treatment scheme into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
calculating a difference between the second predicted lifetime and the first predicted lifetime;
and outputting a result of the auxiliary MDT clinical decision according to the difference.
2. The machine learning-based aided MDT clinical decision method of claim 1, the method is characterized in that the MDT clinical decision-making treatment scheme comprises any one of surgery treatment, local regional treatment, chemotherapy, radiotherapy, surgery combined chemotherapy, local regional treatment combined radiotherapy, local regional treatment combined chemotherapy, immunotherapy or other new treatment schemes, and the clinical data of the cancer patient to be detected and any one of the surgery treatment, local regional treatment, chemotherapy, radiotherapy, surgery combined chemotherapy, local regional treatment combined radiotherapy, local regional treatment combined chemotherapy, immunotherapy or other new treatment schemes are input into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme.
3. The method as claimed in claim 2, wherein the clinical data of the cancer patient to be tested and any one of the MDT clinical decision-making treatment schemes are inputted into a pre-trained treatment evaluation model, so as to obtain a second predicted survival time of the cancer patient to be tested after each MDT clinical decision-making treatment scheme, the second predicted survival time corresponding to the treatment scheme with the longest second predicted survival time is retained, and the difference between the second predicted survival time and the first predicted survival time is calculated.
4. The machine-learning-based aided MDT clinical decision method of claim 1, wherein the clinical data comprises hospitalization information data, pathology data, imaging examination data, laboratory examination data.
5. The machine learning-based auxiliary MDT clinical decision making method according to claim 1, wherein the pre-trained life-cycle prediction model is constructed by: acquiring a clinical data set of a cancer patient as a training set, comparing the predicted survival time of the cancer patient with the actual survival time, generating a loss value, performing back propagation, and optimizing a survival prediction model to obtain the pre-trained survival prediction model; optionally, the algorithm adopted by the pre-trained lifetime prediction model is selected from multiple machine learning algorithms, and the machine learning algorithm is selected from one or more of the following machine learning algorithms: random forest, logistic regression, linear regression, polynomial regression, stepwise regression, ridge regression, lasso regression, elastic regression.
6. The method of claim 1, further comprising deriving an MDT clinical decision based on the first predicted survival time, wherein the deriving the MDT clinical decision based on the first predicted survival time is comparing the first predicted survival time with a median cancer survival time of the cancer patient to be tested, and when the first predicted survival time is greater than the median cancer survival time, deriving the MDT clinical decision as a conventional treatment plan, and inputting clinical data of the cancer patient to be tested and the conventional treatment plan into a pre-trained treatment evaluation model to derive a second predicted survival time of the cancer patient to be tested after the treatment plan; when the first predicted survival time is smaller than the median of the cancer survival period, acquiring an MDT clinical decision as an immunotherapy scheme, and inputting the clinical data of the cancer patient to be detected and the immunotherapy scheme into a pre-trained treatment evaluation model to acquire a second predicted survival time of the cancer patient to be detected after the therapy scheme; optionally, the conventional treatment scheme includes surgery, regional local treatment, chemotherapy, radiotherapy, surgery combined chemotherapy, regional local treatment combined radiotherapy, and regional local treatment combined chemotherapy; the immunotherapy regimen includes immunotherapy such as immune checkpoint, targeted therapy or other novel therapeutic regimens.
7. The machine-learning-based assisted MDT clinical decision system according to claim 1, wherein the system comprises:
the acquiring unit is used for acquiring clinical data of a cancer patient to be detected;
the first prediction unit is used for inputting the clinical data into a pre-trained life cycle prediction model to obtain a first predicted life cycle of the cancer patient to be detected;
the second prediction unit is used for inputting the clinical data of the cancer patient to be detected and the treatment scheme of the MDT clinical decision into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be detected after the treatment scheme;
a calculating unit for calculating a difference between the second predicted lifetime and the first predicted lifetime;
and the output unit is used for outputting the result of the auxiliary MDT clinical decision according to the difference.
8. The machine-learning-based auxiliary MDT clinical decision system of claim 7, further comprising a determination unit, the determination unit being located after the first prediction unit and configured to compare the first predicted survival time with the median number of cancer survival time of the cancer patient to be tested, the second prediction unit comprising a traditional treatment plan second prediction unit and an immunotherapy plan second prediction unit, the traditional treatment plan second prediction unit being configured to input the clinical data of the cancer patient to be tested and a traditional treatment plan into a pre-trained treatment evaluation model to obtain a second predicted survival time of the cancer patient to be tested after the treatment plan, the immunotherapy plan second prediction unit being configured to input the clinical data of the cancer patient to be tested and an immunotherapy plan into the pre-trained treatment evaluation model, obtaining a second predicted survival time of the cancer patient to be tested after the treatment regimen; when the first predicted survival time is longer than the median of the cancer survival period, acquiring an MDT clinical decision as a traditional treatment scheme, inputting the clinical data of the cancer patient to be detected and the traditional treatment scheme into a pre-trained treatment evaluation model, and acquiring a second predicted survival time of the cancer patient to be detected after the treatment scheme; and when the first predicted survival time is less than the median of the cancer survival period, obtaining an MDT clinical decision as an immunotherapy scheme, inputting the clinical data of the cancer patient to be detected and the immunotherapy scheme into a pre-trained treatment evaluation model, and obtaining a second predicted survival time of the cancer patient to be detected after the therapy scheme.
9. A machine learning-based assisted MDT clinical decision making apparatus, comprising: a memory and a processor;
the memory is to store program instructions;
the processor is configured to invoke program instructions that, when executed, implement the machine learning-based assisted MDT clinical decision making method steps of any of claims 1-6.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method steps of the machine learning-based assisted MDT clinical decision making method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210296871.7A CN114613498B (en) | 2022-03-24 | 2022-03-24 | Machine learning-based MDT (minimization drive test) clinical decision making assisting method, system and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210296871.7A CN114613498B (en) | 2022-03-24 | 2022-03-24 | Machine learning-based MDT (minimization drive test) clinical decision making assisting method, system and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114613498A true CN114613498A (en) | 2022-06-10 |
| CN114613498B CN114613498B (en) | 2022-12-13 |
Family
ID=81865302
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210296871.7A Active CN114613498B (en) | 2022-03-24 | 2022-03-24 | Machine learning-based MDT (minimization drive test) clinical decision making assisting method, system and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114613498B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115440383A (en) * | 2022-09-30 | 2022-12-06 | 中国医学科学院北京协和医院 | System for predicting curative effect of PD-1/PD-L1 monoclonal antibody of advanced cancer patient |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109259780A (en) * | 2018-07-24 | 2019-01-25 | 南方医科大学南方医院 | The aided assessment system and method that gastric cancer prognosis and chemotherapy based on enhancing CT images group benefit |
| CN110551819A (en) * | 2019-08-23 | 2019-12-10 | 伯克利南京医学研究有限责任公司 | Application of group of ovarian cancer prognosis related genes |
| CN110580956A (en) * | 2019-09-19 | 2019-12-17 | 青岛市市立医院 | liver cancer prognosis markers and application thereof |
| CN113241183A (en) * | 2021-04-30 | 2021-08-10 | 深圳睿心智能医疗科技有限公司 | Treatment scheme prediction method and device |
| CN113444804A (en) * | 2021-07-14 | 2021-09-28 | 武汉大学中南医院 | Cervical cancer prognosis related gene and application thereof in preparation of cervical cancer prognosis prediction and diagnosis product |
-
2022
- 2022-03-24 CN CN202210296871.7A patent/CN114613498B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109259780A (en) * | 2018-07-24 | 2019-01-25 | 南方医科大学南方医院 | The aided assessment system and method that gastric cancer prognosis and chemotherapy based on enhancing CT images group benefit |
| CN110551819A (en) * | 2019-08-23 | 2019-12-10 | 伯克利南京医学研究有限责任公司 | Application of group of ovarian cancer prognosis related genes |
| CN110580956A (en) * | 2019-09-19 | 2019-12-17 | 青岛市市立医院 | liver cancer prognosis markers and application thereof |
| CN113241183A (en) * | 2021-04-30 | 2021-08-10 | 深圳睿心智能医疗科技有限公司 | Treatment scheme prediction method and device |
| CN113444804A (en) * | 2021-07-14 | 2021-09-28 | 武汉大学中南医院 | Cervical cancer prognosis related gene and application thereof in preparation of cervical cancer prognosis prediction and diagnosis product |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115440383A (en) * | 2022-09-30 | 2022-12-06 | 中国医学科学院北京协和医院 | System for predicting curative effect of PD-1/PD-L1 monoclonal antibody of advanced cancer patient |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114613498B (en) | 2022-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11869202B2 (en) | Method and system for processing multi-modality image | |
| JP5523342B2 (en) | Method and apparatus for refining similar case search | |
| US7606405B2 (en) | Dynamic tumor diagnostic and treatment system | |
| US11037070B2 (en) | Diagnostic test planning using machine learning techniques | |
| US7653227B2 (en) | Hierarchical modeling in medical abnormality detection | |
| BRPI1006388A2 (en) | clinical decision support system (adc) and clinical decision support method (adc) implemented by an adc system | |
| CN115036002B (en) | Treatment effect prediction method based on multi-mode fusion model and terminal equipment | |
| JP2006500075A (en) | Patient data mining for lung cancer discrimination | |
| US11790528B2 (en) | Preoperative survival prediction method based on enhanced medical images and computing device using thereof | |
| JP2019526869A (en) | CAD system personalization method and means for providing confidence level indicators for CAD system recommendations | |
| CN112542247A (en) | Method and system for predicting probability of complete remission of pathology after breast cancer neoadjuvant chemotherapy | |
| CN114613498B (en) | Machine learning-based MDT (minimization drive test) clinical decision making assisting method, system and equipment | |
| WO2006119340A2 (en) | Dynamic tumor diagnostic and treatment system | |
| CN107256344A (en) | Data processing method, device and radiotherapy management system | |
| WO2022191781A1 (en) | Method and system for assessing nonalcoholic steatohepatitis | |
| Scapicchio et al. | A multicenter evaluation of a deep learning software (LungQuant) for lung parenchyma characterization in COVID-19 pneumonia | |
| CN110234270B (en) | Method and apparatus for providing quantitative volume maps of organs or assessment of organ health | |
| CN110992312B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
| Gross et al. | Improved performance and consistency of deep learning 3D liver segmentation with heterogeneous cancer stages in magnetic resonance imaging | |
| Gong et al. | Unified ICH quantification and prognosis prediction in NCCT images using a multi-task interpretable network | |
| CN116259396A (en) | Treatment expense prediction method, system, equipment and storage medium based on machine learning | |
| CN113706541B (en) | Image processing method and device | |
| CN113284619A (en) | Method and equipment for predicting biochemical recurrence of prostate cancer patient after operation | |
| CN114927216A (en) | Method and system for predicting treatment effect of PD-1 of melanoma patient based on artificial intelligence | |
| Yang et al. | Lung Nodule Segmentation and Uncertain Region Prediction with an Uncertainty-Aware Attention Mechanism |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |