CN111297474B - Individualized positioning and mapping system for auricular fibrillation focus - Google Patents
Individualized positioning and mapping system for auricular fibrillation focus Download PDFInfo
- Publication number
- CN111297474B CN111297474B CN201911315008.6A CN201911315008A CN111297474B CN 111297474 B CN111297474 B CN 111297474B CN 201911315008 A CN201911315008 A CN 201911315008A CN 111297474 B CN111297474 B CN 111297474B
- Authority
- CN
- China
- Prior art keywords
- model
- voltage
- data
- focus
- value
- 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.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B90/00—Instruments, implements or accessories specially adapted for surgery or diagnosis and not covered by any of the groups A61B1/00 - A61B50/00, e.g. for luxation treatment or for protecting wound edges
- A61B90/10—Instruments, implements or accessories specially adapted for surgery or diagnosis and not covered by any of the groups A61B1/00 - A61B50/00, e.g. for luxation treatment or for protecting wound edges for stereotaxic surgery, e.g. frame-based stereotaxis
-
- 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
- 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
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
- A61B2034/101—Computer-aided simulation of surgical operations
- A61B2034/105—Modelling of the patient, e.g. for ligaments or bones
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/10—Computer-aided planning, simulation or modelling of surgical operations
- A61B2034/107—Visualisation of planned trajectories or target regions
Abstract
The invention relates to an individualized positioning and mapping system for atrial fibrillation focuses, which comprises a data collection module: the system is used for collecting data, establishing a database by using the collected data, and carrying out data coding and label establishment on the collected data; an analysis module: obtaining healthy atrial voltage data when atrial fibrillation does not occur, and obtaining a healthy atrial model; a verification module: the method is used for data quality control, and the established model is corrected to obtain a corrected and matched healthy atrium model; a calculation module: establishing a multiple regression estimation model by using relative voltage; identifying and mapping module: outputting a visual focus area map; a feedback module: and receiving a follow-up database, and further judging the accuracy of modeling (R, rc model) during the operation according to the standard of the operation result and giving a correction suggestion. The invention can effectively realize the positioning and the mapping of the atrial fibrillation focus and improve the curative effect of the atrial fibrillation radio frequency ablation.
Description
Technical Field
The invention relates to a positioning and mapping system, in particular to an individualized positioning and mapping system for atrial fibrillation focuses, and belongs to the technical field of atrial fibrillation positioning.
Background
Transcatheter ablation therapy can be used as an effective treatment for the recovery and maintenance of sinus rhythm in patients with atrial fibrillation. In recent studies, transcatheter ablation therapy has been shown to be advantageous in terms of efficacy and improvement in the quality of life of patients compared to antiarrhythmic drug therapy.
Currently, transcatheter rf ablation is mainly non-individualized CPVI (cyclic nerve vessel ablation), linear ablation, progressive ablation, vagal nerve ablation, and individualized ablation such as fractionated potential ablation, EAVM (electrokinetic Voltage Mapping) guided ablation, ROTOR (trochanter) ablation, and the like.
In a recent STAR-AF II study, ablation methods of persistent atrial fibrillation were compared: non-individualized ablation (CPVI alone, PVI + left Atrial apex line and mitral isthmus line), and individualized ablation (PVI + CFAE (Complex sectioned attached electrical ablation). The STAR-AF study concluded that non-individualized CPVI ablation success was highest (< 60%), with all other approaches being low due to the lack of accurate and effective techniques for individualized localization of atrial fibrillation lesions.
In general, a quantifiable mapping technique for individualized localization of atrial fibrillation lesions is currently lacking. In the current individuation method, the standard on which the EAVM ablation depends is non-individuation and is not unified at present, the mechanism of the ROTOR ablation is not clear at present, the unified standard is not available, the success rate is not repeatable, and the dispute is large.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide an individualized positioning and mapping system for atrial fibrillation focuses, which can effectively realize the positioning and mapping of the atrial fibrillation focuses and improve the curative effect of atrial fibrillation radio frequency ablation.
According to the technical scheme provided by the invention, the system for individually positioning and mapping the auricular fibrillation focus comprises
A data collection module: the system is used for collecting data, establishing a database by using the collected data, and carrying out data coding and label establishment on the collected data;
an analysis module: the method comprises the steps of simulating an individualized healthy atrial voltage state to obtain healthy atrial voltage data when atrial fibrillation does not occur, and obtaining a healthy atrial model;
a verification module: the method is used for data quality control, and the established model is corrected to obtain a corrected and matched healthy atrium model;
a calculation module: establishing a multiple regression estimation model by using relative voltage;
identifying and mapping module: outputting a visual focus area graph, and drawing a focus area boundary with a certain tolerance;
a feedback module: and receiving a follow-up database, further judging the accuracy of modeling (R, rc model) during the operation according to the standard of the operation result and giving a correction suggestion.
The data encoded in the data collection module includes individual health data, individual anatomical data, and individual electrophysiological data.
Within the verification module, according to [ Hc]＝[SI]*[SM]* Correcting the healthy atrial model by the Hv matrix algorithm to obtain a corrected and matched healthy heartHouse model (H) C A model); wherein, [ SI ]]For the data quality score matrix, [ SM]The model quality score matrix, hv is the healthy atrial model.
The relative voltage in the calculation module is determined as follows:
T v ＝β D *D*(F v *SRVI*(1-β S *S)+∑β tm *X m )，
wherein, T v Critical value of voltage of lesion region, D is atrial dissection area, beta D Coefficient of D, β S Is the coefficient of the relative model R, S is the rhythm state parameter, beta tm Coefficient of personal health factor, anatomical factor and electrophysiological factor to be examined in a multiple regression estimation model using a voltage value at a sampling point position as a dependent variable, F v For the reference voltage collected, SRVI is the standard relative voltage index.
The calculation module specifically determines the lesion as:
when q × Hc (min)/Av (k) >1, mv (k) =1, the position is estimated to be a lesion;
mv (k) =1 when Q is Hc (min)/Av (k) ≦ 1, and p is Hc (k) -Av (k) > Q, and the position is estimated to be a lesion; otherwise, mv (k) =0, deny the location as a lesion;
therefore, under the M model, each sampling point of the heart tissue is represented by 0,1 values which respectively represent that the heart tissue is estimated to be not a focus and is estimated to be a focus; wherein A is v (k) For the voltage value obtained by sampling the point at the actual position, hc (k) is the voltage value estimated in the non-atrial fibrillation crowd according to the position after matching, mv (k) is according to { Av } k And { Hc } k The obtained k position is a possible characteristic value of the focus; that is, mv (k) is a quantitative value of whether the position of the sampling point has the focus characteristics through a health matching mode; min represents the minimum value; q and p are correction values of a judging program for judging whether a certain sampling point position is a focus or not, and the initial value is set to be 1; the Q initial value is the variation width of the voltage of the non-atrial fibrillation population obtained according to the current research and is set to be 7;
RVI (k) represents the relative voltage index at the k position; satisfies A v(k) ＝β D *D*(F v *RVI (k) +∑β km *X m )；
Rv (k) according to RVI (k) and SRVI, obtaining k position as possible characteristic value of the focus; under the R model, a relative voltage index RVI (k) of a k position is provided, and meanwhile, the voltage state of the position is estimated by comparing the relation between the RVI (k) and the SRVI to judge whether the focus is detected;
when RVI (k) ≦ SRVI, rv (k) =1, estimate the location as a lesion;
rv (k) =0 when RVI (k) > SRVI, deny the location as a lesion;
the Rc (k) obtains the matched label quality of the Hc (k) and the Rv (k) and the Mv (k), and the obtained k position is a possible characteristic value of the focus; and correction according to matching accuracy by Ca (Correct by accurateness) and correction according to consistency by Cc (Correct by consistency);
Ca＝g(N,n s ,n d ,L)
n represents the size of the current non-atrial fibrillation human group database; n is s Indicating the amount of sample included after matching the tag; l represents the label grade; n is d The number of matched labels of d grades is represented;
when only one tag combination is considered for matching; when a plurality of (w types of) label combinations exist in Ca for matching, and the Ca value of each combination is calculated;
1)N≥10 4
2)N<10 4
for Cc, cc =0 when Rv (k) = M (k); rc (k) = Rv (k) = M (k);
when Rv (k) ≠ M (k), cc =1; rc (k) = [ (Ca-0.5) × (M (k) -R (k)) ]
Rc (k) is a characteristic value which is obtained according to the matched label quality of Hc (k) and the obtained Rv (k) and Mv (k), and the obtained k position is possible to focus; the value range of Rc (k) is {0,1};
when Rc (k) =1, the estimated location is a lesion;
when Rc (k) =0, deny the position as a lesion;
therefore, the judgment of the focus characteristics at each position can have three model states;
mv (k), rv (k), rc (k), mv (k), health match; rv (k), self-associated; rc (k), weight synthesis;
correspondingly, when the k position is judged as a focus, av (k) can be obtained; for all V (k) =1 (estimated as the location of the lesion), the maximum Av (k) value under each model can be used as the basis for the threshold voltage value.
The invention has the advantages that: the quantitative mapping technology for individualized positioning of atrial fibrillation focuses is provided, the mechanism is clear, the standard is unified, the quantification and the repeatability can be realized, the success rate of the ablation guided by a module and a system supporting the technology through clinical verification is over 80 percent, and the curative effect of the current atrial fibrillation radio frequency ablation can be obviously improved.
Drawings
FIG. 1 is a block diagram of the system of the present invention.
FIG. 2 is a schematic diagram of the present invention.
Detailed Description
The invention is further illustrated by the following specific figures and examples.
As shown in fig. 1 and 2: in order to effectively realize the positioning and mapping of the atrial fibrillation focus and improve the curative effect of the atrial fibrillation radio frequency ablation, the invention comprises the following steps:
a data collection module: the system is used for collecting data, establishing a database by using the collected data, and carrying out data coding and label establishment on the collected data;
specifically, data from device output and manual input can be accepted through the data collection module, and the input data is coded and labeled according to the internal requirement format of the system. Information codes (caseNO. + info No.) and labels (labelNo.) of individual characteristics of cases are established. The data of the information coding is mainly divided into three large blocks including individual health data, individual anatomical data and individual electrophysiological data.
See in particular the data collection module description below:
and (3) encoding content: 1) case number, 2) data encoding, 3) tag definition
And (3) an encoding mode:
the symbols of the case numbers indicate: caseno. XX, XX is number, specifically number NNNNNYYYYMMDDHHMMSS; NNNNNNN is serial number (different serial numbers are distributed under the same date and time); YYYY is the year of three-dimensional electroanatomical mapping; MM is the month of three-dimensional point anatomical mapping; DD is the date of performing three-dimensional electroanatomical mapping; HH is the hour of performing three-dimensional electroanatomical mapping; MM is the minutes for performing three-dimensional electrodissection mapping; SS is the number of seconds for three-dimensional electroanatomical mapping.
Noxx:
data is divided into three large blocks: 1) individual health data, 2) individual anatomical data, 3) individual electrophysiological data; and (3) encoding format: position coding + value coding; position code = section + entry; wherein, the plate block is: 01-individual health data; 02-individual anatomical data; 03-individual electrophysiological data
2) Items and corresponding codes under each version block: example (c):
value encoding:
and (3) label definition: labelNo. XX; the label meaning is as follows: dividing the input information into classes according to the information value and giving the labels; setting label grade:
when the main value to be considered is the atrial voltage, different grades are given to the magnitude of the effect of the factors on the atrial voltage according to the theory at present. Example (c):
and (3) label coding: tag rank-tag item-tag value; example (c): for dataThe values define a demarcation criterion, for example: the BMI (Body Mass Index) tag encodes 0201; the complete encoding of individual BMI tag is obtained by data encoding, BMI value = data (01-01-13)/(01-01-12) 2 。
The BMI was 30 when 25 was used as a cutoff for obesity. The individual's BMI tag is encoded as labelno.020103.
An analysis module: the method comprises the steps of simulating an individualized healthy atrial voltage state to obtain healthy atrial voltage data when atrial fibrillation does not occur, and obtaining a healthy atrial model;
in particular for individualized simulation of healthy atrial voltage states. And matching the existing database of the crowd without atrial fibrillation in the system (the database is established as above, but the crowd is different and is a non-atrial fibrillation crowd) according to the obtained individual characteristic information number and label. According to the formula Hv = f hv (labelNo.((β i *X i ),W j ) (Hv, which is the matched simulated healthy atrial voltage, f hv Is a simulation model established according to a system database), and obtains the healthy atrial voltage data of the atrial fibrillation patient under simulation when the atrial fibrillation patient does not have atrial fibrillation. I.e. to establish an individual matched healthy atrial model (H-model).
Description of model H:
1) And the principle is as follows: there are several major factors (e.g., age, left atrial size, cardiac function, BMI) that are currently considered to affect changes in atrial voltage, and several factors are currently considered to affect clinically meaningful voltage values of the atria as discussed in the prior art. However, as the database is improved, the judgment of influencing factors, such as genotype, is expected to be improved, and when the research factors are not included in the original research and the database is planned to be built, the parameters are included, and the weight values of the parameters are variable at the beginning and are gradually fixed as the database is imported. The development of low voltage in atrial fibrillation lesions is gradual.
2) And significance: by regulating the factors (X) i ) And rank weight (W) j ) I represents each factor, j represents the grade of each grade weight, and the condition of the atrial voltage of the atrial fibrillation patient without atrial fibrillation is simulated as much as possible,and obtaining a low voltage threshold value of the atrial fibrillation lesion estimated in the investigation index, wherein beta is a coefficient, and the value of i is consistent with the number of the included factors.
3) And establishing:
matching: matching atrial states of non-atrial fibrillation populations primarily by tags, i.e. representing factor X under consideration by tag i Label level indicates a level weight W j . Obtaining the distribution rule of the atrial voltage of people with the same label in the investigation range, namely obtaining the label-matched ideal voltage point K of the people without atrial fibrillation (K = atrial area/(electrode spacing) 2 ) Then, the voltage values of all points are used for simulating the voltage value distribution curve of the incorporated atrial fibrillation individual without the occurrence of atrial fibrillation, namely, the f is established hv And (4) an equation.
The equation: 1. and obtaining a non-atrial fibrillation population consistent with the label of the atrial fibrillation individual by matching the investigated factors. And obtaining a voltage distribution state corresponding to the ideal voltage point number under the label set through parameter estimation. Hv (k) is present; k ∈ [1,K ] andgatez, Z denotes an integer set of global points, k denotes each point. 2. Optimizing the factors to be investigated, carrying out multiple regression analysis on the voltage corresponding to the voltage (k) through linear regression, and correcting the factors. The method comprises the steps of establishing a regression equation by taking Hv (k) as a dependent variable and taking an original heart rhythm state and an included factor as independent variables, judging whether residual errors accord with normal distribution, if the residual errors have the trend that obvious key variables are not included in the equation, tentatively adding new factors according to theoretical updating for matching, and repeating the process to obtain the optimal matching model.
Example (c): the labels 010101 and 010202, 020101 are used as matching factors. Meaning that the conditions for inclusion in individuals with atrial fibrillation fit the characteristics indicated in the label. Meanwhile, a first-level matching (010101 and 010202) crowd is obtained from a non-atrial fibrillation crowd, and a crowd with simultaneous second-level matching (020101) can be subdivided and separated.
Model Hv = f hv (labelNo.((β i *X i ),W j ) Initial form and its specific parameters include, but are not limited to, the following forms:
Hv＝(β 1 *X 1 *LNNORMINV((β 2 *X 2 ),(2.929+β 3 *X 3 +β 4 *X 4 +β 0 ),1.296))*W^(-1)
LNNORMINV is the inverse of a lognormal distribution function, where NORMINV is F -1 (x) It is an inverse function of the normal cumulative distribution function F (x), which is a standard normal cumulative distribution function.
Other parameters in the formula are as follows:
with the improvement of the database, the function form, the parameters and the initial values of the above formula are supplemented continuously.
A verification module: the method is used for data quality control, and the established model is corrected to obtain a corrected and matched healthy atrium model;
specifically, the method is used for data quality control including 1) and data quality control to eliminate bias caused by system errors or error errors, and the content includes (background stability of measured data s, accuracy a of a measuring method, degree of difference upsilon of data measurement times/data measurement times, checking times r of data input, and the like). 2) The quality control of individual model establishment comprises the steps of carrying out program statistics on SI (Score information matrix), namely data quality scoring matrix, and SM (Score model matrix), namely model quality scoring matrix according to s, a, upsilon and r, and carrying out program statistics according to [ Hc ]]＝[SI]*[SM]* The Hv matrix algorithm corrects the model to obtain a corrected matched healthy atrial model (H) C A model).
A calculation module: establishing a multiple regression estimation model by adopting relative voltage;
in particular, a personalized mapping strategy for obtaining atrial fibrillation lesions. Conventional EAVM-guided ablation uses a fixed reference value, i.e., an absolute voltage value, for both low-voltage decision thresholds, but this is not a true individualized mapping and has no uniform criteria. In the embodiment of the invention, the atrial voltage is defined and quantified individually again by adopting a Relative Voltage Index (RVI) which is firstly proposed by a preliminary test study, and a Standard Relative Voltage Index (SRVI) is taken as an individual voltage boundary value standard. The voltage-related indicator does not use absolute voltage values (such as, but not limited to, unipolar voltages, bipolar voltage values, voltage cutoff values, average values, etc. in volts) but rather uses relative values.
Description of relative voltages:
1) Principle and meaning: the voltage of the heart tissue of an individual follows a general rule that there is a range of normal values. But also individual variability. For the population suffering from atrial fibrillation, the atrial stroma is changed, and the accurate identification of the low-voltage focus is particularly important, so that the accurate individual difference has great significance for the population suffering from atrial fibrillation. The critical voltage value is searched to distinguish a focus area from a healthy area, and the focus area and the healthy area can be obtained according to a matched healthy population rule (H model) or can be calculated according to a relation between a healthy tissue and a pathological tissue which are relatively constant based on the critical voltage value (RVI, relative voltage index acquisition).
2) And (3) calculating: the present invention utilizes this rule discovered by preliminary studies. The study was based on analysis of the voltage in each region of the cardiac tissue to obtain a relatively stable voltage region and a representative value F v (reference voltage, where "voltage representative of a relatively stable voltage region" includes, but is not limited to, a percentile value of the voltage of an anatomical region of the atrium for reference, such as the atrial forewall region, the atrial appendage region, etc., which was found in tests to have no statistically significant change between the non-atrial fibrillation "healthy" population and the atrial fibrillation population, and therefore can be used as a reference voltage, and as the database was completed, a more stable reference voltage can be found), and a lesion region voltage threshold (T _ threshold) v : threshold Voltage Threshold value, which should be equal to a certain percentile of the lower value of the clinically normal healthy atrial Voltage, which can distinguish between healthy and diseased), and a more stable relationship between the Threshold value and the representative value of the stable region Voltage, the initial form and its specific parameters include but are not limited to the following forms:
T v ＝β D *D*(F v *SRVI*(1-β S *S)+∑β tm *X m )，
the SRVI found in the pilot study was a constant, defined in the examples of the present invention as the "Standard relative Voltage index", which is the value of the "relative value" of the T per case v Or F v When the absolute voltage value changes, the change can be used as a unified standard for judging the atrial lesion and the healthy tissue. Based on the RVI, i.e. the concept of "relative voltage index", the RVI corresponding to a certain sampling point is calculated and compared with the SRVI, and the result can be used to judge whether the point is a focus. The RVI satisfies: a. The v(k) ＝f(RVI (k) ) In the case of a liquid crystal display device, in particular,
A v(k) ＝β D *D*(F v *RVI (k) (1-β S *S)+∑β km *X m )，
wherein, A v : actual absolute voltage value of a certain sampling point position of the Absolute voltage; RVI: relative Voltage Index value. In the above two formulas, D is the atrial anatomical partition, beta D Is the coefficient of D. S is a rhythm state parameter, beta S Is the coefficient of R. t: critical position, the initial value is the anatomical position corresponding to the 5 th percentile of healthy atrial voltage. k is the position of the sampling point. m: each incorporating a factor. As well as factors affecting atrial voltage. When the database is just built, the factors needing to be considered in theory and the factors actually considered finally can be influenced because of the completeness of the data of the included population.
As the database is refined, the considerations in the Hc model and the Rc model will tend to be consistent and close to reality. When is treated with A v (k) As dependent variables, fv and X m And setting up a coefficient of a regression equation for the independent variable. Beta is a tm And beta km The method is used for estimating coefficients of personal health factors, anatomical factors and electrophysiological factors considered in a multiple regression model taking a voltage value of a sampling point position as a dependent variable, wherein the electrophysiological parameters comprise but are not limited to regional distribution characteristics of various electrophysiological indexes.
As the database is perfected, the m (inclusion factor) and the previous i (i.e. the inclusion factor in the H model) tend to be consistent and close to the real situation(ii) a The T value (critical position) corresponds to a critical value point in all k (sampling positions), and the T corresponding to the point is calculated by the formula v The value can represent the lower critical value of the healthy atrium absolute voltage value matched with the current case, and can be used as the absolute voltage value standard for judging the individuation of the atrium focus and the healthy tissue (the actual measurement A corresponding to a certain sampling position v Value and T v The values are compared, and the result can judge whether the point is a focus or not).
The parameters in these two equations are illustrated in the following table:
with the improvement of the database, the function form, the parameters and the initial values of the formula are continuously corrected and supplemented.
3) Is characterized in that: that is, any algorithm using relative voltage index instead of absolute voltage value or relative voltage index as parameter, and any algorithm using other electrophysiological parameters proposed by the present invention, and its application and system are within the scope of the present invention.
Relative model (R model) and description of its application:
1) And position voltage value description, k represents a voltage sampling position, and A, hc, M and R represent model states.
2) M Model, match Model (healthy) matching Model-lesion determination.
3) R Model, relative Model-Relative voltage.
4) The Rc Model, the Relative Model after the correction of Relative Model.
Wherein, the position voltage and the model relation:
A v (k) The method comprises the following steps And actually taking the voltage value obtained by the point at the position. (A: absolute value of Absolute voltage).
Hc (k): and according to the estimated voltage value of the matched position in the non-atrial fibrillation crowd.
Mv (k): according to A v (k) With Hc (k), the k position obtained is probably a lesionThe characteristic value of (2). That is, mv (k) is a quantitative value indicating whether or not a lesion feature is present at the sampling point position by the healthy matching method.
min represents the minimum value. q and p are correction values of a judging program for judging whether a certain sampling point position is a focus or not, and the initial value is set to be 1. The initial value of Q is 7, which is the variation width of the voltage of the non-atrial fibrillation population obtained from the current study.
When q × Hc (min)/Av (k) >1, mv (k) =1, and the position is estimated to be a lesion.
When Q × Hc (min)/Av (k) ≦ 1, p × Hc (k) -Av (k) > Q, mv (k) =1, and the location is estimated to be a lesion. Otherwise, mv (k) =0, denying that the location is a lesion.
Thus, under the M model, each sampling point of the cardiac tissue has two values 0,1, which indicate whether the cardiac tissue is estimated to be a lesion or not and whether the cardiac tissue is estimated to be a lesion.
RVI (k): indicating the relative voltage index at the k position. Satisfies A v(k) ＝β D *D*(F v *RVI (k) +∑β km *X m )。
Rv (k): the k positions obtained from RVI (k) and SRVI are likely characteristic values of the lesion.
Under the R model, the relative voltage index of the k position, RVI (k), is provided, and the voltage state of the position is estimated by comparing the relationship between RVI (k) and SRVI to judge whether the focus is present.
When RVI (k) ≦ SRVI, rv (k) =1, the location is estimated to be a lesion.
When RVI (k) > SRVI, rv (k) =0, denying the location as a lesion.
Rc (k): and according to the label quality matched with the obtained Hc (k) and the obtained Rv (k) and Mv (k), obtaining a k position as a possible characteristic value of the focus. Ca (Correct by accutaness), correction according to matching accuracy. Cc (Correct by consistency), according to the correction of consistency.
Description of Ca:
the principle is as follows: the obtained Hc model is a Hc (k) position-voltage value set obtained by selecting a label, obtaining a non-atrial fibrillation human group database and analyzing. The labels are selected based on the characteristics of individuals who incorporate atrial fibrillation. The number and grade of tags selected, as well as the size of the early database, will affect the accuracy of Hc, as represented by the degree of modeling of the state of the heart tissue before atrial fibrillation occurs in individuals who incorporate atrial fibrillation. Therefore, by introducing the Ca value, confidence is given to the lesion judgment provided by Mv (k).
And (3) calculating:
Ca＝g(N,n s ,n d ,l)
and N represents the size of the current non-atrial fibrillation human group database. n is s Indicating the amount of sample included after matching the tag. And l represents the grade of the label. n is d The number of matching tags representing d levels.
When only one tag combination is considered for matching. Ca is matched with a plurality of label combinations (w types), and the Ca value of each combination is calculated.
1)N≥10 4
2)N<10 4
In the section "Ca description", the purpose of calculating Ca is described. That is, by introducing Ca value, the lesion judgment provided by Mv (k) is provided with confidence.
Cc illustrates:
the principle is as follows: whether the Mv (k) health contrast-matching judgment is carried out, a focus critical value is estimated through a non-atrial fibrillation population database; or the self-contrast-correlation judgment of Rv (k), and the lesion critical value is estimated by the correlation between healthy tissue and lesion-capable tissue with relatively stable self-heart tissue, and the early influencing factor is selected (label selection and grade assignment and beta) j Parameters), are limited by the size of the database, and the knowledge limitation is the portion to be perfected, and the portion can be perfected by the expansion of the size of the database. Therefore, the inconsistent components between Mv (k) and Rv (k) can be accurately determined at an early stage andand analysis can optimize the M model and the R model. Therefore, the main purpose is to optimize the RVI expression and judge the focus area by adopting a self-contrast mode as far as possible, thereby reducing the calculated amount and ensuring the accuracy.
And (3) calculating:
when Rv (k) = M (k), cc =0.Rc (k) = Rv (k) = M (k)
Cc =1 when Rv (k) ≠ M (k).
Rc (k) = [ (Ca-0.5) × (M (k) -R (k)) ], i.e., rc (k) can be calculated by rounding up.
Description of the correction:
similarly, rc (k) is the label quality matched with the obtained Hc (k) and Rv (k) and Mv (k), and the obtained k position is the possible feature value of the lesion. The value range of Rc (k) is {0,1}, as can be seen from the formula.
When Rc (k) =1, the estimated position is a lesion.
When Rc (k) =0, the position is denied as a lesion.
Therefore, the judgment of the focus characteristics at each position can have three model states.
Mv (k), rv (k), rc (k), mv (k), health match; rv (k), self-associated; rc (k), weight synthesis.
Correspondingly, when the k position is judged as a focus, av (k) can be obtained. For all v (k) =1 (estimated lesion position), where v (k) =1 is the voltage determination in each mode, the maximum Av (k) value in each model can be used as the basis for the threshold voltage value.
Identifying and mapping module: outputting a visual focus area graph, and drawing a focus area boundary with a certain tolerance;
in particular, an output system. The method is used for guiding the radio frequency ablation operation and visually positioning the atrial lesion area of the patient with atrial fibrillation. Meanwhile, an intuitive focus area map is output (namely, a low-voltage area is drawn by using different colors or textures on an existing heart three-dimensional model), and the focus area boundary is drawn by using the electrode spacing as the tolerance, namely, the maximum range represented by the acquired points is used as the tolerance.
A feedback module: and receiving follow-up database information, judging the operation result, further judging the accuracy of modeling (R, rc model) during the operation and giving a correction suggestion.
Rc (Relative Model correct) is the correction of the R Model. The above corrective instruction explains the model.
Follow-up database description:
the meaning is as follows: the individual ablation strategy has better results in preliminary experiments, and can be perfected and accurate in application. Under the current condition that the ablation of the atrial fibrillation patient has low success rate and high recurrence rate, the postoperative heart rhythm state of the atrial fibrillation patient is judged by follow-up visit, the evidence property of an individualized ablation strategy is enhanced on the individual level, possible reasons influencing prognosis and population characteristics are analyzed on the population level, and targeted model optimization is provided when the individualized ablation strategy is customized.
After the atrial fibrillation patient is ablated, a better effect can be achieved in the field. That is, the immediate effect of ablation is successful. However, long-term effects need to be judged by follow-up. If patients with atrial fibrillation relapse after ablation, it is assumed that the population labeled 010203 is likely to relapse, [ i.e., in this example, a population with moderate atrial augmentation ]. The atrial augmentation may be considered to be increased to a parameter. As the case may be. The content is as follows: whether atrial fibrillation occurs again after operation or not; if so, how long it takes to occur; the number of arrhythmic events occurring within 3 months, 6 months, 12 months, 3 years, 5 years, and the type. The state of cardiac function. The application comprises the following steps: judging and optimizing the effectiveness of the individual strategy: and (5) inspecting indexes and feeding back opinions.
Claims (3)
1. An individualized positioning and mapping system for auricular fibrillation focuses is characterized in that: comprises that
A data collection module: the system is used for collecting data, establishing a database by using the collected data, and carrying out data coding and label establishment on the collected data;
an analysis module: the method comprises the steps of simulating an individualized healthy atrial voltage state to obtain healthy atrial voltage data when atrial fibrillation does not occur, and obtaining a healthy atrial model;
a verification module: the method is used for data quality control, and the established model is corrected to obtain a corrected and matched healthy atrium model;
a calculation module: establishing a multiple regression estimation model by using relative voltage;
identifying and mapping the module: outputting a visual focus area graph, and drawing a focus area boundary with a certain tolerance;
a feedback module: receiving a follow-up database, further judging the accuracy of modeling (R, rc model) during surgery according to the standard of the surgery result, and giving a correction suggestion;
the relative voltage in the calculation module is determined as follows:
T v ＝β D *D*(F v *SRVI*(1-β S *S)+∑β tm *X m )，
wherein, T v Critical value of voltage of lesion region, D is atrial dissection area, beta D Coefficient of D, β S Is the coefficient of the relative model R, S is the rhythm state parameter, beta tm Coefficient of personal health factor, anatomical factor and electrophysiological factor to be investigated in a multiple regression estimation model with the voltage value of the sampling point position as a dependent variable, F v For the collected reference voltage, SRVI is a standard relative voltage index;
the judgment of the calculation module on the focus is specifically as follows:
when q × Hc (min)/Av (k) >1, mv (k) =1, the position is estimated to be a lesion;
mv (k) =1 when Q is Hc (min)/Av (k) ≦ 1, and p is Hc (k) -Av (k) > Q, and the position is estimated to be a lesion; otherwise, mv (k) =0, deny the location as a lesion;
therefore, under the M model, each sampling point of the heart tissue is represented by 0,1 values which respectively represent that the heart tissue is estimated to be not a focus and is estimated to be a focus; wherein A is v (k) For the voltage value obtained by sampling the point at the actual position, hc (k) is the voltage value estimated in the non-atrial fibrillation crowd according to the position after matching, mv (k) is according to { Av } k And { Hc } k The obtained k position is a possible characteristic value of the focus; that is, mv (k) is a quantitative value of whether the position of the sampling point has the focus characteristics through a health matching mode; min represents the minimum value; q and p are all in a certain miningSetting the initial value as 1 if the point position is the correction value of the judging program of the focus; the Q initial value is the variation width of the voltage of the non-atrial fibrillation population obtained according to the current research and is set to be 7;
RVI (k) represents the relative voltage index at the k position; satisfies A v(k) ＝β D *D*(F v *RVI (k) +∑β km *X m )；
Rv (k) according to RVI (k) and SRVI, obtaining k position as possible characteristic value of the focus; under the R model, a relative voltage index RVI (k) of a k position is provided, and meanwhile, the voltage state of the position is estimated by comparing the relation between the RVI (k) and the SRVI to judge whether the focus is detected;
when RVI (k) ≦ SRVI, rv (k) =1, estimate the location as a lesion;
rv (k) =0 when RVI (k) > SRVI, deny the location is a lesion;
the Rc (k) obtains the matched label quality of the Hc (k) and the Rv (k) and the Mv (k), and the obtained k position is a possible characteristic value of the focus; and correction according to matching accuracy by Ca (Correct by accurateness) and correction according to consistency by Cc (Correct by consistency);
Ca＝g(N,n s ,n d ,L)
n represents the size of the current non-atrial fibrillation human group database; n is a radical of an alkyl radical s Indicating the amount of sample included after matching the tag; l represents the label grade; n is d The number of matched labels of d level is represented;
when only one tag combination is considered for matching; when a plurality of label combinations (w types) exist in Ca for matching, and the Ca value of each combination is calculated;
1)N≥10 4
2)N<10 4
for Cc, cc =0 when Rv (k) = M (k); rc (k) = Rv (k) = M (k);
cc =1 when Rv (k) ≠ M (k);
rc (k) is a characteristic value which is obtained according to the matched label quality of Hc (k) and the obtained Rv (k) and Mv (k), and the obtained k position is possible to focus; the value range of Rc (k) is known from the formula to be {0,1};
when Rc (k) =1, the estimated location is a lesion;
when Rc (k) =0, deny the position as a lesion;
therefore, the judgment of the focus characteristics at each position can have three model states;
mv (k), rv (k), rc (k), mv (k), health match; rv (k), self-associated; rc (k), weight synthesis;
correspondingly, when the k position is judged as a focus, av (k) can be obtained; for all v (k) =1 (estimated as the location of the lesion), the maximum Av (k) value under each model can be used as the basis for the threshold voltage value.
2. The system of claim 1, wherein the system is configured to perform mapping by: the data encoded in the data collection module includes individual health data, individual anatomical data, and individual electrophysiological data.
3. The system of claim 1, wherein the system is configured to perform mapping by: within the verification module, according to [ Hc]＝[SI]*[SM]* The Hv matrix algorithm corrects the healthy atrial model to obtain a corrected and matched healthy atrial model (H) C A model); wherein, [ SI ]]For the data quality score matrix, [ SM]The model quality score matrix, hv is the healthy atrial model.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911315008.6A CN111297474B (en) | 2019-12-19 | 2019-12-19 | Individualized positioning and mapping system for auricular fibrillation focus |
| PCT/CN2020/094506 WO2021120536A1 (en) | 2019-12-19 | 2020-06-05 | Individualized positioning mapping system for atrial fibrillation lesions |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911315008.6A CN111297474B (en) | 2019-12-19 | 2019-12-19 | Individualized positioning and mapping system for auricular fibrillation focus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111297474A CN111297474A (en) | 2020-06-19 |
| CN111297474B true CN111297474B (en) | 2022-11-08 |
Family
ID=71158028
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911315008.6A Active CN111297474B (en) | 2019-12-19 | 2019-12-19 | Individualized positioning and mapping system for auricular fibrillation focus |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111297474B (en) |
| WO (1) | WO2021120536A1 (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008035070A2 (en) * | 2006-09-20 | 2008-03-27 | Imperial Innovations Limited | Atrial fibrillation analysis |
| WO2018135948A1 (en) * | 2017-01-23 | 2018-07-26 | Academisch Ziekenhuis Leiden H.O.D.N. | Conditionally immortalized cells and methods for their preparation |
| CN110167432A (en) * | 2017-01-13 | 2019-08-23 | 圣犹达医疗用品心脏病学部门有限公司 | System and method for generating ventricular premature beat Electrophysiological mapping figure |
Family Cites Families (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| IL139655A0 (en) * | 2000-11-14 | 2002-02-10 | Hillman Yitzchak | A method and a system for combining automated medical and psychiatric profiling from combined input images of brain scans with observed expert and automated interpreter using a neural network |
| US20080140371A1 (en) * | 2006-11-15 | 2008-06-12 | General Electric Company | System and method for treating a patient |
| JP5918038B2 (en) * | 2011-06-29 | 2016-05-18 | 学校法人東邦大学 | Sustained atrial fibrillation model, method for inducing sustained atrial fibrillation and manufacturing method thereof |
| WO2014113555A2 (en) * | 2013-01-17 | 2014-07-24 | Cardioinsight Technologies, Inc. | Focal point identification and mapping |
| CN103394160A (en) * | 2013-08-01 | 2013-11-20 | 侯月梅 | Method for establishing dog vagal trunk nerve stimulation-induced AF (atrial fibrillation) experimental model |
| WO2015027191A1 (en) * | 2013-08-22 | 2015-02-26 | Cardionxt, Inc. | Methods, systems, and apparatus for identification and characterization of rotors associated with atrial fibrillation |
| WO2015120064A1 (en) * | 2014-02-04 | 2015-08-13 | Cardioinsight Technologies, Inc. | Integrated analysis of electrophysiological data |
| US10925511B2 (en) * | 2014-07-24 | 2021-02-23 | Cardiosolv Ablation Technologies, Inc. | System and method for cardiac ablation |
| US10687898B2 (en) * | 2014-11-14 | 2020-06-23 | The Johns Hopkins University | Systems and methods for atrial fibrillation treatment and risk assessment |
| JP6150825B2 (en) * | 2015-02-05 | 2017-06-21 | ユニオンツール株式会社 | Atrial fibrillation detection system |
| WO2016181320A1 (en) * | 2015-05-12 | 2016-11-17 | Navix International Limited | Fiducial marking for image-electromagnetic field registration |
| DE112016005886T5 (en) * | 2015-12-22 | 2018-08-30 | The Regents Of The University Of California | Computational localization of fibrillation sources |
| US10136828B2 (en) * | 2016-03-31 | 2018-11-27 | Biosense Webster (Israel) Ltd. | Mapping of atrial fibrillation |
| JP6697579B2 (en) * | 2016-05-03 | 2020-05-27 | セント・ジュード・メディカル，カーディオロジー・ディヴィジョン，インコーポレイテッド | A system for damage prediction based in part on tissue characterization. |
| CN110151308A (en) * | 2018-02-11 | 2019-08-23 | 北京航空航天大学 | For treating the ablating device and system of auricular fibrillation |
| EP3795176A4 (en) * | 2018-05-15 | 2022-01-12 | Astellas Pharma Inc. | Pharmaceutical composition for suppressing atrial fibrillation having as active ingredient anti-human ngf antibody or antigen-binding fragment thereof |
| CN109770893B (en) * | 2019-03-08 | 2022-11-18 | 东南大学 | Method and device for rapidly positioning atrial fibrillation position in Holter analysis system |
| CN110151164B (en) * | 2019-03-14 | 2024-03-19 | 无锡市人民医院 | Zhou Changbiao assay for atrial fibrillation |
| CN110464896B (en) * | 2019-08-22 | 2022-01-28 | 成都市第三人民医院 | Magnetomotive ventricular assist system |
| CN110522512A (en) * | 2019-09-04 | 2019-12-03 | 上海科技大学 | Radio-frequency ablation procedure auxiliary system based on model |
-
2019
- 2019-12-19 CN CN201911315008.6A patent/CN111297474B/en active Active
-
2020
- 2020-06-05 WO PCT/CN2020/094506 patent/WO2021120536A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008035070A2 (en) * | 2006-09-20 | 2008-03-27 | Imperial Innovations Limited | Atrial fibrillation analysis |
| CN110167432A (en) * | 2017-01-13 | 2019-08-23 | 圣犹达医疗用品心脏病学部门有限公司 | System and method for generating ventricular premature beat Electrophysiological mapping figure |
| WO2018135948A1 (en) * | 2017-01-23 | 2018-07-26 | Academisch Ziekenhuis Leiden H.O.D.N. | Conditionally immortalized cells and methods for their preparation |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021120536A1 (en) | 2021-06-24 |
| CN111297474A (en) | 2020-06-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Brod et al. | Conceptualization of quality of life in dementia. | |
| Grosheva et al. | Prognostic value of electroneurography and electromyography in facial palsy | |
| CN108198615B (en) | Online cognitive evaluation system | |
| Boyd et al. | Evaluating trauma care: the TRISS method | |
| Bakheit et al. | The rate and extent of improvement with therapy from the different types of aphasia in the first year after stroke | |
| Gunn et al. | Why some women look young for their age | |
| Jiang et al. | Objective acoustic analysis of pathological voices from patients with vocal nodules and polyps | |
| US8965497B2 (en) | Electrophysiological analysis system and method | |
| Hupp et al. | Electrophysiological multimodal assessments improve outcome prediction in traumatic cervical spinal cord injury | |
| US20020107455A1 (en) | Method and apparatus for estimating degree of neuronal impairment in brain cortex | |
| Sandhu et al. | Brief pain inventory–facial minimum clinically important difference | |
| Fenton et al. | Prediction of postoperative facial nerve function in acoustic neuroma surgery | |
| Fanton et al. | Critical study of observation of the sternal end of the right 4th rib | |
| Ho et al. | Inter-rater reliability of seven neurolaryngologists in laryngeal EMG signal interpretation | |
| JP2017530849A (en) | Method and system for predicting continuous cardiac output (CCO) of a patient based on physiological data | |
| CN111297474B (en) | Individualized positioning and mapping system for auricular fibrillation focus | |
| Liu et al. | The application of the machine learning method in electromyographic data | |
| CN106645516A (en) | Method for detecting neurotransmitter in hair | |
| CN101596125A (en) | A kind of health and fitness information display system, method and interface thereof that possesses demonstration directly perceived | |
| CN114981900A (en) | Method for optimizing intracranial implantation scheme of a set of electroencephalogram electrodes | |
| CN108877931B (en) | Shoulder joint rehabilitation evaluation method, device and system | |
| Guzman-Vargas et al. | Influence of the loss of time-constants repertoire in pathologic heartbeat dynamics | |
| Koenigsberg et al. | Development of a scale for measuring techniques in the psychotherapy of borderline patients | |
| Bonomi et al. | “Divide and conquer”: assessing energy expenditure following physical activity type classification | |
| CN106096274A (en) | A kind of sectional linear fitting health evaluating method based on normal distribution |
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 |