CN115331817A - Early pregnancy stage premature type preeclampsia risk screening device - Google Patents
Early pregnancy stage premature type preeclampsia risk screening device Download PDFInfo
- Publication number
- CN115331817A CN115331817A CN202210863444.2A CN202210863444A CN115331817A CN 115331817 A CN115331817 A CN 115331817A CN 202210863444 A CN202210863444 A CN 202210863444A CN 115331817 A CN115331817 A CN 115331817A
- Authority
- CN
- China
- Prior art keywords
- marker
- pregnant woman
- value
- risk
- mom
- 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
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/43—Detecting, measuring or recording for evaluating the reproductive systems
- A61B5/4306—Detecting, measuring or recording for evaluating the reproductive systems for evaluating the female reproductive systems, e.g. gynaecological evaluations
- A61B5/4343—Pregnancy and labour monitoring, e.g. for labour onset detection
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
- A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Physics & Mathematics (AREA)
- Veterinary Medicine (AREA)
- Heart & Thoracic Surgery (AREA)
- Biophysics (AREA)
- Animal Behavior & Ethology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- Pregnancy & Childbirth (AREA)
- Gynecology & Obstetrics (AREA)
- Reproductive Health (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention provides a device for screening risks of premature type preeclampsia in the early pregnancy stage, and belongs to the technical field of computer application. The device comprises: an acquisition unit; the calculating unit is used for calculating the gestational age, a regression formula of the median of the marker relative to the gestational age, the MOM value of the marker and a corrected value thereof; the modeling unit is used for establishing a prior risk calculation model, establishing a multi-dimensional Gaussian distribution model based on a premature type preeclampsia pregnant woman group and a normal pregnant woman group, and establishing a marker likelihood ratio calculation model; a correction unit and a decision unit. The invention can quantitatively express the difference of the combined measured values of various markers between the premature type preeclampsia pregnant woman population and the normal pregnant woman population through the marker combined multidimensional probability density distribution function, thereby improving the accuracy of the risk assessment result.
Description
Technical Field
The invention relates to the technical field of disease risk prediction, in particular to a risk screening device for early-stage preterm type preeclampsia in the early pregnancy stage.
Background
Preeclampsia (preeclampsia) is a specific multisystemic disease during pregnancy that usually affects 2-5% of pregnant women and is one of the major causes of morbidity and mortality in both pregnant and perinatal women. Preeclampsia is manifested by new hypertension of pregnant women, and at least one new disease is accompanied after 20 weeks or 20 weeks of gestation: urinary protein, dysfunction of other organs of the mother body and placenta dysfunction seriously threaten the life of the mother and fetus. Preeclampsia is classified as early onset (patient's delivery week <34 weeks + 0), premature onset (patient's delivery week <37 weeks + 0), late onset (patient's delivery week > 34 weeks + 0), full onset (patient's delivery week > 37 weeks + 0), or full onset (patient's delivery week > 37 weeks + 0), as defined by the international Federation of obstetrics and gynecology. In consideration of incidence and harmfulness, the clinical screening of the pre-term eclampsia is more concerned, so that the accurate screening of the pre-term eclampsia is particularly important.
At present, the method adopted by Chinese preeclampsia screening mainly comprises clinical experience and risk assessment software developed abroad. The defects of incapability of quantification, fuzzy screening measurement indexes, unsatisfactory screening effect caused by inconsistent levels of main doctors and the like exist in manual clinical experience screening; the risk assessment software developed abroad is mainly applied to screening in the aspects of early onset and late onset of eclampsia, the basic data of modeling is derived from foreign normal people and sick people, although certain progress and success are achieved in the aspect of screening in the foreign preeclampsia, the difference exists in MOM value (median multiple) probability distribution, weight probability distribution and the like between home and abroad due to factors such as race, weight of pregnant women, regions and the like, the value of the serological index of the domestic normal pregnant women is obviously different from that of the foreign normal pregnant women, so that the deviation of a detection result is caused, if the foreign screening method is directly adopted, the median multiple value of the obtained serological index is applied to prenatal screening of the pregnant women, so that false positive is increased, a large number of positive pregnant women are missed, and the screening efficiency is reduced.
Disclosure of Invention
The invention aims at solving the problem of how to effectively improve the screening efficiency of the risk screening of the premature type preeclampsia in the early pregnancy stage (11 weeks +0 days to 13 weeks +6 days of pregnancy) of Chinese pregnant woman groups.
In order to solve the above problems, the present invention provides a device for screening risk of early stage preterm birth type preeclampsia in pregnancy, comprising:
the device comprises an acquisition unit, a data acquisition unit and a data acquisition unit, wherein the acquisition unit is used for acquiring basic data of a pregnant woman 11 to 13 weeks pregnant, and the basic data comprises basic personal information, medical history information, measurement values of markers, ultrasonic top-hip diameter detection results, ultrasonic double top-diameter detection results, last menstrual time and pregnancy outcome, wherein the markers comprise MAP, UTPI and PLGF;
the calculation unit is used for acquiring the gestational age based on the pregnancy date according to the ultrasonic top-hip diameter detection result or the ultrasonic double-top diameter detection result or the last menstrual time;
the calculation unit is further used for respectively obtaining a first regression formula of the median of each marker relative to the gestational age according to the basic data of the pregnant woman with normal pregnancy outcome;
the calculation unit is further configured to obtain an MOM value of each marker according to each first regression formula and a measured value of each marker;
the calculation unit is further used for obtaining basic characteristics and medical history factors of the pregnant woman which have significant influence on the MOM value of each marker through significance analysis based on the basic personal information and the medical history information, obtaining a second regression formula which takes the basic characteristics and the medical history factors as variables through a regression model, and obtaining MOM correction values of each marker according to the MOM value of each marker and the second regression formula;
the analysis unit is used for carrying out logarithmic transformation on the MOM correction value of each marker to obtain a logarithmic value of the MOM value of each marker, and carrying out statistical analysis on the logarithmic value of the MOM value of each marker to obtain the self distribution characteristic of each marker and the correlation among the markers;
the modeling unit is used for analyzing an influence weight coefficient of a risk factor on the pregnancy outcome according to a logistic regression model, optimizing the influence weight coefficient by adopting a maximum likelihood estimation method and establishing a prior risk calculation model;
the modeling unit is further used for respectively constructing a multi-dimensional Gaussian distribution model based on the premature type preeclampsia pregnant woman population and the normal pregnant woman population according to the measured values of the markers, the self distribution characteristics of the markers and the correlation among the markers, and establishing a marker likelihood ratio calculation model;
the correction unit is used for correcting the prior risk calculation model according to the marker likelihood ratio calculation model to obtain the posterior risk probability;
and the decision unit is used for acquiring an ROC curve according to the posterior risk probability and the pregnancy outcome of the pregnant woman and acquiring the optimal risk cutoff value of the marker under different combination conditions according to the principle of the maximum Youden index.
In one embodiment, the logistic regression model is:
wherein: p is the probability that the pregnancy outcome is pre-term eclampsia;
x 1 、x 2 、x 3 、x n the risk factors include the age, height, weight of the pregnant woman, and whether the pregnant woman has a history of hypertension, diabetes, autoimmune disease, and whether the pregnant woman is a menstruating woman, whether the previous fetus has a history of preeclampsia, and a time interval from the previous fetus;
w 1 、w 2 、w 3 、w n said weight coefficient of influence on said pregnancy outcome for each said risk factor, wherein w 0 A logarithmic value representing the ratio of the probability of the pregnant woman to be sick and not to be sick;
n is the number of the risk factors;
transforming the logistic regression model to:
wherein: p (x) is the windA risk factor results in a probability that the pregnancy outcome of the pregnant woman is pre-term type eclampsia, wherein w = [) 0 ，w 1 ，w 2 ，w 3 …w n ]x＝[1，x 1 ，x 2 ，x 3 …x n ]。
In an embodiment, the modeling unit is further configured to determine a likelihood function, the likelihood function being:
wherein: m is the number of the pregnant women;
x j a risk factor vector for a jth pregnant woman;
y j the pregnancy outcome for the jth pregnant woman.
In one embodiment, the analysis unit is configured to perform statistical analysis on logarithmic values of the MOM values of the markers, and the statistical analysis includes:
obtaining distribution forms of logarithmic values of the MOM values of the markers in the premature type preeclampsia pregnant woman population and the normal pregnant woman population respectively, fitting the distribution forms of the markers by adopting a normal distribution function, obtaining optimal parameters of a measured value normal distribution model of the markers through least square optimization, and obtaining self distribution characteristics of the markers.
In one embodiment, the analysis unit is configured to perform statistical analysis on logarithmic values of the MOM values of the markers, and further includes:
statistically analyzing the correlation between the two markers, and respectively calculating the correlation coefficient of each marker in the premature type preeclampsia pregnant woman population and the normal pregnant woman population.
In one embodiment, the marker likelihood ratio calculation model is:
wherein: LR is the marker likelihood ratio;
G N a multi-dimensional Gaussian distribution function of the normal pregnant woman population marker;
G P a multidimensional Gaussian distribution function of the markers for the pre-term type eclampsia pregnant woman population;
μ is the median of the MOM value pairs for the markers;
sigma is a covariance matrix of the marker;
z j is the logarithmic value of the MOM value of the marker for the jth pregnant woman.
In one embodiment, the modifying unit is configured to modify the prior risk calculation model according to the marker likelihood ratio calculation model to obtain a posterior risk probability, and includes:
calculating the posterior risk probability by adopting a Bayesian model, wherein the Bayesian model is as follows:
wherein: p (y = 1|z) is the log of the MOM value of the marker equal to the probability that z causes the pregnant woman to suffer from pre-term eclampsia;
p (z) is the probability that the logarithm of the MOM value of the marker equals z;
p (y = 1) is the prior risk probability of the pregnant woman suffering from pre-term type eclampsia;
p (z | y = 0) is the probability that the logarithmic value of the MOM value of the marker in the normal pregnant population is equal to z;
p (z | y = 1) is the probability that the logarithmic value of the MOM value of the marker equals z in the population of pregnant women with pre-term eclampsia;
In one embodiment, the basic characteristics and medical history factors include: the age, height, weight of the pregnant woman, and whether the pregnant woman has a history of smoking, hypertension and diabetes.
In one embodiment, the calculating unit is configured to obtain a first regression formula of median relative gestational age of each marker according to basic data of pregnant women with normal pregnancy outcome, and the first regression formula includes:
and obtaining a first regression formula of the median of each marker relative to the gestational age by using MATLAB regression analysis software and adopting any one model including a polynomial model, an exponential model and a logarithmic model according to the basic data of the pregnant woman with normal pregnancy outcome.
Compared with the prior art, the early pregnancy stage premature type preeclampsia risk screening device has the beneficial effects that:
the invention can quantitatively express the difference of the combined measured values of various markers between the premature type eclampsia pre-pregnant woman population and the normal pregnant woman population through the marker combined multidimensional probability density distribution function, thereby further improving the accuracy of the risk assessment result;
the pregnant woman data modeled by the method is derived from domestic pregnant woman detection data, so that the influence of individual difference on the model evaluation result is reduced;
the method can screen the pregnant women which are possibly subjected to preeclampsia before 37 weeks, the evaluation result takes the premature preeclampsia as a screening target, and the screening mode is more favorable for clinical screening evaluation;
according to the risk calculation model disclosed by the invention, the corresponding risk cutoff value is optimized according to different marker detection combinations, so that the application range of the model is wider.
In one embodiment, the early stage preterm type preeclampsia risk screening device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor when executing the computer program implementing the steps of:
acquiring basic data of a pregnant woman 11 to 13 weeks pregnant, wherein the basic data comprises basic personal information, medical history information, measurement values of markers, ultrasonic top hip diameter detection results, ultrasonic double top diameter detection results, last menstrual time and pregnancy outcome, and the markers comprise MAP, UTPI and PLGF;
determining gestational age based on pregnancy according to the ultrasonic peak-hip diameter detection result or the ultrasonic double peak-diameter detection result or the last menstrual time;
respectively obtaining a first regression formula of the median of each marker relative to the gestational age according to the basic data of the pregnant woman with normal pregnancy outcome;
respectively obtaining the MOM value of each marker according to each first regression formula and the measured value of each marker;
obtaining basic characteristics and medical history factors of the pregnant woman which have significant influence on the MOM value of each marker through significance analysis based on the basic personal information and the medical history information, obtaining a second regression formula which takes the basic characteristics and the medical history factors as variables through a regression model, and obtaining the MOM correction value of each marker according to the MOM value of each marker and the second regression formula;
carrying out logarithmic transformation on the MOM corrected value of each marker to obtain a logarithmic value of the MOM value of each marker, and carrying out statistical analysis on the logarithmic value of the MOM value of each marker to obtain the self distribution characteristic of each marker and the correlation among the markers;
analyzing influence weight coefficients of the risk factors on the pregnancy outcome by adopting a logistic regression model, optimizing the influence weight coefficients by adopting a maximum likelihood estimation method, and establishing a prior risk calculation model;
respectively constructing a multi-dimensional Gaussian distribution model based on a premature type preeclampsia pregnant woman group and a normal pregnant woman group according to the measured values of the markers, the self distribution characteristics of the markers and the correlation among the markers, and establishing a marker likelihood ratio calculation model;
correcting the prior risk calculation model according to the marker likelihood ratio calculation model to obtain a posterior risk probability;
and acquiring an ROC curve according to the posterior risk probability and the pregnancy outcome of the pregnant woman, and acquiring the optimal risk cutoff value of the marker under different combination conditions according to the principle of maximum Jordan index.
Drawings
FIG. 1 is a diagram of an environment in which an early pregnancy stage pre-term preterm eclampsia risk screening device in an embodiment of the present invention may be used;
FIG. 2 is a schematic diagram of an early pregnancy stage pre-term onset preeclampsia risk screening device in an embodiment of the present invention;
fig. 3 is a schematic diagram of the internal layout of the early pregnancy stage preterm type preeclampsia risk screening device in the embodiment of the invention.
Detailed Description
The method for screening the preeclampsia in China mainly comprises the following steps: screening by human clinical experience and screening by adopting risk assessment software developed abroad.
In the early pregnancy screening, the artificial clinical experience screening is to comprehensively judge the possibility of the pregnant women to generate preeclampsia through maternal characteristics and medical history and combining with the measured levels of maternal Mean Arterial Pressure (MAP), uterine artery pulsatility index (UTPI) and placental growth factor (PLGF). The disadvantages of this approach are: the screening standards are not uniform; quantitative judgment cannot be carried out; the final judgment result completely depends on the understanding of doctors on each risk factor and the experience of previous cases; the screening result only can indicate whether the pregnant woman is likely to have preeclampsia, and the indexes cannot be further refined.
The risk assessment software developed abroad is mainly applied to screening in the aspects of early onset and late onset of eclampsia, a screening model of the risk assessment software calculates basic risk through maternal characteristics and medical history, then calculates the probability of the preeclampsia of the pregnant woman under the condition of the current measurement result according to the measured values of MAP, UTPI and PLGF, and synthesizes the basic risk to obtain the final risk of the pregnant woman. The disadvantages of this approach are: the modeling data source is foreign pregnant woman population, and the applicability of the risk calculation model is not strong due to group difference; and the detection rate and the false positive rate are not ideal because the risk cutoff value is optimized according to the incidence rate of the preeclampsia in China; risk calculations for early onset and late onset preeclampsia are only lacking, as is the risk of premature preeclampsia.
In order to solve the problem of risk screening of the premature type preeclampsia of Chinese pregnant woman groups, the invention constructs a risk evaluation model for screening the premature type preeclampsia in the early pregnancy stage according to the Chinese pregnant woman group data, and optimizes a risk cutoff value to improve the screening efficiency of the premature type preeclampsia.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
FIG. 1 is a diagram of an environment in which an early pregnancy stage pre-term preterm eclampsia risk screening device in an embodiment of the present invention may be used. Referring to fig. 1, the risk screening device is applied to a risk screening system for early stage of pregnancy pre-term type preeclampsia. The risk screening system includes a terminal 110 and a server 120. The terminal 110 and the server 120 are connected through a network. The terminal 110 may specifically be a desktop terminal or a mobile terminal, and the mobile terminal may specifically be at least one of a mobile phone, a tablet computer, a notebook computer, and the like. The server 120 may be implemented as a stand-alone server or a server cluster comprising a plurality of servers.
In one embodiment, as shown in fig. 2, a pre-term pre-eclampsia risk screening device for early pregnancy is provided. The embodiment is mainly exemplified by the application of the apparatus to the terminal 110 (or the server 120) in fig. 1. Referring to fig. 2, the device for screening risk of pre-term eclampsia during early pregnancy according to the embodiment of the present invention includes:
an acquisition unit 210 for acquiring the following information of a pregnant woman from 11 to 13 weeks gestation: basic personal information, medical history information, measured values of markers, ultrasonic top-hip diameter (CRL) detection results, ultrasonic double top diameter (BPD) detection results, last menstrual time and pregnancy outcome of the pregnant woman, and of course, ultrasonic detection time and blood sampling time are included. Wherein the basic personal information comprises the age, name, gestational age, weight and the like of the pregnant woman, the medical history information comprises diabetes history, hypertension history and the like, the markers comprise MAP, UTPI and PLGF, and the pregnancy outcome of the pregnant woman comprises normal outcome and the outcome of premature type preeclampsia; in addition, the week of pregnancy is 11 weeks to 13 weeks, which means 11 weeks plus 0 days of pregnancy to 13 weeks plus 6 days of pregnancy.
And the calculating unit 220 is used for determining the gestational age based on the ultrasonic peak-hip diameter detection result or the ultrasonic double peak-diameter detection result or the last menstrual time. Specifically, the detection result of CRL or BPD of ultrasonic examination is preferentially adopted in the gestational age calculation, the gestational age based on days is estimated through a gestational age calculation formula, and when the ultrasonic detection result is not acquired, the gestational age based on days is estimated by adopting the last menstrual time;
the calculation unit is also used for determining median regression formula of the markers MAP, UTPI and PLGF about the change of gestational age: the method comprises the steps of taking pregnant woman measurement data with normal pregnancy outcome, utilizing a matlab regression analysis tool, and obtaining a first regression formula of the median of each marker relative to the gestational age by adopting methods such as a polynomial model, an exponential model, a logarithmic model and the like due to different marker regression models. It should be noted that the "first" in the first regression formula is only for distinguishing from the subsequent regression formula, and it is understood that the first regression formula includes a regression formula of the median of the marker MAP with respect to the gestational age, a regression formula of the median of the marker UTPI with respect to the gestational age, and a regression formula of the median of the marker PLGF with respect to the gestational age.
The calculating unit is also used for calculating the median multiple of the markers, namely MOM value: calculating the median multiple (MOM) value of the serological index of each examined pregnant woman according to a first regression formula, wherein the calculation formula is as follows:
the system comprises an acquisition unit, a calculation unit and a calculation unit, wherein the 'detected pregnant woman marker measured value' is the marker measured value acquired by the acquisition unit, and the 'corresponding fetal age marker median regression formula calculated value' is the regression formula of the marker median relative to the fetal age acquired by the calculation unit;
the calculation unit is further configured to modify the MOM value: and respectively carrying out data distribution on the obtained MOM value of the pregnant woman marker according to basic characteristics and medical history factors of the pregnant woman, wherein the basic characteristics and the medical history factors comprise the age, the height, the weight, the smoking history, the hypertension history and the diabetes history of the pregnant woman, the data distribution such as significance analysis is carried out, and the factors of the age, the height, the weight, the smoking history, the hypertension history and the diabetes history of the pregnant woman, which have greater influence on the distribution of the MOM value, are obtained through statistical analysis, so that the MOM value needs to be further corrected. First, the second regression formulas respectively taking the age, height, weight, smoking history, hypertension history and diabetes history of the pregnant woman as variables are obtained through the regression model, it should be noted that the "second" in the second regression formulas is only to distinguish from the aforementioned regression formulas, and it is understood that the second regression formulas herein include, for example, a regression formula taking the age of the pregnant woman as a variable, that is, a regression formula of the age of the pregnant woman to the MOM value correction coefficient described below, for example, a regression formula taking the height of the pregnant woman as a variable, and the like. After obtaining the second regression equation, the MOM correction value is then calculated according to the following equation:
MOM correction value = MOM × k 1 (age) × k 2 (height) × k 3 (body weight) × k 4 (smoking). Times.k 5 (hypertension) × k 6 (diabetes mellitus);
in the formula: MOM is calculated in the step (4);
k 1 (age) is a regression formula of the age of the pregnant woman to the MOM value correction coefficient, namely a regression formula which is obtained by a regression model and takes the age of the pregnant woman as a variable;
k 2 the height is a regression formula of the height of the pregnant woman to the MOM value correction coefficient, namely the regression formula which takes the height of the pregnant woman as a variable is obtained through a regression model;
k 3 (weight) is a regression formula of the weight of the pregnant woman to the MOM value correction coefficient, namely the regression formula which takes the weight of the pregnant woman as a variable is obtained through a regression model;
k 4 (smoking) is a regression formula of whether the smoking history is applied to the MOM value correction coefficient, namely a regression formula which is obtained by a regression model and takes whether the pregnant woman has the smoking history as a variable;
k 5 (hypertension) is a regression formula of whether the hypertension history is subjected to MOM value correction coefficient, namely the regression formula which is obtained by a regression model and takes the hypertension history of the pregnant woman as a variable;
k 6 (diabetes) is a regression formula of whether the diabetes history is to the MOM value correction coefficient, namely the regression formula which takes the diabetes history of the pregnant women as a variable is obtained through a regression model;
a modeling unit 240, configured to establish an a priori risk calculation model: and marking the measurement data of the pregnant woman according to the pregnancy outcome, marking the outcome of the pregnant woman with the premature eclampsia as 1, and otherwise, marking as 0. The a priori risk factors include: age, height, weight, history of hypertension, history of diabetes, whether autoimmune disease, whether it is a pregnant woman, whether it has a history of preeclampsia in the previous fetus, and the time interval of the previous fetus. The influence weight coefficients of the risk factors on the outcome of pregnancy, i.e. the coefficients of the risk factors that may cause the pregnant woman to be ill (pre-term type preeclampsia), are analyzed using a logistic regression model, and the influence weight coefficients are optimized using a maximum likelihood estimation method, e.g. the probability that the ill may be caused by the weight of the pregnant woman, which is the regression coefficient.
Wherein, the formula of the logistic regression model is as follows:
in the formula: p is the probability that the pregnancy outcome is pre-term eclampsia;
x 1 、x 2 、x 3 、x i 、x n the risk factors are respectively age, height, weight, history of hypertension or not, history of diabetes or not, self immunological diseases or not, whether the pregnant woman is a puerpera or not, whether the previous fetus has a preeclampsia history or not, time interval of the previous fetus and the like;
w 1 、w 2 、w 3 、w i 、w n weight coefficient of influence of each risk factor on pregnancy outcome, wherein w 0 A logarithmic value representing the ratio of the probability of illness and the probability of non-illness of the pregnant woman in general;
n is the number of risk factors;
the logistic regression model was converted to:
in the formula: w = [ w = 0 ，w 1 ，w 2 ，w 3 …w n ]x＝[1，x 1 ，x 2 ，x 3 …x n ]；
And optimizing the weight coefficient by adopting a maximum likelihood estimation method, and constructing a likelihood function as follows:
in the formula: m is the number of pregnant women to be tested;
x j a risk factor vector for a jth pregnant woman;
y j is the pregnancy outcome of the jth pregnant woman, wherein y j =1 denotes pregnancy outcome as pre-term onset preeclampsia, y j =0 for pregnancy outcome non-preterm pre-eclampsia;
and the analysis unit 230 is used for taking logarithm of the MOM value according to the calculated MOM value of the marker measured value, and statistically analyzing the distribution form of each logarithmic MOM value of the marker in the premature type preeclampsia pregnant woman population and the normal pregnant woman population respectively, wherein the abscissa in the distribution graph is the logarithm value of the MOM value of the marker, and the ordinate in the distribution graph is the frequency statistical value. Fitting the distribution form of each marker by adopting a normal distribution function, and optimizing by a least square method to obtain the optimal parameters of a normal distribution model of the measured value of each marker; and statistically analyzing the correlation between the two markers, and respectively calculating the correlation coefficients of the markers in the premature type preeclampsia pregnant woman population and the normal pregnant woman population.
The modeling unit 240 is further configured to build a marker likelihood ratio calculation model: according to the number of the markers, the distribution characteristics of the markers and the correlation among the markers, a multi-dimensional Gaussian distribution model, namely a multi-dimensional probability density distribution function, based on the premature type preeclampsia pregnant woman population and the normal pregnant woman population is respectively constructed. The likelihood ratio calculation formula is as follows:
in the formula: LR is marker likelihood ratio;
G N a multi-dimensional Gaussian distribution function of the marker for the normal pregnant woman population;
G P a multi-dimensional Gaussian distribution function of a marker of a pregnant woman population in the early-term eclampsia;
mu is the median of the value pair of the marker MOM value;
sigma is a covariance matrix of the marker;
z j is a logarithmic value of the jth pregnant woman marker MOM value;
the dimensionality of the multi-dimensional Gaussian distribution function depends on the number of the marker measurements of the pregnant woman, and the likelihood ratio calculation model can automatically match and generate a corresponding dimensionality Gaussian distribution function according to the number of the marker measurements;
and a correcting unit 250, configured to further correct the prior risk according to the likelihood ratio to obtain a final risk probability. And (3) adopting a Bayesian model to complete the posterior risk probability calculation:
in the formula: p (y = 1|z) is the log of the marker MOM value equal to the probability that z causes a pregnant woman to suffer from pre-term eclampsia;
p (z) is the probability that the logarithm of the value of the MOM of the marker equals z;
p (y = 1) is the prior risk probability of a pregnant woman suffering from pre-term type eclampsia;
p (z | y = 0) is the probability that the logarithmic value of the marker MOM in the normal pregnant population is equal to z;
p (z | y = 1) is the probability that the logarithmic value of the MOM value of the marker equals z in the population of pregnant women with pre-term eclampsia;
And the decision unit 260 is used for acquiring an ROC curve through SPSS software according to the posterior risk probability value of the pregnant woman and the pregnancy outcome of the pregnant woman, wherein the ROC curve is a test subject working characteristic curve and is a curve drawn by taking the true positive rate (sensitivity) as the ordinate and the false positive rate (1-specificity) as the abscissa. The principle of risk cutoff optimization is the maximum jordan Index (Youden Index), also called the correct Index, which is a commonly used method when the harmfulness of false negative (missed diagnosis rate) and false positive (misdiagnosis rate) is assumed to have equal significance, and reflects the true total ability of patients and non-patients. The johnson index is the sum of the sensitivity and the specificity minus 1, the greater the johnson index is, the greater the authenticity is, the predicted value corresponding to the johnson index is found, then the numerical value corresponding to the data entry is the optimal cutoff value, and the cutoff value is the boundary value for judging the test to be positive and negative, namely the boundary value for disease and normal in the embodiment. Therefore, according to the risk cutoff value optimization principle, the corresponding optimal cutoff values under different conditions of the number of the markers can be obtained through optimization respectively.
According to the embodiment of the invention, MAP, UTPI and PLGF are taken as markers, the acquisition unit is used for acquiring the measured values of the three markers, the calculation unit is used for calculating the median of the three markers according to the gestational age and the relationship between the median of the markers and the gestational age, so that the median multiples of the three markers are further obtained, then the age, the height, the weight, the smoking, the hypertension, the diabetes history and other factors of the pregnant woman are obtained through significance analysis, and the median multiples of each marker are corrected by considering the factors. And acquiring the possibility that each risk factor may cause illness by using a modeling unit according to a logistic regression model and combining a maximum likelihood estimation method. And after taking the logarithm of the corrected median multiple, analyzing the distribution condition of the logarithm in the population of sick and normal pregnant women, fitting the distribution form through normal distribution, optimizing through a least square method to respectively obtain the optimal parameters of normal distribution models of three markers, then analyzing the correlation between every two markers, and the correlation coefficient of each marker in the population of normal and sick pregnant women, constructing a multi-dimensional probability density distribution function combining MAP, UTPI and PLGF based on the population of sick and normal pregnant women, further correcting the prior risk by using the constructed probability density function to obtain the final probability of the disease caused by the median logarithm, and then obtaining an ROC curve through software to perform subsequent analysis.
The device provided by the embodiment of the invention can screen the pregnant women which are possibly eclamptic before 37 weeks, in addition, because the probability distribution forms of MAP, UTPI and PLGF measured values in different gestational weeks are different, different probability density models are established according to different gestational weeks measured by markers (different probability distribution models are established according to different gestational weeks), namely, different Gaussian distribution parameters of the markers are obtained according to the gestational weeks before the joint probability density function is established, and different joint distribution functions are established through different parameters. The difference of the combined measurement values of the multiple markers in the premature type preeclampsia pregnant woman population and the normal pregnant woman population can be quantitatively expressed through the marker combined multidimensional probability density distribution function, so that the accuracy of the risk assessment result is further improved.
In some embodiments, the risk screening device for early stage preterm type preeclampsia includes a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor when executing the computer program implementing the steps of:
acquiring basic data of a pregnant woman 11 to 13 weeks pregnant, wherein the basic data comprises basic personal information, medical history information, measurement values of markers, ultrasonic top hip diameter detection results, ultrasonic double top diameter detection results, last menstrual time and pregnancy outcome, and the markers comprise MAP, UTPI and PLGF;
determining gestational age based on pregnancy according to the ultrasonic peak-hip diameter detection result or the ultrasonic double peak-diameter detection result or the last menstrual time;
respectively obtaining a first regression formula of the median of each marker relative to the gestational age according to the basic data of the pregnant woman with normal pregnancy outcome;
respectively obtaining the MOM value of each marker according to each first regression formula and the measured value of each marker;
obtaining basic characteristics and medical history factors of the pregnant woman which have significant influence on the MOM value of each marker through significance analysis based on the basic personal information and the medical history information, obtaining second regression formulas respectively taking the basic characteristics and the medical history factors as variables through regression models, and obtaining MOM correction values of each marker according to the MOM value and the second regression formulas of each marker;
carrying out logarithmic transformation on the MOM correction value of each marker to obtain a logarithmic value of the MOM value of each marker, and carrying out statistical analysis on the logarithmic value of the MOM value of each marker to obtain the self distribution characteristic of each marker and the correlation among the markers;
analyzing an influence weight coefficient of a risk factor on the pregnancy outcome by adopting a logistic regression model, optimizing the influence weight coefficient by adopting a maximum likelihood estimation method, and establishing a prior risk calculation model;
respectively constructing a multi-dimensional Gaussian distribution model based on a premature type preeclampsia pregnant woman population and a normal pregnant woman population according to the measured value of each marker, the self distribution characteristics of each marker and the correlation among the markers, and establishing a marker likelihood ratio calculation model;
correcting the prior risk calculation model according to the marker likelihood ratio calculation model to obtain a posterior risk probability;
and acquiring an ROC curve according to the posterior risk probability and the pregnancy outcome of the pregnant woman, and acquiring the optimal risk cutoff value of the marker under different combination conditions according to the principle of maximum Youden index.
Specifically, the method comprises the following steps:
(1) Acquiring data: data were derived from Ningbo Olympic for 5124 sample data of the Chinese pregnant woman preeclampsia prediction research project (ChiPERM) by Biotechnology, inc. Acquiring basic personal information and medical history information of the pregnant woman, and MAP, UTPI, a serum marker PLGF detection result, an ultrasonic double apical diameter (CRL) detection result, an ultrasonic crown-to-hip distance (BPD) detection result, ultrasonic detection time and blood sampling time, last menstrual time and pregnancy outcome of the pregnant woman from 11 weeks +0 days to 13 weeks +6 days of pregnancy.
Wherein: MAP is calculated from systolic (sBP) and diastolic (dBP), MAP = dBP + (sBP-dBP)/3. The examinee takes the sitting position, the arm height is flush with the heart, after resting for 5min, the blood pressure of the two arms is measured simultaneously, and two sets of records are made at an interval of 1 min;
the UTPI conditional patients should perform abdominal ultrasonic examination on uterine artery pulse index UTPI from 11 weeks plus 0 day to 13 weeks plus 6 days (fetus top hip diameter is 42-84 mm) of gestation;
PLGF is a glycodimeric glycoprotein secreted by trophoblasts, is a part of a vascular endothelial growth factor family, and is used for detecting the content of PLGF in the serum of pregnant women by using a chemiluminescent detection instrument and reagents produced by Ningbo Olympic Biotech Ltd.
(2) Determination of Gestational Age (GA) based on gestational day: the optimization of fetal age calculation adopts the detection result of CRL or BPD of ultrasonic examination in the step (1), and then estimates the day-based fetal age through a fetal age calculation formula (such as the first row and the second row in the table 1), wherein CRL is the value of ultrasonic double apical diameter in the step (1), and BPD is the value of ultrasonic crown-hip distance in the step (1).
When no ultrasound examination was acquired, last Menstrual Period (LMP) was used to estimate day-based gestational age (third row in table 1);
TABLE 1
| CRL tireAge calculationFormula (II) | GA＝23.620+5.113*sqrt(2.745*crl) |
| BPD tireAge calculationFormula (II) | GA＝32.599+2.912*bpd-0.0222*bpd^2+0.000199*bpd^3 |
| LMP tireAge calculationFormula (II) | GA = check time-last menstrual time |
(3) Median regression formula for determining changes in gestational age for markers MAP, UTPI, PLGF: taking pregnant woman measurement data with normal pregnancy outcome, obtaining a first regression formula of Median (media) of each marker relative to gestational age by utilizing matlab regression analysis tool and adopting methods such as polynomial model, exponential model, logarithmic model and the like, as shown in table 2:
TABLE 2
| MAP median calculation formula | Median＝10^(1.6064+0.006153*GA-0.00003292*GA^2) |
| UTPI median calculation formula | Median＝10^(0.6827-0.005017*GA) |
| PLGF median calculation formula | Median＝10^(0.6978+0.07128*GA) |
(4) Calculating median fold of marker, namely MOM value: calculating the median multiple of the serological index of each detected pregnant woman by using the first regression formula obtained in the step (3), namely the MOM value, wherein the calculation formula is as follows:
(5) Correcting the MOM value: firstly, respectively carrying out significance analysis on the MOM value of the pregnant woman marker obtained in the step (4) according to the basic characteristics and medical history factors of the pregnant woman, obtaining the factors of the age, the height, the weight, the smoking history, the hypertension history and the diabetes history of the pregnant woman through statistical analysis, wherein the factors have great influence on the distribution of the MOM value and need to further correct the MOM value, and the correction formula of the MOM value is as follows:
MOM correction value = MOM × k 1 (age) × k 2 (height) × k 3 (body weight) × k 4 (smoking). Times.k 5 (hypertension) × k 6 (diabetes mellitus);
the regression formulas of the basic characteristics and the medical history factors of the pregnant woman on the MOM value correction coefficients of different markers are shown in tables 3, 4 and 5, wherein the table 3 is the regression formula of the factors such as the Age, the Height, the Weight, the smoking history or not, the hypertension history and the diabetes history on the MOM value correction coefficient of MAP, the table 4 is the regression formula of the factors such as the Age, the Height, the Weight, the smoking history or not, the hypertension history and the diabetes history on the MOM value correction coefficient of UTPI, the table 5 is the regression formula of the factors such as the Age, the Height, the Weight, the smoking history or not, the hypertension history and the diabetes history on the MOM value correction coefficient of PLGF, taking the table 3 as an example, "Age" represents the Age of the pregnant woman, "Height" represents the Height of the pregnant woman, and "Weight" represents the Weight of the pregnant woman, which are the basic information of the pregnant woman collected in the step (1), and when the pregnant woman has the smoking history, k represents the basic information of the MOM value correction coefficient of different markers 4 The value is 1.0602, and k is when the pregnant woman has no history of smoking 4 Value of 1,k 5 、k 6 Value of and k 4 Similarly.
TABLE 3
TABLE 4
TABLE 5
(6) Establishing a prior risk calculation model: and marking the measurement data of the pregnant woman according to the pregnancy outcome, marking the outcome of the pregnant woman with the premature eclampsia as 1, and otherwise, marking as 0. The a priori risk factors include: age, height, weight, history of hypertension, history of diabetes, whether autoimmune disease, whether it is a pregnant woman, whether it has a history of preeclampsia in the previous fetus, and the time interval of the previous fetus. The logistic regression model is adopted, and the model formula is as follows:
x 1 -x 9 the age, height (unit: cm), weight (unit: kg), history of hypertension, history of diabetes, autoimmune disease, whether it is a pregnant woman (based on 2 years), whether it has a history of preeclampsia in the previous fetus, and the time interval (unit: year) of the previous fetus are respectively shown.
The weight coefficients were optimized using the maximum likelihood estimation method, and the regression coefficients after optimization are shown in table 6:
TABLE 6
(7) Establishing a marker likelihood ratio calculation model: and (5) calculating the MOM value of the measured value of the marker according to the step (5), taking logarithm of the MOM value, and respectively counting the histogram distribution of the logarithmic MOM value of each marker in each pregnancy week in the pre-term eclampsia pregnant woman population and the normal pregnant woman population, wherein the abscissa in the histogram is the logarithm value of the MOM value of the marker, and the ordinate is the frequency counting value. And fitting the distribution form of each marker in each gestational week by adopting a normal distribution function, and optimizing by a least square method to obtain the optimal parameters of a normal distribution model of each marker measured value. And statistically analyzing the correlation between different pregnancy measurement values of the two markers, and respectively calculating the correlation coefficient of the markers in a premature type preeclampsia pregnant woman population and a normal pregnant woman population. According to the number of the markers, the distribution characteristics of the markers and the correlation among the markers, a multi-dimensional Gaussian distribution model based on the premature type preeclampsia pregnant woman population and the normal pregnant woman population is respectively constructed, and a marker likelihood ratio calculation model is established.
Taking the statistical parameters of the marker of pregnancy week 12 as an example, the steps of constructing a MAP, UTPI and PLGF triple likelihood ratio calculation model are as follows:
the construction of a multidimensional Gaussian function of normal pregnant women is as follows:
in the formula: z is a radical of j Logarithmic vectors of MOM values of MAP, UTPI and PLGF measured values of the jth pregnant woman are obtained by real-time measurement and conversion;
mu N is Gaussian distribution median vector of MAP, UTPI and PLGF of normal pregnant women in 12 weeks of pregnancy,
μ N ＝[μ N_MAP ，μ N_UTPI ，μ N_PLGF ]；
∑ N the covariance matrix of MAP, UTPI and PLGF of normal pregnant women in 12 weeks is composed of Gaussian distribution standard deviation (see table 7) and correlation coefficient (see table 8);
d is the number of markers.
The construction of a multi-dimensional Gaussian function for a pregnant woman population with pre-term eclampsia is shown as follows:
in the formula: z is a radical of j Logarithmic vectors of MOM values of MAP, UTPI and PLGF measured values of the jth pregnant woman are obtained by real-time measurement and conversion;
μ P is Gauss distribution median vector of MAP, UTPI and PLGF of sick pregnant women in 12 weeks of pregnancy P ＝[μ P_MAP ，μ P_UTPI ，μ P_PLGH ]；
∑ P The covariance matrix of MAP, UTPI and PLGF of the sick pregnant women in 12 weeks of pregnancy consists of Gaussian distribution standard deviation and correlation coefficients;
d is the number of markers.
According to the multi-dimensional Gaussian distribution model of the premature type preeclampsia pregnant woman population and the normal pregnant woman population, the likelihood ratio is calculated, and the formula is as follows:
wherein: the 12-week-gestation gaussian distribution parameters for each marker are shown in table 7:
TABLE 7
From Table 7, it can be seen that N ＝[0，0，0]；μ P ＝[0.056，0.196，-0.286]；σ N_MAP Pregnant 12-week normal pregnant women as MAP markersPopulation standard deviation of Gaussian distribution, σ N_MAP =0.062; similarly, σ N_UTPI ＝0.133；σ N_PLGF ＝0.134；σ P_MAP ＝0.228；σ P_UTPI ＝0.082；σ P_PLGF ＝0.111。
The correlation coefficient r between markers at 12 weeks of gestation is shown in table 8:
TABLE 8
| Marker substance | Normal pregnant woman group | Population of pregnant women with premature type eclampsia |
| MAP-UTPI | －0.0611 | 0.0303 |
| MAP-PLGF | 0.0084 | 0.0111 |
| UTPI-PLGF | -0.0903 | -0.1227 |
R in the above covariance matrix N_MAP_UTPI Representing the correlation coefficient, r, of the two markers MAP and UTPI in the normal pregnant woman population N_MAP_UTPI ＝-0.0611；r P_MAP_UTPI Representing the correlation coefficient, r, of the two markers MAP and UTPI in a population of pregnant women with pre-term eclampsia P_MAP_UTPI =0.0303; similarly, r N_MAP_PLGF ＝0.0084；r P_MAP_PLGF ＝0.0111；r N_UTPI_PLGF ＝-0.0903；r P_UTPI_PLGF ＝-0.1227。
(8) Calculating posterior risk: the posterior risk calculation is further correction of prior risk according to likelihood comparison to obtain final risk probability. And (3) adopting a Bayesian model to complete the posterior risk probability calculation:
in the formula: p (y = 1) is the prior risk probability of a pregnant woman suffering from pre-term eclampsia;
p (z | y = 0) is the probability that the log value of the marker MOM value = z in the normal pregnant population;
p (z | y = 1) is the probability that the log of the value of the marker MOM = z in the population of pregnant women with pre-term eclampsia;
the ratio of P (z | y = 1) to P (z | y = 1) is the likelihood ratio in step (7);
(9) Optimizing a risk cutoff value: and (5) calculating the risk probability value of the detected pregnant woman and the pregnancy outcome of the pregnant woman by the step (8), and acquiring an ROC curve through SPSS software. According to the maximum principle of the johnson index, the optimal cut-off values of different risk assessment models are respectively shown in table 9:
TABLE 9
| Type of model | AUC | Cutoff value |
| Prior risk model + MAP | 0.842 | 178 |
| Prior risk model + (MAP, PLGF) | 0.865 | 82 |
| Prior risk model + (MAP, UTPI, PLGF) | 0.883 | 112 |
In table 9, "prior risk model + MAP" represents the final model obtained by correcting prior risk according to MAP marker likelihood ratio, and its AUC is 0.842, the optimal cutoff value is 178, and AUC is the area enclosed by the coordinate axis under the ROC curve. Similarly, "a priori risk model + (MAP, PLGF)" represents a final model obtained by modifying the priori risk according to the MAP and PLGF joint likelihood ratio, and "a priori risk model + (MAP, UTPI, PLGF)" represents a final model obtained by modifying the priori risk according to the MAP, UTPI, PLGF joint likelihood ratio.
According to the table 9, the AUC of the model obtained by correcting the prior risk according to the triple likelihood ratio of MAP, UTPI and PLGF is closer to 1, and the detection authenticity is higher.
FIG. 3 illustrates an internal block diagram of an early pregnancy stage pre-term preterm eclampsia risk screening device in one embodiment. The risk screening device may specifically be a terminal 110 (or a server 120) for use in fig. 1. As shown in fig. 3, the risk screening device includes a processor, a memory, a network interface, an input device, and a display screen connected by a system bus. The memory comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium of the risk screening device stores an operating system and may also store a computer program that, when executed by the processor, causes the processor to implement an early pregnancy stage pre-term type preeclampsia risk screening method. The internal memory may also have stored therein a computer program that, when executed by the processor, causes the processor to perform the step of screening for risk of pre-term preterm eclampsia during the early stages of pregnancy. The display screen of the device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the risk screening device can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the device, an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the configuration shown in fig. 3 is a block diagram of only a portion of the configuration associated with the inventive arrangements and is not intended to limit the devices to which the inventive arrangements may be applied, and that a particular device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, the processor when executing the computer program further performs the above-described early stage pregnancy pre-term type preeclampsia risk screening step.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring basic data of a pregnant woman 11 to 13 weeks pregnant, wherein the basic data comprises basic personal information, medical history information, measurement values of markers, ultrasonic top hip diameter detection results, ultrasonic double top diameter detection results, last menstrual time and pregnancy outcome, and the markers comprise MAP, UTPI and PLGF;
determining gestational age based on pregnancy according to the ultrasonic peak-hip diameter detection result or the ultrasonic double peak-diameter detection result or the last menstrual time;
respectively obtaining a first regression formula of the median of each marker relative to the gestational age according to the basic data of the pregnant woman with normal pregnancy outcome;
respectively obtaining the MOM value of each marker according to each first regression formula and the measured value of each marker;
obtaining basic characteristics and medical history factors of the pregnant woman which have significant influence on the MOM value of each marker through significance analysis based on the basic personal information and the medical history information, obtaining a second regression formula which takes the basic characteristics and the medical history factors as variables through a regression model, and obtaining the MOM correction value of each marker according to the MOM value of each marker and the second regression formula;
carrying out logarithmic transformation on the MOM correction value of each marker to obtain a logarithmic value of the MOM value of each marker, and carrying out statistical analysis on the logarithmic value of the MOM value of each marker to obtain the self distribution characteristic of each marker and the correlation among the markers;
analyzing influence weight coefficients of the risk factors on the pregnancy outcome by adopting a logistic regression model, optimizing the influence weight coefficients by adopting a maximum likelihood estimation method, and establishing a prior risk calculation model;
respectively constructing a multi-dimensional Gaussian distribution model based on a premature type preeclampsia pregnant woman population and a normal pregnant woman population according to the measured value of each marker, the self distribution characteristics of each marker and the correlation among the markers, and establishing a marker likelihood ratio calculation model;
correcting the prior risk calculation model according to the marker likelihood ratio calculation model to obtain a posterior risk probability;
and acquiring an ROC curve according to the posterior risk probability and the pregnancy outcome of the pregnant woman, and acquiring the optimal risk cutoff value of the marker under different combination conditions according to the principle of maximum Jordan index.
In one embodiment, the computer program when executed by the processor further performs the above-described steps of early stage pregnancy preterm type preeclampsia risk screening.
It will be understood by those skilled in the art that all or part of the processes for implementing the above embodiments may be implemented by hardware related to instructions of a computer program, where the computer program may be stored in a non-volatile computer readable storage medium, and when executed, the computer program may include the processes of the above embodiments. Any reference to memory, storage, databases, or other media used in embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct Rambus Dynamic RAM (DRDRAM), and Rambus Dynamic RAM (RDRAM), among others.
Although the present disclosure has been described above, the scope of the present disclosure is not limited thereto. Various changes and modifications may be effected therein by one of ordinary skill in the pertinent art without departing from the spirit and scope of the present disclosure, and these changes and modifications are intended to be within the scope of the present disclosure.
Claims (10)
1. A device for screening for risk of early stage preterm birth type preeclampsia in an early stage of pregnancy comprising:
the device comprises an acquisition unit, a data acquisition unit and a data acquisition unit, wherein the acquisition unit is used for acquiring basic data of a pregnant woman 11 to 13 weeks pregnant, and the basic data comprises basic personal information, medical history information, measurement values of markers, ultrasonic top-hip diameter detection results, ultrasonic double top-diameter detection results, last menstrual time and pregnancy outcome, wherein the markers comprise MAP, UTPI and PLGF;
the calculation unit is used for acquiring gestational age based on a pregnancy day according to the ultrasonic peak-to-hip diameter detection result or the ultrasonic double peak-to-diameter detection result or the last menstrual time;
the calculation unit is further used for respectively obtaining a first regression formula of the median of each marker relative to the gestational age according to the basic data of the pregnant woman with normal pregnancy outcome;
the calculation unit is further configured to obtain an MOM value of each marker according to each first regression formula and a measured value of each marker;
the calculation unit is further used for obtaining basic characteristics and medical history factors of the pregnant woman which have significant influence on the MOM value of each marker through significance analysis based on the basic personal information and the medical history information, obtaining a second regression formula which takes the basic characteristics and the medical history factors as variables through a regression model, and obtaining MOM correction values of each marker according to the MOM value of each marker and the second regression formula;
the analysis unit is used for carrying out logarithmic transformation on the MOM correction value of each marker to obtain a logarithmic value of the MOM value of each marker, and carrying out statistical analysis on the logarithmic value of the MOM value of each marker to obtain the self distribution characteristic of each marker and the correlation among the markers;
the modeling unit is used for analyzing an influence weight coefficient of a risk factor on the pregnancy outcome according to a logistic regression model, optimizing the influence weight coefficient by adopting a maximum likelihood estimation method and establishing a prior risk calculation model;
the modeling unit is further used for respectively constructing a multi-dimensional Gaussian distribution model based on the premature type preeclampsia pregnant woman population and the normal pregnant woman population according to the measured values of the markers, the self distribution characteristics of the markers and the correlation among the markers, and establishing a marker likelihood ratio calculation model;
the correction unit is used for correcting the prior risk calculation model according to the marker likelihood ratio calculation model to obtain the posterior risk probability;
and the decision unit is used for acquiring an ROC curve according to the posterior risk probability and the pregnancy outcome of the pregnant woman and acquiring the optimal risk cutoff value of the marker under different combination conditions according to the principle of the maximum Youden index.
2. The early pregnancy stage preterm birth type preeclampsia risk screening device of claim 1, wherein the logistic regression model is:
wherein: p is the probability that the pregnancy outcome is pre-term eclampsia;
x 1 、x 2 、x 3 、x n the risk factors include the age, height, weight of the pregnant woman, and whether the pregnant woman has a history of hypertension, diabetes, autoimmune disease, and whether the pregnant woman is a menstruating woman, whether the previous fetus has a history of preeclampsia, and a time interval from the previous fetus;
w 1 、w 2 、w 3 、w n said weight coefficient of influence on said pregnancy outcome for each said risk factor, wherein w 0 A logarithmic value representing the ratio of the probability of the pregnant woman to be affected to the probability of not being affected;
n is the number of the risk factors;
transforming the logistic regression model to:
wherein: p (x) is the probability that the pregnancy outcome of the pregnant woman due to the risk factor is pre-term type eclampsia, wherein w = [ w = 0 ，w 1 ，w 2 ，w 3 …w n ]，x＝[1，x 1 ，x 2 ，x 3 …x n ]。
3. The early stage preterm birth type preeclampsia risk screening device of claim 2, wherein the modeling unit is further configured to determine a likelihood function, the likelihood function being:
wherein: m is the number of the pregnant women;
x j a risk factor vector for a jth pregnant woman;
y j the pregnancy outcome for the jth pregnant woman.
4. The early pregnancy stage preterm birth type preeclampsia risk screening device of claim 1, wherein the analysis unit is configured to perform statistical analysis on the logarithmic values of the MOM values of each of the markers, and comprises:
obtaining distribution forms of logarithmic values of the MOM values of the markers in the premature type preeclampsia pregnant woman population and the normal pregnant woman population respectively, fitting the distribution forms of the markers by adopting a normal distribution function, obtaining optimal parameters of a measured value normal distribution model of the markers through least square optimization, and obtaining self distribution characteristics of the markers.
5. The early pregnancy stage preterm birth type preeclampsia risk screening device of claim 4, wherein the analysis unit is configured to perform statistical analysis on the logarithmic values of the MOM values of each of the markers, further comprising:
statistically analyzing the correlation between the two markers, and respectively calculating the correlation coefficient of each marker in the premature type preeclampsia pregnant woman population and the normal pregnant woman population.
6. The early stage preterm birth type preeclampsia screening device of claim 1, wherein the marker likelihood ratio calculation model is:
wherein: LR is the marker likelihood ratio;
G N a multi-dimensional Gaussian distribution function of the normal pregnant woman population marker;
G P is said pre-term type of eclampsiaThe early pregnant woman population marker multi-dimensional Gaussian distribution function;
μ is the median of the MOM value pairs for the markers;
sigma is a covariance matrix of the marker;
z j is the logarithmic value of the MOM value of the marker for the jth pregnant woman.
7. The early pregnancy stage premature type preeclampsia risk screening device of claim 6, wherein the revising unit is configured to revise the prior risk calculation model according to the marker likelihood ratio calculation model to obtain a posterior risk probability, and comprises:
calculating the posterior risk probability by adopting a Bayesian model, wherein the Bayesian model is as follows:
wherein: p (y = 1|z) is the log of the MOM value of the marker equal to the probability that z causes the pregnant woman to suffer from pre-term type eclampsia;
p (z) is the probability that the logarithm of the MOM value of the marker equals z;
p (y = 1) is the prior risk probability of the pregnant woman suffering from pre-term type eclampsia;
p (z | y = 0) is the probability that the logarithmic value of the MOM value of the marker in the normal pregnant population is equal to z;
p (z | y = 1) is the probability that the logarithmic value of the MOM value of the marker equals z in the population of pregnant women with pre-term eclampsia;
8. The early pregnancy stage preterm birth type preeclampsia screening device of claim 1, wherein the primary characteristics and history factors comprise: the age, height, weight of the pregnant woman, and whether the pregnant woman has a history of smoking, hypertension and diabetes.
9. The device for screening risk of pre-term birth type preeclampsia as set forth in claim 1, wherein the computing unit is configured to obtain a first regression formula of median versus gestational age of each marker respectively according to basic data of pregnant women with normal pregnancy outcome, and comprises:
and obtaining a first regression formula of the median of each marker relative to the gestational age by using MATLAB regression analysis software and adopting any one model including a polynomial model, an exponential model and a logarithmic model according to the basic data of the pregnant woman with normal pregnancy outcome.
10. The early stage preterm birth type preeclampsia risk screening device of claim 1, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the computer program when executing the computer program performing the steps of:
acquiring basic data of a pregnant woman 11 to 13 weeks pregnant, wherein the basic data comprises basic personal information, medical history information, measurement values of markers, ultrasonic top hip diameter detection results, ultrasonic double top diameter detection results, last menstrual time and pregnancy outcome, and the markers comprise MAP, UTPI and PLGF;
determining gestational age based on pregnancy according to the ultrasonic peak-hip diameter detection result or the ultrasonic double peak-diameter detection result or the last menstrual time;
respectively obtaining a first regression formula of the median of each marker relative to the gestational age according to the basic data of the pregnant woman with normal pregnancy outcome;
respectively obtaining the MOM value of each marker according to each first regression formula and the measured value of each marker;
obtaining basic characteristics and medical history factors of the pregnant woman which have significant influence on the MOM value of each marker through significance analysis based on the basic personal information and the medical history information, obtaining a second regression formula which takes the basic characteristics and the medical history factors as variables through a regression model, and obtaining the MOM correction value of each marker according to the MOM value of each marker and the second regression formula;
carrying out logarithmic transformation on the MOM correction value of each marker to obtain a logarithmic value of the MOM value of each marker, and carrying out statistical analysis on the logarithmic value of the MOM value of each marker to obtain the self distribution characteristic of each marker and the correlation among the markers;
analyzing influence weight coefficients of the risk factors on the pregnancy outcome by adopting a logistic regression model, optimizing the influence weight coefficients by adopting a maximum likelihood estimation method, and establishing a prior risk calculation model;
respectively constructing a multi-dimensional Gaussian distribution model based on a premature type preeclampsia pregnant woman group and a normal pregnant woman group according to the measured values of the markers, the self distribution characteristics of the markers and the correlation among the markers, and establishing a marker likelihood ratio calculation model;
correcting the prior risk calculation model according to the marker likelihood ratio calculation model to obtain a posterior risk probability;
and acquiring an ROC curve according to the posterior risk probability and the pregnancy outcome of the pregnant woman, and acquiring the optimal risk cutoff value of the marker under different combination conditions according to the principle of maximum Youden index.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210863444.2A CN115331817B (en) | 2022-07-21 | 2022-07-21 | Early pregnancy stage premature delivery type preeclampsia risk screening device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210863444.2A CN115331817B (en) | 2022-07-21 | 2022-07-21 | Early pregnancy stage premature delivery type preeclampsia risk screening device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115331817A true CN115331817A (en) | 2022-11-11 |
| CN115331817B CN115331817B (en) | 2023-03-17 |
Family
ID=83920020
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210863444.2A Active CN115331817B (en) | 2022-07-21 | 2022-07-21 | Early pregnancy stage premature delivery type preeclampsia risk screening device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115331817B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117747110A (en) * | 2023-12-26 | 2024-03-22 | 南京鼓楼医院 | Preeclampsia risk prediction method and preeclampsia risk prediction system based on maternal factors and early pregnancy biomarkers |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150087553A1 (en) * | 2012-04-13 | 2015-03-26 | Bioscreening and Diagnostics, LLC | Early trimester screening for early- and late-onset preeclampsia |
| CN105069277A (en) * | 2015-07-09 | 2015-11-18 | 宁波大学 | Down's syndrome risk assessment method for prenatal screening in pregnant metaphase |
| US20210033619A1 (en) * | 2018-02-09 | 2021-02-04 | Metabolomic Diagnostics Limited | Methods of predicting pre term birth from preeclampsia using metabolic and protein biomarkers |
| CN112820406A (en) * | 2020-12-30 | 2021-05-18 | 复旦大学附属妇产科医院 | Method for predicting early pregnancy onset risk of epilepsy |
-
2022
- 2022-07-21 CN CN202210863444.2A patent/CN115331817B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150087553A1 (en) * | 2012-04-13 | 2015-03-26 | Bioscreening and Diagnostics, LLC | Early trimester screening for early- and late-onset preeclampsia |
| CN105069277A (en) * | 2015-07-09 | 2015-11-18 | 宁波大学 | Down's syndrome risk assessment method for prenatal screening in pregnant metaphase |
| US20210033619A1 (en) * | 2018-02-09 | 2021-02-04 | Metabolomic Diagnostics Limited | Methods of predicting pre term birth from preeclampsia using metabolic and protein biomarkers |
| CN112820406A (en) * | 2020-12-30 | 2021-05-18 | 复旦大学附属妇产科医院 | Method for predicting early pregnancy onset risk of epilepsy |
Non-Patent Citations (2)
| Title |
|---|
| 季建生等: "UtA-PI、MAP和PLGF对不同孕期子痫前期诊断意义", 《中国妇幼健康研究》 * |
| 谢真等: "早孕期血清妊娠相关血浆蛋白A和胎盘生长因子水平与子痫前期的相关性研究", 《中国妇幼保健》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117747110A (en) * | 2023-12-26 | 2024-03-22 | 南京鼓楼医院 | Preeclampsia risk prediction method and preeclampsia risk prediction system based on maternal factors and early pregnancy biomarkers |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115331817B (en) | 2023-03-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Molina et al. | Validation of Delphi procedure consensus criteria for defining fetal growth restriction | |
| CN114464322B (en) | Female pelvic floor dysfunction disease risk early warning model and construction method and system thereof | |
| Khalil et al. | Longitudinal changes in maternal hemodynamics in a population at risk for pre‐eclampsia | |
| Mackie et al. | Early prognostic factors of outcomes in monochorionic twin pregnancy: systematic review and meta-analysis | |
| US7315787B2 (en) | Multi-marker screening protocol for fetal abnormalities | |
| CN113077900B (en) | Diabetes early risk assessment method, device, computer equipment and medium | |
| Cavoretto et al. | Reference ranges of uterine artery pulsatility index from first to third trimester based on serial Doppler measurements: longitudinal cohort study | |
| CN115331817B (en) | Early pregnancy stage premature delivery type preeclampsia risk screening device | |
| CN112331340B (en) | Intelligent prediction method and system for pregnancy probability of pregnant couple | |
| Guy et al. | Maternal cardiac function at 35–37 weeks' gestation: relationship with birth weight | |
| US10572808B2 (en) | System and method for predicting fetal and maternal health risks | |
| Phan et al. | Arterial stiffness for the early prediction of pre‐eclampsia compared with blood pressure, uterine artery Doppler and angiogenic biomarkers: a prospective cohort study | |
| CN113724873A (en) | Preeclampsia risk prediction method based on MLP multi-platform calibration | |
| CN112992353A (en) | Method and device for accurately predicting due date, computer equipment and storage medium | |
| CN116631622A (en) | AI-based pregnant woman characteristic-related fetal weight prediction method and system | |
| Mansukhani et al. | Ophthalmic artery Doppler at 36 weeks' gestation in prediction of pre‐eclampsia: validation and update of previous model | |
| Zheng et al. | Construction of prediction model for fetal growth restriction during first trimester in an Asian population | |
| Ho et al. | Use of the myocardial performance index in decreased fetal movement assessment: a case-control study | |
| CN117954083B (en) | Preeclampsia prediction method based on machine learning algorithm | |
| Vergani et al. | Risk assessment for Down syndrome with genetic sonogram in women at risk | |
| Hanoon et al. | The use of abdominal aorta length in the fetuses’ after 34 weeks of gestation for calculating gestation age among primigravida healthy women | |
| RU2796073C1 (en) | Method for predicting vaginal delivery in pregnant women after one caesarean section performed by a transverse incision in the lower segment of the uterus | |
| Xu et al. | A machine learning model to predict spontaneous vaginal delivery failure for term nulliparous women: An observational study | |
| Syaharutsa et al. | A Scoring System for Preeclampsia Screening Based on Maternal and Biophysical Factors: Result from a 3-Month Cohort Study in Jakarta, Indonesia | |
| He et al. | The potential of repeated mean arterial pressure measurements for predicting early‐and late‐onset pre‐eclampsia in twin pregnancies: Prediction model study |
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 |