AORTIC STENOSIS CLASSIFICATION

Abstract

A system (102) includes a digital information repository(s) (104) configured to store an aortic valve area measurement, a mean transaortic pressure gradient measurement, and a peak aortic jet velocity measurement for a subject of interest. The system further includes a computing apparatus (106). The computing apparatus comprises a memory (110) configured to store instructions (120) for an aortic stenosis classifier (122). The computing apparatus further comprises a processor (108) configured to execute the stored instructions for the aortic stenosis classifier to classify a severity of an aortic stenosis of the subject of interest based at least on the aortic valve area measurement, the mean transaortic pressure gradient measurement, and the peak aortic jet velocity measurement for the subject of interest. The computing apparatus further comprises a display configured to display the severity.

Claims

1. A system, comprising: a digital information repository(s) configured to store an aortic valve area measurement, a mean transaortic pressure gradient measurement, and a peak aortic jet velocity measurement for a subject of interest; a computing apparatus, comprising: a memory configured to store instructions for an aortic stenosis classifier; and a processor configured to execute the stored instructions for the aortic stenosis classifier to classify a severity of an aortic stenosis of the subject of interest based at least on the aortic valve area measurement, the mean transaortic pressure gradient measurement, and the peak aortic jet velocity measurement for the subject of interest; and a display configured to display the severity.

2. The system of claim 1, wherein the digital information repository(s) is further configured to store information about subjects with aortic stenoses, including at least aortic valve area measurements, mean transaortic pressure gradient measurements, and peak aortic jet velocity measurements thereof, wherein the aortic stenosis classifier is trained with at least the aortic valve area measurements, the mean transaortic pressure gradient measurements, and the peak aortic jet velocity measurements of the subjects with aortic stenoses to provide a trained classifier

3. The system of claim 2, wherein the instructions further includes an individual-level visualizer, and the processor is further configured to execute the instructions for the individual-level visualizer to construct a two-dimensional graph of aortic stenosis versus time based on historical aortic stenosis diagnoses and cause the display monitor to display the two-dimensional graph.

4. The system of claim 2, wherein the instructions further includes a population-level visualizer, and the processor is further configured to: execute the instructions for the population-level visualizer to construct a three-dimensional graph of aortic valve area versus mean transaortic pressure gradient versus and peak aortic jet velocity, including: a data point for the aortic valve area measurement, the mean transaortic pressure gradient measurement, and the peak aortic jet velocity measurement for the subject of interest; data points for the aortic valve area measurements, the mean transaortic pressure gradient measurements, and the peak aortic jet velocity measurements of the subjects with aortic stenoses; and an aortic stenosis threshold plane identifying combinations of values of the aortic valve area, the mean transaortic pressure gradient and the peak aortic jet velocity that indicate severe aortic stenosis; and cause the display monitor to display the three-dimensional graph.

5. The system of claim 4, wherein the processor is further configured to construct a two-dimensional graph including the aortic valve area and the mean transaortic pressure gradient and cause the display monitor to display the two-dimensional graph.

6. The system of claim 4, wherein the processor is further configured to construct a two-dimensional graph including the aortic valve area and the peak aortic jet velocity and cause the display monitor to display the two-dimensional graph.

7. The system of claim 2, wherein the processor is further configured to: extract information from the digital information repository(s) for subjects with aortic stenoses that do not have a prosthetic valve; process the extracted information to at least one of remove outliers, impute missing information, represent repeated measurements, or extract free text; and perform an analysis on the processed extracted data to determine a set of risk factors of aortic stenosis progression.

8. The system of claim 7, wherein the analysis includes performing a univariate analysis for each subject of the subjects to determine initial risk factors associated with aortic valve area, mean transaortic pressure gradient and peak aortic jet velocity, followed by a multivariate analysis of the initial risk factors to determine the set of risk factors associated with the aortic valve area, the mean transaortic pressure gradient and the peak aortic jet velocity.

9. The system of claim 7, wherein the processor is further configured to model each of aortic valve area, mean transaortic pressure gradient and peak aortic jet velocity based on the set of risk factors and predict a severity of aortic stenosis of the subject of interest based on the models.

10. The system of claim 9, wherein the processor is further configured to classify the severity of an aortic stenosis of the subject of interest based on the model each of the aortic valve area, the mean transaortic pressure gradient, and the peak aortic jet velocity.

11. A computer-implemented method, comprising: obtaining information about a subject, including at least an aortic valve area measurement, a mean transaortic pressure gradient measurement, and a peak aortic jet velocity measurement for the subject; obtaining instructions for an aortic stenosis classifier; executing the instructions to classify a severity of an aortic stenosis of the subject of interest based at least on the aortic valve area measurement, the mean transaortic pressure gradient measurement, and the peak aortic jet velocity measurement for the subject of interest; and visually presenting the classified severity.

12. The computer-implemented method of claim 11, further comprising: extracting information about subjects with aortic stenoses, including at least aortic valve area measurements, mean transaortic pressure gradient measurements, and peak aortic jet velocity measurements; and training the aortic stenosis classifier with at least the aortic valve area measurements, the mean transaortic pressure gradient measurements, and the peak aortic jet velocity measurements of the subjects with aortic stenoses.

13. The computer-implemented method of claim 12, further comprising: constructing at least one of a two-dimensional graph of aortic stenosis versus time or a three-dimensional graph of three-dimensional graph of aortic valve area versus mean transaortic pressure gradient versus and peak aortic jet velocity; and displaying the constructed the at least one of the two-dimensional graph or the three-dimensional graph.

14. The computer-implemented method of claim 11, further comprising: extracting information for subjects with aortic stenoses that do not have a prosthetic valve; processing the extracted data to at least one of remove outliers, impute missing information, represent repeated measurements, or extract free text; and performing an analysis on the processed extracted data to determine a set of risk factors of aortic stenosis progression.

15. The computer-implemented method of claim 14, further comprising: modelling each of aortic valve area, mean transaortic pressure gradient, and peak aortic jet velocity based on the set of risk factors; and predicting a severity of aortic stenosis for the subject based on the models.

16. A computer-readable storage medium storing computer executable instructions which when executed by a processor of a computer cause the processor to: obtain information about a subject, including at least an aortic valve area measurement, a mean transaortic pressure gradient measurement, and a peak aortic jet velocity measurement for the subject, from a digital information repository; obtain instructions for an aortic stenosis classifier; execute the instructions to classify a severity of an aortic stenosis of the subject of interest based at least on the aortic valve area measurement, the mean transaortic pressure gradient measurement, and the peak aortic jet velocity measurement for the subject of interest; and visually present the classified severity.

17. The computer-readable storage medium of claim 16, wherein the computer executable instructions further cause the processor to: extract information about subjects with aortic stenoses, including at least aortic valve area measurements, mean transaortic pressure gradient measurements, and peak aortic jet velocity measurements; and train the aortic stenosis classifier with at least the aortic valve area measurements, the mean transaortic pressure gradient measurements, and the peak aortic jet velocity measurements of the subjects with aortic stenoses.

18. The computer-readable storage medium of claim 17, wherein the computer executable instructions further cause the processor to: construct at least one of a two-dimensional graph of aortic stenosis versus time or a three-dimensional graph of three-dimensional graph of aortic valve area versus mean transaortic pressure gradient versus and peak aortic jet velocity; and display the constructed the at least one of the two-dimensional graph or the three-dimensional graph.

19. The computer-readable storage medium of claim 16, wherein the computer executable instructions further cause the processor to: extract information for subjects with aortic stenoses that do not have a prosthetic valve; process the extracted data to at least one of remove outliers, impute missing information, represent repeated measurements, or extract free text; and perform an analysis on the processed extracted data to determine a set of risk factors of aortic stenosis progression.

20. The computer-readable storage medium of claim 19, wherein the computer executable instructions further cause the processor to: model each of aortic valve area, mean transaortic pressure gradient, and peak aortic jet velocity based on the set of risk factors; and predict a severity of aortic stenosis for the subject based on the models.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The invention may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the embodiments and are not to be construed as limiting the invention.

[0009] FIG. 1 diagrammatically illustrates an example system configured for aortic stenosis classification, classification visualization, risk factor determination, and/or severity prediction, in accordance with an embodiment(s) herein.

[0010] FIG. 2 diagrammatically illustrates an example of the aortic stenosis classifier, in accordance with an embodiment(s) herein.

[0011] FIG. 3 diagrammatically illustrates example training of the aortic stenosis classifier, in accordance with an embodiment(s) herein.

[0012] FIG. 4 diagrammatically illustrates further example training of the aortic stenosis classifier, in accordance with an embodiment(s) herein.

[0013] FIG. 5 diagrammatically illustrates an example of the classification AS visualization engine, in accordance with an embodiment(s) herein.

[0014] FIG. 6 graphically shows example display of individual-level classification in two-dimensions, in accordance with an embodiment(s) herein.

[0015] FIG. 7 graphically shows example display of population-level data classification in three-dimensions, in accordance with an embodiment(s) herein.

[0016] FIG. 8 graphically shows a two-dimensional view of the population-level classification of FIG. 7, in accordance with an embodiment(s) herein.

[0017] FIG. 9 graphically shows another two-dimensional view of the population-level classification of FIG. 7, in accordance with an embodiment(s) herein.

[0018] FIG. 10 diagrammatically illustrates an example of the AS factor identifier, in accordance with an embodiment(s) herein.

[0019] FIG. 11 diagrammatically illustrates an example of the AS severity predictor, in accordance with an embodiment(s) herein.

[0020] FIG. 12 diagrammatically illustrates a variation of the aortic stenosis classifier of FIG. 1, in accordance with an embodiment(s) herein.

[0021] FIG. 13 illustrates an example method, in accordance with an embodiment(s) herein.

[0022] FIG. 14 illustrates another example method, in accordance with an embodiment(s) herein.

[0023] FIG. 15 illustrates yet another example method, in accordance with an embodiment(s) herein.

[0024] FIG. 16 illustrates still another example method, in accordance with an embodiment(s) herein.

DETAILED DESCRIPTION OF EMBODIMENTS

[0025] FIG. 1 diagrammatically illustrates an example system 102. The system 102 includes a digital information repository(s) 104 and a computing apparatus 106 (e.g., a computer or the like). The digital information repository(s) 104 includes a physical storage medium that stores digital information. This includes local, remote, distributed, and/or other physical storage medium. In one instance, the digital information repository(s) 104 includes at least information for subjects diagnosed with aortic stenosis, including historical aortic stenosis diagnosis, measurements (e.g., AVA, MPG and Vmax), and/or other information.

[0026] The illustrated computing apparatus 106 includes a processor 108 (e.g., a central processing unit (CPU), a microprocessor (μCPU), and/or other processor) and computer readable storage medium (“memory”) 110 (which excludes transitory medium) such as a physical storage device like a hard disk drive, a solid-state drive, an optical disk, and/or the like. The processor 108 is configured to execute the instructions. Input/output (“I/O”) 114 is configured for communication between the computing apparatus 106 and the digital information repository(s) 104, including receiving data from and/or transmitting a signal to the digital information repository(s) 104.

[0027] A human readable output device(s) 118, such as a display, is in electrical communication with the computing apparatus 106. In one instance, the human readable output device(s) 118 is a separate device configured to communicate with the computing apparatus 106 through a wireless and/or a wire-based interface. In another instance, the human readable output device(s) 118 is part of the computing apparatus 106. An input device(s) 116, such as a keyboard, mouse, a touchscreen, etc., is also in electrical communication with the computing apparatus 106.

[0028] The memory 110 includes instructions 120 at least for an aortic stenosis (AS) classifier 122, a AS classification visualization engine 124, an AS risk factor identifier 126, and/or AS severity predictor 128. As described in greater detail below, in one example, the aortic stenosis classifier 122 classifies aortic stenosis for a subject(s), the AS classification visualization engine 124 causes visualization of such classification, the AS risk factor identifier 126 identifies risk factors related to the progression of AS, and/or the AS severity predictor 128 predicts a severity of AS for the subject(s) based on the identified risk factors and/or provides information to the aortic stenosis classifier 122. In one instance, this provides information for a more accurate assessment of AS severity and/or a progression thereof, relative to a configuration in which the instructions 120 are not utilized and/or absent.

[0029] FIG. 2 diagrammatically illustrates an example of the aortic stenosis classifier 122. A data extractor 202 receives a subject identification (ID) of a subject of interest, e.g., from the input device(s) 116. The data extractor 202 extracts parameters for the subject of interest relevant to classifying a severity of an aortic stenosis of the subject of interest, e.g., AVA, MPG, and Vmax measurements, from the digital information repository(s) 104. A trained classifier 204 receives the extracted parameters. The trained classifier 204 classifies the aortic stenosis of the subject of interest based on the received extracted parameters. In this example, the aortic stenosis classifier 204 causes the display of the classification via a display monitor of the output device(s) 118.

[0030] In one instance, the trained classifier 204 includes a machine learning algorithm(s) such as linear regression, logistic regression, KNN classification, Support Vector Machine (SVM), decision trees, random forest, artificial neural network, K-means clustering, naive Bayes theorem, Recurrent Neural Networks (RNN) algorithm, and/or other machine learning and/or other artificial intelligence algorithm(s). In one example, the trained classifier 204 utilizes majority voting based on multiple classifiers. By using majority voting, the trained classifier 204, in one instance, can provide better performance, e.g., when evaluating different classification problems, relative to an algorithm not including majority voting.

[0031] FIG. 3 diagrammatically illustrates example training of the aortic stenosis classifier 122. A data extractor 302 extracts parameters for a subject's diagnosed aortic stenosis from the digital information repository(s) 104. The data extractor 302 can be part of the instructions 112 of the computing apparatus 106 and/or other instructions of a different computing apparatus. In one instance, the parameters are extracted from a single healthcare entity/site. In another instance, the parameters are extracted from multiple healthcare entities / sites.

[0032] In this example, the extracted parameters represent a training set of data that includes historical aortic stenosis diagnosis of subjects in the digital information repository(s) 104, including at least the AVA, MPG, and Vmax measurements, fed to the aortic stenosis classifier 122. In one instance, the aortic stenosis classifier 122 is configured to determine relations between variables in making a solution, such as decision tree, logistic regression, etc. The aortic stenosis classifier 122 will identify values for each variable and/or combinations of values for variables that lead to each one of the classification values.

[0033] FIG. 4 diagrammatically illustrates further example training of the aortic stenosis classifier 122. In this example, the input data and classification results for a particular subject are provided to the aortic stenosis classifier 122 and used to dynamically update the trained classifier 204. In one instance, the trained classifier 204 is dynamically updated as classification results become available. In another instance, the trained classifier 204 is dynamically updated with classification results on demand. In yet another instance, the trained classifier 204 is dynamically updated with classification results based on a schedule.

[0034] FIG. 5 diagrammatically illustrates an example of the AS classification visualization engine 124. The illustrated AS classification visualization engine 124 includes an individual-level visualizer 502 and/or a population-level visualizer 504. The AS classification visualization engine 124 causes display of information to assist clinicians in checking the parameters of a subject of interest. This information includes historical diagnoses of the subject of interest and hemodynamic parameters of the subject of interest and other subjects with aortic stenosis. In one instance, this allows a clinician to visually observe similar cases to assist diagnosis.

[0035] FIG. 6 graphically shows an example display of individual-level data 600 generated by the individual-level visualizer 502 of FIG. 5 in two-dimensions. In this example, the individual level data 600 includes historical diagnosis of a subject. A first axis 602 represents severity and a second axis 604 represents time. In this example, the first axis 602 includes: a mild severity level 606, a mild to moderate severity level 608, a moderate severity level 610, a moderate to severe severity level 612 and a severe severity level 614, and the second axis 604 indicates dates 616, 618 and 620 on which severity levels 622, 624 and 626 were determined. This display represents the progress of the disease and, in one instance, assists clinicians with understanding how fast and how severe the AS develops.

[0036] FIG. 7 graphically shows an example display of population-level data 700 generated by the population-level visualizer 504 of FIG. 5 in three-dimensions. A first axis 702 represents AVA, a second axis 704 represents MPG, and a third axis 706 represents Vmax. Each circle 708 represents a data point for each of the population subjects. A circle 709 represents a data point for the subject of interest. An AVA threshold plane 710 represents an aortic stenosis threshold, where the circles 708 under the AVA threshold plane 510 represent subjects likely with severe aortic stenosis and the circles 708 above the AVA threshold plane 510 represent subjects likely with non-severe aortic stenosis. In FIG. 7, a gray level shading is used with the circles 708 and 709 to distinguish between levels of severity with respect to the AVA threshold plane 510.

[0037] In this example, the AVA threshold plane 710 is not a horizontal plane at 1 cm.sup.2, as recommended by the guideline. That is, the four corners of the AVA threshold plane 710 do not intersect the vertical axes at the same point. On the first axis, a level 712 represents the guideline value of 1 cm.sup.2. In another instance, the AVA threshold plane 710 is a horizontal plane. In this instance, the AVA threshold plane 710 is at the guideline value of 1 cm.sup.2. In another instance, it is not at the guideline value of 1 cm.sup.2. The population-level data 700 visually assists clinicians with understanding the distribution of the three parameters (AVA, MPG and Vmax) and the classified groups with respect to the whole population.

[0038] FIG. 8 graphically shows a two-dimensional view 800 of two axes of the population-level data 700 of FIG. 7, showing the first axis 702 (AVA) and the second axis 704 (MPG). The AVA threshold plane 710 is shown in two dimensions.

[0039] FIG. 9 graphically shows another two-dimensional view 900 of a different combination of two axes of the population level data 700 of FIG. 7, showing the first axis 702 (AVA) and the third axis 706 (Vmax). The AVA threshold plane 710 is shown in two dimensions. In another instance, the graphical display shows the second axis 704 (MPG) and the third axis 706 (Vmax) in two dimensions.

[0040] The above examples are described with particular application to aortic stenosis. However, it is to be understood that the above can be utilized for other diseases. For example, the approach described herein also applies to cardiovascular diseases that are diagnosed based only on cutoffs of a series of parameters, e.g., mitral regurgitation, mitral stenosis, etc. The degree of mitral regurgitation depends on, e.g., the volume of blood that flow through the mitral valve, aortic valve and the regurgitant orifice area, etc. The degree of mitral stenosis depends on, e.g., mean gradient, mitral valve area, etc.

[0041] FIG. 10 diagrammatically illustrates an example of the AS factor determiner 126 in connection with the digital information repository(s) 104. In this example, the AS factor determiner 126 includes a data extractor 1002, a pre-processor 1004 and a statistical analyzer 1006.

[0042] The data extractor 1002 is configured to extract certain data from the digital information repository(s) 104. For example, in one instance the data extractor 1002 is configured to include information for subjects who have been subject to an adult transthoracic echocardiogram (TTE) and diagnosed with at least mild AS and exclude information for subjects that have prosthetic valves. The included information includes hemodynamic parameters measured by echocardiogram (i.e. AVA, NPG, Vmax, etc.), findings made by echo cardiologists, lab test results, vital signs, medications, problem lists, demographics, clinical notes, and/or other information.

[0043] The pre-processor 1004 is configured to pre-process the data extracted from the digital information repository(s) 104 with the data extractor 1002. This can be based on rules and/or otherwise. In one instance, this includes removing outliers. An example of an outlier includes erroneous data, e.g., due to recording and/or other errors. Taking AVA, for example, since a subject with AS cannot get better without surgery, if an AVA measurement has a more than a 15% increase comparing with the previous measurement, this measurement is considered an outlier and removed, discarded, ignored, and/or otherwise not utilized.

[0044] Additionally, or alternatively, this includes imputing missing information. This includes utilizing an algorithm tailored to the particular missing data. Additionally, or alternatively, this includes representing repeated data. This also includes utilizing an algorithm tailored to the repeated data. For example, where multiple temperatures taken at different points in time, the median and/or other value of a specified time window can be used to represent the temperature. Measurements like temperature, blood pressure, etc. are commonly measured repetitively. In one instance, natural language processing is used to extract features from information represented in free-text such as free-text discharge summary, clinical notes, etc.

[0045] The statistical analyzer 1006 includes a univariate analyzer 1008 and a multivariate analyzer 1010. The univariate analyzer 1008 is configured to employ a statistical model on each of the features from the pre-processor 1004. In one instance, this includes analyzing, for each factor, its interaction with time interval by a Linear Mixed-Effects (LME) model. For example, in one instance a LME model is used to consider a random intercept and a random slope. In this instance, if a feature is significantly associated with the hemodynamic parameters AVA, MPEG and Vmax (e.g., significance level=0.05), the feature is provided to the multivariate analyzer 1010. Otherwise, it is not. The multivariate analyzer 1010 is configured to employ a statistical model on the features from the univariate analyzer 1008. In one instance, an LME model is used to evaluate the significance of the interaction terms of each feature and the time interval. Features considered significant features (e.g., significance level=0.05) output as risk factors.

[0046] Table 1 below shows an example for ΔAVA. In this example, age, gender, left ventricular (LV) systolic function, and AVA calculated by the continuity equation were extracted for subgroup analysis. The AVA from subsequent echocardiograms of the same patient are compared to determine an annual rate of AVA change (ΔAVA). To estimate the annual AVA progression rate for each risk factor, the LME model includes the risk factor and the time interval between the measurements. The fixed effects include the risk factor, the time interval and their interaction term. The random effects include the time interval and a random intercept in order to account for the differing baseline AVA and progression rates due to individual characteristics that are not explained by the risk factor.

[0047] The significance of the risk factor upon the progression rate, in one instance, is analyzed from the p-value of the interaction term in the fixed effects. A median age at the time of the initial echocardiogram was 75 years, and 44% were male. A rate of progression of AS was −0.062±0.003 cm.sup.2/year. This rate was not influenced by age or LV function (P-value>0.05). However, the rate was influenced by gender (P-value<0.05) being more rapid in men compared to women. There is an inverse relationship between initial severity of AS and progression rate. The information in Table 1 provides clinical information on the expected interval before intervention is needed for severe AS.

TABLE-US-00001 TABLE 1 Risk Factor Identification. ΔAVA (cm.sup.2/year) n Annual rate of change (β) P-value All subjects 916 −0.062 ± 0.003 Age Age ≥75 yr 468 −0.057 ± 0.004 0.08 Age <75 yr 448 −0.066 ± 0.004 Gender Male 404 −0.068 ± 0.004 0.03 Female 512 −0.057 ± 0.003 AVA Mild (>1.5 cm.sup.2) 292   .sup. −0.082 ± 0.004 (*) <0.001 (* vs {circumflex over ( )}) Moderate (1.0-1.5 cm.sup.2) 466  −0.057 ± 0.003 ({circumflex over ( )}) Severe (<1.0 cm.sup.2) 158   −0.023 ± 0.008 (+) <0.001 ({circumflex over ( )} vs +) LV Normal 773 −0.061 ± 0.003 0.29 Function Reduced 143 −0.069 ± 0.008

[0048] FIG. 11 diagrammatically illustrates an example of the AS severity predictor 128 in connection with the AS factor determiner 126 and the output device(s) 118. The AS severity predictor 12, in this example, utilizes an LME model for each of the hemodynamic parameters AVA, MPG and Vmax. The following provides an example for AVA that can also be used for MPG and/or Vmax. In Table 1, since all the risk factors are categorical, the p-value using student's t-test is used to determine statistical significance of the progression rate, compared to the reference level of each risk factor.

[0049] Using this LME model, by imputing the time interval, the three hemodynamic parameters can be predicted in the future. In one instance, a predictive performance of a Logistic Regression classifier predicting severity of the AS is utilized using unseen test data. Given the hemodynamic parameters AVA, MPG and Vmax, the accuracy of identifying severe versus non-severe AS was 93%. The area under the curve (AUC), sensitivity, specificity, positive predictive value, negative predictive value, f score of the classifier were 0.98, 0.94, 0.94, 0.91, 0.95, and 0.92, respectively.

[0050] FIG. 12 diagrammatically illustrates an example in which the output of the LME model for each of the hemodynamic parameters AVA, MPG and Vmax is provided to the aortic stenosis classifier 122 of FIG. 2 and utilized thereby, in addition or alternative to, the information extracted from the digital information repository(s) 104, as described above. For example, in one instance the aortic stenosis classifier 122 utilizes the output of the LME model, in addition or alternative to, the AVA, MPG and Vmax measurements from the digital information repository(s) 104 to classify the aortic stenosis of the subject of interest based thereon. Similar to above, the aortic stenosis classifier 122, in one example, causes the display of the classification via the output device(s) 118.

[0051] FIGS. 13, 14, 15 and 16 illustrate example methods in accordance with an embodiment(s) herein. It is to be appreciated that the ordering of the acts of one or more of the method is not limiting. As such, other orderings are contemplated herein. In addition, one or more acts may be omitted, and/or one or more additional acts may be included.

[0052] FIG. 13 illustrates an example method in accordance with an embodiment(s) herein. A training step 1302 trains a classifier, as described herein and/or otherwise. For example, in one instance the training step 1302 can train the classifier with AVA, MPG, Vmax and/or other information of subjects with aortic stenosis from the digital information repository(s) 104. A classifying step 1304 classifies aortic severity of a subject of interest, as described herein and/or otherwise. For example, in one instance classifying step 1304 classifies aortic severity of a subject of interest based on at least with AVA, MPG, Vmax of the subject of interest. A visualizing step 1306 visualizes the classification, as described herein and/or otherwise. For example, in one instance the visualizing step 1306 constructs and displays two and/or three dimensional graphs of aortic stenosis progression and/or AVA, MPG and/or Vmax.

[0053] FIG. 14 illustrates another example method in accordance with an embodiment(s) herein. A data collection step 1402 extracts certain data, as described herein and/or otherwise. For example, in one instance the data collection step 1402 extracts hemodynamic parameters of subjects who have had an adult transthoracic echocardiogram (TTE) and was diagnosed with at least mild AS, the parameters including AVA, NPG and Vmax, findings made by echo cardiologists, lab test results, vital signs, medications, problem lists, demographics, clinical notes, and/or other information.

[0054] A data pre-processing step 1404 pre-processes the extracted data, as described herein and/or otherwise. For example, in one instance the data pre-processing step 1404 at least one of removes outliers, imputes missing information, represents repeated measurements, or extracts from free text. An analysis step 1406 analyzes the pre-processed extracted data to determine a set of risk factors, as described herein and/or otherwise. For example, in one instance the analysis step 1406 filters the pre-processed extracted data via univariate analysis followed by multivariate analysis to keep only the factors considered significant based on a predetermined threshold.

[0055] FIG. 15 illustrates yet another example method in accordance with an embodiment(s) herein. A modelling step 1502 models hemodynamic parameters based on the risk factors of FIG. 14, as described herein and/or otherwise. For example, in one instance, an LME model is built for each of the hemodynamic parameters AVA, MPG and Vmax. A predicting step 1504 predicts a severity of aortic stenosis of a subject based on the hemodynamic parameters, as described herein and/or otherwise.

[0056] FIG. 16 illustrates still another example method in accordance with an embodiment(s) herein. A modelling step 1602 models hemodynamic parameters based on the risk factors of FIG. 14, as described herein and/or otherwise. For example, in one instance, an LME model is built for each of the hemodynamic parameters AVA, MPG and Vmax. A predicting step 1604 predicts a severity of aortic stenosis of a subject based on the hemodynamic parameters, as described herein and/or otherwise. A classifying step 1604 classifies aortic severity, as described herein and/or otherwise. For example, the classification step of FIG. 13 can additionally include training the classifier with the modelled hemodynamic parameters and then classifying aortic stenosis severity.

[0057] One or more of the above may be implemented by way of computer readable instructions, encoded or embedded on computer readable storage medium, which, when executed by a computer processor(s), cause the processor(s) to carry out the described acts. Additionally, or alternatively, at least one of the computer readable instructions is carried out by a signal, carrier wave or other transitory medium, which is not computer readable storage medium.

[0058] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

[0059] The word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage.

[0060] A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope.