WO2021069271A1 - Computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure - Google Patents
Computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure Download PDFInfo
- Publication number
- WO2021069271A1 WO2021069271A1 PCT/EP2020/077301 EP2020077301W WO2021069271A1 WO 2021069271 A1 WO2021069271 A1 WO 2021069271A1 EP 2020077301 W EP2020077301 W EP 2020077301W WO 2021069271 A1 WO2021069271 A1 WO 2021069271A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- bias
- measurement
- measurement data
- medical images
- anatomical structure
- 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.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/08—Clinical applications
- A61B8/0883—Clinical applications for diagnosis of the heart
-
- 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
- G16H70/00—ICT specially adapted for the handling or processing of medical references
- G16H70/20—ICT specially adapted for the handling or processing of medical references relating to practices or guidelines
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/46—Ultrasonic, sonic or infrasonic diagnostic devices with special arrangements for interfacing with the operator or the patient
- A61B8/467—Ultrasonic, sonic or infrasonic diagnostic devices with special arrangements for interfacing with the operator or the patient characterised by special input means
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/5215—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data
- A61B8/5223—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data for extracting a diagnostic or physiological parameter from medical diagnostic data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06398—Performance of employee with respect to a job function
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
- G06T7/0014—Biomedical image inspection using an image reference approach
-
- 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
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Definitions
- the present invention relates to a computer-implemented method for automated analysis of a user-specific bias in measurements performed on medical images of an anatomical structure, a computer program, and an automated measurement tool configured to perform the method.
- Medical ultrasound measurements are used in a wide field of medical treatment.
- echocardiographic measurements are typically used as an indirect method to assess the dimensions and the function of a human' s or animal' s anatomical structure, specifically of the human's heart.
- measurement results are subject to inter-observer and intra-observer discrepancies.
- the Bland-Altman analysis is a common method used in the prior art for determining both the variability and the bias between two different observers (such as users or physicians).
- labs that perform measurements specifically echocardiography labs
- this accreditation includes: Interpretive Quality Review (Physician Interpretation Variability).
- the facility must evaluate the quality and accuracy of the interpretation based on the acquired images.
- a minimum of two cases per modality per quarter must be evaluated for the quality and accuracy of the interpretation of the acquired images.
- the cases must represent as many physicians as possible. Differences in interpretation must be reconciled to achieve uniform examination interpretation. In daily clinical routine, this translates into an additional manual and organizational workload. This is even worse in case of long-term clinical studies in which the measurement observers often change during the realization of the study. Therefore, there is a problem of inter-observer variability and consensus training.
- US 2018150598 A1 discloses to compare the anatomical measurement acquired using machine learning with a predetermined threshold and to generate advisory information to adjust the measurement.
- Vikram Chalana focuses in “A Methodology for Evaluation of Boundary Detection Algorithms on Medical Images” IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 16, NO. 5, OCTOBER 1997 on the comparison of differences in contouring. That is, it compares computer-generated boundaries to multiple observer-created boundaries and calculates a Cohen Kappa-coefficient of agreement between two observers.
- US 2014341449 A1 discloses to reduce inter observer and intra observer variability in contouring images by verifying the consistency using consensus contours generated by machine learning. Further, according to US 2014341449 A1 a degree of deviation of the user created contour from the consensus contour is evaluated.
- the echocardiographic measurements are one example of an indirect method to assess the dimensions and the function of the human heart. Between the measurements of two observers (inter observer) or even between two measurement instances of the same observer (intra observer), the results are bound to vary.
- a common method for the analysis of both the variability (i.e. the margin of error, or 95% confidence interval, of the difference between two observers) and the bias (i.e. the systematic error, or the mean difference between two observers) is the Bland-Altman-Analysis.
- the Bland-Altman method can be used to evaluate the agreement between two series of measured values. In particular, it provides information on the influence of the height of the measured values on the size of the deviations/differences.
- the differences (measurement differences) are represented as a function of the averages (measured value height) in a scatter diagram (Bland-Altman plot).
- a computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure is presented in claim 1. Useful embodiments are set out in the dependent claims.
- a computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure comprises the steps of: receiving a plurality of medical images, wherein the plurality of medical images includes at least a first plurality of medical images and a second plurality of medical images, and each plurality of medical images includes the same anatomical structure of the same or different patient(s), receiving, for the first plurality of medical images, first measurement data of a feature of the anatomical structure, wherein the first measurement data is obtained by a first user, determining, for the first plurality of medical images, a first baseline measurement data of the feature of the anatomical structure by means of an automated measurement tool, determining a first measurement bias for the first user based on the first measurement data and the first baseline measurement data, receiving, for the second plurality of medical images, second measurement data of a feature of the anatomical structure, wherein the second measurement data is obtained by a second user, the second user being different from the first user, determining, for the second plurality of medical
- the invention provides a method in which the same medical images are analysed by both a human user and an automated measurement tool, and the results are compared with each other to obtain a measurement bias for the human user.
- two different measurement instances or observes may be used on the same medical images, and a bias is determined between them. This bias may then be used to estimate the bias between different humans, such as the first and the second user.
- a bias may be a measure of how much the measurement data differs from the true underlying quantitative parameter being estimated. For example, a measured value is biased if it is determined in such a way that it is systematically different from the population parameter being estimated.
- the bias may be a discrepancy metric (i.e. deviation metric) or a margin of error, in other words, the systematic error or the mean difference between two users (i.e. observers).
- the present invention is particularly useful in the latter cases. That is, according to the present invention the baseline measurement data (to be described below) is attained by an automated measurement tool (which might be considered as said different apparatus). Further, the automated measurement tool may be also referred to as a baseline machine. In other words, the present invention provides a calibration/normalization tool in which the measurement bias of a given user may be associated to the systematic (and subjective) difference that said user generates when doing a predetermined measurement on the anatomical structure included in the plurality of medical images.
- the first and second measurement bias may be an absolute measurement bias or a relative measurement bias.
- the first and second absolute measurement bias for the first and second users are preferably determined by calculating the difference between the first/second measurement data and the first/second baseline measurement data, respectively, i.e. by subtracting one from the other.
- the relative measurement bias may be defined by being normalized with respect to the measurement data (i.e. to the kind of measurement). In other words, the relative measurement bias may be determined based on a relative difference of the measurement data and the baseline measurement data instead of the difference of the measurement data and the baseline measurement data (i.e. as compared to the absolute measurement bias).
- the relative measurement bias may be determined by the following equation:
- Rx represents the measurement data
- Cx represents the baseline measurement data
- the absolute bias may have a functional relationship to data and/or physical properties of the measurement data on which it is based.
- the absolute bias may have a functional relationship to properties of the modality used in creating the medical images, to properties of the measurement tool used in creating the measurement data, to circumstances present during the measurement and/or during the creating of the medical images and/or to absolute values of the measurement data. Therefore, the absolute measurement bias may not be readily compared with other absolute measurement biases that are obtained based on different measurement data (e.g. on measurement data that includes a measurement of a different feature of the anatomical structure).
- the relative bias may be independent from such functional relationships. As a result, the relative measurement bias may be compared with other relative measurement biases based on different measurement data.
- a relative difference between the measurement data and baseline measurement data may be determined. That is, by determining the relative difference between the measurement data and baseline measurement data, the relative bias may be determined (for example as outlined above).
- the relative bias is especially useful in the following case: If a relatively small dimension is measured (for example the thickness of a septum of a heart), a 10 % deviation from the baseline measurement data results in a relatively small absolute bias. On the other hand, if a relatively large dimension is measured (for example the diameter of the left ventricle of the heart), a 10 % deviation results in a relatively high absolute bias. As a result, although both measurements have the same percentage deviation, the absolute bias for each measurement differs.
- the relative bias may be determined.
- relative measurement biases may be compared with each other independent of the measurement data on which the measurement biases are based. In other words, independent of the feature of the anatomical structure which is measured, for example.
- bias In the following description, for convenience of explanation, the absolute bias and the relative bias are simply referred to as bias. However, the term bias always refers to both the absolute bias and the relative bias.
- a per-user bias may be determined.
- the per-user bias may be represented by a mean value of all measurement biases of a specific user.
- Medical images may be generated by any medical imaging modality, such as ultrasound, in particular echocardiography, or MRI (Magnetic Resonance Imaging), X-Ray Imaging, CT (Computed Tomography), PET (Positron Emission Tomography), and so on. When using ultrasound, the images may also be obtained by Colour Doppler, Tissue Doppler, Continuous Wave Doppler, Pulsed Wave Doppler or M-Mode, i.e., there is a variety of acquisition modes that may be the basis for manual or automated measurements.
- the medical images may be 2D images or 3D images.
- the medical images (or one set of medical images) may be one or several or all of a sequence of frames of 2D or 3D images that constitute a 4D image, wherein time is the fourth dimension.
- the anatomical structure may be any anatomical structure of the human or animal body, such as a limb, an organ or a blood vessel, for example the heart, the liver, or a fetus.
- the same structure means that if the first plurality of first medical images includes the heart as the anatomical structure, the second plurality of medical images also includes the heart as the anatomical structure.
- the medical images may be derived from the same patient or from different patients, that is, the medical images may be also derived from the same patient but at different points of time. Further, the method or modality used for obtaining the medical images may be different from one image to the other.
- the first plurality of medical images may be obtained using an echocardiography system, wherein the second plurality of medical images may be obtained using an MR system, for example.
- the first plurality of medical images and the second plurality of medical images may not necessarily have to be received at the same location. That is, the inventive method may be partly executed at different locations, i.e. distributed over several locations.
- only the step of comparing the first measurement bias with the second measurement bias may be centralized (i.e.
- the method step of receiving, for the first plurality of medical images, first measurement data of a feature of the anatomical structure, wherein the first measurement data is obtained by a first user, determining, for the first plurality of medical images, a first baseline measurement data of the feature of the anatomical structure by means of an automated measurement tool and determining a first measurement bias for the first user based on the first measurement data and the first baseline measurement data may be executed at one location.
- the corresponding method steps related to the second plurality of medical images may be executed at a different location. As a result, only the first measurement bias and the second measurement bias have to be transmitted for being compared with each other.
- the first medical images may be received and processed (i.e. receiving the first measurement data and determining the first baseline measurement data) at a core lab in Europe while the second medical images may be received and processed (i.e. receiving the second measurement data and determining the second baseline measurement data) at a core lab in the US.
- the comparison of the first measurement bias and the second measurement bias may be performed at a different location or at any of the above locations. That is, only the first and second measurement biases may have to be transmitted for comparing them. The transmission may be performed via the internet, for example.
- the invention may be extended to a third and further pluralities of medical images, and related third and further measurement data and baseline measurement data, to determine a third and further measurement biases of a third and further users.
- the measurement data represents a physical parameter, e.g. a dimension, of the anatomical structure that is to be analysed, in particular a length, area or volume. It may also be a flow rate, time period or other physical parameter that can be measured from medical images.
- the measurement data may include a diameter and/or a circumference of a heart valve such as the mitral valve and/or the aortic valve.
- the measurement data may be manually taken by at least one user.
- the measurement data may be attained by the user by tracing a structure in the medical image, for example.
- the measurement data may be additionally or alternatively generated by the user by drawing a 3D free hand line in a 3D data set.
- the measurement data may include a contour of an anatomical structure.
- the acquisition of the medical images may have been performed by the same user, who will also carry out the (manual) measurements to obtain the measurement data.
- the examination may also be performed by other clinical staff, as is usually the case in MR examinations.
- each plurality of medical images includes one or several sets of medical images, wherein each set of medical images is obtained during one examination session from one patient.
- one measurement data of a feature of the anatomical structure and one baseline measurement data of the feature of the anatomical structure is obtained from each set of medical images, and preferably each plurality of medical images includes several sets.
- each plurality of medical images will thus include sets of medical images from a plurality of patients.
- a plurality of medical images may also include two or more sets of medical images acquired at different examination sessions from the same patient, as will be the case in longitudinal studies.
- one set of medical images may include all ultrasound images obtained during one examination session from the heart of a patient, thus including a number of 2D and/or 3D images from various viewing angles and with different field of views.
- the user may be any person working in the field of medical image analysis such as a doctor, a nurse, an employee in a measurement lab and so on.
- the first and second user are usually different persons. Possibly, the first and second user may be the same person that creates the measurement data at different times or with different measurement techniques.
- the first (second) measurement data of the feature of the anatomical structure are created by the first (second) user performing manual measurements on one or several of the first (second) plurality of images, for example by measuring a physical parameter, e.g. a length or area.
- a physical parameter e.g. a length or area.
- Such measurement will usually be carried out manually, i.e. requiring user intervention, but with the aid of a computer, e.g. by the user clicking on specific landmarks in the medical image displayed on a computer screen, and the distance between the landmarks being calculated automatically by the computer.
- the measurement data may be obtained using a computer-implemented measurement technique requiring user intervention, e.g. such that the user is responsible for the correctness of the measurement data.
- Receiving may mean that the measurement data is received in real time, that is, directly after the measurement data is obtained.
- the measurement data may be received from a storage, that is, previously generated and stored measurement data may be analysed in hindsight. As a result, older studies performed in the past, may be newly analysed and thus further processed using the inventive method.
- the measurement data may be received by the automated measurement tool.
- the first measurement data may be received by a first automated measurement tool and the second measurement data may be received by a second automated measurement tool, wherein the first and second automated measurement tool may be separated tools.
- the first and second baseline measurement data of the feature of the anatomical structure is obtained automatically, i.e. without user intervention, by the automated measurement tool.
- the user may select a particular medical image, on which the baseline measurement is to be taken, e.g. one medical image from a set of medical images obtained during one examination session.
- the baseline measurement data are obtained by a computer-implemented algorithm.
- algorithm may be a classical algorithm, for example an algorithm using techniques such as feature tracking, segmentation techniques (e.g. thresholding, region growing, pixel histogram analysis etc.).
- it may be a machine-learning algorithm, in particular, an algorithm based on deep-learning techniques and including one or several artificial neural networks.
- Such artificial neural networks may be trained on images from a reference database and measurement data obtained by an expert user, optionally verified by a second user, to minimize any errors.
- Baseline measurement data are determined by means of the automated measurement tool, which may include all standard measurements, that are usually performed on the anatomical structure. Further, the automated measurement tool may determine the baseline measurement data by an algorithm (included in the automated measurement tool) including rules based on a reference database.
- the reference database may be created by at least one expert (human) and the algorithm may be configured to measure bias free and/or with a constant bias with respect to said reference database.
- the results of the automated measurement tool are reproducible, that is, in case the same measurement data are available, the same results (i.e. baseline measurement data) are obtained by the automated measurement tool.
- the first and second baseline measurement data may be determined by the first and second automated measurement tool.
- the automated measurement tool may be operated in the background of a currently executed application so as to determine the baseline measurement data hidden from the user. That is, the user must not necessarily notice that the baseline measurement data is currently being determined.
- the automated measurement tool is preferably consistent in its results, that is, it produces the same results (output), for the same image or sequence of images (input).
- the standard measurements may include all important measurements performed on a specific anatomical structure. In case the anatomical structure is the heart, the important measurements include the measurement of the diameter of the heart valve, the contour of a heart chamber, for example.
- the automated measurement tool may be connected to a PACS (Picture Archiving and Communication System) of a clinic or a lab, for example.
- PACS Picture Archiving and Communication System
- the automated measurement tool may be included in a modality for creating medical images and/or for executing measurements on medical images.
- the measurement bias may be determined directly within the respective modality (i.e. in situ). Therefore, only the measurement bias may have to be transmitted for the purpose of comparing it with further measurement biases. Moreover, even the step of comparing two measurement biases with each other may be executed within the modality and the result of such comparison may be outputted.
- the baseline measurement data may be determined by the automated measurement tool for each of the measurement data. Then the difference between each measurement data and the baseline measurement data may be determined, so there is a baseline measurement data for each measurement data.
- the first measurement bias may then be determined as the average or mean of each of the baseline measurement data for the first user.
- the first measurement bias may be a representation of the systematic difference between measurements performed by the first user, compared to the baseline measurement data obtained. The same holds for the second measurement bias.
- the first measurement bias is compared with the second measurement bias, for example by subtracting one from the other, or by simply reviewing the measurement bias of each user.
- the method of the invention may be extended to further users, e.g. by repeating the steps for a third, fourth and further users.
- the bias comparison results can be used to normalize/calibrate the measurement values between users.
- an automated analysis of two or more observers may be executed.
- the mean bias between the two users can be calculated for all patients.
- This result may be used to give feedback to all involved physicians (users) for routine clinical procedures and/or clinical long-term studies without the need to measure identical studies.
- the medical images measured by the user may be used to perform the analysis using the automated measurement tool in order to minimize error sources between the user and the automated measurement tool.
- the bias comparison results can be used to normalize the measurement values between observers in case of the clinical study scenario.
- measurement results of users may be normalized and/or calibrated.
- the method is applicable worldwide, that is, it is not locally bounded. Further, the different users do not have to be at the same location. Therefore, the method provides an increased flexibility in its application.
- the method may further comprise the step of determining a global bias represented by a mean value of all available measurement biases, and comparing the global bias with one of the measurement biases.
- the global bias may be obtained by transmitting locally determined measurement biases to a centralized institution.
- the global bias may be determined by calculating the average or mean value of all locally determined measurement biases.
- measurement biases may be transmitted to the centralized institution from PACS and/or modalities including the automated measurement tool.
- biases may be used that are already a mean value of a plurality of biases.
- the automated measurement tool could have a systematic bias against all human observers (i.e. all users). By determining the global bias, this systematic bias may be determined.
- each user may compare his/her own bias with the global bias in order to assess his own bias.
- a measurement bias for standard values for specific measurements of anatomical structures may be determined and compared with the global bias. Further, a user may compare his/her own bias with the global bias. Finally, the user may compare both comparison results in order to evaluate himself/herself with respect to the standard values. Thereby, each user may verify himself/herself.
- the first measurement data may be determined based on one specific medical image out of the first plurality of medical images and the first baseline measurement data may be determined based on the same specific medical image.
- the second measurement data may be determined based on one specific medical image out of the second plurality of medical images and the second baseline measurement data may be determined based on the same specific medical image.
- the user that performs a measurement on an anatomical structure in order to obtain measurement data may choose a specific image (i.e. a specific view of the anatomical structure) which promises the most accurate measurement among the set of medical images. Further, the chosen image may be then used to perform the measurement. In order to minimize error sources between human (i.e. the user) and machine (i.e. the automated measurement tool) the image chosen by the user may be also used to perform the analysis by the automated measurement tool. According to this, the process is highly efficient and needs fewer computing resources and less time. However, all available images of the plurality or of each set of medical images may be analysed by the automated measurement tool. As a result, more measurements than in the manual case may be generated, thus leading to a more stable system with less statistical noise in the analysis.
- each measurement data and baseline measurement data is determined based on the same specific medical image from each set of medical images acquired during one examination session of a patient.
- the measurement bias may be continuously updated once new measurement data is available, wherein all available measurement data are used to update the measurement bias, or wherein all available measurement data obtained during a pre-determined past period of time are used to update the measurement bias.
- updated may mean that the bias is newly determined based on new available measurement data. Further, continuously may mean that the measurement bias is frequently updated. Moreover, update may mean that the old measurement bias is replaced by a newly determined measurement bias.
- This update procedure may be executed automatically without any specific actuation by the user. Further, the update may be executed such that the user does not notice that there is an update going on.
- the update may be performed by calculating a new average, which incorporates the new difference between new baseline measurement data and measurement data. The update is carried out as often as possible once new measurement data is available. In some cases, where the method steps are executed at different locations, there may be a delay in updating the measurement bias because no connection for transmitting of the newly determined measurement bias or measurement data exits. Then, the update may be carried out if the newly determined measurement bias or measurement data could be transmitted.
- the measurement biases for each user may be based on measurement data obtained during a pre-determined past period, for example the past one to three months. Older measurement data thus do not contribute anymore. Rather, at any point in time, the measurement bias of a user represents his average bias in all measurements performed during the pre-determined past period. That is, the measurement bias may be updated based on measurement data created in a pre-defmed period of time, so as to update the measurement bias using a moving average of the measurement biases in said period of time. Alternatively, all available measurement data are considered by updating the measurement bias. That is, the newly obtained measurement data (and thus the newly determined measurement bias) may be considered in addition to all old measurement bias/measurement data.
- the measurement bias may be automatically adapted to new situations, that is, to an increased/decreased measuring accuracy of the user (e.g. after the user has attended a training).
- the method may further comprise the step of determining an institution wide bias represented by a mean value of all measurement biases of the first, second and further users belonging to one predetermined group or institution.
- the institution wide bias may be a mean value of all measurement biases of the users belonging to said institution, wherein the institution may be a core lab, clinic, research team etc.
- the institution wide bias may be the mean bias of a group of users constituted by a predetermined number of users.
- Mean value may be the sum of all values divided by the number of values.
- the method may further comprise the step of comparing the institution wide bias with one of the measurement biases so as to obtain a per-user bias relative to the institution wide bias.
- the per-user bias may be the individual bias of a user.
- the institution wide bias may be the same as outlined above. Accordingly, each user may be evaluated with respect to the mean value of all user' s measurement biases belonging to the institution. Further, the evaluation is facilitated in that it is easily driveable which user improves the institution wide bias and which user degrades the institution wide bias.
- the step of comparing the institution wide bias with one of the measurement biases may include: determining a difference by subtracting each of the measurement biases by the institution wide bias so as to determine the difference for each user belonging to the same group or institution, and determining the user or users having the highest difference(s) among all users.
- an additional value is calculated (i.e. the difference) representing the difference between the mean value of the institution wide bias and the individual per-user measurement bias. Accordingly, the determination which user has to improve his/her measurement accuracy is facilitated.
- the above embodiment allows to create an agenda on which user needs to increase his/her accuracy in order to improve the institution wide bias to a predetermined level.
- the determination of at least the first baseline measurement data and the second baseline measurement data may take place at different locations.
- the measurement data may be further processed (including the determination of the baseline measurement data and the determination of the measurement bias) at the same location where the measurement data were obtained ore received, for example. That is, the automated measurement tool may be present at the respective location where the measurement data is obtained or received so as to locally determine the baseline measurement data.
- the measurement bias may be also locally determined at the same location where the measurement data is obtained or received. As a result, only the baseline measurement data and/or the measurement bias have to be transmitted for the further processes.
- the automated measurement tool may use artificial intelligence for determining the baseline measurement. That is, the automated measurement tool may operate in a rule-based manner. For example, the automated measurement tool may automatically perform a grey value analysis of the medical image so as to determine boundaries between tissue and blood flow. Further, the automated measurement tool may perform an image analysis of the medical image so as to find specific geometrical objects. Based on those analyses, the measurement tool may identify the feature of the anatomical structure that is to be measured. Subsequently, the automated measurement tool may perform the measurement so as to obtain the baseline measurement data.
- the automated measurement tool may use artificial intelligence for determining the baseline measurement. That is, the automated measurement tool may operate in a rule-based manner. For example, the automated measurement tool may automatically perform a grey value analysis of the medical image so as to determine boundaries between tissue and blood flow. Further, the automated measurement tool may perform an image analysis of the medical image so as to find specific geometrical objects. Based on those analyses, the measurement tool may identify the feature of the anatomical structure that is to be measured. Subsequently, the automated measurement
- the automated measurement tool may use machine learning for determining the baseline measurement, wherein the automated measurement tool may be trained by a reference data base including reference images of the anatomical structure and measurement data obtained by at least one expert user for said reference medical images.
- the reference data base includes examples of standard measurements (i.e. example measurements dates) generated by experts.
- the standard measurements may include all important measurements performed on a specific anatomical structure.
- the automated measurement tool may measure an arbitrary anatomical structure based on the example dates included in the reference data base.
- the automated measurement tool may determine the baseline measurement data such that the bias with respect to the reference data base is minimized and/or constant. Further, the determination of the baseline measurement data may be reproducible. That is, if the same input data (i.e. medical images) are inputted, the automated measurement tool outputs the same output (i.e. the baseline measurement data). In other words, the results determined by the automated measurement tool may be reproducible. Therefore, the automated measurement tool may be suited for a reliable baseline measurement.
- the automated measurement tool may use an artificial neural network for determining the baseline measurement data.
- artificial neural networks may be computing systems that "learn" to perform tasks by considering examples, generally without being programmed with task-specific rules.
- connectionist systems may be computing systems that "learn" to perform tasks by considering examples, generally without being programmed with task-specific rules.
- the artificial neural networks might learn to identify images that contain a specific content by analysing example images that have been manually labelled to contain the specific content and using the results to identify the specific content in other images. The artificial neural networks do this without any prior knowledge of the specific content.
- the artificial neural networks analyse the reference database.
- the reference database includes examples of standard measurements performed by experienced experts having a very low inter-observer bias.
- the artificial neural networks may perform the measurement on unknown (i.e. newly obtained medical images) having substantially the same bias as the experienced experts.
- the artificial neural networks may analyse all images of the plurality of medical images in order to generate a mean bias which is less susceptible to measurement errors that have occurred in a specific image.
- the artificial neural networks may be based on a collection of connected units or nodes called artificial neurons. Each connection, like the synapses in a biological brain, can transmit a signal to other neurons. An artificial neuron that receives a signal then processes it and can signal neurons connected to it.
- the mean value (i.e. the baseline measurement data) outputted by analysing the plurality of medical images of the same anatomical structure is less susceptible to measurement errors that have occurred in a specific image, that is, the baseline measurement data needs to be accurate only in the mean value, whereas a single outlier is innocuous.
- the measurement data may include at least one of a dimension such as a length, area or volume, or volume flow, velocity of blood and velocity of tissue.
- a dimension such as a length, area or volume, or volume flow, velocity of blood and velocity of tissue.
- the present invention may be applied to a wide field of measurement data that may be obtained from medical images.
- the measurement data may include a dimension of a fetus, such as at least one of crown-rump- length, nuchal translucency, biparietal diameter, femur length, humerus length, head circumference, abdominal circumference.
- the measurement data may include Doppler flow profiles, cardiac valve diameter and/or ventricle dimensions.
- a new measurement bias and an old measurement bias may be determined, wherein the new measurement bias may be determined when a predetermined period of time has elapsed since the old measurement bias was determined, or the new measurement bias may be determined when a predetermined amount of new measurement data are available since the old measurement bias was determined, and wherein the new measurement bias may be compared with the old measurement bias so as to obtain a drift in the measurement bias.
- the old measurement bias may be a measurement bias that is determined previous to the new measurement bias.
- the two measurement biases may be compared with each other, wherein each of the measurement biases may be separated from the other one either by a predetermine period of time (e.g. one to three months) or by a predetermined amount of measurement data (e.g. 10 to 100 measurement data, preferably 20 to 50) that are obtained since the previous measurement bias was determined.
- a predetermine period of time e.g. one to three months
- a predetermined amount of measurement data e.g. 10 to 100 measurement data, preferably 20 to 50
- an analysis of trends may be performed so as to determine in which way the measurement bias changes from the old measurement bias to the new measurement bias.
- a drift in relation to a previously defined consensus may be recognized. That is, the consensus may be defined at the same time as the old measurement bias was determined (e.g. at the beginning of the collection of data). This knowledge may be in particular interesting for core labs and for long-term studies.
- the invention provides a computer program comprising program code instructions, which, when executed by a processor, enables the processor to carry out the inventive method.
- the computer program may be in any code, in particular a code suited for computer vision applications, in particular for measurement purposes of features of anatomical structures in medical images.
- computer vision also referred as machine vision
- machine vision is a technology used to provide imaging-based automatic inspection and analysis. Therefore, this technology is in particular useful for use in the automated measurement tool.
- the invention is directed to a computer-readable medium comprising an above defined computer program.
- the computer-readable medium may be any digital data storage device, such as a USB-stick, hard disk, CD-ROM, SD card or SSD card.
- the computer program need not be stored on such a computer-readable medium to be supplied to the costumers, but may be downloaded via the internet.
- the invention is directed to an automated measurement tool configured to execute the method according to the above outlined, wherein the automated measurement tool includes an interface configured to receive a medical image generated by an imaging system and to output results determined by executing the inventive method. Any features or useful embodiments described in connection with the inventive method also apply to the automated measurement tool.
- the automated measurement tool may be included in an imaging modality for obtaining medical images and or in a measuring device for creating measurement data based on the medical images.
- the measurement data may be transferred to the interface of the automated measurement tool as an input to the automated measurement tool. The transfer may be established via cable, near field connection, Wi-Fi, W-Lan or other suitable connections.
- the automated measuring tool may be locally applied in locations where measurement data are obtained or received.
- the measurement tool may be connected to a database (such as a PACS) storing a plurality of measurement data.
- the measurement tool may be separately provided like a server connected to several clients (e.g. imaging modalities, measuring modalities, data storage, separate users, institutions, groups, etc.).
- the method steps may be executed by more than one automated measurement tool. That is, one automated measurement tool may determine the measurement bias and another automated measurement tool may update said measurement bias or may determine the global bias, for example. Further, the automated measurement tools may be present at different locations.
- the medical images may be internally transmitted from the imaging system to the automated measurement tool and the measurement bias may be calculated in situ within the system itself by the automated measurement tool.
- the measurement bias has to be transmitted (outputted from the automated measurement tool) for further processing. Therefore, the amount of data that is to be transmitted may be significantly reduced and it is not necessary to establish a high capacity transmission line for the purpose evaluating the measurement bias.
- the automated measurement toll may be applied on a workstation as well as in PACS (Picture Archiving and Communication System) systems.
- the imaging system may be an ultrasound system.
- Fig. 1 shows a schematic cross-section through a human heart (4-chamber view);
- Fig. 2 is a diagram schematically showing an automated measurement tool according to an embodiment of the present invention.
- Fig. 3 is another schematic diagram showing the measuring results of the user with respect to the baseline measurement data obtained by the automated measurement tool according to the embodiment of the present invention
- Fig. 4 is a flow chart that illustrates the steps of the method according to an embodiment of the present invention.
- Fig. 5 shows a system including the automated measurement tool according to an embodiment of the present invention.
- FIG. 1 illustrates the structure of the human heart as an example of the anatomical structure.
- the blood coming from the lungs flows into the left atrium 2 and from there through the mitral valve 3 into the left ventricle 4. From there, it is pumped through the aortic valve 5 into the aorta 6. This part is also termed left ventricular outflow tract (LVOT).
- LVOT left ventricular outflow tract
- the blood coming from the body flows into the right atrium 7 and is pumped through the tricuspid valve 8 into the right ventricle 9. From there, it is pumped through the pulmonary valve into the pulmonary artery 11.
- Heart wall 12 is made of muscular tissue surrounding the heart chambers 2, 4, 7 and 9.
- the heart has a complex shape, and in addition is constantly moving with the heartbeat, i.e. it is a dynamic anatomical structure.
- Such an anatomical structure specifically the human' s heart is recorded using medical imaging systems like echocardiography systems, ultra sound systems, MR, CT, etc. in order to attain medical images.
- the medical images are subsequently used to measure features of the anatomical structure, for example, during routine clinical processes and during clinical long-term studies. Due to the complexity of the anatomical structures, measuring of shapes, contours, distances and/or areas such as of the mitral valve 3 in order to analyse the feature of the anatomical structure is difficult and prone to individual errors.
- measurement data may be generated in clinics or other medical facilities.
- these facilities e.g. echocardiography labs
- these facilities have to undergo a standardization procedure (as outlined above in the introductory part).
- the inventive method allows to assess and to minimize intra-institutional bias between readers (i.e. users) in the same institution, e.g. a clinic or a core lab without additional workload.
- readers i.e. users
- the same institution e.g. a clinic or a core lab without additional workload.
- two users from the same core lab read studies independently, that is, measure the same anatomical structure of a first patient 15 and of a second patient 16 independently.
- the first user obtains the first measurement data RI and the second user obtains the second measurement data R2.
- the core lab now wants to improve quality in terms of inter-user variability and needs to document that as well. But the lab needs consistency with previous studies, thus an overall drift of measurements cannot be tolerated. Therefore, an automated measurement tool 1 (baseline machine) reads all studies in parallel.
- an automated pendant i.e. the first baseline measurement data Cl and the second baseline measurement data C2
- the baseline measurement data Cl, C2 is not displayed to the respective user.
- a first measurement bias B1 based on the first measurement data R1 and the first baseline measurement data Cl, and for the second user a second measurement bias B2 based on the second measurement data R2 and second baseline measurement data C2 is determined.
- the first and the second measurement bias Bl, B2 are determined based on the difference between the first and second measurement data Rl, R2 and the first and second baseline measurement data Cl, C2, respectively, as illustrated in Fig. 3.
- the relative biases Bl, B2 are determined based on a relative difference of the first and second measurement data Rl, R2 and of the first and second measurement baseline data Cl, C2. Furthermore, a per-user bias Bx is determined and continuously updated. In the present embodiment a measurement bias Bl for the first user and a measurement bias B2 for the second user.
- the mean bias between the two users can be calculated. This result can be used to give feedback to all involved users for both a routine clinical echocardiography scenario and a long-term clinical study scenario without the need to measure identical studies. Therefore, the use of the automated measurement tool 1 for determining the measurement bias for each user provides a location independent (i.e. may be used anywhere on earth) reference, similar to the “metre des Archives” with respect to the length measurement. For the above functionality, the medical image that is chosen by the users can be used to perform the analysis using the automated measurement tool lin order to minimize error sources between user and the automated measurement tool 1.
- an institution-wide bias Be between all users and the automated measurement tool l is determined and continuously updated.
- Monitoring the institution-wide bias Be over time ensures consistent reading behavior of the entire core lab.
- Monitoring the per-user bias Bx over time ensures consistent reading of each user x - or intended changes due to consensus training. Documentation of all these values can be used to document quality of reading, planning of training and assessing training effects.
- the determination of the automated measurement tool l is substantially executed in a hidden manner, that is, none of the users is aware of the determination being executed in the background.
- the automated measurement tool 1 could also be used to minimize bias between institutions.
- the automated measurement tool 1 does not necessarily need to be in line with the average way of reading a study in the core lab. That is, the institution wide bias Be might be large. In other words, the bias might be large without losing its usability as a reference between users and/or between groups of users.
- TOMTEC ® Arena could be one landing spot of such an invention, but IntelliSpace Cardiovascular (ISCV), which is an image and information management system, would represent a good option as well. Furthermore, such a system could also be integrated in ultrasound systems as communication of the values of the per-user bias Bx would be sufficient to operate it. That is, no transfer of image data would be required.
- ISCV IntelliSpace Cardiovascular
- Fig. 4 the process of the method for automated analysis of the bias in measurements performed on medical images of an anatomical structure will be described in more detail.
- a plurality of medical images is received.
- the plurality of medical images includes at least a first plurality of medical images and a second plurality of medical images, and each plurality of medical images includes a same anatomical structure of a same or different patient(s).
- the plurality of medical images is not received by the same automated measurement tool 1 but by two automated measurement tools 1. That is, the first plurality of medical images is received by the first automated measurement tool 1 and the second plurality of medical images is received by the second automated measurement tool 1. Further, the automated measurement tools 1 are located at different locations.
- step 92 for the first plurality of medical images, first measurement data R1 of a feature of the anatomical structure is received.
- the first measurement data R1 is obtained by a first user.
- step 94 for the first plurality of medical images, a first baseline measurement data Cl of the feature of the anatomical structure is determined by the automated measurement tool 1.
- step 96 a first measurement bias B1 for the first user based on the first measurement data R1 and the first baseline measurement data Cl is determined.
- step 98 for the second plurality of medical images, second measurement data R2 of a feature of the anatomical structure is received.
- the second measurement data R2 is obtained by a second user.
- a second baseline measurement data C2 of the feature of the anatomical structure is determined by the other automated measurement tool 1.
- a second measurement bias B2 for the second user based on the second measurement data R2 and the second baseline measurement data C2 is determined.
- step 104 the first measurement bias B1 is compared with the second measurement bias B2.
- the method additionally comprises the step 106 of determining an institution wide bias Be represented by a mean value of all measurement biases B 1, B2, ... , Bx of the first, second and further users belonging to one predetermined group or institution after the step 104.
- the method additionally comprises the step 108 of comparing the institution wide bias Be with one of the measurement biases Bl, B2,..., Bx so as to obtain a per-user bias relative to the institution wide bias Be after the step 106.
- the step 108 of comparing the institution wide bias Be with one of the measurement bias Bl, B2,..., Bx additionally includes: determining a difference dxl, dx2,..., dxx by subtracting each of the measurement biases Bl, B2,..., Bx by the institution wide bias Be so as to determine the difference dxl, dx2,.. ,dxx for each user, and determining the user or users having the highest difference(s) dxl, dx2,..., dxx among all users.
- a global bias may be determined represented by a mean value of all available measurement biases Bx. Subsequently, the global bias is compared with one of the measurement biases Bx.
- FIG. 5 schematically illustrates an imagining system which in this case is part of an ultrasound imaging machine or ultrasound scanner.
- the imagining system includes the automated measurement tool 1. Further, the imagining system includes a computer screen 20 having a touch-based display 50. An ultrasound image 40 is currently displayed on the display 50. The imagining system is able to control any images or elements displayed on the display 50.
- the automated measurement tool lincludes a processor 60, such as a CPU, which is configured to execute an embodiment of the inventive method.
- the imagining system further includes a data storage medium 80, such as a hard disc, connected to the automated measurement tool 1. Further, the automated measurement tool 1 includes a storage on which a computer program necessary for executing the invention on the automated measurement tool 1 or its processor 60 may be stored.
- the imagining system may include an ultrasound control unit 200, which controls the acquisition of ultrasound images by an ultrasound probe 120.
- a patient bed 14 is provided for a subject, e.g. a patient, reclining while ultrasound images 40 are being acquired by the user 17.
- the user 17 may be any person who wishes to perform an accurate image evaluation, which in the case of medical images will often be a radiologist or radiographer, but can also be a specialist in any other field, such as gynaecologist, a cardiologist etc.
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Public Health (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Biomedical Technology (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Pathology (AREA)
- Biophysics (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Heart & Thoracic Surgery (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Primary Health Care (AREA)
- Educational Administration (AREA)
- Epidemiology (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Bioethics (AREA)
- Physiology (AREA)
- Cardiology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP20776198.2A EP4042368A1 (en) | 2019-10-08 | 2020-09-30 | Computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure |
| JP2022521041A JP7649303B2 (ja) | 2019-10-08 | 2020-09-30 | 解剖学的構造物の医療画像において実施される測定におけるバイアスの自動分析のためのコンピュータにより実施される方法 |
| CN202080071041.3A CN114554964A (zh) | 2019-10-08 | 2020-09-30 | 计算机实施的用于自动分析在解剖结构的医学图像上执行的测量中的偏差的方法 |
| US17/765,874 US20220392648A1 (en) | 2019-10-08 | 2020-09-30 | Computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP19202004.8A EP3806031A1 (en) | 2019-10-08 | 2019-10-08 | Computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure |
| EP19202004.8 | 2019-10-08 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2021069271A1 true WO2021069271A1 (en) | 2021-04-15 |
Family
ID=68290152
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/EP2020/077301 Ceased WO2021069271A1 (en) | 2019-10-08 | 2020-09-30 | Computer implemented method for automated analysis of the bias in measurements performed on medical images of an anatomical structure |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US20220392648A1 (cg-RX-API-DMAC7.html) |
| EP (2) | EP3806031A1 (cg-RX-API-DMAC7.html) |
| JP (1) | JP7649303B2 (cg-RX-API-DMAC7.html) |
| CN (1) | CN114554964A (cg-RX-API-DMAC7.html) |
| WO (1) | WO2021069271A1 (cg-RX-API-DMAC7.html) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024238568A1 (en) * | 2023-05-15 | 2024-11-21 | The Regents Of The University Of California | Observer sequential bias mitigation in medical imaging |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140341449A1 (en) | 2011-09-23 | 2014-11-20 | Hamid Reza TIZHOOSH | Computer system and method for atlas-based consensual and consistent contouring of medical images |
| US20180150598A1 (en) | 2016-11-30 | 2018-05-31 | General Electric Company | Methods and systems for compliance accreditation for medical diagnostic imaging |
Family Cites Families (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6058322A (en) * | 1997-07-25 | 2000-05-02 | Arch Development Corporation | Methods for improving the accuracy in differential diagnosis on radiologic examinations |
| US7466848B2 (en) * | 2002-12-13 | 2008-12-16 | Rutgers, The State University Of New Jersey | Method and apparatus for automatically detecting breast lesions and tumors in images |
| DE10316298A1 (de) * | 2003-04-08 | 2004-11-04 | Mohr, Volker, Dr. | Verfahren und Anordnung zur automatischen Aufbereitung und Auswertung medizinischer Daten |
| US20070156032A1 (en) * | 2006-01-04 | 2007-07-05 | Gordon Linda S | Electronic disease management system |
| US7861159B2 (en) * | 2006-04-07 | 2010-12-28 | Pp Associates, Lp | Report generation with integrated quality management |
| US20080262870A1 (en) * | 2007-04-18 | 2008-10-23 | Siemens Medical Solutions Usa, Inc. | Computerized Treatment Order and Associated Alert Processing System |
| WO2009146703A2 (en) * | 2008-06-02 | 2009-12-10 | Aarhus Universitet | System and method for volumetric analysis of medical images |
| EP2332087B1 (en) * | 2008-07-25 | 2020-04-22 | Fundação D. Anna Sommer Champalimaud E Dr. Carlos Montez Champalimaud | Systems and methods of treating, diagnosing and predicting the occurrence of a medical condition |
| US20100198559A1 (en) * | 2009-02-03 | 2010-08-05 | United States Holdings, Llc | Real time monitoring and control of communications networks and radio frequency distribution networks |
| US20110161097A1 (en) * | 2009-12-28 | 2011-06-30 | General Electric Company | Methods and systems for scheduling appointments in healthcare environments |
| US8744041B2 (en) * | 2010-09-09 | 2014-06-03 | Hologic, Inc. | Methods and systems for dynamically modifying acquisition parameter during image acquisition |
| US20120173258A1 (en) * | 2011-01-03 | 2012-07-05 | Athenahealth, Inc. | Methods and apparatus for quality management of healthcare data |
| US10909674B2 (en) * | 2013-09-27 | 2021-02-02 | Koninklijke Philips N.V. | System and method for context-aware imaging |
| US10835213B2 (en) * | 2014-09-11 | 2020-11-17 | Koninklijke Philips N.V. | Quality metric for multi-beat echocardiographic acquisitions for immediate user feedback |
| US20160081663A1 (en) * | 2014-09-18 | 2016-03-24 | General Electric Company | Method and system for automated detection and measurement of a target structure |
| US20160260190A1 (en) * | 2015-03-02 | 2016-09-08 | Bill Schneider | Provider price and quality index |
| US20200383582A1 (en) * | 2016-05-11 | 2020-12-10 | Tyto Care Ltd. | Remote medical examination system and method |
| US20190051411A1 (en) * | 2017-08-11 | 2019-02-14 | Real Time Medical Systems, LLC | Decision making platform |
| EP3542747A1 (en) * | 2018-03-22 | 2019-09-25 | Koninklijke Philips N.V. | Visualization system for visualizing an alignment accuracy |
| US10726548B2 (en) * | 2018-06-25 | 2020-07-28 | Bay Labs, Inc. | Confidence determination in a medical imaging video clip measurement based upon video clip image quality |
| US10937155B2 (en) * | 2018-12-10 | 2021-03-02 | General Electric Company | Imaging system and method for generating a medical image |
| US11521716B2 (en) * | 2019-04-16 | 2022-12-06 | Covera Health, Inc. | Computer-implemented detection and statistical analysis of errors by healthcare providers |
| US11132793B2 (en) * | 2019-08-01 | 2021-09-28 | International Business Machines Corporation | Case-adaptive medical image quality assessment |
-
2019
- 2019-10-08 EP EP19202004.8A patent/EP3806031A1/en not_active Withdrawn
-
2020
- 2020-09-30 EP EP20776198.2A patent/EP4042368A1/en active Pending
- 2020-09-30 WO PCT/EP2020/077301 patent/WO2021069271A1/en not_active Ceased
- 2020-09-30 CN CN202080071041.3A patent/CN114554964A/zh active Pending
- 2020-09-30 JP JP2022521041A patent/JP7649303B2/ja active Active
- 2020-09-30 US US17/765,874 patent/US20220392648A1/en not_active Abandoned
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140341449A1 (en) | 2011-09-23 | 2014-11-20 | Hamid Reza TIZHOOSH | Computer system and method for atlas-based consensual and consistent contouring of medical images |
| US20180150598A1 (en) | 2016-11-30 | 2018-05-31 | General Electric Company | Methods and systems for compliance accreditation for medical diagnostic imaging |
Non-Patent Citations (9)
| Title |
|---|
| ANONYMOUS: "Man against machine: AI is better than dermatologists at diagnosing skin cancer", 28 May 2018 (2018-05-28), XP055667632, Retrieved from the Internet <URL:https://eurekalert.org/pub_releases/2018-05/esfm-mam052418.php> [retrieved on 20200212] * |
| CHRISTIAN KNACKSTEDT: "Fully Automated Versus Standard Tracking of Left Ventricular Ejection Fraction", JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY, vol. 66, no. 13, 2015, XP029277946, DOI: 10.1016/j.jacc.2015.07.052 |
| HAROLD L. KUNDE: "Measurement of Observer Agreement", RADIOLOGY, vol. 228, 2003, pages 303 - 308 |
| LANG: "Recommendations for Cardiac Chamber Quantification by Echocardiography in Adults: An Update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging", THE JOURNAL OF THE AMERICAN SOCIETY OF ECHOCARDIOGRAPHY, vol. 28 |
| NAGUEH: "Recommendations for the Evaluation of Left Ventricular Diastolic Function by Echocardiography: An Update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging", THE JOURNAL OF THE AMERICAN SOCIETY OF ECHOCARDIOGRAPHY, April 2016 (2016-04-01) |
| THE AMERICAN SOCIETY OF ECHOCARDIOGRAPHY RECOMMENDATIONS FOR CARDIAC CHAMBER QUANTIFICATION IN ADULTS: A QUICK REFERENCE GUIDE FROM THE ASE WORKFLOW AND LAB MANAGEMENT TASK FORCE |
| VIKRAM CHALANA: "A Methodology for Evaluation of Boundary Detection Algorithms on Medical Images", IEEE TRANSACTIONS ON MEDICAL IMAGING, vol. 16, no. 5, October 1997 (1997-10-01), XP011035651 |
| WOLFGANG WEYERS: "Screening for malignant melanoma-a critical assessment in historical perspective", DERMATOLOGY PRACTICAL & CONCEPTUAL, vol. 8, no. 2, 30 April 2018 (2018-04-30), pages 89 - 103, XP055742279, DOI: 10.5826/dpc.0802a06 * |
| ZORAN B. POPOVIC ET AL: "Assessing observer variability: a user's guide", CARDIOVASCULAR DIAGNOSIS AND THERAPY, vol. 7, no. 3, June 2017 (2017-06-01), pages 317 - 324, XP055647283, ISSN: 2223-3652, DOI: 10.21037/cdt.2017.03.12 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114554964A (zh) | 2022-05-27 |
| EP3806031A1 (en) | 2021-04-14 |
| EP4042368A1 (en) | 2022-08-17 |
| JP2022551144A (ja) | 2022-12-07 |
| US20220392648A1 (en) | 2022-12-08 |
| JP7649303B2 (ja) | 2025-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11576621B2 (en) | Plaque vulnerability assessment in medical imaging | |
| US10304569B2 (en) | Systems and methods for associating medical images with a patient | |
| US10402971B2 (en) | Matching patient images and images of an anatomical atlas | |
| CN112640000A (zh) | 根据医学数据对预测性数字孪生体模型的数据驱动的估计 | |
| JP2020080165A (ja) | シミュレーションの正確度および性能に対する画質評価 | |
| CN109727660B (zh) | 在医学成像中针对血液动力学量化的不确定性或敏感性的机器学习预测 | |
| US20160228190A1 (en) | Three-dementional quantitative heart hemodynamics in medical imaging | |
| US20250336548A1 (en) | Medical data processing device, medical data processing method, and non-transitory computer-readable recording medium | |
| US20220151500A1 (en) | Noninvasive quantitative flow mapping using a virtual catheter volume | |
| US20220076430A1 (en) | Heatmap and atlas | |
| Ilyasova et al. | Systems for recognition and intelligent analysis of biomedical images | |
| JP7649303B2 (ja) | 解剖学的構造物の医療画像において実施される測定におけるバイアスの自動分析のためのコンピュータにより実施される方法 | |
| JP7601703B2 (ja) | 医用画像処理装置、医用画像処理方法及びプログラム | |
| Santhiyakumari et al. | Medical decision-making system of ultrasound carotid artery intima–media thickness using neural networks | |
| CN117036302A (zh) | 主动脉瓣膜钙化程度的确定方法和系统 | |
| WO2023235653A1 (en) | Panatomic imaging derived 4d hemodynamics using deep learning | |
| Hasaballa et al. | Sensitivity of myocardial stiffness estimates to inter-observer variability in LV geometric modelling | |
| Sorriento et al. | A novel approach for remote ultrasonography of knee osteoarthritis | |
| Karamali et al. | Tricuspid valve geometry in patients with functional tricuspid regurgitation: A three‐dimensional echocardiographic study | |
| Nyns et al. | Evaluation of knowledge-based reconstruction for magnetic resonance volumetry of the right ventricle in tetralogy of Fallot | |
| Rahimi et al. | Trimodality image registration of ultrasound, cardiac computed tomography, and magnetic resonance imaging for transcatheter aortic valve implantation and replacement image guidance | |
| CN115482914B (zh) | 一种医疗影像数据处理方法,设备及存储介质 | |
| US20250349426A1 (en) | Hyper-personalized treatment based on coronary motion fields and big data | |
| Woznicki et al. | Monitoring Achilles tendon healing progress in ultrasound imaging with convolutional neural networks | |
| Fry et al. | Reproducibility Study Of FAI-Score For Measuring Coronary Inflammation In Clinical Practice |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20776198 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2022521041 Country of ref document: JP Kind code of ref document: A |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| ENP | Entry into the national phase |
Ref document number: 2020776198 Country of ref document: EP Effective date: 20220509 |