Fetal heart uterine contraction scanning session duration

Abstract

The proposed concepts are directed to providing schemes, solutions, concepts, designs, methods and systems relating to predicting the remaining duration of a fetal heart uterine contraction (CTG) scan session of a subject. In particular, CTG data and movement data of an object during a scan session are obtained. The data is then analyzed to identify artifacts in the CTG data associated with movement of the object. From the CTG data, the movement data, and the artifacts, a remaining duration of the CTG scan session may be predicted. In this manner, despite signal loss during the scan session, acquisition of a sufficient amount of CTG data during the scan session may be ensured.

Description

Technical Field

[0001] The present invention relates to predicting cardiotocography (CTG) scan session duration. Background Art

[0002] Typically, women with risky or complicated pregnancies (e.g., those who have had a preterm birth or who have pregnancy-induced hypertension) are monitored in hospitals, either inpatient or outpatient settings. A key aspect of this is the use of cardiotocography (CTG) fetal monitoring to monitor the well-being of the fetus.

[0003] CTG measurements are traditionally performed using a CTG fetal monitor, which has an ultrasound Doppler transducer for detecting fetal heart rate (fHR) and a tocodynamics meter for measuring uterine activity (i.e., maternal contractions). Recently, several electrophysiology-based CTG (eCTG) devices have been commercially released and are already in use by clinicians. eCTG has many advantages over ultrasound Doppler-based CTG, such as higher fHR resolution and simpler operation.

[0004] One of the challenges facing eCTG is that it is susceptible to noise, which means that the fHR may be masked. Specifically, the amplitude of the fHR signal is much lower than the amplitude of the maternal heart rate, where uterine activity and other maternal movements can also mask the fHR signal. Through signal processing techniques, such as blind source separation, the fHR signal is extracted and distinguished from other signals including maternal heart rate and uterine activity. Therefore, fHR signal loss is a common problem for CTG measurements. The fHR signal acquired from ultrasound Doppler-based CTG may also be suboptimal (e.g. due to improper transducer placement, fetal movement, etc.).

[0005] Health care professionals usually evaluate the signals acquired during a scanning session to determine whether the duration of the scanning session is sufficient to obtain data sufficient for interpretation. This duration may depend in particular on the number of fHR signal losses and the overall quality of the data, but also on the state of the fetus, since the fetus should not be in a deep sleep state in order to be able to draw clinical conclusions about the fetal health. However, this process is particularly labor-intensive and is not always accurate.

[0006] Furthermore, assessment of session duration is a particular problem for home-based CTG solutions, where an untrained user performs the scan and then transmits the acquired data to a healthcare professional. It is possible that no healthcare professional may be present (even remotely) to assess the data while it is being collected. Therefore, acquiring a sufficient amount of good quality fHR data may be a matter of luck. Summary of the invention

[0007] The present invention is defined by the claims.

[0008] The concepts presented are aimed at providing scenarios, solutions, concepts, designs, methods, and systems related to the remaining duration of a cardiotocogram (CTG) scan session of a subject. Specifically, CTG data and movement data of the subject during the scan session are obtained. The data is then analyzed to identify artifacts in the CTG data associated with the movement of the subject. Based on the CTG data, movement data, and artifacts, the remaining duration of the CTG scan session can be predicted. In this way, despite signal loss during the scan session, a sufficient amount of CTG data can be ensured to be collected during the scan session.

[0009] According to an example of an aspect of the present invention, a method for predicting the remaining duration of a cardiotocogram (CTG) scan session of a subject is provided, the method comprising:

[0010] Obtaining CTG data associated with the subject during the CTG scan session;

[0011] Obtaining movement data describing the movement of the subject during the CTG scan session;

[0012] Identifying artifacts in the CTG data associated with the movement of the subject based on the CTG data and the movement data; and

[0013] Analyzing the CTG data, the movement data, and the artifacts to predict the remaining duration of the CTG scan session.

[0014] In this way, the amount of time (i.e., the remaining duration) required to collect sufficient CTG data from the CTG scan session can be predicted. Sufficient data can be considered as sufficient data for which the CTG data can be used for diagnostic purposes. For example, there must be a sufficient amount of uninterrupted fetal heart rate (fHR) data such that a healthcare professional can correctly analyze the fHR.

[0015] Typically, the amount of remaining time required to complete the session is determined by healthcare professionals who view the CTG data that has already been collected and make a judgment about how much time is needed. This process is labor-intensive, may be inaccurate, and may not be available (e.g., in self-help / home CTG scan sessions). Of course, one solution would be to conduct the scan session much longer than required. However, this wastes the time of the pregnant patient, the time of the person performing the scan, and reduces the opportunities for other pregnant patients who need to have a CTG scan. Therefore, there is a need for an automatic and accurate method to predict the remaining duration of the scan session.

[0016] Therefore, we have realized that CTG data and movement data from the scan session can be utilized to determine the (required) remaining duration. In fact, by identifying artifacts in the CTG data (as well as movement and CTG data), the number and length of CTG signal losses can be detected. Thus, the availability / signal coverage of the CTG data can be determined, from which the (predicted) remaining duration can be derived.

[0017] Movement of the object (i.e., the pregnant patient) typically results in a degradation of the CTG data quality. This is usually because such movement increases the amount of noise detected by the CTG scanner / sensor, thus reducing the signal-to-noise ratio. In cases where the pregnant patient is highly active, it would be helpful to instruct the patient that they should remain at rest during the measurement to assist in improving the signal quality.

[0018] Fetal movement data is also relevant to determining the session duration, as it is a clinically important parameter and it can indicate whether the fetus is in a deep sleep state. Generally, for the CTG to be interpretable, the fetus should not be in a deep sleep state. If the fetus is in a deep sleep state (which can be determined based on the fHR pattern and the presence of fetal movement), then the CTG session should be extended to capture periods when the fetus is not in a deep sleep. Thus, analyzing the movement data of the pregnant woman and the fetus, as well as the CTG data and artifacts therein, enables prediction of the remaining duration required for sufficient signal coverage.

[0019] In some embodiments, the method may further include processing the CTG data to extract feature data indicative of the availability of the CTG data for diagnosis. In such a case, predicting the remaining duration is also based on the extracted feature data.

[0020] In fact, the key factor in determining how long the scan session needs to last is the quality of the CTG data that has already been acquired. Therefore, it is very important to evaluate the CTG data features indicative of the availability of such data in order to accurately determine how much longer the acquired CTG data needs to last to achieve satisfactory diagnostic availability.

[0021] In addition, the feature data may include at least one of fetal heart rate measurements, uterine activity measurements, and fetal wakefulness status. Specifically, the fetal heart rate measurements may include heart rate characteristics, including fetal heart rate variability, fetal heart rate baseline, fetal heart rate accelerations, and fetal heart rate decelerations.

[0022] All of the above factors may be useful for determining the usability of the obtained CTG data. The fHR measurement (including the features therein) is typically the most important signal for CTG acquisition, and thus evaluating this data is the most important factor on which the remaining session time depends. Uterine activity can also be extracted from the CTG data. The fetal sleep state may also have a positive impact on the usability / quality of the data (i.e., CTG scan sessions are typically performed when the fetus is not in a deep sleep state).

[0023] Therefore, by considering this feature data, a more accurate prediction of the remaining CTG scan session duration can be performed.

[0024] In some embodiments, the method may further include acquiring previous CTG data associated with the subject during at least one previous CTG scan session. Then, the remaining duration may also be predicted based on the historical CTG data.

[0025] In this way, the accuracy of the prediction of the remaining duration can be improved. In fact, various clues may be obtained from the subject's past CTG scan sessions. This may include the average overall scan session time, predictable patterns in the useful CTG data, stop periods in past acquisitions, etc. As a result, if the current CTG scan session appears similar (or different) to the subject's previous CTG scan sessions, the remaining duration can be predicted.

[0026] In addition, the method may further include obtaining physiological data describing one or more characteristics of the subject, and wherein predicting the remaining duration is also based on the physiological data. The physiological data may include at least one of age, gestational age, height, weight, medical condition, and medical history (including obstetric history).

[0027] Certain demographic characteristics of the subject may also provide information useful for predicting the remaining duration of the session. For example, a person with a lower gestational age may require a longer overall scan session duration because the complexity of interpreting fetal heart rate patterns is higher for a more immature fetus. Therefore, by considering the physiological data, a more accurate predicted remaining duration can be provided.

[0028] Specifically, the method may further include matching the subject with a plurality of similar subjects based on the physiological data of the subject and the physiological data of a plurality of other subjects, wherein each of the plurality of similar subjects is associated with historical CTG scan session duration data. Therefore, the predicted remaining duration may also be based on the historical CTG scan session duration data.

[0029] CTG data from the scan sessions of other subjects may also provide information useful for accurately predicting the remaining scan duration. In fact, subjects with similar physiological characteristics (such as age, gestational age, weight, etc.) may require approximately similar scan session durations. Therefore, by considering the scan session durations of other similar subjects, the CTG, motion, and artifact data can be put into context, resulting in a more accurate remaining duration for the scan session.

[0030] Matching the subject may include obtaining physiological data for each of a plurality of other subjects, clustering the plurality of other subjects based on the physiological data to classify similar subjects into one of a plurality of classes, and matching the subject to one of the plurality of classes based on the physiological data of the subject.

[0031] One way to identify subjects similar to the patient under discussion is to cluster similar subjects and then identify which group the patient is most similar to. Thus, the relevant CTG data of more other subjects can be analyzed, thereby improving the accuracy of the predicted remaining duration.

[0032] In some embodiments, the method may further include acquiring context information that describes the context of the CTG scan session. Predicting the remaining duration may also be based on the context information. Specifically, the context information may include at least one of the time of day, location, ambient temperature, and the device used to acquire the CTG data.

[0033] Again, context data (such as environmental data and time-related data) can also provide information about the useful context of the CTG data to be obtained and can indicate the overall remaining duration itself. For example, it may be known that a fetus of a subject is more active at night than in the morning, and thus, to obtain a sufficient amount of useful fHR data, the scan session must be longer in the morning than at night. In this way, by considering the context data, a more accurate remaining duration can be predicted.

[0034] In a particular embodiment, analyzing the CTG data, motion data, and artifacts may include providing the CTG data, motion data, and artifacts to a machine learning algorithm that is trained to predict the remaining duration of the CTG scan session and obtaining the predicted remaining duration from the machine learning algorithm.

[0035] A training algorithm configured to receive training inputs and corresponding known outputs can be used to train the machine learning algorithm, where the training inputs include historical CTG data of the subject's previous CTG scan sessions, and where the corresponding known outputs include the historical CTG scan durations of the subject's previous CTG scan sessions.

[0036] Thus, machine learning algorithms may be able to identify patterns in the acquired data that are not typically detectable by standard models and algorithms. Thus, using this tool, the remaining duration of the scan session can be predicted more accurately.

[0037] In some embodiments, CTG data may be obtained from at least one of an ultrasound Doppler-based CTG sensor or an electrophysiology-based CTG sensor.

[0038] Additionally, movement data may be obtained from at least one of an accelerometer and a muscle electromyogram (EMG) sensor.

[0039] In some embodiments, the method may further include determining the probability of completing the session within at least one of a predetermined number of times based on the predicted remaining duration.

[0040] For example, the probability of completing the session within 5 minutes, 10 minutes, 15 minutes, etc. may be determined based on the predicted remaining duration. This information may be more useful to the object being scanned to understand how long the scan session may take. In fact, providing the probability (instead of just the estimated end time) may be more useful for planning purposes.

[0041] In other embodiments, the method may further include generating a recommendation based on the analysis that describes the actions to be taken to reduce the remaining duration.

[0042] In this way, the overall remaining session duration required to acquire sufficient CTG data can be shortened. For example, by recommending that the object remain still in response to detecting many movement-related artifacts, the quality of the acquired CTG data can be improved. This may be particularly useful when an untrained individual performs a CTG scan session at home. In such cases, the knowledge of a healthcare professional may not be available, so any advice that can be provided may help improve the scan session.

[0043] According to another example of an aspect of the present invention, there is provided a computer program comprising computer program code units that, when the computer program is run on a computer, are adapted to implement a method for predicting the remaining duration of a CTG scan session.

[0044] According to additional examples of aspects of the present invention, there is provided a system for predicting the remaining duration of a cardiotocogram (CTG) scan session, the system comprising: an interface configured to obtain CTG data associated with an object during a CTG scan session and movement data describing the movement of the object during the CTG scan session; and a processor configured to: identify artifacts in the CTG data associated with the movement of the object based on the CTG data and the movement data; analyze the CTG data, the movement data, and the artifacts to predict the remaining duration of the CTG scan session.

[0045] These and other aspects of the present invention will be apparent and will be elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] For a better understanding of the present invention, and to more clearly show how it may be put into practice, reference will now be made, by way of example only, to the accompanying drawings, in which

[0047] Figure 1 a flowchart of a method for predicting the remaining duration of a cardiotocogram (CTG) scan session of an object according to an embodiment is presented;

[0048] Figure 2 a flowchart of a method for matching an object with other similar objects according to an aspect of an embodiment of the present invention is presented;

[0049] Figure 3 a system for predicting the remaining duration of a CTG scan session of an object according to another embodiment of the present invention is presented; and

[0050] Figure 4 is a simplified block diagram of a computer in which one or more parts of the embodiment may be employed. DETAILED DESCRIPTION

[0051] The present invention will be described with reference to the accompanying drawings.

[0052] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the devices, systems, and methods, are intended for illustrative purposes only and are not intended to limit the scope of the present invention. These and other features, aspects, and advantages of the devices, systems, and methods of the present invention will be better understood from the following description, the appended claims, and the accompanying drawings. Although specific measures are recited in mutually different dependent claims, this does not indicate that the combination of these measures cannot be used advantageously.

[0053] It should also be understood that the drawings are merely schematic and are not drawn to scale. It should also be understood that throughout the drawings, the same reference numerals are used to denote the same or similar parts.

[0054] The present invention presents the concept of predicting the remaining duration of a cardiotocogram (CTG) scan session of a subject. Specifically, it is proposed to use CTG data and movement data from the scan session to identify artifacts in the CTG data, and then use the CTG, movement, and artifact data to generate a predicted remaining duration of the CTG scan session. In this way, the total remaining time can be automatically evaluated without the need for a skilled healthcare professional, enabling an accurate assessment of the remaining duration required for a CTG scan to acquire sufficient data for diagnostic purposes. In fact, these concepts may be particularly useful for remote (i.e., at-home) CTG scan sessions.

[0055] By way of explanation, depending on the progress of the CTG measurement session (i.e., the percentage of signal coverage), taking into account signal loss, the remaining duration of the CTG scan session may vary. In other words, due to different levels of signal loss and signal coverage, the total duration of the scan session required for sufficient signal acquisition will vary. Currently, there is no tool available to predict / estimate the total duration of the measurements / scan session required to generate sufficient available data for diagnostic decision-making during the progress of the scan session. This estimation is currently performed by a skilled healthcare professional conducting the CTG scan session.

[0056] Furthermore, it is foreseeable that future developments in electrophysiology-based CTG (eCTG) will improve remote (e.g., at-home) fetal monitoring solutions. When performing CTG at home, the patient or other untrained users will perform the measurements using a sensing device. For these non-professional users, it is even more important to determine the total (remaining) duration of the measurement session that will generate data available for diagnostic decision-making.

[0057] Therefore, embodiments of the proposed invention aim to improve the comfort of users conducting CTG scan sessions by providing them with an automatic estimate of the CTG measurement session duration / remaining duration.

[0058] Although eCTG sensing has a higher demand for such solutions, the solution is also relevant for traditional ultrasound Doppler-based CTG.

[0059] More specifically, CTG monitoring devices based on electrophysiology CTG (eCTG) have recently been provided. These eCTG solutions can provide higher-resolution fetal heart rate information. Specifically, eCTG can provide peak-to-peak fHR, while ultrasound Doppler detection may simply provide the average fHR of several heartbeats. This may enable the derivation of fHR variability and even fetal electrocardiogram waveforms (i.e., clinically relevant information that is currently not obtainable from ultrasound Doppler detection).

[0060] Furthermore, it has been shown that eCTG is more effective for people with a higher BMI and is also more comfortable for pregnant women. It has been shown that performing eCTG is less burdensome for healthcare professionals as it does not require frequent repositioning as in the case of ultrasound Doppler-based CTG fetal monitors.

[0061] Frequent hospital visits or hospital stays for monitoring (including CTG) are a significant burden for pregnant women. One way to reduce this burden is to use mobile hospital midwives to perform CTG scan sessions at the patient's home. However, this is particularly labor-intensive. Therefore, remote monitoring solutions have been developed to perform measurements by the patient themselves at home and transmit the data results to healthcare professionals for review.

[0062] Since eCTG technology is new, remote fetal monitoring based on eCTG technology faces new challenges.

[0063] Specifically, the way eCTG is measured is different from ultrasound Doppler CTG. Generally, a sensing electrocardiogram patch with 5 - 6 adhesive electrodes is attached to the patient's abdomen. When the impedance indicators of all electrodes show positive feedback on the fetal monitor, the measurement is performed by leaving the patch on the patient for 30 to 60 minutes. In contrast, ultrasound Doppler-based CTG solutions typically involve using two circular transducers, one with a piezoelectric element for measuring the fetal heart rate (fHR) and the other with a strain gauge for measuring the pregnant woman's uterine contractions. These sensors are placed on the abdomen of the pregnant patient to locate the fetal heart using auditory feedback. The transducers are then fixed in place by wrapping and tightening an elastic band around the abdomen.

[0064] In the case of ultrasound Doppler CTG, if the fHR signal is lost (e.g., due to the fetus changing position), the fetal monitor produces an audible alarm. This notifies the healthcare professional performing the scan to reposition the transducer to continue the measurement. In standard practice, the healthcare professional will evaluate the measurements of the fHR and UA collected (i.e., via a printout), consider the number of fHR losses and repositioning required, and determine the additional measurement time needed.

[0065] In contrast, when using eCTG to acquire fHR signals, the electrodes of the device pick up electrical signals. This signal corresponds to a mixture of different bioelectrical signals generated by the beating of the fetal heart, the beating of the pregnant woman's heart, uterine contractions, and other physiological processes (muscle activity and movement artifacts). Individual signals related to the maternal heart rate (mHR), fHR, and UA are extracted by processing the raw composite signal (i.e., applying blind source separation techniques). Extracting fHR is particularly difficult because the fHR signal amplitude is significantly (about 5 times) smaller than mHR, which means a relatively poor signal-to-noise ratio. Specifically, at times when the noise level is relatively high (i.e., due to movement of the pregnant patient, high levels of UA, etc.), the fHR signal may not be accurately detected. In other words, signal loss for fHR may be common. Therefore, detecting fHR using eCTG during the second stage of labor is usually very challenging. When this occurs and there is no simple remedy (such as reconnecting loose electrodes), the only option is to wait until fHR detection resumes.

[0066] Accordingly, it is desirable to provide a method for automatically determining / predicting / estimating the remaining duration of a desired scan session so that sufficient data can be acquired during the scan session for diagnostic purposes.

[0067] According to a particular exemplary embodiment, the following elements may be provided:

[0068] (i) A sensing device for signal acquisition, such as a CTG transducer or an eCTG sensing patch;

[0069] (ii) Additional sensors, such as an accelerometer or an additional ECG sensor. These can be used to detect maternal movement, which may be a source of movement artifacts, by acceleration or electromyogram (EMG) of muscles;

[0070] (iii) A processing unit for:

[0071] (a) Extracting different features related to the signals of interest. The signals of interest may include fHR, mHR, UA acceleration, and muscle contractions. Features of such signals include the duration of uninterrupted fHR measurements, the presence or absence of UA signals, etc.;

[0072] (b) Calculating the predicted total (remaining) duration of the CTG scan session based on an algorithm / model. This may take into account the fHR signal, mHR signal, UA signal, additional sensor signals (i.e., maternal movement signals), and the extracted features;

[0073] (iv) A storage device for storing model instructions and ECG, movement, and other data related to the current scan session, as well as other relevant data (i.e., context information, physiological data, and previous ECG data);

[0074] (v) A user interface for displaying the predicted total (remaining) duration of a measurement session. The user interface can also provide the user with measurement progress (i.e., percentage or progress bar), estimated end time, and other forms of feedback. The user interface can also provide information advising a pregnant patient to remain still to prevent excessive fHR interruptions due to their movement;

[0075] (vi) A method / system for a healthcare professional (reviewing CTG data from a remote location) to manually adjust the predicted remaining duration of a scan session;

[0076] (vii) A feedback interface for collecting user feedback, such as information regarding the estimation accuracy.

[0077] At least some of the above system components can be part of a fetal monitor. Thus, a processing unit for data analysis can be implemented on the CPU of the fetal monitor. Alternatively, at least some of the system components can be provided on a personal device, such as a tablet, which has both a screen for the user interface and a CPU for data processing.

[0078] In addition, embodiments of the present invention can provide for detecting uterine activity (UA). Contractions (detected based on uterine activity) provide additional useful data and can help analyze how the fetus responds to the stress of contractions by analyzing fHR decelerations relative to the start and end of the contractions. In addition, contraction data is a useful measurement metric for patients at risk of preterm birth. The model can be programmed to extend the measurement session duration in the event of detected UA. For example, after detecting UA, the measurement time can be extended by an additional 20 minutes.

[0079] In another embodiment, the predicted remaining duration may depend on the sleep - wake state of the fetus, as CTG can generally be interpreted when the fetus is not in a deep sleep state. Typically, this determination is made by a healthcare professional who asks the patient if they feel the fetus move and how frequently the fetus moves. In the case of the present invention, this can be detected by means of fetal movement detection, in the form of movement data.

[0080] In a further embodiment, after accumulating a large amount of data from different subjects, the subjects can be clustered to identify groups / classes of similar subjects based on their physiological characteristics (i.e., demographics, obstetric history, etc.). Then, the average scan session time for the corresponding cluster with the highest similarity to the scanned subject can be used to determine the predicted remaining duration of the measurement session.

[0081] In additional embodiments, previously recorded CTG data of the object being scanned can be used. The machine learning model can combine this data with the extracted features to predict the total / remaining scan session duration. Additionally, feedback from the user (i.e., healthcare professional or patient) regarding the estimation accuracy can be used to train and improve the duration prediction.

[0082] Go to Figure 1 , where a flowchart of a method for predicting the remaining duration of a CTG scan session of an object is provided. In other words, the method is used to predict the amount of remaining time of a CTG scan session required to collect sufficient CTG data of the object.

[0083] It should be clarified that the object herein may refer to a pregnant object, the fetus of a pregnant object, or both. In fact, a CTG scan session must be performed on a pregnant object, but is typically aimed at measuring the fHR.

[0084] The prediction of the remaining duration can be made only once, continuously from the start to the end of the scan session, or selectively (i.e., when a CTG signal loss is detected, or at a predetermined interval). In fact, the remaining duration can be predicted by predicting the total required duration and subtracting the time that has already been executed. The predicted remaining duration can also be performed by adjusting (i.e., increasing / decreasing) the target end time, based on the following.

[0085] In step 110, CTG data associated with the object during the CTG scan session is obtained. The CTG data can be obtained / acquired from an ultrasound Doppler-based CTG sensor and / or an electrophysiology-based CTG sensor. In fact, any device suitable for performing CTG can be used.

[0086] The CTG data may include an fHR signal, a maternal heart rate (mHR) signal, and a uterine activity (UA) signal. Such signals may not be completely different, but can be separated by techniques known to those skilled in the art. Additionally, due to the nature of the CTG scan and the use of such devices, noise may also be present in the CTG data. However, the noise can be reduced through preprocessing, such as via an appropriate filter, to eliminate / suppress / minimize this.

[0087] In step 112, movement data describing the movement of the object during the CTG scan session is obtained. Specifically, the movement data may be related to the pregnant object itself (i.e., repositioning, accidental vibrations, etc.) and / or may be related to the movement of the fetus.

[0088] The movement data can be obtained / acquired from an accelerometer and / or a muscle electromyogram (EMG) sensor. In fact, those skilled in the art should understand that any sensor capable of detecting the movement of one or both of the pregnant patient and the fetus can be used.

[0089] In step 120, artifacts in the CTG data associated with the movement of the object are identified. The identification is based on the CTG data and the movement data.

[0090] In other words, the CTG data is analyzed in the context of the movement data in order to identify losses in the CTG signal associated with the movement data. Such losses mean that the quality of the CTG data may degrade at these times, rendering the CTG data potentially useless (or less useful) for diagnostic purposes. Thus, identifying such artifacts is important for accurately adjusting the predicted remaining duration of the scan session.

[0091] More specifically, the movement of the fetus is important as an indicator of whether the fetus is in a deep sleep state, which may be clinically relevant information for interpreting the CTG data to identify artifacts. Maternal movement is relevant because if the maternal object moves a lot, it becomes difficult to extract the fHR signal from the CTG data due to the presence of noise artifacts. In such cases, it may be appropriate to instruct the object to reduce movement during the measurement and potentially increase the predicted remaining duration to account for such movement. Thus, by identifying artifacts in the CTG data, the movement of the pregnant patient and fetus during the scan session may be useful for predicting the total remaining duration required for the scan session.

[0092] In step 130, the CTG data, the movement data, and the artifacts are analyzed to predict the remaining duration of the CTG scan session. The data can be analyzed using an algorithm configured to predict the remaining duration in response to the input data. For example, the algorithm can be a statistical model (i.e., a Bayesian model) or a machine learning-based model.

[0093] In fact, it has been outlined previously how CTG, movement, and artifact data can contribute to indicating the amount of time remaining in the scan session to ensure that sufficient data is available for diagnosis. For example, the predetermined / average time can be adjusted based on the analysis of the CTG, movement, and artifact data in order to determine the remaining scan session duration. In fact, if there are many artifacts in the CTG data, then the remaining scan session time may be increased.

[0094] In some cases, analyzing the CTG data, the movement data, and the artifacts includes providing the CTG data, the movement data, and the artifacts to a machine learning algorithm that is trained to predict the remaining duration of the CTG scan session and obtaining the predicted remaining duration from the machine learning algorithm. The machine learning algorithm can be any machine learning, artificial intelligence, and / or neural network-based algorithm that can learn from previous data to establish patterns in unknown data.

[0095] Specifically, a machine learning algorithm can be trained using a training algorithm configured to receive a training input array and a corresponding known output, where the training input includes historical CTG data from a subject's previous CTG scan sessions, and where the corresponding known output includes the historical CTG scan duration of the subject's previous CTG scan sessions.

[0096] Thus, historical data can be utilized to establish context for the obtained CTG and movement data, thereby providing an accurate predicted remaining duration. A specific example method of analyzing CTG data is as follows. To utilize CTG information to predict the remaining duration of a CTG scan session, the signal loss recorded since the start of the session can be determined. For CTG interpretation, according to agreed-upon guidelines, a signal loss of less than 20% (i.e., the time when the signal is lost is less than 20%) is considered acceptable. If the signal loss starts to exceed this threshold from the start of the scan session, the predicted remaining duration of the scan session may increase (which typically lasts ~30 - 60 minutes). Thus, this can compensate for excessive signal loss and meet the interpretability criteria.

[0097] The current signal loss percentage is x (i.e., the percentage of the signal that is lost in the obtained signal), and the additional percentage of the total scan duration required to acquire sufficient CTG data is y. Given this, and when x exceeds 20%, the following holds:

[0098]

[0099] Of course, this can also be written as:

[0100] y = 5x - 100

[0101] Assuming no additional signal loss during the remaining scan session time, this calculation will yield the total measurement time (when applied to the initial total scan duration). If further signal loss is detected (denoted by z), the additional scan session duration will increase according to the following formula:

[0102] y = 5(x + z) - 100

[0103] In addition, movement data describing the movement of the pregnant woman (i.e., signals from an accelerometer such as a sensor) is also used. In fact, the movement of the pregnant woman may cause the signal loss percentage to be greater than 20%. In this case, suggestions can be given to the pregnant patient, and the suggestions to remain as stationary as possible to improve signal coverage can be displayed on the user interface.

[0104] Optionally, the method may include one or more steps (steps 114, 116, 118, 122, 140, and 142) provided within the dashed outline.

[0105] In step 122, the CTG data is processed to extract feature data indicative of the usability of the CTG data for diagnosis. In this case, the prediction of the remaining duration is also based on the extracted feature data. In other words, the CTG data is processed by identifying / extracting / determining feature data from the CTG data, thereby determining the usability of the CTG data for diagnosis.

[0106] The feature data can include any data useful for the diagnosis object. For example, the feature data includes at least one of fetal heart rate measurement results, uterine activity measurement results, and fetal sleep state. More specifically, the fetal heart rate measurement results can include heart rate characteristics, including fetal heart rate variability, fetal heart rate baseline, fetal heart rate acceleration, and fetal heart rate deceleration.

[0107] Therefore, if it is determined that fHR variability, fHR baseline, and fHR acceleration and deceleration are derived from the CTG data, it may be determined that only a small amount of session duration remains. In fact, if the feature data is scarce (i.e., there is not much data in the CTG data that may be useful for diagnosis), then the remaining scan session time / duration may increase.

[0108] In step 114, previous CTG data related to / measured from the object during at least one previous CTG scan session is obtained. Then, the prediction of the remaining duration is also based on the previous CTG data.

[0109] In this way, feedback from previous / past / historical scan sessions of the object can be considered. At least one previous scan session may include not only multiple scan sessions of the pregnant object during the current pregnancy, but also previous pregnancies. The CTG data of these previous scan sessions can be collected from a database of previous CTG scan sessions. The CTG data may also include metadata, which contains information about the context of the scan (i.e., time of day, gestational age, acquisition device, etc.), which can provide useful information for evaluating the relevance of the CTG data.

[0110] In step 116, physiological data describing one or more characteristics of the object is obtained. The prediction of the remaining duration is then also based on the physiological data. Physiological data is any data that describes the physical characteristics of the object (i.e., demographic data, past medical conditions, etc.), and each data can indicate the amount of remaining time required for satisfactory CTG data acquisition.

[0111] The physiological data can be obtained from the user or from one or more databases storing information about the pregnant object.

[0112] For example, the physiological data includes at least one of age, gestational age, height, weight, medical condition, and medical history. However, those skilled in the art will understand that additional physiological information that can provide information about the possible quality of the CTG data, and thus provide information about the predicted remaining duration of the scan session, can be provided.

[0113] In addition, the use of the obtained physiological data will be explained in more detail below with reference to Figure 2 more detail.

[0114] In step 118, context information that describes the context of the CTG scan session is obtained. Thus, the prediction of the remaining duration is also based on the context information. Any context information (i.e., data about the surrounding environment and the conditions of the scan session) can provide clues about the length of the scan session required for satisfactory CTG data acquisition.

[0115] Such data can be manually input by the user or collected from additional sensors.

[0116] For example, the context information includes at least one of the time of day, location, device used to acquire the CTG data, and ambient temperature. However, those skilled in the art will understand that additional context data that can provide information about the possible quality of the CTG data, and thus provide information about the predicted remaining duration of the scan session, can be provided.

[0117] Of course, the information obtained in steps 114 - 118 can also be utilized by a machine learning algorithm (and historical data for training the machine learning algorithm), which is configured to predict the remaining duration of the scan session (i.e., similar to the CTG, movement, and artifact data described above). Additionally and / or alternatively, the data can be used to identify artifacts in the CTG data and can also be used to extract the feature data as described above.

[0118] Finally, steps 140 and 142 present the following steps, where the predicted remaining duration and the analysis of the above - mentioned data of the scan session are used to generate an output useful for the object, the user, and / or the healthcare professional.

[0119] In step 140, the probability of completing the session within at least one of a predetermined number of times is determined based on the predicted remaining duration. For example, the probability of completing the session within 5 minutes, 10 minutes, 15 minutes, etc. can be determined according to the predicted remaining duration.

[0120] In step 142, based on the analysis of the CTG, movement, and artifact data, recommendations for actions to be taken to reduce the remaining duration are generated. For example, it can be recommended to remain stationary during the examination, perform the scan session at an alternative time, or re - position the CTG sensor.

[0121] In fact, the data generated in steps 140 and 142 can subsequently be output to the user. Alternative and / or additional outputs can also be considered, including simply displaying the predicted remaining duration.

[0122] Figure 2 Presents a flowchart of a method for matching an object with other similar objects based on physiological data obtained in step 116 of the method described with respect to Figure 1 the description.

[0123] In fact, matching an object with multiple similar objects can be based on the physiological data of the object and the physiological data of multiple other objects, wherein each of the multiple similar objects is associated with historical CTG scan session duration data. In this case, the predicted remaining duration is also based on the historical CTG scan session duration data.

[0124] To give a simple example, when the CTG scan session duration data of multiple other (similar) objects is known, then this duration data can be used to provide information for predicting the remaining session duration. Similar objects may often experience roughly similar scan session durations, and taking this into account may help provide an accurate prediction.

[0125] Specifically, steps 210 - 240 provide a method for matching an object with multiple other objects.

[0126] At 210, physiological data of the object is obtained. This is the same as step 116 described with reference to Figure 1 and thus, for the sake of brevity, further explanation is omitted here.

[0127] At step 220, physiological data of each of the multiple other objects is obtained. This can be obtained from a large database of other objects and their associated CTG data. The physiological data of the other objects may relate to the same or similar characteristics as those described by the physiological data of the object (i.e., relating to age, gestational age, weight, etc.).

[0128] At step 230, the multiple other objects are clustered based on the physiological data such that similar objects are classified into one of multiple categories. Clustering can be performed by any method known to those skilled in the art, clustering the objects based on values related to the same parameters.

[0129] At step 240, the object is matched to one of the multiple categories based on the physiological data of the object. This means that the category of the object with the most similar physiological characteristics is determined, and the data of the objects in that category can be used to predict the remaining duration of the scan session.

[0130] Figure 3 Disclosed is a system 300 for predicting the remaining duration of a CTG scan session of an object. Specifically, the system includes an interface 310 and a processor 310, and may optionally further include an output display 330. Of course, each of the interface 310, the processor 320, and the output display 330 may include sub-components for handling various functions of each module. The interface 310, the processor 320, and the output display 330 may be provided by an integrated device (i.e., a smartphone / tablet).

[0131] First, the interface 310 is at least configured to collect CTG data associated with the object during the CTG scan session, as well as movement data describing the movement of the object during the CTG scan session. Depending on the embodiment, the interface 310 may also obtain physiological data of the object, physiological data of other objects, context data of the scan session, and previous CTG data of the object's previous scan sessions. The interface 310 may store such data locally, or may collect such information from a memory / database of one or more such information sources.

[0132] In addition, the interface 310 may subsequently preprocess the information, or may simply pass the information to the processor.

[0133] The processor 320 is configured to: identify artifacts in the CTG data related to the movement of the object based on the CTG data and the movement data; in fact, the processor 320 may also consider the physiological data of the object, the physiological data of other objects, the context data of the scan session, and the previous CTG data of the object's previous scan sessions to identify artifacts.

[0134] In addition, the processor 320 is configured to analyze the CTG data, the movement data, and the artifacts to predict the remaining duration of the CTG scan session. Moreover, the processor 320 may additionally consider the physiological data of the object, the physiological data of other objects, the context data of the scan session, and the previous CTG data of the object's previous scan sessions to identify artifacts.

[0135] Then, the processor 320 may generate a probability of completing the session within at least one of a predetermined number of times determined based on the predicted remaining duration. In addition, the processor 320 may generate a recommendation based on the analysis of the CTG, movement, and artifact data (and other data, if obtained), the recommendation describing the actions to be taken to reduce the generated remaining duration.

[0136] The predicted remaining duration of a scan session can be provided to an output display 330, which can provide this (and other generated data) to a user (i.e., a pregnant patient, a healthcare professional, and / or a person performing the scan session).

[0137] Figure 4 An example of a computer 1000 in which one or more portions of the embodiments may be employed is illustrated. The various operations discussed above can utilize the functionality of the computer 1000. For example, one or more portions of a system for predicting the remaining duration of a CTG scan session according to another embodiment of the present invention can be incorporated into any of the elements, modules, applications, and / or components discussed herein. In this regard, it should be understood that the system functional blocks can run on a single computer or can be distributed across multiple computers and locations (e.g., via an Internet connection).

[0138] The computer 1000 includes, but is not limited to, a PC, a workstation, a laptop computer, a PDA, a handheld device, a server, a memory, and the like. Generally, in terms of its hardware architecture, the computer 1000 can include one or more processors 1010, a memory 1020, and one or more I / O devices 1030 communicatively coupled via a local interface (not shown). The local interface can be, for example but not limited to, one or more buses or other wired or wireless connections, as known in the art. The local interface can have additional elements such as controllers, caches, drivers, repeaters, and receivers to enable communication. In addition, the local interface can include addressing, control, and / or data connections to enable proper communication between the above components.

[0139] The processor 1010 is a hardware device for running software that can be stored in the memory 1020. The processor 1010 can actually be any custom or commercial processor, a central processing unit (CPU), a digital signal processor (DSP), or an auxiliary processor associated with the computer 1000, and the processor 1010 can be a semiconductor-based microprocessor (in the form of a microchip) or a microprocessor.

[0140] The memory 1020 may include any one or a combination of volatile memory elements (e.g., random access memory (RAM), such as dynamic random access memory (DRAM), static random access memory (SRAM), etc.) and non-volatile memory elements (e.g., ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic tape, compact disc read-only memory (CD-ROM), magnetic disk, floppy disk, cartridge, cassette tape, etc.). In addition, the memory 1020 may comprise electronic, magnetic, optical, and / or other types of storage media. Note that the memory 1020 may have a distributed architecture, where various components are located far from each other, but can be accessed by the processor 1010.

[0141] The software in the memory 1020 may include one or more separate programs, each program including an ordered list of executable instructions for implementing logical functions. According to an exemplary embodiment, the software in the memory 1020 includes a suitable operating system (O / S) 1050, a compiler 1060, source code 1070, and one or more application programs 1080. As shown, the application program 1080 includes a plurality of functional components for implementing the features and operations of the exemplary embodiment. According to an exemplary embodiment, the application program 1080 of the computer 1000 may represent various application programs, computing units, logics, functional units, processes, operations, virtual entities, and / or modules, but the application program 1080 is not intended to be limiting.

[0142] The operating system 1050 controls the execution of other computer programs and provides scheduling, input-output control, file and data management, memory management, and communication control and related services. The inventors anticipate that the application program 1080 for implementing the exemplary embodiment may be applicable to all commercially available operating systems.

[0143] The application program 1080 may be a source program, an executable program (object code), a script, or any other entity including a set of instructions to be executed. When it is a source program, the program is typically translated by a compiler (such as the compiler 1060), an assembler, an interpreter, etc., which may or may not be included in the memory 1020, in order to operate properly with the operating O / S 1050. In addition, the application program 1080 may be written in an object-oriented programming language having classes of data and methods, or a procedural programming language having routines, subroutines, and / or functions, such as but not limited to C, C++, C#, Pascal, BASIC, API calls, HTML, XHTML, XML, ASP scripts, JavaScript, FORTRAN, COBOL, Perl, Java, ADA,.NET, etc.

[0144] The I / O device 1030 may include input devices such as, but not limited to, a mouse, a keyboard, a scanner, a microphone, a camera, etc. In addition, the I / O device 1030 may also include output devices such as, but not limited to, a printer, a display, etc. Finally, the I / O device 1030 may also include devices that transfer both input and output, such as, but not limited to, a NIC or a modem / demodulator (for accessing remote devices, other files, devices, systems, or networks), a radio frequency (RF) or other transceiver, a telephone interface, a bridge, a router, etc. The I / O device 1030 also includes components for communicating over various networks such as the Internet or an intranet.

[0145] If the computer 1000 is a PC, a workstation, a smart device, etc., the software in the memory 1020 may also include a basic input / output system (BIOS) (omitted for simplicity). The BIOS is a set of basic software routines that initialize and test the hardware at startup, start the O / S 1050, and support data transfer between hardware devices. The BIOS is stored in some kind of read-only memory such as ROM, PROM, EPROM, EEPROM, etc., so as to run the BIOS when the computer 800 starts up.

[0146] When the computer 1000 is in operation, the processor 1010 is configured to run the software stored in the memory 1020 to transfer data to and from the memory 1020 and generally control the operation of the computer 1000 according to the software. The application program 1080 and the O / S 1050 are all or partially read by the processor 1010, may be buffered within the processor 1010, and then run.

[0147] When the application program 1080 is implemented in software, it should be noted that the application program 1080 can be stored on almost any computer-readable medium for use by or in conjunction with any computer-related system or method. In the context of this document, a computer-readable medium can be an electronic, magnetic, optical, or other physical device or device that can contain or store a computer program for use by or in conjunction with a computer-related system or method.

[0148] The application 1080 can be implemented in a variety of computer-readable media for use by or in conjunction with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems capable of retrieving and executing instructions from an instruction execution system, apparatus, or device. In the context of this document, "computer-readable media" can be any means that can store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Computer-readable media can be, for example, but not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, devices, or propagation media.

[0149] Regarding Figure 1 and Figure 2 the method described and regarding Figure 3 the (one or more) systems described can be implemented in hardware or software or a combination of both (e.g., as firmware running on a hardware device). To the extent that an embodiment is implemented partially or fully in software, the functional steps shown in the process flow diagrams can be executed by a suitably programmed physical computing device, such as one or more central processing units (CPUs) or graphics processing units (GPUs). Each process and its individual component steps (as shown in the flowcharts) can be executed by the same or different computing devices. According to various embodiments, a computer-readable storage medium stores a computer program including computer program code configured to cause one or more physical computing devices to execute the encoding or decoding method described above when the program runs on the one or more physical computing devices.

[0150] The storage medium may include volatile and non-volatile computer memories, such as RAM, PROM, EPROM, and EEPROM, optical discs (such as CDs, DVDs, BDs), and magnetic storage media (such as hard disks and tapes). The various storage media can be fixed within the computing device or can be removable, such that one or more programs stored thereon can be loaded into the processor.

[0151] For embodiments implemented partially or fully in hardware, Figure 3The blocks shown in the block diagrams can be separate physical components, or logical subdivisions of a single physical component, or can all be implemented in an integrated manner in one physical component. The functions of a single block shown in the drawings can be divided among multiple components in an embodiment, or the functions of multiple blocks shown in the drawings can be combined in a single component in an embodiment. Hardware components suitable for use in embodiments of the present invention include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field programmable gate arrays (FPGAs). One or more blocks can be implemented as a combination of dedicated hardware for performing some functions and one or more programmed microprocessors and associated circuitry for performing other functions.

[0152] Those skilled in the art can understand and implement variations of the disclosed embodiments when practicing the claimed invention by studying the drawings, the disclosure, and the claims. In the claims, the word "comprising" does not exclude other elements or steps, and the words "a" or "one" do not exclude a plurality. A single processor or other unit can perform the functions of several items recited in the claims. Although specific measures are recited in mutually different dependent claims, this does not indicate that the combination of these measures cannot be used advantageously. If a computer program is described above, it can be stored / distributed on a suitable medium such as an optical storage medium or a solid state medium provided together with or as part of other hardware, but can also be distributed in other forms such as via the Internet or other wired or wireless telecommunication systems. If the term "adapted to" is used in the claims or the specification, it should be noted that the term "adapted to" is intended to be equivalent to the term "configured to". Any reference signs in the claims should not be construed as limiting the scope.

[0153] The flowcharts and block diagrams in the drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram can represent a module, segment, or portion of instructions that contains one or more executable instructions for implementing the specified logical function(s). In some alternative embodiments, the functions noted in the blocks may occur in the order shown in the figures. For example, depending on the functions involved, two blocks shown successively may actually be executed substantially simultaneously, or the blocks may sometimes be executed in the reverse order. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware system that performs the specified functions or actions or a combination of dedicated hardware and computer instructions.

Claims

1. A method (100) for predicting the remaining duration of a cardiotocogram (CTG) scan session of an object, the method comprising: Obtain (110) CTG data associated with the object during the CTG scan session; Obtain (112) movement data describing the movement of the object during the CTG scan session; Identify (120) artifacts in the CTG data associated with the movement of the object based on the CTG data and the movement data; And Analyze (130) the CTG data, the movement data, and the artifacts to predict the remaining duration of the CTG scan session.

2. The method according to claim 1, further comprising: Process (122) the CTG data to extract feature data indicative of the usability of the CTG data for diagnosis, wherein predicting the remaining duration is further based on the extracted feature data.

3. The method according to claim 2, wherein, The feature data includes at least one of fetal heart rate measurements, uterine activity measurements, and fetal wakefulness states, and optionally wherein the fetal heart rate measurements include heart rate characteristics, the heart rate characteristics including fetal heart rate variability, fetal baseline heart rate, fetal heart rate accelerations, and fetal heart rate decelerations.

4. The method according to any one of claims 1 - 3, further comprising: Obtain (114) previous CTG data associated with the object during at least one previous CTG scan session, and wherein predicting the remaining duration is further based on the previous CTG data.

5. The method according to any one of claims 1 - 4, further comprising: Obtain (116) physiological data describing one or more characteristics of the object, and wherein predicting the remaining duration is further based on the physiological data, and optionally wherein the physiological data includes at least one of age, gestational age, height, weight, medical condition, and medical history.

6. The method according to claim 5, further comprising: Match (200) the object with a plurality of similar objects based on the physiological data of the object and the physiological data of the plurality of other objects, wherein each of the plurality of similar objects is associated with historical CTG scan session duration data; and wherein the predicted remaining duration is further based on the historical CTG scan session duration data.

7. The method according to claim 6, wherein, Matching (200) the object includes: Obtain (220) the physiological data of each of the plurality of other objects; Cluster (230) the plurality of other objects based on the physiological data to classify similar objects into one of a plurality of categories; and Match (240) the object to one of the plurality of categories based on the physiological data of the object.

8. The method according to any one of claims 1 - 7, further comprising obtaining (118) context information describing the context of the CTG scan session, and wherein, Predicting the remaining duration is further based on the context information, and optionally wherein the context information includes at least one of the time of day, location, ambient temperature, and the device used to obtain the CTG data.

9. The method according to any one of claims 1 - 8, wherein, Analyzing the CTG data, the movement data, and the artifacts includes: Providing the CTG data, the movement data, and the artifacts to a machine learning algorithm trained to predict the remaining duration of the CTG scan session; and Obtaining the predicted remaining duration from the machine learning algorithm.

10. The method according to claim 9, wherein, The machine learning algorithm is trained using a training algorithm configured to receive an array of training inputs and corresponding known outputs, where the training inputs include historical CTG data from previous CTG scan sessions of the subject, and where the corresponding known outputs include the historical CTG scan durations of the previous CTG scan sessions of the subject.

11. The method according to any one of claims 1 - 10, wherein, The CTG data is obtained from at least one of an ultrasound Doppler-based CTG sensor or an electrophysiology-based CTG sensor, and where the movement data is obtained from at least one of an accelerometer and a muscle electromyogram (EMG) sensor.

12. The method according to any one of claims 1 - 11, further comprising determining (140) the probability of session completion within at least one of a predetermined number of times based on the predicted remaining duration.

13. The method according to any one of claims 1 - 12, further comprising generating (142) a recommendation based on the analysis that describes actions to be taken to reduce the remaining duration.

14. A computer program comprising computer program code modules which, when the computer program is run on a computer, are adapted to implement the method according to any one of claims 1 - 13.

15. A system (300) for predicting the remaining duration of a cardiotocogram CTG scan session, the system comprising: An interface (310) configured to obtain CTG data associated with a subject during a CTG scan session, and movement data describing the movement of the subject during the CTG scan session; and A processor (320) configured to: Identify artifacts in the CTG data associated with the movement of the subject based on the CTG data and the movement data; Analyze the CTG data, the movement data, and the artifacts to predict the remaining duration of the CTG scan session.

Images