Method and system for estimating arterial input function
Examples of the present disclosure include a method and system for deriving a computer implemented trained artificial intelligence (AI) model that is capable of predicting arterial input function (AIF) from blood and myocardial signal intensity curves of a subject obtained using a magnetic resonance (MR) scanner during injection of a single high dosage bolus of contrast agent that would otherwise cause signal saturation. The method and system requires an input data set of a large number of prior obtained dual bolus or dual sequence sets of scan measurements to be used as training data for the AI model. Once the computer implemented trained AI model has been obtained, it is deployed in a further method and system which receives MR data for a particular subject for which myocardial perfusion parameters such as the AIF is to found, the MR data being obtained from an MR scan taken during injection of a single bolus of contrast agent at a concentration sufficient to give MR signal saturation. The trained AI model is then able to predict, from the single bolus saturated MR data, myocardial perfusion parameters such as the AIF for the subject. The trained AI model is also able to convert between dual bolus datasets and duals sequence datasets.
1 . A computer implemented method of generating a trained machine learning model capable of predicting arterial input function parameters of a subject, the method comprising:
receiving a training data set comprising a plurality of sets of: (i) cardiac magnetic resonance (CMR) scan data sets obtained using a dual bolus contrast agent CMR imaging protocol and/or a dual sequence CMR imaging protocol, the CMR scan data sets comprising:
first CMR scan data obtained from a CMR scan undertaken during the administration of a first bolus of diluted contrast agent which does not cause MR signal saturation, and second CMR scan data obtained from a CMR scan undertaken during the administration of a second bolus of contrast agent, the second bolus of contrast agent having a concentration which causes MR signal saturation, and (ii) arterial input function parameters, the CMR scan data sets having associated arterial input function parameters obtained from the CMR scan data sets;
training a machine learning model using the training data set to provide a trained artificial intelligence model capable of predicting arterial input function parameters for a subject based on a single input CMR data set obtained from a single CMR scan of the subject undertaken during the administration of a single bolus of contrast agent sufficient to cause MR signal saturation; and
storing and/or outputting the trained machine learning model for future use.
2 . The method according to claim 1 , wherein the machine learning model is a neural network.
3 . The method according to claim 1 , wherein the arterial input function is then used to determine myocardial perfusion parameters for the subject.
4 . The method according to claim 1 , wherein the training data set comprises at least 10, more preferably at least 100, 1000, 3000 or 5000 sets of CMR scan data sets and associated arterial input function parameters.
5 . The method according to claim 1 , wherein the training data set comprises at least 100 sets of CMR scan data sets and associated arterial input function parameters.
6 . The method according to claim 1 , wherein the training data set comprises at least 1000 sets of CMR scan data sets and associated arterial input function parameters.
7 . The method according to claim 1 , wherein the training data set comprises at least 3000, and more preferably at least 5000, sets of CMR scan data sets and associated arterial input function parameters.
8 . The method according to claim 1 , wherein the dual bolus CMR dataset is derived from dual sequence CMR dataset or vice versa.
9 . A computer implemented method of determining an arterial input function for a subject from a cardiac magnetic resonance (CMR) scan, the method comprising:
acquiring saturated CMR scan data from a single CMR scan of the subject undertaken during the administration of a single bolus of contrast agent sufficient to cause MR signal saturation;
inputting the saturated CMR scan data into a trained machine learning model, the trained machine learning model having been trained using a training data set comprising a plurality of sets of: (i) cardiac magnetic resonance (CMR) scan data sets obtained using a dual bolus contrast agent CMR imaging protocol and/or a dual sequence CMR imaging protocol, the CMR scan data sets comprising: first CMR scan data obtained from a CMR scan undertaken during the administration of a first bolus of diluted contrast agent which does not cause MR signal saturation, and second CMR scan data obtained from a CMR scan undertaken during the administration of a second bolus of contrast agent, the second bolus of contrast agent having a concentration which causes MR signal saturation, and (ii) arterial input function parameters, the CMR scan data sets having associated arterial input function parameters obtained from the CMR scan data sets; and
receiving, as an output from the trained machine learning model, a prediction of arterial input function parameters in the subject.
10 . The method according to claim 9 , wherein the prediction of the arterial input function parameters comprises a simulated or predicted non saturated MR data set corresponding to an MR data set that would have been obtained from the subject had a diluted bolus of contrast agent which did not cause MR signal saturation been used during the scan of the subject.
11 . The method according to claim 9 , wherein the prediction of the arterial input function parameters comprises a simulated or predicted non saturated MR data set corresponding to an MR data set that would have been obtained from the subject had a dual sequence acquisition method been used during the scan of the subject.
12 . The method according to claim 10 , wherein the arterial input function parameters for the subject are derived from the simulated or predicted non saturated MR data set, and the acquired saturated CMR scan data.
13 . The method according to claim 9 , wherein the dual bolus CMR dataset is derived from dual sequence CMR dataset.
14 . The method according to claim 9 , wherein the dual sequence CMR dataset is derived from dual bolus CMR dataset or vice versa.
15 . The method according to claim 9 , wherein the arterial input function parameters include the arterial input function for the subject.
16 . The method according to claim 9 , wherein the trained machine learning model is a trained neural network.
17 . The method according to claim 9 , wherein the trained machine learning model is trained by:
receiving a training data set comprising a plurality of sets of cardiac magnetic resonance (CMR) scan data sets obtained using a dual bolus contrast agent CMR imaging protocol and/or a dual sequence CMR imaging protocol and associated arterial input function parameters obtained therefrom;
training a machine learning model using the training data set to provide a trained artificial intelligence model capable of predicting arterial input function parameters for a subject based on a single input CMR data set obtained using a single bolus of contrast agent; and
storing and/or outputting the trained machine learning model for future use.
18 . The method according to claim 9 , wherein the acquiring comprises undertaking a single CMR scan of the subject using an MR scanner, the subject having had administered a single bolus of contrast agent sufficient to cause MR signal saturation prior to undertaking the scan.
19 . A system for generating a trained machine learning model capable of predicting arterial input function parameters of a subject, the system comprising:
a processor; and
a computer readable medium storing computer readable instructions that when executed by the processor cause the system to operate to:
receive a training data set comprising a plurality of sets of: (i) cardiac magnetic resonance (CMR) scan data sets obtained using a dual bolus contrast agent CMR imaging protocol and/or a dual sequence CMR imaging protocol, the CMR scan data sets comprising: first CMR scan data obtained from a CMR scan undertaken during the administration of a first bolus of diluted contrast agent which does not cause MR signal saturation, and second CMR scan data obtained from a CMR scan undertaken during the administration of a second bolus of contrast agent, the second bolus of contrast agent having a concentration which causes MR signal saturation, and (ii) arterial input function parameters, the CMR scan data sets having associated arterial input function parameters obtained from the CMR scan data sets;
train a machine learning model using the training data set to provide a trained artificial intelligence model capable of predicting arterial input function parameters for a subject based on a single input CMR data set obtained from a single CMR scan of the subject undertaken during the administration of a single bolus of contrast agent sufficient to cause MR signal saturation; and
store and/or output the trained machine learning model for future use.
20 . A system for determining arterial input function parameters for a subject from a cardiac magnetic resonance (CMR) scan, the system comprising:
a processor; and
a computer readable medium storing computer readable instructions that when executed by the processor cause the system to operate to:
acquire saturated CMR scan data from a single CMR scan undertaken during the administration of a single bolus of contrast agent sufficient to cause MR signal saturation;
input the saturated CMR scan data into a trained machine learning model, the trained machine learning model having been trained using a training data set comprising a plurality of sets of (i) cardiac magnetic resonance (CMR) scan data sets obtained using a dual bolus contrast agent CMR imaging protocol and/or a dual sequence CMR imaging protocol, the CMR scan data sets comprising: first CMR scan data obtained from a CMR scan undertaken during the administration of a first bolus of diluted contrast agent which does not cause MR signal saturation, and second CMR scan data obtained from a CMR scan undertaken during the administration of a second bolus of contrast agent, the second bolus of contrast agent having a concentration which causes MR signal saturation, and (ii) arterial input function parameters, the CMR scan data sets having associated arterial input function parameters obtained from the CMR scan data sets; and
receive, as an output from the trained machine learning model, data indicative of arterial input function parameters in the subject.