Abstract: According to an aspect, there is provided an apparatus for analysing a medical image, the apparatus comprising: a memory comprising instruction data representing a set of instructions; and a processor configured to communicate with the memory and to execute the set of instructions. The set of instructions, when executed by the processor, cause the processor to: i) use a first ML model to predict a label for the image; ii) obtain a saliency map indicating relevancy of different regions of the image to the first ML model when the first ML model predicted the label for the image; iii) obtain a segmentation of the image; and iv) determine an indication of a feature in the image that led to the label being predicted by the first ML model, using the segmentation and the saliency map in combination.

Abstract: Methods and systems are provided for assessing an area of interest in the circulatory system of a subject. The method comprises: obtaining (210) a temporal sequence of ultrasound image frames of the area of interest; obtaining (220, 320) one or more trained models; inputting (250, 350) the temporal sequence of ultrasound image frames into the one or more trained models; and analysing (260, 360) the temporal sequence of ultrasound image frames, using the one or more trained models, to produce output data characterising a condition of the area of interest, wherein the output data characterises an extent of atherosclerosis or calcification of the area of interest, and wherein the area of interest is the aortic valve. An associated training method for training the one or more models is also provided.

Abstract: A method (100) for classifying a patient's diastolic function, comprising: (i) receiving (120), from an ultrasound device (280), a plurality of 2D echocardiographic images of the patient's heart; (ii) analyzing (150), by a trained diastolic function prediction algorithm, the plurality of 2D echocardiographic images of the patient's heart to estimate left ventricular end-diastolic pressure (LVEDP); (iii) classifying (160) the patient's diastolic function as normal or abnormal based on the estimated LVEDP; and (iv) providing (170), to a user via a user interface, an indication of the patient's diastolic function as normal or abnormal.

Abstract: An ultrasound imaging system may receive ultrasound data of a heart, wherein the ultrasound data was acquired across at least a portion of a cardiac cycle of the heart. The ultrasound imaging system may generate strain measurements based, at least in part on the ultrasound data. The ultrasound imaging system may analyze, using a cardiotoxicity detection algorithm, at least the strain measurements to determine a cardiotoxicity level of the heart. In some examples, the cardiotoxicity detection algorithm comprises a support vector machine (SVM) model trained on one or more of left ventricular (LV) strain data, left atrial (LA) strain data, right ventricular (RV) strain data, or right atrial (RA) strain data. In some cases, the cardiotoxicity detection algorithm uses one or more global left ventricular (LV) strain curves as an input to determine the cardiotoxicity level of the heart.

Abstract: An ultrasound imaging system may receive ultrasound data of a heart, wherein the ultrasound data was acquired across at least a portion of a cardiac cycle of the heart. The ultrasound imaging system may generate a plurality of heart based measurements based, at least in part on the ultrasound data. The ultrasound imaging system may analyze, using a correlation algorithm, the plurality of heart based measurements to determine a value of cardiac pressure. In some examples, the correlation algorithm is a support vector machine (SVM) model trained on one or more of a left atrial (LA) index, a left ventricular (LV) index, LA volume parameters, echo parameters, LV strain data, LA strain data, right atrial (RA) strain data, and right ventricular (RV) strain data. In some cases, the ultrasound imaging system may weight one or more of the plurality of heart based measurements.

Abstract: A method (100) for providing an analysis of coronary artery disease (CAD), comprising: (i) receiving (120) patient metadata about the patient; (ii) receiving (130) a temporal sequence of 2D and/or 3D ultrasound images of the patient's heart; (iii) selecting (140), by the CAD prediction system, a plurality of ultrasound images from the temporal sequence; (iv) processing (150), using a trained AI algorithm of the CAD prediction system, the selected plurality of ultrasound images to generate a feature map of the selected plurality of ultrasound images; (v) analyzing (160) the generated feature map and the received patient metadata using a trained algorithm of the CAD prediction system to generate a CAD prediction output; (vi) providing (170), via a user interface of the CAD prediction system, the generated CAD prediction output.

Abstract: A system and method for providing an enhancing or contrast agent advisability indicator. Subject data for a subject is processed using an AI-based model to obtain an indication of whether a medical imaging procedure for the subject requires an enhancing or contrast agent. The enhancing or contrast agent advisability indicator is generated with respect to one or more target pathologies assessed in the medical imaging procedure.

Abstract: A computer implemented method for collating patient data for analysis comprises receiving a set of input data comprising a plurality of patient data records, wherein the plural patient data records comprise medical imaging data and at least one other patient data type; and generating a vector for each of the plural patient data records by processing each patient data record using a corresponding encoding algorithm, wherein the encoding algorithm used to generate the vector is selected based on the type of patient data record and wherein the vectors are for use by a machine learning model.

Abstract: Methods and systems are provided for assessing an area of interest in the circulatory system of a subject. The method comprises: obtaining (210) a temporal sequence of ultrasound image frames of the area of interest; obtaining (220, 320) one or more trained models; inputting (250, 350) the temporal sequence of ultrasound image frames into the one or more trained models; and analysing (260, 360) the temporal sequence of ultrasound image frames, using the one or more trained models, to produce output data characterising a condition of the area of interest, wherein the output data characterises an extent of atherosclerosis or calcification of the area of interest, and wherein the area of interest is the aortic valve. An associated training method for training the one or more models is also provided.

Abstract: Embodiments discussed herein facilitate segmentation of vascular plaque, training a deep learning model to segment vascular plaque, and/or informing clinical decision-making based on segmented vascular plaque. One example embodiment accessing vascular imaging data for a patient, wherein the vascular imaging data comprises a volume of interest; pre-process the vascular imaging data to generate pre-processed vascular imaging data; provide the pre-processed vascular imaging data to a deep learning model trained to segment a lumen and a vascular plaque; and obtain segmented vascular imaging data from the deep learning model, wherein the segmented vascular imaging data comprises a segmented lumen and a segmented vascular plaque in the volume of interest.

Abstract: Embodiments discussed herein facilitate classification of vascular plaque. One example embodiment can: access vascular imaging data comprising one or more slices, wherein each slice comprises a plurality of A-lines of that slice; for each A-line of the plurality of A-lines of each slice of the one or more slices: extract one or more features for that A-line, wherein the one or more features for that A-line comprise at least one of: one or more features extracted from that A-line, one or more features extracted from the slice comprising that A-line, or one or more features extracted from the vascular imaging data; provide the one or more features for that A-line to at least one classifier; and generate a classification of that A-line via the at least one classifier, wherein the classification of that A-line is one of fibrocalcific, fibrolipidic, or other.

Abstract: Embodiments discussed herein facilitate segmentation of vascular plaque, training a deep learning model to segment vascular plaque, and/or informing clinical decision-making based on segmented vascular plaque. One example embodiment accessing vascular imaging data for a patient, wherein the vascular imaging data comprises a volume of interest; pre-process the vascular imaging data to generate pre-processed vascular imaging data; provide the pre-processed vascular imaging data to a deep learning model trained to segment a lumen and a vascular plaque; and obtain segmented vascular imaging data from the deep learning model, wherein the segmented vascular imaging data comprises a segmented lumen and a segmented vascular plaque in the volume of interest.

Abstract: Embodiments discussed herein facilitate classification of vascular plaque. One example embodiment can: access vascular imaging data comprising one or more slices, wherein each slice comprises a plurality of A-lines of that slice; for each A-line of the plurality of A-lines of each slice of the one or more slices: extract one or more features for that A-line, wherein the one or more features for that A-line comprise at least one of: one or more features extracted from that A-line, one or more features extracted from the slice comprising that A-line, or one or more features extracted from the vascular imaging data; provide the one or more features for that A-line to at least one classifier; and generate a classification of that A-line via the at least one classifier, wherein the classification of that A-line is one of fibrocalcific, fibrolipidic, or other.