Integrated diagnostic imaging system with automated protocol selection and analysis
Disclosed in an integrated diagnostic imaging system (IDIS) designed to optimize the process of medical imaging and disease diagnosis. The IDIS can incorporate an intuitive disease selection interface that allows clinicians to select potential diseases of interest to screen for. Subsequently, a dynamic protocol selection algorithm can utilize a comprehensive imaging protocol database to determine the optimal imaging protocol corresponding to the selected diseases. The chosen protocols can be then executed by an imaging system, which can include different imaging modalities. Following the image acquisition, image analysis processes specific to the targeted diseases can process the images and identify pertinent features. The diagnostic results can be displayed in an intuitive and interactive results display, enhancing clinician interpretation and decision-making. The IDIS can streamline the diagnostic imaging workflow, facilitate accurate disease detection, and ensure adaptability to advancements in medical image analysis.
1 . A non-transitory computer readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform a retinal imaging method comprising:
receiving one or more health selections from a user interface, wherein the one or more health selections comprise a selection of a first disease and a second disease from a plurality of diseases;
identifying a first imaging protocol associated with the first disease and identifying a second imaging protocol associated with the second disease from a plurality of imaging protocols, wherein the first imaging protocol and the second imaging protocol each specifies one or more retinal view identifiers for acquisition by a retinal imaging camera, the one or more retinal view identifiers comprising at least one of a macula-centered view identifier or an optic-disc-centered view identifier;
generating a combined imaging protocol based on combining the first imaging protocol and the second imaging protocol, the combining comprising removing a redundancy between the first and second imaging protocols, wherein the redundancy comprises a duplicate retinal view identifier specified by both the first imaging protocol and the second imaging protocol;
obtaining, with the retinal imaging camera, a plurality of images of a retina according to the combined imaging protocol, wherein obtaining the plurality of images comprises, based on the combined imaging protocol, (i) setting one or more parameters for imaging the retina comprising at least one of illumination, focus, aperture selection, or image capture, and (ii) selectively activating a fixation target of the retinal imaging camera comprising at least one light emitting diode to direct gaze for at least one retinal view identifier specified by the combined imaging protocol; and
providing an indication of a presence of the first disease and the second disease based on analyzing the plurality of images of the retina, wherein analyzing comprises selecting, from a model registry, one or more image analysis algorithms associated with the first disease and the second disease and executing the selected one or more image analysis algorithms,
wherein the one or more processors comprise one or more processors of the retinal imaging camera and one or more processors of a computing device configured to be communicatively coupled to the retinal imaging camera.
2 . The non-transitory computer readable medium of claim 1 , wherein the first imaging protocol comprises a first sequence of images and the second imaging protocol comprises a second sequence of images, and wherein the redundancy comprises an overlap between the first and second sequences of images.
3 . The non-transitory computer readable medium of claim 2 , wherein the identifying comprises selecting the first imaging protocol from a first plurality of imaging protocols associated with the first disease and selecting the second imaging protocol from a second plurality of imaging protocols associated with the second disease, and wherein the first and second imaging protocols are selected responsive to determining that the overlap between the first and second imaging protocols exceeds overlaps between other pairs of imaging protocols from the first and second pluralities of imaging protocols.
4 . The non-transitory computer readable medium of claim 1 , wherein the indication of the presence of the first disease and the second disease is performed by a classifier.
5 . The non-transitory computer readable medium of claim 4 , wherein the classifier comprises a machine learning model configured to detect in the plurality of images one or more features relevant to the first disease and the second disease.
6 . The non-transitory computer readable medium of claim 1 , wherein providing the indication comprises displaying the indication on a display.
7 . The non-transitory computer readable medium of claim 1 , wherein providing the indication comprises communicating the indication to a remote computing system.
8 . A retinal imaging method comprising:
receiving one or more health selections from a user interface, wherein the one or more health selections comprise a selection of a first disease and a second disease from a plurality of diseases;
identifying a first imaging protocol associated with the first disease and identifying a second imaging protocol associated with the second disease from a plurality of imaging protocols, wherein the first imaging protocol and the second imaging protocol each specifies one or more retinal view identifiers for acquisition by a retinal imaging camera, the one or more retinal view identifiers comprising at least one of a macula-centered view identifier or an optic-disc-centered view identifier;
generating a combined imaging protocol based on combining the first imaging protocol and the second imaging protocol, the combining comprising removing a redundancy between the first and second imaging protocols, wherein the redundancy comprises a duplicate retinal view identifier specified by both the first imaging protocol and the second imaging protocol;
obtaining, with the retinal imaging camera, a plurality of images of a retina according to the combined imaging protocol, wherein obtaining the plurality of images comprises, based on the combined imaging protocol, (i) setting one or more parameters for imaging the retina comprising at least one of illumination, focus, aperture selection, or image capture, and (ii) selectively activating a fixation target of the retinal imaging camera comprising at least one light emitting diode to direct gaze for at least one retinal view identifier specified by the combined imaging protocol; and
providing an indication of a presence of the first disease and the second disease based on analyzing the plurality of images of the retina, wherein analyzing comprises selecting, from a model registry, one or more image analysis algorithms associated with the first disease and the second disease and executing the selected one or more image analysis algorithms,
wherein the retinal imaging method is performed by one or more processors of the retinal imaging camera and one or more processors of at least one computing device communicatively coupled to the retinal imaging camera.
9 . The retinal imaging method of claim 8 , wherein the first imaging protocol comprises a first sequence of images and the second imaging protocol comprises a second sequence of images, and wherein the redundancy comprises an overlap between the first and second sequences of images.
10 . The retinal imaging method of claim 9 , wherein the identifying comprises selecting the first imaging protocol from a first plurality of imaging protocols associated with the first disease and selecting the second imaging protocol from a second plurality of imaging protocols associated with the second disease, and wherein the first and second imaging protocols are selected responsive to determining that the overlap between the first and second imaging protocols exceeds overlaps between other pairs of imaging protocols from the first and second pluralities of imaging protocols.
11 . The retinal imaging method of claim 8 , wherein the indication of the presence of the first disease and the second disease is performed by a classifier.
12 . The retinal imaging method of claim 11 , wherein the classifier comprises a machine learning model configured to detect in the plurality of images one or more features relevant to the first disease and the second disease.
13 . The retinal imaging method of claim 8 , wherein providing the indication comprises displaying the indication on a display.
14 . The retinal imaging method of claim 8 , wherein providing the indication comprises communicating the indication to a remote computing system.