Diagnosis assistance system and method

Abstract

In this invention, a system (10) for providing diagnosis assistance information (110) regarding the sickness of a patient (5) comprises: an acquisition unit (11) that acquires patient information (105) including actual image data (15) from an MR image of the patient, which contains at least a reference region containing a portion that includes an evaluation region; and an information provision unit (12) that provides diagnosis assistance information (110) based on pseudo-PET image data (115) of the evaluation region, which is generated from actual image data of patient-specific MR images by an image processing model (60). The image processing model (60) is built by machine learning on the basis of training data (70) comprising, from a plurality of test subjects, actual image data (71) of MR images of the reference region and actual image data (72) of PET images containing the evaluation region, so as to generate pseudo-PET image data (75) of an evaluation region from actual image data (71) of a reference region MR image.

Description

technical field [0001] The present invention relates to a system and method for assisting the diagnosis of complex systems such as living organisms. Background technique [0002] As devices for diagnosing the form and function of an object to be examined (subject), known CT (Computed Tomography: computerized tomography), MRI (Magnetic Resonance Imaging: magnetic resonance imaging), PET (Positron Emission Tomography: positron emission Tomography), SPECT (Single Photon Emission Computed Tomography: Single Photon Emission Computed Tomography), PET-CT and other tomography devices (modalities), and are actually used in the diagnosis of dementia, Parkinson's disease and other diseases. In PET and SPECT, a radiopharmaceutical is administered into the body of a subject by intravenous injection or the like, and radiation released from the drug is captured inside the body to generate an image. According to the image using the medicine, not only the shape of each part of the body can ...

AI Technical Summary

Problems solved by technology

For example, imaging of images using drugs is generally more expensive than imaging such as MRI that does not use drugs, so the financial burden on the subject or the health insurance association that bears the examination costs of the subject may increase.

In addition, in imaging such as PET images using radioactive drugs, radioactive substances must be ingested into the body even in small amounts, and there is a possibility of exposure to radiation.

In addition, people with certain diseases such as kidney disease may not be able to use medicine to take images to avoid adverse effects

[0004] In Apoorva Sikka et al., "From MRI to FDG-PET: Cross-modal synthesis using 3D U-Net for multimodal Alzheimer's classification (MRI to FDG-PET : Cross-ModalSynthesis Using 3D U-Net For Multi-Modal Alzheimer's Classification)” ([Online], Cornell University Library (Cornell University Library) [Accessed 31 August 2018], Internet (URL: https: / / arxiv.Org / abs / 1807.10111v1) (hereinafter referred to as Sikka et al.), reveals magnetic resonance imaging (MRI) for measuring brain atrophy and positron emission tomography for quantifying metabolic decline Combined analysis of photography (PET) to improve the diagnostic accuracy of Alzheimer's disease, noting that this method is limited by the availability of scans corresponding to each modality

Images