US 12,462,382 B2
Diagnosis assistance device, machine learning device, diagnosis assistance method, machine learning method, machine learning program, and Alzheimer's prediction program
Akihiko Shiino, Otsu (JP)
Assigned to NATIONAL UNIVERSITY CORPORATION SHIGA UNIVERSITY OF MEDICAL SCIENCE, Otsu (JP); and ERISA CO., LTD., Matsue (JP)
Appl. No. 17/920,396
Filed by NATIONAL UNIVERSITY CORPORATION SHIGA UNIVERSITY OF MEDICAL SCIENCE, Otsu (JP); and ERISA CO., LTD., Matsue (JP)
PCT Filed Apr. 22, 2021, PCT No. PCT/JP2021/016283
§ 371(c)(1), (2) Date Oct. 21, 2022,
PCT Pub. No. WO2021/215494, PCT Pub. Date Oct. 28, 2021.
Claims priority of application No. 2020-076563 (JP), filed on Apr. 23, 2020.
Prior Publication US 2023/0162351 A1, May 25, 2023
Int. Cl. G06T 7/00 (2017.01); G06T 7/11 (2017.01); G06T 7/62 (2017.01); G16H 50/20 (2018.01)
CPC G06T 7/0012 (2013.01) [G06T 7/11 (2017.01); G06T 7/62 (2017.01); G16H 50/20 (2018.01); G06T 2207/10088 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/30016 (2013.01)] 17 Claims
OG exemplary drawing
 
17. A non-transitory storage device including an Alzheimer's prediction program that causes a computer to execute steps comprising:
a teacher data generation step of generating teacher data from brain images of multiple persons and diagnosis results indicating whether each person has developed Alzheimer's disease before the end of a prescribed period from the acquisition of the brain image,
a learning step of learning a prediction algorithm based on the teacher data, and
a prediction step of predicting, according to the prediction algorithm, a possibility that a subject who has Alzheimer's disease neuropathologic change will develop Alzheimer's disease within the prescribed period;
wherein the teacher data generation step comprises:
separating gray matter from the brain images acquired from the persons,
setting multiple regions of interest in the gray matter,
calculating the volume of each region of interest,
calculating a z-value of each region of interest based on the volume, and
associating the diagnosis results with the z-values to generate the teacher data; and
wherein the prediction step comprises:
separating gray matter from a brain image acquired from the subject,
setting multiple regions of interest in the gray matter,
calculating the volume of each region of interest,
calculating a z-value of each region of interest based on the volume, and
predicting the possibility based on the z-values.