US 11,955,243 B2
Using unstructured temporal medical data for disease prediction
Carlos Federico Arteta Montilva, Oxford (GB); Nicholas Dowson, Oxford (GB); Timor Kadir, Oxford (GB); and Jerome Declerck, Oxford (GB)
Assigned to Optellum Limited, Oxford (GB)
Filed by Optellum Limited, Oxford (GB)
Filed on Nov. 11, 2020, as Appl. No. 17/094,949.
Prior Publication US 2022/0148733 A1, May 12, 2022
Int. Cl. G16H 50/30 (2018.01); A61B 5/00 (2006.01); A61B 5/055 (2006.01); A61B 5/08 (2006.01); A61B 6/03 (2006.01); A61B 8/00 (2006.01); G06T 7/00 (2017.01); G16H 10/40 (2018.01); G16H 30/20 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G16H 50/50 (2018.01); G16H 50/70 (2018.01)
CPC G16H 50/30 (2018.01) [A61B 5/004 (2013.01); A61B 5/055 (2013.01); A61B 5/08 (2013.01); A61B 5/7264 (2013.01); A61B 5/7275 (2013.01); A61B 5/7405 (2013.01); A61B 5/742 (2013.01); A61B 5/746 (2013.01); A61B 5/7475 (2013.01); A61B 6/032 (2013.01); A61B 6/037 (2013.01); A61B 8/00 (2013.01); G06T 7/0016 (2013.01); G16H 10/40 (2018.01); G16H 30/20 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G16H 50/50 (2018.01); G16H 50/70 (2018.01); G06T 2207/10081 (2013.01); G06T 2207/10088 (2013.01); G06T 2207/10104 (2013.01); G06T 2207/10108 (2013.01); G06T 2207/10116 (2013.01); G06T 2207/10132 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30061 (2013.01)] 16 Claims
OG exemplary drawing
 
1. A method for providing a lung disease risk measure in a Computer Aided Diagnosis system comprising the steps of:
receiving a plurality of inputs in an input sequence for a subject at an input circuit, each input comprising at least one image showing all or part of the lungs of a patient and a time stamp for the image, where the inputs are obtained at varying intervals;
analysing the inputs using a machine learning model to assess temporal changes in the images using an input data encoder and a time stamp encoder, wherein the input data encoder identifies patterns in the input data and encodes the result of the input processing as an input data descriptor; and the time stamp encoder parses time information about the image as a vector tN;
wherein the output from at least one encoder is provided to a system state calculator, and the system state calculator encodes a summary of all the input data received at that time; wherein system state calculator is updated via a feedback system, to take account of data from successive images, where for each new input data the system state calculator calculates a new system state using input units remaining in the sequence of input units until the sequence of input units is exhausted;
inputting the output of the system state calculator to a score calculator to calculate a risk score; and
outputting the risk score from an output circuit indicating the lung disease risk for the subject, wherein the risk score is the risk score obtained after the final input unit has been presented to the system.