US 12,076,120 B2
Systems, methods and media for estimating compensatory reserve and predicting hemodynamic decompensation using physiological data
Robert W. Techentin, Rochester, MN (US); Timothy B. Curry, Rochester, MN (US); Michael J. Joyner, Rochester, MN (US); Clifton R. Haider, Rochester, MN (US); David R. Holmes, III, Rochester, MN (US); Christopher L. Felton, Rochester, MN (US); Barry K. Gilbert, Rochester, MN (US); Charlotte Sue Van Dorn, Rochester, MN (US); William A. Carey, Rochester, MN (US); and Victor A. Convertino, San Antonio, TX (US)
Assigned to Mayo Foundation for Medical Education and Research, Rochester, MN (US); and The Government of the United States, as Represented by the Secretary of the Army, Frederick, MD (US)
Filed by Mayo Foundation for Medical Education and Research, Rochester, MN (US); and The Government of The United States, as Represented by the Secretary of the Army, Frederick, MD (US)
Filed on Jul. 21, 2020, as Appl. No. 16/934,805.
Claims priority of provisional application 62/877,145, filed on Jul. 22, 2019.
Prior Publication US 2021/0022620 A1, Jan. 28, 2021
Int. Cl. A61B 5/02 (2006.01); A61B 5/00 (2006.01); A61B 5/021 (2006.01); A61B 5/024 (2006.01); G06N 3/004 (2023.01); G06N 3/008 (2023.01)
CPC A61B 5/02042 (2013.01) [A61B 5/0004 (2013.01); A61B 5/02028 (2013.01); A61B 5/02108 (2013.01); A61B 5/02416 (2013.01); A61B 5/7264 (2013.01); A61B 5/7275 (2013.01); G06N 3/004 (2013.01); G06N 3/008 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A system for estimating compensatory reserve, the system comprising:
at least one hardware processor that is programmed to:
receive a blood pressure waveform of a subject;
generate a first sample of the blood pressure waveform, wherein the first sample comprises a time series of blood pressure values having a first duration;
provide the first sample as input to a trained one-dimensional (1D) convolutional neural network (CNN),
wherein the 1D CNN was trained as a regression model using samples of the first duration from blood pressure waveforms recorded from a plurality of subjects while decreasing the respective subject's central blood volume,
wherein each sample used to train the 1D CNN is a one-dimensional time series data structure that was associated with a compensatory reserve metric based on a decrease of the respective subject's central blood volume at a time the respective sample was recorded, and
wherein an output layer of the trained 1D CNN is a linear layer that outputs a compensatory reserve metric value;
receive, from the trained 1D CNN, a first compensatory reserve metric based on the first sample, wherein the first compensatory reserve metric is a single quantitative value indicating a percentage of compensatory reserve in the subject; and
cause information indicative of remaining compensatory reserve to be presented.