System and methods to predict serum lactate level

• US 10,327,709 B2
• Filed: 08/12/2016
• Issued: 06/25/2019
• Est. Priority Date: 08/12/2015
• Status: Active Grant

• Current Assignee: Massachusetts Institute of Technology, The General Hospital Corporation (Mass General Brigham, Inc.)
• Original Assignee: Massachusetts Institute of Technology
• Inventors: Heldt, Thomas, Dunitz, Max H., Verghese, George Cheeran, Reisner, Andrew Tomas, Filbin, Michael
• Primary Examiner(s): Natnithithadha, Navin
• Application Number:
  US15/236,114
• Publication Number:
  US 20170042483A1
• Time in Patent Office: 1,047 Days
• Field of Search: 600481, 600483-486, 600488, 600490-503
• CPC Class Codes: A61B 5/02028 Determining haemodynamic parameters not otherwise provided for, e.g. cardiac contractility or left ventricular ejection fraction
• CPC Class Codes: A61B 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/02108 from analysis of pulse wave characteristics
• CPC Class Codes: A61B 5/024 Detecting, measuring or recording pulse rate or heart rate A61B5/0205, A61B5/021 take precedence
• CPC Class Codes: A61B 5/14546 for measuring analytes not otherwise provided for, e.g. ions, cytochromes
• CPC Class Codes: A61B 5/222 combined with detection or measurement of physiological parameters, e.g. heart rate
• CPC Class Codes: A61B 5/412 Detecting or monitoring sepsis
• CPC Class Codes: A61B 5/4866 Evaluating metabolism A61B5/083 takes precedence
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
• CPC Class Codes: A61B 5/7278 Artificial waveform generation or derivation, e.g. synthesising signals from measured signals A61B5/327 takes precedence
• CPC Class Codes: A61B 5/7405 using sound
• CPC Class Codes: A61B 5/742 using visual displays A61B5/339 takes precedence
• CPC Class Codes: A61B 5/746 Alarms related to a physiological condition, e.g. details of setting alarm thresholds or avoiding false alarms alarms per se G08B
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

A method to quantitatively predict a patient'"'"'s serum lactate level, comprising measuring arterial blood pressure and heart rate from the patient, computing estimates of one or more cardiovascular parameters from the measured arterial blood pressure and heart rate, providing one or more classifiers that have been trained on a training data set including a reference set of arterial blood pressure, heart rate, and serum lactate levels and using the one or more classifiers to estimate the serum lactate level of the patient.

Claims

1. A method to estimate a patient'"'"'s serum lactate level, comprising:

measuring arterial blood pressure and heart rate from the patient;

computing estimates of one or more cardiovascular parameters from the measured arterial blood pressure and heart rate;

extracting one or more static parameters from history associated with the patient;

providing one or more classifiers that have been trained on a training data set obtained from a group comprising patients that are not the patient, the training data set including;

a reference set of arterial blood pressure, heart rate, and serum lactate levels recorded from the group,

estimates of one or more cardiovascular parameters from the reference set that correspond to the one or more cardiovascular parameters estimated from the measured arterial blood pressure and heart rate of the patient, and

one or more static parameters extracted from histories associated with the group that correspond to the one or more static parameters extracted from the history associated with the patient; and

using, at an application stage, the one or more classifiers and one or more of the measured arterial blood pressure and heart rate, the estimates of the one or more cardiovascular parameters from the measured blood pressure and heart rate from the patient, and the one or more static parameters extracted from the history associated with the patient to estimate the serum lactate level of the patient,

wherein the one or more classifiers has been trained at least in part on one or more selected subsets of the training data set, the selected subsets formed at least in part based on the one or more static parameters extracted from the history associated with the patient.

2. The method of claim 1, wherein a risk of sepsis is assigned to the patient depending on the estimated serum lactate level.

3. The method of claim 1, wherein the one or more cardiovascular parameters estimated from the measured arterial blood pressure and heart rate of the patient include at least one of a total peripheral resistance, cardiac output, and stroke volume.

4. The method of claim 1, wherein the history associated with the patient includes demographic information and one or more lab values.

5. The method of claim 1, wherein the selected subsets of the training data set comprise patients with similar history.

6. The method of claim 1, wherein the arterial blood pressure and heart rate are determined from real-time wave forms.

7. The method of claim 1, wherein an alarm is generated if the estimated serum lactate level of the patient crosses a predetermined threshold.

8. The method of claim 1, wherein a classifier from the one or more classifiers is selected to estimate serum lactate level for a patient, wherein the selected classifier is based on a set of features including a median systolic blood pressure and a log-ratio that compares a first median heart rate for a length of time as measured at a beginning of a measurement and a second median heart rate for the length of time as measured at an end of the measurement.

9. The method of claim 1, wherein at least one of the one or more classifiers has a high median area under the curve value and/or a high equal error rate sensitivity in comparison to other classifiers in the one or more classifiers.

10. The method of claim 1, wherein the estimate of the patient'"'"'s serum lactate level is an estimate of the patient'"'"'s current serum lactate level.

11. The method of claim 1, wherein the estimate of the patient'"'"'s serum lactate level is a prediction of a future serum lactate level of the patient.

12. The method of claim 1, wherein the estimate of the patient'"'"'s serum lactate level is an estimate of the past serum lactate level of the patient.

13. A system to estimate a patient'"'"'s serum lactate level, the system comprising at least one processor configured to:

measure arterial blood pressure and heart rate from the patient;

compute estimates of one or more cardiovascular parameters from the measured arterial blood pressure and heart rate;

extract one or more static parameters from history associated with the patient;

provide one or more classifiers that have been trained on a training data set obtained from a group comprising patients that are not the patient, the training data set including;

a reference set of arterial blood pressure, heart rate, and serum lactate levels recorded from the group,

estimates of one or more cardiovascular parameters from the reference set that correspond to the one or more cardiovascular parameters estimated from the measured arterial blood pressure and heart rate of the patient, and

one or more static parameters extracted from histories associated with the group that correspond to the one or more static parameters extracted from the history associated with the patient; and

use, at an application stage, the one or more classifiers and one or more of the measured arterial blood pressure and heart rate, the estimates of the one or more cardiovascular parameters from the measured blood pressure and heart rate from the patient, and the one or more static parameters extracted from the history associated with the patient to estimate the serum lactate level of the patient,

wherein the one or more classifiers has been trained at least in part on one or more selected subsets of the training data, the selected subsets formed at least in part based on the one or more static parameters extracted from the history associated with the patient.

14. The system of claim 13, wherein a risk of sepsis is assigned to the patient depending on the estimated serum lactate level.

15. The system of claim 13, wherein the one or more cardiovascular parameters estimated from the measured arterial blood pressure and heart rate of the patient include at least one of a total peripheral resistance, cardiac output, and stroke volume.

16. The system of claim 13, wherein the history associated with the patient includes demographic information and one or more lab values.

17. The system of claim 13, wherein the selected subsets of the training data set comprise patients with similar history.

18. The system of claim 13, wherein the arterial blood pressure and heart rate are determined from real-time wave forms.

19. The system of claim 13, wherein an alarm is generated if the estimated serum lactate level of the patient crosses a predetermined threshold.

20. The system of claim 13, wherein a classifier from the one or more classifiers is selected to estimate serum lactate level for a patient, wherein the selected classifier is based on a set of features including a median systolic blood pressure and a log-ratio that compares a first median heart rate for a length of time as measured at a beginning of a measurement and a second median heart rate for the length of time as measured at an end of the measurement.

21. The system of claim 13, wherein at least one of the one or more classifiers has a high median area under the curve value and/or a high equal error rate sensitivity in comparison to other classifiers in the one or more classifiers.

22. The system of claim 13, wherein the estimate of the patient'"'"'s serum lactate level is an estimate of the patient'"'"'s current serum lactate level.

23. The system of claim 13, wherein the estimate of the patient'"'"'s serum lactate level is a prediction of a future serum lactate level of the patient.

24. The system of claim 13, wherein the estimate of the patient'"'"'s serum lactate level is an estimate of the past serum lactate level of the patient.

25. A method for treating a patient, comprising:

performing an intervention on a patient if an estimated serum lactate level of the patient is increased past a predetermined threshold, wherein the estimated serum lactate level is determined based at least in part on;

measurement of arterial blood pressure and heart rate from the patient,

a computation of estimates of one or more cardiovascular parameters from the measured arterial blood pressure and heart rate,

extraction of one or more static parameters from history associated with the patient, and

use of, at an application stage, one or more classifiers and one or more of the measured arterial blood pressure and heart rate, the estimates of the one or more cardiovascular parameters from the measured blood pressure and heart rate from the patient, and the one or more static parameters extracted from the history associated with the patient;

wherein;

the one or more classifiers have been trained on a training data set obtained from a group comprising patients that are not the patient, the training data set including;

a reference set of arterial blood pressure, heart rate, and serum lactate levels recorded from the group,

estimates of one or more cardiovascular parameters from the reference set that correspond to the those estimated from the measured arterial blood pressure and heart rate of the patient, and

one or more static parameters extracted from histories associated with the group that correspond to those extracted from the history associated with the patient; and

the one or more classifiers have been trained at least in part on one or more selected subsets of the training data, the subsets formed at least in part based on the one or more static parameters extracted from the history associated with the patient.