EXPERT SYSTEM FOR DETERMINING PATIENT TREATMENT RESPONSE

• US 20140279746A1
• Filed: 02/17/2014
• Published: 09/18/2014
• Est. Priority Date: 02/20/2008
• Status: Active Grant

• Current Assignee: Digital Medical Experts Inc.
• Original Assignee: Digital Medical Experts Inc.
• Inventors: De Bruin, Hubert, Hasey, Gary, Khodayari-Rostamabad, Ahmad, Maccrimmon, Duncan, Reilly, James
• Granted Patent: US 9,460,400 B2
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: A61B 5/00 Measuring for diagnostic purposes radiation diagnosis A61B6/00; diagnosis by ultrasonic, sonic or infrasonic waves A61B8/00; Identification of persons
• CPC Class Codes: A61B 5/369 Electroencephalography [EEG] devices for psychotechnics A61B5/16
• CPC Class Codes: A61B 5/372 Analysis of electroencephalograms
• CPC Class Codes: A61B 5/4088 Diagnosing of monitoring cognitive diseases, e.g. Alzheimer, prion diseases or dementia
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G16H 10/40 for data related to laboratory analysis, e.g. patient specimen analysis
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 20/10 relating to drugs or medications, e.g. for ensuring correct administration to patients
• CPC Class Codes: G16H 20/70 relating to mental therapies, e.g. psychological therapy or autogenous training
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

A medical digital expert system to predict a patient'"'"'s response to a variety of treatments (using pre-treatment information) is described. The system utilizes data fusion, advanced signal/information processing and machine learning/inference methodologies and technologies to integrate and explore diverse sets of attributes, parameters and information that are available to select the optimal treatment choice for an individual or for a subset of individuals suffering from any illness or disease including psychiatric, mental or neurological disorders and illnesses. The methodology and system can also be used to determine or confirm medical diagnosis, estimate the level, index, severity or critical medical parameters of the illness or condition, or provide a list of likely diagnoses for an individual suffering/experiencing any illness, disorder or condition.

Claims

1-42. -42. (canceled)

1. A computer-implemented method for predicting patient response to treatment for a subject patient, the method comprising:

receiving, for a subject patient, a subject patient dataset including features obtained for a reduced feature dataset; and

comparing the subject patient dataset to a feature data scheme to predict a response for the subject patient;

wherein the feature data scheme comprises;

(a) a diagnostic model for determining a diagnosis of one of a plurality of known disorders indicated by an individual patient dataset; and

(b) a plurality of diagnosis-specific treatment response models, each of the diagnosis-specific treatment response models corresponding to a specific diagnosis of one of the known disorders, the treatment response models configured to use feature data to predict treatment response;

wherein comparing the subject patient dataset to the feature data scheme to predict the response for the subject patient comprises;

applying the diagnostic model to the subject patient dataset to determine a subject patient diagnosis of one of the known disorders indicated for the subject patient by the subject patient dataset; and

selecting and applying one of the diagnosis-specific treatment response models to the subject patient dataset to predict the response for the subject patient;

wherein the selected diagnosis-specific treatment response model corresponds to the determined subject patient diagnosis.

2. The method of claim 1, wherein comparing the subject patient dataset to the feature data scheme to predict the response for the subject patient is carried out on-site where the subject patient dataset is collected.

3. The method of claim 1, wherein:

the subject patient dataset is received at a central processing site remote from a site where the subject patient dataset is collected; and

comparing the subject patient dataset to the feature data scheme to predict the response for the subject patient is carried out at the central processing site.

4. The method of claim 1, wherein the feature data scheme comprises:

a plurality of predictor architectures, each predictor architecture independently producing a preliminary treatment response prediction given the measured patient related data; and

a data fusion architecture for combining the preliminary treatment response predictions.

5. The method of claim 1, wherein the feature data scheme is generated by:

storing, on a system database, a first level training dataset comprising a plurality of records comprising measured patient related data from a large number of patients, the measured patient related data including clinical and/or laboratory data, diagnoses of presence or absence of known disorders and data relating to patient treatment response;

processing the measured patient related data to extract features from the measured patient related data to generate an extracted feature dataset;

processing the extracted feature dataset to derive the feature data scheme;

the feature data scheme including the reduced feature dataset whose cardinality is less than that of the extracted feature dataset;

the reduced feature dataset being derived by processing the data to obtain features appearing to discriminate for a useful prediction.

6. A method according to claim 5, wherein processing the extracted feature dataset includes segregating the feature data to discriminate for treatment response.

7. A method according to claim 5, wherein the feature data scheme allocates feature data values to a treatment response category.

8. A method according to claim 5, wherein the feature data scheme is determined by solving a multi-dimensional or joint optimization problem, according to a criterion of optimality.

9. A method according to claim 8, wherein the optimization problem is constrained by one of model accuracy and model structure.

10. A method according to claim 5, wherein the feature data scheme is derived by defining one or more decision or estimation functions between feature data values to segregate or discriminate between response categories.

11. A method according to claim 5, wherein processing the extracted feature dataset to derive the feature data scheme comprises defining one or more non-linear boundaries between feature data values.

12. A method according to claim 5, wherein processing the extracted feature dataset comprises processing according to a criterion of optimality.

13. A method according to claim 5, wherein one or more of the clinical and/or laboratory data are derived by means of a brain activity or brain structure monitoring system.

14. A method according to claim 5, wherein one or more of the clinical and/or laboratory data are derived by means of an electromagnetic monitoring system.

15. A method according to claim 5 wherein one or more of the clinical and/or laboratory data are derived by means of an EEG system wherein EEG data is derived from the patient(s) prior to undergoing the treatment for which response is being predicted.

16. A method according to claim 5, wherein one or more of the clinical and/or laboratory data relates to one or more of demographic, mood, anxiety or other psychiatric or neurological symptomatology, medical co-morbidity, pharmacological, biochemical, hematological, hepatic, renal, endocrine, immunologic, nutritional, metabolomic, genetic markers, cognitive or personality assessment information derived from the subject patient.

17. A method according to claim 5, wherein processing the extracted feature dataset comprises defining boundaries between feature data values to segregate the values into different classes.

18. A method according to claim 5, wherein a joint or multivariate criterion of optimality is employed to establish a decision and/or regression function to minimize classification and/or modeling error.

19. A method according to claim 5, wherein processing the extracted feature dataset comprises one of kernelization, mapping, linear data transformation and non-linear data transformation.

20. A method according to claim 5, wherein the feature data scheme is refined using a feedback step in which the training dataset is updated with further follow-up training data from patients.

21. A method according to claim 20, wherein the further follow-up training data comprises patient datasets in addition to patient response information.

22. A method according to claim 20, wherein the reduced feature dataset can change in response to updating of the training dataset with the further follow-up training data.

23. A method according to claim 20, wherein follow-up information regarding the response outcome that is determined to be either unreliable or invalid after screening is excluded from further use as part of the updated training data set.

24. A method according to claim 5, wherein processing the extracted feature dataset reduces the number of features to less than 20 from a number of training dataset candidate features in excess of 500.

25. A method according to claim 5, wherein processing the extracted feature dataset utilizes at least one of probabilistic estimation methods, information-theoretic decision or estimation procedure, and graphical network methods, to generate quantitative features or attributes characterizing at least one of statistical, geometrical, temporal, graphical, network, hierarchical or dynamic model properties of the measured data.

26. A method according to claim 5, wherein processing the extracted feature dataset utilizes a maximum mutual information technique.

27. A method according to claim 5, wherein processing the extracted feature dataset includes a procedure for ensuring the reduced feature dataset has minimum mutual redundancy.

28. A method according to claim 5, in which one or more of the clinical and/or laboratory data are derived by means of a brain activity or structure monitoring system, and one or more of the clinical and/or laboratory data elates to biological test results or demographic, mood, anxiety or other psychiatric symptomatology, medical co-morbidity, pharmacological, biochemical, hematological, hepatic, renal, endocrine, immunologic, nutritional, metabolomic, genetic markers, cognitive or personality assessment information derived from the subject patient.

29. A medical system comprising:

a first computer, wherein the first computer comprises;

storage for storing a feature data scheme, wherein the feature data scheme comprises;

(a) a diagnostic model for determining a diagnosis of one of a plurality of known disorders indicated by an individual patient dataset; and

(b) a plurality of diagnosis-specific treatment response models, each of the diagnosis-specific treatment response models corresponding to a specific diagnosis of one of the known disorders, the treatment response models configured to use feature data to predict treatment response; and

at least one processor coupled to the storage and to an input device, the processor executing instructions for;

receiving, for a subject patient, a subject patient dataset including features obtained for a reduced feature dataset; and

comparing the subject patient dataset to a feature data scheme to predict a response for the subject patient;

wherein comparing the subject patient dataset to the feature data scheme to predict the response for the subject patient comprises;

applying the diagnostic model to the subject patient dataset to determine a subject patient diagnosis of one of the known disorders indicated for the subject patient by the subject patient dataset; and

selecting and applying one of the diagnosis-specific treatment response models to the subject patient dataset to predict the response for the subject patient;

and wherein the selected diagnosis-specific treatment response model corresponds to the determined subject patient diagnosis.

30. The medical system of claim 29, wherein the feature data scheme is generated by:

storing, on a system database, a first level training dataset comprising a plurality of records comprising measured patient related data from a large number of patients, the measured patient related data including clinical and/or laboratory data, diagnoses of presence or absence of known disorders and data relating to patient treatment response;

processing the measured patient related data to extract features from the measured patient related data to generate an extracted feature dataset;

processing the extracted feature dataset to derive the feature data scheme;

the feature data scheme including the reduced feature dataset whose cardinality is less than that of the extracted feature dataset;

the reduced feature dataset being derived by processing the data to obtain features appearing to discriminate for a useful prediction.

31. A medical system according to claim 30, wherein the system includes a feedback arrangement in which the training database is updated with further training data to update the feature data scheme.

32. A medical system according to claim 29, further including means for inputting patient derived data.

33. A medical system according to claim 29, further comprising a brain activity or brain structure monitoring device for inputting patient derived feature data wherein the brain derived data is derived from the patient(s) prior to undergoing the treatment for which response is being predicted.

34. A computer-implemented method for assessment of status of a patient, the method comprising:

receiving, for a subject patient, a subject patient dataset including features obtained for a reduced feature dataset; and

comparing the subject patient dataset to a feature data scheme to allocate an assessment status for the subject patient;

wherein the feature data scheme comprises;

(a) a diagnostic model for determining a diagnosis of one of a plurality of known disorders indicated by an individual patient dataset; and

(b) a plurality of diagnosis-specific assessment status models, each of the diagnosis-specific assessment status models corresponding to a specific diagnosis of one of the known disorders, the assessment status models configured to use feature data to allocate an assessment status;

wherein comparing the subject patient dataset to the feature data scheme to allocate the assessment status for the subject patient comprises;

applying the diagnostic model to the subject patient dataset to determine a subject patient diagnosis of one of the known disorders indicated for the subject patient by the subject patient dataset; and

selecting and applying one of the diagnosis-specific assessment status models to the subject patient dataset to allocate the assessment status for the subject patient;

and wherein the selected diagnosis-specific assessment status model corresponds to the determined subject patient diagnosis.

35. The method of claim 34, wherein comparing the subject patient dataset to the feature data scheme to allocate the assessment status for the subject patient is carried out on-site where the subject patient dataset is collected.

36. The method of claim 34, wherein:

the subject patient dataset is received at a central processing site remote from a site where the subject patient dataset is collected; and

comparing the subject patient dataset to the feature data scheme to allocate the assessment status for the subject patient is carried out at the central processing site.

37. The method of claim 34, wherein the feature data scheme comprises:

a plurality of estimator architectures, each estimator architecture independently producing a preliminary assessment allocation given the measured patient related data; and

a data fusion architecture for combining the preliminary assessment allocations.

38. The method of claim 34, wherein the feature data scheme is generated by:

storing, on a system database, a first level training dataset comprising a plurality of records comprising measured patient related data from a large number of patients, the measured patient related data including clinical and/or laboratory data and diagnoses of presence or absence of known disorders;

processing the measured patient related data to extract features from the measured patient related data to generate an extracted feature dataset;

processing the extracted feature dataset to derive the feature data scheme;

the feature data scheme including the reduced feature dataset whose cardinality is less than that of the extracted feature dataset;

the reduced feature dataset being derived by processing the data to obtain features appearing to discriminate for a useful prediction.

39. A method according to claim 38, wherein processing the first level training dataset includes segregating the feature data to discriminate between assessment status.

40. A method according to claim 34, wherein the feature data scheme allocates feature data values to an assessment status category.

41. A method according to claim 38, wherein the feature data scheme is derived by defining one or more decision or estimation functions between feature data values to segregate or discriminate between assessment status categories.

42. A method according to claim 38, wherein the feature data scheme is derived by defining one or more non-linear boundaries between feature data values.

43. A method according to claim 38, wherein processing the first level training dataset comprises processing the first level training dataset according to a criterion of optimality.

44. A medical system comprising:

a computer, wherein the computer comprises;

storage for storing a feature data scheme, wherein the feature data scheme comprises;

(a) a diagnostic model for determining a diagnosis of one of a plurality of known disorders indicated by an individual patient dataset; and

(b) a plurality of diagnosis-specific assessment status models, each of the diagnosis-specific assessment status models corresponding to a specific diagnosis of one of the known disorders, the assessment status models configured to use feature data to allocate an assessment status;

at least one processor coupled to the storage and to an input device, the processor executing instructions for;

receiving, for a subject patient, a subject patient dataset including features obtained for a reduced feature dataset; and

comparing the subject patient dataset to a feature data scheme to allocate an assessment status for the subject patient;

wherein comparing the subject patient dataset to the feature data scheme to allocate the assessment status for the subject patient comprises;

applying the diagnostic model to the subject patient dataset to determine a subject patient diagnosis of one of the known disorders indicated for the subject patient by the subject patient dataset; and

selecting and applying one of the diagnosis-specific assessment status models to the subject patient dataset to allocate the assessment status for the subject patient;

and wherein the selected diagnosis-specific assessment status model corresponds to the determined subject patient diagnosis.

45. The medical system of claim 44, wherein the feature data scheme is generated by:

storing, on a system database, a first level training dataset comprising a plurality of records comprising measured patient related data from a large number of patients, the measured patient related data including clinical and/or laboratory data, diagnoses of presence or absence of known disorders and data relating to patient assessment status;

processing the measured patient related data to extract features from the measured patient related data to generate an extracted feature dataset;

processing the extracted feature dataset to derive the feature data scheme;

the feature data scheme including the reduced feature dataset whose cardinality is less than that of the extracted feature dataset;

the reduced feature dataset being derived by processing the data to obtain features appearing to discriminate for a useful prediction.