Selective differentiating diagnostic process based on broad data bases

• US 5,687,716 A
• Filed: 11/15/1995
• Issued: 11/18/1997
• Est. Priority Date: 11/15/1995
• Status: Expired due to Fees

• Current Assignee: H.Ang.Kan Beving, Nils U. Olsson, Peter Kaufmann
• Original Assignee: H.Ang.Kan Beving, Nils U. Olsson, Peter Kaufmann
• Inventors: Olsson, Nils U., Beving, H.ang.kan, Kaufmann, Peter
• Primary Examiner(s): JAWORSKI, FRANCIS J
• Application Number:
  US08/559,032
• Time in Patent Office: 734 Days
• Field of Search: 128/630, 128/653.1, 364/413.01, 364/413.02, 364/413.13, 395/21, 395/23, 395/24
• US Class Current: 600/300
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass
• CPC Class Codes: Y10S 706/924 Medical

Abstract

A process for analyzing for a variety of physical medical disorders, which comprises:

a) clinical testing by a common analysis, patients free of such disorders and those patients that possess such one or more disorders, which analysis distinquishes any such disorder thereby obtaining the characterization of a collection of such disorders as numerical data;

b) scaling the matrix of said numerical data in a computer;

c) configuring in a computer, an artificial neural network prescribed by the number of variables and the number of disorders, which artificial neural network is electronic data that possesses an input layer, one or more hidden layers and an output layer;

d) fitting the artificial neural network electronic data in a computer to the numerical data according to adjustable parameters of the artificial neural network, including one or more of

i) the number of neurons in a hidden layer;

ii) the number of hidden layers;

iii) the type of transfer functions in the layers; and

iv) the weights connecting the neurons;

e) whereby the artificial neural network is trained in the computer to automatically provide an analytic model;

f) withdrawing from a new patient that has not been diagnosed for such disorder, a sample or samples of a kind taken from said reference patients and subjecting said diagnostic sample to said clinical testing to obtain new numerical data;

g) automatically scaling in the computer the new data to the data derived from the reference patients;

h) feeding the scaled new data to the trained artificial neural network and the analytic model thereof, and

i) automatically obtaining a diagnosis of the new patient with respect to such disorders.

Claims

1. A process for analyzing for a variety of physical medical disorders, which comprises:

a) clinical testing by common analyses, samples of patients free of certain physical medical disorders encompassed by said variety of physical medical disorders and samples of patients that possess one or more of such variety of disorders, to obtain clinical testing of such variety of disorders as derived from such samples, which analyses distinguish such disorders thereby obtaining the characterization of a collection of such disorders as numerical data;

b) scaling the matrix of said numerical data in a computer;

c) configuring in a computer, an artificial neural network prescribed by the number of variables and the number of disorders, which artificial neural network is electronic data that possesses an input layer, one or more hidden layers and an output layer;

d) fitting the artificial neural network electronic data in a computer to the numerical data according to adjustable parameters of the artificial neural network;

e) whereby the artificial neural network is trained in the computer to automatically provide a analytic model;

f) withdrawing from a new patient that has not been diagnosed for such disorder, a sample or samples of a kind taken from said reference patients and subjecting said diagnostic sample to said clinical testing to obtain new numerical data;

g) automatically scaling in the computer the new data to the data derived from the reference patients;

h) feeding the scaled new data to the trained artificial neural network and the analytic model thereof, and

i) automatically obtaining a diagnosis of the new patient with respect to such disorders.

2. The process of claim 1 wherein the clinical testing involves diagnostic procedures that involve determining the nature and cause of 1 disease or injury through evaluation of patient history, examination, and review of laboratory data.

3. The process of claim 2 wherein the laboratory data includes the diagnostic analysis of bodily fluids, cells, soft and hard tissues, and organs.

4. The process of claim 3 wherein the data is from one or more of immunoassays, wet and dry chemical analysis, hematological analysis, cell size analysis, chromatographic analysis, spectrographic analysis, electrochemical analysis, pH analysis, ion selective potentiometry, NIR, ISS, SIMS, AES, ICP-MS, AAS, LC/MS, nuclear magnetic resonance spectrometry, nuclear magnetic imaging techniques, chemical elemental analysis, and gas analysis including gas chromatography, antibody analysis, blotting techniques, electrophoresis, amino acid sequencing, and DNA analysis.

5. The process of claim 1 wherein the obtained numerical data characterizes deviations from the data of healthy patients as well as data that correlates with the data of the reference group patients that possess the disorder.

6. The process of claim 1 wherein the obtained diagnosis provides the best matching reference group and thereby assigning the particular disorder for the physician.

7. The process of claim 1 wherein if no matching disorder is provided by the obtained diagnosis, any abnormality of the sample is flagged.

8. The process of claim 1 wherein autoscaling is done by analog or digital means and the translations are computed by computer.

9. The process of claim 8 wherein translating the data is effected on a computer through digital conversions.

10. The process of claim 1 wherein the artificial neural network is optimized by (i) determining the right number of hidden layers, (ii) the right number of neurons in the hidden layers, and (iii) overall training of the network.

11. The process of claim 10 wherein (i) and (ii) are chosen by configuring the network by a pre-chosen number and training the network, calculating prediction error between real and model values, storing this data and choosing the next higher number for number (i) and (ii) above.

12. The process of claim 11 wherein the pre-chosen number is at least one hidden layer and at least two neurons.

13. The process of claim 12 wherein the new network is trained, prediction error is calculated and stored, and this process is repeated until a minimum prediction error is attained.