Point of care diagnostic systems

Abstract

Systems and methods for medical diagnosis or risk assessment for a patient are provided. These systems and methods are designed to be employed at the point of care, such as in emergency rooms and operating rooms, or in any situation in which a rapid and accurate result is desired. The systems and methods process patient data, particularly data from point of care diagnostic tests or assays, including immunoassays, electrocardiograms, X-rays and other such tests, and provide an indication of a medical condition or risk or absence thereof. The systems include an instrument for reading or evaluating the test data and software for converting the data into diagnostic or risk assessment information.

1 . A method of classifying an image, comprising: 
 reducing the image to a set of derived parameters that can be used to reconstruct the image within a specified degree of tolerance;    inputting the derived parameters into a classification means; and    determining the classification of the image based on the output of the classification means.    
     
     
         2 . The method of  claim 1 , wherein the classification means is a neural network.  
     
     
         3 . The method of  claim 1 , wherein the means of reducing the image to a set of derived parameters comprises: 
 defining a mathematical function containing a plurality of parameters representative of the image; and    optimizing the parameters of the mathematical function using a numerical technique that minimizes the error between the image and a reconstruction of the image using the mathematical function.    
     
     
         4 . The method of  claim 1 , wherein the means of reducing the image to a set of derived parameters comprises: 
 inputting the image to a trained neural network, wherein the inputs to the network represent the image, the network comprises a hidden layer where the number of hidden elements is smaller than the number of inputs to the network, and the outputs of the network represent the reconstruction of the image; and    setting the derived parameters to the output values of the trained neural network.    
     
     
         5 . The method of  claim 1 , wherein the means of reducing the image to a set of derived parameters comprises: 
 defining a neural network wherein the inputs to the network are the coordinates of a point in the image, a hidden layer contains a plurality of elements, and the output of the network represents the reconstruction of the associated point in the image;    training the neural network so that the error between the network output and the image are minimized for all points in the image; and    setting the derived parameters to the weights of the hidden layer of the trained neural network.    
     
     
         6 . A method for determining results from a test strip, comprising: 
 (a) measuring light reflected from the surface of the test strip to obtain a reflectance signal, wherein the reflectance signal is indicative of the presence of the analyte; and    (b) processing the data obtained from the reflectance signal using data processing software employing data reduction and curve fitting algorithms and/or a trained neural network to convert the reflectance signal into a result indicative of the presence or absence or a threshold concentration of analyte in a sample by a process of  claim 1 .    
     
     
         7 . The method of  claim 6 , wherein the classification means is a neural network.  
     
     
         8 . The method of  claim 6 , wherein the means of reducing the image to a set of derived parameters comprises: 
 defining a mathematical function containing a plurality of parameters representative of the image; and    optimizing the parameters of the mathematical function using a numerical technique that minimizes the error between the image and a reconstruction of the image using the mathematical function.    
     
     
         9 . The method of  claim 6 , wherein the means of reducing the image to a set of derived parameters comprises: 
 inputting the image to a trained neural network, wherein the inputs to the network represent the image, the network comprises a hidden layer where the number of hidden elements is smaller than the number of inputs to the network, and the outputs of the network represent the reconstruction of the image; and    setting the derived parameters to the output values of the trained neural network.    
     
     
         10 . The method of  claim 6 , wherein the means of reducing the image to a set of derived parameters comprises: 
 defining a neural network wherein the inputs to the network are the coordinates of a point in the image, a hidden layer contains a plurality of elements, and the output of the network represents the reconstruction of the associated point in the image;    training the neural network so that the error between the network output and the image are minimized for all points in the image; and    setting the derived parameters to the weights of the hidden layer of the trained neural network.    
     
     
         11 . The method of  claim 6 , wherein the processing of data process, comprises: 
 (i) reducing the raw reflectance data;    (ii) plotting the reduced data to generate a second image of the data;    (iii) expressing the second image as a polynomial mathematical function to generate parameters that define this image; and    (iv) comparing the parameters to parameters generated from a reference sample, whereby a positive, negative or quantitative result is obtained.    
     
     
         12 . The method of  claim 6 , wherein the data is processed by a process, comprising: 
 (i) optionally correcting the reflectance readings to correct for light leakage;    (ii) reducing the raw reflectance data using a ratiometric formula;    (iii) generating a second image of the test data by plotting the reduced data;    (iv) expressing the second image as a polynomial mathematical function, and generating parameters that define the image;    (v) comparing the scanned image and second image by solving the linear regression through the curves;    (vi) validating the parameters obtained from the curve-fitting and the peak heights obtained to obtain a validated result; and    (vii) classifying the validated result as positive or negative by comparing peak heights of a clinical sample to reference samples.    
     
     
         13 . The method of  claim 12 , wherein the data processing process, further comprises: 
 (viii) inputting the validated result into a decision-support system to generate a medical diagnosis or risk assessment.    
     
     
         14 . The method of  claim 6 , wherein the data processing process, further comprises: 
 (viii) inputting the result into a decision-support system to generate a medical diagnosis or risk assessment.    
     
     
         15 . A method for determining results from a test strip, comprising: 
 (a) scanning the test strip in a reflectance reader to obtain a scanned image, wherein the signal is indicative of the presence of the analyte; and    (b) processing the data obtained from the reflectance signals using data processing software employing data reduction and curve fitting algorithms and/or a trained neural network to convert the reflectance signal obtained from reading the test strip into data indicative of the test results' presence or absence or a threshold concentration of analyte in a sample by a process of  claim 1.