Classification of factor deficiencies from coagulation assays using neural networks.

Activated partial thromboplastin time (APTT) and prothrombin time (PT) assays are widely used to screen for coagulation disorders and to monitor administration of therapeutic drugs. The analysis of data from coagulation assays has traditionally concentrated on determination of clot times (for APTT and PT) and magnitude of signal change during coagulation (e.g. for PT-based fibrinogen quantitation). The purpose of this study was to determine if the diagnostic power of these assays could be increased by using neural networks to interpret multiple parameters from these assays. Error back-propagation neural networks were trained using multiple variables derived from APTT and PT optical data for 200 normal and abnormal patient specimens. These networks were used to: (1) classify samples as either deficient or non-deficient with respect to individual blood components; and (2) estimate the approximate concentration of specific coagulation factors. Results indicated that these networks could be successfully trained to identify specific factor deficiencies at less than 30% normal levels with good specificity and variable sensitivity, but that they estimated actual concentrations poorly in most cases. These results support possible applications for neural networks identifying specific coagulation abnormalities from non-specific APTT and PT assays using expanded data parameter sets.

Properties of optical data from activated partial thromboplastin time and prothrombin time assays.

Changes in characteristics of optical transmittance data from coagulation assays were examined as a function of concentration of coagulation proteins or anticoagulants. Transmittance data were collected for activated partial thromboplastin time (APTT) and prothrombin time (PT) assays from: 1) plasmas prepared by mixing normal plasmas with deficient plasmas to give varying levels of coagulation proteins; 2) plasmas containing added heparin; and 3) 200 specimen plasmas that were also assayed for fibrinogen, coagulation factors, and other components. Optical profiles were characterized using a set of parameters describing onset and completion of coagulation, magnitude of signal change, rate of coagulation and other properties. Results indicated that parameters other than those typically reported for APTT and PT are associated with individual deficiencies, but that diagnosis of specimen status on the basis of optical data is complex. These results suggest possibilities for expanded interpretation of PT/APTT optical data for clinical or research applications.

Predicting the presence of plasma heparin using neural networks to analyze coagulation screening assay optical profiles.

A method for predicting the presence of heparin from coagulation screening assays is described and data are presented. This method incorporates the use of a multilayer perceptron trained through an error back-propagation algorithm in analyzing clotting optical data profiles. This method may lead to the identification of abnormalities from screening assays that might otherwise go undetected, or require additional testing to isolate.

Improvements in accuracy and reproducibility of quantitative clotting factor assays by use of a novel approach for modeling reference curves.

A method for creating a reference model in the quantitative assay of specific clotting factor activities is described. This method incorporates the use of a piecewise function with two component polynomials. This function allows more accurate representation of the global coagulation reaction, a sequential activation of multiple serine protease enzymes and cofactors, leading to improvements over traditional methods in range, accuracy, precision and robustness in reported activity levels. Clotting factor assay results using this method are compared with traditional and other candidate methods.