Dosage individualization of erythropoietin using a profile-dependent support vector regression

• Published in: IEEE transactions on biomedical engineering Vol. 50; no. 10; pp. 1136 - 1142
• Main Authors: Martin-Guerrero, J.D., Camps-Valls, G., Soria-Olivas, E., Serrano-Lopez, A.J., Perez-Ruixo, J.J., Jimenez-Torres, N.V.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.10.2003; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adult; Aged; Aged, 80 and over; Algorithms; Anemia, Hemolytic - blood; Anemia, Hemolytic - drug therapy; Anemia, Hemolytic - etiology; Biological and medical sciences; Cohort Studies; Cost function; Digital signal processing; Drug Therapy, Computer-Assisted - methods; Erythropoietin - administration & dosage; Hemoglobins - analysis; Hospitals; Humans; Immune system; Injections, Subcutaneous; Kidney Failure, Chronic - blood; Kidney Failure, Chronic - complications; Kidney Failure, Chronic - therapy; Medical diagnostic imaging; Medical sciences; Medical treatment; Middle Aged; Multilayer perceptrons; Neural Networks (Computer); Pathology; Recombinant Proteins; Regression (Psychology); Renal Dialysis; Sensitivity analysis; Studies; Treatment Outcome; support vector method; time series prediction; Anemia; chronic renal failure; drug monitoring; neural networks; clinical pharma- cokinetics; erythropoietin
• ISSN: 0018-9294; 1558-2531
• DOI: 10.1109/TBME.2003.816084

The external administration of recombinant human erythropoietin is the chosen treatment for those patients with secondary anemia due to chronic renal failure in periodic hemodialysis. The objective of this paper is to carry out an individualized prediction of the EPO dosage to be administered to those patients. The high cost of this medication, its side-effects and the phenomenon of potential resistance which some individuals suffer all justify the need for a model which is capable of optimizing dosage individualization. A group of 110 patients and several patient factors were used to develop the models. The support vector regressor (SVR) is benchmarked with the classical multilayer perceptron (MLP) and the Autoregressive Conditional Heteroskedasticity (ARCH) model. We introduce a priori knowledge by relaxing or tightening the /spl epsiv/-insensitive region and the penalization parameter depending on the time period of the patients' follow-up. The so-called profile-dependent SVR (PD-SVR) improves results of the standard SVR method and the MLP. We perform sensitivity analysis on the MLP and inspect the distribution of the support vectors in the input and feature spaces in order to gain knowledge about the problem.