Assessment Of 'Breath Print' In Patients With Chronic Kidney Disease During Dialysis By Non-Invasive Breath Screening Of Exhaled Volatile Compounds Using An Electronic Nose

摘要

sensors, coupled with principal components Analysis (PCA), Discriminant Function Analysis (DFA) and Support Vector Machines (SVMs), as alternative tools for the resolution of complex classification situations, were used. As results, the application of PCA, DFA and SVMs methods revealed a good separation between clusters of the breath samples of healthy subjects and CKD patients before, during and after HD. Consequently, the data treatment by DFA have demonstrated its efficiency to discriminate between CKD at pre-HD and post-HD. These findings indicate that the proposed e-nose could be a noninvasive diagnosis tool of breath analysis for real time monitoring of HD treatment.