Analysis of chromatographic signals from urine samples for prostate cancer analysis using signal processing.

Abstract

The early stoppage of prostate cancer is of vital importance for the start-up of the treatment of diagnosed patient, it is for them
that the present project has as objective to look for an alternative tool that allows the possibility of presenting a tool of support for detecting
prostate cancer, developing a method of diagnosis based on artificial intelligence, through chromatography signals of flatulence’s come from
urine sample for the study of the area of urology. In this project used mathematical techniques such as SVM and RNA, in order to extract and
verify the pattern found. The algorithm of present project was trained 10 signals chromatography came from urine sample, which five were
obtained of groups of control patients, and five of sick patients. For the development of analisys used Matlab 2014ª student version. The
Obtaining of positive results took with it the application of techniques of pre and processing about chromatography signals, among which are
clipping of interest area, filtering and base line correction whose propose of application allowed the search of typical characteristic patterns of
each group of patients, due to the presence of abnormal or cancerous cells in the prostate. The system was appreciated, doing used of blind dates
and the result contrasted with the expert doctor in the area, allowing of this manner to arrive to punctual conclusions. The software specificity
was 92.86%, rate obtained in the validation of software, whose signals entered come from of urine sample of analyzed patients, but different
from the ones used to the treatment.