A hierarchical classifier based on human blood plasma fluorescence for non-invasive colorectal cancer screening.

Authors:
Felipe Soares
Felipe Soares
Institute of Informatics - Universidade Federal do Rio Grande do Sul
Brazil
Karin Becker
Karin Becker
University of North Dakota
Grand Forks | United States
Michel J Anzanello
Michel J Anzanello
Federal University of Rio Grande do Sul
Brazil

Artif Intell Med 2017 Oct 20;82:1-10. Epub 2017 Sep 20.

Department of Industrial Engineering - Universidade Federal do Rio Grande do Sul, Av. Osvaldo Aranha, 99-5° andar, Porto Alegre, RS, Brazil. Electronic address:

Colorectal cancer (CRC) a leading cause of death by cancer, and screening programs for its early identification are at the heart of the increasing survival rates. To motivate population participation, non-invasive, accurate, scalable and cost-effective diagnosis methods are required. Blood fluorescence spectroscopy provides rich information that can be used for cancer identification. The main challenges in analyzing blood fluorescence data for CRC classification are related to its high dimensionality and inherent variability, especially when analyzing a small number of samples. In this paper, we present a hierarchical classification method based on plasma fluorescence to identify not only CRC, but also adenomas and other non-malignant colorectal findings that may require further medical investigation. A feature selection algorithm is proposed to deal with the high dimensionality and select discriminant fluorescence wavelengths. These are used to train a binary support vector machine (SVM) in the first level to identify the CRC samples. The remaining samples are then presented to a one-class SVM trained on healthy subjects to detect deviant samples, and thus non-malignant findings. This hierarchical design, together with the one class-SVM, aims to reduce the effects of small samples and high variability. Using a dataset analyzed in previous studies comprised of 12,341 wavelengths, we achieved much superior results. Sensitivity and specificity are 0.87 and 0.95 for CRC detection, and 0.60 and 0.79 for non-malignant findings, respectively. Compared to related work, the proposed method presented a better accuracy, required fewer features, and provides a unified approach that expands CRC detection to non-malignant findings.

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.artmed.2017.09.004DOI Listing

Still can't find the full text of the article?

Sign up to send a request to the authors directly.
October 2017
6 Reads

Publication Analysis

Top Keywords

non-malignant findings
12
high dimensionality
8
plasma fluorescence
8
colorectal cancer
8
identify crc
8
cancer screening
8
crc detection
8
blood fluorescence
8
crc
6
samples
5
fluorescence
5
vector machine
4
healthy subjects
4
machine svm
4
binary support
4
fluorescence wavelengths
4
wavelengths train
4
train binary
4
svm level
4
support vector
4

Similar Publications