A fuzzy measurement-based assessment of breast cancer prognostic markers

The paper aims to assess breast cancer prognostic markers and to determine an optimum subset that can yield high prediction accuracy for an individual breast cancer patient's prognosis by means of a fuzzy measurement derived from the fuzzy k-nearest neighbour algorithm (FK-NN). The analyses are...

Full description

Saved in:
Bibliographic Details
Main Authors: Seker, H., Odetayo, M., Petrovic, D., Naguib, R.N.G., Bartoli, C., Alasio, L., Lakshmi, M.S., Sherbet, G.V.
Format: Conference Proceeding
Language:English
Subjects:
Accuracy
Artificial intelligence
Biomedical computing
Biomedical measurements
Breast cancer
Diseases
DNA
Medical diagnostic imaging
Oncology
USA Councils
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The paper aims to assess breast cancer prognostic markers and to determine an optimum subset that can yield high prediction accuracy for an individual breast cancer patient's prognosis by means of a fuzzy measurement derived from the fuzzy k-nearest neighbour algorithm (FK-NN). The analyses are carried out for both nodal involvement and five-year survival. The data set used for the analysis of breast cancer prognosis consists of seven input markers (histology type, grade, DNA ploidy, S-Phase Fraction (SPF), G/sub 0/G/sub 1//G/sub 2/M ratio, minimum and maximum nuclear pleomorphism indices (NPI)) and two corresponding outputs to be predicted (negative or positive nodal status in the case of nodal involvement assessment, and whether the patient is alive or dead within 5 years of diagnosis for survival analysis). The highest predictive accuracy is 78% with the fuzzy measurement of 0.7254 for nodal involvement assessment, and 88% with the fuzzy measurement of 0.8183 for survival analysis. The best results are obtained from the subset (Histology type, Grade, DNA. Ploidy, SPF (%), G/sub 0/G/sub 1//G/sub 2/M Ratio) for survival prediction and the subset (Grade, SPF, minimum NPI) for nodal involvement analysis.
DOI:10.1109/ITAB.2000.892381