Document Type

Conference Proceeding

Publication Date

5-2016

Keywords

Statistical Learning theory; Pleural effusion; atypical cell classification; clinical decision support systems (CDSS)

Abstract

Pleural effusion occurs when fluid accumulates in the pleural cavity surrounding the lung. This condition is commonly caused by infection, but can also be associated with the presence of a metastatic tumor. Samples of pleural fluid are used to analyze the morphologies of mesothelial cells and can typically be used to make a diagnosis between benignity and malignancy. Atypical pleural effusion samples are not easily identified as benign or malignant due to a lack of differentiable visual features, and such a problem has a significant influence in clinicians' decision making. In this paper, the goal is to develop a clinical decision support system (CDSS) using computer imaging and machine learning techniques for diagnosing atypical pleural effusion. The proposed approach involves four steps for analyzing slides of pleural effusion samples: image processing, feature measurement, feature selection, and classification. Processing and measurement of images produced a preliminary data set of 500 samples; each is described by 398 features. A genetic algorithm was applied for feature selection and identified a subset of 39 important features. The experimental results showed that the selected features can distinguish atypical nuclei as benign or malignant with a five-fold cross validation accuracy of 91%.

Publisher Attribution

2016 Industrial and Systems Engineering Research Conference/IIE Annual Conference & Expo 2016, At Anaheim, CA, USA

Download

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.