Multichannel fluorescence microscopy images CTC detection: A deep learning approach

The analysis of Circulating Tumour Cells (CTCs) provides continuous and real-time information for disease monitoring of cancer patients in a minimally invasive way, beeing established as predictive biomarkers by clinical experts. However, the identification and enumeration of these cells require a v...

Full description

Saved in:

Bibliographic Details
Main Authors:	Calvo-Almeida, Shaila, Serrano-Llabrés, Ignacio, Cal-González, Victoria M., Piairo, Paulina, Pires, Liliana R., Diéguez, Lorena, González-Castro, Lorena
Format:	Conference Proceeding
Language:	eng
Subjects:	Artificial neural networks Biomarkers Computer vision Enumeration Fluorescence Machine learning Medical imaging
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	The analysis of Circulating Tumour Cells (CTCs) provides continuous and real-time information for disease monitoring of cancer patients in a minimally invasive way, beeing established as predictive biomarkers by clinical experts. However, the identification and enumeration of these cells require a very complex and time-consuming manual p rocess. Thus the need arises to automate this detection and enumeration process in a way that facilitates this highly operator-dependent analysis. Computer Vision techniques have been successfully applied in this type of task, using Convolutional Neural Networks (CNNs) to achieve good performance in the detection of these type of cells. In this work, we present the training and evaluation of several CNNs models focused on the detection of CTCs from multi-channel images obtained from a fluorescence m icroscope. The best model obtained in this study is a trained DetectoRS CNN architecture that has been able to identify 23 out of 24 CTCs in the test dataset, offering a recall of 95.83%.
ISSN:	0094-243X 1551-7616