Discriminative bag-of-cells for imaging-genomics

Benjamin Chidester, Minh N Do, Jian Ma

Research output: Contribution to journalConference article

Abstract

Connecting genotypes to image phenotypes is crucial for a comprehensive understanding of cancer. To learn such connections, new machine learning approaches must be developed for the better integration of imaging and genomic data. Here we propose a novel approach called Discriminative Bag-of-Cells (DBC) for predicting genomic markers using imaging features, which addresses the challenge of summarizing histopathological images by representing cells with learned discriminative types, or codewords. We also developed a reliable and efficient patch-based nuclear segmentation scheme using convolutional neural networks from which nuclear and cellular features are extracted. Applying DBC on TCGA breast cancer samples to predict basal subtype status yielded a class-balanced accuracy of 70% on a separate test partition of 213 patients. As data sets of imaging and genomic data become increasingly available, we believe DBC will be a useful approach for screening histopathological images for genomic markers. Source code of nuclear segmentation and DBC are available at: https://github.com/bchidest/DBC.

Original languageEnglish (US)
Pages (from-to)319-330
Number of pages12
JournalPacific Symposium on Biocomputing
Volume0
Issue number212669
DOIs
StatePublished - Jan 1 2018
Event23rd Pacific Symposium on Biocomputing, PSB 2018 - Kohala Coast, United States
Duration: Jan 3 2018Jan 7 2018

Fingerprint

Genomics
Imaging techniques
Learning systems
Screening
Neural networks
Genotype
Breast Neoplasms
Phenotype
Neoplasms

Keywords

  • Computational pathology
  • Histopathological image analysis
  • Imaging-genomics

ASJC Scopus subject areas

  • Medicine(all)

Cite this

Discriminative bag-of-cells for imaging-genomics. / Chidester, Benjamin; Do, Minh N; Ma, Jian.

In: Pacific Symposium on Biocomputing, Vol. 0, No. 212669, 01.01.2018, p. 319-330.

Research output: Contribution to journalConference article

Chidester, Benjamin ; Do, Minh N ; Ma, Jian. / Discriminative bag-of-cells for imaging-genomics. In: Pacific Symposium on Biocomputing. 2018 ; Vol. 0, No. 212669. pp. 319-330.
@article{a2418c0f061d4a52b5d3f67ee3af1f69,
title = "Discriminative bag-of-cells for imaging-genomics",
abstract = "Connecting genotypes to image phenotypes is crucial for a comprehensive understanding of cancer. To learn such connections, new machine learning approaches must be developed for the better integration of imaging and genomic data. Here we propose a novel approach called Discriminative Bag-of-Cells (DBC) for predicting genomic markers using imaging features, which addresses the challenge of summarizing histopathological images by representing cells with learned discriminative types, or codewords. We also developed a reliable and efficient patch-based nuclear segmentation scheme using convolutional neural networks from which nuclear and cellular features are extracted. Applying DBC on TCGA breast cancer samples to predict basal subtype status yielded a class-balanced accuracy of 70{\%} on a separate test partition of 213 patients. As data sets of imaging and genomic data become increasingly available, we believe DBC will be a useful approach for screening histopathological images for genomic markers. Source code of nuclear segmentation and DBC are available at: https://github.com/bchidest/DBC.",
keywords = "Computational pathology, Histopathological image analysis, Imaging-genomics",
author = "Benjamin Chidester and Do, {Minh N} and Jian Ma",
year = "2018",
month = "1",
day = "1",
doi = "10.1142/9789813235533_0030",
language = "English (US)",
volume = "0",
pages = "319--330",
journal = "Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing",
issn = "2335-6936",
number = "212669",

}

TY - JOUR

T1 - Discriminative bag-of-cells for imaging-genomics

AU - Chidester, Benjamin

AU - Do, Minh N

AU - Ma, Jian

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Connecting genotypes to image phenotypes is crucial for a comprehensive understanding of cancer. To learn such connections, new machine learning approaches must be developed for the better integration of imaging and genomic data. Here we propose a novel approach called Discriminative Bag-of-Cells (DBC) for predicting genomic markers using imaging features, which addresses the challenge of summarizing histopathological images by representing cells with learned discriminative types, or codewords. We also developed a reliable and efficient patch-based nuclear segmentation scheme using convolutional neural networks from which nuclear and cellular features are extracted. Applying DBC on TCGA breast cancer samples to predict basal subtype status yielded a class-balanced accuracy of 70% on a separate test partition of 213 patients. As data sets of imaging and genomic data become increasingly available, we believe DBC will be a useful approach for screening histopathological images for genomic markers. Source code of nuclear segmentation and DBC are available at: https://github.com/bchidest/DBC.

AB - Connecting genotypes to image phenotypes is crucial for a comprehensive understanding of cancer. To learn such connections, new machine learning approaches must be developed for the better integration of imaging and genomic data. Here we propose a novel approach called Discriminative Bag-of-Cells (DBC) for predicting genomic markers using imaging features, which addresses the challenge of summarizing histopathological images by representing cells with learned discriminative types, or codewords. We also developed a reliable and efficient patch-based nuclear segmentation scheme using convolutional neural networks from which nuclear and cellular features are extracted. Applying DBC on TCGA breast cancer samples to predict basal subtype status yielded a class-balanced accuracy of 70% on a separate test partition of 213 patients. As data sets of imaging and genomic data become increasingly available, we believe DBC will be a useful approach for screening histopathological images for genomic markers. Source code of nuclear segmentation and DBC are available at: https://github.com/bchidest/DBC.

KW - Computational pathology

KW - Histopathological image analysis

KW - Imaging-genomics

UR - http://www.scopus.com/inward/record.url?scp=85048094878&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048094878&partnerID=8YFLogxK

U2 - 10.1142/9789813235533_0030

DO - 10.1142/9789813235533_0030

M3 - Conference article

C2 - 29218893

AN - SCOPUS:85048094878

VL - 0

SP - 319

EP - 330

JO - Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing

JF - Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing

SN - 2335-6936

IS - 212669

ER -