Predictive classification of correlated targets with application to detection of metastatic cancer using functional CT imaging

Yuan Wang; Brian P Hobbs; Jianhua Hu; Chaan S Ng; Kim-Anh Do

doi:10.1111/biom.12304

Back

Predictive classification of correlated targets with application to detection of metastatic cancer using functional CT imaging

Journal article

Open access

Peer reviewed

Predictive classification of correlated targets with application to detection of metastatic cancer using functional CT imaging

Yuan Wang, Brian P Hobbs, Jianhua Hu, Chaan S Ng and Kim-Anh Do

Biometrics, Vol.71(3), pp.792-802

09/2015

DOI: https://doi.org/10.1111/biom.12304

Handle:

https://hdl.handle.net/2376/112673

PMCID: PMC4575264

PMID: 25851056

Files and links (1)

url

https://doi.org/10.1111/biom.12304View

Published (Version of record) Open

Abstract

Data Interpretation, Statistical

Reproducibility of Results

Humans

Middle Aged

Tomography, X-Ray Computed - methods

Male

Neoplasm Metastasis - diagnostic imaging

Statistics as Topic

Models, Statistical

Liver Neoplasms - diagnostic imaging

Radiographic Image Interpretation, Computer-Assisted - methods

Angiography - methods

Computer Simulation

Sensitivity and Specificity

Bayes Theorem

Female

Perfusion computed tomography (CTp) is an emerging functional imaging modality that uses physiological models to quantify characteristics pertaining to the passage of fluid through blood vessels. Perfusion characteristics provide physiological correlates for neovascularization induced by tumor angiogenesis. Thus CTp offers promise as a non-invasive quantitative functional imaging tool for cancer detection, prognostication, and treatment monitoring. In this article, we develop a Bayesian probabilistic framework for simultaneous supervised classification of multivariate correlated objects using separable covariance. The classification approach is applied to discriminate between regions of liver that contain pathologically verified metastases from normal liver tissue using five perfusion characteristics. The hepatic regions tend to be highly correlated due to common vasculature. We demonstrate that simultaneous Bayesian classification yields dramatic improvements in performance in the presence of strong correlation among intra-subject units, yet remains competitive with classical methods in the presence of weak or no correlation.

Metrics

9 Record Views

Details

Title: Predictive classification of correlated targets with application to detection of metastatic cancer using functional CT imaging
Creators: Yuan Wang - Department of Biostatistics, University of Texas M.D. Anderson Cancer Center, Houston, Texas, U.S.A
Brian P Hobbs - Department of Biostatistics, University of Texas M.D. Anderson Cancer Center, Houston, Texas, U.S.A
Jianhua Hu - Department of Biostatistics, University of Texas M.D. Anderson Cancer Center, Houston, Texas, U.S.A
Chaan S Ng - Department of Diagnostic Radiology, University of Texas M.D. Anderson Cancer Center, Houston, Texas, U.S.A
Kim-Anh Do - Department of Biostatistics, University of Texas M.D. Anderson Cancer Center, Houston, Texas, U.S.A
Publication Details: Biometrics, Vol.71(3), pp.792-802
Academic Unit: Mathematics and Statistics, Department of
Publisher: United States
Grant note: R01-CA158113 / NCI NIH HHS R01 CA157458 / NCI NIH HHS P30 CA016672 / NCI NIH HHS P50-CA140388 / NCI NIH HHS R01 CA158113 / NCI NIH HHS P30-CA016672 / NCI NIH HHS P50 CA140388 / NCI NIH HHS R01-CA157458 / NCI NIH HHS
Identifiers: 99900547669801842
Language: English
Resource Type: Journal article