Using logistic regression to improve the prognostic value of microarray gene expression data sets: Application to early-stage squamous cell carcinoma of the lung and triple negative breast carcinoma

David W. Mount, Charles W. Putnam, Sara M. Centouri, Ann M. Manziello, Ritu Pandey, Linda L Garland, Jesse D Martinez

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Background: Numerous microarray-based prognostic gene expression signatures of primary neoplasms have been published but often with little concurrence between studies, thus limiting their clinical utility. We describe a methodology using logistic regression, which circumvents limitations of conventional Kaplan Meier analysis. We applied this approach to a thrice-analyzed and published squamous cell carcinoma (SQCC) of the lung data set, with the objective of identifying gene expressions predictive of early death versus long survival in early-stage disease. A similar analysis was applied to a data set of triple negative breast carcinoma cases, which present similar clinical challenges. Methods. Important to our approach is the selection of homogenous patient groups for comparison. In the lung study, we selected two groups (including only stages I and II), equal in size, of earliest deaths and longest survivors. Genes varying at least four-fold were tested by logistic regression for accuracy of prediction (area under a ROC plot). The gene list was refined by applying two sliding-window analyses and by validations using a leave-one-out approach and model building with validation subsets. In the breast study, a similar logistic regression analysis was used after selecting appropriate cases for comparison. Results: A total of 8594 variable genes were tested for accuracy in predicting earliest deaths versus longest survivors in SQCC. After applying the two sliding window and the leave-one-out analyses, 24 prognostic genes were identified; most of them were B-cell related. When the same data set of stage I and II cases was analyzed using a conventional Kaplan Meier (KM) approach, we identified fewer immune-related genes among the most statistically significant hits; when stage III cases were included, most of the prognostic genes were missed. Interestingly, logistic regression analysis of the breast cancer data set identified many immune-related genes predictive of clinical outcome. Conclusions: Stratification of cases based on clinical data, careful selection of two groups for comparison, and the application of logistic regression analysis substantially improved predictive accuracy in comparison to conventional KM approaches. B cell-related genes dominated the list of prognostic genes in early stage SQCC of the lung and triple negative breast cancer.

Original languageEnglish (US)
Article number33
JournalBMC Medical Genomics
Volume7
Issue number1
DOIs
StatePublished - Jun 10 2014

Fingerprint

Squamous Cell Carcinoma
Logistic Models
Breast Neoplasms
Gene Expression
Lung
Genes
Regression Analysis
B-Lymphocytes
Triple Negative Breast Neoplasms
Datasets
Kaplan-Meier Estimate
Transcriptome
Patient Selection
Breast
Neoplasms

ASJC Scopus subject areas

  • Genetics(clinical)
  • Genetics
  • Medicine(all)

Cite this

Using logistic regression to improve the prognostic value of microarray gene expression data sets : Application to early-stage squamous cell carcinoma of the lung and triple negative breast carcinoma. / Mount, David W.; Putnam, Charles W.; Centouri, Sara M.; Manziello, Ann M.; Pandey, Ritu; Garland, Linda L; Martinez, Jesse D.

In: BMC Medical Genomics, Vol. 7, No. 1, 33, 10.06.2014.

Research output: Contribution to journalArticle

@article{26c267497b354639aba4a9f6977a5761,
title = "Using logistic regression to improve the prognostic value of microarray gene expression data sets: Application to early-stage squamous cell carcinoma of the lung and triple negative breast carcinoma",
abstract = "Background: Numerous microarray-based prognostic gene expression signatures of primary neoplasms have been published but often with little concurrence between studies, thus limiting their clinical utility. We describe a methodology using logistic regression, which circumvents limitations of conventional Kaplan Meier analysis. We applied this approach to a thrice-analyzed and published squamous cell carcinoma (SQCC) of the lung data set, with the objective of identifying gene expressions predictive of early death versus long survival in early-stage disease. A similar analysis was applied to a data set of triple negative breast carcinoma cases, which present similar clinical challenges. Methods. Important to our approach is the selection of homogenous patient groups for comparison. In the lung study, we selected two groups (including only stages I and II), equal in size, of earliest deaths and longest survivors. Genes varying at least four-fold were tested by logistic regression for accuracy of prediction (area under a ROC plot). The gene list was refined by applying two sliding-window analyses and by validations using a leave-one-out approach and model building with validation subsets. In the breast study, a similar logistic regression analysis was used after selecting appropriate cases for comparison. Results: A total of 8594 variable genes were tested for accuracy in predicting earliest deaths versus longest survivors in SQCC. After applying the two sliding window and the leave-one-out analyses, 24 prognostic genes were identified; most of them were B-cell related. When the same data set of stage I and II cases was analyzed using a conventional Kaplan Meier (KM) approach, we identified fewer immune-related genes among the most statistically significant hits; when stage III cases were included, most of the prognostic genes were missed. Interestingly, logistic regression analysis of the breast cancer data set identified many immune-related genes predictive of clinical outcome. Conclusions: Stratification of cases based on clinical data, careful selection of two groups for comparison, and the application of logistic regression analysis substantially improved predictive accuracy in comparison to conventional KM approaches. B cell-related genes dominated the list of prognostic genes in early stage SQCC of the lung and triple negative breast cancer.",
author = "Mount, {David W.} and Putnam, {Charles W.} and Centouri, {Sara M.} and Manziello, {Ann M.} and Ritu Pandey and Garland, {Linda L} and Martinez, {Jesse D}",
year = "2014",
month = "6",
day = "10",
doi = "10.1186/1755-8794-7-33",
language = "English (US)",
volume = "7",
journal = "BMC Medical Genomics",
issn = "1755-8794",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - Using logistic regression to improve the prognostic value of microarray gene expression data sets

T2 - Application to early-stage squamous cell carcinoma of the lung and triple negative breast carcinoma

AU - Mount, David W.

AU - Putnam, Charles W.

AU - Centouri, Sara M.

AU - Manziello, Ann M.

AU - Pandey, Ritu

AU - Garland, Linda L

AU - Martinez, Jesse D

PY - 2014/6/10

Y1 - 2014/6/10

N2 - Background: Numerous microarray-based prognostic gene expression signatures of primary neoplasms have been published but often with little concurrence between studies, thus limiting their clinical utility. We describe a methodology using logistic regression, which circumvents limitations of conventional Kaplan Meier analysis. We applied this approach to a thrice-analyzed and published squamous cell carcinoma (SQCC) of the lung data set, with the objective of identifying gene expressions predictive of early death versus long survival in early-stage disease. A similar analysis was applied to a data set of triple negative breast carcinoma cases, which present similar clinical challenges. Methods. Important to our approach is the selection of homogenous patient groups for comparison. In the lung study, we selected two groups (including only stages I and II), equal in size, of earliest deaths and longest survivors. Genes varying at least four-fold were tested by logistic regression for accuracy of prediction (area under a ROC plot). The gene list was refined by applying two sliding-window analyses and by validations using a leave-one-out approach and model building with validation subsets. In the breast study, a similar logistic regression analysis was used after selecting appropriate cases for comparison. Results: A total of 8594 variable genes were tested for accuracy in predicting earliest deaths versus longest survivors in SQCC. After applying the two sliding window and the leave-one-out analyses, 24 prognostic genes were identified; most of them were B-cell related. When the same data set of stage I and II cases was analyzed using a conventional Kaplan Meier (KM) approach, we identified fewer immune-related genes among the most statistically significant hits; when stage III cases were included, most of the prognostic genes were missed. Interestingly, logistic regression analysis of the breast cancer data set identified many immune-related genes predictive of clinical outcome. Conclusions: Stratification of cases based on clinical data, careful selection of two groups for comparison, and the application of logistic regression analysis substantially improved predictive accuracy in comparison to conventional KM approaches. B cell-related genes dominated the list of prognostic genes in early stage SQCC of the lung and triple negative breast cancer.

AB - Background: Numerous microarray-based prognostic gene expression signatures of primary neoplasms have been published but often with little concurrence between studies, thus limiting their clinical utility. We describe a methodology using logistic regression, which circumvents limitations of conventional Kaplan Meier analysis. We applied this approach to a thrice-analyzed and published squamous cell carcinoma (SQCC) of the lung data set, with the objective of identifying gene expressions predictive of early death versus long survival in early-stage disease. A similar analysis was applied to a data set of triple negative breast carcinoma cases, which present similar clinical challenges. Methods. Important to our approach is the selection of homogenous patient groups for comparison. In the lung study, we selected two groups (including only stages I and II), equal in size, of earliest deaths and longest survivors. Genes varying at least four-fold were tested by logistic regression for accuracy of prediction (area under a ROC plot). The gene list was refined by applying two sliding-window analyses and by validations using a leave-one-out approach and model building with validation subsets. In the breast study, a similar logistic regression analysis was used after selecting appropriate cases for comparison. Results: A total of 8594 variable genes were tested for accuracy in predicting earliest deaths versus longest survivors in SQCC. After applying the two sliding window and the leave-one-out analyses, 24 prognostic genes were identified; most of them were B-cell related. When the same data set of stage I and II cases was analyzed using a conventional Kaplan Meier (KM) approach, we identified fewer immune-related genes among the most statistically significant hits; when stage III cases were included, most of the prognostic genes were missed. Interestingly, logistic regression analysis of the breast cancer data set identified many immune-related genes predictive of clinical outcome. Conclusions: Stratification of cases based on clinical data, careful selection of two groups for comparison, and the application of logistic regression analysis substantially improved predictive accuracy in comparison to conventional KM approaches. B cell-related genes dominated the list of prognostic genes in early stage SQCC of the lung and triple negative breast cancer.

UR - http://www.scopus.com/inward/record.url?scp=84905226601&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84905226601&partnerID=8YFLogxK

U2 - 10.1186/1755-8794-7-33

DO - 10.1186/1755-8794-7-33

M3 - Article

C2 - 24916928

AN - SCOPUS:84905226601

VL - 7

JO - BMC Medical Genomics

JF - BMC Medical Genomics

SN - 1755-8794

IS - 1

M1 - 33

ER -