Prediction of survival in patients with esophageal carcinoma using artificial neural networks

Fumiaki Sato, Yutaka Shimada, Florin Selaru, David Shibata, Masato Maeda, Go Watanabe, Yuriko Mori, Sanford A. Stass, Masayuki Imamura, Stephen Meltzer

Research output: Contribution to journalArticle

Abstract

BACKGROUND. Accurate estimation of outcome in patients with malignant disease is an important component of the clinical decision-making process. To create a comprehensive prognostic model for esophageal carcinoma, artificial neural networks (ANNs) were applied to the analysis of a range of patient-related and tumor-related variables. METHODS. Clinical and pathologic data were collected from 418 patients with esophageal carcinoma who underwent resection with curative intent. A data base that included 199 variables was constructed. Using ANN-based sensitivity analysis, the optimal combination of variables was determined to allow creation of a survival prediction model. The accuracy (area under the receiver operator characteristic curve [AUR]) of this ANN model subsequently was compared with the accuracy of the conventional statistical technique: linear discriminant analysis (LDA). RESULTS. The optimal ANN models for predicting outcomes at 1 year and 5 years consisted of 65 variables (AUR = 0.883) and 60 variables (AUR = 0.884), respectively. These filtered, optimal data sets were significantly more accurate (P <0.0001) than the original data set of 199 variables. The majority of ANN models demonstrated improved accuracy compared with corresponding LDA models for 1-year and 5-year survival predictions. Furthermore, ANN models based on the optimal data set were superior predictors of survival compared with a model based solely on TNM staging criteria (P <0.0001). CONCLUSIONS. ANNs can be used to construct a highly accurate prognostic model for patients with esophageal carcinoma. Sensitivity analysis based on ANNs is a powerful tool for seeking optimal data sets.

Original languageEnglish (US)
Pages (from-to)1596-1605
Number of pages10
JournalCancer
Volume103
Issue number8
DOIs
StatePublished - Apr 15 2005
Externally publishedYes

Fingerprint

Neural Networks (Computer)
Carcinoma
Survival
Discriminant Analysis
Neoplasm Staging
Databases
Datasets
Neoplasms

Keywords

  • Artificial neural network
  • Esophageal carcinoma
  • Prognosis
  • Sensitivity analysis
  • Survival

ASJC Scopus subject areas

  • Cancer Research
  • Oncology

Cite this

Prediction of survival in patients with esophageal carcinoma using artificial neural networks. / Sato, Fumiaki; Shimada, Yutaka; Selaru, Florin; Shibata, David; Maeda, Masato; Watanabe, Go; Mori, Yuriko; Stass, Sanford A.; Imamura, Masayuki; Meltzer, Stephen.

In: Cancer, Vol. 103, No. 8, 15.04.2005, p. 1596-1605.

Research output: Contribution to journalArticle

Sato, F, Shimada, Y, Selaru, F, Shibata, D, Maeda, M, Watanabe, G, Mori, Y, Stass, SA, Imamura, M & Meltzer, S 2005, 'Prediction of survival in patients with esophageal carcinoma using artificial neural networks', Cancer, vol. 103, no. 8, pp. 1596-1605. https://doi.org/10.1002/cncr.20938
Sato, Fumiaki ; Shimada, Yutaka ; Selaru, Florin ; Shibata, David ; Maeda, Masato ; Watanabe, Go ; Mori, Yuriko ; Stass, Sanford A. ; Imamura, Masayuki ; Meltzer, Stephen. / Prediction of survival in patients with esophageal carcinoma using artificial neural networks. In: Cancer. 2005 ; Vol. 103, No. 8. pp. 1596-1605.
@article{f7221365999f495d89c9a881e55ca73d,
title = "Prediction of survival in patients with esophageal carcinoma using artificial neural networks",
abstract = "BACKGROUND. Accurate estimation of outcome in patients with malignant disease is an important component of the clinical decision-making process. To create a comprehensive prognostic model for esophageal carcinoma, artificial neural networks (ANNs) were applied to the analysis of a range of patient-related and tumor-related variables. METHODS. Clinical and pathologic data were collected from 418 patients with esophageal carcinoma who underwent resection with curative intent. A data base that included 199 variables was constructed. Using ANN-based sensitivity analysis, the optimal combination of variables was determined to allow creation of a survival prediction model. The accuracy (area under the receiver operator characteristic curve [AUR]) of this ANN model subsequently was compared with the accuracy of the conventional statistical technique: linear discriminant analysis (LDA). RESULTS. The optimal ANN models for predicting outcomes at 1 year and 5 years consisted of 65 variables (AUR = 0.883) and 60 variables (AUR = 0.884), respectively. These filtered, optimal data sets were significantly more accurate (P <0.0001) than the original data set of 199 variables. The majority of ANN models demonstrated improved accuracy compared with corresponding LDA models for 1-year and 5-year survival predictions. Furthermore, ANN models based on the optimal data set were superior predictors of survival compared with a model based solely on TNM staging criteria (P <0.0001). CONCLUSIONS. ANNs can be used to construct a highly accurate prognostic model for patients with esophageal carcinoma. Sensitivity analysis based on ANNs is a powerful tool for seeking optimal data sets.",
keywords = "Artificial neural network, Esophageal carcinoma, Prognosis, Sensitivity analysis, Survival",
author = "Fumiaki Sato and Yutaka Shimada and Florin Selaru and David Shibata and Masato Maeda and Go Watanabe and Yuriko Mori and Stass, {Sanford A.} and Masayuki Imamura and Stephen Meltzer",
year = "2005",
month = "4",
day = "15",
doi = "10.1002/cncr.20938",
language = "English (US)",
volume = "103",
pages = "1596--1605",
journal = "Cancer",
issn = "0008-543X",
publisher = "John Wiley and Sons Inc.",
number = "8",

}

TY - JOUR

T1 - Prediction of survival in patients with esophageal carcinoma using artificial neural networks

AU - Sato, Fumiaki

AU - Shimada, Yutaka

AU - Selaru, Florin

AU - Shibata, David

AU - Maeda, Masato

AU - Watanabe, Go

AU - Mori, Yuriko

AU - Stass, Sanford A.

AU - Imamura, Masayuki

AU - Meltzer, Stephen

PY - 2005/4/15

Y1 - 2005/4/15

N2 - BACKGROUND. Accurate estimation of outcome in patients with malignant disease is an important component of the clinical decision-making process. To create a comprehensive prognostic model for esophageal carcinoma, artificial neural networks (ANNs) were applied to the analysis of a range of patient-related and tumor-related variables. METHODS. Clinical and pathologic data were collected from 418 patients with esophageal carcinoma who underwent resection with curative intent. A data base that included 199 variables was constructed. Using ANN-based sensitivity analysis, the optimal combination of variables was determined to allow creation of a survival prediction model. The accuracy (area under the receiver operator characteristic curve [AUR]) of this ANN model subsequently was compared with the accuracy of the conventional statistical technique: linear discriminant analysis (LDA). RESULTS. The optimal ANN models for predicting outcomes at 1 year and 5 years consisted of 65 variables (AUR = 0.883) and 60 variables (AUR = 0.884), respectively. These filtered, optimal data sets were significantly more accurate (P <0.0001) than the original data set of 199 variables. The majority of ANN models demonstrated improved accuracy compared with corresponding LDA models for 1-year and 5-year survival predictions. Furthermore, ANN models based on the optimal data set were superior predictors of survival compared with a model based solely on TNM staging criteria (P <0.0001). CONCLUSIONS. ANNs can be used to construct a highly accurate prognostic model for patients with esophageal carcinoma. Sensitivity analysis based on ANNs is a powerful tool for seeking optimal data sets.

AB - BACKGROUND. Accurate estimation of outcome in patients with malignant disease is an important component of the clinical decision-making process. To create a comprehensive prognostic model for esophageal carcinoma, artificial neural networks (ANNs) were applied to the analysis of a range of patient-related and tumor-related variables. METHODS. Clinical and pathologic data were collected from 418 patients with esophageal carcinoma who underwent resection with curative intent. A data base that included 199 variables was constructed. Using ANN-based sensitivity analysis, the optimal combination of variables was determined to allow creation of a survival prediction model. The accuracy (area under the receiver operator characteristic curve [AUR]) of this ANN model subsequently was compared with the accuracy of the conventional statistical technique: linear discriminant analysis (LDA). RESULTS. The optimal ANN models for predicting outcomes at 1 year and 5 years consisted of 65 variables (AUR = 0.883) and 60 variables (AUR = 0.884), respectively. These filtered, optimal data sets were significantly more accurate (P <0.0001) than the original data set of 199 variables. The majority of ANN models demonstrated improved accuracy compared with corresponding LDA models for 1-year and 5-year survival predictions. Furthermore, ANN models based on the optimal data set were superior predictors of survival compared with a model based solely on TNM staging criteria (P <0.0001). CONCLUSIONS. ANNs can be used to construct a highly accurate prognostic model for patients with esophageal carcinoma. Sensitivity analysis based on ANNs is a powerful tool for seeking optimal data sets.

KW - Artificial neural network

KW - Esophageal carcinoma

KW - Prognosis

KW - Sensitivity analysis

KW - Survival

UR - http://www.scopus.com/inward/record.url?scp=20144387857&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=20144387857&partnerID=8YFLogxK

U2 - 10.1002/cncr.20938

DO - 10.1002/cncr.20938

M3 - Article

C2 - 15751017

AN - SCOPUS:20144387857

VL - 103

SP - 1596

EP - 1605

JO - Cancer

JF - Cancer

SN - 0008-543X

IS - 8

ER -