An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma

Andrew S. Jones, Azzam G F Taktak, Timothy R. Helliwell, John E. Fenton, Martin A. Birchall, David J. Husband, Anthony C. Fisher

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

The accepted method of modelling and predicting failure/survival, Cox's proportional hazards model, is theoretically inferior to neural network derived models for analysing highly complex systems with large datasets. A blinded comparison of the neural network versus the Cox's model in predicting survival utilising data from 873 treated patients with laryngeal cancer. These were divided randomly and equally into a training set and a study set and Cox's and neural network models applied in turn. Data were then divided into seven sets of binary covariates and the analysis repeated. Overall survival was not significantly different on Kaplan-Meier plot, or with either test model. Although the network produced qualitatively similar results to Cox's model it was significantly more sensitive to differences in survival curves for age and N stage. We propose that neural networks are capable of prediction in systems involving complex interactions between variables and non-linearity.

Original languageEnglish (US)
Pages (from-to)541-547
Number of pages7
JournalEuropean Archives of Oto-Rhino-Laryngology
Volume263
Issue number6
DOIs
StatePublished - Jun 2006
Externally publishedYes

Fingerprint

Squamous Cell Carcinoma
Proportional Hazards Models
Neural Networks (Computer)
Survival
Laryngeal Neoplasms

Keywords

  • Artificial intelligence
  • Chaos
  • Complex systems modeling
  • Laryngeal carcinoma

ASJC Scopus subject areas

  • Otorhinolaryngology

Cite this

Jones, A. S., Taktak, A. G. F., Helliwell, T. R., Fenton, J. E., Birchall, M. A., Husband, D. J., & Fisher, A. C. (2006). An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma. European Archives of Oto-Rhino-Laryngology, 263(6), 541-547. https://doi.org/10.1007/s00405-006-0021-2

An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma. / Jones, Andrew S.; Taktak, Azzam G F; Helliwell, Timothy R.; Fenton, John E.; Birchall, Martin A.; Husband, David J.; Fisher, Anthony C.

In: European Archives of Oto-Rhino-Laryngology, Vol. 263, No. 6, 06.2006, p. 541-547.

Research output: Contribution to journalArticle

Jones, AS, Taktak, AGF, Helliwell, TR, Fenton, JE, Birchall, MA, Husband, DJ & Fisher, AC 2006, 'An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma', European Archives of Oto-Rhino-Laryngology, vol. 263, no. 6, pp. 541-547. https://doi.org/10.1007/s00405-006-0021-2
Jones, Andrew S. ; Taktak, Azzam G F ; Helliwell, Timothy R. ; Fenton, John E. ; Birchall, Martin A. ; Husband, David J. ; Fisher, Anthony C. / An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma. In: European Archives of Oto-Rhino-Laryngology. 2006 ; Vol. 263, No. 6. pp. 541-547.
@article{7b61c58ac4d641e3919286f6742c0564,
title = "An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma",
abstract = "The accepted method of modelling and predicting failure/survival, Cox's proportional hazards model, is theoretically inferior to neural network derived models for analysing highly complex systems with large datasets. A blinded comparison of the neural network versus the Cox's model in predicting survival utilising data from 873 treated patients with laryngeal cancer. These were divided randomly and equally into a training set and a study set and Cox's and neural network models applied in turn. Data were then divided into seven sets of binary covariates and the analysis repeated. Overall survival was not significantly different on Kaplan-Meier plot, or with either test model. Although the network produced qualitatively similar results to Cox's model it was significantly more sensitive to differences in survival curves for age and N stage. We propose that neural networks are capable of prediction in systems involving complex interactions between variables and non-linearity.",
keywords = "Artificial intelligence, Chaos, Complex systems modeling, Laryngeal carcinoma",
author = "Jones, {Andrew S.} and Taktak, {Azzam G F} and Helliwell, {Timothy R.} and Fenton, {John E.} and Birchall, {Martin A.} and Husband, {David J.} and Fisher, {Anthony C.}",
year = "2006",
month = "6",
doi = "10.1007/s00405-006-0021-2",
language = "English (US)",
volume = "263",
pages = "541--547",
journal = "European Archives of Oto-Rhino-Laryngology",
issn = "0937-4477",
publisher = "Springer Verlag",
number = "6",

}

TY - JOUR

T1 - An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma

AU - Jones, Andrew S.

AU - Taktak, Azzam G F

AU - Helliwell, Timothy R.

AU - Fenton, John E.

AU - Birchall, Martin A.

AU - Husband, David J.

AU - Fisher, Anthony C.

PY - 2006/6

Y1 - 2006/6

N2 - The accepted method of modelling and predicting failure/survival, Cox's proportional hazards model, is theoretically inferior to neural network derived models for analysing highly complex systems with large datasets. A blinded comparison of the neural network versus the Cox's model in predicting survival utilising data from 873 treated patients with laryngeal cancer. These were divided randomly and equally into a training set and a study set and Cox's and neural network models applied in turn. Data were then divided into seven sets of binary covariates and the analysis repeated. Overall survival was not significantly different on Kaplan-Meier plot, or with either test model. Although the network produced qualitatively similar results to Cox's model it was significantly more sensitive to differences in survival curves for age and N stage. We propose that neural networks are capable of prediction in systems involving complex interactions between variables and non-linearity.

AB - The accepted method of modelling and predicting failure/survival, Cox's proportional hazards model, is theoretically inferior to neural network derived models for analysing highly complex systems with large datasets. A blinded comparison of the neural network versus the Cox's model in predicting survival utilising data from 873 treated patients with laryngeal cancer. These were divided randomly and equally into a training set and a study set and Cox's and neural network models applied in turn. Data were then divided into seven sets of binary covariates and the analysis repeated. Overall survival was not significantly different on Kaplan-Meier plot, or with either test model. Although the network produced qualitatively similar results to Cox's model it was significantly more sensitive to differences in survival curves for age and N stage. We propose that neural networks are capable of prediction in systems involving complex interactions between variables and non-linearity.

KW - Artificial intelligence

KW - Chaos

KW - Complex systems modeling

KW - Laryngeal carcinoma

UR - http://www.scopus.com/inward/record.url?scp=33745108811&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745108811&partnerID=8YFLogxK

U2 - 10.1007/s00405-006-0021-2

DO - 10.1007/s00405-006-0021-2

M3 - Article

C2 - 16767468

AN - SCOPUS:33745108811

VL - 263

SP - 541

EP - 547

JO - European Archives of Oto-Rhino-Laryngology

JF - European Archives of Oto-Rhino-Laryngology

SN - 0937-4477

IS - 6

ER -