Artificial neural networks and survival prediction in ovarian carcinoma

Eur J Gynaecol Oncol. 2000;21(6):583-4.


The standard use of known survival predictors for ovarian cancer in clinical practice are primarily based on disease stage. This does not permit a real individualization of a patient's potential outcome. This study assessed the value of neural networks to refine the prediction of survival based only on information gleaned at primary surgery. The possibility exists that such methods may permit further elucidation of outcome and influence management.

MeSH terms

  • England / epidemiology
  • Female
  • Humans
  • Neural Networks, Computer*
  • Ovarian Neoplasms / diagnosis*
  • Ovarian Neoplasms / mortality*
  • Predictive Value of Tests
  • Prognosis
  • ROC Curve
  • Survival Analysis