Artificial neural networks in oncology
- Dr. G. Schwarzer, FDM, Institute of Medical Biometry and Medical Informatics
Prof. Dr. W. Vach, Department of Statistics, Odense University
Prof. Dr. M. Schumacher, FDM, Institute of Medical Biometry and Medical Informatics
- Summary of the project
The relationship between ANNs and statistical methods, especially logistic regression models, has been evaluated in a previous project (1,2). Simulation experiences have been conducted to illustrate that uncritical use of ANNs can lead to functions describing the probability of class membership that are far from plausible. This is due to the flexibility of ANNs which is often cited as an advantage; one could also argue that it must be seen as a major concern. Problems associated with the estimation of misclassification probabilities and the application of ANNs to survival data that often occur in oncology are examined (3). In addition to these general, methodological considerations a literature search on the application of ANNs in oncology has been conducted in order to identify and to report the most frequently occuring mistakes. Another literature search has been conducted in the specialized field of prostate cancer (4).
It is concluded that there is no evidence so far that applications of ANNs represent real progress in the field of diagnosis and prognosis in oncology.
- Publications
Neural networks and logistic regression. Part I, Computational Statistics and Data Analysis, 21, 1996, 661-82 [FDM-Preprint Nr. 2]
(2) Vach W, Roßner R, Schumacher M
Neural networks and logistic regression. Part II,
Computational Statistics and Data Analysis, 21, 1996, 683-701 [FDM-Preprint Nr. 2]
(3) Schwarzer G, Vach W, Schumacher M
On the misuses of artificial neural networks for prognostic and diagnostic classification in oncology,
Statistics in Medicine, 19, 2000, 541-561 [FDM-Preprint Nr. 46]
(4) Schwarzer G, Schumacher M
Artificial neural networks for diagnosis and prognosis in prostate cancer,
Seminars in Urological Oncology, 20, 2002, 89-95