Identification of women with high–grade histopathology results after conisation by artificial neural networks

Authors

  • Marko Mlinaric Zasebna ginekoloska ambulanta Marko Mlinaric, dr.med.
  • Miljenko Križmarič
  • Iztok Takač
  • Alenka Repše Fokter

Abstract

Background. The aim of this study was to evaluate if artificial neural networks can predict high–grade histopathology results after conisation from risk factors and their combinations in patients undergoing conisation because of pathological changes on uterine cervix.

 

Materials and methods.1475 patients who had conisation surgery at the University Clinic for Gynaecology and Obstetrics of University Clinical Centre Maribor from 1993–2005. We arranged database in different datasets to deal with unbalance data and enhance classification performance. Weka open–source software was used for analysis with artificial neural networks. Last PAP smear and risk–factors for development of cervical dysplasia and carcinoma were used as input and high dysplasia Yes/No as output result. 10–fold cross validation was used for defining training and holdout set for analysis.

 

Results. Baseline classification and multiple runs of artificial neural network on various risk factors settings were performed. We achieved 84.19% correct classifications, area under the curve 0.87, kappa 0.64, F–measure 0.884 and MCC 0.515 in model, where baseline prediction was 69.79%.

 

Conclusions. With artificial neural networks we were able to identify more patients who developed high–grade squamous intraepithelial lesion on final histopathology result of conisation as with baseline prediction but characteristics of 1475 patients who had conisation in years 1993–2005 at the University Clinical Centre Maribor did not allow reliable prediction of with artificial neural networks for every–day clinical practice.

Downloads

Published

2022-09-08

How to Cite

Mlinaric, M., Križmarič, M., Takač, I., & Repše Fokter, A. (2022). Identification of women with high–grade histopathology results after conisation by artificial neural networks. Radiology and Oncology, 56(3), 355–364. Retrieved from https://radioloncol.com/index.php/ro/article/view/3833

Issue

Section

Clinical oncology