Estimation of bladder muscle invasion in transitional cell carcinoma by using artificial neural networks: A study based on prebiopsy imaging findings

  • Authors:

  • Mahmut Tokmakçı;Nuri Erdogan;Nurettin Sahin;Hulya Akgun;Oguz Ekmekcioglu

  • Affiliations:

  • Erciyes University, Engineering Faculty, Department of Electronics Engineering, 38039 Kayseri, Turkey;Erciyes University, Faculty of Medicine, Department of Radiology, 38039 Kayseri, Turkey;Erciyes University, Faculty of Medicine, Department of Urology, 38039 Kayseri, Turkey;Erciyes University, Faculty of Medicine, Department of Pathology, 38039 Kayseri, Turkey;Erciyes University, Faculty of Medicine, Department of Urology, 38039 Kayseri, Turkey

  • Venue:
  • Expert Systems with Applications: An International Journal
  • Year:
  • 2008

Quantified Score

Hi-index 12.05

Visualization

Abstract

Purpose: To identify the capability of ANNs in estimation of muscle invasive disease in transitional cell carcinomas (TCC). Materials and methods: In this study, we developed a MLP based ANN to detect muscle invasive disease in transitional cell carcinoma of bladder through the analysis of prebiopsy imaging data. The study includes 172 patients (116 males and 56 females; mean age, 63.92 years; range, 31-92 years) who had had the definitive diagnosis of Transitional cell carcinomas (TCC) based on biopsy results. Results: In the test group, 34 out of 35 cases were correctly classified by he MLP based Neural Network with only one false negative case. The sensitivity, specificity, positive predictive and negative predictive values calculated from the output data were 100%, 96.1%, 90%, and 100%, respectively. Conclusion: The proposed algorithm produced high sensitivity and specifity in predicting the histopathologic results, which shows that this method has a promising value in estimation of bladder muscle invasion in TCC.