Predicting the prognosis of Conservative Treatment in Acute Cholecystitis using an Artificial Neural Network
Introduction: In majority of the cases clinicians usually choose the conservative mode of treatment as the first line of treatment in case of acute cholecystitis. But studies have shown that in most of the cases patients do not show any significant improvement and eventually are referred to cholecystectomy. This delay in referral catalyses complications like septic conditions and development of gangrenous cholecystitis. Computational structures like artificial neural networks are excellent tools for predicting outcomes apriori. So a neural network was employed to predict the prognosis of conservative treatment mode in cholecystitis and avoid the otherwise imminent complications.
Materials and Methods: A neural network was developed and trained on clinically significant data from a set of 150 medical records pertaining to patients who presented themselves with diagnosis of acute cholecystitis at a tertiary care hospital. The performance of the network was tested on this training set and a separate validation set comprising of data from 100 such patients.
Results: The developed network demonstrated excellent prognostic capabilities with respect to the success/failure of conservative treatment in acute cholecystitis.
Conclusion: A neural network can predict the outcome of conservative treatment in cholecystitis with significant accuracy and could prove to be an indispensable tool to clinicians
2. Papi C, D' Ambrosio L, Capurso L. Timing of cholecystectomy for acute calculous cholecystitis: a meta-analysis. Am J Gasterenterol 2004; 99: 147-55.
3. Fagan SP, Awad SS, Rahwan K, et al. Prognostic factors for the development of gangrenous cholecystitis. Am J Surg 2003; 186: 481-5.
4. Baxt WG. Use of an artificial neural network for the diagnosis of myocardial infarction. Ann Intern Med 191; 115: 843-848.
5. Rutledge R, Fakhry S. The use of neural networking and polynomial regression to predict the outcome in the critically ill: a comparison with APACHE II[abstract]. Crit Care Med 1993; 21(suppl): S213.
6. Specht, D. A general regression neural network. IEEE Trans Neural Networks, 1991; 2:6, 568-576.
7. Weiss SM, Kulikowski CA. How to estimate true performance of a learning system. In: Computer systems that learn. Palo Alto, CA: Morgan Kaufmann, 1991; 17-49.
8. Bottaci L, Drew PJ, Hartley JE, et al. Artificial neural network applied to outcome prediction for colorectal cancer patients in separate institutions. Lancet 1997; 350: 469-472.