University of Thessaly (UTH)
Η χρήση των τεχνητών νευρωνικών δικτύων για την εκτίμηση της εξωσωματικής λιθοτριψίας ως θεραπεία της λιθίασης του ουροποιητικού στη Θεσσαλία
Abstract
dc:descriptionExtracorporeal shockwave lithotripsy (ESWL), a medical method that attempts to partition urinary stones by external means without surgery, is a medical milestone that brought revolution in Urology.Its evaluation, as a method of choice for the treatment of lithiasis, was designed to be done in this thesis, by the development of an artificial neural network. To accomplish this goal, the medical records of 716 patients, who were treated by ESWL, were recorded and 2 main features were distinguished: the occurrence of complications at the end of the procedure (i.e. the safety of the method) and the successfulness of lithotripsy (i.e. the effectiveness of the method). We define as complications of the method, the occurrence of pain or infection in the patient. The effectiveness of the method is the result of ESWL, i.e. the successful defragmentation of the stone or not.Patients’ medical data were collected and entered into a database with the use of MS Excel, by the authorized and special trained personnel that is responsible and handles the lithotripter, in a specially designed area at the University Hospital of Larissa, under the guidance and supervision of the medical staff, in a period of a decade. Then, data were formatted for the purpose of this thesis, in order to come up in a compatible form by the computer programs used. Statistical analysis of data took place in the direction of univariate and multivariate analysis. There was an effort to find possible correlations between each input variable and the result and then, with the use of the statistically significant variables, to create a multivariate model. The importance of statistically significant variables was noticed during the analysis. The IBM SPSS statistical software was used.Data were inserted into an artificial neural network that was designed especially for this purpose. A unique algorithm that was programmed in MATLAB environment, trained the network with the use of relative functions and then evaluated it with propriate sets of patients’ data. The same procedure was followed for both the evaluation of safety and the evaluation of efficacy of lithotripsy. The behavior and the results of the network in evaluation set were evaluated in both cases, initially with input data, that were consisted of all categorial variables and then only with the statistically significant variables, that came out from the statistical analysis and showed correlation with the result (output values).From the training and evaluation results it became clear that the artificial neural network can provide valuable results, especially in the case of large amounts of data that include statistically significant variables.
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tsitsiflis, Athanasios
- Τσιτσιφλής, Αθανάσιος
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/53428