| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font
First page > Show document

Show document Help

Title:Identifikacija žensk za visokorizičen izvid po konizaciji z uporabo nevronskih mrež
Authors:ID Mlinarič, Marko (Author)
ID Takač, Iztok (Mentor) More about this mentor... New window
ID Repše Fokter, Alenka (Comentor)
Files:.pdf DOK_Mlinaric_Marko_2023.pdf (4,12 MB)
MD5: E7487FD96FEE84BD3590A3D1BD0E73B6
 
Language:Slovenian
Work type:Doctoral dissertation
Typology:2.08 - Doctoral Dissertation
Organization:MF - Faculty of Medicine
Abstract:Na svetu je rak materničnega vratu (RMV) četrti najpogostejši rak pri ženskah. V Sloveniji je bil pri ženskah leta 2017 na sedemnajstem mestu. Z ustreznim presejanjem, zgodnjim odkrivanjem predrakavih sprememb in njihovim zdravljenjem, ga je mogoče preprečiti. Metode umetne inteligence ('artificial intelligence – AI') bi lahko postale orodje, ki bi pripomoglo k rešitvi problema RMV. Z našo študijo smo želeli preveriti, ali lahko z umetnimi nevronskimi mrežami na podlagi dejavnikov tveganja za razvoj ploščatocelične intraepitelijske lezije (PIL) na materničnem vratu (MV) in RMV ter izvida zadnjega brisa materničnega vratu (BMV) napovemo, ali bo končni histološki izvid konusa PIL visoke stopnje (PIL-VS) oziroma RMV ali ne. Poleg nevronskih mrež smo preizkusili tudi nekatere druge klasifikacijske sisteme. Za analizo smo uporabili podatke 1475 pacientk, ki so imele narejeno konizacijo na Kliniki za ginekologijo in perinatologijo Univerzitetnega kliničnega centra Maribor v letih 1993–2005. Vse podatke smo anonimizirali. Uporabili smo metode za uravnoteženje manjšinskega in večinskega razreda. Za analizo smo oblikovali več baz, izvedli pa smo jo z odprtokodnim programskim paketom za podatkovno rudarjenje WEKA. Nevronske mreže so bile uspešnejše pri napovedovanju tveganih pacientk kot večinski algoritem. V naši študiji se je klasifikacijski algoritem Random Forest s sestavljeno metodo 'bagging' izkazal kot najuspešnejši in bi bil primeren za klinično uporabo.
Keywords:rak materničnega vratu, ploščatocelična intraepitelijska lezija visoke stopnje, umetna inteligenca, umetne nevronske mreže, napovedovanje tveganja
Place of publishing:Maribor
Year of publishing:2023
PID:20.500.12556/DKUM-82180 New window
COBISS.SI-ID:162081283 New window
Publication date in DKUM:24.08.2023
Views:456
Downloads:40
Metadata:XML DC-XML DC-RDF
Categories:MF
:
MLINARIČ, Marko, 2023, Identifikacija žensk za visokorizičen izvid po konizaciji z uporabo nevronskih mrež [online]. Doctoral dissertation. Maribor. [Accessed 13 April 2025]. Retrieved from: https://dk.um.si/IzpisGradiva.php?lang=eng&id=82180
Copy citation
  
Average score:
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
(0 votes)
Your score:Voting is allowed only for logged in users.
Share:Bookmark and Share


Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
Link:http://creativecommons.org/licenses/by-nc-nd/4.0/
Description:The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.
Licensing start date:26.07.2022

Secondary language

Language:English
Title:Identification of women with high risk histopatology after conisation by neural networks
Abstract:Cervical cancer is the fourth most common cancer in women globally. In Slovenia, it was the seventeenth most common cancer in women in 2017. Cervical cancer can be prevented with early detection and treatment of precancerous lesions. Artificial intelligence (AI), therefore, has the potential to be an important tool for eliminating the problem of cervical cancer. The aim of our study was to evaluate if artificial neural networks (ANN) can identify women who have high-grade final histopathology of the cone only on the basis of known risk factors for the development of cervical squamous intraepithelial lesion (SIL) and cancer, and last PAP smear result. Other classification algorithms were also tested. Data from 1475 patients who had conization at the Clinic for gynaecology and perinatology of University Clinical Centre Maribor from 1993-2005 was used for analysis. Data was anonymized. Methods to deal with imbalanced classes were used. Multiple databases were constructed for analysis with WEKA open-source program for data mining. Neural networks outperformed the majority algorithm in predicting high-risk patients. In our study, Random Forest algorithm with bagging method proved to be the best algorithm for the task and is suitable for clinical use.
Keywords:cervical cancer, high-grade squamous intraepithelial lesion, artificial intelligence, neural network, forecasting


Comments

Leave comment

You must log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica