Back to search

National and Kapodistrian University of Athens

Η αξία μεθοδολογιών λήψης απόφασης στην ανάλυση της διαγνωστικής ακρίβειας της κυτταρολογίας ενδομητρίου

Abstract

dc:description

OBJECTIVE: Endometrial cytology could be an alternative perspective for the diagnosis of endometrial cancer. Objective of this study is to investigate the potential of an Artificial Neural Network (ANN) based on Multi Layer Perceptron (MPL), for the discrimination of benign from malignant endometrial nuclei and lesions. Secondary it is examined the potential risk factors for strenuousness in endometrial cytology sampling. STUDY DESIGN: 416 histological confirmed liquid based cytological smears were collected, specifically: 117 within normal limits cases, 66 malignant cases, 37 hyperplasias without atypia and 8 cases of hyperplasia with atypia. From each case were extracted nuclear morphometric features from about 90 nuclei using a custom image analysis system. Subsequently nuclei from 50% of the cases were used to train an MPL ANN model that classified each nucleus as benign or malignant. The remaining data from the 50% of the cases were used to evaluate the ANN performance and stability on unknown data. Based on the results of nucleus classification process are proposed two algorithms to discriminate endometrium cases as benign or malignant. A subgroup of 181 women participated in the secondary study. Strenuousness in obtaining the sample was graded into a five-level scale-score. Various parameters were assessed in association with the strenuousness score. Multivariate ordinal logistic regression analysis was performed. RESULTS: The ANN-MPL had an overall accuracy for the classification of endometrial nuclei equal to 83.02%, specificity 85.09% and sensitivity 77.01%. For the case classification the overall accuracy was 92.98%, specificity 92.86% and sensitivity 93.24%. As far as the secondary study is concerned, postmenopausal status (adjusted OR=2.63, 95%CI: 1.52-4.56, p=0.001) and previous invasive/surgical procedures in the cervix (adjusted OR=2.15, 95%CI: 1.10-4.24, p=0.026) were associated with higher strenuousness score. Participants’ age at sampling, phase of the menstrual cycle, endometrial thickness, obesity, current hormonal use and the reproductive history of women were not significantly associated with the strenuousness of the procedure. CONCLUSION: Computerized systems based on ANNs have the potential to be used for the classification of endometrial nuclei and lesions. Increased difficulty during endometrial sampling is noted in postmenopausal women, and in patients with previous surgical procedures in the cervix.

Degree

thesis:*
Grantor dc:publisher
National and Kapodistrian University of Athens
Year dc:date
2016

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Makris, Georgios-Marios
  • Μακρής, Γεώργιος-Μάριος

Subjects

dc:subject × 10

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/37713

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Makris, Georgios-Marios; Μακρής, Γεώργιος-Μάριος. Η αξία μεθοδολογιών λήψης απόφασης στην ανάλυση της διαγνωστικής ακρίβειας της κυτταρολογίας ενδομητρίου. National and Kapodistrian University of Athens, 2016. http://hdl.handle.net/10442/hedi/37713