National and Kapodistrian University of Athens
Η συμβολή μεθοδολογιών λήψης απόφασης στην αύξηση της διαγνωστικής ακρίβειας της κυτταρολογίας ενδομητρίου
Abstract
dc:descriptionAim: Objective of this study is to investigate the potential of a computerized system ofclassification and regression trees (CARTs) for the discrimination of benign frommalignant endometrial nuclei and lesions.Methods: 228 histologically confirmed liquid based cytological smears were collected,specifically: 117 benign cases, 66 malignant cases, 37 hyperplasias without atypiaand 8 cases of hyperplasia with atypia.From each case we measured about 100 nuclei using an image analysis system, intotal we collected 22.783 nuclei and we extracted nuclear morphometric featuresusing a custom image analysis system. Initially we performed feature selection andsubsequently we applied a logistic regression model that classified each nucleus asbenign or malignant. Based on the results of nucleus classification process weconstructed an algorithm to discriminate endometrium cases as benign or malignant.The nuclei from 50% of the cases (the training set) were used to construct a CARTmodel that was used for knowledge extraction. The nuclei from the remaining cases(test set) were used to evaluate the stability and performance of the CART onunknown data. Based on the results of the CART for nuclei classification we proposedtwo classification methods to discriminate cases as benign or malignant.Results: The logistic regression system had an overall accuracy for the classificationof endometrial nuclei equal to 83.02%, specificity 85.09% and sensitivity 77.01%. Forthe case classification the overall accuracy was 92.98%, specificity 92.86% andsensitivity 93.24%.The CART model had an overall accuracy for the classification of endometrial nuclei equal to 85%, specificity 90.68% and sensitivity 72.05%. Both methods for cases classification had similar performance: overall accuracy in the range 94-95%, specificity 95% and sensitivity 91-94%.Conclusion: The proposed computerized system can be applied for the classification of endometrial nuclei and lesions as outperformed the standard cytological diagnosis. This study highlights interesting diagnostic features of endometrial nuclear morphology and the proposed method can be a useful tool in the everyday practice of the cytological laboratory.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2015
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Zygouris, Dimitrios
- Ζυγούρης, Δημήτριος
Subjects
dc:subject × 18- Κυτταρολογία ενδομητρίου
- Τεχνητή νοημοσύνη
- Δένδρα ταξινόμησης
- Λογιστική παλινδρόμηση
- Κυτταρολογία υγρής μορφής
- Καρκίνος ενδομητρίου
- Endometrial cytology
- Artificial intelligence
- Classification and regression trees
- Logistic regression
- Liquid based cytology
- Endometrial cancer
- Ιατρική και Επιστήμες Υγείας
- Κλινική Ιατρική
- Ιατρική Βιοτεχνολογία
- Medical and Health Sciences
- Clinical Medicine
- Medical Biotechnology
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/41742