Dataset Information

Evaluation of machine learning methods with Fourier Transform features for classifying ovarian tumors based on ultrasound images.

ABSTRACT: INTRODUCTION:Ovarian tumors are the most common diagnostic challenge for gynecologists and ultrasound examination has become the main technique for assessment of ovarian pathology and for preoperative distinction between malignant and benign ovarian tumors. However, ultrasonography is highly examiner-dependent and there may be an important variability between two different specialists when examining the same case. The objective of this work is the evaluation of different well-known Machine Learning (ML) systems to perform the automatic categorization of ovarian tumors from ultrasound images. METHODS:We have used a real patient database whose input features have been extracted from 348 images, from the IOTA tumor images database, holding together with the class labels of the images. For each patient case and ultrasound image, its input features have been previously extracted using Fourier descriptors computed on the Region Of Interest (ROI). Then, four ML techniques are considered for performing the classification stage: K-Nearest Neighbors (KNN), Linear Discriminant (LD), Support Vector Machine (SVM) and Extreme Learning Machine (ELM). RESULTS:According to our obtained results, the KNN classifier provides inaccurate predictions (less than 60% of accuracy) independently of the size of the local approximation, whereas the classifiers based on LD, SVM and ELM are robust in this biomedical classification (more than 85% of accuracy). CONCLUSIONS:ML methods can be efficiently used for developing the classification stage in computer-aided diagnosis systems of ovarian tumor from ultrasound images. These approaches are able to provide automatic classification with a high rate of accuracy. Future work should aim at enhancing the classifier design using ensemble techniques. Another ongoing work is to exploit different kind of features extracted from ultrasound images.

SUBMITTER: Martinez-Mas J

PROVIDER: S-EPMC6660116 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Evaluation of machine learning methods with Fourier Transform features for classifying ovarian tumors based on ultrasound images.

Martínez-Más José J Bueno-Crespo Andrés A Khazendar Shan S Remezal-Solano Manuel M Martínez-Cendán Juan-Pedro JP Jassim Sabah S Du Hongbo H Al Assam Hisham H Bourne Tom T Timmerman Dirk D

PloS one 20190726 7

<h4>Introduction</h4>Ovarian tumors are the most common diagnostic challenge for gynecologists and ultrasound examination has become the main technique for assessment of ovarian pathology and for preoperative distinction between malignant and benign ovarian tumors. However, ultrasonography is highly examiner-dependent and there may be an important variability between two different specialists when examining the same case. The objective of this work is the evaluation of different well-known Machi ...[more]

PMID: 31348783

Similar Datasets

Project description:Fourier transform infrared (FTIR) spectroscopy, a technology traditionally used in chemistry to determine the molecular composition of a wide range of sample types, has gained growing interest in microbial typing. It is based on the different vibrational modes of the covalent bonds between atoms of a given sample, as bacterial cells, induced by the absorption of infrared radiation. This technique has been largely used for the study of pathogenic species, especially in the clinical field, and has been proposed also for the typing at different subspecies levels. The high throughput, speed, low cost, and simplicity make FTIR spectroscopy an attractive technique also for industrial applications, in particular, for probiotics. The aim of this study was to compare FTIR spectroscopy with established genotyping methods, pulsed-field gel electrophoresis (PFGE), whole-genome sequencing (WGS), and multilocus sequence typing (MLST), in order to highlight the FTIR spectroscopy potential discriminatory power at strain level. Our study focused on bifidobacteria, an important group of intestinal commensals generally recognized as probiotics. For their properties in promoting and maintaining health, bifidobacteria are largely marketed by the pharmaceutical, food, and dairy industries. Strains belonging to Bifidobacterium longum subsp. longum and Bifidobacterium animalis subsp. lactis were taken into consideration together with some additional type strains. For B. longum subsp. longum, it was possible to discriminate the strains with all the methods used. Although two isolates were shown to be strictly phylogenetically related, constituting a unique cluster, based on PFGE, WGS, and MLST, no clustering was observed with FTIR. For B. animalis subsp. lactis group, PFGE, WGS, and MLST were non-discriminatory, and only one strain was easily distinguished. On the other hand, FTIR discriminated all the isolates one by one, and no clustering was observed. According to these results, FTIR analysis is not only equivalent to PFGE, WGS, and MLST, but also for some strains, in particular, for B. animalis subsp. lactis group, more informative, being able to differentiate strains not discernible with the other two methods based on phenotypic variations likely deriving from certain genetic changes. Fourier transform infrared spectroscopy has highlighted the possibility of using the cell surface as a kind of barcode making tracing strains possible, representing an important aspect in probiotic applications. Furthermore, this work constitutes the first investigation on bifidobacterial strain typing using FTIR spectroscopy.

Dataset Information

Evaluation of machine learning methods with Fourier Transform features for classifying ovarian tumors based on ultrasound images.

Publications

Evaluation of machine learning methods with Fourier Transform features for classifying ovarian tumors based on ultrasound images.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure