Redes neurais convolucionais para a classificação de nódulos tireoidianos através de ultrassonografia

• Igor Machado Seixas PUC Minas
• Alexei Manso Correa Machado PUC Minas / UFMG

DOI: https://doi.org/10.5753/sbcas.2023.229645

Abstract

Premature detection of malignant nodes in the thyroid is critical for effective treatment. In this study a computer-aided diagnosis system is proposed to classify malign and benign nodes of the thyroid based on ultrasound images, as well as in the scale of the Thyroid Imaging Reporting and Data System (TI-RADS). The experiments implement 5 convolutional network and 3 support vector machines aplied to a public dataset. Preliminary results indicate the MobileNet as the best binary classifier with 89% of accuracy and the DenseNet121 with 56% of accuracy for the 4 TI-RADS categories.

References

J. Chi and E. Walia. Thyroid nodule classification in ultrasound images by fine-tuning deep convolutional neural network. J Digit Imaging, 30(4):477–486, 2017.

J. Ding, H. Cheng, C. Ning, J. Huang, and Y. Zhang. Quantitative measurement for thyroid cancer characterization based on elastography. J Ultrasound Med, 30(9):1259–66, 2011.

B. Gopinath and N. Shanthi. Support vector machine based diagnostic system for thyroid cancer using statistical texture features. Asian Pac J Cancer Prev, 14(1):97–102, 2013.

S. Guth, U. Theune, J. Aberle, A. Galach, and C. Bamberger. Very high prevalence of thyroid nodules detected by high frequency (13 mhz) ultrasound examination. Eur J Clin Invest, 39(8):699–706, 2009.

Y. Hang. Thyroid nodule classification in ultrasound images by fusion of conventional features and res-gan deep features. Journal of Healthcare Engineering, 22:9917538, 2021.

J. A. John and N. R. Draper. An alternative family of transformations. Journal of the Royal Statistical Society. Series C (Applied Statistics), 29(2):190–197, 1980.

J. Kwak, K. Han, and J. Yoon. Thyroid imaging reporting and data system for us features of nodules: a step in establishing better stratification of cancer risk. Radiology, 260(3): 892–9, 2011.

W. Li, P. Cao, D. Zhao, and J. Wang. Pulmonary nodule classification with deep convolutional neural networks on computed tomography images. Comput Math Methods Med, 216, 2016.

R. Lingam, M. Qarib, and N. Tolley. Evaluating thyroid nodules: predicting and selecting malignant nodules for fine-needle aspiration (fna) cytology. Insights Imaging, 4:617–624, 2013.

O. Moussa, H. Khachnaoui, R. Guetari, and N. Khlifa. Thyroid nodules classification and diagnosis in ultrasound images using fine-tuning deep convolutional neural network. Int J Imaging Syst Technol, 30(1):185–195, 2020.

V. Park, K. Han, and Y. Seong. Diagnosis of thyroid nodules: Performance of a deep learning convolutional neural network model vs. radiologists. Scientific Reports, 9, 2019.

L. Pedraza. An open access thyroid ultrasound-image database. Proceedings of the SPIE, 9287, 2015.

F. A. Spanhol, L. S. Oliveira, C. Petitjean, and L. Heutte. Breast cancer histopathological image classification using convolutional neural networks. In 2016 International Joint Conference on Neural Networks (IJCNN), pages 2560–2567, 2016.

M. Stib and I. Pan. Thyroid nodule malignancy risk stratification using a convolutional neural network. Ultrasound Q., 36(2), 2020.