Computational vision and pattern recognition for radionics biomarkers identification

• Grant number: 18/07765-6
• Support type: Research Grants - Visiting Researcher Grant - International
• Duration: September 05, 2018 - September 23, 2018
• Field of knowledge: Engineering - Biomedical Engineering
• Principal Investigator: Paulo Mazzoncini de Azevedo Marques
• Grantee: Paulo Mazzoncini de Azevedo Marques
• Visiting researcher: Rangaraj Mandayam Rangayyan
• Visiting researcher institution: University of Calgary, Canada
• Home Institution: Faculdade de Medicina de Ribeirão Preto (FMRP). Universidade de São Paulo (USP). Ribeirão Preto , SP, Brazil
• Associated research grant: 16/17078-0 - Mining, indexing and visualizing Big Data in clinical decision support systems (MIVisBD), AP.TEM

Abstract

A recent and important advance in the quantitative analysis of images is the concept of Radiomics. Radiomics is an emerging field of radiology that is based on the conversion of image data into a characteristic space of high dimension, using different characteristic extraction algorithms. The rationale for radiomics is that a broader characterization of the underlying tumor phenotypes can be obtained by extracting a large number of imaging features, which may correlate with specific clinical outcomes. The extraction of quantitative characteristics of medical images and their use for pattern recognition and decision support contrasts with a more traditional and still common radiology practice based (almost) exclusively on visual interpretation. Although the concept of radiomics is a natural extension of the concept of computer-aided diagnosis (CAD), there are significant differences between them. CAD systems are targeted to deliver a unique response (a second opinion) to the specialist. Radiomics, in turn, combining data extracted from images with other patient characteristics, as available, to increase the power of decision support models. Quantitative features based on the gray levels intensity, shape, texture, size and volume extracted from the images can, for example, provide information about the phenotype of the microenvironment of a tumor, which are different from those obtained from clinical reports, laboratory test results or obtained by genomic or proteomic methods. These attributes, combined with other information, can correlate with clinical findings, allowing for evidence-based decision making. In other words, radiomics presents the potential to provide biomarkers based on images that allow the diagnosis and definition of prognosis, prediction of response to treatments and monitoring the state of a given disease. In this context, this project proposes the investigation of techniques of computer vision and pattern recognition for the identification of radiomics biomarkers that support the diagnosis, prognosis and therapeutic decision making in the evaluation and treatment of cancer and autoimmune rheumatic diseases. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

FALEIROS, MATHEUS CALIL; NOGUEIRA-BARBOSA, MARCELLO HENRIQUE; DALTO, VITOR FAEDA; FERREIRA JUNIOR, JOSE RANIERY; MAGALHAES TENORIO, ARIANE PRISCILLA; LUPPINO-ASSAD, RODRIGO; LOUZADA-JUNIOR, PAULO; RANGAYYAN, RANGARAJ MANDAYAM; DE AZEVEDO-MARQUES, PAULO MAZZONCINI. Machine learning techniques for computer-aided classification of active inflammatory sacroiliitis in magnetic resonance imaging. ADVANCES IN RHEUMATOLOGY, v. 60, n. 1 MAY 7 2020. Web of Science Citations: 1.

MATHEUS CALIL FALEIROS; MARCELLO HENRIQUE NOGUEIRA-BARBOSA; VITOR FAEDA DALTO; JOSÉ RANIERY FERREIRA JÚNIOR; ARIANE PRISCILLA MAGALHÃES TENÓRIO; RODRIGO LUPPINO-ASSAD; PAULO LOUZADA JUNIOR; RANGARAJ MANDAYAM RANGAYYAN; PAULO MAZZONCINI DE AZEVEDO-MARQUES. Machine learning techniques for computer-aided classification of active inflammatory sacroiliitis in magnetic resonance imaging. ADVANCES IN RHEUMATOLOGY, v. 60, p. -, 2020.