Advanced search
Start date

Development of radiographic quantitative markers for Precision Medicine regarding acute respiratory syndrome

Grant number: 20/14180-4
Support Opportunities:Scholarships in Brazil - Post-Doctoral
Effective date (Start): November 01, 2020
Effective date (End): March 06, 2022
Field of knowledge:Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
Principal Investigator:Agma Juci Machado Traina
Grantee:Karem Daiane Marcomini
Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
Associated research grant:20/07200-9 - Analyzing complex data from COVID-19 to support decision making and prognosis, AP.R


The new SARS-CoV-2 coronavirus pandemic has already confirmed over 40 million cases worldwide, accounting for more than 1,000,000 of deaths. In Brazil, the COVID-19 stage, named for the syndrome caused by SARS-CoV-2, has infected more than one million people and led to over 150 thousand deaths. Patients diagnosed with Severe Acute Respiratory Syndromes (SARS) are routinely subjected to medical imaging exams, especially chest radiography (CXR), to assess disease progression, such as COVID-19. However, the radiographic characteristics of SARS are very similar, making it difficult to proceed with clinical decision-making. Radiomics, in turn, can help the radiological assessment of SARS by associating quantitative characteristics of the images with the disease outcomes, such as etiology, severity, response to treatment, among others. Therefore, this project proposes to investigate biomarkers and radiomics models based on CXR to assist the diagnosis, prognosis and clinical decisions of COVID-19, as well as the therapeutic response of SARS. Different cohorts of patients with SARS will be used for the development and validation of the investigated methods. CXR images of patients will be automatically processed and lung lesions will be segmented by algorithms already developed. Patients will be characterized massively by quantitative characteristics of the segmented CXRs, including gray level histogram, texture matrices and transformed in the image frequency domain. Multivariate statistical analysis and state-of-the-art artificial intelligence techniques will be used in predictive modeling to increase radiomics performance. Predictive performance will be assessed by ROC curves, sensitivity and specificity measures and Kaplan-Meier curves for the probability of an event occurring. Performance metrics will be evaluated statistically by methods such as Mann-Whitney for diagnostic biomarkers, log-rank for prognostic indexes and DeLong for predictive models. Thus, the potential markers and radiomics models may play a fundamental role in supporting the clinical decisions of acute respiratory syndromes. (AU)

News published in Agência FAPESP Newsletter about the scholarship:
Articles published in other media outlets (0 total):
More itemsLess items

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)
MARCOMINI, KAREM DAIANE; CARDONA CARDENAS, DIEGO ARMANDO; MACHADO TRAINA, AGMA JUCI; KRIEGER, JOSE EDUARDO; GUTIERREZ, MARCO ANTONIO; DRUKKER, K; IFTEKHARUDDIN, KM. A deep learning approach for COVID-19 screening and localization on Chest X-Ray images. MEDICAL IMAGING 2022: COMPUTER-AIDED DIAGNOSIS, v. 12033, p. 9-pg., . (20/07200-9, 16/17078-0, 20/14180-4)
DE AGUIAR, ERIKSON J.; MARCOMINI, KAREM D.; QUIRINO, FELIPE A.; GUTIERREZ, MARCO A.; TRAINA, CAETANO, JR.; TRAINA, AGMA J. M.; DRUKKER, K; IFTEKHARUDDIN, KM. Evaluation of the impact of physical adversarial attacks on deep learning models for classifying covid cases. MEDICAL IMAGING 2022: COMPUTER-AIDED DIAGNOSIS, v. 12033, p. 7-pg., . (20/07200-9, 16/17078-0, 20/14180-4)

Please report errors in scientific publications list using this form.