Neural networks in the detection of telangiectasias: evaluation of sclerotherapy results and comparison with manual analysis

• Grant number: 18/20855-4
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): June 01, 2019
• Effective date (End): July 31, 2020
• Field of knowledge: Health Sciences - Medicine - Surgery
• Principal Investigator: Matheus Bertanha
• Grantee: Rafael Guimarães Kanda
• Host Institution: Faculdade de Medicina (FMB). Universidade Estadual Paulista (UNESP). Campus de Botucatu. Botucatu , SP, Brazil

Abstract

Background: Telangiectasias are small, abnormally dilated veins and visible to the naked eye often associated with important aesthetic complaints. Chemical sclerotherapy is the treatment of choice and its efficacy is determined by the visual elimination of treated vessels, usually assessed by qualitative photographic analysis. This technique is marked by a subjectivity that prevents the elaboration of precise studies about the extension of vessels and effectiveness of the therapies used, being an open area for new methodological approaches. The use of machine learning with neural networks (deep learning) has become a major trend in data analysis, with growing applications in the medical field. In search of a more reliable method of evaluation, we hypothesized that telangiectasias can be studied through neural networks. Objective: To explore the efficacy of computerized analysis with neural networks in the detection of telangiectasias compared with manual method in the quantification of treatment by sclerotherapy. Methods: from pre and post-sclerotherapy results of 98 patients submitted to two different treatments, with 0.2% polidocanol + 70% Glucose (from previous studies on telangiectasia treatment of lower limbs). Group 1) versus 75% pure Glucose (Group 2). The photographs of the participants remain under our control and will be analyzed again, now with the use of neural networks. These results will be compared with those previously obtained in order to determine the applicability of Deep Learning in the evaluation of telangiectasias.