Evaluation of algorithms for automatic segmentation of SPEC/CT nuclear medicine images

• Grant number: 24/07137-6
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: October 01, 2024
• End date: September 30, 2025
• Field of knowledge: Health Sciences - Medicine - Medical Radiology
• Principal Investigator: Diana Rodrigues de Pina Miranda
• Grantee: João Pedro Papacidero Borges
• Host Institution: Faculdade de Medicina (FMB). Universidade Estadual Paulista (UNESP). Campus de Botucatu. Botucatu , SP, Brazil

Abstract

The segmentation and quantification of medical images through computational tools bring several benefits to clinical practice. There are various approaches to segmenting medical images today, ranging from the latest and most complex, which use machine learning and artificial intelligence, to classical methods such as threshold segmentation, which require less processing power and computational resources. However, there is no definitive method to address this problem, and the choice of the best segmentation technique should take into account the image modality and the type of information desired. Due to their characteristics, Nuclear Medicine images pose a challenge for the automatic segmentation process, and even today, human visual inspection is the gold standard for most procedures in this specialty. This research proposes to evaluate different algorithms for automatic segmentation available in free software and compare them with segmentation performed by trained observers to assess which algorithm is most suitable for this image modality. With this result, it will be possible to support the development of more accurate segmentation and quantification strategies in Nuclear Medicine, adding new tools for the evaluation of medical images and contributing to a more efficient and safe clinical practice for professionals and patients