FINE: Improving time and precision of segmentation techniques for vertebral compression fractures in MRI

Author(s)

Ramos, Jonathan S. ; Cazzolato, Mirela T. ; Nogueira-Barbosa, Marcello H. ; Traina, Agma J. M. ; ACM

• Document type: Journal article
• Source: PROCEEDINGS OF THE 35TH ANNUAL ACM SYMPOSIUM ON APPLIED COMPUTING (SAC'20); v. N/A, p. 4-pg., 2020-01-01.

Abstract

Lower back pain is often related to spinal diseases. In particular, Vertebral Compression Fractures (VCFs) can impair mobility and compromise quality of life. In a Computer-Aided Diagnosis (CAD) context, the segmentation of VCFs is a challenging task due to non-homogeneous intensities within the same vertebral body. Semiautomatic segmentation methods have been employed to cope with this challenge. However, these methods require inside and outside annotation, which is not practical when analyzing a more significant number of exams. Aimed at minimizing the time spent on manual annotation, we proposed Fast INside Estimation (FINE), which automatically estimates the inside seeds based on the outside seeds. The experimental results with a representative dataset showed that FINE does not demand manual inside annotation, what the competitors methods do, and achieve higher Recall and Dice Score, on average, 97% and 96%, respectively. Higher Recall is particularly essential on features extraction and classification of VCFs. Therefore, FINE speeds up the manual annotation process while allowing more accurate semiautomatic segmentation. (AU)

• FAPESP's process: 18/24414-2 - A framework for integration of feature extraction techniques and complex databases for MIVisBD
• Grantee: Mirela Teixeira Cazzolato
• Support Opportunities: Scholarships in Brazil - Technical Training Program - Technical Training
• FAPESP's process: 16/17078-0 - Mining, indexing and visualizing Big Data in clinical decision support systems (MIVisBD)
• Grantee: Agma Juci Machado Traina
• Support Opportunities: Research Projects - Thematic Grants
• FAPESP's process: 17/23780-2 - Content-based retrieval of medical images to aid the clinical decision using radiomics
• Grantee: Jonathan da Silva Ramos
• Support Opportunities: Scholarships in Brazil - Doctorate