Cooperative strategy for a dynamic ensemble of classification models in clinical applications: the case of MRI vertebral compression fractures

Abstract

The performances have been also compared with those obtained with those of standard ensemble classification algorithms showing superior results.

Machine learning techniques for computer-aided classification of active inflammatory sacroiliitis in magnetic resonance imaging

“…Machine learning is a branch of artificial intelligence, which allows the extraction of meaningful patterns from examples [5,6]. The artificial intelligence approach has been widely used in medical image classification tasks, such as melanoma [7], discrimination of smoking status based on deep learning with MRI [8], classification of dermatological ulcers [9], evaluation of breast cancer [10], lung diseases [11,12], and vertebral compression fractures [13,14]. Computer-assisted analysis can be based on different approaches, such as statistical methods, instancebased analysis, decision trees, and artificial neural networks (ANNs).…”

BGrowth

“…Figure 1 shows an example of manual segmentation over five lumbar vertebral bodies (L1-L5). Computer-Aided Diagnosis (CAD) systems with automatic or semi-automatic segmentation methods are quite useful in this case, as long as they present fast results and delineation closer/similar to the manual segmentation (8,9). Hence, an accurate segmentation algorithm plays a significant role and can assist the medical specialist in surgical planning and evaluation of suitable treatments (10), for example, segmentation of Lung nodule (11,12) and cortical bone (13,14).…”

Analysis of medical images to support decision-making in the musculoskeletal field

“…A very recurrent disease among older adults is Vertebral Compression Fractures (VCF), which, in general, is caused by bone-weakening diseases. The literature reports the works (BARBIERI et al, 2015;CASTI et al, 2017;FRIGHETTO-PEREIRA et al, 2015;FRIGHETTO-PEREIRA et al, 2015;AZEVEDO-MARQUES et al, 2015;FRIGHETTO-PEREIRA et al, 2016a) that employ computational methods for VCF analysis from MRIs exams. The works used, unless stated otherwise, a database with 63 T1 weighted exams (detailed in Section 3.2.1).…”

“…The works can be divided into segmentation and classification: Segmentation: In general, VCFs segmentation is a challenging task due to non-homogeneous gray intensities within the same vertebral body, which hardens boundaries delineation. The literature presents the works (BARBIERI et al, 2015;CASTI et al, 2017) that dealt with the challenge of VCF segmentation. Both techniques used only one seed point (a mouse click) approximately in the center of the vertebral body in a single slice.…”

“…Classification: the works (FRIGHETTO-PEREIRA et al, 2015;FRIGHETTO-PEREIRA et al, 2015;AZEVEDO-MARQUES et al, 2015;FRIGHETTO-PEREIRA et al, 2016a) dealt with the classification of VCFs using manually segmented vertebral bodies in 2D (slices) exams. Casti et al (2017) proposed a cooperative strategy that exploited the dynamic contribution of the classification models, presenting an AUC of 94% using QDA classifier and considering the reference manual segmentation (ground-truth).…”