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

Abstract

Background: Currently, magnetic resonance imaging (MRI) is used to evaluate active inflammatory sacroiliitis related to axial spondyloarthritis (axSpA). The qualitative and semiquantitative diagnosis performed by expert radiologists and rheumatologists remains subject to significant intrapersonal and interpersonal variation. This encouraged us to use machine-learning methods for this task. Methods: In this retrospective study including 56 sacroiliac joint MRI exams, 24 patients had positive and 32 had negative findings for inflammatory sacroiliitis according to the ASAS group criteria. The dataset was randomly split with~80% (46 samples, 20 positive and 26 negative) as training and~20% as external test (10 samples, 4 positive and 6 negative). After manual segmentation of the images by a musculoskeletal radiologist, multiple features were extracted. The classifiers used were the Support Vector Machine, the Multilayer Perceptron (MLP), and the Instance-Based Algorithm, combined with the Relief and Wrapper methods for feature selection. Results: Based on 10-fold cross-validation using the training dataset, the MLP classifier obtained the best performance with sensitivity = 100%, specificity = 95.6% and accuracy = 84.7%, using 6 features selected by the Wrapper method. Using the test dataset (external validation) the same MLP classifier obtained sensitivity = 100%, specificity = 66.7% and accuracy = 80%. Conclusions: Our results show the potential of machine learning methods to identify SIJ subchondral bone marrow edema in axSpA patients and are promising to aid in the detection of active inflammatory sacroiliitis on MRI STIR sequences. Multilayer Perceptron (MLP) achieved the best results.

Evaluating the Impact of High Intensity Interval Training on Axial Psoriatic Arthritis Based on MR Images

“…Mean values of textural features were different in voxels from BME compared to voxels from healthy bone marrow ( Table 4 ). These observed differences are consistent with a previous study, where textural features of MR images have been applied in machine learning to classify active inflammation in sacroiliac joints [ 35 ]. Choices of textural features are also important for successful tissue discrimination [ 30 ].…”

“…When discriminating the post-radiation lesions edema, fatty conversion, and hemorrhage, Romanos et al found that GLCM textural features comprised four out of five features in the optimal design of the classification scheme [ 48 ]. Sacroiliitis could be classified based on extracted features from STIR MR images and machine learning [ 35 ]. The features’ maximum pixel values and LH components from the two-level Haar wavelet decompositions, which depict horizontal traits of the image, were important to discriminate instances.…”

“…Additional information from MR images can be quantified by extracting spatial variations of grey-level intensity in an image [ 29 ]. Such a texture analysis may be more sensitive to change than visual inspection and has been applied in studies of BME [ 30 , 31 , 32 , 33 , 34 , 35 , 36 ]. It has demonstrated the potential to classify BME from MR images [ 32 , 33 , 34 ], and to detect changes in bone marrow after physical activity [ 36 ].…”

Radiomics Assessment of SPAIR and STIR MRI Sequences to Predict Axial and Peripheral Spondyloarthritis

“…After anonymization, this work selected six consecutive images from each MRI exam to be used in the analysis, as previously used in other studies [Faleiros et al 2020. Each image had the sacroiliac joints manually segmented by a musculoskeletal radiologist and processed by the warp geometric transform for feature extraction.…”

“…Besides that, this work also compared statistically the performance of neural networks modeled with features extracted from both MRI sequences of SPAIR and STIR on the prediction of SpA. Previous works have already recognized inflammatory patterns from sacroiliac joints to predict SpA, but they used a single MRI sequence (either STIR or SPAIR), and hence, did not compare the performances of different imaging acquisition protocols [Faleiros et al 2020].…”