Textural analysis of MR images as an additional evaluation tool of Parotid Glands in Sjögren's Syndrome in children and adolescents - initial findings

Abstract

Abstract Magnetic Resonance Imaging (MRI) is often a reference imaging method in soft tissue pathologies, particularly of the head and neck region. A significant increase in MRI popularity is observed, especially concerning salivary gland pathologies like neoplasm or chronic inflammatory processes of salivary glands like in Sjogren syndrome. The development of MRI techniques like MR sialography set a new non-invasive method alternative to classical sialography, especially in advanced cases. Another rising MRI technique - texture analysis - appears to be a promising tool in assessing structural changes in parenchymal organs and can be an additional parameter for assessing the inflammatory process in salivary glands. This research aimed to explore the correlation between apparent diffusion coefficient (ADC), based on T2 weighted images MRI sialography of parotid in children with and without Sjögren's Syndrome (SS). Both groups (20 healthy and 36 patients with SS) underwent 3.0 Tesla MRI, including diffusion-weighted imaging (DWI) (b = 0, 500 and 1000 s/mm2) and T2-weighted imaging. First-order statistics (FSO), second-order, and higher-order statistical texture analysis of the bilateral parotid glands were performed through texture analysis. A multimodal analysis, including texture analysis (TA) clinical classification; MRI evaluation for this disease entity to identify factors of disease activity. The examination group with SS was divided by the Tonami scale. The parotid ADC values from whole volume ROI were significantly lower in group 1 than in the higher activity group. The lowest kurtosis values in the highest grade of the Tonami Scale were found. The FSO parameters showed a positive correlation in TA from T2 weighted images- a moderate correlation with individual parameters. Our findings make TA a promising tool for assessing the grade of parotid gland inflammation. However, the effect of binning and ROI (region of interest) size is yet to be determined and needs further studies.