Assessment of in vivo bone microarchitecture changes in an anti-TNFα treated psoriatic arthritic patient.

OBJECTIVE: Psoriatic arthritis (PsA) is an inflammatory rheumatic disease, mediated in part by TNFα and associated with bone loss. Anti-TNFα treatment should inhibit this phenomenon and reduce the systemic bone loss. Ultra-high field MRI (UHF MRI) may be used to quantify bone microarchitecture (BM) in-vivo. In this study, we quantified BM using UHF MRI in a PsA patient and followed up the changes related to anti-TNFα treatment. SUBJECTS AND METHODS: A non-treated PsA patient with knee arthritis and 7 gender-matched controls were scanned using a gradient re-echo sequence at UHF MRI. After a year of Adalimumab treatment, the patient underwent a second UHF MRI. A PET-FNa imaging was performed before and after treatment to identify and localize the abnormal metabolic areas. BM was characterized using typical morphological parameters quantified in 32 regions of interest (ROIs) located in the patella, proximal tibia, and distal femur. RESULTS: Before treatment, the BM parameters were statistically different from controls in 24/32 ROIs with differences reaching up to 38%. After treatment, BM parameters were normalized for 15 out of 24 ROIs. The hypermetabolic areas disclosed by PET-FNa before the treatment partly resumed after the treatment. CONCLUSION: Thanks to UHF MRI, we quantified in vivo BM anomalies in a PsA patient and we illustrated a major reversion after one year of treatment. Moreover, BM results highlighted that the abnormalities were not only localized in hypermetabolic regions identified by PET-FNa, suggesting that the bone loss was global and not related to inflammation.

Deep learning-based fully automatic segmentation of wrist cartilage in MR images.

The study objective was to investigate the performance of a dedicated convolutional neural network (CNN) optimized for wrist cartilage segmentation from 2D MR images. CNN utilized a planar architecture and patch-based (PB) training approach that ensured optimal performance in the presence of a limited amount of training data. The CNN was trained and validated in 20 multi-slice MRI datasets acquired with two different coils in 11 subjects (healthy volunteers and patients). The validation included a comparison with the alternative state-of-the-art CNN methods for the segmentation of joints from MR images and the ground-truth manual segmentation. When trained on the limited training data, the CNN outperformed significantly image-based and PB-U-Net networks. Our PB-CNN also demonstrated a good agreement with manual segmentation (Sørensen-Dice similarity coefficient [DSC] = 0.81) in the representative (central coronal) slices with a large amount of cartilage tissue. Reduced performance of the network for slices with a very limited amount of cartilage tissue suggests the need for fully 3D convolutional networks to provide uniform performance across the joint. The study also assessed inter- and intra-observer variability of the manual wrist cartilage segmentation (DSC = 0.78-0.88 and 0.9, respectively). The proposed deep learning-based segmentation of the wrist cartilage from MRI could facilitate research of novel imaging markers of wrist osteoarthritis to characterize its progression and response to therapy.