QU-BraTS: MICCAI BraTS 2020 Challenge on Quantifying Uncertainty in Brain Tumor Segmentation - Analysis of Ranking Scores and Benchmarking Results

Deep learning (DL) models have provided state-of-the-art performance in various medical imaging benchmarking challenges, including the Brain Tumor Segmentation (BraTS) challenges. However, the task of focal pathology multi-compartment segmentation (e.g., tumor and lesion sub-regions) is particularly challenging, and potential errors hinder translating DL models into clinical workflows. Quantifying the reliability of DL model predictions in the form of uncertainties could enable clinical review of the most uncertain regions, thereby building trust and paving the way toward clinical translation. Several uncertainty estimation methods have recently been introduced for DL medical image segmentation tasks. Developing scores to evaluate and compare the performance of uncertainty measures will assist the end-user in making more informed decisions. In this study, we explore and evaluate a score developed during the BraTS 2019 and BraTS 2020 task on uncertainty quantification (QU-BraTS) and designed to assess and rank uncertainty estimates for brain tumor multi-compartment segmentation. This score (1) rewards uncertainty estimates that produce high confidence in correct assertions and those that assign low confidence levels at incorrect assertions, and (2) penalizes uncertainty measures that lead to a higher percentage of under-confident correct assertions. We further benchmark the segmentation uncertainties generated by 14 independent participating teams of QU-BraTS 2020, all of which also participated in the main BraTS segmentation task. Overall, our findings confirm the importance and complementary value that uncertainty estimates provide to segmentation algorithms, highlighting the need for uncertainty quantification in medical image analyses. Finally, in favor of transparency and reproducibility, our evaluation code is made publicly available at https://github.com/RagMeh11/QU-BraTS.

Videofluoroscopic Evaluation of the Influence of a Gradually Reducing Femoral Radius on Joint Kinematics During Daily Activities in Total Knee Arthroplasty

BACKGROUND

Paradoxical anterior translation in midflexion is reduced in total knee arthroplasties (TKAs) with a gradually reducing femoral radius, when compared to a 2-radii design. This reduction has been shown in finite element model simulations, in vitro tests, intraoperatively, and recently also in vivo during a lunge and unloaded flexion-extension. However, TKA kinematics are task dependent and this reduction has not been tested for gait activities.

METHODS

Thirty good outcome subjects (≥1 year postoperatively) with a unilateral cruciate-retaining TKA with a gradually reducing (n = 15) or dual (n = 15) femoral radius design were assessed during 5 complete cycles of level walking, stair descent (0.18-m steps), deep knee bend, and sitting down onto and standing up from a chair, using a moving fluoroscope (25 Hz, 1 ms shutter time). Kinematic data were extracted by 2D/3D image registration.

RESULTS

Tibiofemoral ranges of motion for flexion-extension, abduction-adduction, internal-external rotation, and anteroposterior (AP) translation were similar for both groups, whereas the pattern of AP translation-flexion-coupling differed. The subjects with the dual-radii design showed a sudden change in direction of AP translation around 30° of flexion, which was not present in the subjects with the gradually reducing femoral radius design.

CONCLUSION

Through the unique ability of moving fluoroscopy, the present study confirmed that the gradually reducing femoral radii eliminated the paradoxical sudden anterior translation at 30° present in the dual-radii design in vivo during daily activities, including gait and stair descent.

Development of VariLeg, an exoskeleton with variable stiffness actuation: first results and user evaluation from the CYBATHLON 2016

BACKGROUND

Powered exoskeletons are a promising approach to restore the ability to walk after spinal cord injury (SCI). However, current exoskeletons remain limited in their walking speed and ability to support tasks of daily living, such as stair climbing or overcoming ramps. Moreover, training progress for such advanced mobility tasks is rarely reported in literature. The work presented here aims to demonstrate the basic functionality of the VariLeg exoskeleton and its ability to enable people with motor complete SCI to perform mobility tasks of daily life.

METHODS

VariLeg is a novel powered lower limb exoskeleton that enables adjustments to the compliance in the leg, with the objective of improving the robustness of walking on uneven terrain. This is achieved by an actuation system with variable mechanical stiffness in the knee joint, which was validated through test bench experiments. The feasibility and usability of the exoskeleton was tested with two paraplegic users with motor complete thoracic lesions at Th4 and Th12. The users trained three times a week, in 60 min sessions over four months with the aim of participating in the CYBATHLON 2016 competition, which served as a field test for the usability of the exoskeleton. The progress on basic walking skills and on advanced mobility tasks such as incline walking and stair climbing is reported. Within this first study, the exoskeleton was used with a constant knee stiffness.

RESULTS

Test bench evaluation of the variable stiffness actuation system demonstrate that the stiffness could be rendered with an error lower than 30 Nm/rad. During training with the exoskeleton, both users acquired proficient skills in basic balancing, walking and slalom walking. In advanced mobility tasks, such as climbing ramps and stairs, only basic (needing support) to intermediate (able to perform task independently in 25% of the attempts) skill levels were achieved. After 4 months of training, one user competed at the CYBATHLON 2016 and was able to perform 3 (stand-sit-stand, slalom and tilted path) out of 6 obstacles of the track. No adverse events occurred during the training or the competition.

CONCLUSION

Demonstration of the applicability to restore ambulation for people with motor complete SCI was achieved. The CYBATHLON highlighted the importance of training and gaining experience in piloting an exoskeleton, which were just as important as the technical realization of the robot.