Experimental models of bone marrow lesions in ovine femoral condyles.

OBJECTIVE: To develop an in vivo experimental model for bone marrow lesions (BMLs) in ovine femorotibial joints. STUDY DESIGN: Randomized, prospective experimental study. ANIMALS: Eighteen healthy, skeletally-mature Dorper cross ewes. METHODS: One medial femoral condyle was penetrated with a 1.1 mm pin, and the contralateral medial femoral condyle was treated with transcutaneous extracorporeal shockwave (ESW) at 0.39 ± 0.04 mJ/mm2 . Clinical examination, magnetic resonance imaging (MRI), computed tomography (CT), and histopathological analyses were used to detect and characterize the development and progression of BMLs in the medial femoral condyle at 4, 8, and 12 weeks post-surgery. RESULTS: Pin penetration induced a BML detected on MRI within 2 weeks and lasted at least 12 weeks. BMLs were not observed in ESW-treated condyles. Histologically, BMLs were characterized by hemorrhage and inflammatory cellular infiltrate, and progressed to more dense fibrous tissue over time. Pathological changes were not observed in the articular cartilage overlying the region of BMLs. CONCLUSIONS: Direct, focal trauma to all layers of the osteochondral unit was sufficient to create an experimentally-induced BML which persisted for at least 90 days. The protocol used for ESW in this study did not induce BMLs. CLINICAL SIGNIFICANCE: Experimental induction of BMLs is possible and mimicked naturally occurring disease states. Volumetric imaging is a sensitive method for characterization of the dynamic nature of these lesions.

AI Radar Sensor: Creating Radar Depth Sounder Images Based on Generative Adversarial Network.

Significant resources have been spent in collecting and storing large and heterogeneous radar datasets during expensive Arctic and Antarctic fieldwork. The vast majority of data available is unlabeled, and the labeling process is both time-consuming and expensive. One possible alternative to the labeling process is the use of synthetically generated data with artificial intelligence. Instead of labeling real images, we can generate synthetic data based on arbitrary labels. In this way, training data can be quickly augmented with additional images. In this research, we evaluated the performance of synthetically generated radar images based on modified cycle-consistent adversarial networks. We conducted several experiments to test the quality of the generated radar imagery. We also tested the quality of a state-of-the-art contour detection algorithm on synthetic data and different combinations of real and synthetic data. Our experiments show that synthetic radar images generated by generative adversarial network (GAN) can be used in combination with real images for data augmentation and training of deep neural networks. However, the synthetic images generated by GANs cannot be used solely for training a neural network (training on synthetic and testing on real) as they cannot simulate all of the radar characteristics such as noise or Doppler effects. To the best of our knowledge, this is the first work in creating radar sounder imagery based on generative adversarial network.

Automation and Simplification: Drivers of Innovative Collection and Use of Patient-Reported Outcomes Data.

The aim was to develop an electronic data capture (EDC) system to capture patient-reported outcome (PRO) measures successfully by automating processes identified as barriers to implementation. Clinical success, research impact, and patient acceptance of this system were evaluated during a pilot and a follow-up period 2 years later. During the pilot, there were 44,831 eligible visits. Capture rate was 99.0% (44,374 visits) and completion rate was 99.4% (44,108 visits). Capture rate was 99.4% and completion rate was 95.2% during the follow-up period. Zero help desk tickets were put in for the EDC system during either time period. Patients accepted the EDC system both during the pilot (1.4% refusal rate) and follow-up period (1.2%). An automated Structured Query Language server feed provided data used to produce numerous abstracts and manuscripts. Automation was crucial to overcoming implementation barriers and delivering PRO scores to the electronic health record in real time with minimal impact on clinical workflow. Automation also has supported PRO research.