Ultrasound-guided determination demonstrates influence of age, sex and type of sport on medial femoral condyle cartilage thickness in children and adolescents.

PURPOSE: To analyse the reliability of ultrasound-guided measurement of the cartilage thickness at the medial femoral condyle in athletically active children and adolescents before and after mechanical load in relation to age, sex and type of sport. METHODS: Three successive measurements were performed in 157 participants (median/min-max age: 13.1/6.0-18.0 years, 106 males) before and after mechanical load by squats at the same site of the medial femoral condyle by defined transducer positioning. Test-retest reliability was examined using Cronbach's α $\alpha $ calculation. Differences in cartilage thickness were analysed with respect to age, sex and type of practiced sports, respectively. RESULTS: Excellent reliability was achieved both before and after mechanical load by 30 squats with a median cartilage thickness of 1.9 mm (range: 0.5-4.8 mm) before and 1.9 mm (0.4-4.6 mm) after mechanical load. Male cartilages were thicker (p < 0.01) before (median: 2.0 mm) and after (2.0 mm) load when compared to female cartilage (before: 1.6 mm; after: 1.7 mm). Median cartilage thickness was about three times higher in karate athletes (before: 2.3 mm; after: 2.4 mm) than in sports shooters (0.7; 0.7 mm). Cartilage thickness in track and field athletes, handball players and soccer players were found to lay in-between. Sport type related thickness changes after mechanical load were not significant. CONCLUSION: Medial femoral condyle cartilage thickness in childhood correlates with age, sex and practiced type of sports. Ultrasound is a reliable and simple, pain-free approach to evaluate the cartilage thickness in children and adolescents. LEVEL OF EVIDENCE: Level III.

Comparison of the scientific performance in hip and knee arthroplasty between the leading continents.

Background: Scientific progress in the field of knee and hip arthroplasty has enabled the preservation of mobility and quality of life in the case of patients with many primary degenerative and (post-) traumatic joint diseases. This comparative study aims to investigate differences in scientific performance between the leading continents in the field of hip and knee arthroplasty. Methods: Using specific search terms all studies published by the scientific leading continents Europe, North America, Asia and Oceania listed in the Web of Science databases were included. All identified publications were analysed and comparative conclusions were drawn regarding the qualitative and quantitative scientific merit of each continent. Results: Europe, followed by North America, Asia, and Oceania, had the highest overall number of publications in the field of arthroplasty. Since 2000, there has been a strong increase in knee arthroplasty publication rate, particular pronounced in Asia. Studies performed and published in North America and those on knee arthroplasty received the highest number of fundings. Publications regarding hip arthroplasty achieved the highest average citation rate. In contradistinction to the others, in North America most funding was provided by private agencies. Conclusion: Although Europe showed the highest total number of publications, authors and institutions, arthroplasty research from North America received greater scientific attention and financial support. Measured by citations, publications on hip arthroplasty attained higher scientific interest and studies on knee arthroplasty received higher economic affection.

Subgroup analysis of scientific performance in the field of arthroplasty.

Introduction: Arthroplasty is the final treatment option for maintaining mobility and quality of life in many primary degenerative and (post-) traumatic joint diseases. Identification of research output and potential deficits for specific subspecialties may be an important measure to achieve long-term improvement of patient care in this field. Methods: Using specific search terms and Boolean operators, all studies published since 1945 to the subgroups of arthroplasty listed in the Web of Science Core Collection were included. All identified publications were analysed according to bibliometric standards, and comparative conclusions were drawn regarding the scientific merit of each subgroup. Results: Most publications investigated the subgroups of septic surgery and materials followed by approach, navigation, aseptic loosening, robotic and enhanced recovery after surgery (ERAS). In the last 5 years, research in the fields of robotic and ERAS achieved the highest relative increase in publications In contrast, research on aseptic loosening has continued to lose interest over the last 5 years. Publications on robotics and materials received the most funding on average while those on aseptic loosening received the least. Most publications originated from USA, Germany, and England, except for research on ERAS in which Denmark stood out. Relatively, publications on aseptic loosening received the most citations, whereas the absolute scientific interest was highest for the topic infection. Discussion: In this bibliometric subgroup analysis, the primary scientific outputs focused on septic complications and materials research in the field of arthroplasty. With decreasing publication output and the least financial support, intensification of research on aseptic loosening is urgently recommended.

[Digital tools in musculoskeletal rehabilitation]. / Digitale Hilfsmittel in der muskuloskelettalen Rehabilitation.

The demographic transition in combination with the increasing demands of society and a growing shortage of skilled workers are leading to a shortage of care in musculoskeletal rehabilitation, especially in times of the pandemic. Digital interventions represent an opportunity to reintegrate patients with musculoskeletal dysfunctions into everyday life. The changes to the legal basis enable physicians and therapists to support the rehabilitation of their patients with reimbursable apps and digital applications and to permanently integrate learned skills into their daily lives. Telerehabilitation technologies, apps, telerobotics and mixed reality offer the opportunity to complement and optimize existing care structures and to redesign specialized therapeutic home visits with modern technology in a new and contemporary way.

COVID-19: a catalyst for the digitization of surgical teaching at a German University Hospital.

BACKGROUND: The summer semester 2020, had to be restructured due to the SARS-CoV-2 pandemic and the associated contact restrictions. Here, for the first time, the established lectures in lecture halls and small group seminars could not be conducted in presence as usual. A possible tool for the implementation of medical teaching, offers the use of eLearning, online webinars and learning platforms. At present it is unclear how the SARS-CoV-2 pandemic will affect surgical teaching, how digitization will be accepted by students, and how virtual teaching can be expanded in the future. METHODS: The teaching, which was previously delivered purely through face-to-face lectures, was completely converted to digital media. For this purpose, all lectures were recorded and were available to students on demand. The seminars were held as a twice a week occurring online webinar. The block internship was also conducted as a daily online webinar and concluded with an online exam at the end. At the end of the semester, a survey of the students was carried out, which was answered by n = 192 students with an anonymized questionnaire. The questionnaire inquires about the previous and current experience with eLearning, as well as the possibility of a further development towards a purely digital university. RESULTS: There were n = 192 students in the study population. For 88%, the conversion of classes to web-based lectures represented their first eLearning experience. For 77% of all students, the digitization of teaching led to a change in the way they prepare for class. 73% of the participating students are of the opinion that eLearning lectures should continue to be offered. 54% of the students felt that eLearning lectures made more sense than face-to-face lectures. A purely virtual university could be imagined by 41% of the students. CONCLUSION: The conversion of teaching represented the first contact with eLearning for most students. Overall, the eLearning offering was experienced as positive. Due to the new teaching structure, the way of learning had already changed during the semester. Based on the new eLearning content, the already existing formats can be further expanded in the future. Nevertheless, it turned out that the practical-surgical contents and skills cannot be adequately represented by purely online offers; for this, the development of hybrid practice-oriented teaching concepts is necessary.

Learn to Predict Sets Using Feed-Forward Neural Networks.

This paper addresses the task of set prediction using deep feed-forward neural networks. A set is a collection of elements which is invariant under permutation and the size of a set is not fixed in advance. Many real-world problems, such as image tagging and object detection, have outputs that are naturally expressed as sets of entities. This creates a challenge for traditional deep neural networks which naturally deal with structured outputs such as vectors, matrices or tensors. We present a novel approach for learning to predict sets with unknown permutation and cardinality using deep neural networks. In our formulation we define a likelihood for a set distribution represented by a) two discrete distributions defining the set cardinally and permutation variables, and b) a joint distribution over set elements with a fixed cardinality. Depending on the problem under consideration, we define different training models for set prediction using deep neural networks. We demonstrate the validity of our set formulations on relevant vision problems such as: 1) multi-label image classification where we outperform the other competing methods on the PASCAL VOC and MS COCO datasets, 2) object detection, for which our formulation outperforms popular state-of-the-art detectors, and 3) a complex CAPTCHA test, where we observe that, surprisingly, our set-based network acquired the ability of mimicking arithmetics without any rules being coded.

RefineNet: Multi-Path Refinement Networks for Dense Prediction.

Recently, very deep convolutional neural networks (CNNs) have shown outstanding performance in object recognition and have also been the first choice for dense prediction problems such as semantic segmentation and depth estimation. However, repeated subsampling operations like pooling or convolution striding in deep CNNs lead to a significant decrease in the initial image resolution. Here, we present RefineNet, a generic multi-path refinement network that explicitly exploits all the information available along the down-sampling process to enable high-resolution prediction using long-range residual connections. In this way, the deeper layers that capture high-level semantic features can be directly refined using fine-grained features from earlier convolutions. The individual components of RefineNet employ residual connections following the identity mapping mindset, which allows for effective end-to-end training. Further, we introduce chained residual pooling, which captures rich background context in an efficient manner. We carry out comprehensive experiments on semantic segmentation which is a dense classification problem and achieve good performance on seven public datasets. We further apply our method for depth estimation and demonstrate the effectiveness of our method on dense regression problems.

Multi-Target Tracking by Discrete-Continuous Energy Minimization.

The task of tracking multiple targets is often addressed with the so-called tracking-by-detection paradigm, where the first step is to obtain a set of target hypotheses for each frame independently. Tracking can then be regarded as solving two separate, but tightly coupled problems. The first is to carry out data association, i.e., to determine the origin of each of the available observations. The second problem is to reconstruct the actual trajectories that describe the spatio-temporal motion pattern of each individual target. The former is inherently a discrete problem, while the latter should intuitively be modeled in continuous space. Having to deal with an unknown number of targets, complex dependencies, and physical constraints, both are challenging tasks on their own and thus most previous work focuses on one of these subproblems. Here, we present a multi-target tracking approach that explicitly models both tasks as minimization of a unified discrete-continuous energy function. Trajectory properties are captured through global label costs, a recent concept from multi-model fitting, which we introduce to tracking. Specifically, label costs describe physical properties of individual tracks, e.g., linear and angular dynamics, or entry and exit points. We further introduce pairwise label costs to describe mutual interactions between targets in order to avoid collisions. By choosing appropriate forms for the individual energy components, powerful discrete optimization techniques can be leveraged to address data association, while the shapes of individual trajectories are updated by gradient-based continuous energy minimization. The proposed method achieves state-of-the-art results on diverse benchmark sequences.

Continuous energy minimization for multitarget tracking.

Many recent advances in multiple target tracking aim at finding a (nearly) optimal set of trajectories within a temporal window. To handle the large space of possible trajectory hypotheses, it is typically reduced to a finite set by some form of data-driven or regular discretization. In this work, we propose an alternative formulation of multitarget tracking as minimization of a continuous energy. Contrary to recent approaches, we focus on designing an energy that corresponds to a more complete representation of the problem, rather than one that is amenable to global optimization. Besides the image evidence, the energy function takes into account physical constraints, such as target dynamics, mutual exclusion, and track persistence. In addition, partial image evidence is handled with explicit occlusion reasoning, and different targets are disambiguated with an appearance model. To nevertheless find strong local minima of the proposed nonconvex energy, we construct a suitable optimization scheme that alternates between continuous conjugate gradient descent and discrete transdimensional jump moves. These moves, which are executed such that they always reduce the energy, allow the search to escape weak minima and explore a much larger portion of the search space of varying dimensionality. We demonstrate the validity of our approach with an extensive quantitative evaluation on several public data sets.