In the last decades, the use of artificial intelligence (AI) has been increasingly investigated in intervertebral disc degeneration (IDD) and chronic low back pain (LBP) research. To date, several AI-based cutting-edge technologies, such as computer vision, computer-assisted diagnosis, decision support system and natural language processing have been utilized to optimize LBP prevention, diagnosis, and treatment. This talk will provide an outline on contemporary AI applications to IDD and LBP research, with a particular attention towards actual knowledge gaps and promising innovative tools

Introduction and Objective

Up to 30% of thoracolumbar (TL) fractures are missed in the emergency room. Failure to identify these fractures can result in neurological injuries up to 51% of the casesthis article aimed to clarify the incidence and risk factors of traumatic fractures in China. The China National Fracture Study (CNFS. Obtaining sagittal and anteroposterior radiographs of the TL spine are the first diagnostic step when suspecting a traumatic injury. In most cases, CT and/or MRI are needed to confirm the diagnosis. These are time and resource consuming. Thus, reliably detecting vertebral fractures in simple radiographic projections would have a significant impact. We aim to develop and validate a deep learning tool capable of detecting TL fractures on lateral radiographs of the spine. The clinical implementation of this tool is anticipated to reduce the rate of missed vertebral fractures in emergency rooms.

Introduction

Orthopedics is experiencing a significant transformation with the introduction of technologies such as robotics and apps. These, integrated into the post-operative rehabilitation process, promise to improve clinical outcomes, patient satisfaction, and the overall efficiency of the healthcare system. This study examines the impact of an app called Mymobility and intra-operative data collected via the ROSA® robotic system on the functional recovery of patients undergoing robot-assisted knee arthroplasty.

Artificial Intelligence (AI) is becoming more powerful but is barely used to counter the growth in health care burden. AI applications to increase efficiency in orthopedics are rare. We questioned if (1) we could train machine learning (ML) algorithms, based on answers from digitalized history taking questionnaires, to predict treatment of hip osteoartritis (either conservative or surgical); (2) such an algorithm could streamline clinical consultation. Multiple ML models were trained on 600 annotated (80% training, 20% test) digital history taking questionnaires, acquired before consultation. Best performing models, based on balanced accuracy and optimized automated hyperparameter tuning, were build into our daily clinical orthopedic practice. Fifty patients with hip complaints (>45 years) were prospectively predicted and planned (partly blinded, partly unblinded) for consultation with the physician assistant (conservative) or orthopedic surgeon (operative). Tailored patient information based on the prediction was automatically sent to a smartphone app. Level of evidence: IV. Random Forest and BernoulliNB were the most accurate ML models (0.75 balanced accuracy). Treatment prediction was correct in 45 out of 50 consultations (90%), p<0.0001 (sign and binomial test). Specialized consultations where conservatively predicted patients were seen by the physician assistant and surgical patients by the orthopedic surgeon were highly appreciated and effective. Treatment strategy of hip osteoartritis based on answers from digital history taking questionnaires was accurately predicted before patients entered the hospital. This can make outpatient consultation scheduling more efficient and tailor pre-consultation patient education

Gait analysis is an indispensable tool for scientific assessment and treatment of individuals whose ability to walk is impaired. The high cost of installation and operation are a major limitation for wide-spread use in clinical routine. Advances in Artificial Intelligence (AI) could significantly reduce the required instrumentation. A mobile phone could be all equipment necessary for 3D gait analysis. MediaPipe Pose provided by Google Research is such a Machine Learning approach for human body tracking from monocular RGB video frames that is detecting 3D-landmarks of the human body. Aim of this study was to analyze the accuracy of gait phase detection based on the joint landmarks identified by the AI system. Motion data from 10 healthy volunteers walking on a treadmill with a fixed speed of 4.5km/h (Callis, Sprintex, Germany) was sampled with a mobile phone (iPhone SE 2nd Generation, Apple). The video was processed with Mediapipe Pose (Version 0.9.1.0) using custom python software. Gait phases (Initial Contact - IC and Toe Off - TO) were detected from the angular velocities of the lower legs. For the determination of ground truth, the movement was simultaneously recorded with the AS-200 System (LaiTronic GmbH, Innsbruck, Austria). The number of detected strides, the error in IC detection and stance phase duration was calculated. In total, 1692 strides were detected from the reference system during the trials from which the AI-system identified 679 strides. The absolute mean error (AME) in IC detection was 39.3 ± 36.6 ms while the AME for stance duration was 187.6 ± 140 ms. Landmark detection is a challenging task for the AI-system as can clearly be seen be the rate of only 40% detected strides. As mentioned by Fadillioglu et al., error in TO-detection is higher than in IC-detection

Introduction. With advances in artificial intelligence, the use of computer-aided detection and diagnosis in clinical imaging is gaining traction. Typically, very large datasets are required to train machine-learning models, potentially limiting use of this technology when only small datasets are available. This study investigated whether pretraining of fracture detection models on large, existing datasets could improve the performance of the model when locating and classifying wrist fractures in a small X-ray image dataset. This concept is termed “transfer learning”. Method. Firstly, three detection models, namely, the faster region-based convolutional neural network (faster R-CNN), you only look once version eight (YOLOv8), and RetinaNet, were pretrained using the large, freely available dataset, common objects in context (COCO) (330000 images). Secondly, these models were pretrained using an open-source wrist X-ray dataset called “Graz Paediatric Wrist Digital X-rays” (GRAZPEDWRI-DX) on a (1) fracture detection dataset (20327 images) and (2) fracture location and classification dataset (14390 images). An orthopaedic surgeon classified the small available dataset of 776 distal radius X-rays (Arbeidsgmeischaft für Osteosynthesefragen Foundation / Orthopaedic Trauma Association; AO/OTA), on which the models were tested. Result. Detection models without pre-training on the large datasets were the least precise when tested on the small distal radius dataset. The model with the best accuracy to detect and classify wrist fractures was the YOLOv8 model pretrained on the GRAZPEDWRI-DX fracture detection dataset (mean average precision at intersection over union of 50=59.7%). This model showed up to 33.6% improved detection precision compared to the same models with no pre-training. Conclusion. Optimisation of machine-learning models can be challenging when only relatively small datasets are available. The findings of this study support the potential of transfer learning from large datasets to improve model performance in smaller datasets. This is encouraging for wider application of machine-learning technology in medical imaging evaluation, including less common orthopaedic pathologies

Osteoarthritis (OA) is a degenerative joint disease affecting millions worldwide. Early detection of OA and monitoring its progression is essential for effective treatment and for preventing irreversible damage. Although sensors have emerged as a promising tool for monitoring analytes in patients, their application for monitoring the state of pathology is currently restricted to specific fields (such as diabetes). In this study, we present the development of an optical sensor system for real-time monitoring of inflammation based on the measurement of nitric oxide (NO), a molecule highly produced in tissues during inflammation. Single-walled carbon nanotubes (SWCNT) were functionalized with a single-stranded DNA (ssDNA) wrapping designed using an artificial intelligence approach and tested using S-nitroso-N-acetyl penicillamine (SNAP) as a standard released-NO marker. An optical SWIR reader with LED excitation at 650 nm, 730 nm and detecting emission above 1000 nm was developed to read the fluorescence signal from the SWCNTs. Finally, the SWCNT was embedded in GelMa to prove the feasibility of monitoring the release of NO in bovine chondrocyte and osteochondral inflamed cultures (1–10 ng/ml IL1β) monitored over 48 hours. The stability of the inflammation model and NO release was indirectly validated using the Griess and DAF-FM methods. A microfabricated sensor tag was developed to explore the possibility of using ssDNA-SWCNT in an ex vivo anatomic set-up for surgical feasibility, the limit of detection, and the stability under dynamic flexion. The SWCNT sensor was sensitive to NO in both in silico and in vitro conditions during the inflammatory response from chondrocyte and osteochondral plug cultures. The fluorescence signal decreased in the inflamed group compared to control, indicating increased NO concentration. The micro-tag was suitable and stable in joints showing a readable signal at a depth of up to 6 mm under the skin. The ssDNA-SWCNT technology showed the possibility of monitoring inflammation continuously in an in vitro set-up and good stability inside the joint. However, further studies in vivo are needed to prove the possibility of monitoring disease progression and treatment efficacy in vivo. Acknowledgments: The project was co-financed by Innosuisse (grant nr. 56034.1 IP-LS)

Introduction. Patients (2.7M in EU) with positive cancer prognosis frequently develop metastases (≈1M) in their remaining lifetime. In 30-70% cases, metastases affect the spine, reducing the strength of the affected vertebrae. Fractures occur in ≈30% patients. Clinicians must choose between leaving the patient exposed to a high fracture risk (with dramatic consequences) and operating to stabilise the spine (exposing patients to unnecessary surgeries). Currently, surgeons rely on their sole experience. This often results in to under- or over-treatment. The standard-of-care are scoring systems (e.g. Spine Instability Neoplastic Score) based on medical images, with little consideration of the spine biomechanics, and of the structure of the vertebrae involved. Such scoring systems fail to provide clear indications in ≈60% patients. Method. The HEU-funded METASTRA project is implemented by biomechanicians, modellers, clinicians, experts in verification, validation, uncertainty quantification and certification from 15 partners across Europe. METASTRA aims to improve the stratification of patients with vertebral metastases evaluating their risk of fracture by developing dedicated reliable computational models based on Explainable Artificial Intelligence (AI) and on personalised Physiology-based biomechanical (VPH) models. Result. The METASTRA-AI model is expected to be able to stratify most patients with limited effort end cost, based on parameters extracted semi-automatically from the medical files and images. The cases which are not reliably stratified through the AI model, are examined through a more detailed and personalised biomechanical VPH model. These METASTRA numerical tools are trained through an unprecedentedly large multicentric retrospective study (2000 cases) and validated against biomechanical ex vivo experiments (120 specimens). Conclusion. The METASTRA decision support system is tested in a multicentric prospective observational study (200 patients). The METASTRA approach is expected to cut down the indeterminate diagnoses from the current 60% down to 20% of cases. METASTRA project funded by the European Union, HEU topic HLTH-2022-12-01, grant 101080135

The objectives of this study are to evaluate the impact of the CoVID-19 pandemic on the development of relevant emerging digital healthcare trends and to explore which digital healthcare trend does the health industry need most to support HCPs. A web survey using 39 questions facilitating Five-Point Likert scales was performed from 1.8.2020 – 31.10.2020. Of 260 participants invited, 90 participants answered the questionnaire. The participants were located in the Hospital/HCP sector in 11.9%, in other healthcare sectors in 22.2%, in the pharmaceutical sector in 11.1%, in the medical device and equipment industry in 43.3%. The Five-Point Likert scales were in all cases fashioned as from 1 (strongly disagree) to 5 (strongly agree). As the top 3 most impacted digital health care trends strongly impacted by CoVID-19, respondents named:. - remote management of patients by telemedicine, mean answer 4.44. - shared data governance under patient control, mean answer 3.80. - new virtual interaction between HCP´s and medical industry, mean answer 3.76. Respondents were asked which level of readiness of the healthcare system currently possess to cope with the current trend impacted by CoVID-19. - Digital and efficient healthcare logistics, mean answer 1.54. - Integrated health care, mean answer 1.73. - Use of big data and artificial intelligence, mean answer 2.03. Asked if collaborative research in the form of digital data platforms for research data sharing and increasing collaboration with multi-centric consortia would have a positive impact on the healthcare sector, the agreement was high with a value of mean 4.10 on the scale. We can conclude that the impact of COVID-19 appears to be a high agreement of necessary advances in digitalization in the health care sector and in the collaboration of HCPs with the health care industry. Health care professional are unsure, in how far the national health care sector is capable of transformation in healthcare logistics and integrated health care

Precision medicine tailoring the patient pathway based on the risk, prognosis, and treatment response may bring benefits to the patients. To identify risk factors contributing to the early failure of treatment (development of events of interest) and when possible to change the prognosis via modifying these factors may improve the outcome and/or lower the risk of complications. There is an emerging goal to identify such parameters in total knee arthroplasty (TKA) thus lower the risk of revision surgery. The goal of this study was to identify factors explaining the risk for early revision of TKA using an artificial intelligence method appropriate for this task. We applied a patient similarity network (PSN) for the identification of risk factors associated with early reoperations (n=109, 5.8%) in patients with TKA (n=1885). Next, an algorithm based on formal concept analysis was developed to support the patient decision on how to change modifying personal characteristics with respect to the estimated probability of reoperations. The early reoperations were less frequent in women (4.4%, median time to reoperation 4.5 mo) than in men (8.2%, 10 mo), reaching the highest incidence in younger men (10.9%)