Background and Objectives: Fracture healing is currently assessed through qualitative evaluation of radiographic images, which is highly subjective in nature. Radiographs can only provide snapshots in time, which are limited due to logistics and radiation exposure. We recently proposed assessing the bone healing status through continuous monitoring of the implant load, utilizing an implanted sensor system, the Fracture Monitor. The device telemetrically transmits statistically derived implant parameters via the patient’s mobile phone to assist physicians in diagnostics and treatment decision-making. This preclinical study aims to systematically investigate the device safety and performance in an animal setting. Materials and Methods: Mid-shaft tibial osteotomies of different sizes (0.6–30 mm) were created in eleven Swiss mountain sheep. The bones were stabilized with either a conventional Titanium or stainless-steel locking plate equipped with a Fracture Monitor. Data were continuously collected over the device’s lifetime. Conventional radiographs and clinical CT scans were taken longitudinally over the study period. The radiographs were systematically scored and CTs were evaluated for normalized bone volume in the defect. The animals were euthanized after 9 months. The sensor output was correlated with the radiologic parameters. Tissue samples from the device location were histologically examined. Results: The sensors functioned autonomously for 6.5–8.4 months until energy depletion. No macroscopic or microscopic adverse effects from device implantation were observed. The relative implant loads at 4 and 8 weeks post-operation correlated significantly with the radiographic scores and with the normalized bone volume metric. Conclusions: Continuous implant load monitoring appears as a relevant approach to support and objectify fracture healing assessments and carries a strong potential to enable patient-tailored rehabilitation in the future.

Artificial intelligence is a revolutionary technology that promises to assist clinicians in improving patient care. In radiology, deep learning (DL) is widely used in clinical decision aids due to its ability to analyze complex patterns and images. It allows for rapid, enhanced data, and imaging analysis, from diagnosis to outcome prediction. The purpose of this study was to evaluate the current literature and clinical utilization of DL in spine imaging. This study is a scoping review and utilized the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology to review the scientific literature from 2012 to 2021. A search in PubMed, Web of Science, Embased, and IEEE Xplore databases with syntax specific for DL and medical imaging in spine care applications was conducted to collect all original publications on the subject. Specific data was extracted from the available literature, including algorithm application, algorithms tested, database type and size, algorithm training method, and outcome of interest. A total of 365 studies (total sample of 232,394 patients) were included and grouped into 4 general applications: diagnostic tools, clinical decision support tools, automated clinical/instrumentation assessment, and clinical outcome prediction. Notable disparities exist in the selected algorithms and the training across multiple disparate databases. The most frequently used algorithms were U-Net and ResNet. A DL model was developed and validated in 92% of included studies, while a pre-existing DL model was investigated in 8%. Of all developed models, only 15% of them have been externally validated. Based on this scoping review, DL in spine imaging is used in a broad range of clinical applications, particularly for diagnosing spinal conditions. There is a wide variety of DL algorithms, database characteristics, and training methods. Future studies should focus on external validation of existing models before bringing them into clinical use.

Background and Objectives: Spinal fusion is an effective and widely accepted intervention. However, complications such as non-unions and hardware failures are frequently observed. Radiologic imaging and physical examination are still the gold standards in the assessment of spinal fusion, despite multiple limitations including radiation exposure and subjective image interpretation. Furthermore, current diagnostic methods only allow fusion assessment at certain time points and require the patient’s presence at the hospital or medical practice. A recently introduced implantable sensor system for continuous and wireless implant load monitoring in trauma applications carries the potential to overcome these drawbacks, but transferability of the principle to the spine has not been demonstrated yet. Materials and Methods: The existing trauma sensor was modified for attachment to a standard pedicle-screw-rod system. Two lumbar segments (L2 to L4) of one Swiss white alpine sheep were asymmetrically instrumented. After facetectomy, three sensors were attached to the rods between each screw pair and activated for measurement. The sheep was euthanized 16 weeks postoperatively. After radiological assessment the spine was explanted and loaded in flexion-extension to determine the range of motion of the spinal segments. Sensor data were compared with mechanical test results and radiologic findings. Results: The sensors measured physiological rod loading autonomously over the observation period and delivered the data daily to bonded smartphones. At euthanasia the relative rod load dropped to 67% of the respective maximum value for the L23 segment and to 30% for the L34 segment. In agreement, the total range of motion of both operated segments was lower compared to an intact reference segment (L23: 0.57°; L34: 0.49°; intact L45: 4.17°). Radiologic assessment revealed fusion mass in the facet joint gaps and bilateral bridging bone around the joints at both operated segments. Conclusions: Observations of this single-case study confirm the basic ability of continuous rod load measurement to resolve the spinal fusion process as indicated by a declining rod load with progressing bone fusion. A strong clinical potential of such technology is eminent, but further data must be collected for final proof of principle.

Background: The current standard of care for intervertebral disc (IVD) herniation, surgical discectomy, does not repair annulus fibrosus (AF) defects, which is partly due to the lack of effective methods to do so and is why new repair strategies are widely investigated and tested preclinically. There is a need to develop a standardized IVD injury model in large animals to enable comparison and interpretation across preclinical study results. The purpose of this study was to compare in vivo IVD injury models in sheep to determine which annulus fibrosus (AF) defect type combined with partial nucleus pulposus (NP) removal would better mimic degenerative human spinal pathologies.Methods: Six skeletally mature sheep were randomly assigned to one of the two observation periods (1 and 3 months) and underwent creation of 3 different AF defect types (slit, cruciate, and box‐cut AF defects) in conjunction with 0.1 g NP removal in three lumbar levels using a lateral retroperitoneal surgical approach. The spine was monitored by clinical CT scans pre‐ and postoperatively, at 2 weeks and euthanasia, and by magnetic resonance imaging (MRI) and histology after euthanasia to determine the severity of degeneration (disc height loss, Pfirrmann grading, semiquantitative histopathology grading).Results: All AF defects led to significant degenerative changes detectable on CT and MR images, produced bulging of disc tissue without disc herniation and led to degenerative and inflammatory histopathological changes. However, AF defects were not equal in terms of disc height loss at 3 months postoperatively; the cruciate and box‐cut AF defects showed significantly decreased disc height compared to their preoperative height, with the box‐cut defect creating the greatest disc height loss, while the slit AF defect showed restoration of normal preoperative disc height.Conclusions: The tested IVD injury models do not all generate comparable disc degeneration but can be considered suitable IVD injury models to investigate new treatments. Results of the current study clearly indicate that slit AF defect should be avoided if disc height is used as one of the main outcomes; additional confirmatory studies may be warranted to generalize this finding.

Objective: To provide a systematic assessment of the efficacy of preoperative skin asepsis using chlorhexidine versus povidone‐iodine based protocols for surgical site infection (SSI) prevention in veterinary surgery.Study design: Systematic meta‐analytical review according to PRISMA‐P guidelines.Sample population: Studies comparing preoperative skin asepsis protocols using chlorhexidine versus povidone‐iodine in veterinary surgery identified by systematic search between 1990 and 2020.Methods: A search using MEDLINE/Pubmed, Web of Science and CAB Abstracts was performed, followed by secondary searches of Google Scholar, Proquest Dissertation and Theses, and relevant bibliographic articles. Primary and secondary outcome measures were the efficacy of skin asepsis protocols using chlorhexidine versus povidone‐iodine on SSI incidence and skin bacterial colonization, respectively. A meta‐analysis was performed with a random‐effect model, with effect size calculated as risk ratio (RR) or mean standard deviation (MSD) with 95% CI. Statistical significance was set at P < .05.Results: Among 1067 publications that met the initial search criteria, 9 relevant studies were eligible for analysis. No difference in the incidence of postoperative SSI or skin bacterial colonization between preoperative asepsis protocols using chlorhexidine versus povidone‐iodine was found. Insufficient information and detail were frequent among studies and precluded a clear assessment of bias.Conclusion: This study showed that asepsis protocols using chlorhexidine were comparable to povidone‐iodine in preventing postoperative SSI and reducing skin bacterial colonization.Clinical significance: Given the limitations of the studies that were included in terms of both quality and quantity, more high‐quality randomized controlled trials are needed to confirm these conclusions.