Augmented reality (AR) is an advanced technology and emerging field that has been adopted into spine surgery to enhance care and outcomes. AR superimposes a three-dimensional computer-generated image over the normal anatomy of interest in order to facilitate visualization of deep structures without the ability to directly see them. To summarize the latest literature and highlight AR from the annual “Spinal Navigation, Emerging Technologies and Systems Integration” meeting lectures presented by the Seattle Science Foundation (SSF) on the development and use of augmented reality in spinal surgery. We performed a comprehensive literature review from 2016 to 2020 on PubMed to correlate with lectures given at the annual “Emerging Technologies” conferences. After the exclusion of papers that concerned non-spine surgery specialties, a total of 54 papers concerning AR in spinal applications were found. The articles were then categorized by content and focus. The 54 papers were divided into six major focused topics: training, proof of concept, feasibility and usability, clinical evaluation, state of technology, and nonsurgical applications. The greatest number of papers were published during 2020. Each paper discussed varied topics such as patient rehabilitation, proof of concept, workflow, applications in neurological and orthopedic spine surgery, and outcomes data. The recent literature and SSF lectures on AR provide a solid base and demonstrate the emergence of an advanced technology that offers a platform for an advantageous technique that is superior, in that it allows the operating surgeon to focus directly on the patient rather than a guidance screen.

Objective: To provide guidelines to healthcare workers for performing a focused neurological examination via telemedicine during the coronavirus disease-2019 (COVID-2019) pandemic. Methods: We reviewed our department’s outpatient clinic visits after the implementation of a telemedicine protocol in response to the COVID-19 crisis. Crossover rates from telehealth to in-person visits were evaluated and guidelines for performing a telemedicine neurological exam were created based on the consensus of 16 neurosurgical attending providers over a four-month period. Results: From March 23, 2020 to July 20, 2020, some 2157 telehealth visits were performed in our department. Some 26 were converted to in-person visits by the provider request with the most cited reason for conversion being the need for a more detailed patient evaluation. Based on these experiences, we created a graphical tutorial to address the key components of the neurological exam with adaptations specific to the telehealth visit. Conclusions: In response to the global coronavirus pandemic, telemedicine has become an integral part of neurosurgeons’ daily practice. Telemedicine failures remain low but primarily occur due to a need for more comprehensive evaluations. We provide guidelines for the neurosurgical exam during telehealth visits in an effort to assuage some of these issues.

Study Design.: Laboratory investigation with phantom spine models. Objectives.: The aim of this study was to demonstrate the ability of Augmented Reality system to track instruments from different companies without major modifications. Summary of Background Data.: Augmented Reality is an emergent technology with applications in industrial, military, gaming, and medical fields. AR applications in Spine surgery are actively being developed. Features of headpiece ergonomics, digital processing power, intuitive interface, and reliable accuracy are being optimized for successful adaptation of technology into the field. System versatility across various instrumentation sets is important for cost-effectiveness and efficiency in application. Methods.: In this project, five phantom spine models were instrumented L1-S1 with pedicle screws from five major companies. AR assistance was used for all. Each screwdriver was equipped with a generic 3D printed navigation marker for tracking. Results.: Every instrumentation set was successfully paired with AR navigation imaging. Sixty pedicle screws were inserted with an average time of 1.6 min/screw. There was an evidence of learning curve with fastest time achieved of 1 min/screw. All five systems had equivocal radiographic outcomes. There were two breached screws (3%). Conclusion.: Any currently available instrumentation set can readily pair for tracking with Augmented Reality system. Active tracking of the drivers allowed for improved accuracy making AR system very attractive as an adjunct to the current instrumentation techniques. Level of Evidence: 3

Lumbar disc degeneration, a progressive structural wear and tear of lumbar intervertebral disc, is regarded as an essential role on low back pain, a significant global health concern. Automated lumbar spine geometry reconstruction from MR images will enable fast measurement of medical parameters to evaluate the lumbar status, in order to determine a suitable treatment. Existing image segmentation-based techniques often generate erroneous segments or unstructured point clouds, unsuitable for medical parameter measurement. In this work, we present $\textit{UNet-DeformSA}$ and $\textit{TransDeformer}$: novel attention-based deep neural networks that reconstruct the geometry of the lumbar spine with high spatial accuracy and mesh correspondence across patients, and we also present a variant of $\textit{TransDeformer}$ for error estimation. Specially, we devise new attention modules with a new attention formula, which integrate image features and tokenized contour features to predict the displacements of the points on a shape template without the need for image segmentation. The deformed template reveals the lumbar spine geometry in an image. Experiment results show that our networks generate artifact-free geometry outputs, and the variant of $\textit{TransDeformer}$ can predict the errors of a reconstructed geometry. Our code is available at https://github.com/linchenq/TransDeformer-Mesh.

Ambiguity is inevitable in medical images, which often results in different image interpretations (e.g. object boundaries or segmentation maps) from different human experts. Thus, a model that learns the ambiguity and outputs a probability distribution of the target, would be valuable for medical applications to assess the uncertainty of diagnosis. In this paper, we propose a powerful generative model to learn a representation of ambiguity and to generate probabilistic outputs. Our model, named Coordinate Quantization Variational Autoencoder (CQ-VAE) employs a discrete latent space with an internal discrete probability distribution by quantizing the coordinates of a continuous latent space. As a result, the output distribution from CQ-VAE is discrete. During training, Gumbel-Softmax sampling is used to enable backpropagation through the discrete latent space. A matching algorithm is used to establish the correspondence between model-generated samples and "ground-truth" samples, which makes a trade-off between the ability to generate new samples and the ability to represent training samples. Besides these probabilistic components to generate possible outputs, our model has a deterministic path to output the best estimation. We demonstrated our method on a lumbar disk image dataset, and the results show that our CQ-VAE can learn lumbar disk shape variation and uncertainty.