Deep learning-based recognition of key anatomical structures during robot-assisted minimally invasive esophagectomy.

OBJECTIVE: To develop a deep learning algorithm for anatomy recognition in thoracoscopic video frames from robot-assisted minimally invasive esophagectomy (RAMIE) procedures using deep learning. BACKGROUND: RAMIE is a complex operation with substantial perioperative morbidity and a considerable learning curve. Automatic anatomy recognition may improve surgical orientation and recognition of anatomical structures and might contribute to reducing morbidity or learning curves. Studies regarding anatomy recognition in complex surgical procedures are currently lacking. METHODS: Eighty-three videos of consecutive RAMIE procedures between 2018 and 2022 were retrospectively collected at University Medical Center Utrecht. A surgical PhD candidate and an expert surgeon annotated the azygos vein and vena cava, aorta, and right lung on 1050 thoracoscopic frames. 850 frames were used for training of a convolutional neural network (CNN) to segment the anatomical structures. The remaining 200 frames of the dataset were used for testing the CNN. The Dice and 95% Hausdorff distance (95HD) were calculated to assess algorithm accuracy. RESULTS: The median Dice of the algorithm was 0.79 (IQR = 0.20) for segmentation of the azygos vein and/or vena cava. A median Dice coefficient of 0.74 (IQR = 0.86) and 0.89 (IQR = 0.30) were obtained for segmentation of the aorta and lung, respectively. Inference time was 0.026 s (39 Hz). The prediction of the deep learning algorithm was compared with the expert surgeon annotations, showing an accuracy measured in median Dice of 0.70 (IQR = 0.19), 0.88 (IQR = 0.07), and 0.90 (0.10) for the vena cava and/or azygos vein, aorta, and lung, respectively. CONCLUSION: This study shows that deep learning-based semantic segmentation has potential for anatomy recognition in RAMIE video frames. The inference time of the algorithm facilitated real-time anatomy recognition. Clinical applicability should be assessed in prospective clinical studies.

Computer-aided anatomy recognition in intrathoracic and -abdominal surgery: a systematic review.

BACKGROUND: Minimally invasive surgery is complex and associated with substantial learning curves. Computer-aided anatomy recognition, such as artificial intelligence-based algorithms, may improve anatomical orientation, prevent tissue injury, and improve learning curves. The study objective was to provide a comprehensive overview of current literature on the accuracy of anatomy recognition algorithms in intrathoracic and -abdominal surgery. METHODS: This systematic review is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline. Pubmed, Embase, and IEEE Xplore were searched for original studies up until January 2022 on computer-aided anatomy recognition, without requiring intraoperative imaging or calibration equipment. Extracted features included surgical procedure, study population and design, algorithm type, pre-training methods, pre- and post-processing methods, data augmentation, anatomy annotation, training data, testing data, model validation strategy, goal of the algorithm, target anatomical structure, accuracy, and inference time. RESULTS: After full-text screening, 23 out of 7124 articles were included. Included studies showed a wide diversity, with six possible recognition tasks in 15 different surgical procedures, and 14 different accuracy measures used. Risk of bias in the included studies was high, especially regarding patient selection and annotation of the reference standard. Dice and intersection over union (IoU) scores of the algorithms ranged from 0.50 to 0.98 and from 74 to 98%, respectively, for various anatomy recognition tasks. High-accuracy algorithms were typically trained using larger datasets annotated by expert surgeons and focused on less-complex anatomy. Some of the high-accuracy algorithms were developed using pre-training and data augmentation. CONCLUSIONS: The accuracy of included anatomy recognition algorithms varied substantially, ranging from moderate to good. Solid comparison between algorithms was complicated by the wide variety of applied methodology, target anatomical structures, and reported accuracy measures. Computer-aided intraoperative anatomy recognition is an upcoming research discipline, but still at its infancy. Larger datasets and methodological guidelines are required to improve accuracy and clinical applicability in future research. TRIAL REGISTRATION: PROSPERO registration number: CRD42021264226.

[Tardive dyskinesia in a child treated with atypical antipsychotics]. / Tardieve dyskinesie bij een kind behandeld met atypische antipsychotica.

We describe the case of a 12-year-old boy that developed invalidating motor symptoms after starting and gradually increasing the dose of quetiapine with extended release and phasing out aripiprazole. The symptoms met the definition of tardive dyskinesia given their duration and presentation. Symptoms decreased spontaneously after discontinuation of neuroleptic treatment. The literature and knowledge among clinicians concerning the occurrence of tardive dyskinesia in children and adolescents treated with atypical antipsychotics are limited. We give an overview of the available scientific findings with special attention to the presentation, the incidence and treatment possibilities.

Identification and in vivo characterization of a brain-penetrating nanobody.

BACKGROUND: Preclinical models to determine blood to brain transport ability of therapeutics are often ambiguous. In this study a method is developed that relies on CNS target-engagement and is able to rank brain-penetrating capacities. This method led to the discovery of an anti-transferrin receptor nanobody that is able to deliver a biologically active peptide to the brain via receptor-mediated transcytosis. METHODS: Various nanobodies against the mouse transferrin receptor were fused to neurotensin and injected peripherally in mice. Neurotensin is a neuropeptide that causes hypothermia when present in the brain but is unable to reach the brain from the periphery. Continuous body temperature measurements were used as a readout for brain penetration of nanobody-neurotensin fusions after its peripheral administration. Full temperature curves were analyzed using two-way ANOVA with Dunnett multiple comparisons tests. RESULTS: One anti-transferrin receptor nanobody coupled to neurotensin elicited a drop in body temperature following intravenous injection. Epitope binning indicated that this nanobody bound a distinct transferrin receptor epitope compared to the non-crossing nanobodies. This brain-penetrating nanobody was used to characterize the in vivo hypothermia model. The hypothermic effect caused by neurotensin is dose-dependent and could be used to directly compare peripheral administration routes and various nanobodies in terms of brain exposure. CONCLUSION: This method led to the discovery of an anti-transferrin receptor nanobody that can reach the brain via receptor-mediated transcytosis after peripheral administration. This method could be used to assess novel proteins for brain-penetrating capabilities using a target-engaging readout.

The influence of gait and speed on the dynamic navicular drop - A cross sectional study on healthy subjects.

INTRODUCTION: Variations of gait speed influence kinematic variables that may have an effect on dynamic foot deformation. The influence of gait speed on the navicular drop has not yet been investigated. METHODS: The navicular drop was evaluated in static and dynamic conditions using a 3D-motion capture system. The dynamic navicular drop was evaluated on a treadmill while walking and running at three different speeds. A repeated measures ANOVA and post-hoc tests were conducted to evaluate the differences in dynamic navicular drop, corresponding unloaded navicular height at foot strike and loaded navicular height during stance. RESULTS: Higher walking speed led to a significant decrease in navicular height at foot strike and a subsequent decrease of dynamic navicular drop (p=0.006). Across increasing running speeds, minimum navicular height was significantly decreased which in consequence led to an increased dynamic navicular drop (p=0.015). For walking and running at the same speed, there was a large effect of gait style with an increase of dynamic navicular drop by 3.5mm (p<0.001) during running. DISCUSSION: The change of gait from walking to running at the same speed had a large effect on dynamic navicular drop. The values of navicular height at foot strike and minimum navicular height during stance should be taken into account for the interpretation of dynamic navicular drop measures. Static and dynamic navicular drop measures differ substantially.