Background and objectivesAutomated segmentation and tracking of surgical instruments and catheters under X-ray fluoroscopy hold the potential for enhanced image guidance in catheter-based endovascular procedures. This article presents a novel method for real-time segmentation of catheters and guidewires in 2d X-ray images. We employ Convolutional Neural Networks (CNNs) and propose a transfer learning approach, using synthetic fluoroscopic images, to develop a lightweight version of the U-Net architecture. Our strategy, requiring a small amount of manually annotated data, streamlines the training process and results in a U-Net model, which achieves comparable performance to the state-of-the-art segmentation, with a decreased number of trainable parameters. MethodsThe proposed transfer learning approach exploits high-fidelity synthetic images generated from real fluroscopic backgrounds. We implement a two-stage process, initial end-to-end training and fine-tuning, to develop two versions of our model, using synthetic and phantom fluoroscopic images independently. A small number of manually annotated in-vivo images is employed to fine-tune the deepest 7 layers of the U-Net architecture, producing a network specialized for pixel-wise catheter/guidewire segmentation. The network takes as input a single grayscale image and outputs the segmentation result as a binary mask against the background. ResultsEvaluation is carried out with images from in-vivo fluoroscopic video sequences from six endovascular procedures, with different surgical setups. We validate the effectiveness of developing the U-Net models using synthetic data, in tests where fine-tuning and testing in-vivo takes place both by dividing data from all procedures into independent fine-tuning/testing subsets as well as by using different in-vivo sequences. Accurate catheter/guidewire segmentation (average Dice coefficient of ~ 0.55, ~ 0.26 and ~ 0.17) is obtained with both U-Net models. Compared to the state-of-the-art CNN models, the proposed U-Net achieves comparable performance ( ± 5% average Dice coefficients) in terms of segmentation accuracy, while yielding a 84% reduction of the testing time. This adds flexibility for real-time operation and makes our network adaptable to increased input resolution. ConclusionsThis work presents a new approach in the development of CNN models for pixel-wise segmentation of surgical catheters in X-ray fluoroscopy, exploiting synthetic images and transfer learning. Our methodology reduces the need for manually annotating large volumes of data for training. This represents an important advantage, given that manual pixel-wise annotations is a key bottleneck in developing CNN segmentation models. Combined with a simplified U-Net model, our work yields significant advantages compared to current state-of-the-art solutions.

The loss of a hand disrupts the sophisticated neural pathways between the brain and the hand, severely affecting the level of independence of the patient and the ability to carry out daily work and social activities. Recent years have witnessed a rapid evolution of surgical techniques and technologies aimed at restoring dexterous motor functions akin to those of the human hand through bionic solutions, mainly relying on probing of electrical signals from the residual nerves and muscles. Here, we report the clinical implementation of an interface aimed at achieving this goal by exploiting muscle deformation, sensed through passive magnetic implants: the myokinetic interface. One participant with a transradial amputation received an implantation of six permanent magnets in three muscles of the residual limb. A truly self-contained myokinetic prosthetic arm embedding all hardware components and the battery within the prosthetic socket was developed. By retrieving muscle deformation caused by voluntary contraction through magnet localization, we were able to control in real time a dexterous robotic hand following both a direct control strategy and a pattern recognition approach. In just 6 weeks, the participant successfully completed a series of functional tests, achieving scores similar to those achieved when using myoelectric controllers, a standard-of-care solution, with comparable physical and mental workloads. This experience raised conceptual and technical limits of the interface, which nevertheless pave the way for further investigations in a partially unexplored field. This study also demonstrates a viable possibility for intuitively interfacing humans with robotic technologies.

Magnetic localizers have been widely investigated in the biomedical field, especially for intra-body applications, because they don’t require a free line-of-sight between the implanted magnets and the magnetic field sensors. However, while researchers have focused on narrow and specific aspects of the localization problem, no one has comprehensively searched for general design rules for accurately localizing multiple magnetic objectives. In this study, we sought to systematically analyse the effects of remanent magnetization, number of sensors, and geometrical configuration (i.e. distance among magnets—L_inter-MM—and between magnets and sensors—L_MM-sensor) on the accuracy of the localizer in order to unveil the basic principles of the localization problem. Specifically, through simulations validated with a physical system, we observed that the accuracy of the localization was mainly affected by a specific angle ( $$\theta$$θ = tan⁻¹(L_inter-MM / L_MM-sensor)), descriptive of the system geometry. In particular, while tracking nine magnets, errors below ~ 1 mm (10% of the length of the simulated trajectory) and around 9° were obtained if θ ≥ ~ 31°. The latter proved a general rule across all tested conditions, also when the number of magnets was doubled. Our results are interesting for a whole range of biomedical engineering applications exploiting multiple-magnets tracking, such as human–machine interfaces, capsule endoscopy, ventriculostomy interventions, and endovascular catheter navigation.

The quest for an intuitive and physiologically appropriate human machine interface for the control of dexterous prostheses is far from being completed. In the last decade, much effort has been dedicated to explore innovative control strategies based on the electrical signals generated by the muscles during contraction. In contrast, a novel approach, dubbed myokinetic interface, derives the control signals from the localization of multiple magnetic markers (MMs) directly implanted into the residual muscles of the amputee. Building on this idea, here we present an embedded system based on 32 magnetic field sensors and a real time computation platform. We demonstrate that the platform can simultaneously localize in real-time up to five MMs in an anatomically relevant workspace. The system proved highly linear (R² = 0.99) and precise (1% repeatability), yet exhibiting short computation times (4 ms) and limited cross talk errors (10% the mean stroke of the magnets). Compared to a previous PC implementation, the system exhibited similar precision and accuracy, while being ~75% faster. These results proved for the first time the viability of using an embedded system for magnet localization. They also suggest that, by using an adequate number of sensors, it is possible to increase the number of simultaneously tracked MMs while introducing delays that are not perceivable by the human operator. This could allow to control more degrees of freedom than those controllable with current technologies.

Objective : The search for a physiologically appropriate interface for the control of dexterous hand prostheses is an ongoing challenge in bioengineering. In this context, we proposed an interface, named myokinetic control interface , based on the localization of magnets implanted in the residual limb muscles, to monitor their contractions and send appropriate commands to the artificial hand. As part of such concept, this interface requires a transcutaneous magnet localizer that can be integrated in a self-contained limb prosthesis, a feature yet to be realized within the current state of the art. Methods: In an attempt to cover this gap, here we present a modular embedded system consisting of a computation unit able to acquire synchronized samples captured by up to eight acquisition units, so to localize multiple magnets. Results: The system exhibits short computation times (<60ms) and power consumption (0.6-1.2W) which are suitable for use in a clinically viable prosthetic arm. The system proved able to localize magnets while moving at speeds in the range of physiological movements (<0.24m/s), with high accuracy (<1mm) and precision (<0.5mm). Conclusion: We demonstrated a system suitable for the implementation of a self-contained myokinetic prosthetic hand. Significance: These results pave the way towards the clinical implementation of the myokinetic interface, with amputees controlling an artificial arm by means of implanted magnets.