Phantom study of SPECT/CT augmented reality for intraoperative localization of sentinel lymph nodes in head and neck melanoma.

OBJECTIVE: To show that augmented reality (AR) visualization of single-photon emission computed tomography (SPECT)/computed tomography (CT) data in 3D can be used to accurately localize targets in the head and neck region. MATERIALS AND METHODS: Eight head and neck styrofoam phantoms were painted with a mixture of radioactive solution (Tc-99m) detectable with a handheld gamma probe and fluorescent ink visible only under ultraviolet (UV) light to create 10-20 simulated lymph nodes on their surface. After obtaining SPECT/CT images of these phantoms, virtual renderings of the nodes were generated from the SPECT/CT data and displayed using a commercially available AR headset. For each of three physician evaluators, the time required to localize lymph node targets was recorded (1) using the gamma probe alone and (2) using the gamma probe while wearing the AR headset. In addition, the surface localization accuracy when using the AR headset was evaluated by measuring the misalignment between the locations visually marked by the evaluators and the ground truth locations identified using UV stimulation of the ink at the site of the nodes. RESULTS: For all three evaluators, using the AR headset significantly reduced the time to detect targets (P = 0.012, respectively) compared to using the gamma probe alone. The average misalignment between the location marked by the evaluators and the ground truth location was 8.6 mm. CONCLUSION: AR visualization of SPECT/CT data in 3D allows for accurate localization of targets in the head and neck region, and may reduce the localization time of targets.

A Patient-Specific Mixed-Reality Visualization Tool for Thoracic Surgical Planning.

PURPOSE: Identifying small lung lesions during minimally invasive thoracic surgery can be challenging. We describe 3-dimensional mixed-reality visualization technology that may facilitate noninvasive nodule localization. DESCRIPTION: A software application and medical image processing pipeline were developed for the Microsoft HoloLens to incorporate patient-specific data and provide a mixed-reality tool to explore and manipulate chest anatomy with a custom-designed user interface featuring gesture and voice recognition. EVALUATION: A needs assessment between engineering and clinical disciplines identified the potential utility of mixed-reality technology in facilitating safe and effective resection of small lung nodules. Through an iterative process, we developed a prototype employing a wearable headset that allows the user to (1) view a patient's original preoperative imaging; (2) manipulate a 3-dimensional rendering of that patient's chest anatomy including the bronchial, osseus, and vascular structures; and (3) simulate lung deflation and surgical instrument placement. CONCLUSIONS: Mixed-reality visualization during surgical planning may facilitate accurate and rapid identification of small lung lesions during minimally invasive surgeries and reduce the need for additional invasive preoperative localization procedures.

MR imaging of magnetic ink patterns via off-resonance sensitivity.

PURPOSE: Printed magnetic ink creates predictable B0 field perturbations based on printed shape and magnetic susceptibility. This can be exploited for contrast in MR imaging techniques that are sensitized to off-resonance. The purpose of this work was to characterize the susceptibility variations of magnetic ink and demonstrate its application for creating MR-visible skin markings. METHODS: The magnetic susceptibility of the ink was estimated by comparing acquired and simulated B0 field maps of a custom-built phantom. The phantom was also imaged using a 3D gradient echo sequence with a presaturation pulse tuned to different frequencies, which adjusts the range of suppressed frequencies. Healthy volunteers with a magnetic ink pattern pressed to the skin or magnetic ink temporary flexible adhesives applied to the skin were similarly imaged. RESULTS: The volume-average magnetic susceptibility of the ink was estimated to be 131 ± 3 parts per million across a 1-mm isotropic voxel (13,100 parts per million assuming a 10-µm thickness of printed ink). Adjusting the saturation frequency highlights different off-resonant regions created by the ink patterns; for example, if tuned to suppress fat, fat suppression will fail near the ink due to the off-resonance. This causes magnetic ink skin markings placed over a region with underlying subcutaneous fat to be visible on MR images. CONCLUSION: Patterns printed with magnetic ink can be imaged and identified with MRI. Temporary flexible skin adhesives printed with magnetic ink have the potential to be used as skin markings that are visible both by eye and on MR images.