Noise-robust breathing-phase estimation on marker-free, ultra low dose X-ray projections for real-time tumor localization via surrogate structures.

PURPOSE: This paper presents a novel strategy for feature-based breathing-phase estimation on ultra low-dose X-ray projections for tumor motion control in radiation therapy. METHODS: Coarse-scaled Curvelet coefficients are identified as motion sensitive but noise-robust features for this purpose. For feature-based breathing-phase estimation, an ensemble strategy with two classifiers is used. This consensus-based estimation substantially increases tracking reliability by rejection of false positives. The algorithm is evaluated on both synthetic and measured phantom data: Monte Carlo simulated ultra low dose projections for a C-arm X-ray and on the basis of 4D-chest-CTs of eight patients on one hand side and real measurements based on a motion phantom. RESULTS: To achieve an accuracy of breathing-phase estimation of more than 95% a fluence between 20 and 400 photons per pixel (open field) is required depending on the patient. Furthermore, the algorithm is evaluated on real ultra low dose projections from an XVI R5.0 system (Elekta AB, Stockholm, Sweden) using an additional lead filter to reduce fluence. The classifiers-consensus-based-gating method estimated the correct position of the test projections in all test cases at a fluence of ∼180 photons per pixel and 92% at a fluence of ∼40 photons per pixel. The deposited dose to patient per image is in the range of nGy. CONCLUSIONS: A novel method is presented for estimation of breathing-phases for real-time tumor localization at ultra low dose both on a simulation and a phantom basis. Its accuracy is comparable to state of the art X-ray based algorithms while the released dose to patients is reduced by two to three orders of magnitude compared to conventional template-based approaches. This allows for continuous motion control during irradiation without the need of external markers.

Augmented Reality with HoloLens® in Parotid Tumor Surgery: A Prospective Feasibility Study.

INTRODUCTION: Augmented reality can improve planning and execution of surgical procedures. Head-mounted devices such as the HoloLens® (Microsoft, Redmond, WA, USA) are particularly suitable to achieve these aims because they are controlled by hand gestures and enable contactless handling in a sterile environment. OBJECTIVES: So far, these systems have not yet found their way into the operating room for surgery of the parotid gland. This study explored the feasibility and accuracy of augmented reality-assisted parotid surgery. METHODS: 2D MRI holographic images were created, and 3D holograms were reconstructed from MRI DICOM files and made visible via the HoloLens. 2D MRI slices were scrolled through, 3D images were rotated, and 3D structures were shown and hidden only using hand gestures. The 3D model and the patient were aligned manually. RESULTS: The use of augmented reality with the HoloLens in parotic surgery was feasible. Gestures were recognized correctly. Mean accuracy of superimposition of the holographic model and patient's anatomy was 1.3 cm. Highly significant differences were seen in position error of registration between central and peripheral structures (p = 0.0059), with a least deviation of 10.9 mm (centrally) and highest deviation for the peripheral parts (19.6-mm deviation). CONCLUSION: This pilot study offers a first proof of concept of the clinical feasibility of the HoloLens for parotid tumor surgery. Workflow is not affected, but additional information is provided. The surgical performance could become safer through the navigation-like application of reality-fused 3D holograms, and it improves ergonomics without compromising sterility. Superimposition of the 3D holograms with the surgical field was possible, but further invention is necessary to improve the accuracy.

Augmented reality with HoloLens in parotid surgery: how to assess and to improve accuracy.

PURPOSE: Augmented reality improves planning and execution of surgical procedures. The aim of this study was to evaluate the feasibility of a 3D augmented reality hologram in live parotic surgery. Another goal was to develop an accuracy measuring instrument and to determine the accuracy of the system. METHODS: We created a software to build and manually align 2D and 3D augmented reality models generated from MRI data onto the patient during surgery using the HoloLens® 1 (Microsoft Corporation, Redmond, USA). To assess the accuracy of the system, we developed a specific measuring tool applying a standard electromagnetic navigation device (Fiagon GmbH, Hennigsdorf, Germany). RESULTS: The accuracy of our system was measured during real surgical procedures. Training of the experimenters and the use of fiducial markers significantly reduced the accuracy of holographic system (p = 0.0166 and p = 0.0132). Precision of the developed measuring system was very high with a mean error of the basic system of 1.3 mm. Feedback evaluation demonstrated 86% of participants agreed or strongly agreed that the HoloLens will play a role in surgical education. Furthermore, 80% of participants agreed or strongly agreed that the HoloLens is feasible to be introduced in clinical routine and will play a role within surgery in the future. CONCLUSION: The use of fiducial markers and repeated training reduces the positional error between the hologram and the real structures. The developed measuring device under the use of the Fiagon navigation system is suitable to measure accuracies of holographic augmented reality images of the HoloLens.

Conditional random fields for phase-based lung feature tracking with ultra-low-dose x-rays.

PURPOSE: During radiation therapy, a continuous internal tumor monitoring without additional imaging dose is desirable. In this study, a sequential feature-based position estimation with ultra-low-dose (ULD) kV x rays using linear-chain conditional random fields (CRFs) is performed. METHODS: Four-dimensional computed tomography (4D-CTs) of eight patients serve as a-priori information from which ULD projections are simulated using a Monte Carlo method. CRFs are trained with Local Energy-based Shape Histogram features extracted from the ULD images to estimate one out of ten breathing phases from the 4D-CT associated with the tumor position. RESULTS: Compared to a mean accuracy for ±1 breathing phase of 0.867 using a support vector machine (SVM), a mean accuracy of 0.958 results for the CRF with ten incident photons per pixel. This corresponds to a position estimation with a discretization error of 2.4-5.3 mm assuming a linear displacement relation between the breathing phases and a systematic error of 2.0-4.4 mm due to motion underestimation of the 4D-CT. CONCLUSIONS: The tumor position estimation is comparable to state-of-the-art methods despite its low imaging dose. Training CRFs further allows a prediction of the following phase and offers a precise post-treatment evaluation tool when decoding the full image sequence.