Assessment of the axial resolution of a compact gamma camera with coded aperture collimator.

PURPOSE: Handheld gamma cameras with coded aperture collimators are under investigation for intraoperative imaging in nuclear medicine. Coded apertures are a promising collimation technique for applications such as lymph node localization due to their high sensitivity and the possibility of 3D imaging. We evaluated the axial resolution and computational performance of two reconstruction methods. METHODS: An experimental gamma camera was set up consisting of the pixelated semiconductor detector Timepix3 and MURA mask of rank 31 with round holes of 0.08 mm in diameter in a 0.11 mm thick Tungsten sheet. A set of measurements was taken where a point-like gamma source was placed centrally at 21 different positions within the range of 12-100 mm. For each source position, the detector image was reconstructed in 0.5 mm steps around the true source position, resulting in an image stack. The axial resolution was assessed by the full width at half maximum (FWHM) of the contrast-to-noise ratio (CNR) profile along the z-axis of the stack. Two reconstruction methods were compared: MURA Decoding and a 3D maximum likelihood expectation maximization algorithm (3D-MLEM). RESULTS: While taking 4400 times longer in computation, 3D-MLEM yielded a smaller axial FWHM and a higher CNR. The axial resolution degraded from 5.3 mm and 1.8 mm at 12 mm to 42.2 mm and 13.5 mm at 100 mm for MURA Decoding and 3D-MLEM respectively. CONCLUSION: Our results show that the coded aperture enables the depth estimation of single point-like sources in the near field. Here, 3D-MLEM offered a better axial resolution but was computationally much slower than MURA Decoding, whose reconstruction time is compatible with real-time imaging.

Point Cloud Registration for Measuring Shape Dependence of Soft Tissue Deformation by Digital Twins in Head and Neck Surgery.

Introduction: A 2½ D point cloud registration method was developed to generate digital twins of different tissue shapes and resection cavities by applying a machine learning (ML) approach. This demonstrates the feasibility of quantifying soft tissue shifts. Methods: An ML model was trained using simulated surface scan data obtained from tumor resections in a pig head cadaver model. It hereby uses 438 2½ D scans of the tissue surface. Tissue shift was induced by a temperature change from 7.91 ± 4.1°C to 36.37 ± 1.28°C. Results: Digital twins were generated from various branched and compact resection cavities (RCs) and cut tissues (CT). A temperature increase induced a tissue shift with a significant volume increase of 6 mL and 2 mL in branched and compact RCs, respectively (p = 0.0443; 0.0157). The volumes of branched and compact CT were decreased by 3 and 4 mL (p < 0.001). In the warm state, RC and CT no longer fit together because of the significant tissue deformation. Although not significant, the compact RC showed a greater tissue deformation of 1 µL than the branched RC with 0.5 µL induced by the temperature change (p = 0.7874). The branched and compact CT forms responded almost equally to changes in temperature (p = 0.1461). Conclusions: The simulation experiment of induced soft tissue deformation using digital twins based on 2½ D point cloud models proved that our method helps to quantify shape-dependent tissue shifts.

Diffuse domain approach for flexible needle insertion and relaxation.

Needle insertion simulations play an important role in medical training and surgical planning. Most simulations require boundary conforming meshes, while the diffuse domain approach, currently limited to stiff needles, eliminates the need for meshing geometries. In this article the diffuse domain approach for needle insertion simulations is first extended to the use of flexible needles with bevel needle tips, which are represented by an Euler-Bernoulli beam. The model parameters are tuned and the model is evaluated on a real-world phantom experiment. Second, a new method for the relaxation of the needle-tissue system after the user releases the needle is introduced. The equilibrium state of the system is determined by minimizing the potential energy. The convergence rate of the coupled Laplace equations for solving the Euler-Bernoulli beam is 1.92 ± 0.14 for decreasing cell size. The diffuse penalty method for the application of Dirichlet boundary conditions results in a convergence rate of 0.73 ± 0.21 for decreasing phase field width. The simulated needle deviates on average by 0.29 mm compared to the phantom experiment. The error of the tissue deformation is below 1 mm for 97.5% of the attached markers. Two additional experiments demonstrate the feasibility of the relaxation process. The simulation method presented here is a valuable tool for patient-specific medical simulations using flexible needles without the need for boundary conforming meshing. To the best of the authors' knowledge this is the first work to introduce a relaxation model, which is a major step for simulating accurate needle-tissue positioning during realistic medical interventions.

Real-time definition of single seed placement sensitivity in low-dose-rate prostate brachytherapy.

PURPOSE: In low-dose-rate brachytherapy, iodine-125 seeds are implanted based on a treatment plan, generated with respect to different dose constraints. The quality of the dose distribution depends on a precise seed placement, however, during treatment planning the impact on the dose parameters when certain seeds fail to be placed precisely is not clear. METHODS AND MATERIALS: We developed a method using automatic differentiation to calculate gradients of dose parameters with regard to the seeds' positions. Thus, we understand their sensitivity with respect to the seed placement. A statistical analysis is performed on a data set with 35 prostate brachytherapy patients. RESULTS: The most sensitive seeds regarding the dosimetric parameters of both rectum and urethra are close to the corresponding organ. Their gradient directions are mainly orthogonal to their surfaces. However, not all seeds close to the surface are equally sensitive with regard to the dose parameter. The most sensitive seeds regarding the prostate's dose parameters are distributed throughout the prostate and the direction of the gradients are mainly parallel to its surface. A linear regression with respect to different patient parameters shows that dose constraints which are barely fulfilled have large gradients and thus are additionally sensitive to misplacement. CONCLUSION: Automatic differentiation can be used to analyze dose parameter sensitivity with respect to seed placement. Integrating this into treatment planning systems is valuable as it speeds up the planning procedure, making it more robust and less dependent on user experience while showing the operating physician which needle placements require greater accuracy than others.

Refocusing of Attention on Positive Events Using Monitoring-Based Feedback and Microinterventions for Patients With Chronic Musculoskeletal Pain in the PerPAIN Randomized Controlled Trial: Protocol for a Microrandomized Trial.

BACKGROUND: Chronic musculoskeletal pain (CMSP) affects between 13% and 47% of the population, with a global growth rate of 20.3% within the last 15 years, suggesting that there is a high need for effective treatments. Pain diaries have long been a common tool in nonpharmacological pain treatment for monitoring and providing feedback on patients' symptoms in daily life. More recently, positive refocusing techniques have come to be used, promoting pain-free episodes and positive outcomes rather than focusing on managing the pain. OBJECTIVE: This study aims to evaluate the feasibility (ie, acceptability, intervention adherence, and fidelity) and initial signals of efficacy of the PerPAIN app, an ecological momentary intervention for patients with CMSP. The app comprises digitalized monitoring using the experience sampling method (ESM) and feedback. In addition, the patients receive 3 microinterventions targeted at refocusing of attention on positive events. METHODS: In a microrandomized trial, we will recruit 35 patients with CMSP who will be offered the app for 12 weeks. Participants will be prompted to fill out 4 ESM monitoring questionnaires a day assessing information on their current context and the proximal outcome variables: absence of pain, positive mood, and subjective activity. Participants will be randomized daily and weekly to receive no feedback, verbal feedback, or visual feedback on proximal outcomes assessed by the ESM. In addition, the app will encourage participants to complete 3 microinterventions based on positive psychology and cognitive behavioral therapy techniques. These microinterventions are prompts to report joyful moments and everyday successes or to plan pleasant activities. After familiarizing themselves with each microintervention individually, participants will be randomized daily to receive 1 of the 3 exercises or none. We will assess whether the 2 feedback types and the 3 microinterventions increase proximal outcomes at the following time point. The microrandomized trial is part of the PerPAIN randomized controlled trial (German Clinical Trials Register DRKS00022792) investigating a personalized treatment approach to enhance treatment outcomes in CMSP. RESULTS: Approval was granted by the Ethics Committee II of the University of Heidelberg on August 4, 2020. Recruitment for the microrandomized trial began in May 2021 and is ongoing at the time of submission. By October 10, 2022, a total of 24 participants had been enrolled in the microrandomized trial. CONCLUSIONS: This trial will provide evidence on the feasibility of the PerPAIN app and the initial signals of efficacy of the different intervention components. In the next step, the intervention would need to be further refined and investigated in a definitive trial. This ecological momentary intervention presents a potential method for offering low-level accessible treatment to a wide range of people, which could have substantial implications for public health by reducing disease burden of chronic pain in the population. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/43376.

Artificial intelligence directed development of a digital twin to measure soft tissue shift during head and neck surgery.

Digital twins derived from 3D scanning data were developed to measure soft tissue deformation in head and neck surgery by an artificial intelligence approach. This framework was applied suggesting feasibility of soft tissue shift detection as a hitherto unsolved problem. In a pig head cadaver model 104 soft tissue resection had been performed. The surface of the removed soft tissue (RTP) and the corresponding resection cavity (RC) was scanned (N = 416) to train an artificial intelligence (AI) with two different 3D object detectors (HoloLens 2; ArtecEva). An artificial tissue shift (TS) was created by changing the tissue temperature from 7,91±4,1°C to 36,37±1,28°C. Digital twins of RTP and RC in cold and warm conditions had been generated and volumes were calculated based on 3D surface meshes. Significant differences in number of vertices created by the different 3D scanners (HoloLens2 51313 vs. ArtecEva 21694, p<0.0001) hence result in differences in volume measurement of the RTC (p = 0.0015). A significant TS could be induced by changing the temperature of the tissue of RC (p = 0.0027) and RTP (p = <0.0001). RC showed more correlation in TS by heating than RTP with a volume increase of 3.1 µl or 9.09% (p = 0.449). Cadaver models are suitable for training a machine learning model for deformable registration through creation of a digital twin. Despite different point cloud densities, HoloLens and ArtecEva provide only slightly different estimates of volume. This means that both devices can be used for the task.TS can be simulated and measured by temperature change, in which RC and RTP react differently. This corresponds to the clinical behaviour of tumour and resection cavity during surgeries, which could be used for frozen section management and a range of other clinical applications.

A theoretical study of H2 O2 as the surrogate of dose in minibeam radiotherapy, with a diffusion model considering radical removal process.

BACKGROUND: Minibeam radiation therapy (MBRT) is an innovative dose delivery method with the potential to spare normal tissue while achieving similar tumor control as conventional radiotherapy. However, it is difficult to use a single dose parameter, such as mean dose, to compare different patterns of MBRT due to the spatially fractionated radiation. Also, the mechanism leading to the biological effects is still unknown. PURPOSE: This study aims to demonstrate that the hydrogen peroxide (H2 O2 ) distribution could serve as a surrogate of dose distribution when comparing different patterns of MBRT. METHODS: A free diffusion model (FDM) for H2 O2 developed with Fick's second law was compared with a previously published model based on Monte Carlo & convolution method. Since cells form separate compartments that can eliminate H2 O2 radicals diffusing inside the cell, a term describing the elimination was introduced into the equation. The FDM and the diffusion model considering removal (DMCR) were compared by simulating various dose rate irradiation schemes and uniform irradiation. Finally, the DMCR was compared with previous microbeam and minibeam animal experiments. RESULTS: Compared with a previous Monte Carlo & Convolution method, this analytical method provides more accurate results. Furthermore, the new model shows H2 O2 concentration distribution instead of the time to achieve a certain H2 O2 uniformity. The comparison between FDM and DMCR showed that H2 O2 distribution from FDM varied with dose rate irradiation, while DMCR had consistent results. For uniform irradiation, FDM resulted in a Gaussian distribution, while the H2 O2 distribution from DMCR was close to the dose distribution. The animal studies' evaluation showed a correlation between the H2 O2 concentration in the valley region and treatment outcomes. CONCLUSION: DMCR is a more realistic model for H2 O2 simulation than the FDM. In addition, the H2 O2 distribution can be a good surrogate of dose distribution when the minibeam effect could be observed.

Simulation-to-real generalization for deep-learning-based refraction-corrected ultrasound tomography image reconstruction.

Objective. The image reconstruction of ultrasound computed tomography is computationally expensive with conventional iterative methods. The fully learned direct deep learning reconstruction is promising to speed up image reconstruction significantly. However, for direct reconstruction from measurement data, due to the lack of real labeled data, the neural network is usually trained on a simulation dataset and shows poor performance on real data because of the simulation-to-real gap.Approach. To improve the simulation-to-real generalization of neural networks, a series of strategies are developed including a Fourier-transform-integrated neural network, measurement-domain data augmentation methods, and a self-supervised-learning-based patch-wise preprocessing neural network. Our strategies are evaluated on both the simulation dataset and real measurement datasets from two different prototype machines.Main results. The experimental results show that our deep learning methods help to improve the neural networks' robustness against noise and the generalizability to real measurement data.Significance. Our methods prove that it is possible for neural networks to achieve superior performance to traditional iterative reconstruction algorithms in imaging quality and allow for real-time 2D-image reconstruction. This study helps pave the path for the application of deep learning methods to practical ultrasound tomography image reconstruction based on simulation datasets.

Augmented reality during parotid surgery: real-life evaluation of voice control of a head mounted display.

PURPOSE: Augmented Reality can improve surgical planning and performance in parotid surgery. For easier application we implemented a voice control manual for our augmented reality system. The aim of the study was to evaluate the feasibility of the voice control in real-life situations. METHODS: We used the HoloLens 1® (Microsoft Corporation) with a special speech recognition software for parotid surgery. The evaluation took place in a audiometry cubicle and during real surgical procedures. Voice commands were used to display various 3D structures of the patient with the HoloLens 1®. Commands had different variations (male/female, 65 dB SPL)/louder, various structures). RESULTS: In silence, 100% of commands were recognized. If the volume of the operation room (OR) background noise exceeds 42 dB, the recognition rate decreases significantly, and it drops below 40% at > 60 dB SPL. With constant speech volume at 65 dB SPL male speakers had a significant better recognition rate than female speakers (p = 0.046). Higher speech volumes can compensate this effect. The recognition rate depends on the type of background noise. Mixed OR noise (52 dB(A)) reduced the detection rate significantly compared to single suction noise at 52 dB(A) (p ≤ 0.00001). The recognition rate was significantly better in the OR than in the audio cubicle (p = 0.00013 both genders, 0.0086 female, and 0.0036 male). CONCLUSIONS: The recognition rate of voice commands can be enhanced by increasing the speech volume and by singularizing ambient noises. The detection rate depends on the loudness of the OR noise. Male voices are understood significantly better than female voices.

Are We There Yet? The Value of Deep Learning in a Multicenter Setting for Response Prediction of Locally Advanced Rectal Cancer to Neoadjuvant Chemoradiotherapy.

This retrospective study aims to evaluate the generalizability of a promising state-of-the-art multitask deep learning (DL) model for predicting the response of locally advanced rectal cancer (LARC) to neoadjuvant chemoradiotherapy (nCRT) using a multicenter dataset. To this end, we retrained and validated a Siamese network with two U-Nets joined at multiple layers using pre- and post-therapeutic T2-weighted (T2w), diffusion-weighted (DW) images and apparent diffusion coefficient (ADC) maps of 83 LARC patients acquired under study conditions at four different medical centers. To assess the predictive performance of the model, the trained network was then applied to an external clinical routine dataset of 46 LARC patients imaged without study conditions. The training and test datasets differed significantly in terms of their composition, e.g., T-/N-staging, the time interval between initial staging/nCRT/re-staging and surgery, as well as with respect to acquisition parameters, such as resolution, echo/repetition time, flip angle and field strength. We found that even after dedicated data pre-processing, the predictive performance dropped significantly in this multicenter setting compared to a previously published single- or two-center setting. Testing the network on the external clinical routine dataset yielded an area under the receiver operating characteristic curve of 0.54 (95% confidence interval [CI]: 0.41, 0.65), when using only pre- and post-therapeutic T2w images as input, and 0.60 (95% CI: 0.48, 0.71), when using the combination of pre- and post-therapeutic T2w, DW images, and ADC maps as input. Our study highlights the importance of data quality and harmonization in clinical trials using machine learning. Only in a joint, cross-center effort, involving a multidisciplinary team can we generate large enough curated and annotated datasets and develop the necessary pre-processing pipelines for data harmonization to successfully apply DL models clinically.

Natural Language Processing in Diagnostic Texts from Nephropathology.

INTRODUCTION: This study investigates whether it is possible to predict a final diagnosis based on a written nephropathological description-as a surrogate for image analysis-using various NLP methods. METHODS: For this work, 1107 unlabelled nephropathological reports were included. (i) First, after separating each report into its microscopic description and diagnosis section, the diagnosis sections were clustered unsupervised to less than 20 diagnostic groups using different clustering techniques. (ii) Second, different text classification methods were used to predict the diagnostic group based on the microscopic description section. RESULTS: The best clustering results (i) could be achieved with HDBSCAN, using BoW-based feature extraction methods. Based on keywords, these clusters can be mapped to certain diagnostic groups. A transformer encoder-based approach as well as an SVM worked best regarding diagnosis prediction based on the histomorphological description (ii). Certain diagnosis groups reached F1-scores of up to 0.892 while others achieved weak classification metrics. CONCLUSION: While textual morphological description alone enables retrieving the correct diagnosis for some entities, it does not work sufficiently for other entities. This is in accordance with a previous image analysis study on glomerular change patterns, where some diagnoses are associated with one pattern, but for others, there exists a complex pattern combination.

Remote sensing array (RSA) for linac beam monitoring.

The purpose of the present work is to evaluate the feasibility of a novel real-time beam monitoring device for medical linacs which remotely senses charge carriers produced in air by the beam without intersecting and attenuating the beamline. The primary goal is to elaborate a theoretical concept of a possible detector geometry and underlying physical model that allows for determination of clinically relevant beam data in real time, namely MLC leaf positions and dose rate. The detector consists of two opposing electrode arrays arranged in two possible orientations around the beamline. Detection of charge carriers is governed by electromagnetic principles described by Shockley-Ramo theorem. Ions produced by ionization of the air column upstream of patient move laterally in an external electric field. According to the method of images, mirror charges and mirror currents are formed in the strip electrodes. Determination of MU rate and MLC positions using the measured signal requires solution of an inverse problem. In the present work we adopted a Least-Square approach and characterized detector response and sensitivity to detection of beam properties for different electrode geometries and MLC shapes. Results were dependent on MLC field shape and the leaf position within the active volume. The accuracy of determination of leaf positions were in the sub-mm range (up to 0.25-1 mm). Additionally, detector sensitivity was quantified by simulating ions/pulse delivered with a radiation transport deterministic computation in 1D in CEPXS/ONEDANT. For a 6 MV linac pulse, signal amplitude per pulse was estimated to be in the lower pA to fA range. We computationally demonstrated feasibility of the remote sensing detector capable of measuring beam parameters such as MLC leaf positions and dose range for each pulse. Future work should focus on optimizing the electrode geometry to increase sensitivity and better reconstruction algorithms to provide more accurate solutions of the inverse problem.

Normalization of HE-stained histological images using cycle consistent generative adversarial networks.

BACKGROUND: Histological images show strong variance (e.g. illumination, color, staining quality) due to differences in image acquisition, tissue processing, staining, etc. This can impede downstream image analysis such as staining intensity evaluation or classification. Methods to reduce these variances are called image normalization techniques. METHODS: In this paper, we investigate the potential of CycleGAN (cycle consistent Generative Adversarial Network) for color normalization in hematoxylin-eosin stained histological images using daily clinical data with consideration of the variability of internal staining protocol variations. The network consists of a generator network GB that learns to map an image X from a source domain A to a target domain B, i.e. GB:XA→XB. In addition, a discriminator network DB is trained to distinguish whether an image from domain B is real or generated. The same process is applied to another generator-discriminator pair (GA,DA), for the inverse mapping GA:XB→XA. Cycle consistency ensures that a generated image is close to its original when being mapped backwards (GA(GB(XA))≈XA and vice versa). We validate the CycleGAN approach on a breast cancer challenge and a follicular thyroid carcinoma data set for various stain variations. We evaluate the quality of the generated images compared to the original images using similarity measures. In addition, we apply stain normalization on pathological lymph node data from our institute and test the gain from normalization on a ResNet classifier pre-trained on the Camelyon16 data set. RESULTS: Qualitative results of the images generated by our network are compared to original color distributions. Our evaluation indicates that by mapping images to a target domain, the similarity training images from that domain improves up to 96%. We also achieve a high cycle consistency for the generator networks by obtaining similarity indices greater than 0.9. When applying the CycleGAN normalization to HE-stain images from our institute the kappa-value of the ResNet-model that is only trained on Camelyon16 data is increased more than 50%. CONCLUSIONS: CycleGANs have proven to efficiently normalize HE-stained images. The approach compensates for deviations resulting from image acquisition (e.g. different scanning devices) as well as from tissue staining (e.g. different staining protocols), and thus overcomes the staining variations in images from various institutions.The code is publicly available at https://github.com/m4ln/stainTransfer_CycleGAN_pytorch . The data set supporting the solutions is available at https://doi.org/10.11588/data/8LKEZF .

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.

Feasibility of interstitial stepping-source electronic brachytherapy to locally inoperable tumors.

PURPOSE: Radiotherapy is the mainstay in the treatment of locally inoperable tumors. Interstitial electronic needle-based kilovoltage brachytherapy (EBT) could be an economic alternative to high-dose-rate (HDR) brachytherapy or permanent seed implantation (PSI). In this work, we evaluated if locally inoperable tumors treated with PSI at our institution may be suitable for EBT. MATERIAL AND METHODS: A total of 10 post-interventional computed tomography (CT) scans of patients, who received PSI and simulated stepping-source EBT applied with Intrabeam system and needle applicator were used. EBT treatment planning software with 3-dimensional image and projection of applicator were applied for designing trajectories and establishing dwell positions. Dwell position doses were summarized, and doses covering 90% of the target volume (D90) achieved with stepping-source EBT were compared to those of PSI. Additionally, conformality of dose distributions and total irradiation time were assessed using conformation number (CN) or conformal index (COIN). RESULTS: In all patients, D90 of EBT exceeded the prescribed dose or D90 of PSI on average by 4.7% or 21.3% relative to the prescribed dose, respectively. Mean number of trajectories was 5.0 for EBT and 6.9 for PSI. Average CN/COIN for EBT was 0.69, with a mean irradiation time of 27.8 minutes for standardized dose of 13 Gy. CONCLUSIONS: Stepping-source EBT allowed for a conformal treatment of inoperable interstitial tumors with similar D90. Fewer trajectories were required for EBT in majority of cases.

Imaging and Characterization of Sustained Gadolinium Nanoparticle Release from Next Generation Radiotherapy Biomaterial.

Smart radiotherapy biomaterials (SRBs) present a new opportunity to enhance image-guided radiotherapy while replacing routinely used inert radiotherapy biomaterials like fiducials. In this study the potential of SRBs loaded with gadolinium-based nanoparticles (GdNPs) is investigated for magnetic resonance imaging (MRI) contrast. GdNP release from SRB is quantified and modelled for accurate prediction. SRBs were manufactured similar to fiducials, with a cylindrical shell consisting of poly(lactic-co-glycolic) acid (PLGA) and a core loaded with GdNPs. Magnetic resonance imaging (MRI) contrast was investigated at 7T in vitro (in agar) and in vivo in subcutaneous tumors grown with the LLC1 lung cancer cell line in C57/BL6 mice. GdNPs were quantified in-phantom and in tumor and their release was modelled by the Weibull distribution. Gd concentration was linearly fitted to the R1 relaxation rate with a detection limit of 0.004 mmol/L and high confidence level (R2 = 0.9843). GdNP loaded SRBs in tumor were clearly visible up to at least 14 days post-implantation. Signal decrease during this time showed GdNP release in vivo, which was calculated as 3.86 ± 0.34 µg GdNPs release into the tumor. This study demonstrates potential and feasibility for SRBs with MRI-contrast, and sensitive GdNP quantification and release from SRBs in a preclinical animal model. The feasibility of monitoring nanoparticle (NP) concentration during treatment, allowing dynamic quantitative treatment planning, is also discussed.

Ultrasound transmission tomography image reconstruction with a fully convolutional neural network.

Image reconstruction of ultrasound computed tomography based on the wave equation is able to show much more structural details than simpler ray-based image reconstruction methods. However, to invert the wave-based forward model is computationally demanding. To address this problem, we develop an efficient fully learned image reconstruction method based on a convolutional neural network. The image is reconstructed via one forward propagation of the network given input sensor data, which is much faster than the reconstruction using conventional iterative optimization methods. To transform the ultrasound measured data in the sensor domain into the reconstructed image in the image domain, we apply multiple down-scaling and up-scaling convolutional units to efficiently increase the number of hidden layers with a large receptive and projective field that can cover all elements in inputs and outputs, respectively. For dataset generation, a paraxial approximation forward model is used to simulate ultrasound measurement data. The neural network is trained with a dataset derived from natural images in ImageNet and tested with a dataset derived from medical images in OA-Breast Phantom dataset. Test results show the superior efficiency of the proposed neural network to other reconstruction algorithms including popular neural networks. When compared with conventional iterative optimization algorithms, our neural network can reconstruct a 110 × 86 image more than 20 times faster on a CPU and 1000 times faster on a GPU with comparable image quality and is also more robust to noise.

Diffuse domain method for needle insertion simulations.

We present a new strategy for needle insertion simulations without the necessity of meshing. A diffuse domain approach on a regular grid is applied to overcome the need for an explicit representation of organ boundaries. A phase field function captures the transition of tissue parameters and boundary conditions are imposed implicitly. Uncertainties of a volume segmentation are translated in the width of the phase field, an approach that is novel and overcomes the problem of defining an accurate segmentation boundary. We perform a convergence analysis of the diffuse elastic equation for decreasing phase field width, compare our results to deformation fields received from conforming mesh simulations and analyze the diffuse linear elastic equation for different widths of material interfaces. Then, the approach is applied to computed tomography data of a patient with liver tumors. A three-class U-Net is used to automatically generate tissue probability maps serving as phase field functions for the transition of elastic parameters between different tissues. The needle tissue interaction forces are approximated by the absolute gradient of a phase field function, which eliminates the need for explicit boundary parameterization and collision detection at the needle-tissue interface. The results show that the deformation field of the diffuse domain approach is comparable to the deformation of a conforming mesh simulation. Uncertainties of tissue boundaries are included in the model and the simulation can be directly performed on the automatically generated voxel-based probability maps. Thus, it is possible to perform easily implementable patient-specific elastomechanical simulations directly on voxel data.