Search

Magnetic resonance imaging (MRI) provides enhanced soft tissue contrast and high spatial resolution. However, the relationship between intensity values among soft tissues in MRI is inconsistent, even when obtained under the same conditions (e.g., vendors and acquisition protocols). This inconsistency hinders accurate medical image segmentation and disease classification. Therefore, we propose a framework to harmonize multi-vendor MRI using a novel radiomics approach for robust segmentation. The proposed model comprises a cycle-consistent adversarial network (CycleGAN)-based network and a segmentation network. The CycleGAN-based network harmonizes MRI with the support of a radiomics-based method (radiomic feature (RF) loss function newly designed for this study). The segmentation network encourages the CycleGAN-based network to enhance intervertebral disc (IVD) segmentation features using dice loss functions during harmonization. Furthermore, publicly available datasets and diverse MRI scans provided by a collaborating hospital were used to make our model more robust to MRI variability. The proposed model was evaluated for segmentation using the Dice coefficient, intersection-over-union (IoU), F1 score, precision, and recall. It outperformed other segmentation methods (Dice = 0.920, IoU = 0.853, F1 score = 0.920, precision = 0.940, and recall = 0.902), even on diverse test datasets with disease information. The harmonization performance was assessed using the relative error of the RF values between the target (standard) and harmonized data. It achieved the four best scores ( $\mathrm {\approx 0}$ ) among the five features in a relative error of RF compared to other harmonization methods (e.g., conventional histogram-based method and deep learning model).

Although spinal surgeries with minimal incisions and a minimal amount of X-ray exposure (MIMA) mostly occur in a prone posture on a Wilson table, the prone posture’s effects on spinal muscles have not been investigated. Thus, this study used ultrasound shear-wave elastography (SWE) to compare the material properties of the erector spinae and multifidus muscles when subjects lay on the Wilson table used for spinal surgery and the flat table as a control condition. Thirteen male subjects participated in the study. Using ultrasound SWE, the shear elastic moduli (SEM) of the erector spinae and multifidus muscles were investigated. Significant increases were found in the SEM of erector spinae muscle 1, erector spinae muscle 2, and multifidus muscles on the Wilson table (W) compared to in the flat table (F; W:22.19 ± 7.15 kPa, F:10.40 ± 3.20 kPa, p < 0.001; W:12.10 ± 3.31 kPa, F: 7.17 ± 1.71 kPa, p < 0.001; W: 18.39 ± 4.80 kPa, F: 11.43 ± 2.81 kPa, p < 0.001, respectively). Our results indicate that muscle material properties measured by SWE can be changed due to table posture, which should be considered in biomechanical modeling by guiding surgical planning to develop minimal-incision surgical procedures.

Knowing the material properties of the musculoskeletal soft tissue could be important to develop rehabilitation therapy and surgical procedures. However, there is a lack of devices and information on the viscoelastic properties of soft tissues around the lumbar spine. The goal of this study was to develop a portable quantifying device for providing strain and stress curves of muscles and ligaments around the lumbar spine at various stretching speeds. Each sample was conditioned and applied for 20 repeatable cyclic 5 mm stretch-and-relax trials in the direction and perpendicular direction of the fiber at 2, 3 and 5 mm/s. Our device successfully provided the stress and strain curve of the samples and our results showed that there were significant effects of speed on the young’s modulus of the samples (p < 0.05). Compared to the expensive commercial device, our lower-cost device provided comparable stress and strain curves of the sample. Based on our device and findings, various sizes of samples can be measured and viscoelastic properties of the soft tissues can be obtained. Our portable device and approach can help to investigate young’s modulus of musculoskeletal soft tissues conveniently, and can be a basis for developing a material testing device in a surgical room or various lab environments.

Cardiovascular-related diseases are one of the leading causes of death worldwide. An understanding of heart movement based on images plays a vital role in assisting postoperative procedures and processes. In particular, if shape information can be provided in real-time using electrocardiogram (ECG) signal information, the corresponding heart movement information can be used for cardiovascular analysis and imaging guides during surgery. In this paper, we propose a 3D+t cardiac coronary artery model which is rendered in real-time, according to the ECG signal, where hierarchical cage-based deformation modeling is used to generate the mesh deformation used during the procedure. We match the blood vessel’s lumen obtained from the ECG-gated 3D+t CT angiography taken at multiple cardiac phases, in order to derive the optimal deformation. Splines for 3D deformation control points are used to continuously represent the obtained deformation in the multi-view, according to the ECG signal. To verify the proposed method, we compare the manually segmented lumen and the results of the proposed method for eight patients. The average distance and dice coefficient between the two models were 0.543 mm and 0.735, respectively. The required time for registration of the 3D coronary artery model was 23.53 s/model. The rendering speed to derive the model, after generating the 3D+t model, was faster than 120 FPS.

Finite element (FE) modeling of the passive ligamentous spine is widely used to assess various biomechanical behaviors. Currently, FE models that incorporate the vertebrae, ligaments, and the personalized geometry of the bony spine may be used in conjunction with external loads from the muscles. However, while the muscles place a load (moment) on the spine and support it simultaneously, the effect of the passive support from the adjacent spinal muscles has not been considered. This study thus aims to investigate the effect of passive support from the psoas major, quadratus lumborum, and erector muscles on the range of motion (RoM) and intradiscal pressure (IDP) of the lumbar spine. Various L2-sacrum spinal models that differed only in their muscle properties were constructed and loaded with a pure moment (2.5–15.0 Nm) alone or combined with a compressive (440 or 1000 N) follower load. The RoM and IDP of the model that excluded the effect of muscles closely matched previous FE results under the corresponding load conditions. When the muscles (40–160 kPa) were included in the FE model, the RoM at L2 was reduced by up to 6.57% under a pure moment (10 Nm). The IDP was reduced by up to 6.45% under flexion and 6.84% under extension. It was also found that the erector muscles had a greater effect than the psoas major and quadratus muscles.

The creation of 3D models for cardiac mapping systems is time-consuming, and the models suffer from issues with repeatability among operators. The present study aimed to construct a double-shaped model composed of the left ventricle and left atrium. We developed cascaded-regression-based segmentation software with probabilistic point and appearance correspondence. Group-wise registration of point sets constructs the point correspondence from probabilistic matches, and the proposed method also calculates appearance correspondence from these probabilistic matches. Final point correspondence of group-wise registration constructed independently for three surfaces of the double-shaped model. Stochastic appearance selection of cascaded regression enables the effective construction in the aspect of memory usage and computation time. The two correspondence construction methods of active appearance models were compared in terms of the paired segmentation of the left atrium (LA) and left ventricle (LV). The proposed method segmented 35 cardiac CTs in six-fold cross-validation, and the symmetric surface distance (SSD), Hausdorff distance (HD), and Dice coefficient (DC), were used for evaluation. The proposed method produced 1.88 ± 0.37 mm of LV SSD, 2.25 ± 0.51 mm* of LA SSD, and 2.06 ± 0.34 mm* of the left heart (LH) SSD. Additionally, DC was 80.45% ± 4.27%***, where * p < 0.05, ** p < 0.01, and *** p < 0.001. All p values derive from paired t-tests comparing iterative closest registration with the proposed method. In conclusion, the authors developed a cascaded regression framework for 3D cardiac CT segmentation.

Augmented reality (AR) surgical navigation systems have attracted considerable attention as they assist medical professionals in visualizing the location of ailments within the human body that are not readily seen with the naked eye. Taking medical imaging with a parallel C-shaped arm (C-arm) as an example, surgical sites are typically targeted using an optical tracking device and a fiducial marker in real-time. These markers then guide operators who are using a multifunctional endoscope apparatus by signaling the direction or distance needed to reach the affected parts of the body. In this way, fiducial markers are used to accurately protect the vessels and nerves exposed during the surgical process. Although these systems have already shown potential for precision implantation, delamination of the fiducial marker, which is a critical component of the system, from human skin remains a challenge due to a mechanical mismatch between the marker and skin, causing registration problems that lead to poor position alignments and surgical degradation. To overcome this challenge, the mechanical modulus and stiffness of the marker patch should be lowered to approximately 150 kPa, which is comparable to that of the epidermis, while improving functionality. Herein, we present a skin-conformal, stretchable yet breathable fiducial marker for the application in AR-based surgical navigation systems. By adopting pore patterns, we were able to create a fiducial marker with a skin-like low modulus and breathability. When attached to the skin, the fiducial marker was easily identified using optical recognition equipment and showed skin-conformal adhesion when stretched and shrunk repeatedly. As such, we believe the marker would be a good fiducial marker candidate for patients under surgical navigation systems.

Image-based surgical techniques can support minimally invasive keyhole neurosurgery through the segmentation of computed tomography or magnetic resonance imaging, optimal surgical planning, and surgical guidance. Keyhole neurosurgery requires automatic surgical planning and accuracy guidance for a target of the non-invasive or minimally invasive treatment of an anatomical structure. For supporting these requirements, we propose a markerless surgical robotic guidance system for keyhole neurosurgery. Our proposed robotic system consists of a six-degree of freedom (DoF) robotic arm, a needle guidance device, a mobile cart, a control workstation, and a 3D surface scanner. A 3D surface scanner is used for markerless registration between a preoperative reference image and a patient’s face in the operating room. This system configuration has merits of minimally invasive surgery. Traditional frame-based registration requires a stereotactic frame onto the patient’s head. Also, it requires an additional CT/MRI scanning to identify the coordinate system of the frame. However, the proposed markerless system can register without an additional CT scan by using a surface scanner. First, we show the system configuration to perform robotic-assisted surgery precisely. Next, we analyze subcomponents and calibrations such as the repeatability of the robotic arm, hand-tool calibration error for the relationship between needle guidance device and flange of the robotic arm, 3D scanning accuracy of the surface scanner, and hand-eye calibration error for the relationship between the surface scanner and the flange of the robotic arm. After calibrating each component of proposed system, the system accuracy is affected by the propagation of small errors. We sampled 15 paths located in the working space to check and compensate the residual transformation. We also test the total system accuracy with a phantom model; the proposed system allows us to obtain translation error (0.75 ± 0.38 mm) and rotation error (0.85 ± 0.16°) as a residual median error, which satisfies a surgical requirement (< 1 mm). With placements of medical devices to the planned pathway, it is crucial to automatically calculate a surgical path for keyhole neurosurgery. Thus, we proposed a path-planning algorithm that can support surgical decision-making process for keyhole neurosurgery. In the proposed algorithm, a possible trajectory is defined by a pair of a possible entry point on skull surface and a target point inside of target region. Then, the surgical paths are evaluated to satisfy the surgical rules such as the avoidance of vital organs, the maximization of the coverage of the target volume, and orthogonal insertion into the scalp surface.

Venipuncture is an important health diagnosis process. Although venipuncture is one of the most commonly performed procedures in medical environments, locating the veins of infants, obese, anemic, or colored patients is still an arduous task even for skilled practitioners. To solve this problem, several devices using infrared light have recently become commercially available. However, such devices for venipuncture share a common drawback, especially when visualizing deep veins or veins of a thick part of the body like the cubital fossa. This paper proposes a new vein-visualizing device applying a new penetration method using near-infrared (NIR) light. The light module is attached directly on to the declared area of the skin. Then, NIR beam is rayed from two sides of the light module to the vein with a specific angle. This gives a penetration effect. In addition, through an image processing procedure, the vein structure is enhanced to show it more accurately. Through a phantom study, the most effective penetration angle of the NIR module is decided. Additionally, the feasibility of the device is verified through experiments in vivo. The prototype allows us to visualize the vein patterns of thicker body parts, such as arms.