Showing papers presented at "Computer Assisted Radiology and Surgery in 2014"

Toward embedded detection of polyps in WCE images for early diagnosis of colorectal cancer

Abstract: Purpose Wireless capsule endoscopy (WCE) is commonly used for noninvasive gastrointestinal tract evaluation, including the detection of mucosal polyps. A new embeddable method for polyp detection in wireless capsule endoscopic images was developed and tested.

Surgical process modelling: a review

Abstract: Surgery is continuously subject to technological and medical innovations that are transforming daily surgical routines. In order to gain a better understanding and description of surgeries, the field of surgical process modelling (SPM) has recently emerged. The challenge is to support surgery through the quantitative analysis and understanding of operating room activities. Related surgical process models can then be introduced into a new generation of computer-assisted surgery systems. In this paper, we present a review of the literature dealing with SPM. This methodological review was obtained from a search using Google Scholar on the specific keywords: “surgical process analysis”, “surgical process model” and “surgical workflow analysis”. This paper gives an overview of current approaches in the field that study the procedural aspects of surgery. We propose a classification of the domain that helps to summarise and describe the most important components of each paper we have reviewed, i.e., acquisition, modelling, analysis, application and validation/evaluation. These five aspects are presented independently along with an exhaustive list of their possible instantiations taken from the studied publications. This review allows a greater understanding of the SPM field to be gained and introduces future related prospects.

Brain tumor detection and segmentation in a CRF (conditional random fields) framework with pixel-pairwise affinity and superpixel-level features.

Abstract: Detection and segmentation of a brain tumor such as glioblastoma multiforme (GBM) in magnetic resonance (MR) images are often challenging due to its intrinsically heterogeneous signal characteristics. A robust segmentation method for brain tumor MRI scans was developed and tested. Simple thresholds and statistical methods are unable to adequately segment the various elements of the GBM, such as local contrast enhancement, necrosis, and edema. Most voxel-based methods cannot achieve satisfactory results in larger data sets, and the methods based on generative or discriminative models have intrinsic limitations during application, such as small sample set learning and transfer. A new method was developed to overcome these challenges. Multimodal MR images are segmented into superpixels using algorithms to alleviate the sampling issue and to improve the sample representativeness. Next, features were extracted from the superpixels using multi-level Gabor wavelet filters. Based on the features, a support vector machine (SVM) model and an affinity metric model for tumors were trained to overcome the limitations of previous generative models. Based on the output of the SVM and spatial affinity models, conditional random fields theory was applied to segment the tumor in a maximum a posteriori fashion given the smoothness prior defined by our affinity model. Finally, labeling noise was removed using “structural knowledge” such as the symmetrical and continuous characteristics of the tumor in spatial domain. The system was evaluated with 20 GBM cases and the BraTS challenge data set. Dice coefficients were computed, and the results were highly consistent with those reported by Zikic et al. (MICCAI 2012, Lecture notes in computer science. vol 7512, pp 369–376, 2012). A brain tumor segmentation method using model-aware affinity demonstrates comparable performance with other state-of-the art algorithms.

Surgical planning tool for robotically assisted hearing aid implantation

Abstract: For the facilitation of minimally invasive robotically performed direct cochlea access (DCA) procedure, a surgical planning tool which enables the surgeon to define landmarks for patient-to-image registration, identify the necessary anatomical structures and define a safe DCA trajectory using patient image data (typically computed tomography (CT) or cone beam CT) is required. To this end, a dedicated end-to-end software planning system for the planning of DCA procedures that addresses current deficiencies has been developed. Efficient and robust anatomical segmentation is achieved through the implementation of semiautomatic algorithms; high-accuracy patient-to-image registration is achieved via an automated model-based fiducial detection algorithm and functionality for the interactive definition of a safe drilling trajectory based on case-specific drill positioning uncertainty calculations was developed. The accuracy and safety of the presented software tool were validated during the conduction of eight DCA procedures performed on cadaver heads. The plan for each ear was completed in less than 20 min, and no damage to vital structures occurred during the procedures. The integrated fiducial detection functionality enabled final positioning accuracies of $$0.15\pm 0.08$$ mm. Results of this study demonstrated that the proposed software system could aid in the safe planning of a DCA tunnel within an acceptable time.

Assessing performance in brain tumor resection using a novel virtual reality simulator

Abstract: NeuroTouch is a virtual reality (VR) simulator developed for neurosurgical skill training. Validation demonstrating that the system is useful and reliable is required for formal adoption into training curriculums. Face and content validity have been demonstrated for some neurosurgical simulators, but construct validity remains difficult to establish. A pilot validation study was conducted for a NeuroTouch training exercise. Participants completed the internal resection of a simulated convexity meningioma and filled out questionnaires to provide feedback on the experience. Performance metrics included volume of tissues removed, tool path lengths, duration of excessive forces applied and efficient use of the aspirator. Results were analyzed according to participants’ level of training, gender, handedness, surgical experience in meningioma removal and hours/week playing musical instruments or video games. Seventy-two participants (10 medical students, 18 junior residents and 44 senior residents) were enrolled. Analyses demonstrated statistically significant increase in tumor removed and efficiency of ultrasonic aspirator use between medical students and residents, but not between junior and senior residents. After covariate adjustment for the number of meningioma cases operated on, multivariate analysis of the level of training became nonsignificant. Participants judged the exercise appropriate and realistic, desiring use of the system in current training programs. We have conducted a pilot validation study for the NeuroTouch tumor resection scenario and demonstrated for the first time, face, content and construct validity of a VR neurosurgical simulation exercise. Future full-scale studies will be conducted in noncompetitive settings and incorporate expert participants.

Optimization of breast mass classification using sequential forward floating selection (SFFS) and a support vector machine (SVM) model.

Abstract: Improving radiologists’ performance in classification between malignant and benign breast lesions is important to increase cancer detection sensitivity and reduce false-positive recalls. For this purpose, developing computer-aided diagnosis schemes has been attracting research interest in recent years. In this study, we investigated a new feature selection method for the task of breast mass classification. We initially computed 181 image features based on mass shape, spiculation, contrast, presence of fat or calcifications, texture, isodensity, and other morphological features. From this large image feature pool, we used a sequential forward floating selection (SFFS)-based feature selection method to select relevant features and analyzed their performance using a support vector machine (SVM) model trained for the classification task. On a database of 600 benign and 600 malignant mass regions of interest, we performed the study using a tenfold cross-validation method. Feature selection and optimization of the SVM parameters were conducted on the training subsets only. The area under the receiver operating characteristic curve $$(\hbox {AUC}) = 0.805\pm 0.012$$ was obtained for the classification task. The results also showed that the most frequently selected features by the SFFS-based algorithm in tenfold iterations were those related to mass shape, isodensity, and presence of fat, which are consistent with the image features frequently used by radiologists in the clinical environment for mass classification. The study also indicated that accurately computing mass spiculation features from the projection mammograms was difficult, and failed to perform well for the mass classification task due to tissue overlap within the benign mass regions. In conclusion, this comprehensive feature analysis study provided new and valuable information for optimizing computerized mass classification schemes that may have potential to be useful as a “second reader” in future clinical practice.

Volumetric texture features from higher-order images for diagnosis of colon lesions via CT colonography

Abstract: Purpose Differentiation of colon lesions according to underlying pathology, e.g., neoplastic and non-neoplastic lesions, is of fundamental importance for patient management. Image intensity-based textural features have been recognized as useful biomarker for the differentiation task. In this paper, we introduce texture features from higher-order images, i.e., gradient and curvature images, beyond the intensity image, for that task.

Experimental evaluation of ultrasound-guided 3D needle steering in biological tissue

Abstract: Purpose In this paper, we present a system capable of automatically steering bevel tip flexible needles under ultrasound guidance toward stationary and moving targets in gelatin phantoms and biological tissue while avoiding stationary and moving obstacles. We use three-dimensional (3D) ultrasound to track the needle tip during the procedure. Methods Our system uses a fast sampling-based path planner to compute and periodically update a feasible path to the target that avoids obstacles. We then use a novel control algorithm to steer the needle along the path in a manner that reduces the number of needle rotations, thus reducing tissue damage. We present experimental results for needle insertion procedures for both stationary and moving targets and obstacles for up to 90 mm of needle insertion. Results We obtained a mean targeting error of 0.32±0.10 and 0.38±0.19 mm in gelatin-based phantom and biological tissue, respectively. Conclusions The achieved submillimeter accuracy suggests that our approach is sufficient to target the smallest lesions ( ϕ 2 mm) that can be detected using state-of-the-art ultrasound imaging systems.

ROC operating point selection for classification of imbalanced data with application to computer-aided polyp detection in CT colonography.

Abstract: Purpose Computer-aided detection and diagnosis (CAD) of colonic polyps always faces the challenge of classifying imbalanced data. In this paper, three new operating point selection strategies based on receiver operating characteristic curve are proposed to address the problem.

Variable dose interplay effects across radiosurgical apparatus in treating multiple brain metastases

Abstract: Normal brain tissue doses have been shown to be strongly apparatus dependent for multi-target stereotactic radiosurgery. In this study, we investigated whether inter-target dose interplay effects across contemporary radiosurgical treatment platforms are responsible for such an observation. For the study, subsets ( $$n = 3, 6, 9,$$ and 12) of a total of 12 targets were planned at six institutions. Treatment platforms included the (1) Gamma Knife Perfexion (PFX), (2) CyberKnife, (3) Novalis linear accelerator equipped with a 3.0-mm multi-leaf collimator (MLC), and the (4) Varian Truebeam flattening-filter-free (FFF) linear accelerator also equipped with a 2.5 mm MLC. Identical dose–volume constraints for the targets and critical structures were applied for each apparatus. All treatment plans were developed at individual centers, and the results were centrally analyzed. We found that dose–volume constraints were satisfied by each apparatus with some differences noted in certain structures such as the lens. The peripheral normal brain tissue doses were lowest for the PFX and highest for TrueBeam FFF and CyberKnife treatment plans. Comparing the volumes of normal brain receiving 12 Gy, TrueBeam FFF, Novalis, and CyberKnife were 180–290 % higher than PFX. The mean volume of normal brain-per target receiving 4-Gy increased by approximately 3.0 cc per target for TrueBeam, 2.7 cc per target for CyberKnife, 2.0 cc per target for Novalis, and 0.82 cc per target for PFX. The beam-on time was shortest with the TrueBeam FFF (e.g., 6–9 min at a machine output rate of 1,200 MU/min) and longest for the PFX (e.g., 50–150 mins at a machine output rate of 350 cGy/min). The volumes of normal brain receiving 4 and 12 Gy were higher, and increased more swiftly per target, for Linac-based SRS platforms than for PFX. Treatment times were shortest with TrueBeam FFF.

GPU accelerated segmentation and centerline extraction of tubular structures from medical images.

Abstract: To create a fast and generic method with sufficient quality for extracting tubular structures such as blood vessels and airways from different modalities (CT, MR and US) and organs (brain, lungs and liver) by utilizing the computational power of graphic processing units (GPUs). A cropping algorithm is used to remove unnecessary data from the datasets on the GPU. A model-based tube detection filter combined with a new parallel centerline extraction algorithm and a parallelized region growing segmentation algorithm is used to extract the tubular structures completely on the GPU. Accuracy of the proposed GPU method and centerline algorithm is compared with the ridge traversal and skeletonization/thinning methods using synthetic vascular datasets. The implementation is tested on several datasets from three different modalities: airways from CT, blood vessels from MR, and 3D Doppler Ultrasound. The results show that the method is able to extract airways and vessels in 3–5 s on a modern GPU and is less sensitive to noise than other centerline extraction methods. Tubular structures such as blood vessels and airways can be extracted from various organs imaged by different modalities in a matter of seconds, even for large datasets.

Delineation of blood vessels in pediatric retinal images using decision trees-based ensemble classification

Abstract: Automatic segmentation of the retinal vasculature is a first step in computer-assisted diagnosis and treatment planning. The extraction of retinal vessels in pediatric retinal images is challenging because of comparatively wide arterioles with a light streak running longitudinally along the vessel’s center, the central vessel reflex. A new method for automatic segmentation was developed and tested. A supervised method for retinal vessel segmentation in the images of multi-ethnic school children was developed based on ensemble classifier of bootstrapped decision trees. A collection of dual Gaussian, second derivative of Gaussian and Gabor filters, along with the generalized multiscale line strength measure and morphological transformation is used to generate the feature vector. The feature vector encodes information to handle the normal vessels as well as the vessels with the central reflex. The methodology is evaluated on CHASE_DB1, a relatively new public retinal image database of multi-ethnic school children, which is a subset of retinal images from the Child Heart and Health Study in England (CHASE) dataset. The segmented retinal images from the CHASE_DB1 database produced best case accuracy, sensitivity and specificity of 0.96, 0.74 and 0.98, respectively, and worst case measures of 0.94, 0.67 and 0.98, respectively. A new retinal blood vessel segmentation algorithm was developed and tested with a shared database. The observed accuracy, speed, robustness and simplicity suggest that the algorithm may be a suitable tool for automated retinal image analysis in large population-based studies.

Multi-trajectories automatic planner for StereoElectroEncephaloGraphy (SEEG)

Abstract: StereoElectroEncephaloGraphy (SEEG) is done to identify the epileptogenic zone of the brain using several multi-lead electrodes whose positions in the brain are pre-operatively defined. Intracranial hemorrhages due to disruption of blood vessels can cause major complications of this procedure ( $$<$$ 1 %). In order to increase the intervention safety, we developed and tested planning tools to assist neurosurgeons in choosing the best trajectory configuration. An automated planning method was developed that maximizes the distance of the electrode from the vessels and avoids the sulci as entry points. The angle of the guiding screws is optimized to reduce positioning error. The planner was quantitatively and qualitatively compared with manually computed trajectories on 26 electrodes planned for three patients undergoing SEEG by four neurosurgeons. Quantitative comparison was performed computing for each trajectory using (a) the Euclidean distance from the closest vessel and (b) the incidence angle. Quantitative evaluation shows that automatic planned trajectories are safer in terms of distance from the closest vessel with respect to manually planned trajectories. Qualitative evaluation performed by four neurosurgeons showed that the automatically computed trajectories would have been preferred to manually computed ones in 30 % of the cases and were judged good or acceptable in about 86 % of the cases. A significant reduction in time required for planning was observed with the automated system (approximately 1/10). The automatic SEEG electrode planner satisfied the essential clinical requirements, by providing safe trajectories in an efficient timeframe.

Automated aorta segmentation in low-dose chest CT images

Abstract: Abnormalities of aortic surface and aortic diameter can be related to cardiovascular disease and aortic aneurysm. Computer-based aortic segmentation and measurement may aid physicians in related disease diagnosis. This paper presents a fully automated algorithm for aorta segmentation in low-dose non-contrast CT images. The original non-contrast CT scan images as well as their pre-computed anatomy label maps are used to locate the aorta and identify its surface. First a seed point is located inside the aortic lumen. Then, a cylindrical model is progressively fitted to the 3D image space to track the aorta centerline. Finally, the aortic surface is located based on image intensity information. This algorithm has been trained and tested on 359 low-dose non-contrast CT images from VIA-ELCAP and LIDC public image databases. Twenty images were used for training to obtain the optimal set of parameters, while the remaining images were used for testing. The segmentation result has been evaluated both qualitatively and quantitatively. Sixty representative testing images were used to establish a partial ground truth by manual marking on several axial image slices. Compared to ground truth marking, the segmentation result had a mean Dice Similarity Coefficient of 0.933 (maximum 0.963 and minimum 0.907). The average boundary distance between manual segmentation and automatic segmentation was 1.39 mm with a maximum of 1.79 mm and a minimum of 0.83 mm. Both qualitative and quantitative evaluations have shown that the presented algorithm is able to accurately segment the aorta in low-dose non-contrast CT images.

Clinical, functional, and radiographic outcomes following total knee arthroplasty with patient-specific instrumentation, computer-assisted surgery, and manual instrumentation: a short-term follow-up study

Abstract: Purpose The purpose of this study was to evaluate clinical, functional, and radiographic outcomes following total knee arthroplasty (TKA) performed with patient-specific instrumentation (PSI), computer-assisted surgery (CAS), and manual instruments at short-term follow-up.

Variable stiffness outer sheath with “Dragon skin” structure and negative pneumatic shape-locking mechanism

Abstract: Natural orifice transluminal endoscopic surgery (NOTES) offers the possibility of surgery without visible scars. To overcome the limitations of NOTES, we developed a novel surgical device called the rigid and flexible outer sheath with multi-piercing surgery (MPS). This sheath can switch between flexible and rigid modes using a novel “Dragon skin” structure and a negative pneumatic shape-locking mechanism. In addition, it has an integrated bending structure that enables it to flex in four directions at the distal end. The insertion part of the prototype is 575 mm long with a 20 mm outer diameter. The rigid and flexible shaft is separated into two parts. The primary and secondary shape-locking shafts are 300 and 200 mm long, respectively. The two parts of shape-locking shaft can be locked independently, including both being locked simultaneously. In addition, the model was equipped with one 7- and 1.7-mm-diameter, and three 3-mm-diameter working channels. When the sheath approaches the target, the surgeon locks the shape and then easily inserts flexible instruments through the path created by the sheath. In this study, we evaluated the stiffness of the prototype. Furthermore, an in vivo partial gastrectomy experiment was also performed. The experimental results regarding the shape property suggest that the rigidity of the shaft was considerably increased using this mechanism. An in vivo partial gastrectomy experiment performed using a swine was successfully performed using the outer sheath for assistance. The outer sheath was inserted through a percutaneous route during the in vivo experiment. The evaluation results show that the advanced outer sheath system has great promise for solving access and stability problems with NOTES.

A laboratory comparison of computer navigation and individualized guides for distal radius osteotomy.

Abstract: This article presents the results of a multiuser, randomized laboratory trial comparing the accuracy and precision of image-based navigation against individualized guides for distal radius osteotomy (DRO). Six surgeons each performed four DROs using image-based navigation and four DROs using individualized guides in a laboratory setting with plastic phantom replicas of radii from patients who had received DRO as treatment for radial deformity. Time required and correction errors of ulnar variance, radial inclination, and volar tilt were measured. There were no statistically significant differences in the average correction errors. There was a statistically significant difference in the standard deviation of ulnar variance error (2.0 mm for navigation vs. 0.6 mm for guides). There was a statistically significant difference in the standard deviation of radial inclination error ( $$6.1^{\circ }$$ for navigation vs. $$1.4^{\circ }$$ for guides). There were statistically significant differences in the times required (705 s for navigation vs. 214 s for guides) and their standard deviations (144 s for navigation vs. 98 s for guides). Compared to navigated DRO, individualized guides were easier to use, faster, and produced more precise correction of ulnar variance and radial inclination. The combination of true three-dimensional planning, ease of use, and accurate and precise corrective guidance makes the individualized guide technique a promising approach for performing corrective osteotomy of the distal radius.

Electrode localization for planning surgical resection of the epileptogenic zone in pediatric epilepsy

Abstract: Purpose In planning for a potentially curative resection of the epileptogenic zone in patients with pediatric epilepsy, invasive monitoring with intracranial EEG is often used to localize the seizure onset zone and eloquent cortex. A precise understanding of the location of subdural strip and grid electrodes on the brain surface, and of depth electrodes in the brain in relationship to eloquent areas is expected to facilitate pre-surgical planning.

Intraoperative visualization of cerebral oxygenation using hyperspectral image data: a two-dimensional mapping method.

Abstract: Superficial temporal artery (STA)—middle cerebral artery (MCA) bypass is an important technique for cerebrovascular reconstruction. Intraoperative hemodynamic imaging is needed to perform cerebrovascular reconstruction safely and effectively. Optical intrinsic signal (OIS) imaging is commonly used for assessing cerebral hemodynamics in experimental studies, because it can provide high-resolution mapping images. However, OIS is not used clinically due to algorithm, instrumentation and spectral resolution limitations. We tested the feasibility of a hyperspectral camera (HSC) for assessment of cortical hemodynamics with spectral imaging of the cerebral cortex in rats and in vivo humans. A hyperspectral camera (HSC) was tested in a rat model of cerebral ischemia (middle cerebral artery occlusion) and during human revascularization surgery (STA–MCA anastomosis). Changes in cortical oxygen saturation were derived from spectral imaging data (400–800 nm) collected by exposing the cortex to Xenon light. Reflected light was sampled using the HSC. The system was then tested intraoperatively during superficial temporal artery to middle cerebral artery anastomosis procedures. Comparison with single-photon emission computed tomography (SPECT) imaging data was done. During middle cerebral artery occlusion in rats, the HSC technique showed a significant decrease in cortical oxygen saturation in the ischemic hemisphere. In clinical cases, the cortical oxygen saturation was increased after STA–MCA anastomosis, which agreed with the SPECT imaging data. Continuous collection of imaging spectroscopic data is feasible and may provide reliable quantification of the hemodynamic responses in the brain. The HSC system may be useful for monitoring intraoperative changes in cortical surface hemodynamics during revascularization procedures in humans.

Vertebra identification using template matching modelmp and K-means clustering

Abstract: Accurate vertebra detection and segmentation are essential steps for automating the diagnosis of spinal disorders. This study is dedicated to vertebra alignment measurement, the first step in a computer-aided diagnosis tool for cervical spine trauma. Automated vertebral segment alignment determination is a challenging task due to low contrast imaging and noise. A software tool for segmenting vertebrae and detecting subluxations has clinical significance. A robust method was developed and tested for cervical vertebra identification and segmentation that extracts parameters used for vertebra alignment measurement. Our contribution involves a novel combination of a template matching method and an unsupervised clustering algorithm. In this method, we build a geometric vertebra mean model. To achieve vertebra detection, manual selection of the region of interest is performed initially on the input image. Subsequent preprocessing is done to enhance image contrast and detect edges. Candidate vertebra localization is then carried out by using a modified generalized Hough transform (GHT). Next, an adapted cost function is used to compute local voted centers and filter boundary data. Thereafter, a $$K$$ -means clustering algorithm is applied to obtain clusters distribution corresponding to the targeted vertebrae. These clusters are combined with the vote parameters to detect vertebra centers. Rigid segmentation is then carried out by using GHT parameters. Finally, cervical spine curves are extracted to measure vertebra alignment. The proposed approach was successfully applied to a set of 66 high-resolution X-ray images. Robust detection was achieved in 97.5 % of the 330 tested cervical vertebrae. An automated vertebral identification method was developed and demonstrated to be robust to noise and occlusion. This work presents a first step toward an automated computer-aided diagnosis system for cervical spine trauma detection.

MRI-compatible positioning device for guiding a focused ultrasound system for transrectal treatment of prostate cancer

Abstract: High-intensity focused ultrasound (HIFU) is a promising treatment method for many common cancers, including prostate cancer. Magnetic resonance image (MRI) guidance of HIFU permits targeting and monitoring of therapy. A prototype MRI-compatible positioning device that navigates a HIFU transducer was designed, fabricated and tested. The positioning device has two PC-controlled and one manually driven stage that allow endorectal access to the prostate. The positioning device was constructed using a 3-D rapid prototype manufacturing device. Software was developed that controls the motion of the positioning device and enables activation of a HIFU transducer. In vitro testing of the system was performed in a 1.5T MRI scanner using ex vivo turkey tissue. Optical encoders were employed to enhance the accuracy of this positioning device. The positioning device was successfully tested for MRI compatibility. The movement error of the positioning device is approximately 20 $$\upmu $$ m. The robot has the ability to accurately move the transducer for creation of discrete and overlapping lesions. An MRI-compatible HIFU positioning system was developed that has the ability to create thermal lesions with MRI guidance for endorectal treatment of prostate cancer.

An automated insertion tool for cochlear implants with integrated force sensing capability

Abstract: Minimally invasive cochlear implantation and residual hearing preservation require both the surgical approach to the cochlea as well as the implant insertion to be performed in an atraumatic fashion. Considering the geometric limitations of this approach, specialized instrumentation is required to insert the electrode while preserving intracochlear membranes carrying the sensory hair cells. An automated insertion tool for cochlear implants, which is capable of sensing insertion forces with a theoretical resolution of $$30\,\upmu \mathrm{N}$$ , is presented. In contrast to previous designs, the custom force sensor is integrated in the insertion mechanism. Moreover, a test bench for insertion studies under constant and reproducible boundary conditions is proposed. It is used to experimentally validate the force sensing insertion tool, which is achieved by comparing the acquired forces to a ground truth measurement. The results of insertion studies on both an acrylic cochlear phantom and temporal bone specimen are given and discussed. Results reveal that friction, occurring between the electrode carrier and the inside of the insertion tool guide tube, is likely to affect the force output of the proposed sensor. An appropriate method to compensate for these disturbances is presented and experimentally validated. Using the proposed approach to friction identification, a mean accuracy of $$(4.0\pm 3.2)\, \hbox {mN}$$ is observed. The force information provided by the proposed, automated insertion tool can be used to detect complications during electrode insertion. However, in order to obtain accurate results, an identification of frictional forces prior to insertion is mandatory. The insertion tool is capable of automatically executing the appropriate trajectories.

Needle deflection estimation: prostate brachytherapy phantom experiments

Abstract: The performance of a fusion-based needle deflection estimation method was experimentally evaluated using prostate brachytherapy phantoms. The accuracy of the needle deflection estimation was determined. The robustness of the approach with variations in needle insertion speed and soft tissue biomechanical properties was investigated. A needle deflection estimation method was developed to determine the amount of needle bending during insertion into deformable tissue by combining a kinematic deflection model with measurements taken from two electromagnetic trackers placed at the tip and the base of the needle. Experimental verification of this method for use in prostate brachytherapy needle insertion procedures was performed. A total of 21 beveled tip, 18 ga, 200 mm needles were manually inserted at various speeds through a template and toward different targets distributed within 3 soft tissue mimicking polyvinyl chloride prostate phantoms of varying stiffness. The tracked positions of both the needle tip and base were recorded, and Kalman filters were applied to fuse the sensory information. The estimation results were validated using ground truth obtained from fluoroscopy images. The manual insertion speed ranged from 8 to 34 mm/s, needle deflection ranged from 5 to 8 mm at an insertion depth of 76 mm, and the elastic modulus of the soft tissue ranged from 50 to 150 kPa. The accuracy and robustness of the estimation method were verified within these ranges. When compared to purely model-based estimation, we observed a reduction in needle tip position estimation error by $$52\pm 17$$ % (mean $$\pm $$ SD) and the cumulative deflection error by $$57\pm 19$$ %. Fusion of electromagnetic sensors demonstrated significant improvement in estimating needle deflection compared to model-based methods. The method has potential clinical applicability in the guidance of needle placement medical interventions, particularly prostate brachytherapy.

Image-guided placement of the Bonebridge™ without surgical navigation equipment.

Abstract: Most of the current Bonebridge surgeries undergo preoperative simulation planning in a computer. However, surgeons usually use the landmarks on the bone surface to determine the location where to implant the device, using the simulation image in the computer only as a reference (conventional method). We developed an image-guided method for precisely replicating simulation surgery upon performing Bonebridge implantation. Based on our previous development of the surface template-assisted marker positioning (STAMP) method for performing image-guided otologic surgery, we fabricated templates that fit only at the designated location on the patient’s temporal bone surface. The Bonebridge STAMP (BB-STAMP) plate shows the exact location where to start drilling. The BB-STAMP was also combined with a perforator-guiding sleeve, so that the location, direction and depth of the cylindrical well could be precisely replicated as simulated. We also created a STAMP plate for confirmation that fits only after sufficient drilling at the correct location is finished. To evaluate the proposed methods, we performed simulation surgery on four cadaveric temporal bones and their 12 replicas (three each for four bones). The time used and the degree of mismatch between the simulated location and the drilled location were compared. A feasibility study was successfully conducted using the proposed BB-STAMP methods and the conventional method. The amount of time required for the procedure did not differ significantly between the surgical methods, although using the BB-STAMP and perforator guide was always quicker. The degree of mismatch between the simulation and resected models had tendency to be smaller when the surgery was guided by the BB-STAMP with or without a perforator guide, although the difference was not statistically significant. The proposed BB-STAMP is a promising method for replicating exactly what is performed during simulation without using a surgical navigation system.

Interventional radiology virtual simulator for liver biopsy

Abstract: Training in Interventional Radiology currently uses the apprenticeship model, where clinical and technical skills of invasive procedures are learnt during practice in patients. This apprenticeship training method is increasingly limited by regulatory restrictions on working hours, concerns over patient risk through trainees’ inexperience and the variable exposure to case mix and emergencies during training. To address this, we have developed a computer-based simulation of visceral needle puncture procedures. A real-time framework has been built that includes: segmentation, physically based modelling, haptics rendering, pseudo-ultrasound generation and the concept of a physical mannequin. It is the result of a close collaboration between different universities, involving computer scientists, clinicians, clinical engineers and occupational psychologists. The technical implementation of the framework is a robust and real-time simulation environment combining a physical platform and an immersive computerized virtual environment. The face, content and construct validation have been previously assessed, showing the reliability and effectiveness of this framework, as well as its potential for teaching visceral needle puncture. A simulator for ultrasound-guided liver biopsy has been developed. It includes functionalities and metrics extracted from cognitive task analysis. This framework can be useful during training, particularly given the known difficulties in gaining significant practice of core skills in patients.

Hybrid method for the detection of pulmonary nodules using positron emission tomography/computed tomography: a preliminary study

Abstract: In this study, an automated scheme for detecting pulmonary nodules using a novel hybrid PET/CT approach is proposed, which is designed to detect pulmonary nodules by combining data from both sets of images. Solitary nodules were detected on CT by a cylindrical filter that we developed previously, and in the PET imaging, high-uptake regions were detected automatically using thresholding based on standardized uptake values along with false-positive reduction by means of the anatomical information obtained from the CT images. Initial candidate nodules were identified by combining the results. False positives among the initial candidates were eliminated by a rule-based classifier and three support vector machines on the basis of the characteristic features obtained from CT and PET images. We validated the proposed method using 100 cases of PET/CT images that were obtained during a cancer-screening program. The detection performance was assessed by free-response receiver operating characteristic (FROC) analysis. The sensitivity was 83.0 % with the number of false positives/case at 5.0, and it was 8 % higher than the sensitivity of independent detection systems using CT or PET images alone. Detection performance indicates that our method may be of practical use for the identification of pulmonary nodules in PET/CT images.

A low-cost tracked C-arm (TC-arm) upgrade system for versatile quantitative intraoperative imaging

Abstract: C-arm fluoroscopy is frequently used in clinical applications as a low-cost and mobile real-time qualitative assessment tool. C-arms, however, are not widely accepted for applications involving quantitative assessments, mainly due to the lack of reliable and low-cost position tracking methods, as well as adequate calibration and registration techniques. The solution suggested in this work is a tracked C-arm (TC-arm) which employs a low-cost sensor tracking module that can be retrofitted to any conventional C-arm for tracking the individual joints of the device. Registration and offline calibration methods were developed that allow accurate tracking of the gantry and determination of the exact intrinsic and extrinsic parameters of the imaging system for any acquired fluoroscopic image. The performance of the system was evaluated in comparison to an Optotrak $$^\mathrm{TM}$$ motion tracking system and by a series of experiments on accurately built ball-bearing phantoms. Accuracies of the system were determined for 2D–3D registration, three-dimensional landmark localization, and for generating panoramic stitched views in simulated intraoperative applications. The system was able to track the center point of the gantry with an accuracy of $$1.5 \pm 1.2$$ mm or better. Accuracies of 2D–3D registrations were $$2.3 \pm 1.1$$ mm and $$0.2 \pm 0.2^{\circ }$$ . Three-dimensional landmark localization had an accuracy of $$3.1 \pm 1.3\%$$ of the length (or $$4.4 \pm 1.9$$ mm) on average, depending on whether the landmarks were located along, above, or across the table. The overall accuracies of the two-dimensional measurements conducted on stitched panoramic images of the femur and lumbar spine were 2.5 $$\pm $$ 2.0 % $$(3.1 \pm 2.5 \hbox { mm})$$ and $$0.3 \pm 0.2^{\circ }$$ , respectively. The TC-arm system has the potential to achieve sophisticated quantitative fluoroscopy assessment capabilities using an existing C-arm imaging system. This technology may be useful to improve the quality of orthopedic surgery and interventional radiology.

Simultaneous extraction of carotid artery intima-media interfaces in ultrasound images: assessment of wall thickness temporal variation during the cardiac cycle

Abstract: Objectives The aim of this work is to present and evaluate a novel segmentation method for localizing the contours of the intima-media complex in the carotid artery wall through longitudinal ultrasound B-mode imaging. The method is used to investigate the association between atherosclerosis risk factors and the cyclic variation of the intima-media thickness during the heart beat. Methods The framework introduced is based on two main features. The first is a simultaneous extraction of both the lumen-intima and the media-adventitia interfaces, using the combination of an original shape-adapted filter bank and a specific dynamic programming scheme. The second is an innovative spatial transformation that eases the extraction of skewed and curved contours, and exploits the result from the previous image as a priori information, when processing the current image. The intima-media thickness is automatically derived from the estimated contours for each time step during the cardiac cycle. Our method was evaluated in vivo on 57 healthy volunteers and 25 patients at high cardiovascular risk. Reference contours were generated for each subject by averaging the tracings performed by three experienced observers. Results Segmentation errors were 29±27\upmum 29 ± 27 μ m for the lumen-intima interface, 42±38\upmum 42 ± 38 μ m for the media-adventitia interface, and 22±16\upmum 22 ± 16 μ m for the intima-media thickness. This uncertainty was similar to inter- and intra-observer variability. Furthermore, the amplitude of the temporal variation in thickness of the intima-media layers during the cardiac cycle was significantly higher in at-risk patients compared to healthy volunteers (79±36 ( 79 ± 36 vs. 64±26\upmum,p=0.032) 64 ± 26 μ m , p = 0.032 ) . Conclusion The method proposed may provide a relevant diagnostic aid for atherosclerosis screening in clinical studies.

Analysis of electrode deformations in deep brain stimulation surgery

Abstract: Deep brain stimulation (DBS) surgery is used to reduce motor symptoms when movement disorders are refractory to medical treatment. Post-operative brain morphology can induce electrode deformations as the brain recovers from an intervention. The inverse brain shift has a direct impact on accuracy of the targeting stage, so analysis of electrode deformations is needed to predict final positions. DBS electrode curvature was evaluated in 76 adults with movement disorders who underwent bilateral stimulation, and the key variables that affect electrode deformations were identified. Non-linear modelling of the electrode axis was performed using post-operative computed tomography (CT) images. A mean curvature index was estimated for each patient electrode. Multivariate analysis was performed using a regression decision tree to create a hierarchy of predictive variables. The identification and classification of key variables that determine electrode curvature were validated with statistical analysis. The principal variables affecting electrode deformations were found to be the date of the post-operative CT scan and the stimulation target location. The main pathology, patient’s gender, and disease duration had a smaller although important impact on brain shift. The principal determinants of electrode location accuracy during DBS procedures were identified and validated. These results may be useful for improved electrode targeting with the help of mathematical models.

Automated liver lesion detection in CT images based on multi-level geometric features

Abstract: Due to the increasing number of liver cancer cases in clinical practice, there is a significant need for efficient tools for computer-assisted liver lesion analysis. A wide range of clinical applications, such as lesion characterization, quantification and follow-up, can be facilitated by automated liver lesion detection. Liver lesions vary significantly in size, shape, density and heterogeneity, which make them difficult to detect automatically. The goal of this work was to develop a method that can detect all types of liver lesions with high sensitivity and low false positive rate within a short run time. The proposed method identifies abnormal regions in liver CT images based on their intensity using a multi-level segmentation approach. The abnormal regions are analyzed from the inside-out using basic geometric features (such as asymmetry, compactness or volume). Using this multi-level shape characterization, the abnormal regions are classified into lesions and other region types (including vessel, liver boundary). The proposed analysis also allows defining the contour of each finding. The method was trained on a set of 55 cases involving 120 lesions and evaluated on a set of 30 images involving 59 (various types of) lesions, which were manually contoured by a physician. The proposed algorithm demonstrated a high detection rate (92 %) at a low (1.7) false positive per case (precision 51 %), when the method was started from a manually contoured liver. The same level of false positive per case (1.6) and precision (51 %) was achieved at a somewhat lower detection rate (85 %), when the volume of interest was defined by a fully automated liver segmentation. The proposed method can efficiently detect liver lesions irrespective of their size, shape, density and heterogeneity within half a minute. According to the evaluation, its accuracy is competitive with the actual state-of-the-art approaches, and the contour of the detected findings is acceptable in most of the cases. Future work shall focus on more precise lesion contouring so that the proposed method can be a solid basis for fully automated liver tumour burden estimation.