This JMI Editorial describes upcoming changes to the journal leadership for 2018.

This guest editorial introduces the special section on Visions of Safety: Perspectives on Radiation Exposure and Risk in Medical Imaging

The aim of this paper is to review the available published studies from African countries on patient doses and medical radiation protection and identify strengths, weaknesses, and challenges. Papers on radiation doses to patients published until 2016 pertaining to studies in African countries were reviewed. Radiography, interventional radiology, computed tomography (CT), and mammography modalities were covered. In radiography, the entrance surface air kerma values were below the established diagnostic reference levels (DRLs) provided by the International Atomic Energy Agency, European Commission, and National Council on Radiation Protection and Measurements. Patient and staff doses in interventional procedures were not on the higher side when compared with other published reports from developed countries. The dose length product values in CT in many situations were higher than established DRLs. In mammography, the variations of clinical image quality and dose to standard breast between African countries and other countries were insignificant. In conclusion, like in any continent, not all countries in Africa are active, but some have produced good results. The potential for optimization of radiation protection using simple and inexpensive techniques has been demonstrated. The lack of medical physicists is one of the important challenges.

Radiation backscattered from the patient can contribute substantially to skin dose in fluoroscopically guided interventions (FGIs). The distribution of backscatter is not spatially uniform, and use of a single backscatter factor cannot provide an accurate determination of skin dose. This study evaluates a method to determine the backscatter spatial distribution through convolution of a backscatter-to-primary (BP) point spread function (PSFn). The PSFn is derived for a pencil beam using EGSnrc Monte Carlo software and is convolved with primary distributions using a dose-tracking system. The backscatter distribution calculated using the convolution method is validated with Monte Carlo-derived distributions for three different size “uniform” fields and with XR-QA2 Gafchromic film for nonuniform x-ray fields obtained using region-of-interest (ROI) attenuators and compensation filters, both with homogenous poly-methyl methacrylate and nonhomogenous head phantoms. The BP ratios inside uniform fields were calculated within ±2% of that determined using EGSnrc. For shaped fields, the BP ratio in the unattenuated ROI was calculated within ±3% of that measured with film; in the beam-attenuated periphery, agreement was within ±17%, due to the larger uncertainty of the dose-response curve of the film in the low-dose region. This backscatter PSFn convolution method is much faster than performing full-field Monte Carlo calculations and provides improved accuracy in skin dose distribution determination for FGI procedures.

The purpose of this work was to assess the: (a) impact of regulations on radiation safety and development of radiation safety culture, (b) perceived safety level in the participating facilities, (c) future needs regarding strengthening of regulations, patient dose consideration, and exposure tracking, in different countries around the world. Appropriate questionnaires probing the above-mentioned themes were sent to radiologic professionals working in healthcare facilities. A total of 257 responses from 25 countries were received and analyzed. Average scores for the three different sections/themes of the questionnaire ranged from 64.6% to 74.9%. Higher scores indicated stronger agreement of the survey participants with the theme in question. Statistical comparisons among different professional groups revealed that professionals in developing countries perceived regulations to be significantly more important for improving safety than their counterparts in developed countries did. Radiologic professionals believe that regulation enhances radiation safety and the development of safety culture. However, there is still room for improvement regarding the implementation of regulatory tools. Safety levels were perceived to be satisfactory, and future needs point toward strengthening of regulations regarding patient dose control, patient dose consideration, and patient exposure tracking. The results of this study should be interpreted keeping possible selection bias in view.

The purpose of this study was to extend the concept of weighted CT dose index ( CTDI w ) to the elliptical phantoms. Based on the published body dimension data, eight body aspect ratios were chosen between 1 (perfectly circular) and 1.72 (extremely elliptical). For each aspect ratio, two elliptical cylinders were created digitally to represent adult and pediatric bodies. Their cross-sectional areas were identical to the standard 32- and 16-cm CTDI phantoms. For each phantom, CTDI 100 at center and periphery were simulated for tube voltages between 70 and 140 kVp using a validated Monte Carlo program. The simulations also provided the average dose over the cross-sectional area, CTDI xsec . Values of CTDI xsec and CTDI 100 allowed linear systems of equations to be established, from which central and peripheral weighting coefficients were solved. Regardless of phantom shape, only two weighting coefficients were needed: w 1 for the central CTDI 100 and w 2 for the average of the four peripheral CTDI 100

Much of the emphasis on radiation protection about 2 decades ago accrued from the need for protection of radiation workers and collective doses to populations from medical exposures. With the realization that individual patient doses were rising and becoming an issue, the author had propagated the concept of a smart card for radiation exposure history of individual patients. During the last 7 years, much has happened wherein radiation exposure and the dose history of individual patients has become a reality in many countries. In addition to dealing with overarching questions, such as “Why track, what to track, and how to track?,” this review elaborates on a number of points such as attitudes toward tracking, review of practices in large parts of the world, description of various elements for exposure and dose tracking, how to use the information available from tracking, achievements and stumbling blocks in implementation to date, templates for implementation of tracking at different levels of health care, the role of picture archiving and communication systems and eHealth, the role of tracking in justification and optimization of protection, comments on cumulative dose, how referrers can use this information, current provisions in international standards, and future actions.

The purpose of this study was to investigate relationships between patient attributes and organ dose for a population of computational phantoms for 20 tomosynthesis and radiography protocols. Organ dose was estimated from 54 adult computational phantoms (age: 18 to 78 years, weight 52 to 117 kg) using a validated Monte-Carlo simulation (PENELOPE) of a system capable of performing tomosynthesis and radiography. The geometry and field of view for each exam were modeled to match clinical protocols. For each protocol, the energy deposited in each organ was estimated by the simulations, converted to dose units, and then normalized by exposure in air. Dose to radiosensitive organs was studied as a function of average patient thickness in the region of interest and as a function of body mass index. For tomosynthesis, organ doses were also studied as a function of x-ray tube position. This work developed comprehensive information for organ dose dependencies across a range of tomosynthesis and radiography protocols. The results showed a protocol-dependent exponential decrease with an increasing patient size. There was a variability in organ dose across the patient population, which should be incorporated in the metrology of organ dose. The results can be used to prospectively and retrospectively estimate organ dose for tomosynthesis and radiography.

This study aimed to estimate the organ dose reduction potential for organ-dose-based tube current modulated (ODM) thoracic computed tomography (CT) with a wide dose reduction arc. Twenty-one computational anthropomorphic phantoms (XCAT) were used to create a virtual patient population with clinical anatomic variations. The phantoms were created based on patient images with normal anatomy (age range: 27 to 66 years, weight range: 52.0 to 105.8 kg). For each phantom, two breast tissue compositions were simulated: 50/50 and 20/80 (glandular-to-adipose ratio). A validated Monte Carlo program (PENELOPE, Universitat de Barcelona, Spain) was used to estimate the organ dose for standard tube current modulation (TCM) (SmartmA, GE Healthcare) and ODM (GE Healthcare) for a commercial CT scanner (Revolution, GE Healthcare) using a typical clinical thoracic CT protocol. Both organ dose and CTDI_vol-to-organ dose conversion coefficients (h factors) were compared between TCM and ODM. ODM significantly reduced all radiosensitive organ doses (p<0.01). The breast dose was reduced by 30±2%. For h factors, organs in the anterior region (e.g., thyroid and stomach) exhibited substantial decreases, and the medial, distributed, and posterior region saw either an increase of less than 5% or no significant change. ODM significantly reduced organ doses especially for radiosensitive superficial anterior organs such as the breasts.

The purpose of this study was to formulate a systematic, evidence-based method to relate quantitative diagnostic performance to radiation dose, enabling a multidimensional system to optimize computed tomography imaging across pediatric populations. Based on two prior foundational studies, radiation dose was assessed in terms of organ doses, effective dose (E), and risk index for 30 patients within nine color-coded pediatric age-size groups as a function of imaging parameters. The cases, supplemented with added noise and simulated lesions, were assessed in terms of nodule detection accuracy in an observer receiving operating characteristic study. The resulting continuous accuracy–dose relationships were used to optimize individual scan parameters. Before optimization, the nine protocols had a similar E of 2.2±0.2  mSv with accuracy decreasing from 0.89 for the youngest patients to 0.67 for the oldest. After optimization, a consistent target accuracy of 0.83 was established for all patient categories with E ranging from 1 to 10 mSv. Alternatively, isogradient operating points targeted a consistent ratio of accuracy–per-unit-dose across the patient categories. The developed model can be used to optimize individual scan parameters and provide for consistent diagnostic performance across the broad range of body sizes in children.

In some medical-imaging procedures using cone-beam CT (CBCT) and fluoroscopy, only the center of the field of view (FOV) may be needed to be visualized with optimal image quality. To reduce the dose to the patient while maintaining visualization of the entire FOV, a Cu attenuator with a circular aperture for the region of interest (ROI) is used. The potential organ and effective dose reductions of ROI imaging when applied to CBCT and interventional fluoroscopic procedures were determined using EGSnrc Monte Carlo code. The Monte Carlo model was first validated by comparing the surface dose distribution in a solid-water block phantom with measurement by Gafchromic film. The dependence of dose reduction on the ROI attenuator thickness, the opening size of the ROI, the axial beam position, and the location of the different organs for both neuro and thoracic imaging was evaluated. The results showed a reduction in most organ doses of 45% to 70% and in effective dose of 46% to 66% compared to the dose in a CBCT scan and in an interventional procedure without the ROI attenuator. This work provides evidence of a substantial reduction of organ and effective doses when using an ROI attenuator during CBCT and fluoroscopic procedures.

Cardiac imagers worldwide are bracing for increased utilization of cardiac computed tomography (CT) in clinical practice. This expanding opportunity brings along a responsibility to produce diagnostic quality images with optimized radiation dose. The following review aims to address the dose reduction strategies in cardiac CT in light of recent scientific evidence and technical developments.

This study investigates the dosimetry methodology proposed by the American Association of Physicists in Medicine (AAPM) task group 111 and compares with the computed tomography dose index (CTDI) method and the SEDENTEXCT DI method on one clinical multislice CT and two dental cone beam CT (CBCT) scanners using adult, adolescent, and child head phantoms. Following the AAPM method, the normalized (100 mAs) equilibrium doses (Deq) for Toshiba Aquilion One MSCT computed using dose measurements from the central hole of the phantom (Deq,c), the peripheral hole of the phantom, (Deq,p), and by the CTDIw equation (Deq,w) are in the range from 20 to 25 mGy. For i-CAT Next Generation dental CBCT, the normalized Deq,c, Deq,p, Deq,w, and Deq′s by the two SEDENTEXCT DI methods are in the range from 12 to 15 mGy. Fitting the AAPM equation is not possible for the limited scan lengths available on the CS 9300 dental CBCT. This study offers a simple CTDI-like measurement that can approximate the AAPM Deq in clinical CBCT scanners capable of providing four or more scan lengths.

The use of iterative reconstruction (IR) algorithms in CT generally decreases image noise and enables dose reduction. However, the amount of dose reduction possible using IR without sacrificing diagnostic performance is difficult to assess with conventional image quality metrics. Through this investigation, achievable dose reduction using a commercially available IR algorithm without loss of low contrast spatial resolution was determined with a channelized Hotelling observer (CHO) model and used to optimize a clinical abdomen/pelvis exam protocol. A phantom containing 21 low contrast disks—three different contrast levels and seven different diameters—was imaged at different dose levels. Images were created with filtered backprojection (FBP) and IR. The CHO was tasked with detecting the low contrast disks. CHO performance indicated dose could be reduced by 22% to 25% without compromising low contrast detectability (as compared to full-dose FBP images) whereas 50% or more dose reduction significantly reduced detection performance. Importantly, default settings for the scanner and protocol investigated reduced dose by upward of 75%. Subsequently, CHO-based protocol changes to the default protocol yielded images of higher quality and doses more consistent with values from a larger, dose-optimized scanner fleet. CHO assessment provided objective data to successfully optimize a clinical CT acquisition protocol.

Optimization is one of the key concepts of radiation protection in medical imaging. In practice, it involves compromising between the image quality and dose to the patient; the dose should not be higher than necessary to achieve an image quality (or diagnostic information) needed for the clinical task. Monitoring patient dose is a key requirement toward optimization. The concept of diagnostic reference level (DRL) was introduced by the International Commission on Radiological Protection as a practical tool for optimization. Unfortunately, this concept has not been applied consistently worldwide. To review the current strengths and weaknesses worldwide and to promote improvements, the International Atomic Energy Agency organized a Technical Meeting on patient dose monitoring and the use of DRLs on May 2016. This paper reports a summary of the findings and conclusions from the meeting. The strengths and weaknesses were generally different in less-developed countries compared with developed countries. Possible improvements were suggested in six areas: human resources and responsibilities, training, safety and quality culture, regulations, funding, and tools and methods. An overall conclusion was that radiation protection requires a patient-centric approach and a transfer from purely reactive to increasingly proactive optimization, whereby the best outcome is expected from good teamwork.

We propose an efficient optical tomography with discretized path integral. We first introduce the primal-dual approach to solve the inverse problem formulated as a constraint optimization problem. Next, we develop efficient formulations for computing Jacobian and Hessian of the cost function of the constraint nonlinear optimization problem. Numerical experiments show that the proposed formulation is faster (23.14±1.32  s) than the previous work with the log-barrier interior point method (91.17±1.48  s) for the Shepp–Logan phantom with a grid size of 24×24, while keeping the quality of the estimation results (root-mean-square error increasing by up to 12%).

Stand-alone cone beam computed tomography (CT) and single-photon emission computed tomography (SPECT) systems capable of complex acquisition trajectories have previously been developed for breast imaging. Fully three-dimensional (3-D) motions of SPECT systems provide views into the chest wall and throughout the entire volume. The polar tilting capability of the CBCT system has shown improvement in sampling close to the chest wall, while eliminating cone beam artifacts. Here, a single hybrid SPECT–CT system, with each individual modality capable of independently traversing complex trajectories around a common pendant breast volume, was developed. We present the practical implementation of this design and preliminary results of the CT system. The fully 3-D SPECT was nested inside the suspended CT gantry and oriented perpendicular to the CT source–detector pair. Both subsystems were positioned on a rotation stage, with the combined polar and azimuthal motions enabling spherical trajectories. Six trajectories were used for initial evaluation of the tilt capable CT system. The developed system can achieve polar tilt angles with a <0.02-deg positioning error and no hysteresis. Initial imaging results demonstrate that additional off-axis projection views of various geometric resolution phantoms facilitate more complete sampling, more consistent attenuation value recovery, and markedly improved reconstructions. This system could have various applications in diagnostic or therapeutic breast imaging.

Interactive segmentation is a promising approach to solving the pervasive shortage of reference annotations for automated medical image processing. We focus on the challenging task of glioblastoma segmentation in magnetic resonance imaging using a random forest pixel classifier trained iteratively on scribble annotations. Our experiments use data from the MICCAI Multimodal Brain Tumor Segmentation Challenge 2013 and simulate expert interactions using different approaches: corrective annotations, class-balanced corrections, annotations where classifier uncertainty is high, and corrections where classifier uncertainty is high/low. We find that it is better to correct the classifier than to provide annotations where the classifier is uncertain, resulting in significantly better Dice scores in the edema (0.662 to 0.686) and necrosis (0.550 to 0.676) regions after 20 interactions. It is also advantageous to balance inputs among classes, with significantly better Dice in the necrotic (0.501 to 0.676) and nonenhancing (0.151 to 0.235) regions compared to fully random corrections. Corrective annotations in regions of high classifier uncertainty provide no additional benefit, low uncertainty corrections perform worst. Preliminary experiments with real users indicate that those with intermediate proficiency make a considerable number of annotation errors. The performance of corrective approaches suffers most strongly from this, leading to a less profound difference to uncertainty-based annotations.

Multiatlas segmentation offers an exceedingly convenient process by which image segmentation tools can be created from a series of labeled atlases (i.e., raters). However, creation of the atlases is exceedingly time consuming and prone to shifts in clinical/research demands as anatomical definitions are refined, combined, or subdivided. Hence, a process by which atlases from distinct, but complementary, anatomical “protocols” could be combined would allow for greater innovation in structural analysis and efficiency of data (re)use. Recent innovation in protocol fusion has shown that propagation of information across distinct protocols is feasible. However, how to effectively include this information in simultaneous truth and performance level estimation (STAPLE) has been elusive. We present a generalization of the STAPLE framework to account for multiprotocol rater performance (i.e., accuracy of registered atlases). This approach, multiset STAPLE (MS-STAPLE), provides a statistical framework for combining label information from atlases that have been labeled with distinct protocols (i.e., whole brain versus subcortical) and is compatible with the current local, nonlocal, probabilistic, log-odds, and hierarchical innovations in STAPLE theory. Using the MS-STAPLE approach, information from a broad range of datasets can be combined so that each available dataset contributes in a spatially dependent manner to local labels. We evaluate the model in simulations and in the context of an experiment where an existing set of whole-brain labels (14 structures) is refined to include parcellation of subcortical structures (26 structures). In the empirical results, we see significant improvement in the Dice similarity coefficient when comparing MS-STAPLE to STAPLE and nonlocal MS-STAPLE to nonlocal STAPLE.

Diabetic retinopathy (DR) is a consequence of diabetes and is the leading cause of blindness among 18- to 65-year-old adults. Regular screening is critical to early detection and treatment of DR. Computer-aided diagnosis has the potential of improving the practice in DR screening or diagnosis. An automated and unsupervised approach for retrieving clinically relevant images from a set of previously diagnosed fundus camera images for improving the efficiency of screening and diagnosis of DR is presented. Considering that DR lesions are often localized, we propose a multiclass multiple-instance framework for the retrieval task. Considering the special visual properties of DR images, we develop a feature space of a modified color correlogram appended with statistics of steerable Gaussian filter responses selected by fast radial symmetric transform points. Experiments with real DR images collected from five different datasets demonstrate that the proposed approach is able to outperform existing methods.

Low-field magnetic resonance imaging (MRI) has recently been integrated with radiation therapy systems to provide image guidance for daily cancer radiation treatments. The main benefit of the low-field strength is minimal electron return effects. The main disadvantage of low-field strength is increased image noise compared to diagnostic MRIs conducted at 1.5 T or higher. The increased image noise affects both the discernibility of soft tissues and the accuracy of further image processing tasks for both clinical and research applications, such as tumor tracking, feature analysis, image segmentation, and image registration. An innovative method, adaptive anatomical preservation optimal denoising (AAPOD), was developed for optimal image denoising, i.e., to maximally reduce noise while preserving the tissue boundaries. AAPOD employs a series of adaptive nonlocal mean (ANLM) denoising trials with increasing denoising filter strength (i.e., the block similarity filtering parameter in the ANLM algorithm), and then detects the tissue boundary losses on the differences of sequentially denoised images using a zero-crossing edge detection method. The optimal denoising filter strength per voxel is determined by identifying the denoising filter strength value at which boundary losses start to appear around the voxel. The final denoising result is generated by applying the ANLM denoising method with the optimal per-voxel denoising filter strengths. The experimental results demonstrated that AAPOD was capable of reducing noise adaptively and optimally while avoiding tissue boundary losses. AAPOD is useful for improving the quality of MRIs with low-contrast-to-noise ratios and could be applied to other medical imaging modalities, e.g., computed tomography.

Grating-based Talbot–Lau x-ray interferometry is a popular method for measuring absorption, phase shift, and small-angle scattering. The standard acquisition method for this modality is phase stepping, where the Talbot pattern is reconstructed from multiple images acquired at different grating positions. We review the implicit assumptions in phase-stepping reconstruction, and find that the assumptions of perfectly known grating positions and homoscedastic noise variance are violated in some scenarios. Additionally, we investigate a recently reported estimation bias in the visibility and dark-field signal. To adapt the phase-stepping reconstruction to these findings, we propose three improvements to the reconstruction. These improvements are (a) to use prior knowledge to compute more accurate grating positions to reduce moiré artifacts, (b) to utilize noise variance information to reduce dark-field and phase noise in high-visibility acquisitions, and (c) to perform correction of an estimation bias in the interferometer visibility, leading to more quantitative dark-field imaging in acquisitions with a low signal-to-noise ratio. We demonstrate the benefit of our methods on simulated data, as well as on images acquired with a Talbot–Lau interferometer.

Extraction of blood vessels in retinal images is an important step for computer-aided diagnosis of ophthalmic pathologies. We propose an approach for blood vessel tracking and diameter estimation. We hypothesize that the curvature and the diameter of blood vessels are Gaussian processes (GPs). Local Radon transform, which is robust against noise, is subsequently used to compute the features and train the GPs. By learning the kernelized covariance matrix from training data, vessel direction and its diameter are estimated. In order to detect bifurcations, multiple GPs are used and the difference between their corresponding predicted directions is quantified. The combination of Radon features and GP results in a good performance in the presence of noise. The proposed method successfully deals with typically difficult cases such as bifurcations and central arterial reflex, and also tracks thin vessels with high accuracy. Experiments are conducted on the publicly available DRIVE, STARE, CHASEDB1, and high-resolution fundus databases evaluating sensitivity, specificity, and Matthew’s correlation coefficient (MCC). Experimental results on these datasets show that the proposed method reaches an average sensitivity of 75.67%, specificity of 97.46%, and MCC of 72.18% which is comparable to the state-of-the-art.

Assessment of three-dimensional (3-D) morphology and volume of breast masses is important for cancer diagnosis, staging, and treatment but cannot be derived from conventional mammography. Digital breast tomosynthesis (DBT) provides data from which 3-D mass segmentation could be obtained. Our method combined Gaussian mixture models based on intensity and a texture measure indicative of in-focus structure, gray-level variance. Thresholding these voxel probabilities, weighted by distance to the estimated mass center, gave the final 3-D segmentation. Evaluation used 40 masses annotated twice by a consultant radiologist on in-focus slices in two diagnostic views. Human intraobserver variability was assessed as the overlap between repeated annotations (median 77% and range 25% to 91%). Comparing the segmented mass outline with probability-weighted ground truth from these annotations, median agreement was 68%, and range was 7% to 88%. Annotated and segmented diameters correlated well with histological mass size (both Spearman’s rank correlations ρ=0.69). The volumetric segmentation demonstrated better agreement with tumor volumes estimated from pathology than volume derived from radiological annotations (95% limits of agreement −16 to 11 ml and −23 to 41 ml, respectively). We conclude that it is feasible to assess 3-D mass morphology and volume from DBT, and the method has the potential to aid breast cancer management.

We present a fully automatic method for segmenting orbital structures (globes, optic nerves, and extraocular muscles) in CT images. Prior anatomical knowledge, such as shape, intensity, and spatial relationships of organs and landmarks, were utilized to define a volume of interest (VOI) that contains the desired structures. Then, VOI was used for fast localization and successful segmentation of each structure using predefined rules. Testing our method with 30 publicly available datasets, the average Dice similarity coefficient for right and left sides of [0.81, 0.79] eye globes, [0.72, 0.79] optic nerves, and [0.73, 0.76] extraocular muscles were achieved. The proposed method is accurate, efficient, does not require training data, and its intuitive pipeline allows the user to modify or extend to other structures.

Early stage diagnosis of laryngeal squamous cell carcinoma (SCC) is of primary importance for lowering patient mortality or after treatment morbidity. Despite the challenges in diagnosis reported in the clinical literature, few efforts have been invested in computer-assisted diagnosis. The objective of this paper is to investigate the use of texture-based machine-learning algorithms for early stage cancerous laryngeal tissue classification. To estimate the classification reliability, a measure of confidence is also exploited. From the endoscopic videos of 33 patients affected by SCC, a well-balanced dataset of 1320 patches, relative to four laryngeal tissue classes, was extracted. With the best performing feature, the achieved median classification recall was 93% [interquartile range (IQR)=6%]. When excluding low-confidence patches, the achieved median recall was increased to 98% (IQR=5%), proving the high reliability of the proposed approach. This research represents an important advancement in the state-of-the-art computer-assisted laryngeal diagnosis, and the results are a promising step toward a helpful endoscope-integrated processing system to support early stage diagnosis.

In ultrasound (US)-guided medical procedures, accurate tracking of interventional tools is crucial to patient safety and clinical outcome. This requires a calibration procedure to recover the relationship between the US image and the tracking coordinate system. In literature, calibration has been performed on passive phantoms, which depend on image quality and parameters, such as frequency, depth, and beam-thickness as well as in-plane assumptions. In this work, we introduce an active phantom for US calibration. This phantom actively detects and responds to the US beams transmitted from the imaging probe. This active echo (AE) approach allows identification of the US image midplane independent of image quality. Both target localization and segmentation can be done automatically, minimizing user dependency. The AE phantom is compared with a crosswire phantom in a robotic US setup. An out-of-plane estimation US calibration method is also demonstrated through simulation and experiments to compensate for remaining elevational uncertainty. The results indicate that the AE calibration phantom can have more consistent results across experiments with varying image configurations. Automatic segmentation is also shown to have similar performance to manual segmentation.

Intraoperative soft tissue deformation, referred to as brain shift, compromises the application of current image-guided surgery navigation systems in neurosurgery. A computational model driven by sparse data has been proposed as a cost-effective method to compensate for cortical surface and volumetric displacements. We present a mock environment developed to acquire stereoimages from a tracked operating microscope and to reconstruct three-dimensional point clouds from these images. A reconstruction error of 1 mm is estimated by using a phantom with a known geometry and independently measured deformation extent. The microscope is tracked via an attached tracking rigid body that facilitates the recording of the position of the microscope via a commercial optical tracking system as it moves during the procedure. Point clouds, reconstructed under different microscope positions, are registered into the same space to compute the feature displacements. Using our mock craniotomy device, realistic cortical deformations are generated. When comparing our tracked microscope stereo-pair measure of mock vessel displacements to that of the measurement determined by the independent optically tracked stylus marking, the displacement error was ∼2  mm on average. These results demonstrate the practicality of using tracked stereoscopic microscope as an alternative to laser range scanners to collect sufficient intraoperative information for brain shift correction.

Brain shift during tumor resection compromises the spatial validity of registered preoperative imaging data that is critical to image-guided procedures. One current clinical solution to mitigate the effects is to reimage using intraoperative magnetic resonance (iMR) imaging. Although iMR has demonstrated benefits in accounting for preoperative-to-intraoperative tissue changes, its cost and encumbrance have limited its widespread adoption. While iMR will likely continue to be employed for challenging cases, a cost-effective model-based brain shift compensation strategy is desirable as a complementary technology for standard resections. We performed a retrospective study of n=9 tumor resection cases, comparing iMR measurements with intraoperative brain shift compensation predicted by our model-based strategy, driven by sparse intraoperative cortical surface data. For quantitative assessment, homologous subsurface targets near the tumors were selected on preoperative MR and iMR images. Once rigidly registered, intraoperative shift measurements were determined and subsequently compared to model-predicted counterparts as estimated by the brain shift correction framework. When considering moderate and high shift (<3  mm, n=13±6 measurements per case), the alignment error due to brain shift reduced from 5.7±2.6 to 2.3±1.1  mm, representing ∼59% correction. These first steps toward validation are promising for model-based strategies.

The structural similarity index (SSIM) family is a set of metrics that has demonstrated good agreement with human observers in tasks using reference images. These metrics analyze the viewing distance, edge information between the reference and the test images, changed and preserved edges, textures, and structural similarity of the images. Eight metrics based on that family are proposed. This new set of metrics, together with another eight well-known SSIM family metrics, was tested to predict human performance in some specific tasks closely related to the evaluation of radiological medical images. We used a database of radiological images, comprising different acquisition techniques (MRI and plain films). This database was distorted with different types of distortions (Gaussian blur, noise, etc.) and different levels of degradation. These images were analyzed by a board of radiologists with a double-stimulus methodology, and their results were compared with those obtained from the 16 metrics analyzed and proposed in this research. Our experimental results showed that the readings of human observers were sensitive to the changes and preservation of the edge information between the reference and test images, changes and preservation in the texture, structural component of the images, and simulation of multiple viewing distances. These results showed that several metrics that apply this multifactorial approach (4-G-SSIM, 4-MS-G-SSIM, 4-G-r*, and 4-MS-G-r*) can be used as good surrogates of a radiologist to analyze the medical quality of an image in an environment with a reference image.

Visual x-ray image processing (XRIP) represents a fundamental component of catheter-based cardiovascular interventions (CBCVIs). To date, no data are available to define XRIP in this setting. To characterize CBCVI XRIP, we developed a computer-based method allowing continuous temporal–spatial analysis of data recorded by a head-mounted eye-tracking device. Quantitative analysis of gaze duration of an expert operator (EO) revealed that the average time in minutes spent viewing the images on the display screen was 39.5%±13.6% and 41.5%±18.3% of the total recorded time in coronary angiography (CA) and in CA followed by CBCVI, respectively. Qualitative analysis of gaze data of the EO revealed consistent focus on the center point of the screen. Only if suspicious findings were detected did gaze move toward the target. In contrast, a novice operator (NO) observing a subset of cases viewed coronary artery segments separately and sequentially. The developed methodology allows continuous registration and analysis of gaze data for analysis of XRIP strategies of EOs in live-cases scenarios and may assist in the transfer of experts’ reading skills to novices.

We use high-resolution μCT data in multiple experiments to estimate the sources of error during coregistration of images acquired on separate preclinical instruments. In combination with experiments with phantoms, we completed in vivo imaging on mice, aimed at identifying the possible sources of registration errors, caused either by transport of the animal, movement of the animal itself, or methods of coregistration. The same imaging cell was used as a holder for phantoms and animals. For all procedures, rigid coregistration was carried out using a common landmark coregistration system, placed inside the imaging cell. We used the fiducial registration error and the target registration error to analyze the coregistration accuracy. We found that moving an imaging cell between two preclinical devices during a multimodal procedure gives an error of about 200  μm at most. Therefore, it could not be considered a source of coregistration errors. Errors linked to spontaneous movements of the animal increased with time, to nearly 1 mm at most, excepted for body parts that were properly restrained. This work highlights the importance of animal intrinsic movements during a multiacquisition procedure and demonstrates a simple method to identify and quantify the sources of error during coregistration.

Our goal was to ascertain how fatigue affects performance in reading computed tomography (CT) examinations of patients with multiple injuries. CT images with multiple fractures from a previous study of satisfaction of search (SOS) were read by radiologists after a day of clinical work. Performance in this study with fatigued readers was compared to a previous study in which readers were not fatigued. Detection accuracy for obvious injuries was not affected by fatigue, but accuracy for subtle fractures was reduced (P=0.016). An SOS effect on decision thresholds was evident mirroring recent studies. Without fatigue, readers spent more time interpreting and reporting findings as the number of the injuries increased. When fatigued, readers did not increase reading time as fracture number increased. Without fractures, reading time for not-fatigued and fatigued readers was the same (P=0.493) but was significant (P=0.016) with an added subtle fracture. The difference increased with a major injury (P=0.003) and increased further with both a major injury and subtle fracture (P=0.0007). Fatigue and multiple abnormalities have independent effects on detection performance but do interact in determining search time.

To advance magnetic resonance imaging (MRI) technologies further for in vivo tissue characterization with histopathologic validation, we investigated the feasibility of ex vivo tissue imaging of a surgically removed human brain tumor as a comprehensive approach for radiology–pathology correlation in histoanatomically identical fashion in a rare case of pigmented ganglioglioma with complex paramagnetic properties. Pieces of surgically removed ganglioglioma, containing melanin and hemosiderin pigments, were imaged with a small bore 7-T MRI scanner to obtain T1-, T2-, and T2*-weighted image and diffusion tensor imaging (DTI). Corresponding histopathological slides were prepared for routine hematoxylin and eosin stain and special stains for melanin and iron/hemosiderin to correlate with MRI signal characteristics. Furthermore, mean diffusivity (MD) maps were generated from DTI data and correlated with cellularity using image analysis. While the presence of melanin was difficult to interpret in in vivo MRI with certainty due to concomitant hemosiderin pigments and calcium depositions, ex vivo tissue imaging clearly demonstrated pieces of tissue exhibiting the characteristic MR signal pattern for melanin with pathologic confirmation in a histoanatomically identical location. There was also concordant correlation between MD and cellularity. Although it is still in an initial phase of development, ex vivo tissue imaging is a promising approach, which offers radiology–pathology correlation in a straightforward and comprehensive manner.

A registration method to fuse two-dimensional (2-D) echocardiography images with cardiac computed tomography (CT) volume is presented. The method consists of two major procedures: temporal and spatial registrations. In temporal registration, the echocardiography frames at similar cardiac phases as the CT volume were interpolated based on electrocardiogram signal information, and the noise of the echocardiography image was reduced using the speckle reducing anisotropic diffusion technique. For spatial registration, an intensity-based normalized mutual information method was applied with a pattern search optimization algorithm to produce an interpolated cardiac CT image. The proposed registration framework does not require optical tracking information. Dice coefficient and Hausdorff distance for the left atrium assessments were <inline-formula< <mml:math display="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"< <mml:mrow< <mml:mn<0.87</mml:mn< <mml:mo<±</mml:mo< <mml:mn<0.04</mml:mn< </mml:mrow< </mml:math< </inline-formula< and <inline-formula< <mml:math display="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"< <mml:mrow< <mml:mn<1.23</mml:mn< <mml:mo<±</mml:mo< <mml:mn<0.32</mml:mn< <mml:mtext<  </mml:mtext< <mml:mi<mm</mml:mi< </mml:mrow< </mml:math< </inline-formula<, respectively; for left ventricle, they were <inline-formula< <mml:math display="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"< <mml:mrow< <mml:mn<0.82</mml:mn< <mml:mo<±</mml:mo< <mml:mn<0.07</mml:mn< </mml:mrow< </mml:math< </inline-formula< and <inline-formula< <mml:math display="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"< <mml:mrow< <mml:mn<1.14</mml:mn< <mml:mo<±</mml:mo< <mml:mn<0.18</mml:mn< <mml:mtext<  </mml:mtext< <mml:mi<mm</mml:mi< </mml:mrow< </mml:math< </inline-formula<, respectively. There was no significant difference in the mitral valve annulus diameter measurement between the manually and automatically registered CT images. The transformation parameters showed small deviations (<inline-formula< <mml:math display="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"< <mml:mrow< <mml:mo<≤</mml:mo< <mml:mn<1.14</mml:mn< <mml:mtext<  </mml:mtext< <mml:mi<mm</mml:mi< </mml:mrow< </mml:math< </inline-formula< deviation in translation and <inline-formula< <mml:math display="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"< <mml:mrow< <mml:mo form="prefix"<<</mml:mo< <mml:mn<2</mml:mn< <mml:mtext<  </mml:mtext< <mml:mi<deg</mml:mi< </mml:mrow< </mml:math< </inline-formula< for rotation) between manual and automatic registrations. The proposed method aids the physician in diagnosing mitral valve disease as well as provides surgical guidance during the treatment procedure.