Welcome to the second issue of the SPIE Journal of Medical Imaging (JMI). As of this writing, we have received 146 submissions from 28 countries—from major institutions worldwide. I believe we are on track in accomplishing our goals for the journal as the manuscripts submitted thus far span the various fields of medical imaging. For example, this second issue includes excellent papers on image processing; computer-aided diagnosis; image-guided procedures, robotic interventions and modeling; image perception, observer performance, and technology; ultrasonic imaging and tomography; digital pathology; and biomedical applications in molecular, structural, and functional imaging. We also publish our first book review.

We describe a complete pipeline for the detection and accurate automatic segmentation of the optic disc in digital fundus images. This procedure provides separation of vascular information and accurate inpainting of vessel-removed images, symmetry-based optic disc localization, and fitting of incrementally complex contour models at increasing resolutions using information related to inpainted images and vessel masks. Validation experiments, performed on a large dataset of images of healthy and pathological eyes, annotated by experts and partially graded with a quality label, demonstrate the good performances of the proposed approach. The method is able to detect the optic disc and trace its contours better than the other systems presented in the literature and tested on the same data. The average error in the obtained contour masks is reasonably close to the interoperator errors and suitable for practical applications. The optic disc segmentation pipeline is currently integrated in a complete software suite for the semiautomatic quantification of retinal vessel properties from fundus camera images (VAMPIRE).

Multiatlas methods have been successful for brain segmentation, but their application to smaller anatomies remains relatively unexplored. We evaluate seven statistical and voting-based label fusion algorithms (and six additional variants) to segment the optic nerves, eye globes, and chiasm. For nonlocal simultaneous truth and performance level estimation (STAPLE), we evaluate different intensity similarity measures (including mean square difference, locally normalized cross-correlation, and a hybrid approach). Each algorithm is evaluated in terms of the Dice overlap and symmetric surface distance metrics. Finally, we evaluate refinement of label fusion results using a learning-based correction method for consistent bias correction and Markov random field regularization. The multiatlas labeling pipelines were evaluated on a cohort of 35 subjects including both healthy controls and patients. Across all three structures, nonlocal spatial STAPLE (NLSS) with a mixed weighting type provided the most consistent results; for the optic nerve NLSS resulted in a median Dice similarity coefficient of 0.81, mean surface distance of 0.41 mm, and Hausdorff distance 2.18 mm for the optic nerves. Joint label fusion resulted in slightly superior median performance for the optic nerves (0.82, 0.39 mm, and 2.15 mm), but slightly worse on the globes. The fully automated multiatlas labeling approach provides robust segmentations of orbital structures on magnetic resonance imaging even in patients for whom significant atrophy (optic nerve head drusen) or inflammation (multiple sclerosis) is present.

Most medical image registration algorithms suffer from a directionality bias that has been shown to largely impact subsequent analyses. Several approaches have been proposed in the literature to address this bias in the context of nonlinear registration, but little work has been done for global registration. We propose a symmetric approach based on a block-matching technique and least-trimmed square regression. The proposed method is suitable for multimodal registration and is robust to outliers in the input images. The symmetric framework is compared with the original asymmetric block-matching technique and is shown to outperform it in terms of accuracy and robustness. The methodology presented in this article has been made available to the community as part of the NiftyReg open-source package.

We investigated the benefits of incorporating texture features into an existing computer-aided diagnosis (CAD) system for classifying benign and malignant lesions in automated three-dimensional breast ultrasound images. The existing system takes into account 11 different features, describing different lesion properties; however, it does not include texture features. In this work, we expand the system by including texture features based on local binary patterns, gray level co-occurrence matrices, and Gabor filters computed from each lesion to be diagnosed. To deal with the resulting large number of features, we proposed a combination of feature-oriented classifiers combining each group of texture features into a single likelihood, resulting in three additional features used for the final classification. The classification was performed using support vector machine classifiers, and the evaluation was done with 10-fold cross validation on a dataset containing 424 lesions (239 benign and 185 malignant lesions). We compared the classification performance of the CAD system with and without texture features. The area under the receiver operating characteristic curve increased from 0.90 to 0.91 after adding texture features (p<0.001).

Automatic change detection methods for identifying the changes of serial MR images taken at different times are of great interest to radiologists. The majority of existing change detection methods in medical imaging, and those of brain images in particular, include many preprocessing steps and rely mostly on statistical analysis of magnetic resonance imaging (MRI) scans. Although most methods utilize registration software, tissue classification remains a difficult and overwhelming task. Recently, dictionary learning techniques are being used in many areas of image processing, such as image surveillance, face recognition, remote sensing, and medical imaging. We present an improved version of the EigenBlockCD algorithm, named the EigenBlockCD-2. The EigenBlockCD-2 algorithm performs an initial global registration and identifies the changes between serial MR images of the brain. Blocks of pixels from a baseline scan are used to train local dictionaries to detect changes in the follow-up scan. We use PCA to reduce the dimensionality of the local dictionaries and the redundancy of data. Choosing the appropriate distance measure significantly affects the performance of our algorithm. We examine the differences between L₁ and [i]L₂ norms as two possible similarity measures in the improved EigenBlockCD-2 algorithm. We show the advantages of the L₂ norm over the L₁ norm both theoretically and numerically. We also demonstrate the performance of the new EigenBlockCD-2 algorithm for detecting changes of MR images and compare our results with those provided in the recent literature. Experimental results with both simulated and real MRI scans show that our improved EigenBlockCD-2 algorithm outperforms the previous methods. It detects clinical changes while ignoring the changes due to the patient’s position and other acquisition artifacts.

Ultrasound can provide real-time image guidance of radiation therapy, but the probe-induced tissue deformations cause local deviations from the treatment plan. If placed during treatment planning, the probe causes streak artifacts in required computed tomography (CT) images. To overcome these challenges, we propose robot-assisted placement of an ultrasound probe, followed by replacement with a geometrically identical, CT-compatible model probe. In vivo reproducibility was investigated by implanting a canine prostate, liver, and pancreas with three 2.38-mm spherical markers in each organ. The real probe was placed to visualize the markers and subsequently replaced with the model probe. Each probe was automatically removed and returned to the same position or force. Under position control, the median three-dimensional reproducibility of marker positions was 0.6 to 0.7 mm, 0.3 to 0.6 mm, and 1.1 to 1.6 mm in the prostate, liver, and pancreas, respectively. Reproducibility was worse under force control. Probe substitution errors were smallest for the prostate (0.2 to 0.6 mm) and larger for the liver and pancreas (4.1 to 6.3 mm), where force control generally produced larger errors than position control. Results indicate that position control is better than force control for this application, and the robotic approach has potential, particularly for relatively constrained organs and reproducibility errors that are smaller than established treatment margins.

Ovarian cancer is particularly deadly because it is usually diagnosed after it has metastasized. We have previously identified features of ovarian cancer using optical coherence tomography (OCT) and second-harmonic generation (SHG) microscopy (targeting collagen). OCT provides an image of the ovarian microstructure, while SHG provides a high-resolution map of collagen fiber bundle arrangement. Here, we investigated the diagnostic potential of dual-modality OCT and SHG imaging. We conducted a fully crossed, multireader, multicase study using seven human observers. Each observer classified 44 ex vivo mouse ovaries (16 normal and 28 abnormal) as normal or abnormal from OCT, SHG, and simultaneously viewed, coregistered OCT and SHG images and provided a confidence rating on a six-point scale. We determined the average receiver operating characteristic (ROC) curves, area under the ROC curves (AUC), and other quantitative figures of merit. The results show that OCT has diagnostic potential with an average AUC of 0.91±0.06. The average AUC for SHG was less promising at 0.71±0.13. The average AUC for simultaneous OCT and SHG was not significantly different from OCT alone, possibly due to the limited SHG field of view. The high performance of OCT and coregistered OCT and SHG warrants further investigation.

The objective of this investigation was to propose techniques for determining which patients are likely to benefit from quantitative respiratory-gated imaging by correlating respiratory patterns to changes in positron emission tomography (PET) metrics. Twenty-six lung and liver cancer patients underwent PET/computed tomography exams with recorded chest/abdominal displacements. Static and adaptive amplitude-gated [¹⁸F]fluoro-D-glucose (FDG) PET images were generated from list-mode acquisitions. Patients were grouped by respiratory pattern, lesion location, or degree of lesion attachment to anatomical structures. Respiratory pattern metrics were calculated during time intervals corresponding to PET field of views over lesions of interest. FDG PET images were quantified by lesion maximum standardized uptake value (SUV_max). Relative changes in SUV_max between static and gated PET images were tested for association to respiratory pattern metrics. Lower lung lesions and liver lesions had significantly higher changes in SUV_max than upper lung lesions (14 versus 3%, p<0.0001). Correlation was highest (0.42±0.10, r²=0.59, p<0.003) between changes in SUV_max and nonstandard respiratory pattern metrics. Lesion location had a significant impact on changes in PET quantification due to respiratory gating. Respiratory pattern metrics were correlated to changes in SUV_max, though sample size limited statistical power. Validation in larger cohorts may enable selection of patients prior to acquisition who would benefit from respiratory-gated PET imaging.

Quantitative sparse array vascular elastography visualizes the shear modulus distribution within vascular tissues, information that clinicans could use to reduce the number of strokes each year. However, the low transmit power sparse array (SA) imaging could hamper the clinical usefulness of the resulting elastograms. In this study, we evaluated the performance of modulus elastograms recovered from simulated and physical vessel phantoms with varying attenuation coefficients (0.6, 1.5, and 3.5  cm ⁻¹ ) and modulus contrasts (−12.04 , −6.02 , and −2.5  dB ) using SA imaging relative to those obtained with conventional linear array (CLA) and plane-wave (PW) imaging techniques. Plaques were visible in all modulus elastograms, but those produced using SA and PW contained less artifacts. The modulus contrast-to-noise ratio decreased rapidly with increasing modulus contrast and attenuation coefficient, but more quickly when SA imaging was performed than for CLA or PW. The errors incurred varied from 10.9% to 24% (CLA), 1.8% to 12% (SA), and ≈4% (PW). Modulus elastograms produced with SA and PW imagings were not significantly different (p<0.05 ). Despite the low transmit power, SA imaging can produce useful modulus elastograms in superficial organs, such as the carotid artery.

We propose a workflow for color reproduction in whole slide imaging (WSI) scanners, such that the colors in the scanned images match to the actual slide color and the inter-scanner variation is minimum. We describe a new method of preparation and verification of the color phantom slide, consisting of a standard IT8-target transmissive film, which is used in color calibrating and profiling the WSI scanner. We explore several International Color Consortium (ICC) compliant techniques in color calibration/profiling and rendering intents for translating the scanner specific colors to the standard display (sRGB) color space. Based on the quality of the color reproduction in histopathology slides, we propose the matrix-based calibration/profiling and absolute colorimetric rendering approach. The main advantage of the proposed workflow is that it is compliant to the ICC standard, applicable to color management systems in different platforms, and involves no external color measurement devices. We quantify color difference using the CIE-DeltaE2000 metric, where DeltaE values below 1 are considered imperceptible. Our evaluation on 14 phantom slides, manufactured according to the proposed method, shows an average inter-slide color difference below 1 DeltaE. The proposed workflow is implemented and evaluated in 35 WSI scanners developed at Philips, called the Ultra Fast Scanners (UFS). The color accuracy, measured as DeltaE between the scanner reproduced colors and the reference colorimetric values of the phantom patches, is improved on average to 3.5 DeltaE in calibrated scanners from 10 DeltaE in uncalibrated scanners. The average inter-scanner color difference is found to be 1.2 DeltaE. The improvement in color performance upon using the proposed method is apparent with the visual color quality of the tissue scans.

The main author of this book is Dr. Wolfgang Birkfellner (Center for Medical Physics and Biomedical Engineering, Medical University Vienna), and coauthors are Johann Hummel (chapter 1), Michael Figl (chapter 10), and Ziv Yaniv and Özgür Güler (chapter 11). Birkfellner’s main research interest is medical image processing for therapeutic applications such as computer-aided surgery and image-guided radiation therapy. Dr. Birkfellner has coauthored over 50 peer-reviewed research papers, several book chapters, and 2 books.