Peyton H. Bland

Demonstration of accuracy and clinical versatility of mutual information for automatic multimodality image fusion using affine and thin-plate spline warped geometric deformations

This paper applies and evaluates an automatic mutual information-based registration algorithm across a broad spectrum of multimodal volume data sets. The algorithm requires little or no pre-processing, minimal user input and easily implements either affine, i.e. linear or thin-plate spline (TPS) warped registrations. We have evaluated the algorithm in phantom studies as well as in selected cases where few other algorithms could perform as well, if at all, to demonstrate the value of this new method. Pairs of multimodal gray-scale volume data sets were registered by iteratively changing registration parameters to maximize mutual information. Quantitative registration errors were assessed in registrations of a thorax phantom using PET/CT and in the National Library of Medicines Visible Male using MRI T2-/T1-weighted acquisitions. Registrations of diverse clinical data sets were demonstrated including rotate-translate mapping of PET/MRI brain scans with significant missing data, full affine mapping of thoracic PET/CT and rotate-translate mapping of abdominal SPECT/CT. A five-point thin-plate spline (TPS) warped registration of thoracic PET/CT is also demonstrated. The registration algorithm converged in times ranging between 3.5 and 31 min for affine clinical registrations and 57 min for TPS warping. Mean error vector lengths for rotate-translate registrations were measured to be subvoxel in phantoms. More importantly the rotate-translate algorithm performs well even with missing data. The demonstrated clinical fusions are qualitatively excellent at all levels. We conclude that such automatic, rapid, robust algorithms significantly increase the likelihood that multimodality registrations will be routinely used to aid clinical diagnoses and post-therapeutic assessment in the near future.

Retrospective correction of intensity inhomogeneities in MRI

Medical imaging data sets are often corrupted by multiplicative inhomogeneities, often referred to as nonuniformities or intensity variations, that hamper the use of quantitative analyses. The authors describe an automatic technique that not only improves the worst situations, such as those encountered with magnetic resonance imaging (MRI) surface coils, but also corrects typical inhomogeneities encountered in routine volume data sets, such as MRI head scans, without generating additional artifact. Because the technique uses only the patient data set, the technique can be applied retrospectively to all data sets, and corrects both patient independent effects, such as rf coil design, and patient dependent effects, such as attenuation of overlying tissue experienced both in high field MRI and X-ray computed tomography (CT). The authors show results for several MRI imaging situations including thorax, head, and breast. Following such corrections, region of interest analyses, volume histograms, and thresholding techniques are more meaningful. The value of such correction algorithms may increase dramatically with increased use of high field strength magnets and associated patient-dependent rf attenuation in overlying tissues.

Motion Correction in fMRI via Registration of Individual Slices Into an Anatomical Volume

An automated retrospective image registration based on mutual information is adapted to a multislice functional magnetic resonance imaging (fMRI) acquisition protocol to provide accurate motion correction. Motion correction is performed by mapping each slice to an anatomic volume data set acquired in the same fMRI session to accommodate inter‐slice head motion. Accuracy of the registration parameters was assessed by registration of simulated MR data of the known truth. The widely used rigid body volume registration approach based on stacked slices from the time series data may hinder statistical accuracy by introducing inaccurate assumptions of no motion between slices for multislice fMRI data. Improved sensitivity and specificity of the fMRI signal from mapping‐each‐slice‐to‐volume method is demonstrated in comparison with a stacked‐slice correction method by examining functional data from two normal volunteers. The data presented in a standard anatomical coordinate system suggest the reliability of the mapping‐each‐slice‐to‐volume method to detect the activation signals consistent between the two subjects. Magn Reson Med 41:964–972, 1999.

Semiautomatic registration of volumetric ultrasound scans

We demonstrate the ability to register easily and accurately volumetric ultrasound scans without significant data preprocessing or user intervention. Two volumetric ultrasound breast scan data sets were acquired from two different patients with breast cancer. Volumetric scan data were acquired by manually sweeping a linear array transducer mounted on a linear slider with a position encoder. The volumetric data set pairs consisted of color flow and/or power mode Doppler data sets acquired serially on the same patients. A previously described semiautomatic registration method based on maximizing mutual information was used to determine the transform between data sets. The results suggest that, even for the deformable breast, three-dimensional full affine transforms can be sufficient to obtain clinically useful registrations; warping may be necessary for increased registration accuracy. In conclusion, mutual information-based automatic registration as implemented on modern workstations is capable of yielding clinically useful registrations in times <35 min.

Least biased target selection in probabilistic atlas construction

Probabilistic atlas has broad applications in medical image segmentation and registration. The most common problem building a probabilistic atlas is picking a target image upon which to map the rest of the training images. Here we present a method to choose a target image that is the closest to the mean geometry of the population under consideration as determined by bending energy. Our approach is based on forming a distance matrix based on bending energies of all pair-wise registrations and performing multidimensional scaling (MDS) on the distance matrix.

Quantitative assessment of cartilage surface roughness in osteoarthritis using high frequency ultrasound

Osteoarthritis (OA) is a common disease which affects nearly 50% of people over age 60. Histologic evaluation suggests that fibrillations approximately 20-150 microns are among the earliest changes in the articular cartilage. We propose a technique to quantify these surface fibrillatory changes in osteoarthritic articular cartilage by considering the angular distribution of the envelope-detected backscattered pressure field from an incident 30-MHz focused transducer. The angular distribution of the scattered acoustic field from an inosonifying source will directly relate to the distribution of surface fibrillatory changes. Data are presented for three different grades (400, 500 and 600 grit) of commercially available emory paper and three samples of osteoarthritic femoral head articular cartilage, which were visually assessed as having smooth, intermediate and rough surfaces, respectively. Our preliminary results indicate a probable monotonic relationship between articular cartilage roughening and the degree of broadening in the angle-dependent pressure amplitude. When applied to the emory paper, the technique indicates sensitivity to differences as small as approximately 5-10 microns in mean roughness. This procedure may provide an extremely sensitive and reproducible means of quantifying and following the cartilage changes observed in early osteoarthritis.

Adaptive registration using local information measures

Rapidly advancing registration methods increasingly employ warping transforms. High degrees of freedom (DOF) warpings can be specified by manually placing control points or instantiating a regular, dense grid of control points everywhere. The former approach is laborious and prone to operator bias, whereas the latter is computationally expensive. We propose to improve upon the latter approach by adaptively placing control points where they are needed. Local estimates of mutual information (MI) and entropy are used to identify local regions requiring additional DOF.

Voxel-by-Voxel Functional Diffusion Mapping for Early Evaluation of Breast Cancer Treatment

Quantitative isotropic diffusion MRI and voxel-based analysis of the apparent diffusion coefficient (ADC) changes have been demonstrated to be able to accurately predict early response of brain tumors to therapy. The ADC value changes measured during pre- and posttherapy interval are closely correlated to treatment response. This work was demonstrated using a voxel-based analysis of ADC change during therapy in the brains of both rats and humans, following rigidly registering pre- and post-therapeutic ADC MRI exams. The primary goal of this paper is to extend this voxel-by-voxel analysis to assess therapeutic response in breast cancer. Nonlinear registration (with higher degrees of freedom) between the pre- and post-treatment exams is needed to ensure that the corresponding voxels actually contain similar cellular partial contributions due to soft tissue deformations in the breast and compartmental tumor changes during treatment as well. With limited data sets, we have observed the correlation between changes of ADC values and treatment response also exists in breast cancers. With diffusion scans acquired at three different timepoints (pre-treatment, early post-treatment and late post-treatment), we have also shown that ADC changes across responders within 5 weeks are a function of time interval after the initiation of treatment. Comparison of the experimental results with pathology shows that ADC changes can be used to evaluate early response of breast cancer treatment.

Determination of liver volume from CT scans using histogram cluster analysis

The histogram cluster analysis procedure (HICAP), which was developed by NASA for processing satellite images, classifies images into discrete clusters of pixels according to one or more arbitrary imaging variables. We incorporated this nonparametric, multivariate procedure in a semiautomatic computer algorithm for calculating total liver volume from CT scans and compared its performance with that of a human observer. Total liver volumes were calculated from CT scans in adult patients by the algorithm and by an experienced radiologist using the trackball controlled cursor at the CT console. Variability in the computer calculated volumes was determined by repeating calculations three times over the course of 3–12 months. Using HICAP in the univariate mode, we calculated total liver volumes from 28 contrast enhanced CT scans in 27 patients. Liver volumes calculated by the semiautomatic and manual methods had a median absolute difference of 3.6% (Vcomputer = 1.08 * Vmanual – 99.52 cc; r2 = 0.99). Median day-to-day variability of the computer calculated volumes was 1.9% (95% confidence interval: 1.3–2.7%). Using HICAP in a bivariate mode to illustrate its ability to incorporate two image features in one analysis, we studied an additional patient and compared total liver volume calculated from the univariate data set defined by the contrast enhanced CT scan with that calculated from the bivariate data set defined by nonenhanced and contrast enhanced CT scans. The HICAP errors were 4.1% in the univariate analysis and 0.4% in the bivariate analysis. It is concluded that this statistical clustering algorithm provides a clinically accurate, repeatable, and feasible method of in vivo liver volume determination.

Generating a normalized geometric liver model using warping

RATIONALE AND OBJECTIVES.Automated liver surface determination in abdominal computed tomography scans, currently difficult to achieve, is of interest to determine liver location and size for various medical applications, including radiation therapy treatment planning, surgical planning, and oncologic monitoring. The authors propose to facilitate automation by the addition of a priori shape information in the form of a liver model. METHODS.The normalized geometric liver model is generated by averaging outlines from a set of normal liver studies previously registered using thin-plate spline warping. The model consists of an averaged liver surface, a set of anatomic landmarks, and a deformation function. RESULTS.A liver model is presented and its ability to represent normal liver shapes is demonstrated. CONCLUSIONS.Liver surface warping provides a means of data normalization for model construction and a means of model deformation for representation of liver organ shapes.