Nicholas Dowson

Mutual Information for Lucas-Kanade Tracking (MILK): An Inverse Compositional Formulation

Mutual information (Ml) is popular for registration via function optimization. This work proposes an inverse compositional formulation of Ml for Levenberg-Marquardt optimization. This yields a constant Hessian, which may be precomputed. Speed improvements of 15 percent were obtained, with convergence accuracies similar those of the standard formulation.

A unifying framework for mutual information methods for use in non-linear optimisation

Many variants of MI exist in the literature. These vary primarily in how the joint histogram is populated. This paper places the four main variants of MI: Standard sampling, Partial Volume Estimation (PVE), In-Parzen Windowing and Post-Parzen Windowing into a single mathematical framework. Jacobians and Hessians are derived in each case. A particular contribution is that the non-linearities implicit to standard sampling and post-Parzen windowing are explicitly dealt with. These non-linearities are a barrier to their use in optimisation. Side-by-side comparison of the MI variants is made using eight diverse data-sets, considering computational expense and convergence. In the experiments, PVE was generally the best performer, although standard sampling often performed nearly as well (if a higher sample rate was used). The widely used sum of squared differences metric performed as well as MI unless large occlusions and non-linear intensity relationships occurred. The binaries and scripts used for testing are available online.

Correlation of MRI-derived apparent diffusion coefficients in newly diagnosed gliomas with [18F]-fluoro-L-dopa PET: what are we really measuring with minimum ADC?

BACKGROUND AND PURPOSE: There is significant interest in whether diffusion-weighted MR imaging indices, such as the minimum apparent diffusion coefficient, may be useful clinically for preoperative tumor grading and treatment planning. To help establish the pathologic correlate of minimum ADC, we undertook a study investigating the relationship between minimum ADC and maximum FDOPA PET uptake in patients with newly diagnosed glioblastoma multiforme. MATERIALS AND METHODS: MR imaging and FDOPA PET data were acquired preoperatively from 15 patients who were subsequently diagnosed with high-grade brain tumor (WHO grade III or IV) by histopathologic analysis. ADC and SUVR normalized FDOPA PET maps were registered to the corresponding CE MR imaging. Regions of minimum ADC within the FDOPA-defined tumor volume were anatomically correlated with areas of maximum FDOPA SUVR uptake. RESULTS: Minimal anatomic overlap was found between regions exhibiting minimum ADC (a putative marker of tumor cellularity) and maximum FDOPA SUVR uptake (a marker of tumor infiltration and proliferation). FDOPA SUVR measures for tumoral regions exhibiting minimum ADC (1.36 ± 0.22) were significantly reduced compared with those with maximum FDOPA uptake (2.45 ± 0.88, P = .0001). CONCLUSIONS: There was a poor correlation between minimum ADC and the most viable/aggressive component of high-grade gliomas. This study suggests that other factors, such as tissue compression and ischemia, may be contributing to restricted diffusion in GBM. Caution should be exercised in the clinical use of minimum ADC as a marker of tumor grade and the use of this index for guiding tumor biopsies preoperatively.

Simultaneous modeling and tracking (SMAT) of feature sets

A novel method for the simultaneous modeling and tracking (SMAT) of a feature set during motion sequence is proposed. The method requires no prior information. Instead the a posteriori distribution of appearance and shape is built up incrementally using an exemplar based approach. The resulting model is less optimal than when a priori data is used, but can be built in real-time. Data in any form may be used, provided a distance measure and a means to classify outliers exists. Here, a two tier implementation of SMAT is used: at the feature level, mutual information is used to track image patches; and at the object level, a structure model is built from the feature positions. As experiments demonstrate, the tracker is robust and operates in real-time without requiring prelearned data.

Hypoxia Imaging in Gliomas With 18F-Fluoromisonidazole PET: Toward Clinical Translation

There is significant interest in the development of improved image-guided therapy for neuro-oncology applications. Glioblastomas (GBM) in particular present a considerable challenge because of their pervasive nature, propensity for recurrence, and resistance to conventional therapies. MRI is routinely used as a guide for planning treatment strategies. However, this imaging modality is not able to provide images that clearly delineate tumor boundaries and affords only indirect information about key tumor pathophysiology. With the emergence of PET imaging with new oncology radiotracers, mapping of tumor infiltration and other important molecular events such as hypoxia is now feasible within the clinical setting. In particular, the importance of imaging hypoxia levels within the tumoral microenvironment is gathering interest, as hypoxia is known to play a central role in glioma pathogenesis and resistance to treatment. One of the hypoxia radiotracers known for its clinical utility is (18)F-fluoromisodazole ((18)F-FMISO). In this review, we highlight the typical causes of treatment failure in gliomas that may be linked to hypoxia and outline current methods for the detection of hypoxia. We also provide an overview of the growing body of studies focusing on the clinical translation of (18)F-FMISO PET imaging, strengthening the argument for the use of (18)F-FMISO hypoxia imaging to help optimize and guide treatment strategies for patients with glioblastoma.

Hashed Nonlocal Means for Rapid Image Filtering

Denoising algorithms can alleviate the trade-off between noise-level and acquisition time that still exists for certain image types. Nonlocal means, a recently proposed technique, outperforms other methods in removing noise while retaining image structure, albeit at prohibitive computational cost. Modifications have been proposed to reduce the cost, but the method is still too slow for practical filtering of 3D images. This paper proposes a hashed approach to explicitly represent two summed frequency (hash) functions of local descriptors (patches), utilizing all available image data. Unlike other approaches, the hash spaces are discretized on a regular grid, so primarily linear operations are used. The large memory requirements are overcome by recursing the hash spaces. Additional speed gains are obtained by using a marginal linear interpolation method. Careful choice of the patch features results in high computational efficiency, at similar accuracies. The proposed approach can filter a 3D image in less than a minute versus 15 minutes to 3 hours for existing nonlocal means methods.

Estimating the Joint Statistics of Images Using Nonparametric Windows with Application to Registration Using Mutual Information

Recently, the Non-Parametric (NP) Windows has been proposed to estimate the statistics of real 1D and 2D signals. NP Windows is accurate, because it is equivalent to sampling images at a high (infinite) resolution for an assumed interpolation model. This paper extends the proposed approach to consider joint distributions of image-pairs. Secondly, Greens Theorem is used to simplify the previous NP Windows algorithm. Finally, a resolution aware NP Windows algorithm is proposed, to improve robustness to relative scaling between an image-pair. Comparative testing of 2D image registration was performed using translation-only and affine transformations. Although more expensive than other methods, NP Windows frequently demonstrated superior performance for bias (distance between ground truth and global maximum) and frequency of convergence. Unlike other methods, the number of samples and histogram bin-size has little effect on NP Windows, and the prior selection of a kernel is not required.

PET, MRI, and simultaneous PET/MRI in the development of diagnostic and therapeutic strategies for glioma.

Glioma is the most aggressive brain tumour, resulting in death often within 1-2 years. Current treatment strategies involve surgical resection followed by chemoradiation therapy. Despite continuing improvements in the delivery of adjuvant therapies, there has not been a dramatic increase in survival for glioma. Molecular imaging techniques have become central in the development of new therapeutic strategies in recent years. The multimodal imaging technology of positron emission tomography/magnetic resonance imaging (PET/MRI) has recently been realised on a preclinical scale and the effect of this technology is starting to be observed in preclinical drug development for glioma. Here, we propose that PET/MRI will play an integral part in the development of new diagnostic and therapeutic strategies for glioma.

Optical Differentiation Between Malignant and Benign Lymphadenopathy by Grey Scale Texture Analysis of Endobronchial Ultrasound Convex Probe Images

BACKGROUND Morphologic and sonographic features of endobronchial ultrasound (EBUS) convex probe images are helpful in predicting metastatic lymph nodes. Grey scale texture analysis is a well-established methodology that has been applied to ultrasound images in other fields of medicine. The aim of this study was to determine if this methodology could differentiate between benign and malignant lymphadenopathy of EBUS images. METHODS Lymph nodes from digital images of EBUS procedures were manually mapped to obtain a region of interest and were analyzed in a prediction set. The regions of interest were analyzed for the following grey scale texture features in MATLAB (version 7.8.0.347 [R2009a]): mean pixel value, difference between maximal and minimal pixel value, SEM pixel value, entropy, correlation, energy, and homogeneity. Significant grey scale texture features were used to assess a validation set compared with fluoro-D-glucose (FDG)-PET-CT scan findings where available. RESULTS Fifty-two malignant nodes and 48 benign nodes were in the prediction set. Malignant nodes had a greater difference in the maximal and minimal pixel values, SEM pixel value, entropy, and correlation, and a lower energy (P < .0001 for all values). Fifty-one lymph nodes were in the validation set; 44 of 51 (86.3%) were classified correctly. Eighteen of these lymph nodes also had FDG-PET-CT scan assessment, which correctly classified 14 of 18 nodes (77.8%), compared with grey scale texture analysis, which correctly classified 16 of 18 nodes (88.9%). CONCLUSIONS Grey scale texture analysis of EBUS convex probe images can be used to differentiate malignant and benign lymphadenopathy. Preliminary results are comparable to FDG-PET-CT scan.

Topology-corrected segmentation and local intensity estimates for improved partial volume classification of brain cortex in MRI.

In magnetic resonance imaging (MRI), accuracy and precision with which brain structures may be quantified are frequently affected by the partial volume (PV) effect. PV is due to the limited spatial resolution of MRI compared to the size of anatomical structures. Accurate classification of mixed voxels and correct estimation of the proportion of each pure tissue (fractional content) may help to increase the precision of cortical thickness estimation in regions where this measure is particularly difficult, such as deep sulci. The contribution of this work is twofold: on the one hand, we propose a new method to label voxels and compute tissue fractional content, integrating a mechanism for detecting sulci with topology preserving operators. On the other hand, we improve the computation of the fractional content of mixed voxels using local estimation of pure tissue intensity means. Accuracy and precision were assessed using simulated and real MR data and comparison with other existing approaches demonstrated the benefits of our method. Significant improvements in gray matter (GM) classification and cortical thickness estimation were brought by the topology correction. The fractional content root mean squared error diminished by 6.3% (p<0.01) on simulated data. The reproducibility error decreased by 8.8% (p<0.001) and the Jaccard similarity measure increased by 3.5% on real data. Furthermore, compared with manually guided expert segmentations, the similarity measure was improved by 12.0% (p<0.001). Thickness estimation with the proposed method showed a higher reproducibility compared with the measure performed after partial volume classification using other methods.