Auto-segmentation of normal and target structures in head and neck CT images: a feature-driven model-based approach.

PURPOSE: Intensity modulated radiation therapy (IMRT) allows greater control over dose distribution, which leads to a decrease in radiation related toxicity. IMRT, however, requires precise and accurate delineation of the organs at risk and target volumes. Manual delineation is tedious and suffers from both interobserver and intraobserver variability. State of the art auto-segmentation methods are either atlas-based, model-based or hybrid however, robust fully automated segmentation is often difficult due to the insufficient discriminative information provided by standard medical imaging modalities for certain tissue types. In this paper, the authors present a fully automated hybrid approach which combines deformable registration with the model-based approach to accurately segment normal and target tissues from head and neck CT images. METHODS: The segmentation process starts by using an average atlas to reliably identify salient landmarks in the patient image. The relationship between these landmarks and the reference dataset serves to guide a deformable registration algorithm, which allows for a close initialization of a set of organ-specific deformable models in the patient image, ensuring their robust adaptation to the boundaries of the structures. Finally, the models are automatically fine adjusted by our boundary refinement approach which attempts to model the uncertainty in model adaptation using a probabilistic mask. This uncertainty is subsequently resolved by voxel classification based on local low-level organ-specific features. RESULTS: To quantitatively evaluate the method, they auto-segment several organs at risk and target tissues from 10 head and neck CT images. They compare the segmentations to the manual delineations outlined by the expert. The evaluation is carried out by estimating two common quantitative measures on 10 datasets: volume overlap fraction or the Dice similarity coefficient (DSC), and a geometrical metric, the median symmetric Hausdorff distance (HD), which is evaluated slice-wise. They achieve an average overlap of 93% for the mandible, 91% for the brainstem, 83% for the parotids, 83% for the submandibular glands, and 74% for the lymph node levels. CONCLUSIONS: Our automated segmentation framework is able to segment anatomy in the head and neck region with high accuracy within a clinically-acceptable segmentation time.

A feature-based approach for refinement of model-based segmentation of low contrast structures.

Accuracy and robustness are fundamental requirements of any automated method used for segmentation of medical images. Model-based segmentation (MBS) is a well established technique, where uncertainties in image content can be to a certain extent compensated by the use of prior shape information. This approach is, however, often problematic in cases where image information does not allow for generating a strong feature response, one example being soft tissue organs in CT data, which typically appear in low contrast. In this paper, we enhance our recently proposed framework for voxel classification-based refinement of MBS using a level-set segmentation technique with shape priors. We also introduce a novel feature weighting methodology that improves the performance of the classifier, demonstrating results superior to the previous feature selection method. Results of fully automated segmentation of low contrast organs in head and neck CT are presented. Compared to our previous approach, we have achieved an increase of up to 22% in segmentation accuracy.

A framework for optimizing measurement weight maps to minimize the required sample size.

We propose a fully automatic statistical framework for identifying the non-negative, real-valued weight map that best discriminate between two groups of objects. Given measurements on a spatially defined grid, a numerical optimization scheme is used to find the weight map that minimizes the sample size required to discriminate the two groups. The weight map produced by the method reflects the relative importance of the different areas in the objects, and the resulting sample size reduction is an important end goal in situations where data collection is difficult or expensive. An example is in clinical studies where the cost and the patient burden are directly related to the number of participants needed for the study. In addition, inspection of the weight map might provide clues that can lead to a better clinical understanding of the objects and pathologies being studied. The method is evaluated on synthetic data and on clinical data from knee cartilage MRI. The clinical data contain a total of 159 subjects aged 21-81 years and ranked from zero to four on the Kellgren-Lawrence osteoarthritis severity scale. Compared to a uniform weight map, we achieve sample size reductions up to 58% for cartilage thickness measurements. Based on quantifications from both morphometric and textural based imaging features, we also identify the most pathological areas in the articular cartilage.

Identification of progressors in osteoarthritis by combining biochemical and MRI-based markers.

INTRODUCTION: At present, no disease-modifying osteoarthritis drugs (DMOADS) are approved by the FDA (US Food and Drug Administration); possibly partly due to inadequate trial design since efficacy demonstration requires disease progression in the placebo group. We investigated whether combinations of biochemical and magnetic resonance imaging (MRI)-based markers provided effective diagnostic and prognostic tools for identifying subjects with high risk of progression. Specifically, we investigated aggregate cartilage longevity markers combining markers of breakdown, quantity, and quality. METHODS: The study included healthy individuals and subjects with radiographic osteoarthritis. In total, 159 subjects (48% female, age 56.0 +/- 15.9 years, body mass index 26.1 +/- 4.2 kg/m2) were recruited. At baseline and after 21 months, biochemical (urinary collagen type II C-telopeptide fragment, CTX-II) and MRI-based markers were quantified. MRI markers included cartilage volume, thickness, area, roughness, homogeneity, and curvature in the medial tibio-femoral compartment. Joint space width was measured from radiographs and at 21 months to assess progression of joint damage. RESULTS: Cartilage roughness had the highest diagnostic accuracy quantified as the area under the receiver-operator characteristics curve (AUC) of 0.80 (95% confidence interval: 0.69 to 0.91) among the individual markers (higher than all others, P < 0.05) to distinguish subjects with radiographic osteoarthritis from healthy controls. Diagnostically, cartilage longevity scored AUC 0.84 (0.77 to 0.92, higher than roughness: P = 0.03). For prediction of longitudinal radiographic progression based on baseline marker values, the individual prognostic marker with highest AUC was homogeneity at 0.71 (0.56 to 0.81). Prognostically, cartilage longevity scored AUC 0.77 (0.62 to 0.90, borderline higher than homogeneity: P = 0.12). When comparing patients in the highest quartile for the longevity score to lowest quartile, the odds ratio of progression was 20.0 (95% confidence interval: 6.4 to 62.1). CONCLUSIONS: Combination of biochemical and MRI-based biomarkers improved diagnosis and prognosis of knee osteoarthritis and may be useful to select high-risk patients for inclusion in DMOAD clinical trials.

Resolving crossings in the corticospinal tract by two-tensor streamline tractography: Method and clinical assessment using fMRI.

An inherent drawback of the traditional diffusion tensor model is its limited ability to provide detailed information about multidirectional fiber architecture within a voxel. This leads to erroneous fiber tractography results in locations where fiber bundles cross each other. This may lead to the inability to visualize clinically important tracts such as the lateral projections of the corticospinal tract. In this report, we present a deterministic two-tensor eXtended Streamline Tractography (XST) technique, which successfully traces through regions of crossing fibers. We evaluated the method on simulated and in vivo human brain data, comparing the results with the traditional single-tensor and with a probabilistic tractography technique. By tracing the corticospinal tract and correlating with fMRI-determined motor cortex in both healthy subjects and patients with brain tumors, we demonstrate that two-tensor deterministic streamline tractography can accurately identify fiber bundles consistent with anatomy and previously not detected by conventional single-tensor tractography. When compared to the dense connectivity maps generated by probabilistic tractography, the method is computationally efficient and generates discrete geometric pathways that are simple to visualize and clinically useful. Detection of crossing white matter pathways can improve neurosurgical visualization of functionally relevant white matter areas.

Osteoarthritic cartilage is more homogeneous than healthy cartilage: identification of a superior region of interest colocalized with a major risk factor for osteoarthritis.

RATIONALE AND OBJECTIVES: Cartilage loss as determined by magnetic resonance imaging (MRI) or joint space narrowing as determined by x-ray is the result of cartilage erosion. However, metabolic processes within the cartilage that later result in cartilage loss may be a more sensitive assessment method for early changes. Recently, it was shown that cartilage homogeneity visualized by MRI representing the biochemical changes undergoing in the cartilage is a potential marker for early detection of knee osteoarthritis (OA) and is also able to significantly separate groups of healthy subjects from those with OA. The purpose of this study was twofold. First, we wished to evaluate whether the results on cartilage homogeneity from the previous study can be reproduced using an independent population. Second, based on the homogeneity framework, we present an automatic technique that partitions the region of interest in the cartilage that contributes most to discrimination between healthy and OA subjects and allows for identification of the most implicated areas in early OA. These findings may allow further investigation of whether cartilage homogeneity reveals a predisposition for OA or whether it evolves as a consequence to disease and thereby can be used as a progression biomarker. MATERIALS AND METHODS: A total of 283 right and left knees from 159 subjects aged 21 to 81 years were scanned using a Turbo 3D T1 sequence on a 0.18-T MRI Esaote scanner. The medial compartment of the tibial cartilage sheet was segmented using a fully automatic voxel classification scheme based on supervised learning. From the segmented cartilage sheet, homogeneity was quantified by measuring entropy from the distribution of signal intensities inside the compartment. Each knee was examined by radiography, and the knees were categorized by the Kellgren and Lawrence (KL) Index. Next, based on a gradient descent optimization technique, the cartilage region that contributed to the maximum statistical significance of homogeneity in separating healthy subjects from the diseased was partitioned. The generalizability of the region was evaluated by testing for overfitting. Three different regularization techniques were evaluated for reducing overfitting errors. RESULTS: The P values for separating the different groups based on cartilage homogeneity were 2 x 10(-5) (KL 0 versus KL 1) and 1 x 10(-7) (KL 0 versus KL >0). Using the automatic gradient descent technique, the partitioned region was toward the peripheral part of the cartilage sheet. Using this region, the P values for separating the different groups based on homogeneity were 5 x 10(-9) (KL 0 versus KL 1) and 1 x 10(-15) (KL 0 versus KL >0). The precision of homogeneity for the partitioned region assessed as a test-retest root-mean-square coefficient of variation was 3.3%. Bootstrapping proved to be an effective regularization tool in reducing overfitting errors. CONCLUSION: The validation study supported the use of cartilage homogeneity as a tool for the early detection of knee OA and for separating groups of healthy subjects from those who have disease. Our automatic, unbiased partitioning algorithm based on a general statistical framework outlined the cartilage region of interest that best separated healthy from OA conditions on the basis of homogeneity discrimination. We have shown that OA affects certain areas of the cartilage more distinctly, and these areas are located more toward the peripheral region of the cartilage. We propose that this region corresponds anatomically to cartilage covered by the meniscus in healthy subjects. This finding may provide valuable clues in the early detection and monitoring of OA and thus may improve treatment efficacy.