Performance of topological texture features to classify fibrotic interstitial lung disease patterns.

PURPOSE: Topological texture features were compared in their ability to classify "honeycombing," a morphological pattern that is considered indicative for the presence of fibrotic interstitial lung disease in high-resolution computed tomography (HRCT) images. METHODS: For 14 patients with known occurrence of honeycombing, a stack of 70 axial, lung kernel reconstructed images was acquired from HRCT chest exams. A set of 964 regions of interest of both healthy and pathological (356) lung tissue was identified by an experienced radiologist. Texture features were extracted using statistical features (Stat), six properties calculated from gray-level co-occurrence matrices (GLCMs), Minkowski dimensions (MDs), and three Minkowski functionals (MFs) (e.g., MF.Euler). A naïve Bayes (NB) and k-nearest-neighbor (k-NN) classifier, a multilayer radial basis functions network (RBFN), and a support vector machine with a radial basis function (SVMrbf) kernel were optimized in a tenfold cross-validation for each texture vector, and the classification accuracy was calculated on independent test sets as a quantitative measure of automated tissue characterization. A Wilcoxon signed-rank test was used to compare two accuracy distributions and the significance thresholds were adjusted for multiple comparisons by the Bonferroni correction. RESULTS: The best classification results were obtained by the MF features, which performed significantly better than all the standard Stat, GLCM, and MD features (p < 0.001) for both classifiers. The highest accuracies were found for MF.Euler (93.6%, 94.9%, 94.2%, and 95.0% for NB, k-NN, RBFN, and SVMrbf, respectively). The best groups of standard texture features were a Stat and GLCM ("homogeneity") feature set (up to 91.8%). CONCLUSIONS: The results indicate that advanced topological texture features derived from MFs can provide superior classification performance in computer-assisted diagnosis of fibrotic interstitial lung disease patterns when compared to standard texture analysis methods.

Pulmonary abnormalities in immunocompromised patients: comparative detection with parallel acquisition MR imaging and thin-section helical CT.

PURPOSE: To compare parallel acquisition magnetic resonance (MR) imaging with thin-section helical computed tomography (CT) for depiction of pulmonary abnormalities suggestive of pneumonia in immunocompromised patients. MATERIALS AND METHODS: The institutional review board approved this study; prior consent was obtained. Thirty consecutive neutropenic patients (10 women, 20 men; mean age, 51 years +/- 15 [standard deviation]; range, 25-75 years) with fever of unknown origin or clinical signs and symptoms of lung infection were examined with breath-hold single-shot half-Fourier turbo spin-echo MR imaging. To reduce image blurring and increase MR signal in the lungs, the echo time was shortened with generalized autocalibrating partially parallel acquisition (GRAPPA). Patients underwent thoracic CT (four detector rows and 1-mm section thickness [4 x 1 mm]; pitch, 6) as reference standard. Pulmonary abnormalities (ill-defined nodules, ground-glass opacity areas, and consolidation), their location and distribution, and lesion characteristics were analyzed at MR imaging by three readers, blinded to results of CT, in consensus. Frequencies were calculated for each feature; paired Wilcoxon rank sum test was used to examine whether differences between CT and MR imaging features were statistically significant (alpha < .05). Bonferroni adjustments were performed. Overall sensitivity, specificity, and positive and negative predictive values were determined. RESULTS: Twenty-two patients had pulmonary abnormalities at CT. In 21 (95%) patients, pneumonia was correctly diagnosed with MR imaging. One false-negative finding occurred in a patient with ill-defined nodules smaller than 1 cm at CT. One false-positive finding with MR imaging was the result of blurring and respiratory artifacts (sensitivity, 95%; specificity, 88%; positive predictive value, 95%; negative predictive value, 88%). There was no significant difference in lesion location and distribution. CONCLUSION: With parallel imaging (GRAPPA technique) and fast MR imaging, detection of pulmonary abnormalities is almost as good as with CT. MR imaging has a slight disadvantage in its lower capability to assist in characterization of specific internal features, such as cavitations.

Catheter-based intraluminal optical coherence tomography (OCT) of the ureter: ex-vivo correlation with histology in porcine specimens.

Intraluminal optical coherence tomography (OCT) applies coherent light to provide cross-sectional images with a spatial resolution of 10-25 microm. We compared OCT and matching whole-mount histology microscopy sections of porcine upper ureters ex vivo for visualization and delineation of different tissue layers of the ureteral wall. Porcine ureters (six specimens, 24 quadrants) were flushed with normal saline solution prior to insertion of the OCT catheter (diameter, 0.014 inch, OCT wavelength, 1,300+/-20 nm). Cross-sectional OCT images were obtained in marked locations before specimens were fixed in 4% formalin, cut at marked locations, whole-mounted, and stained with hematoxilin and eosin. Visualization and delineation of different tissue layers of the ureteral wall by OCT was compared with matching histology by two independent observers (O1,O2). OCT distinguished tissue layers of the ureteral wall in all quadrants. In OCT images, O1/O2 delineated urothelium and lamina propria in 23/24 quadrants, lamina propria and muscle layer in 19/16 quadrants, inner and outer muscle layer in 13/0 quadrants, and urothelial cell layers in 13/2 quadrants, respectively. Intraluminal OCT provides histology-like images of the ureter in porcine specimens ex vivo and reliably distinguishes between urothelium and deeper tissue layers of the ureteral wall.

The role of new imaging techniques in diagnosis and staging of malignant pleural mesothelioma.

This review highlights the different imaging modalities for the detection of malignant pleural mesothelioma. The chest film is the initial diagnostic tool of choice because it is easy to perform, inexpensive, and widely available. Unfortunately, it demonstrates malignant pleural mesothelioma in later stages of disease and is not suitable for an early sensitive and specific diagnosis. Computed tomography is capable of distinguishing the different forms of pleural abnormalities. The knowledge of computed tomography findings is important for differentiation of malignant pleural diseases from benign diseases. Contrast-enhanced magnetic resonance imaging in three planes can be clinically useful to differentiate pleural mesothelioma from other malignancies or from benign pleural diseases. Furthermore, magnetic resonance imaging offers the possibility to differentiate invasion of the diaphragm from transdiaphragmatic tumor growth, and, in patients who are surgical candidates and who have questionable areas of local tumor extension on computed tomography, magnetic resonance imaging may provide additional information to plan or avoid surgery. Improvements in the detection of regional and distant metastases are needed to identify patients most likely to benefit from aggressive combined modality treatment regimes. In this context, positron emission tomography is a metabolic imaging technique that offers the possibility to evaluate active malignant cells. Drawbacks to this technique include false-positive findings, which may occur at sides of inflammation, and lesser anatomic detail information. A complete and accurate staging of malignant pleural mesothelioma is essential to evaluate the efficacy of new therapeutic strategies. This implies the need to be familiar with the most recently developed staging system from the International Mesothelioma Interest Group.