Second-harmonic generation imaging analysis can help distinguish sarcoidosis from tuberculoid leprosy.

Sarcoidosis and tuberculoid leprosy (TL) are prototypes of granulomatous inflammation in dermatology, which embody one of the histopathology limitations in distinguishing some diseases. Recent advances in the use of nonlinear optical microscopy in skin have enabled techniques, such as second-harmonic generation (SHG), to become powerful tools to study the physical and biochemical properties of skin. We use SHG images to analyze the collagen network, to distinguish differences between sarcoidosis and TL granulomas. SHG images obtained from skin biopsies of 33 patients with TL and 24 with sarcoidosis retrospectively were analyzed using first-order statistics (FOS) and second-order statistics, such as gray-level co-occurrence matrix (GLCM). Among the four parameters evaluated (optical density, entropy, contrast, and second angular moment), only contrast demonstrated statistical significance, being higher in sarcoidosis (p = 0.02; 4908.31 versus 2822.17). The results may indicate insufficient differentiating power for most tested FOS and GLCM parameters in classifying sarcoidosis and TL granulomas, when used individually. But in combination with histopathology (H&E and complementary stains, such as silver and fast acid stains), SHG analysis, like contrast, can contribute to distinguishing between these diseases. This study can provide a way to evaluate collagen distribution in granulomatous diseases.

Impact of arterial input function selection on the accuracy of dynamic contrast-enhanced MRI quantitative analysis for the diagnosis of clinically significant prostate cancer.

PURPOSE: Using a limited temporal resolution dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) dataset to assess the impact of the arterial input function (AIF) choice on the transfer constant (K(trans) ) to distinguish prostate carcinoma (PCa) from benign tissue. MATERIALS AND METHODS: Thirty-eight patients with clinically important peripheral PCa (≥0.5 cc) were retrospectively studied. These patients underwent 1.5T multiparametric prostate MR with PCa and benign regions of interest (ROIs) selected using a visual registration with morphometric reconstruction obtained from radical prostatectomy. Using three pharmacokinetic (PK) analysis software programs, the mean K(trans) of ROIs was computed using three AIFs: an individual AIF (Ind-AIF) and two literature population average AIFs of Weinmann (W-AIF) and of Fritz-Hansen (FH-AIF). A pairwise comparison of the area under the receiver operating characteristic curves (AUROCC) obtained with different AIFs was performed. RESULTS: AUROCCs obtained with W-AIF (ranging from 0.801 to 0.843) were significantly higher than FH-AIF (ranging from 0.698 to 0.780, 0.002 ≤ P ≤ 0.045) and similar to or higher than Ind-AIF (ranging from 0.591 to 0.839, 0.014 ≤ P ≤ 0.9). Ind-AIF and FH-AIF provided similar AUROCC (0.34 ≤ P ≤ 0.81). The pairwise correlation of K(trans) values was moderate to very strong when comparing W-AIF with FH-AIF (the Spearman's correlation coefficients [SCCs] ranged from 0.55 to 0.93) and very weak to moderate when comparing W-AIF with Ind-AIF (the SCCs ranged from 0.018 to 0.59) or FH-AIF with Ind-AIF (the SCCs ranged from 0.30 to 0.51). CONCLUSION: W-AIF yielded a higher performance than FH-AIF and a similar or higher performance than Ind-AIF in distinguishing PCa from benign tissue.

The ability of the Lauge-Hansen classification to predict ligament injury and mechanism in ankle fractures: an MRI study.

OBJECTIVES: The Lauge-Hansen classification system was designed to predict the mechanism and ligament injury patterns of ankle fractures on the basis of x-rays. The purpose of this study was to evaluate the accuracy of these predicted injury sequences using magnetic resonance imaging (MRI) in a series of patients with ankle fractures. DESIGN: Retrospective cohort. SETTING: Two university level 1 trauma centers. PATIENTS: Fifty-nine patients with operative ankle fractures who were evaluated with both x-ray and MRI were included. INTERVENTION: All patients had a standard 3-view ankle x-ray series before fracture reduction, followed by an MRI. All plain x-rays were assigned to a Lauge-Hansen category by an experienced orthopedic traumatologist. MRI studies were subsequently read by an MRI musculoskeletal radiologist for the integrity of the ankle ligaments. MAIN OUTCOME MEASUREMENTS: After evaluation of the x-rays, fractures were classified according to the system of Lauge-Hansen, and the predicted presence, sequence, and mechanism of injury was determined. These were then compared to the actual injured structures on MRI in each case, and the ability of the Lauge-Hansen system to accurately predict the complete injury pattern was determined for the entire cohort. RESULTS: Average patient age was 59 (range: 18 to 84) years. Of the 59 ankle fractures evaluated, 37 (63%) were classified as supination external rotation, 11 (19%) were pronation external rotation, 1 (2%) was supination adduction, and 10 (17%) were not classifiable on the basis of the Lauge-Hansen system. Of the 49 fractures that fit into Lauge-Hansen categories, 26 (53%) had patterns of ligamentous injury and fracture morphology that did not coincide with the Lauge-Hansen predictions. A common fracture pattern was observed in 8 of the 10 unclassifiable fractures, which included a high spiral fracture of the fibula, vertical shear fracture of the medial malleolus, posterior malleolar fracture, and complete tears of the anterior-inferior tibiofibular ligament and the interosseous membrane. In addition, over 65% of patients in this series had complete ligamentous injury and a fracture of the malleolus to which the ligament attaches. CONCLUSIONS: These results demonstrate that the Lauge-Hansen classification system may have some limitations as a predictor of the mechanism of injury and the presence of soft-tissue damage associated with ankle fractures. The identification of a novel pattern of ankle fracture also illustrates how the system fails to describe all possible fracture patterns. For these reasons, we recommend that the Lauge-Hansen system be used only as a guide in the diagnosis and management of ankle fractures and not solely relied upon for treatment decisions. Although the exact clinical implications of the variety of ligamentous injuries observed on MRI are yet to be determined, this technique may be useful in individual cases in which doubt about joint stability and soft-tissue integrity exists. Additionally, MRI may be helpful in planning surgical approaches in atypical fractures in which injury patterns are less predictable solely on the basis of x-ray.

Support vector machines for temporal classification of block design fMRI data.

This paper treats support vector machine (SVM) classification applied to block design fMRI, extending our previous work with linear discriminant analysis [LaConte, S., Anderson, J., Muley, S., Ashe, J., Frutiger, S., Rehm, K., Hansen, L.K., Yacoub, E., Hu, X., Rottenberg, D., Strother S., 2003a. The evaluation of preprocessing choices in single-subject BOLD fMRI using NPAIRS performance metrics. NeuroImage 18, 10-27; Strother, S.C., Anderson, J., Hansen, L.K., Kjems, U., Kustra, R., Siditis, J., Frutiger, S., Muley, S., LaConte, S., Rottenberg, D. 2002. The quantitative evaluation of functional neuroimaging experiments: the NPAIRS data analysis framework. NeuroImage 15, 747-771]. We compare SVM to canonical variates analysis (CVA) by examining the relative sensitivity of each method to ten combinations of preprocessing choices consisting of spatial smoothing, temporal detrending, and motion correction. Important to the discussion are the issues of classification performance, model interpretation, and validation in the context of fMRI. As the SVM has many unique properties, we examine the interpretation of support vector models with respect to neuroimaging data. We propose four methods for extracting activation maps from SVM models, and we examine one of these in detail. For both CVA and SVM, we have classified individual time samples of whole brain data, with TRs of roughly 4 s, thirty slices, and nearly 30,000 brain voxels, with no averaging of scans or prior feature selection.

Measurement of retinal vessel widths from fundus images based on 2-D modeling.

Changes in retinal vessel diameter are an important sign of diseases such as hypertension, arteriosclerosis and diabetes mellitus. Obtaining precise measurements of vascular widths is a critical and demanding process in automated retinal image analysis as the typical vessel is only a few pixels wide. This paper presents an algorithm to measure the vessel diameter to subpixel accuracy. The diameter measurement is based on a two-dimensional difference of Gaussian model, which is optimized to fit a two-dimensional intensity vessel segment. The performance of the method is evaluated against Brinchmann-Hansen's half height, Gregson's rectangular profile and Zhou's Gaussian model. Results from 100 sample profiles show that the presented algorithm is over 30% more precise than the compared techniques and is accurate to a third of a pixel.