Adaptive speckle reduction for improving the differential diagnosis of breast lesions

A tissue-specific adaptive texture filter for medical ultrasound images

The interpretation of ultrasound images remains a difficult task and the opinion of different doctors is generally not unequivocal. Therefore, there is a growing interest in the field of computer-aided diagnosis. In the field of medical image processing, computer-aided diagnosis includes image enhancement to facilitate visual interpretation, automatic indication of affected areas, organs and other regions of medical interest, the performance of automatic measurements and image registration. In this article, we introduce a new algorithm for ultrasound image enhancement that employs a multivariate texture classifier based on the co-occurrence matrix, which, in combination with an adaptive texture smoothing filter, is used to enhance the visual difference between and improve boundary detection between healthy neonatal brain tissue and tissue affected by periventricular leukomalacia. For a quantitative comparison, we delineate the periventricular leukomalacia-affected regions with two different active contours before and after processing 10 images with the proposed technique and several speckle filters from the literature. The semi-automatic delineations thus obtained are compared with the manual delineations of a neonatologist. In all cases, the average delineation achieved with the proposed technique is closer to that of the manual expert delineation than when the images are processed with the other techniques.

Relevance of sonographic B-mode criteria and computer-aided ultrasonic tissue characterization in differential/diagnosis of solid breast masses

We aimed to evaluate the differential diagnostic value of a method of computer-assisted texture analysis in comparison to established ultrasonographic B-mode characteristics in the examination of solid breast masses. At two centers, 77 patients presenting with a solid mass on B-mode scan were studied at 7.5 MHz. Description of B-mode appearance included assessment of tumor shape, borders, presence of an echogenic rim, tissue architecture, internal echo structure, absorption and elasticity. For statistical pattern recognition, the following parameters were used: form factor, mean grey level, signal-to-noise ratio, mean gradient and correlation from the co-occurrence matrix. At center 1, the most decisive parameter for differential diagnosis was distortion of tissue architecture (sensitivity, SN, 83%; specificity, SP, 92%) and, at center 2, relation to the adjacent tissue (SN 93%, SP 92%). Among texture parameters, best discrimination was achieved for correlation from the co-occurrence matrix at center 1 (SN 58%, SP 73%) and for form factor at center 2 (SN 93%, SP 77%). Among sonographic criteria, the highest contribution to the diagnosis was found for an unsharp border (odds ratio, OR, 12.2), architectural distortion (OR, 8.6), fixation to skin or chest wall (OR, 9.0) and fixation to adjacent breast parenchyma (OR, 8.8), according to texture analysis for parameters form factor (OR 4.0) and correlation from the co-occurrence matrix (OR 4.7). Ultrasonographic texture analysis can be helpful as an additional parameter in differential diagnosis of breast tumors, but did not reach differential diagnostic accuracy of sonomorphologic features.