Heterogeneous Tissue Characterization

“

The paper deals with the topic of ultrasound tissue characterization by examining tumor heterogeneity and assessing its response to treatment. Quantitative analysis of tumor tissue is challenging since changes in tumor texture in response to treatment are subtle, and the radiologists' visual analysis of tumour volumes is tedious and time-consuming. The paper should be of interest to readers in the area of ultrasound tissue characterization (tumor characterization, texture and image analysis). The work in the paper follows a fractal approach for evaluating the performance of different well-known statistical models for an automated detection of liver tumor response to chemotherapy treatment. The approach relates the fractal characteristics of the tissue scatterers to the underlying statistical properties derived from the radio-frequency (RF) envelope-detected signal. The textural measures represented by the fractal dimension (assessing the degree of self-similarity) and the derived Lacunarity measure (quantifying the level of spaces within the tissue texture) were related to the scatterers' spatial distribution and density, respectively. The fractal features were extracted in a multiresolution fashion in order to account for the varying complexity of the tumor tissue, and the modeling of the RF envelope-detected signal was based on a multimodal statistical distribution. Finally the extracted multifractal features were fed to a classifier for automated decision support. Liver tumor ultrasound scanning is recently becoming more recommended as a first diagnosis option for early prediction of response to chemotherapy treatment. The paper shows that employing the most relevant and practical statistical model while taking into consideration the tissue fractal features is substantial for improving clinical implications, and therefore could have potential consequences to the design of an early and effective therapy.

”