What is it about?

Improve the detection and classification of the subtle low activity regions (necrotic regions) within the tumor tissue for early prediction of response to chemotherapy treatment.

Featured Image

Why is it important?

Characteizing tissue via Ultrasound is considered fast, easy, and safe as compared to other imaging modalities; however, the texture in ultrasound tend to have a speckle appearance due to the low SNR and imaging artifacts, making it difficult to characterise the structures within the tissue. This work takes a step further and aims to characterise the variations within the tumor tissue itself - which we call inter-heterogeneity. The work will show that regions with low aggression behavior (which we define as low activity) would correspond to necrotic regions, and hence indicating that tumor is responding to treatment. These low activity regions appear subtle and are usually difficult to discern by eye.

Perspectives

The proposed new multifractal Nakagami-based feature descriptor (MNF) algorithm: - Exploits the surface roughness to overcome speckle tissue intensity variation - Employs an adaptive tissue characterisation analysis at various appropriate scales to account for the distribution mixtures complexities - Computes local parametric Nakagami region of interests via MLE in a volumetric fashion for a more accurate estimation. - Demonstrates good visualization of mass heterogeneity.

Dr Omar S Al-Kadi
University of Jordan

Read the Original

This page is a summary of: Quantification of ultrasonic texture intra-heterogeneity via volumetric stochastic modeling for tissue characterization, Medical Image Analysis, April 2015, Elsevier,
DOI: 10.1016/j.media.2014.12.004.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page