Semi-Automatic Normalized Entropy Characterization of Metastatic Renal Cell Cancer Via Spatio-Textural Tumour Classification

Normalized entropy is a strong predictor of patient survival rate for lung and renal cell cancers in CT images. In this work, we present a semi-automatic algorithm to compute the normalized entropy of lung metastasis from renal cell cancer via spatio-textural tumour classification. The proposed algorithm only requires a single click from the user to calculate the normalized entropy for tumours, thus, minimizing the amount of user intervention required. We evaluated the performance of the proposed approach in computing normalized entropy with respect to the normalized entropy calculated via manual clinical segmentation, and a correlation of 0.87 was achieved.