Characterizing intra-tumoral heterogeneity for response and outcome prediction using radiomic spatial textural descriptor (RADISTAT)

An apparatus for predicting treatment response or survival time in cancerous pathology, the apparatus comprising:

• a processor;
• a memory configured to store a set of digitized images of a region of interest (ROI) demonstrating cancerous pathology, where a member of the set of digitized images has a plurality of pixels, a pixel having an intensity;
• an input/output (I/O) interface;
• a set of circuits comprising an image acquisition circuit, a radiomic feature expression scene circuit, a superpixel clustering circuit, an expression level mapping circuit, a radiomic spatial textural descriptor (RADISTAT) circuit, and a classification circuit;
• and
• an interface that connects the processor, the memory, the I/O interface, and the set of circuits;
• the image acquisition circuit configured to access a member of the set of digitized images of the ROI demonstrating cancerous pathology;
• the radiomic feature expression scene circuit configured:
• extract a set of radiomic features from the ROI, a radiomic feature having a value;
• and
• define a radiomic feature expression scene based on the ROI and the set of radiomic features;
• the superpixel clustering circuit configured to generate a cluster map by superpixel clustering the radiomic feature expression scene;
• the expression level mapping circuit configured to generate an expression map by repartitioning the cluster map into B expression levels, where B is an integer;
• the RADISTAT circuit configured to:
• compute a textural phenotype for the expression map based on the B expression levels;
• compute a spatial phenotype for the expression map based on the B expression levels;
• construct a RADISTAT descriptor by concatenating the textural phenotype with the spatial phenotype;
• and
• compute a first probability that the ROI is a responder or non-responder, or a second probability that the ROI will experience long-term survival or short-term survival, based, at least in part, on the RADISTAT descriptor;
• and
• the classification circuit configured to generate a classification of the ROI as a responder or non-responder, or as long-term survivor or short-term survivor based, at least in part, on the first probability or the second probability, respectively.

A non-transitory computer-readable storage device storing computer-executable instructions that when executed by a computer control the computer to perform a method comprising:

• accessing a set of digitized images of a region of tissue demonstrating cancerous pathology, where a member of the set of digitized images has a plurality of pixels, a pixel having an intensity;
• defining an input radiomic scene I based on a member of the set of digitized images, where the input radiomic scene I includes a spatial grid of pixels, where a pixel is associated with a radiomic feature value;
• generating a cluster map Î by applying superpixel clustering to the input radiomic scene I;
• generating an expression map Ĩ by repartitioning cluster map Î into B expression levels, where B is an integer;
• computing a textural phenotype based on the proportion of each of B expression levels in the expression map Ĩ;
• computing a spatial phenotype based on the adjacency of the B expression levels represented in the expression map Ĩ;
• generating a RADISTAT descriptor by concatenating the textural phenotype with the spatial phenotype;
• and
• generating a classification of the region of tissue as a responder or non-responder, or as a long-term survivor or short-term survivor based, at least in part, on the RADISTAT descriptor.