System and method for tomography-based radiomic mass analysis

A method for classifying an indeterminate lung nodule in a subject as malignant or benign by performing a quantitative analysis on computed tomography (CT) imaging data of the subject's lung, the method including using a processor to perform steps comprising:

• identifying a region of interest in the indeterminate lung nodule in the CT imaging data;
• segmenting the CT imaging data to delimit at least a portion of the indeterminate lung nodule into image segments;
• extracting a set of radiomic parameters indicative of the indeterminate lung nodule, wherein the set of radiomic parameters include:
• one or more nodule location parameter that includes an offset coordinate location associated with a centroid of the indeterminate lung nodule within the CT imaging data, wherein the offset coordinate location is identified relative to an origin in the CT imaging data;
• one or more nodule texture parameter,
• two nodule shape parameters, and
• four nodule surface characteristic parameters;
• generating a report classifying the indeterminate lung nodule as being malignant or benign based on the set of radiomic parameters of the indeterminate lung nodule.

A system comprising:

• a memory having stored therein computed tomography images of a subject;
• a processor configured to access the memory and carry out steps comprising:
• (a) identify a region of interest in the CT images that includes a tissue mass;
• (b) segment the images to delimit at least a portion of the region of interest into segments;
• (c) tag at least a portion of the image segments based on at least one initial radiomic parameter value;
• (d) determine a metric indicative of a nodule texture distribution or nodule surface distribution in the region of interest by comparing the tagged image segments having the at least one initial radiomic parameter value to a candidate image segment using a similarity metric, and perform multi-dimensional scaling (MDS) to project pairwise similarities between the tagged image segments and the candidate image segment to generate the metric indicative of the nodule texture distribution or the nodule surface distribution in the region of interest;
• (e) extract a set of radiomic parameters indicative of the tissue mass, wherein the set of radiomic parameters include one or more of tissue mass location, tissue mass shape, tissue mass surface characteristic, or the metric indicative of the nodule texture distribution or nodule surface distribution;
• and
• (f) generate a report classifying the indeterminate lung nodule as being malignant or benign based on the set of radiomic parameters.