Tumor characterization and outcome prediction through quantitative measurements of tumor-associated vasculature

A non-transitory computer-readable medium storing computer-executable instructions that, when executed, cause a processor to perform operations, comprising:

• accessing a training set comprising a plurality of routine clinical medical imaging scans, wherein each routine clinical medical imaging scan of the training set is associated with a three-dimensional segmentation including a three-dimensionally segmented associated tumor and a three-dimensionally segmented set of tumor-associated vasculature (TAV) associated with the associated tumor, wherein the associated tumor of that routine clinical medical imaging scan is associated with at least one of a known response to a treatment or a known prognosis;
• and
• for each routine clinical medical imaging scan of the training set, extracting associated values for a set of features from the associated TAV for the associated tumor of that routine clinical medical imaging scan, the set of features including one or more of at least one TAV spatial orientation feature or a statistic of the at least one TAV spatial orientation feature and at least one TAV morphology feature or a statistic of the at least one TAV morphology feature, said extracting using one or both of at least a portion of the three-dimensional segmentation or a projection onto a plane of at least a portion of the three-dimensional segmentation, such that the set of features represent spatial orientation and morphology of at least a portion of the associated TAV in three dimensions;
• and
• training a machine learning model based on the associated values extracted from the associated TAV for the associated tumor of that routine clinical medical imaging scan and on the at least one of the known response to the treatment or the known prognosis.

An apparatus, comprising:

• memory configured to store data derived from one or more routine clinical medical imaging scans of including a tumor, wherein the tumor and tumor-associated vasculature (TAV) associated with the tumor are segmented in a three-dimensional segmentation;
• one or more processors configured to:
• extract two or more features from the TAV, the two or more features including at least one TAV spatial orientation feature or a statistic of the at least one TAV spatial orientation feature and at least one TAV morphology feature or a statistic of the at least one TAV morphology feature, the extraction using one or both of at least a portion of the three-dimensional segmentation or a projection onto a plane of at least a portion of the three-dimensional segmentation, such that the two or more features represent spatial orientation and morphology of at least a portion of the TAV in three dimensions;
• provide the ene or more features extracted from the TAV to a trained machine learning model;
• and
• receive, from the trained machine learning model, one of a predicted response to a treatment for the tumor or a prognosis for the tumor.

A method, comprising:

• accessing one or more medical imaging scans of a tumor, wherein the tumor is segmented on the one or more medical imaging scans;
• segmenting tumor-associated vasculature (TAV) associated with the tumor based on the one or more medical imaging scans;
• extracting one or more features from the TAV, the one or more features comprising one or more of at least one TAV spatial orientation feature or a statistic of the at least one TAV spatial orientation feature, wherein the at least one TAV spatial orientation feature comprises one or more of a vessel orientation along a XY projection image, a vessel orientation along a XZ projection image, a vessel orientation along a YZ projection image, a vessel orientation along a rotation-elevation projection image, a vessel orientation along a distance-rotation projection image, or a vessel orientation along a distance-elevation projection image;
• providing the one or more features extracted from the TAV to a trained machine learning model;
• and
• receiving, from the trained machine learning model, one of a predicted response to a treatment for the tumor or a prognosis for the tumor.