System and method for feature extraction and classification on ultrasound tomography images
Disclosed herein systems, processors, or computer-readable media configured with instructions to: receive transmission and/or reflection images of a tissue of a subject, wherein the images are generated from acoustic signals derived from acoustic waveforms transmitted through the tissue; provide a set of prognostic parameters associated with a user selected region of interest; wherein the set of prognostic parameters comprises sound propagation metrics characterizing sound propagation within a tissue; wherein the set of prognostic parameters corresponds to inputs into a tissue classifier model; wherein the set of prognostic parameters comprises a plurality of subsets of related feature groupings; and determine a type of tissue of the subject based on said plurality of subsets of related feature groupings using the classifier model, wherein the type of tissue is a cancerous tumor, a fibroadenoma, a cyst, a nonspecific benign mass, and an unidentifiable mass.
1. A computer implemented method for characterizing a lesion in a volume of tissue, the method comprising:
receiving a plurality of acoustic renderings, the acoustic renderings comprising a representation of sound propagation through the volume of tissue, wherein the plurality of acoustic renderings comprises at least a transmission rendering and a reflection rendering;
determining a set of prognostic parameters, wherein the set of prognostic parameters comprises one or a plurality of sound propagation metrics that are derived from the plurality of acoustic renderings; and
applying a classifier model to classify the volume of tissue, wherein the classifier model is generated from:
a first classifier configured to assign each prognostic parameter of a plurality of prognostic parameters a predictive value and
generate the set of prognostic parameters from the plurality of prognostic parameters based on the predictive value of each prognostic parameter of the plurality of prognostic parameters, wherein the first classifier configured to trim the plurality of prognostic parameters based on the predictive value to generate the set of prognostic parameters; and
a second classifier configured to calculate a score relating to a probability that the lesion is of a classification based on the set of prognostic parameters, wherein said second classifier is a trained classifier trained at least in part on the set of prognostic parameters and a plurality of classified acoustic renderings.
2. The method of claim 1 , wherein the lesion comprises a cancer, a fibroadenoma, a cyst, a nonspecific benign mass, or an unidentifiable mass.
3. The method of claim 1 , wherein the plurality of acoustic renderings comprises one or more of: (i) a plurality of reflection renderings, (ii) a plurality of transmission renderings, or (iii) at least one reflection rendering and at least one transmission rendering.
4. The method of claim 1 , wherein the transmission rendering comprises a sound speed rendering or an attenuation rendering.
5. The method of claim 1 , wherein the plurality of prognostic parameters comprises sound speed metrics relating to a region of interest and (i) is based on the region of interest or (ii) comprises a user-assigned classification of the region of interest.
6. The method of claim 5 , wherein the region of interest (i) is a user selected region of interest, (ii) is partially selected using the set of prognostic parameters, (iii) comprises a portion of a lesion, or (iv) comprises a two-dimensional region of interest.
7. The method of claim 6 , wherein the two-dimensional region of interest is used to generate a three-dimensional region of interest.
8. The method of claim 5 , wherein the user-assigned classification is a mass boundary score.
9. The method of claim 1 , wherein the set of prognostic parameters comprises at least one member selected from the group consisting of: a morphological metric of the lesion, crispiness, a texture metric of the region of interest, and fuzziness.
10. The method of claim 9 , wherein the morphological metric comprises:
(i) at least one of a roundness, an irregularity of a shape, an irregularity of a margin, and a smoothness of a margin;
(ii) the fuzziness of a boundary of the lesion; or
(iii) the crispiness of a margin of the lesion.
11. The method of claim 1 , wherein the one or a plurality of sound propagation metrics characterizes one or more of: (i) sound propagation interior to a region of interest or (ii) sound propagation exterior to a region of interest.
12. The method of claim 1 , wherein the one or the plurality of sound propagation metrics comprises at least one of a mean, a standard deviation, a skewness, and a kurtosis.
13. The method of claim 1 , wherein the one or the plurality of sound propagation metrics characterizes at least one of sound speed, sound attenuation, and sound reflection.
14. The method of claim 1 , wherein the one or the plurality of sound propagation metrics comprises one or more members selected from the group consisting of a sound speed metric, a reflection metric, an attenuation metric, a morphological metric, and a user defined score.
15. The method of claim 1 , wherein the set of prognostic parameters comprises one or more of: (i) one or a plurality of sound propagation metrics characterizing sound speed, (ii) one or a plurality of sound propagation metrics characterizing sound attenuation, or (iii) one or a plurality of sound propagation metrics characterizing sound reflection.
16. The method of claim 1 , wherein the classifier model determines a type of tissue with a sensitivity at least 85% and a specificity of at least 84%.
17. The method of claim 1 , wherein the classifier model determines: (i) a threshold value of one or more prognostic parameters sufficient to classify a tissue, (ii) a relative statistical accuracy of one or more prognostic parameters, (iii) a threshold value of the set of prognostic parameters sufficient to classify a tissue, or a combination thereof.
18. The method of claim 1 , wherein the classifier determines a likelihood that the lesion is a malignant lesion.
19. The method of claim 1 , wherein the trained classifier is generated using a machine learning technique.