Predicting breast cancer responsiveness to hormone treatment using quantitative textural analysis

1. A method of predicting responsiveness of breast tumors to hormone therapy, the method comprising:

performing a quantitative textual analysis (QTA) based on a breast scan for a patient;

determining using the QTA based on the breast scan for the patient, a particular value for a QTA-based parameter;

estimating an Oncotype DX Assay Recurrence Score (RS) value for the patient using a biomarker signature for use in identifying subsequent breast tumors having a high risk Oncotype DX Assay RS, and the particular value for the QTA based parameter,

wherein the biomarker signature was derived through linear modeling of a first population of prior images having low risk RS scores and a second population of prior images having high risk RS scores, the biomarker signature expressed in the form

log(RS)= Mx+B , where:

M is a coefficient;

x is the QTA-based parameter; and

B is a constant; and

predicting responsiveness to hormone therapy for the patient based on the estimated RS value.

2. The method of claim 1 , wherein the biomarker signature is derived using a spatial scale filter (SSF) value of 0.4, and x corresponds to a standard deviation (SD).

3. The method of claim 2 , wherein:

M has a value in the range of −0.0001 to −0.001; and

B has a value in the range of 1 to 2.

4. The method of claim 3 , wherein:

M is about −0.0004177; and

B is about 1.500248.

5. The method of claim 1 , wherein the biomarker signature is further derived using progesterone receptor (PR) status where PR−=0 and PR+=1, and the biomarker signature expressed in the form

log(RS)= N (PR)+ Mx+B

where N is a coefficient.

6. The method of claim 5 , wherein the biomarker signature is derived using a spatial scale filter (SSF) value of 1, and x corresponds to the QTA-based parameter [Skewness-Diff].

7. The method of claim 6 , wherein:

N is in the range of −0.1 to −1;

M is in the range of 0.01 to 0.1; and

B is in the range of 1 to 2.

8. The method of claim 7 , wherein:

N is about −0.3759396;

M is about 0.0785932; and

B is about 1.592273.

9. The method of claim 1 , wherein:

the first and second populations have only PR+ status;

the signature is derived using a spatial scale filter (SSF) value of 0.8; and

x corresponds to [Skewness-Diff].

10. The method of claim 9 , wherein:

M has a value in the range of 0.1 to 1; and

B has a value in the range of 1 to 4.

11. The method of claim 9 , wherein:

M is about 0.3033226; and

B is about 2.83791.

12. A method of predicting responsiveness of breast tumors to hormone therapy, the method comprising:

performing a quantitative textual analysis (QTA) based on a breast scan for a patient;

determining using the QTA based on the breast scan for the patient, a particular value for a QTA-based parameter;

estimating a recurrence score (RS) value for the patient using a biomarker signature for use in identifying breast tumors having a probability p of exhibiting a high risk RS value, the signature derived from high risk and low risk legacy tumor data using logistic modeling and expressed in the form

p =ex/[1+ex] where x=Ay+B , and:

A is a coefficient;

Y is a QTA based parameter; and

B is a constant; and

predicting responsiveness to hormone therapy for the patient based on the estimated RS value.

13. The method of claim 12 , wherein:

the PR status of the legacy tumor data is not considered;

Y corresponds to the QTA parameter [Entropy-Diff];

the logistic model employed an SSF value of 0; and

high risk RS>30.

14. The method of claim 13 , wherein:

A is about −6; and

B is about −2.

15. The method of claim 12 , wherein:

a PR status of the legacy tumor data is not considered;

Y corresponds to a QTA parameter [Mean-Total];

the logistic model employed an SSF value of 0.6; and

high risk RS>25.

16. The method of claim 15 , wherein:

A is about −140; and

B is about −0.3.

17. The method of claim 12 , wherein:

the legacy tumor data have only PR+ status;

the biomarker signature is derived using a spatial scale filter (SSF) value of 0;

a high risk RS>25; and

Y corresponds to [SD-Diff].

18. The method of claim 17 , wherein:

A is about −0.01; and

B is about −0.3.

19. A method of predicting responsiveness of breast tumors to hormone therapy, the method comprising:

performing a quantitative textual analysis (QTA) based on a breast scan for a patient;

determining using the QTA based on the breast scan for the patient, a particular value for a QTA-based parameter;

estimating a recurrence score (RS) value for the patient using a biomarker signature for use in identifying breast tumors having a probability log it(p) of exhibiting a high risk RS value, the signature derived from high risk and low risk legacy tumor data using logistic modeling and expressed in the faun

log it( p )= C (PR)+ Ay+B where:

PR−=0 and PR+=1;

C is a first coefficient;

A is a second coefficient;

y is a QTA-based parameter; and

B is a constant; and

predicting responsiveness to hormone therapy for the patient based on the estimated RS value.

20. The method of claim 19 , wherein:

RS>30; SSF=0;

C is about 3;

A is about −0.01;

y is [SD-Diff]; and

B is about 0.4.

21. The method of claim 19 , wherein:

RS>30; SSF=0;

C is about 3;

A is about −0.01;

y is [SD-Diff]; and

B is about 0.4.