Once a woman's relative risk for breast cancer has been calculated in the Gail model (previous section), it needs to be converted to the absolute risk for the individual patient. Absolute risks are calculated for the periods of 10, 20 and 30 years from the time that the woman is counseled.
Parameters:
(1) relative risk
(2) age of patient at the time of counseling in years
(3) number of biopsies taken
(4) presence or absence of known atypical hyperplasia
Benichout et al showed different figures for each set of parameters. The equations below were based on data extracted from the figures, with the line approximations derived using JMP.
If the age of the woman is > 50, then the figures in Benichou et al do not handle the absolute risk.
10 Year Absolute Risk |
Figure |
Age |
Equation for Absolute Risk from Relative Risk |
---|---|---|---|
no biopsy |
C2a |
20 |
= (0.0671 * RR) - 0.067143 |
|
|
30 |
= (0.42607 * RR) + 0.15393 |
|
|
40 |
= (1.14786 * RR) + 0.345 |
|
|
50 |
= (1.42929 * RR) + 0.4021 |
1 biopsy, no known atypical hyperplasia |
C2b |
20 |
= (0.0427 * RR) + 0.1706 |
|
|
30 |
= (0.412 * RR) + 0.35867 |
|
|
40 |
= (1.116786 * RR) + 0.4971 |
|
|
50 |
= (1.4015 * RR) + 0.633 |
1 biopsy, atypical hyperplasia |
C2c |
20 |
= (0.064286 * RR) - 0.2619 |
|
|
30 |
= (0.44045 * RR) + 0.21418 |
|
|
40 |
= (-0.007456 * (RR^2)) + (1.2671 * RR) + 0.004 |
|
|
50 |
= (-0.008342 * (RR^2)) + (1.50793 * RR) + 0.561157 |
>= 2 biopsies, no known atypical hyperplasia |
C2d |
20 |
= (0.05263 * RR) - 0.0526 |
|
|
30 |
= (0.42615 * RR) + 0.1970303 |
|
|
40 |
= (-0.008 * (RR^2)) + (1.2731 * RR) - 0.06935 |
|
|
50 |
= (1.4465 * RR) + 0.423 |
>= 2 biopsies, atypical hyperplasia |
C2e |
20 |
= (0.05203 * RR) - 0.248448 |
|
|
30 |
= (0.4045238 * RR) + 0.4007 |
|
|
40 |
= (-0.005479 * (RR^2)) + (1.19746 * RR) + 0.195893 |
|
|
50 |
= (-0.00829 * (RR^2)) + (1.49857 * RR) + 0.4 |
where:
• RR = relative risk
20 Year Absolute Risk |
Figure |
Age |
Equation for Absolute Risk |
---|---|---|---|
no biopsy |
C2f |
20 |
= (0.51 * RR) - 0.03 |
|
|
30 |
= (1.592857 * RR) + 0.157 |
|
|
40 |
= (-0.04345 * (RR^2)) + (2.89345 * RR) - 0.125 |
|
|
50 |
= (-0.0411 * (RR^2)) + (3.23839 * RR) + 0.02232 |
1 biopsy, no known atypical hyperplasia |
C2g |
20 |
= (0.5 * RR) |
|
|
30 |
= (1.499286 * RR) + 0.52857 |
|
|
40 |
= (-0.014881 * (RR^2)) + (2.2345 * RR) + 0.4 |
|
|
50 |
= (-0.0607 * (RR^2)) + (3.46857 * RR) - 0.66 |
1 biopsy, atypical hyperplasia |
C2h |
20 |
=(0.465909 * RR) + 0.2045455 |
|
|
30 |
= (-0.00737 * (RR^2)) + (1.575857 * RR) + 0.3416 |
|
|
40 |
= ( -0.017234* (RR^2)) + (2.24285 * RR) + 0.59307 |
|
|
50 |
= (-0.040692 * (RR^2)) + (3.182526 * RR) + 0.16467 |
>= 2 biopsies, no known atypical hyperplasia |
C2i |
20 |
= (0.4639 * RR) + 0.2187879 |
|
|
30 |
= (-0.010384 * (RR^2)) + (1.627 * RR) + 0.27744 |
|
|
40 |
= (-0.01568 * (RR^2)) + (2.0325596 * RR) + 0.26261 |
|
|
50 |
= (-0.04063 * (RR^2)) + (3.160179 * RR) + 0.31 |
>= 2 biopsies, atypical hyperplasia |
C2j |
20 |
= (0.4569 * RR) + 0.18214 |
|
|
30 |
= (-0.010357 * (RR^2)) + (1.65107 * RR) - 0.308929 |
|
|
40 |
= (-0.013167 * (RR^2)) + (1.917 * RR) + 1.1982 |
|
|
50 |
= (-0.040286 * (RR^2)) + (3.16657 * RR) + 0.08 |
30 Year Absolute Risk |
Figure |
Age |
Equation for Absolute Risk |
---|---|---|---|
no biopsy |
C1a |
20 |
= (1.77143 * RR) - 0.4714 |
|
|
30 |
= (2.9357 * RR) + 0.18095 |
|
|
40 |
= (-0.08024 * (RR^2)) + (4.40784 * RR) + 0.023287 |
|
|
50 |
= (-0.08024 * (RR^2)) + (4.40784 * RR) + 0.023287 |
1 biopsy, no known atypical hyperplasia |
C1b |
20 |
= (1.5759 * RR) - 0.0357 |
|
|
30 |
= (-0.02381 * (RR^2)) + (2.6667 * RR) + 0.2857143 |
|
|
40 |
= (-0.046875 * (RR^2)) + (3.495536 * RR) + 0.25 |
|
|
50 |
= (-0.089286 * (RR^2)) + (4.52143 * RR) - 0.7 |
1 biopsy, atypical hyperplasia |
C1e |
20 |
= (-0.01354 * (RR^2)) + (1.744365 * RR) - 0.08409 |
|
|
30 |
= (-0.027697 * (RR^2)) + (2.6822 * RR) + 0.7905 |
|
|
40 |
= (-0.04248 * (RR^2)) + (3.37689 * RR) + 1.05 |
|
|
50 |
= (-0.039177 * (RR^2)) + (3.481 * RR) + 4.4797619 |
>= 2 biopsies, no known atypical hyperplasia |
C1c |
20 |
= (-0.017337* (RR^2)) + (1.8286 * RR) + 0.036 |
|
|
30 |
= (-0.0319* (RR^2)) + (2.7047 * RR) - 1.1318 |
|
|
40 |
= (-0.024 * (RR^2)) + (2.666 * RR) + 1.8 |
|
|
50 |
= (-0.02823 * (RR^2)) + (3.364 * RR) + 3.94 |
>= 2 biopsies, atypical hyperplasia |
C1f |
20 |
= (-0.010405 * (RR^2)) + (1.6617857 * RR) + 0.1553571 |
|
|
30 |
= (-0.01881 * (RR^2)) + (2.3069048 * RR) + 1.4357143 |
|
|
40 |
= (-0.024186 * (RR^2)) + (2.6597381 * RR) + 2.0817857 |
|
|
50 |
= (-0.063836* (RR^2)) + (4.1417669 * RR) + 0.6841035 |
The 95% confidence intervals for each risk prediction (10, 20 or 30 years) could be estimated, using different equations if atypical hyperplasia was known present or not.
95% Confidence Limits |
Figure |
Line |
Equation for 95% CI |
---|---|---|---|
no atypical hyperplasia |
C3 |
lower |
= (-0.002054 * (AR^2)) + (0.76375 * AR) + 0.55 |
|
|
upper |
= (1.36 * AR) - 0.4 |
atypical hyperplasia |
C3 |
lower |
= (-0.00119 * (AR^2)) + (0.6467619 * AR) - 0.357143 |
|
|
upper |
= (-0.00566 * (AR^2)) + (1.651147 * AR) + 0.172727 |
where:
• AR = absolute risk from 10, 20 or 30 year risk equations
If age was not exactly at the decade (20, 30, 40 or 50), then a value between the absolute risk for the decade above and for the decade below were used.
absolute risk when between decades =
= (absolute risk for lower decade) + ((((age) - (lower decade)) / 10) * ((absolute risk for higher decade) - (absolute risk for lower decade)))
Limitations:
• The Gail model was found by Bondy et al to perform well at predicting the risk of breast cancer in women who adhered to the American Cancer Society screening guidelines, However, women who did not adhere to the screening guidelines, it tended to overpredict the risk.
• Bondy et al also noted that the model tended to overpredict the risk for women < 60 and to underpredict the risk for women >= 60 years of age.
Purpose: To convert a woman's relative risk for breast cancer to absolute risk in the Gail model.
Specialty: Hematology Oncology, Surgery, general, Obstetrics & Gynecology
Objective: risk factors
ICD-10: C50-C50,