#### Title

Utilization of the three Magee equations to select patients for genomic testing

#### Recommended Citation

Ali H, Abdel-Rahman Z, Favazza L, and Chitale D. Utilization of the three Magee equations to select patients for genomic testing. Eur J Cancer 2018; 92:S91-S92.

#### Document Type

Conference Proceeding

#### Publication Date

2018

#### Publication Title

Eur J Cancer

#### Abstract

Background: Several genomic tests are available to help in decision making for breast cancer adjuvant therapy. The 21 gene genomic assay oncotype Dx (RS) is a widely utilized assay. The assay determines prognosis and likelihood of benefit from adjuvant chemotherapy. The Magee equation derives a recurrence score using clinical and histological values (RSm). Three equations have been developed, each incorporates slightly different values to derive a recurrence score (RSm1, RSm2 and RSm3) the score follows a 0-100 scale comparable to RS. Concordance between RSm1, RSm2 and RSm3, and RS risk category is 55.8%, 59.4%, and 54.4% for RSm1, RSm2 and RSm3 respectively, and 100%, 98.6%, and 98.7% when the intermediate risk category was eliminated. The Magee equations viewed as a low cost alternative to expensive genomic testing. Materials and Methods: In our study we studied the correlation coefficient “CC” between the RS and RSm at different RS intervals we deemed clinically significant with the aim of identifying a subset of patients in whom RSm is highly predictive of RS allowing it's use in lieu of RS. 664 patients with known RS were included in our study. 544 had sufficient values for equation 2, 226 for equation 1 and 284 for equation 3. The patients were divided into three groups based on their RS. Group one had RS between 0-10, group 2 between >10 to 25 and group 3 was >25. The first and the third groups were expected to contain 97.5% of low risk and high risk patients respectively. Correlation coefficient between RS and RSm1, RSm2 and RSm3 was calculated in each group. Results: The correlation coefficient for each of the subgroups are shown in Table 1. Because RSm3 showed the highest correlation, we analyzed the patients in this cohort who fell outside their assigned RS risk category. We classified shifts as one risk category and two risk category shifts. Of the 19 patients who scored higher than 25 on RSm3, 3 had an RS of less than 25, none of the 3 had an RS less than 10, the lowest RS was 17. Of the 53 patients with RSm3 10 or less, 35 had an RS of more than 10, however 1 patient out of those had a score more than 25 (the RS was 28). Conclusion: Our studyshowsthat the correlation betweentheRSandRSm increases with higher scores. We demonstrated that 2 category shifts in RS (considered clinically significant)wereuncommon whenRSm3is used. These results suggest that RSm3 > 25, and possibly <10, may be useful selection criteria for patients in whom RS is unlikely to change the clinical decision making in a meaningful way, therefore avoiding the cost of genomic testing. (Table presented).

#### Volume

92

#### First Page

S91

#### Last Page

S92