176PD - Prediction of late breast cancer recurrence by the ROR (PAM50) score in postmenopausal women in the TRANSATAC cohort
|Date||01 October 2012|
|Event||ESMO Congress 2012|
|Session||Basic Science and Translational Research II|
|Topics|| Breast Cancer, Early Stage
J. Cuzick1, I. Sestak1, S. Ferree2, J.W. Cowens3, M. Dowsett4
Adjuvant endocrine therapy beyond 5 years is known to be of benefit to some oestrogen receptor (ER) positive breast cancer patients. We have shown previously that the IHC4 score and the ROR score are significantly correlated with distant recurrence in the cohort from the TransATAC study. Here we assess the relationship between the ROR score, RS (Oncotype Dx) and the IHC4 score in predicting distant recurrence beyond 5 years.Methods
We determined the univariate and multivariate prognostic value of ROR score, RS, and IHC4 score for distant recurrence separately in years 0-5 and 5-10 of follow up for all patients (N = 940), and for node-negative (N = 683) and node-positive (N = 257) patients separately using proportional hazard models for time to distant recurrence.Results
In years 0-5, all three scores add significant prognostic information on distant recurrence in a univariate model (ROR: ΔLR-?2 = 41.50, P < 0.0001; RS: ΔLR-?2 = 23.23, P < 0.0001; IHC4: ΔLR-?2 = 36.18, P < 0.0001). When the Clinical Treatment Score (CTS) was added to the model, IHC4 added the most prognostic information in years 0-5 (ΔLR-?2 = 25.22, P < 0.0001). In years 5-10, ROR provided the most prognostic information in a univariate analysis although all scores were still significant in this time period (ROR: ΔLR-?2 = 46.31, P < 0.0001; RS: ΔLR-?2 = 10.95, P = 0.0009; IHC4: ΔLR-?2 = 12.42, P = 0.0004). In the multivariate model, ROR added the most prognostic information both overall (ΔLR-?2 = 16.66, P < 0.0001) and in node-negative patients (ΔLR-?2 = 8.82, P = 0.003). For all patients, smaller contributions were seen for RS (ΔLR-?2 = 5.42, P = 0.02) and IHC4 (ΔLR-?2 = 7.12, P = 0.008) and neither added significant information in node-negative patients. For node-positive patients, ROR added most information in years 5-10 in both the univariate and multivariate model.Conclusion
Overall, all three scores provided significant additional prognostic information in the 5-10 year period, but ROR added most prognostic information in both the univariate and multivariate analysis and in node-negative patients. These results may be used to select patients who could benefit from hormonal therapy beyond 5 years.Disclosure
J. Cuzick: Dr. Cuzick disclosed that he received grant support from and is a consultant for AstraZeneca.
S. Ferree: Dr. Ferree disclosed that he is an employee of and shareholder in NanoString Technologies.
J.W. Cowens: Dr. Cowens disclosed that he is an employees of and shareholder in NanoString Technologies.
M. Dowsett: Dr. Dowsett disclosed that he received grant support from and is on the speaker's bureau, has received grant support and provided legal advice for AstraZeneca and has sponsored IHC4 in a NICE assessment.
All other authors have declared no conflicts of interest.