Does Oncotype DX recurrence score correlate with survival benefit in “good prognosis” patients as estimated by AdjuvantOnline!?
38 Background: An important question for ER positive, node negative breast cancer is the benefit of chemotherapy over hormonal therapy. We sought to explore if a low estimate of chemotherapy benefit, based on AdjuvantOnline, correlated with low risk Oncotype DX score , identifying patients (pts) for whom the “hormone therapy alone” treatment decision would not change. Methods: Our institutional database was reviewed to identify pts with ER positive, HER2 negative, node negative breast cancer submitted for oncotype DX testing. Clinicopathologic data and Oncotype Recurrence Score (RS) were extracted and entered into ‘Adjuvantonline’ to obtain estimated survival benefit gained at 10 years (yrs) (SBG@10). Two datasets were obtained; survival benefit gained at 10 yrs from chemotherapy only (SBG@10+C) and net survival benefit gained at 10 yrs: the absolute difference between combined therapy and hormone therapy (SBG@10 net). Pts were distributed into subgroups SBG@10 ≤3% and >3%. i.e. RS ≤18 versus >18.Correlation between SBG and RS were calculated with Spearman’s Rank method. Distribution of RS were compared with Mann-Whitney and Χ2 tests. Results: We identified 77 pts. For SBG@10+C ≤3% (n=28) and >3% (n=49) subgroups, 16 (57.1%) and 23 (46.7%) had low RS, respectively (p=0.53). Three (10.7%) and 10 (20.4%) in each subgroup had high RS (p=0.48). Distribution of RS between both subgroups was not significantly different, p=0.38. Correlation between RS and SBG@10 in the ≤3% subgroup was weak (rs=0.37, p=0.05). For SBG@10 net ≤3% (n=45) and >3% (n=32) subgroups, 25 (55.6%) and 14 (43.8%) had low RS, respectively (p=0.43). Five (11.1%) and 8 (25%) in each subgroup had high RS (p=0.20). Distribution of RS between both subgroups was not significantly different, p=0.19. Correlation between RS and SGB@10 in the ≤3% subgroup was weak (rs=0.21, p=0.15). Conclusions: Although the decision for additional chemotherapy changed for ≈10% with Oncotype DX, we could not identify a cohort of good prognosis pts for whom Oncotype DX would not significantly alter treatment recommendations. There are no subsets of node negative, ER positive, breast cancer for whom Oncotype DX does not provide independently significant prognostic and predictive information.