Discordance between Oncotype Dx and Saint Gallen criteria, Adjuvant!, NCCN 2011 guidelines, and initial physician treatment recommendation.
e11014 Background: The Oncotype Dx recurrence score (RS) assay quantifies the risk of distant recurrence (rDR) and its use has increased despite the lack of prospective studies. Methods: This is a cross sectional retrospective study of consecutive patients (PTS) from our institution with histologically confirmed invasive breast cancer who underwent surgery with curative intent and in whom Oncotype was performed. The main objectives were to compare (1) the predicted rDR by RS and Adjuvant! (2) Risk allocation by RS and St Gallen Criteria, (3) chemotherapy indication according to RS results and NCCN guidelines and (4) to evaluate the impact of RS results on the recommendation of adjuvant chemotherapy (aCT). Results: Between October/2006-June/2011, 74 PTS were evaluated. Forty seven (63,5%) were EC IA and all had estrogen receptor positive; axillary lymph node involvement was seen in 19 PTS (13 micro and 6 macrometastasis). The rDR by RS was low in 50 PTS (67%), intermediate in 19 (26%) and high in 5 (7%). According to Saint Gallen, 7 (9%), 51 (69%) and 14 PTS (19%) were classified as low, intermediate and high risk, respectively. There was a statistical significant discordance between risk allocation according to RS and Saint Gallen (Kappa coefficient=-0.002; p=0.971). Among the 55 node-negative PTS, there was also a statistical significant discordance between the predicted average rDR, obtained from Oncotype, and Adjuvant! with median risk of 8,5% vs 15,7%, respectively (p = 0.00001 rank sum Mann Whitney test). The NCCN 2011 would have indicated aCT to 62 PTS. Among 55 classified as low and high risk by RS, the NCCN would have indicated aCT to 46 PTS. In other words, 89% (41) of PTS who would receive aCT by NCCN were classified as low risk by RS, with a statistically significant discordance (Kappa coefficient=0.035, p=0.328). Conclusions: Oncotype changed the medical management in 28 (55%) of 51 PTS in which the initial intention of the physician was known. Of these, 93% were spared aCT. We found no statistical significant concordance among the Saint Gallen, Adjuvant! or NCCN guidelines with Oncotype Dx. The rDR may be overestimated by clinicopathological-based classifications.