Estimation of Risk of Recurrence of Early Stage Estrogen Receptor Positive (ER+) Breast Carcinoma (BC) by Pathologists Compared to Oncotype Dx Recurrence Score – A Multi-Institutional Study
G Acs, NN Esposito, C Reynolds, TJ Lawton, I Bleiweiss, C Nagi, O Ioffe, O Hameed, D Dabbs, R Bhargava, MR Quddus, CJ Sung, M Gilcreast, A Sapino, G Cserni, J Kulka, J Jacquemier, P Regitnig, I Baas, P van Diest, C Quinn. Moffitt Cancer Center, Tampa, FL
Background: The Oncotype Dx assay is increasingly being used to predict recurrence in early stage BC; however, it has not been adequately compared to prediction based on traditional clinicopathologic features.
Design: We selected 153 patients with early stage ER+ BC and available Oncotype Dx recurrence score (RS). Clinicopathologic data, including age, menopausal status, tumor size, grade, mitotic activity, lymphatic invasion (LVI), nodal status, hormone receptor and HER2 status on all patients were provided to 21 breast pathologists. Participants were asked to estimate the risk of recurrence based on clinicopathologic data and provide the three most important tumor features their estimates were based on. Risk estimates of participants were compared with RS results.
Results: Based on Oncotype Dx results, 94 (61.4%), 45 (29.4%) and 14 (9.2%) BC were low (RS <18), intermediate (RS 18-30) and high (RS ≥31) risk, respectively. RS values showed significant correlation with tumor grade, mitotic activity, LVI, hormone receptor and HER2 status, while no correlation with age, menopausal status, tumor size and histologic type was found. Participants' risk estimates agreed with the Oncotype Dx assay in 52.5 ± 2.7 % (mean ± SEM, range 29.4 - 71.9) of cases, while the risk was over- and underestimated compared to RS results in 32.6 ± 3.6 (range 8.5 - 64.1) and 15.0 ± 1.4 (range 6.5 - 28.1) %, respectively. The rates of overestimation were significantly higher than those of underestimation (p = 0.0002). Assessment of the concurrence of participants with RS results showed a mean kappa value of 0.27 (range 0 - 0.53). The intraclass correlation coefficient for agreement for all participants was 0.501. Participants ranked tumor stage/nodal status, histologic grade and tumor size as the most important clinicopathologic features for estimating recurrence risk.
Conclusions: Compared to Oncotype Dx, pathologists tend to overestimate the risk of tumor recurrence based on traditional clinicopathologic features alone. While the RS may provide additional information regarding intrinsic biological features of ER+ BC, development of a nomogram to reliably predict RS from available clinicopathologic data would be useful and economical in patient management.
Tuesday, March 23, 2010 1:00 PM
Poster Session IV # 28, Tuesday Afternoon