ER and PR Discordances: Comparison of IHC by Local and Central Laboratory and RT-PCR by Central Laboratory in ECOG Breast Cancer Study 2197
FL Baehner, R Gray, B Childs, T Maddala, S Rowley, N Davidson, S Shak, LJ Goldstein, GW Sledge, JA Sparano, S Badve. University of California, San Francisco, San Francisco, CA; Eastern Cooperative Oncology Group, Indianapolis, IA; Sanofi Aventis, Bridgewater, NJ; GHI, Redwood City, CA
Background: Accurate laboratory assessment of hormone receptors (HR) in breast carcinoma is therapeutically important. To characterize the nature of the discordances between laboratories and methods, we examined discordances for ER and PR between local IHC, central IHC and central RT-PCR using clinical trial breast cancers from E2197.
Design: Tumors from 761 pts enrolled in E2197 were examined; pts had 0-3 positive nodes and received doxorubicin and cyclophosphamide or docetaxel plus hormonal therapy. Blinded ER and PR results were obtained by local IHC (reported by the site), by central IHC (using 1D5 and 636 performed in duplicate on 1.0 mm core TMAs using pre-defined Allred score >2 cutoff); and by central quantitative RT-PCR analysis using Oncotype DX (RNA extracted from formalin fixed paraffin embedded tissue, using pre-defined cutoffs of 6.5 for ER and 5.5 for PR).
Results: Results from local IHC (761 pts) were compared with central IHC (755 pts) and RT-PCR (761 pts). The concordance for determination of hormone receptor status between local or central IHC and central RT-PCR was high and similar to the high concordance between local and central lab IHC (table). Five of 403 (2%) ER+ pts by central IHC and 21 of 429 (5%) ER+ pts by local IHC were ER- by RT-PCR; of the 25 total discordant cases between IHC and RT-PCR, 1 case was positive by both central and local IHC and the remaining 24 cases were positive by only local or only central but not by both. Fifty of 352 (14%) ER- pts by central IHC and 44 of 332 (13%) ER- pts by local IHC were ER+ by RT-PCR; of the 72 total discordant cases between IHC and RT-PCR, 22 cases were negative by both central and local IHC and 50 were negative by only local or only central but not by both.
Conclusions: These results indicate that many of the discordances between IHC and quantitative RT-PCR do not reflect pre-analytic or analytic differences between protein and RNA measurements but may be attributed to inter-laboratory variability in IHC assessment.
|Local IHC vs Central IHC||90 (88, 92)||84 (82, 87)||90 (88, 92)|
|Local IHC vs RT-PCR||91 (89, 93)||88 (85, 90)||91 (89, 93)|
|Central IHC vs RT-PCR||93 (91, 94)||90 (88, 92)||93 (91, 95)|