The Prosigna breast cancer assay provides independent additional prognosic value compared to classic pathohistological and immunohistochemical (IHC) parameters and assists adjuvant therapy decisions. In the present study, we retrospectively evaluated correlation of Prosigna breast cancer assay intrinsic subtype (IST) with immunohistochemical (IHC) IST estimation in a clinical setting.
We included all tumors (150 tumors from 146 women) that were analysed using the Prosigna assay at the Medical University of Vienna between December 2014 and July 2018, and of which detailed IHC status (ER/PR receptor, HER-2 and Ki67) could be accessed. 3 tumors with HER-2 enriched Prosigna IST were excluded from the analysis. For our analysis Ki67 cutoff values alone were used for discriminating Luminal (Lum) A from Lum B IHC IST estimation (Lum B: ≥20%, ≥30% and ≥40% Ki67). In addition the role of the cut off value of KI67 ≥20% and ≥30% in combination with PR receptor expression and grading were evaluated, regarding clinical and genomic correlation for defining subtypes. Statistical analyses were conducted using Cohen’s kappa coefficient.
In total we analysed 150 tumors from 146 women (median age 55.5 years), of whom 60% were lymph node positive. 79.3% (n = 119) of the tumors were of Prosigna Lum B subtype and 52% (n = 78) IHC Lum B IST, when using a cutoff of ≥ 20% Ki67 for Lum B disease. We found only a fair agreement between Prosigna and IHC IST (kappa 0.326). When using KI67 ≥30% as a cutoff, a slight improvement in agreement (kappa 0.411) was noted. Using a cutoff of ≥ 40 % KI67% to define IHC luminal IST, agreement decreased again (kappa 0.227).
Agreement between Prosigna luminal IST and estimation of luminal IST based on IHC over a range of Ki67 cutoff values between 20% to 40% proves to be fair at best. Our study demonstrates, that molecular-genetic tests, such as the Prosigna assay, provide crucial information for adjuvant chemotherapy decision management in cases where Ki67 IST estimation alone is not sufficient for therapy decision. These associations should be addressed by further studies to ensure an exact and cost-effective utilisation of the assay.
Department of Pathology, Medical University of Vienna.
K.A. Tendl: Travel support: Roche, MSD, Novartis, outside the submitted work. F. Fitzal: Travel support, scientific support: Pfizer, Comesa, Novartis, Roche, AstraZeneca, Pfizer, Myriad, NanoString, Lilly und Bondimed outside the submitted work; Editor for Oncoplastic Surgery part I and II, Springer. L. Müllauer: Research support: NanoString, outside the submitted work. Z.A. Bago-Horvath: Grants: Boehringer Ingelheim; Personal fees: Novartis, Roche; Travel support: Roche, outside the submitted work. All other authors have declared no conflicts of interest.