Present Role of Immunohistochemistry in the Classification of Molecular Subtypes Beppe Viale European Institute of Oncology University of Milan Milan-Italy
We know it is many diseases Breast cancer is a heterogeneous disease: Histological features Biological characteristics Clinical outcome Responsiveness to therapies Need for classification
Juliet: "What's in a name? That which we call a rose By any other name would smell as sweet." Romeo and Juliet (II, ii, 1-2)
The perfect classification Clinically useful Prognostic/Predictive Scientifically accurate Applicable Easy to teach, easy to learn Affordable (time and resources) Reproducible
Molecular Classification Unsupervised analysis of global gene expression patterns unveiled distinct and robust molecular subtypes of breast cancer (496 genes) Sørlie T et al PNAS 2001
Molecular Classification 4-6 subtypes Tumours with different prognosis in the same category (e.g., Basal-like) Issues of affordability
Surrogates? PAM50 3 gene signature ER, HER2, Aurora kinasi IHC ER/PgR HER2 Ki-67
St Gallen Consensus 2011 For practical purposes tumour subtype can be ascertained by non-genetic tests for ER, PgR, Her2 and Ki-67? Yes: 82% No: 12.2% A: 4.9%
Subtype Surrogate IHC markers Luminal A ER and/or PgR positive HER2 negative, Ki-67 low (<14%)* Luminal B (HER2 neg) ER and/or PgR positive HER2 negative, Ki-67 high Luminal B (HER2 pos) ER and/or PgR positive HER2 positive HER2 positive (non luminal) Type of therapy Endocrine therapy alone Cytotoxics + endocrine therapy Cytotoxics + anti-her2 + endocrine therapy Cytotoxics + anti-her2 Notes Few require cytotoxics (e.g. high nodal status). Inclusion and type of cytotoxics may depend on perceived risk and patient preference. No data are available to support the omission of cytotoxics in this group. Triple negative (ductal) Cytotoxics Consider DNA disrupting agents. Special histological types * A. Endocrine responsive B. Endocrine non responsive Endocrine therapy Cytotoxics Medullary and adenoid cystic carcinomas may not require any adjuvant cytotoxics.
Are TN and BL the same entity? Only 71-91% of TN have a BL gene expression profile Importance of the cut-off? Only 77% of BL carcinomas have a TN immunophenotype They may express ER and/or HER2 Almost 20% of non-tn have a BL gene expression profile
A triple negative immunophenotype does not overlap the basal like gene expression profile Basal-like Triple negative
BL tumors are heterogeneous IDC NOS, high-grade ILC high-grade, pleomorphic Metaplastic, high-grade Myoepithelial carcinoma High-grade (oat-cell) neuroendocrine Apocrine Medullary Adenoid-cystic Metaplastic, low-grade Low grade adenosquamous Fibromatosis-like Poor prognosis Good prognosis
Immunohistochemical surrogates of basal-like carcinomas ER-, HER2-,CK5/6+, EGFR+ 55-76% sensitivity, 100% specificity Basal Cytokeratins CK5 and/or CK14, and/or CK17 Additional Markers Vimentin c-kit P-cadherin p63 Smooth muscle actin
Caution Only 60% BL express CK5 BL do not express CK14 (!) Only 57% to 72% BL express EGFR 4% to 29% BL express ER and/or HER2 Cut-off for CKs: Any stained cell? 10% immunostained cells? Cut-off for EGFR: Any cell? 10%, 20%, 50%?
EGFR in 284 TNBC:70 months follow-up
EGFR in 284 TNBC: 70 months follow-up
St. Gallen Consensus 2011 Should we aim at identifying basal-like tumours within the TNBC category using basal CKs and/or EGFR immunostainings? No: 80% Yes: 8% Abstain: 12%
A clinically meaningful approach to TNBC Identify special types with better prognosis Adenoid cystic, Medullary, Metaplastic low-grade, Apocrine low-grade Identify truly non endocrine-responsive tumors <1% ER & PgR immunoreactive cells Do not miss candidate patients to anti-her2 interventions Equivocal IHC/FISH (CISH,SISH) Add prognostically relevant markers (?) CK 5, 14, 17 EGFR
But it is already over!
Epilogue No classification -taken alone- is perfect Enthusiasm for the novel assays Biotech pressure Clinical, morphological, immunohistochemical and molecular data should be integrated into a single classification scheme with definite prognostic/predictive value