A 2-Marker IHC Panel of Nestin and INPP4B for Detection of Basal-Like Breast Cancer Defined by Gene Expression
Jennifer R Choo, Dongxia Gao, George Chao, Christine Chow, Samuel Aparicio, Charles M Perou, Torsten O Nielsen. Genetic Pathology Evaluation Centre, Vancouver, Canada; University of British Columbia, Vancouver, Canada; University of North Carolina, NC
Background: Basal-like breast cancer, originally defined by gene expression profiling, is an aggressive subtype unresponsive to available targeted therapies and characterized by poor prognosis. Surrogate immunohistochemical definitions for basal-like breast cancer rely on lack of ER, PR and HER2, and can be improved with additional markers such as cytokeratin 5 and EGFR. Many additional biomarkers have since been proposed, but rarely validated against a gene expression profile gold standard. We recently performed a large survey to assess the most sensitive and specific biomarkers, including nestin (a stem cell marker) and INPP4B (a negative regulator of phosphatidyl inositol signaling). Building on this study, an optimized immunopanel for basal-like breast cancer was determined.
Design: Whenever possible, the scoring system used for each proposed biomarker was taken from the original literature associating that biomarker with basal-like breast cancer. Only cases with complete data were included in preliminary model building for development of an optimized immunopanel. A kappa statistic was used to assess the level of agreement between gold standard and surrogate immunopanels. The strength of agreement was interpreted for kappa values as follows: <0.00 = poor, 0.00-0.20 = slight, 0.21-0.40 = fair, 0.41-0.60 = moderate, 0.61-0.80 = substantial and 0.81-1.00 = almost perfect.
Results: INPP4B possessed the best combination of sensitivity (61%) and specificity (99%) using a 5% cutoff predefined in the literature. With an odds ratio of 108.4, it was the single best biomarker for basal-like breast cancer among those surveyed. The combination of negative INPP4B expression and/or positive nestin expression provided the most sensitive (83%) and specific (98%) panel for detection of basal-like breast cancer against a PAM50 gene expression profile gold standard (kappa = 0.83).
Conclusions: Loss of INPP4B expression is the best single biomarker for basal-like breast carcinoma. The combination of negative INPP4B expression (at 5% cutoff) and/or positive nestin expression (at 1% cutoff) provides the most sensitive and specific 2-marker panel for detection of basal-like breast cancer. Validation on a large independent series is pending.
Monday, March 19, 2012 1:00 PM
Poster Session II # 65, Monday Afternoon