Subtyping of triple-negative breast cancers: its prognostication and implications in diagnosis of breast origin

•Utilizing TNBC IHC surrogate and spatial immune patterns may be a practical means of representing gene profiling subtypes.•Surrogate TNBC subtypes showed distinct features and prognoses.•Marker expression used in the diagnosis of breast origin of metastatic cancers varied among different TNBC subtypes. BackgroundTriple-negative breast cancer (TNBC) subtyping by gene profiling has provided valuable clinical information. Here, we aimed to evaluate the relevance of TNBC subtyping using immunohistochemistry (IHC), which could be a more clinically practical approach, for prognostication and applications in patient management.MethodsA total of 123 TNBC cases were classified using androgen receptor (AR), CD8, Forkhead box C1 protein (FOXC1), and doublecortin-like kinase 1 (DCLK1) into luminal androgen receptor (LAR), basal-like immunosuppressive (BLIS), mesenchymal-like (MES), and immunomodulatory (IM) subtypes. The IM cases were further divided into the IM-excluded and IM-inflamed categories by CD8 spatial distribution. Their clinicopathological and biomarker profiles and prognoses were evaluated.ResultsLAR (28.6%) and MES (11.2%) were the most and least frequent subtypes. The IHC-TNBC subtypes demonstrated distinct clinicopathological features and biomarker profiles, corresponding to the reported features in gene profiling studies. IM-inflamed subtype had the best outcome, while BLIS had a significantly poorer survival. Differential breast-specific marker expressions were found. Trichorhinophalangeal syndrome type 1 (TRPS1) was more sensitive for IM-inflamed and BLIS, GATA-binding protein 3 (GATA3) for IM-excluded and MES, and gross cystic disease fluid protein 15 (GCDFP15) for LAR subtypes.ConclusionsOur findings demonstrated the feasibility of IHC surrogates to stratify TNBC subtypes with distinct features and prognoses. The IM subtype can be refined by its CD8 spatial pattern. Breast-specific marker expression varied among the subtypes. Marker selection should be tailored accordingly. Triple-negative breast cancer (TNBC) subtyping by gene profiling has provided valuable clinical information. Here, we aimed to evaluate the relevance of TNBC subtyping using immunohistochemistry (IHC), which could be a more clinically practical approach, for prognostication and applications in patient management. A total of 123 TNBC cases were classified using androgen receptor (AR), CD8, Forkhead box C1 protein (FOXC1), and doublecortin-like kinase 1 (DCLK1) into luminal androgen receptor (LAR), basal-like immunosuppressive (BLIS), mesenchymal-like (MES), and immunomodulatory (IM) subtypes. The IM cases were further divided into the IM-excluded and IM-inflamed categories by CD8 spatial distribution. Their clinicopathological and biomarker profiles and prognoses were evaluated. LAR (28.6%) and MES (11.2%) were the most and least frequent subtypes. The IHC-TNBC subtypes demonstrated distinct clinicopathological features and biomarker profiles, corresponding to the reported features in gene profiling studies. IM-inflamed subtype had the best outcome, while BLIS had a significantly poorer survival. Differential breast-specific marker expressions were found. Trichorhinophalangeal syndrome type 1 (TRPS1) was more sensitive for IM-inflamed and BLIS, GATA-binding protein 3 (GATA3) for IM-excluded and MES, and gross cystic disease fluid protein 15 (GCDFP15) for LAR subtypes. Our findings demonstrated the feasibility of IHC surrogates to stratify TNBC subtypes with distinct features and prognoses. The IM subtype can be refined by its CD8 spatial pattern. Breast-specific marker expression varied among the subtypes. Marker selection should be tailored accordingly.