Synthesis: Lehmann/Pietenpol subtypes of TNBC

Evidence grades (GRADE-adapted): A high — multiple well-conducted RCTs or systematic reviews converge. B moderate — single pivotal RCT or consistent observational evidence. C limited — single observational study, mechanistic, or expert consensus. D preclinical / hypothesis-generating.

Why TNBC needed sub-typing

Triple-negative breast cancer (TNBC) is defined by absence rather than by any unifying positive feature: ER−, PR−, HER2−. Clinically grouping tumors this way obscures the biological heterogeneity that drives heterogeneous treatment response — the same chemotherapy regimen can produce pathologic complete response in one tumor and rapid progression in another, with no current biomarker reliably predicting which is which. The molecular-subtyping research program of the past 15 years has tried to resolve this heterogeneity into discrete, biologically coherent, and (ideally) therapeutically actionable groups. The Lehmann/Pietenpol framework, first published in 2011 and refined in 2016, is the most widely adopted of these efforts, and its language (“BL1,” “LAR,” “mesenchymal”) is now embedded in the trial design and the literature.

Pre-history: intrinsic subtypes and the basal-like category

The conceptual basis for molecular sub-typing of breast cancer was laid by Perou and Sørlie's 2000 paper in Nature, which used hierarchical clustering of cDNA microarray expression data from 65 breast tumors to define five intrinsic subtypes: luminal A, luminal B, HER2-enriched, basal-like, and normal-like^[1]A. The basal-like subtype showed expression patterns reminiscent of basal/myoepithelial cells of the normal breast (cytokeratins 5/6, EGFR, vimentin) and was independently linked to poor prognosis a year later in a follow-up cohort by the same group^[2]A. The intrinsic-subtype framework was later operationalized as the PAM50 classifier, a 50-gene predictor compatible with FFPE tissue^[3]A.

Approximately 80% of TNBC tumors cluster as basal-like by PAM50, but the overlap is imperfect: a meaningful minority of TNBC tumors are classified as HER2-enriched, luminal, or normal-like, and a meaningful minority of basal-like tumors are not triple-negative^[1]. Equating TNBC with basal-like was therefore recognized early as a useful approximation but not a substitute for tumor-level molecular characterization. The Lehmann/Pietenpol work was an attempt to fill that gap specifically for the TNBC clinical population.

Lehmann 2011: six TNBC subtypes (TNBCtype)

Lehmann, Bauer, Pietenpol and colleagues at Vanderbilt analyzed gene-expression data from 587 TNBC tumors aggregated across 21 publicly deposited datasets, applied k-means clustering on the top differentially expressed genes after removing redundant probes, and identified six stable clusters^[4]A:

• BL1 (basal-like 1) — cell-cycle and DNA-damage-response pathway enrichment; high MKI67, MYC, and homologous recombination machinery expression.
• BL2 (basal-like 2) — growth-factor signaling (EGF, MET, IGF1R, NGF) and glycolytic/gluconeogenesis enrichment; described as expressing basal/myoepithelial differentiation markers.
• IM (immunomodulatory) — high immune-cell signature genes (cytokine signaling, NK-cell, T-cell pathways, immune transcription factors).
• M (mesenchymal) — EMT, motility, ECM-receptor interaction enrichment; growth-factor pathways shared with BL2 but with EMT predominance.
• MSL (mesenchymal stem-like) — low cell-proliferation gene expression alongside stem-cell and mesenchymal signatures; angiogenesis pathway enrichment.
• LAR (luminal androgen receptor) — AR signaling enrichment, with downstream hormone-regulated genes (DHCR24, ALCAM, FKBP5, PIP); also enriched for ERBB family and PI3K pathway signaling.

The paper also profiled 25 TNBC cell lines and assigned each to one of the six subtypes, enabling subtype-specific preclinical drug-sensitivity screening — an unusually high-utility companion resource that drove much of the framework's adoption. BL1 cell lines were preferentially sensitive to cisplatin (consistent with the DDR-pathway enrichment), MSL and BL2 lines showed sensitivity to dasatinib and NVP-BEZ235 (PI3K/mTOR), and LAR lines responded to the AR antagonist bicalutamide^[4].

A companion paper by the Vanderbilt group applied the TNBCtype classifier to neoadjuvant chemotherapy response data from 130 TNBC patients in MD Anderson and other cohorts, reporting that BL1 tumors achieved the highest pCR rate (~52%) and BL2 and LAR the lowest (~0–10%)^[5]B. This was the framework's first explicit clinical-utility claim, and it remains influential despite caveats (single-cohort, retrospective, no adjustment for stage or grade).

Burstein 2015: an independent four-subtype alternative

Burstein and colleagues at Baylor took a different methodological route: integrating gene-expression data from 198 TNBC tumors with DNA methylation, copy-number, and exome-sequencing data, and clustering on the integrated feature set rather than on expression alone. The resulting framework comprises four subtypes^[6]B:

• BLIA (basal-like immune-activated) — STAT signal-transduction genes, immune-effector signatures; reported best prognosis.
• BLIS (basal-like immunosuppressed) — SOX-family transcription factors, immune-down-regulated; reported worst prognosis.
• MES (mesenchymal) — broadly overlaps with Lehmann's M and MSL.
• LAR (luminal androgen receptor) — broadly overlaps with Lehmann's LAR; AR signaling and ERBB family enrichment.

The two frameworks agree about LAR as a discrete entity and they agree about a broadly mesenchymal axis, but they disagree about how to split the basal-predominant subgroup: Lehmann splits BL1 and BL2 by DNA-damage-response vs. growth-factor enrichment; Burstein splits BLIA and BLIS by the immune-active vs. immune-suppressed dichotomy. The disagreement matters because the prognostic implications differ — Burstein's framework predicts better outcomes for the immune-rich subset, anticipating the immunotherapy-era observation that high tumor-infiltrating-lymphocyte (TIL) content predicts response^[7]A.

Lehmann 2016: refinement to four subtypes (TNBCtype-4)

The Lehmann group revisited their 2011 framework after recognizing that the IM and MSL subtypes might be artefacts of tumor sample composition rather than tumor-cell-intrinsic biology. Using laser-capture microdissection to isolate tumor epithelium from infiltrating immune and stromal cells in 84 TNBC tumors, they confirmed that the IM signature derived predominantly from tumor-infiltrating lymphocytes and the MSL signature predominantly from stromal cells; neither remained a distinct cluster when only tumor-cell-intrinsic expression was analyzed^[8]B. They proposed a refined four-subtype classifier — BL1, BL2, M, LAR — with IM and MSL re-cast as a separate microenvironment-derived axis layered on top of the tumor-intrinsic subtype.

The 2016 refinement also reported improved prognostic discrimination of pCR after neoadjuvant chemotherapy: BL1 maintained the highest pCR rates (~41%), LAR the lowest (~21%), with BL2 and M intermediate^[8]. This four-subtype framework (often called “TNBCtype-4”) is the version currently distributed via the public TNBCtype web tool hosted by Vanderbilt.

Reconciliation and subsequent integrative work

Several groups have attempted to reconcile the Lehmann and Burstein frameworks against larger, multi-omic datasets:

• Bareche 2018 — multi-omic analysis (expression, methylation, mutation, copy-number) of 550 TNBC tumors from the METABRIC and TCGA cohorts. Identified four molecular clusters that broadly mapped to Lehmann's BL1/BL2/M/LAR with TIL-derived signatures separable as a distinct immune axis — consistent with Lehmann's 2016 refinement^[9]B.
• Jiang 2019 (FUSCC TNBC cohort) — multi-omic analysis of 465 primary TNBC tumors from Fudan University Shanghai Cancer Center, the largest single-institution cohort. Proposed a four-subtype refinement (LAR, IM, BLIS, MES) and explicitly mapped each subtype to candidate targeted therapies (LAR → AR antagonists, IM → immune checkpoint inhibitors, BLIS → cell-cycle inhibitors and DNA-damaging agents, MES → PI3K/mTOR and EMT-targeted)^[10]B. The FUSCC framework is the basis for the FUTURE-series of biomarker-stratified TNBC trials in China.
• Liu 2016 — long non-coding RNA-based clustering of 165 TNBC tumors identified four subtypes (IM, LAR, MES, BLIS) that broadly aligned with the integrative-omics frameworks above, suggesting the subtype axes are robust to the assay modality^[11]C.

Across these efforts the recurring conclusion is that LAR is the most stable and reproducible subtype across cohorts and methods, mesenchymal/basal axes are reproducible but their boundaries are method-dependent, and the immune-active vs. immune-suppressed split (Burstein's BLIA/BLIS) is increasingly recognized as a tumor-microenvironment axis rather than a tumor-intrinsic one — Lehmann's 2016 view.

Subtype-by-pCR association across cohorts

Numbers harmonized across studies that used different neoadjuvant regimens (mostly anthracycline+taxane); direct comparison is approximate. The IM/BLIA row in particular is sensitive to whether immune signature is treated as a subtype or a separate axis.

Clinical applicability today

Routine clinical use of TNBC molecular subtyping is still limited. The frameworks are used to stratify enrollment in investigational trials and to interpret retrospective subgroup analyses, but no major society guideline currently requires TNBC subtype assignment to drive standard-of-care treatment selection. Three subtype-driven clinical-use cases have advanced furthest:

• LAR → AR antagonists. The strongest subtype-actionable lead. The TBCRC 011 phase II trial of bicalutamide in AR+ metastatic TNBC reported a clinical benefit rate of ~19% in the AR-positive subgroup, defining proof-of-concept^[12]C. Enzalutamide, the second-generation AR antagonist, was tested in AR-expressing advanced TNBC (Traina 2018, n=118) with a 16-week clinical benefit rate of 25% in patients with ≥ 10% AR by IHC^[13]B. TBCRC 032 combined enzalutamide with the PI3K inhibitor taselisib in AR+ metastatic TNBC, reporting a clinical benefit rate of 36% and identifying PIK3CA-mutation enrichment as a sub-response signal^[14]B. Definitive randomized data are still pending.
• Immune-rich subtype → immune checkpoint inhibitors. KEYNOTE-522, KEYNOTE-355, and IMpassion130 / IMpassion131 all preceded routine subtype assignment; subgroup analyses by Lehmann/Burstein subtype were retrospective and underpowered. Tumor-infiltrating lymphocytes per the International TILs Working Group standard remain the more practical proxy for the immune-active axis^[15]A.
• FUTURE-series trials at FUSCC (China) have prospectively biomarker-stratified pre-treated metastatic TNBC patients to subtype-matched targeted therapy with reported response rates substantially higher than historical controls^[16]B. Replication in non-Chinese cohorts is pending.

Subtype-assignment workflows currently rely on RNA-seq or gene-expression microarrays of FFPE-suitable quality. The TNBCtype web tool (Vanderbilt) returns calls from a 101-sample-trained classifier; targeted-panel approximations exist but are not yet a regulatory device. Inter-classifier agreement (Lehmann TNBCtype-4 vs. Burstein vs. FUSCC) on the same patient cohort runs in the 60–75% range — reasonable for research but below the level demanded for treatment selection^[9].

Limitations and methodological caveats

• Sample-composition contamination. The original Lehmann 2011 classifier was derived from bulk-tumor expression data; IM and MSL signals turned out to reflect microenvironment composition rather than tumor-intrinsic biology^[8]. Newer single-cell and spatial transcriptomic methods promise cleaner separation but have not yet replaced bulk-derived classifiers in routine use.
• Cohort bias. The 2011 derivation pooled 21 datasets generated on different platforms (Affymetrix variants, Agilent, custom arrays) with varying preanalytical handling. Cross-platform normalization is imperfect; cluster stability depends on the chosen normalization method^[9].
• Intratumoral heterogeneity. Single-cell RNA-seq has revealed substantial subclonal diversity within individual TNBC tumors, including coexistence of subtype-discordant subpopulations. A bulk-derived subtype assignment may represent the dominant clone but miss therapeutically important minority populations — relevant for relapse and metastasis biology^[17]B.
• Subtype-by-treatment interactions are mostly retrospective. Prospective subtype-stratified clinical trials are scarce; the FUTURE series at FUSCC and a handful of phase II investigator-initiated efforts are the main exceptions. Most subtype-treatment claims rest on subgroup analyses of trials that did not pre-specify subtype as a stratification variable.
• Ancestry and population representation. The 2011 derivation cohorts were predominantly drawn from US and European patient populations. The FUSCC cohort is the largest non-Western validation, and its proposed framework (LAR/IM/BLIS/MES) diverges in proportions from Western cohorts — LAR is more prevalent (~23% vs ~15%), IM is less (~24% vs ~30%). Whether the underlying biology truly differs or whether ascertainment and treatment patterns confound the comparison is an active question^[10].

Open questions and active investigation

• Will single-cell / spatial methods produce a new canonical classifier? Methods that resolve tumor cells from microenvironment should sharpen subtype boundaries; whether the resulting classifier converges on Lehmann/Burstein/FUSCC or supersedes them is an open question.
• Definitive AR-antagonist trial. Despite over a decade of LAR proof-of-concept evidence, no phase III trial has confirmed AR antagonism as standard of care for LAR-subtype TNBC. The 2026 readouts from ongoing combination trials (AR antagonist + CDK4/6 inhibitor; AR antagonist + PARP inhibitor in AR+/HRD+ tumors) may resolve this.
• Subtype-stratified neoadjuvant trial design. Whether KEYNOTE-522's IO + chemo benefit is uniform across subtypes or concentrated in immune-rich tumors remains undetermined. Retrospective Lehmann-stratified analyses of KEYNOTE-522 data would substantially inform de-escalation strategies.
• Multi-omic integration with treatment-response data at scale. The Jiang 2019 FUSCC framework's strength is the explicit subtype-to-therapy mapping; replicating that mapping in larger multi-cohort efforts (e.g., AURORA-US) would substantially raise the clinical decision-relevance of subtype assignment.
• HER2-low and subtype interaction. Whether HER2-low status (now actionable for trastuzumab deruxtecan per DESTINY-Breast04) varies systematically by Lehmann/Burstein subtype, and whether the trastuzumab-deruxtecan benefit varies by subtype within HER2-low TNBC, are immediately relevant biomarker-stratification questions for metastatic care.

For the molecular-definition view of TNBC and the standard-of-care landscape that subtype work feeds into, see the overview synthesis. For the diagnostic boundary between TNBC and HER2-low and the assay heterogeneity behind it, see IHC for ER/PR/HER2 and the TNBC definition. For the patient-layer companion to this subtypes discussion, see What is TNBC?

References

Each citation links to the original publication via DOI. The same records are searchable in the evidence library by title or DOI.

1. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature. 2000;406(6797):747–752. doi:10.1038/35021093. ↩
2. Sørlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001;98(19):10869–10874. doi:10.1073/pnas.191367098. ↩
3. Parker JS, Mullins M, Cheang MCU, et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol. 2009;27(8):1160–1167. doi:10.1200/JCO.2008.18.1370. ↩
4. Lehmann BD, Bauer JA, Chen X, et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J Clin Invest. 2011;121(7):2750–2767. doi:10.1172/JCI45014. ↩
5. Masuda H, Baggerly KA, Wang Y, et al. Differential response to neoadjuvant chemotherapy among 7 triple-negative breast cancer molecular subtypes. Clin Cancer Res. 2013;19(19):5533–5540. doi:10.1158/1078-0432.CCR-13-0799. ↩
6. Burstein MD, Tsimelzon A, Poage GM, et al. Comprehensive genomic analysis identifies novel subtypes and targets of triple-negative breast cancer. Clin Cancer Res. 2015;21(7):1688–1698. doi:10.1158/1078-0432.CCR-14-0432. ↩
7. Denkert C, von Minckwitz G, Darb-Esfahani S, et al. Tumour-infiltrating lymphocytes and prognosis in different subtypes of breast cancer: a pooled analysis of 3771 patients treated with neoadjuvant therapy. Lancet Oncol. 2018;19(1):40–50. doi:10.1016/S1470-2045(17)30904-X. ↩
8. Lehmann BD, Jovanović B, Chen X, et al. Refinement of triple-negative breast cancer molecular subtypes: implications for neoadjuvant chemotherapy selection. PLOS One. 2016;11(6):e0157368. doi:10.1371/journal.pone.0157368. ↩
9. Bareche Y, Venet D, Ignatiadis M, et al. Unravelling triple-negative breast cancer molecular heterogeneity using an integrative multi-omic analysis. Ann Oncol. 2018;29(4):895–902. doi:10.1093/annonc/mdy024. ↩
10. Jiang YZ, Ma D, Suo C, et al. Genomic and transcriptomic landscape of triple-negative breast cancers: subtypes and treatment strategies. Cancer Cell. 2019;35(3):428–440.e5. doi:10.1016/j.ccell.2019.02.001. ↩
11. Liu YR, Jiang YZ, Xu XE, et al. Comprehensive transcriptome analysis identifies novel molecular subtypes and subtype-specific RNAs of triple-negative breast cancer. Breast Cancer Res. 2016;18(1):33. doi:10.1186/s13058-016-0690-8. ↩
12. Gucalp A, Tolaney S, Isakoff SJ, et al. Phase II trial of bicalutamide in patients with androgen receptor-positive, estrogen receptor-negative metastatic breast cancer (TBCRC 011). Clin Cancer Res. 2013;19(19):5505–5512. doi:10.1158/1078-0432.CCR-12-3327. ↩
13. Traina TA, Miller K, Yardley DA, et al. Enzalutamide for the treatment of androgen receptor-expressing triple-negative breast cancer. J Clin Oncol. 2018;36(9):884–890. doi:10.1200/JCO.2016.71.3495. ↩
14. Lehmann BD, Abramson VG, Sanders ME, et al. TBCRC 032 IB/II Multicenter Study: molecular insights to AR antagonist and PI3K inhibitor efficacy in patients with AR+ metastatic triple-negative breast cancer. Clin Cancer Res. 2020;26(9):2111–2123. doi:10.1158/1078-0432.CCR-19-2170. ↩
15. Salgado R, Denkert C, Demaria S, et al. The evaluation of tumor-infiltrating lymphocytes (TILs) in breast cancer: recommendations by an International TILs Working Group 2014. Ann Oncol. 2015;26(2):259–271. doi:10.1093/annonc/mdu450. ↩
16. Jiang YZ, Liu Y, Xiao Y, et al. Molecular subtyping and genomic profiling expand precision medicine in refractory metastatic triple-negative breast cancer: the FUTURE trial. Cell Research. 2021;31(2):178–186. doi:10.1038/s41422-020-0375-9. ↩
17. Karaayvaz M, Cristea S, Gillespie SM, et al. Unravelling subclonal heterogeneity and aggressive disease states in TNBC through single-cell RNA-seq. Nature Communications. 2018;9(1):3588. doi:10.1038/s41467-018-06052-0. ↩

Last reviewed: 2026-05-31. Researcher-layer synthesis page. Evidence grades follow the GRADE-adapted rubric defined at the top of this page. Citations are anchored to the full bibliographic entries above; click the ↩ arrow next to any reference to return to its first citation in the prose.