Synthesis: Tumor microenvironment, TILs, and immune subtyping

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 TILs in TNBC

TILs — lymphocytes infiltrating the tumor bed in histologic sections — have been a recognized prognostic feature in many cancers for decades, but their measurement was poorly standardized until the 2010s. TNBC specifically shows substantial inter-tumor variation in TIL density: some tumors are densely infiltrated (immune-hot); others are nearly devoid of immune cells (immune-cold). This variation has clinical correlates: high-TIL TNBC has better prognosis than low-TIL TNBC at every stage of disease, and high-TIL TNBC has higher response rates to immune checkpoint inhibitor therapy.

The International TILs Working Group standard

Salgado and colleagues, on behalf of the International TILs Working Group 2014, published a standardized methodology for TIL assessment on H&E-stained sections^[1]A. The standard:

• Measure stromal TILs — lymphocytes within the stromal compartment between tumor nests, expressed as a percentage of stromal area. Intratumoral lymphocytes (within tumor nests) are reported separately but less reliably; stromal TILs are the primary measure.
• Score on H&E-stained sections — no special staining needed; the method works on routine clinical pathology slides.
• Assess in invasive carcinoma areas — exclude lymphocyte-rich areas associated with DCIS, necrosis, biopsy-tract scarring, or other non-tumor processes.
• Use the full tumor section, not hotspots — report the average across the invasive tumor area, not the maximum focal density.
• Report as a continuous percentage — though clinical reporting often categorizes (low < 10%, intermediate 10–50%, high ≥ 50% or ≥ 60%).

The Working Group standard has been adopted internationally and incorporated into ESMO and ASCO/CAP recommendations for TNBC pathology reporting. Inter-pathologist agreement on stromal TIL percentage, after training, is reasonable (kappa ~0.4–0.6, depending on study). AI-assisted TIL scoring (machine learning models trained on consensus annotations) is in active development and may further improve reproducibility^[2]B.

Prognostic value in early-stage TNBC

Denkert and colleagues conducted a pooled individual-patient-data analysis of 3,771 patients across six neoadjuvant chemotherapy trials, examining TIL density and outcomes by subtype^[3]A. In the TNBC subset (n=906):

• High-TIL TNBC (≥ 60% stromal TILs) had pCR rate ~50% vs ~30% for low-TIL TNBC (< 10% stromal TILs)
• 5-year disease-free survival ~85% high-TIL vs ~60% low-TIL
• 5-year overall survival ~90% high-TIL vs ~65% low-TIL

Loi and colleagues did a parallel pooled IPD analysis of 2,148 early-stage TNBC patients across 9 trials, focused on patients who did NOT receive immunotherapy^[4]A. Each 10% increase in stromal TILs was associated with a ~20% reduction in distant recurrence risk and a ~20% reduction in mortality, with effect sizes consistent across nodal status, tumor size, and grade. The Loi analysis explicitly tested TILs as a continuous variable, confirming that the prognostic relationship is dose-dependent rather than dichotomous.

The collective message: TIL density is a robust prognostic biomarker in early-stage TNBC. A patient with high-TIL TNBC has substantially better prognosis than a patient with otherwise-equivalent low-TIL TNBC. Clinical pathology reporting now routinely includes stromal TIL percentage for TNBC cases.

Predictive value for immunotherapy benefit

Whether TILs predict benefit from immune checkpoint inhibitor therapy is an active and clinically important question. Several analyses have addressed it:

• KEYNOTE-522 sub-analyses have examined pCR rate by TIL category. High-TIL patients achieved higher pCR rates in both arms (chemotherapy alone and pembrolizumab + chemotherapy), and the absolute pCR benefit from adding pembrolizumab was numerically larger in the intermediate-TIL range than at either extreme — very-high-TIL patients had high pCR with chemotherapy alone, leaving less room for IO benefit; very-low-TIL patients did poorly in both arms.
• IMpassion130 sub-analyses showed that high TIL density correlated with both PD-L1 SP142 IC positivity and atezolizumab benefit. Whether TIL adds predictive value independent of PD-L1 is unclear.
• NeoTRIPaPDL1 retrospective analyses suggested that high-TIL patients gained more from adding atezolizumab to neoadjuvant chemotherapy, even though the overall trial was negative.

The general pattern: high-TIL TNBC has higher response rates to IO + chemotherapy than low-TIL TNBC, and the relative magnitude of IO-attributable benefit is more concentrated in the intermediate-TIL range than at either extreme. Whether TIL should be used as a clinical biomarker for IO selection (analogous to PD-L1 CPS) is debated; in current practice, TILs inform prognosis but do not gate IO eligibility.

The broader TNBC tumor microenvironment

Beyond CD8+ T-cell density, the TNBC tumor microenvironment includes multiple immune cell populations with both pro-tumor and anti-tumor activity^[5]B:

• CD8+ effector T cells — the primary anti-tumor lymphocyte population; correlate with TIL counts.
• CD4+ helper T cells — support effector responses; Th1-skewed cells favor anti-tumor immunity, Th2-skewed cells less so.
• Regulatory T cells (Tregs; CD4+ FOXP3+) — immunosuppressive; high Treg density correlates with worse outcomes despite total TIL counts being high.
• Tumor-associated macrophages (TAMs) — can polarize toward anti-tumor M1 or pro-tumor M2 phenotypes; predominant M2 polarization in TNBC is common and pro-progression.
• Myeloid-derived suppressor cells (MDSCs) — immunosuppressive; elevated MDSC presence in TNBC correlates with worse IO response.
• B cells and tertiary lymphoid structures — recent evidence suggests tertiary lymphoid structures correlate with better IO response in TNBC, similar to the melanoma evidence.
• NK cells — low overall density in TNBC; HLA-class-I-loss tumors may benefit from NK-cell-directed therapy.

The composition of TNBC microenvironment (the balance of effector cells, suppressor cells, and stromal cells) is an active research area; multiparameter spatial profiling (multiplex IHC, imaging mass cytometry, CODEX) is producing rich datasets but is not yet clinically deployed.

TIL-defined immune subtypes of TNBC

The clinical TIL spectrum aligns with the BLIA (immune-activated, high TIL) and BLIS (immune-suppressed, low TIL) categories of the Burstein four-subtype framework — transcriptomic-immune-axis classification and histologic TIL measurement converge on the same biological dichotomy from different methodologies. The 2016 refinement of the Lehmann/Pietenpol framework similarly reassigned the original IM (immunomodulatory) subtype as TIL contamination of the underlying basal-like categories, recognizing that immune content is a microenvironment feature rather than a tumor-cell-intrinsic feature.

Practical implication: TILs are easier to assess in routine pathology than gene-expression-based immune signatures, so TIL scoring serves as a practical surrogate for the broader immune-active vs immune-cold dichotomy that the molecular frameworks identify.

Current clinical applicability

• Prognostic biomarker, early-stage TNBC: well-established. Stromal TIL percentage is routinely reported on TNBC pathology specimens at major centers. The prognostic information is incorporated into multidisciplinary tumor board discussions.
• Predictive biomarker for IO benefit: not currently a gating biomarker for IO eligibility. Pembrolizumab + chemo (KEYNOTE-522) is given irrespective of TIL status in eligible stage II–III TNBC patients. TIL is sometimes used as additional context when discussing prognosis and treatment expectations.
• Predictive biomarker for de-escalation: several ongoing trials test whether very-high-TIL TNBC patients can de-escalate chemotherapy without losing efficacy — the hypothesis being that the tumor microenvironment is doing some of the work that chemotherapy normally does. Results pending.
• Predictive biomarker for adjuvant decisions: not currently used. Adjuvant pembrolizumab continuation, adjuvant capecitabine, adjuvant olaparib decisions are not currently gated by TIL status.

Evidence table — key TIL studies in TNBC

Open questions and active investigation

• Chemo de-escalation in very-high-TIL TNBC. Multiple ongoing trials test whether patients with stromal TIL ≥ 50–75% can safely receive less intensive chemotherapy. The biology is plausible; the trials are accumulating events.
• TIL as gating biomarker for adjuvant capecitabine. Post-KEYNOTE-522 era, decisions about adjuvant capecitabine in residual-disease patients are not currently TIL-stratified. Whether TIL status should refine the capecitabine decision is an active question.
• AI-assisted TIL scoring. Multiple academic and commercial groups are training deep-learning models on consensus-annotated TIL data. Inter-pathologist variability is the main motivation; FDA-cleared standalone TIL-scoring AI tools are likely within 2–3 years.
• Spatial TIL features. Beyond density, the spatial organization of TILs (tumor-cell-contact, immune-excluded vs immune-inflamed patterns, tertiary lymphoid structures) appears to add predictive information. Whether these refinements should enter routine clinical reporting is being tested.
• Multi-marker immune signatures. Beyond stromal TIL percentage, multiplex IHC for CD8, FOXP3, PD-L1, TIM-3, and others produces a richer immune phenotype. Standardization of multiplex panels for clinical use is in early stages.
• Tertiary lymphoid structure quantification. Tertiary lymphoid structures in TNBC predict better IO response in retrospective analyses; quantitation in routine clinical pathology is not yet standardized.

For the molecular subtyping frameworks that capture the immune-active vs immune-cold biology, see the Burstein subtypes synthesis and the Lehmann/Pietenpol subtypes synthesis. For PD-L1 testing in metastatic disease, see the PD-L1 assays synthesis. For the neoadjuvant chemo+IO standard that has incorporated TIL biology by retrospective analysis, see the KEYNOTE-522 synthesis.

References

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

1. 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. ↩
2. Amgad M, Stovgaard ES, Balslev E, et al. Report on computational assessment of tumor infiltrating lymphocytes from the International Immuno-Oncology Biomarker Working Group. NPJ Breast Cancer. 2020;6:16. doi:10.1038/s41523-020-0154-2. ↩
3. 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. ↩
4. Loi S, Drubay D, Adams S, et al. Tumor-Infiltrating Lymphocytes and Prognosis: A Pooled Individual Patient Analysis of Early-Stage Triple-Negative Breast Cancers. J Clin Oncol. 2019;37(7):559–569. doi:10.1200/JCO.18.01010. ↩
5. Bianchini G, De Angelis C, Licata L, Gianni L. Triple-negative breast cancer: an evolving moving target. Nat Rev Clin Oncol. 2022;19(2):91–113. doi:10.1038/s41571-021-00565-2. ↩

Last reviewed: 2026-06-04. Researcher-layer synthesis page. Evidence grades follow the GRADE-adapted rubric defined at the top of this page.