Publication in Med

How tertiary lymphoid structures predict immunotherapy response — and how to induce them in cold tumors

Explicyte collaborated with: Institut Bergonié·University of Bordeaux
Tertiary lymphoid structures and cancer immunotherapy: From bench to bedside
Bone and soft tissueBreastCentral nervous systemEndocrineGastrointestinalGenitourinaryGynecologicHead and neckSkinThoracicBiomarker analysisBiomarker discoveryDiscoveryTLS scoringTrialsFFPE tissueBioinformaticsMultiplex IF/IHC

Immune-checkpoint blockade transforms outcomes for some patients and fails others, and tumor mutational burden alone doesn't explain the gap. Tertiary lymphoid structures (TLSs) — organized immune aggregates that assemble inside tumors — have emerged among the strongest predictors of who responds. This review maps how TLSs are defined, detected, and graded across cancer types; why their maturity and location govern prognosis and immunotherapy response; and the preclinical strategies now being tested to build TLSs in tumors that lack them.

Published in Med (Cell Press), this comprehensive review was led by Florent Peyraud — an Institut Bergonié oncologist who completed his PhD at Explicyte under Prof. Antoine Italiano — with senior authorship from Alban Bessede and Antoine Italiano. Explicyte’s translational team co-authored the work and contributed original digital-pathology and multiplex-immunofluorescence imaging of TLSs at different maturation stages (Figures 1 and 2), drawing on the standardized mature-TLS scoring methodology the group developed and validated across thousands of tumor samples.

The question

Why do tertiary lymphoid structures predict response to immune-checkpoint blockade — and can they be therapeutically induced in tumors that lack them?

Key steps

  1. 1

    Define and grade TLS maturity

    The review organizes TLSs into three stages — early/immature aggregates, primary follicle-like, and fully mature secondary follicle-like structures with germinal centers. It anchors this on a standardized pathology algorithm validated on 357 samples (211 carcinomas, 146 sarcomas), which defines a TLS as an aggregate of more than 50 admixed B and T cells on HES-stained slides and grades maturity by a visible germinal center on HES or a CD23+ follicular dendritic cell network by IHC.

  2. 2

    Map mature-TLS prevalence across tumor types

    Using HES plus CD20 and CD23 IHC on serial sections, Explicyte screened 1,295 tumor samples spanning 33 histotypes and found mature TLSs in 35.7% (427/1,295), with the lowest prevalence in soft-tissue sarcoma and the highest in colorectal, head and neck, and pancreatic tumors (Figure 2). Applied in routine pathology, the same approach identified mature TLSs in 33.5% of 1,383 cases — 36.2% of carcinomas, 24.8% of sarcomas, and 25.0% of neuroendocrine tumors.

  3. 3

    Establish TLS as a checkpoint-response biomarker

    Across NSCLC, STS, HNSCC, RCC, urothelial, and colorectal cancers treated with anti-PD1/PD-L1, mature TLSs — but not immature ones — correlated with higher response rates and longer PFS and OS, independently of PD-L1 expression and CD8+ T-cell infiltration. In the TLS-selected PEMBROSARC sarcoma cohort, pembrolizumab produced a 30% objective response rate and 40% six-month non-progression, versus 2.4% and 4.9% in an unselected cohort.

  4. 4

    Automate TLS detection with deep learning

    The review details AI tools that read TLSs directly from H&E/HES whole-slide images. One detection model was trained on 2,959 manually annotated TLSs from 212 sarcoma patients, generalized to other cancer types, and correlated TLS presence with improved overall survival; a pan-cancer classifier reached 90% sensitivity at 68% specificity, pointing toward standardized, tissue-sparing screening.

  5. 5

    Induce TLS to overcome resistance

    The therapeutic section synthesizes preclinical strategies to build or mature TLSs — CXCL13, CCL21, and LIGHT delivery, STING/CD40/TLR9 agonists, radiotherapy, chemotherapy, and antiangiogenic-plus-PD-L1 combinations. Early clinical translation is underway: a histology-agnostic phase 2 trial of avelumab plus regorafenib in mature-TLS-positive tumors showed durable responses even in typically immunotherapy-resistant settings such as MSS colorectal, biliary tract, and thyroid cancers, with 11 TLS-guided trials tracked in the review.

Impact

For drug developers, TLSs offer a biomarker that predicts immunotherapy benefit where TMB and PD-L1 fall short — and a therapeutic target in their own right.

35.7%
of 1,295 tumors across 33 histotypes harbored mature TLSs (Explicyte screening)
30% vs 2.4%
ORR in TLS-selected vs unselected advanced sarcoma (PEMBROSARC phase 2)
11
ongoing TLS-guided immunotherapy trials cataloged in the review

Mature TLSs predict checkpoint-inhibitor benefit independently of PD-L1 and CD8+ T-cell infiltration, making them a candidate stratifier for trials in tumors where existing biomarkers underperform. For combination-regimen developers, TLS-inducing agents — from CXCL13 and LIGHT to STING and TLR9 agonists — offer a route to convert cold tumors into responsive ones. Both hinge on reliable, standardized TLS detection and scoring, which is where validated grading algorithms and digital pathology become decisive.

Building TLS status into an immunotherapy biomarker strategy or trial-selection design? Let's talk detection and scoring.

Talk to us
Explicyte Oncology CRO logo

Capabilities

Modalities