Cold tumours: deserts and exclusion
Three immune weathers: inflamed (T cells inside), excluded (stuck at the edge), desert (none). Most common cancers are cold. Radiation, viruses, STING agonists and vessel-opening drugs try to warm them; engagers and CAR-T bring their own T cells.
Diagram
Pick a product above a diagram to see the nodes it hits and the escape routes below the block. Hover or tap any node or arrow for what it is; every node opens its target, glossary entry or the pathway page. Violet boxes are druggable targets.
Three kinds of town: one where the police already patrol the streets (inflamed), one where they mill about outside a wall (excluded), and one with no police station at all (desert). Removing the officers' handcuffs (PD-1 blockade) only helps in the first; the second needs a gate, the third needs recruitment.
A smoke detector wired to the fire brigade. DNA in the cytoplasm is smoke; cGAS is the detector; STING is the alarm bell; interferon is the 999 call that brings the immune system. Many tumours have quietly removed the batteries.
A town planner who first refuses all new building (tumour suppressor) and then, corrupted, builds walls and moats around the tumour that keep the police out (immune exclusion).
What happens
In plain words, then the glossary entries the stage rests on. Chapter 6, Escaping the immune system: Every tumour that exists has already beaten the immune system once.
Three immune weathers: inflamed (T cells inside), excluded (stuck at the edge), desert (none). Most common cancers are cold. Radiation, viruses, STING agonists and vessel-opening drugs try to warm them; engagers and CAR-T bring their own T cells.
Cold tumours: immune deserts and exclusion. Tumours come in three immune weathers: inflamed (T cells inside, checkpoint drugs work), excluded (T cells stuck at the edge), and desert (no T cells at all). Most common cancers are excluded or desert, and turning them 'hot' is the central problem of immunotherapy.
cGAS-STING innate sensing. cGAS-STING is the cell's alarm for DNA in the wrong place. Radiation, chemotherapy, and ADCs spill DNA into the cytoplasm; cGAS detects it, STING sounds the alarm, and interferon calls in the immune system.
TGF-β signalling. A signal that stops normal cells from dividing but, once a cancer is established, switches sides: it builds scar-like stroma, walls out immune cells, and pushes cells into a migratory state.
Wnt / β-catenin. Wnt/β-catenin is a developmental pathway hijacked by colorectal cancer. Normally a destruction complex keeps β-catenin low; losing APC lets it flood the nucleus and drive growth genes.
The molecular players
The proteins and genes at this stage, with their role and how many products act on each. Listed players come from the atlas; drawn players sit as nodes in the diagrams above.
The signal tumours use to grow their own blood supply. Blocking it starves tumours and, surprisingly, helps immunotherapy work.
EZH2 is an enzyme that silences genes. The first drug against it treated a rare sarcoma and some lymphomas until it was withdrawn in 2026 for causing second blood cancers.
FAP (fibroblast activation protein) sits on the cancer-associated fibroblasts that scaffold more than 90% of epithelial cancers and is almost absent from normal adult tissue. FAPI PET tracers therefore light up tumours with high contrast, including pancreatic, gastric and low-grade cancers where FDG PET is weak, and FAP-targeted radioligands are in development.
PD-1 is a brake on T cells. Blocking it releases the immune system against the tumour and has cured some previously incurable cancers.
PIK3CA is the most commonly mutated gene in hormone-driven breast cancer. Drugs against it work, but hitting it cleanly without raising blood sugar took years.
The receptor that macrophages depend on; blocking it shrinks tenosynovial giant cell tumour (a CSF1-driven tumour) and depletes tumour-supporting macrophages, though the latter has not yet helped patients with common cancers.
Where medicines act
Products grouped by the node they hit, most advanced first, with the cancers an approved product is linked to. Pick one above the diagram to see it light up.
- AnlotinibApprovedNon-small-cell lung cancerSmall-cell lung cancerSarcomas (soft tissue, bone, GIST)
- AxitinibApprovedRenal cell carcinomaClear cell renal cell carcinomaAdenoid cystic carcinoma
- BevacizumabApprovedColorectal cancerOvarian cancerNon-small-cell lung cancer
- Bevacizumab (glioblastoma use)ApprovedGlioma & glioblastoma
- CabozantinibApprovedHepatocellular carcinomaAdvanced hepatocellular carcinoma (BCLC C)Renal cell carcinoma
- Camrelizumab + rivoceranibApprovedHepatocellular carcinoma
- DonafenibApprovedHepatocellular carcinomaThyroid cancer
- FruquintinibApprovedColorectal cancer
- +21 more at VEGF / VEGFR →
- MevrometostatPhase 3
- XNW5004Phase 2
- TazemetostatWithdrawn
- EB-MF-CAR-NK-01Phase 2
- FAP-2286 (177Lu / 68Ga)Phase 2
- CadonilimabApprovedCervical cancerGastric & gastro-oesophageal junction cancer
- CamrelizumabApprovedOesophageal cancerHepatocellular carcinomaNon-small-cell lung cancer
- CemiplimabApprovedNon-small-cell lung cancerMelanomaPD-L1-high non-small-cell lung cancer without a driver mutation
- DostarlimabApprovedMismatch repair deficient (MSI-high) pancreatic ductal adenocarcinomaEndometrial cancerColorectal cancer
- NivolumabApprovedMelanomaNon-small-cell lung cancerRenal cell carcinoma
- PembrolizumabApprovedMismatch repair deficient (MSI-high) pancreatic ductal adenocarcinomaTriple-negative breast cancer (TNBC)Non-small-cell lung cancer
- PenpulimabApprovedNasopharyngeal carcinomaHodgkin lymphomaRecurrent and metastatic nasopharyngeal carcinoma
- PucotenlimabApprovedMelanomaColorectal cancer
- +20 more at PD-1 →
- AlpelisibApprovedHR-positive / HER2-negative breast cancerHR-positive metastatic breast cancer after CDK4/6 inhibitors
- CapivasertibApprovedHR-positive / HER2-negative breast cancerProstate cancerMetastatic hormone-sensitive prostate cancer
- DuvelisibApprovedChronic lymphocytic leukaemiaPeripheral T-cell lymphomas (including cutaneous T-cell lymphoma)
- EverolimusApprovedHR-positive / HER2-negative breast cancerRenal cell carcinomaNeuroendocrine tumours
- GedatolisibApprovedHR-positive / HER2-negative breast cancerHR-positive metastatic breast cancer after CDK4/6 inhibitors
- IdelalisibApprovedChronic lymphocytic leukaemiaFollicular lymphomaRelapsed or refractory chronic lymphocytic leukaemia
- InavolisibApprovedHR-positive / HER2-negative breast cancerHR-positive metastatic breast cancer after CDK4/6 inhibitors
- therascreen companion diagnostic kits (KRAS, EGFR, PIK3CA, FGFR, BRAF)ApprovedColorectal cancerNon-small-cell lung cancerHR-positive / HER2-negative breast cancer
- +5 more at PIK3CA / PI3K-alpha →
How tumours escape
Records tied to this stage that describe resistance, evasion or tolerance. The resistance atlas lists the routes class by class.
Measured by
Biomarkers, tests and assays in the corpus that read this stage in a patient.
Open questions
What is not known at this stage: the atlas's own questions, the bottlenecks it bears on, and the ideas in the corpus that try to answer them.
- What actually holds T cells at the tumour border: fibroblasts, vessels, or chemokines?
- Why did systemic STING agonists and TGF-β traps disappoint?
- early clinicalindustryClear the suppressive neutrophils out of pancreatic tumours first
Pancreatic tumours are packed with a type of white blood cell that shuts down the immune attack. Blocking the signal that recruits them may open the tumour to immunotherapy.
- early clinicalFAP theranostics as a pan-cancer stromal strategy
Instead of finding a different target for each cancer, hit the scaffolding cells that almost all solid tumours share.
- early clinicalresearchGroup trials by broken mechanism, not by organ or single mutation
Rare cancers often share a broken cellular machine even when they arise in different organs. Grouping patients by that shared fault makes trials possible.
- early clinicalIs aneuploidy itself a druggable vulnerability?
Most cancers have the wrong number of chromosomes; normal cells do not. If that difference creates a specific weakness, a drug against it would spare normal tissue by definition.
- early clinicalMaking microsatellite-stable colorectal cancer immunotherapy-responsive
Ninety-five percent of bowel cancers ignore immunotherapy. Combinations that heat the tumour up (targeted drugs, radiation, new checkpoints) are the main hope.
- early clinicalNeoadjuvant immunotherapy with surgical window for glioblastoma
Give immunotherapy before surgery rather than after, so the tumour is still present to teach the immune system, then look inside it to learn what happened.
- early clinicalindustryReprogramme suppressive macrophages instead of trying to delete them
Tumours fill with immune cells that protect them. Earlier drugs tried to remove those cells and failed. Newer ones aim to switch them to the attacking side.
- early clinicalresearchTurn one tumour into a vaccine to treat all the others
Injecting immune-activating agents into a single tumour, plus a small dose of radiation, can teach the immune system to attack tumours elsewhere in the body.
- early clinicalindustryTurn the map of immune cells inside a tumour into a standardised test
Whether immune cells are next to cancer cells matters more than how many there are. Turning that spatial picture into a reliable, standardised test would predict response better.
- early clinicalresearchUnmask hidden antigens with a short epigenetic course before immunotherapy
Low doses of drugs that change how DNA is packaged can make cancer cells display more of what marks them as abnormal, potentially waking up immunotherapy in cold tumours.
17 more ideas are linked to this stage's pathways, targets and terms; see the rankings →
Key evidence
Papers in the corpus tied to this stage's pathways, targets and terms, newest first.
- 2025rctHARMONi-2: ivonescimab, a PD-1 x VEGF bispecific, beats pembrolizumab head-to-head in PD-L1-positive lung cancerThe Lancet
- 2023rctCOMMANDS: luspatercept versus epoetin alfa as first treatment for anaemia in lower-risk MDS needing transfusionsThe Lancetchanged practice
- 2021rctCLEAR: lenvatinib plus pembrolizumab versus sunitinib as first treatment for advanced kidney cancerNew England Journal of Medicinechanged practice
- 2020rctIMbrave150: atezolizumab plus bevacizumab replaces sorafenib as first treatment for advanced liver cancerNew England Journal of Medicinechanged practice
- 2019rctPAOLA-1: olaparib added to bevacizumab maintenance in newly diagnosed ovarian cancer, with benefit confined to HRD-positive tumoursNew England Journal of Medicinechanged practice
- 2018reviewBinnewies 2018: understanding the tumour immune microenvironment for effective therapyNature Medicine
- 2018rctKEYNOTE-189: pembrolizumab plus chemotherapy as first treatment for non-squamous lung cancer without a driver mutationNew England Journal of Medicinechanged practice
- 2018observationalTGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cellsNature
- 2018basicTGFβ drives immune evasion in genetically reconstituted colon cancer metastasisNature
- 2017reviewElements of cancer immunity and the cancer-immune set pointNature
- 2015observationalMelanoma-intrinsic β-catenin signalling prevents anti-tumour immunityNature
- 2014reviewLamouille 2014: molecular mechanisms of epithelial-mesenchymal transitionNature Reviews Molecular Cell Biology
src/data/mechanics-atlas.ts). Players, medicines, escape routes, tests, ideas and papers are resolved from the knowledge graph at build time through the stage's pathways, targets and terms, so every item here has its own page and sources. Where a section is missing, the corpus has no record tied to the stage yet. Nothing here is medical advice; see about and methodology. Stage 6.3 of 56.