Tumour genome sequencing

Prepared 5 September 2026 by Kya at a friend's request, for a friend and her partner. It explains what sequencing the tumour's DNA actually tells the team, how that is used in treatment decisions after surgery, and what it does and does not buy in chordoma specifically. Articles are listed at the end with PubMed identifiers; the Nature Communications papers are free to read. No prognosis figures. Nothing here changes the plan of biopsy, surgery and proton therapy; sequencing sits alongside it.

What the test is

Whole genome sequencing reads the entire DNA of the tumour and compares it with the DNA in a blood sample. The comparison shows what has gone wrong in the tumour cells and nowhere else: single-letter mutations, genes that have been deleted or duplicated, chromosomes that have broken and rejoined, the overall count of mutations, and patterns of damage that point to a cause. The blood sample also reveals any inherited change she carries. In England this is done through the NHS Genomic Medicine Service, and sarcomas, including chordoma, are on the national test directory. It needs fresh frozen tumour and blood; formalin-fixed tissue cannot be used for the full test.

What chordoma genomes look like

Chordoma has been sequenced in enough patients to know its shape:

• Brachyury. Nearly every chordoma depends on a gene called T, which makes the brachyury protein, the marker used to diagnose it. About a quarter of tumours carry extra copies of the gene, and most patients carry a common inherited variant of it (Pillay 2012; Tarpey 2017). This is the tumour's engine, and it is the target of the vaccine and drug research described in the cutting-edge document, none of which has yet produced a treatment.
• Growth signalling. A minority carry changes in the PI3K pathway, which has drugs in trials for other cancers (Tarpey 2017).
• Chromatin genes. Changes in genes that control how DNA is packaged (PBRM1, SETD2, and in the poorly differentiated subtype loss of SMARCB1, also called INI-1) are recurrent (Tarpey 2017; Bai 2021). SMARCB1 loss defines the poorly differentiated subtype, first described around 2010 and recognised as its own entity in the WHO classification since 2020, with roughly a hundred cases in the literature, mostly children and mostly skull base. It is the one finding that changes treatment today, because it opens the EZH2 inhibitor tazemetostat and a chemotherapy conversation.

[Correction 11 Oct 2026: tazemetostat was withdrawn from all markets in March 2026 (Ipsen, after SYMPHONY-1 showed second blood cancers in 5.7% vs 0); trials and compassionate access stopped. PMID 42599162; research 12 A3. Read anything above or below about it as history, not an option.] - Deletions. Loss of the CDKN2A gene, the p16 brake on cell division, is common, and is the reason CDK4/6 inhibitors have been trialled. - Low mutation count. Chordomas carry few mutations overall, which is why immunotherapy that relies on a mutation-heavy tumour has done little in them. - Skull base specifically. The 2021 whole genome study of skull base chordomas found genomic changes associated with the tumour coming back after treatment and a mutational pattern specific to chordoma (Bai 2021). Those markers are used to decide how closely to follow up and which trials fit, not to choose surgery or protons.

How the result is used after surgery

Honestly, in four ways, in descending order of how often they matter.

1. Subtype certainty. The stains (brachyury, INI-1) give the answer in days; sequencing confirms it and adds the rest of the picture. If the tumour has lost SMARCB1, the plan changes: that is poorly differentiated chordoma, chemotherapy and tazemetostat enter the conversation, and the national MDT will know the trial options. If it has not, which is far more likely at her age and with a tumour of this size, the plan stays surgery then protons.
2. A map for later. If the tumour ever returns after surgery and protons, the sequencing report is what the oncologist reads first. A CDKN2A deletion points to CDK4/6 inhibitors; a PDGFR change points to imatinib; a PI3K change points to trial arms; a match to any arm of DETERMINE, the UK precision-medicine platform with a Southampton site, or MegaMOST in Lyon, gets her into a trial that would otherwise be closed. Without the report, those doors do not exist. With it, they are open but not guaranteed: for conventional chordoma the drugs so far have given stability in some patients rather than shrinkage, and none is licensed.
3. Tumour-agnostic treatments. A handful of drugs are approved for any cancer with a particular genomic feature, regardless of where it started: NTRK fusions, and very high mutation counts or mismatch repair loss for pembrolizumab. Chordoma rarely has these, but the test checks, and finding one would change the picture immediately.
4. Inherited findings. The blood comparison occasionally shows an inherited cancer predisposition. In chordoma the common inherited variant in T is not actionable and does not run in families in a way that needs testing; rare findings would be discussed with clinical genetics.

What sequencing does not do: it does not change the operation, the proton dose or the timing of either. It is a map for the branch of the road that everyone hopes she never takes, plus certainty about the subtype now.

What it buys, precisely

• Now: confirmation of subtype beyond the stains, and the one change of plan that matters if SMARCB1 is lost.
• Later, if needed: eligibility for trials and off-label targeted drugs that are matched to the tumour's own alterations, rather than to the word "chordoma".
• Always: a research-grade record of her tumour that the national MDT, the proton centre and the Chordoma Foundation's biobank can use, and that future trials will ask for.
• Not: a licensed drug for conventional chordoma, or any change to what happens in theatre and in the proton gantry.

DETERMINE in detail

Cancer Research UK's precision-medicine platform for rare cancers, modelled on the Dutch DRUP study. It gives licensed targeted drugs to patients whose tumour carries the matching alteration but whose cancer type was never in the drug's licence, so entry is defined by the tumour's genetics rather than its name. Arms as of September 2026: alectinib (ALK), atezolizumab (very high mutation count or mismatch repair loss), entrectinib (ROS1), trastuzumab with pertuzumab (HER2), vemurafenib with cobimetinib (BRAF), capmatinib (MET), and dabrafenib with trametinib (BRAF); each is a separately registered study (numbers in the cutting-edge document). More than twenty UK sites including Southampton, Bristol, London and Manchester. Route: the treating oncologist contacts a site investigator with the molecular report; a molecular tumour board decides within five working days; a fresh biopsy is mandatory at entry; it is for advanced disease that has grown despite standard treatment, not for a new diagnosis. Honest fit for chordoma: a conventional chordoma rarely carries these particular alterations, so DETERMINE is a possibility rather than a likelihood; MET changes appeared as incidental findings in the palbociclib chordoma trial, which makes that the one arm with a plausible link. The CDKN2A deletion chordomas often carry is not a DETERMINE arm; the European basket MegaMOST in Lyon has one. To be clear, MegaMOST is a drug trial for disease that has returned after standard treatment, not a place to be treated: her surgery and proton therapy should happen in the UK at the unit the national MDT recommends, and a UK sarcoma centre could refer across the border later only if the sequencing and the situation ever made that arm relevant. DETERMINE is the reason the sequencing has to be done properly: no actionable alteration on an NHS report means no referral; with one, the route is short and the drugs are free.

Reading the report when it comes

The report lists the alterations found, marks which are "actionable" with a tier, and notes germline findings. Ask that it is discussed at a molecular tumour board, that a sarcoma medical oncologist explains it to the patient and her partner, and that a copy goes to the national MDT and the proton centre. Ask the two questions that matter: is there anything actionable now, and is there anything that would change what we do if it came back.

Articles

• Tarpey PS et al. The driver landscape of sporadic chordoma. Nature Communications 2017. PubMed 29026114. Free to read. The definitive map of what goes wrong in chordoma DNA: brachyury duplication, PI3K, chromatin genes.
• Bai J et al. Whole genome sequencing of skull-base chordoma reveals genomic alterations associated with recurrence and chordoma-specific mutational signatures. Nature Communications 2021. PubMed 33536423. Free to read. The skull base study; the source of the "markers associated with recurrence" idea.
• Pillay N et al. A common single-nucleotide variant in T is strongly associated with chordoma. Nature Genetics 2012. PubMed 23064415. The inherited brachyury variant most patients carry.
• Gounder MM et al. Tazemetostat, an EZH2 inhibitor, in solid tumors harboring SWI/SNF alterations: a phase II basket study. Nature Communications 2026. PubMed 41882006. Free to read. The trial that makes SMARCB1 loss actionable, with modest results in poorly differentiated chordoma.
• Sosinsky A et al. Insights for precision oncology from the integration of genomic and clinical data of 13,880 tumours from the 100,000 Genomes Cancer Programme. Nature Medicine 2024. PMC10803271. Why whole genome sequencing is now routine in the NHS and what it finds across cancers.
• NHS England, National Genomic Test Directory for cancer (sarcoma whole genome sequencing): https://www.england.nhs.uk/genomics/
• Cancer Research UK, DETERMINE precision-medicine trial: patient and clinician pages linked from the cutting-edge document; Southampton is a site.
• Chordoma Foundation, drug and trial finder and biobank: https://www.chordomafoundation.org/