ESMO ESCAT v2 (2026): Levels of Evidence for Mutations in Tumor Genetic Reports, New Resistance Scale, and Tumor-Agnostic Therapies

There Is a Mutation in the Report, but What Does It Mean? ESMO's Updated Evidence Scale

Genetic testing of the tumour (genomic profiling) has become routine in many cancers today; however, not every mutation that appears in the report changes treatment. The European Society for Medical Oncology (ESMO) has comprehensively updated the ESCAT scale, which grades how reliable a guide a finding is for treatment, for the first time since 2018. ESCAT v2 was published on 22 October 2026 in the journal Annals of Oncology. The new version shows the degree of certainty of the evidence more clearly, introduces criteria for tumour-agnostic therapies and, for the first time, defines a separate scale for biomarkers that indicate drug resistance. We have previously covered a comparison of comprehensive genomic tests, next-generation sequencing (NGS), CeGaT tumour tests and RNA sequencing in separate articles.

Why does this matter?

Comprehensive genomic tests can find dozens, sometimes hundreds, of alterations in a tumour. However, only some of these have an approved drug or one with proven efficacy; some are meaningful only for participation in a clinical trial, and others cannot yet be linked to treatment at all. When patients and their families see a "targetable" mutation in the report, they are understandably hopeful; physicians, in turn, need to explain honestly whether this finding will really bring benefit (cancer mutations).

ESCAT was developed in 2018 to bring a common language to this question and is widely used in molecular tumour boards, laboratory reports and ESMO guidelines. In the meantime, however, tumour-agnostic approvals have multiplied, real-world evidence has grown, mutations indicating drug resistance have become part of clinical decisions, and new types of biomarker have emerged, such as protein staining (immunohistochemistry) and AI-assisted pathology. The authors state that the first version fell short of meeting these developments in five respects.

What is ESCAT v2, and how does it change the tumour genetic report?

ESCAT v2 is ESMO's updated scale that grades, from Tier I to X, how strongly a biomarker detected in the tumour supports treatment with a specific drug. The version published in October 2026 recommends that, for every finding, the answers to both "does it support treatment with this drug?" and "does it indicate resistance to this drug?" be reported together with the level of evidence.

The scale classifies only "actionable" biomarkers, that is, findings whose presence predicts benefit from a specific treatment. Each rating has three components: the biomarker, a specific drug or drug class, and a specific tumour type or clinical setting. For example, the same BRAF V600E mutation can be Tier I in lung cancer and Tier X in colorectal cancer for treatment with a BRAF drug alone. The new version covers not only DNA sequencing but every kind of validated biomarker test, such as protein staining and AI-assisted pathology.

Tier I: the same message, three levels of certainty

In the first version, Tier I was divided into three according to trial design. In the new version, the distinction is based on how certain the evidence is; yet the clinical message of the three subtiers is the same: testing should be done and, if the finding is present, the matched treatment should be chosen. This way, reports can also convey the degree of uncertainty of the evidence without weakening the treatment recommendation.

Examples of Tier I-I are the BRAF V600E mutation in lung cancer and the KRAS G12C mutation in previously treated patients (KRAS G12C targeted therapies). This category matters for rare molecular groups in which a randomised trial is ethically or practically difficult; however, the authors stress that confirmation at the I-P level should continue whenever randomisation is possible. Tier I-P also has a new clarification: if the comparator arm is already a biomarker-based standard treatment (for example, when two different EGFR drugs are compared), a trial showing that the new drug is "not worse" is also considered sufficient (EGFR-mutated lung cancer in 2026).

Tumour-agnostic therapies: the ETAC-S criteria

In recent years, some drugs have been approved on the basis of a specific molecular alteration, regardless of the organ in which the tumour originated (tumour-agnostic therapies). The first ESCAT did not define when evidence obtained in one tumour could be applied to another, and this led to differing interpretations between centres. In ESCAT v2, this decision has been left to the ETAC-S criteria published in 2024. Because the criteria are a separate document, they can be updated without changing ESCAT itself.

Tier I-E has a deliberate flexibility: whereas survival data are required in the single-tumour tiers, here a meaningful response rate and duration of response in each tumour type are considered sufficient. This is because tumour-agnostic evidence often comes from many small patient groups within a single trial. Examples are NTRK fusion and larotrectinib, and MSI-H/dMMR and immunotherapy in tumours other than colorectal cancer (dostarlimab in dMMR/MSI-H tumours). If enough data accumulate in a particular tumour type, that tumour–biomarker pair can also be moved separately to a higher or lower tier.

Tier II: "may be used", but with data recorded

Tier II is for matches whose efficacy has been shown in a specific tumour type but which have not reached the Tier I standard. A treatment in this tier can be used case by case, preferably with the approval of the molecular tumour board. The new version adds an important responsibility: the centre giving the treatment should collect the patient's outcomes through a clinical trial or a prospective registry. In other words, in Tier II the treatment recommendation and data collection go hand in hand. The tier has been divided into three subtiers:

II-C is the most uncertain group in Tier II; it is open to publication and patient selection bias. In this case, the tumour board needs to state clearly that the evidence is limited and that the decision carries greater uncertainty.

In rare situations such as the glioma in the example, structured real-world evidence is seen as the main route by which an efficacy signal can become strong evidence (molecular diagnosis in brain tumours). The criterion separating Tier II-A from I-I is also clear: if a non-randomised trial reports only a response rate, it is II-A; if progression-free or overall survival clearly exceeds the best known outcomes, it is I-I.

Tiers III, IV and X

• Tier III refers to clinical trials: Matches that make biological sense but lack sufficient clinical evidence in that tumour. Even if the tumour board supports it, this is not considered a justification for treatment outside a trial (what is a clinical trial, phase 1 trials). In III-A, the evidence comes from another tumour and ETAC-S is not met; for example, KRAS G12C and sotorasib in ovarian cancer. In III-B, there are no patient data at all, and the alteration is considered functionally similar to a known target; for example, EGFR kinase domain duplication.
• Tier IV is for research only: IV-A has data from cell or animal models, IV-B only computational predictions. The authors stress that the two are not equivalent and that laboratory data are far more valuable than computational predictions.
• Tier X has been split in two: In X-A, the biomarker was tested but the drug's benefit turned out to be independent of the biomarker. For example, sacituzumab govitecan is effective regardless of Trop-2 level; for this reason, approvals do not require a Trop-2 test (sacituzumab in triple-negative breast cancer, sacituzumab in hormone receptor-positive breast cancer). In X-B, despite a plausible target, the drug was ineffective in that disease: for instance, giving a BRAF drug alone for BRAF V600E in colorectal cancer (encorafenib).

Tier V of the first version (response present but no survival benefit, a candidate for combination) is not included in the new version because it was not used in practice.

The new resistance scale: which finding rules out a drug?

Biomarkers can show not only who will benefit but also who will not. Until now, this second dimension had not been tied to a common standard. ESCAT v2 introduces a four-tier scale for resistance biomarkers, built on the same logic as the actionability scale. The scale covers both resistance present before treatment (intrinsic) and resistance that emerges during treatment (acquired) (how cancer cells develop resistance).

Tier I-R can be reached in two ways. For I-A-R, a randomised trial or its biomarker analysis must show that benefit is absent or markedly reduced in biomarker carriers. I-B-R is for resistance that emerges during treatment: because retreatment or randomisation is not ethical here, what is required is a consistent finding in at least two independent retrospective datasets from different centres, a well-defined mechanism, and a tissue or blood (liquid biopsy) sample taken at progression. For example, in EGFR-mutated lung cancer, the T790M mutation that develops against first- and second-generation EGFR drugs also identifies patients who will benefit from osimertinib (osimertinib).

The example of Tier II-R is the STK11 (LKB1) mutation in lung cancer: in data from many centres, patients carrying this mutation have lower response rates and survival with immunotherapy alone; however, some of them still benefit (immunotherapies). For this reason, STK11 alone should not be used to deny immunotherapy. Examples of Tier III-R include RB1 loss and CDK4/6 inhibitors in hormone receptor-positive, HER2-negative breast cancer (genomic testing and CDK4/6 inhibitors), MDM2/MDM4 amplification, which has been associated with rapid disease progression on immunotherapy, and additional chemotherapy in the presence of a POLE mutation in high-risk uterine (endometrial) cancer. An example showing that the scale is dynamic is also given: NRAS mutations against BRAF drugs in melanoma were first described only at the laboratory level (IV-R); today there is clinical evidence as well.

Differences from the first version

Information on the first version is based on the 2018 paper and on the explanations in the ESCAT v2 paper.

What if there is more than one target?

Tumours often have more than one targetable alteration. ESCAT v2 grades each biomarker separately; which one to target first is decided by the tumour board. This decision should take into account the tumour type, the size of the expected benefit, the availability and licensing status of the drug, any coexisting resistance findings, prior treatments and the patient's condition together. Combinations based on lower-tier alterations, on the other hand, should be tested in clinical trials rather than in routine treatment.

Limitations and points of debate

• A consensus document: ESCAT v2 is not a new clinical trial; it is based on the consensus reached by an international expert panel over eight rounds of feedback. As the authors also note, some thresholds are taken from clinical trial conventions rather than from a formal calibration study.
• Applying it requires expertise: The distinctions between types of evidence and endpoints require clinical trial knowledge that not every laboratory has. The authors recommend that, in uncertain cases, laboratories determine the tier together with experts.
• Dependence on knowledge bases: Everyday use of the scale depends on continuously updated databases such as OncoKB and CIViC, and on reporting software.
• Tiers can change: As new data arrive, a match's tier can go up or down; the tier in an older report may no longer be valid today (reading research correctly).
• Independent of approval: This makes the scale usable worldwide; however, a drug being Tier I does not mean that it is licensed or reimbursed in every country.
• Conflicts of interest: The work was funded by ESMO; consulting or speaking relationships of most of the authors with pharmaceutical and diagnostics companies are declared in the paper.

Tips for patients and their families

• Read the report together: Seeing a mutation in your genetic test report may not mean there is an effective drug for it. You can ask your doctor: "What is the ESCAT tier of this finding, and what does it mean for my tumour?"
• Ask about resistance findings too: Some findings show that a particular drug will not work; this information protects you from unnecessary side effects and lost time.
• For Tier II and III: Ask whether the treatment can be given within a clinical trial or a registry programme.
• Retesting: If the disease progresses during treatment, a new biopsy or liquid biopsy can reveal new resistance mechanisms and new treatment options.
• At the decision stage, do not hesitate to get a second opinion from an experienced centre.

Frequently asked questions

What is the ESCAT scale?

ESCAT is the scale of the European Society for Medical Oncology (ESMO) that grades how strongly a genetic or molecular alteration found in the tumour supports treatment with a specific drug. Tier I indicates standard treatment, Tier II case-by-case use, Tier III a clinical trial, Tier IV research only, and Tier X that the match does not work.

What has changed in ESCAT v2?

In the new version published in October 2026, Tier I was divided into three according to the certainty of the evidence, the ETAC-S criteria were introduced for tumour-agnostic use, Tier II was split into three subtiers, Tier V was removed, Tier X was split in two and, for the first time, a four-tier resistance scale was added for biomarkers indicating drug resistance.

If my report shows a Tier I mutation, will the drug definitely work?

No. Tier I shows that the match is part of standard treatment based on strong evidence; however, it does not guarantee a response in every patient. In addition, ESCAT is independent of drug approvals; the drug may not be licensed or reimbursed in Turkey. The decision should be made by your doctor and the tumour board.

What is the resistance tier for?

The resistance tier shows how much a finding reduces the likelihood of benefiting from a specific drug. A finding at the I-R level means that the drug should not be given, II-R means cautious use, and III-R only a preliminary signal. For example, a RAS mutation in colorectal cancer indicates I-R level resistance to anti-EGFR antibodies.

Glossary

Biomarker

A feature that can be measured in the tumour and helps predict response to treatment, such as a mutation, a gene fusion or a protein level.

Actionable

A finding whose presence predicts benefit from a specific treatment.

Molecular tumour board

A board that assesses genetic test results together with oncologists, pathologists, geneticists and bioinformatics specialists.

Tumour-agnostic therapy

Treatment given according to a specific molecular alteration rather than the organ in which the cancer started.

ETAC-S

ESMO's criteria for assessing whether a therapy can be used in a tumour-agnostic way.

Real-world evidence

Patient data collected during everyday treatment, outside clinical trials.

Contraindication

A situation in which a treatment should not be given.

Liquid biopsy

A test performed by analysing tumour DNA circulating in the blood.