Is TMB a good predictive biomarker for immunotherapy?

  • Aurélien Marabelle
Cancer Research ESMO Immuno-Oncology Congress 2023
Aurélien Marabelle

Aurélien Marabelle

Gustave Roussy, Villejuif


Evidence is compelling, but utility may depend on the clinical context

The potential of tumour mutational burden (TMB) as a predictive biomarker for immune checkpoint blockers was highlighted nearly a decade ago in the setting of melanoma (N Engl J Med. 2014;371:2189–2199). It took another 6 years for compelling evidence to be generated by the phase II KEYNOTE-158 trial, which showed that high levels of tissue TMB ([TMB-high] ≥10 mutations/megabase) were associated with a greater objective response rate to the anti-PD1 antibody pembrolizumab than low tissue TMB across 10 tumour types (Lancet Oncol. 2020;21:1353–1365). The KEYNOTE-158 trial led to the US FDA approval in 2020 of pembrolizumab for the treatment of adult and paediatric patients with unresectable or metastatic TMB-high solid tumours, with the FoundationOne CDx assay approved as the companion diagnostic. The predictive utility of TMB has since been shown in a number of different tumour types and for various immune checkpoint blockers targeting PD-1, PD-L1 and CTLA-4.

However, the value of TMB as a universal predictive biomarker of response to immune checkpoint blockers has been questioned on a number of levels. In a study of patients receiving immune checkpoint blockers for advanced colorectal cancer (CRC), the therapeutic benefit associated with TMB-high was not observed in patients with microsatellite stable/mismatch-repair proficient tumours (N Engl J Med. 2021;384:1168–1170). It is worth noting that the findings do not take into account the fact that the presence of liver metastases and high lactate dehydrogenase (LDH) levels – both of which can be expected in patients with advanced CRC – are linked to resistance to immune checkpoint blockade (Eur J Cancer. 2022:177:80–93). Another factor confounding interpretation of TMB’s role across indications relates to the use of chemotherapy in combination with immune checkpoint blockers. The lack of TMB predictive utility noted in such context may not be surprising given the likely negative impact of chemotherapy on immune responses. We also need to consider the likelihood that the threshold of TMB positivity may not be uniform, but may be higher for some cancers than for others.

In terms of wider clinical implementation, one of the greatest obstacles concerns the challenges faced in gaining regulatory approval. The requirement for randomised phase III trial evidence hinders investigation in the tumour-agnostic setting, given the different standards of care required for each individual cancer type. There is a need for regulatory agencies to consider looking at surrogate endpoints as well as to re-evaluate the level of response describing clinical benefit. For instance, in patients failing standard-of-care agents and for whom there are no further active therapy choices, an approach yielding a 30% durable response rate could be regarded as a valuable option. It should be remembered that we are only talking about a relatively small number of patients, given that only around 10% of patients in this setting will have TMB-high tumours.

Another barrier to the wider clinical use of TMB as a biomarker relates to the financial implications of having to use the approved, tissue-based companion diagnostic test. Alternative techniques, such as next-generation sequencing, are being assessed and the use of blood TMB measurements may also avoid the technical problems associated with obtaining sufficient tumour tissue for assaying. However, this raises the issue of identifying the optimal predictive threshold of blood TMB, and levels of ≥16 mutations/megabase have been found to be equivalent to around 13 mutations/megabase in tissue (Nat Med. 2022;28:1831–1839).

For some, the independent relationship between TMB and PD-L1 is a stumbling block. Counterintuitively, the KEYNOTE-158 trial showed no correlation between PD-L1 expression and the level of TMB, such that tumours can be TMB-high/PD-L1-low and vice versa. The biological basis for the increased responsiveness of TMB-high tumours to PD-1 blockers remains unclear. A better knowledge of the underlying pathways involved may help to improve our understanding of when and how TMB should be applied in the clinical setting.

Don't miss:

Is TMB a good predictive biomarker for immunotherapy? ESMO Immuno-Oncology Congress 2023

Controversy Session, 07.12.2023, h. 14:15 – 15:15, Room A

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.