By Anthony M Magliocco MD
A new molecular marker “Tumor Mutational Burden” is rapidly emerging in the immuno-oncology world. New trials are showing that TMB may be superior to PDL1 IHC analysis to determine a patients probability of responding to costly and potentially toxic immuno-therapy treatments such as immune checkpoint inhibitors.
Tumor Mutational Burden
The Cancer Genome Atlas (TCGA) has shown that cancers have significant variation in the burden of genomic mutations they carry. Some tumors such as melanoma have extremely high burdens whereas others such as thyroid cancer have very low loads.
Some tumors have exceptionally high mutational loads which probably represents an underlying DNA repair deficiency such as POLE or MSI abnormalities. It may also reflect the mechanism of oncogenesis as UV induced tumors such as melanoma have very high burdens.
It is thought that TMB actually results in the development of neo-antigens, which are essentially immunogenic. The probability of neo-antigens emerging is proportional to the total tumor mutational burden. However, this is still a probability measurement, its possible that tumors with even low mutational loads might still generate neo-antigens of interest to the immune system.
Recent Clinical Trials Point to TMB as an important pan-cancer marker
There have recently been three interesting trials in advanced lung cancer reported with a significant association between tumor mutational burden (TMB) and response to the PD-L1 inhibitor nivolumab (Opdivo). CheckMate 012 trial, was a single-arm evaluation of the combination of the PD-1 inhibitor nivolumab (Opdivo) and the CTLA-4 inhibitor ipilimumab (Yervoy), reveled benefit in patients with high TMB independent of PDL-1 expression.
At AACR The CheckMate 568 trial, used a TMB cutoff of ≥10 mutations per megabase of DNA (mut/Mb) as the definition of high TMB. Comparing TMB and response rate in 98 patients with untreated stage IV non-small cell lung cancer (NSCLC), investigators found a 44% response rate in association with TMB ≥10 mut/Mb and no further improvement in response with a higher TMB. The response rate fell dramatically with TMB <10 mut/Mb. This finding is interesting as it may there is a “shelf” or a bimodal distribution of TMB
According to Dr Ramalingam of Emory “PD-L1 and TMB identify distinct and independent populations of non-small cell lung cancer that independently are associated with enhanced objective response rate and progression-free survival.”
Investigators in the randomized CheckMate 227 trial prospectively applied the TMB cutoff of ≥10 mut/Mb. There was aa threefold improvement in 12-month PFS (42.6% versus 13.2%) in the subgroup of patients with TMB ≥10 mut/Mb.
The PFS difference persisted across analyses of high and low PD-L1 expression and squamous versus nonsquamous histology.
“CheckMate 227 validates TMB as an important and independent biomarker to be routinely tested in treatment-naive, advanced non-small cell lung cancer,” said Matthew Hellmann, MD, of Memorial Sloan Kettering Cancer Center in New York City
“TMB should be a standard of care in the initial evaluation of the patient with non-small cell lung cancer,” said Naiyer Rizvi, MD, of Columbia University Medical Center in New York City. “PD-L1 as a biomarker remains as a standard of care in concert with TMB. a validated TMB platform needs to be used.”
Problems with TMB
TMB is definitely showing promise, but what are the drawbacks?
NGS is required
First, TMB calculations require that a significant portion of DNA be sequenced, to generate enough sequence information to determine the tumor mutational load. However, some recent studies suggest that even targeted sequencing panels may provide enough sequence information to determine if the load is high. Access to NGS sequencing remains a challenge. Due to low reimbursements and difficulty of implementing the technology many oncologists may have difficulty accessing the technology
The calculation of TMB is currently non-standardized and non-trivial
Second, TMB calculation is not standardized. It is not a trivial bioinformatics process as the bioinformatifcs process needs to determine if a DNA variation is “real” or an artifact of sequencing- this is non-trivial as filters need to be defined to define the criteria to make a “call” ie what the confidence of the read is, what the alleic fraction is and whether the mutation is somatic or germline. In addition it must be determined if the sequence affects the coding region of a gene. Further complicating this is what the denominator might be in an assay- ie does the NGS sequence only coding regions or are there significant non-coding regions. If the non-coding regions are included in the calculation the number may be artifactually low.
Third, thinking from a biological and mechanistic approach, it may matter whether the mutation actually produces a neo-antigen. Again this cannot be easily measured. It involves factors such as whehter the mutation is actually transcribed into protein, and whether the protein conformation is actually altered and neo-antigenic. Further issues include whether the sequence is secreted or made available when the cells degenerate.
TMB are not the only source of neo-antigens
DNA mutations may only account for some of the neo-antigens that a cancer can create. Other sources of neo-antigens in neoplasia include microbes (ie HPV virus in HPV driven cancers such as cervical or head and neck cancer. Other sources include post-translational modifications in cancer such as glycosolation events etc.
Further complicating the impact of neo-antigens include the condition of the host immune system and its capacity to recognize and react to neo-antigens. For example in immune deficiency conditions, neo-antigens may be present but ignored by the immune systems. Or genetic variants in cellular receptors or MHC may affect how neo-antigens are presented to the immune system
Neo-Antigens may be induced in tumors as a therapeutic strategy
Another interesting angle affecting therapy is the possibility that neo-antigens could be induced in a cancer to trigger immune response. This effect may be a side effect of some therapies such as temozolamide or radiation therapy.
Dr Magliocco is Chair of Pathology and director of the Morsani Molecular Laboratories at the Moffitt Cancer Center