For the first time the FDA has approved a drug (Merck & Co.’s PD-1 inhibitor Keytruda) based on biomarker status alone. This move away from indication-based labelling represents an exciting meeting of two central themes in current-day cancer drug development: biomarkers and immunotherapy.

For many years, biomarker-driven prescribing fuelled the development and approval of many oncology therapies. Biomarkers are attractive: they identify the patient population most likely to benefit from a given drug, and spare patients who are unlikely to benefit from enduring side effects. A validated, biomarker-selected population is also more likely to produce positive clinical data in randomized clinical trials, and these therapies also enjoy strong penetration of their eligible populations owing to the compelling rationale for prescribing them. Furthermore, biomarker-associated therapies often win favorable reimbursement in cost-constrained markets, making them good commercial opportunities.

Immunotherapy languished for many years in the shadow of biomarker-driven therapy development in oncology, with just a handful of infrequently-used treatments on the market that caused high toxicities in exchange for impressive but rare efficacy benefits. Over the last few years however, immune checkpoint inhibitors, in particular PD-1/PD-L1 inhibitors, have rocketed immunotherapy into the limelight, rapidly gaining momentum based on impressive efficacy outcomes in a range of tumor types. In 2017, immunotherapy and immune checkpoint inhibitors dominate oncology drug development efforts. Frenzied development of this drug class has seen them secure a multitude of approvals across a diverse range of tumor types, but a key problem has persisted: they do not have a biomarker that reliably predicts response to treatment.


Personalizing immunotherapy: the search for a biomarker

Development of PD-1/PD-L1 inhibitors has been rapid (for example, the first Phase I data for Bristol-Myers Squibb’s Opdivo was shown at ASCO 2012, and the drug was FDA approved by the end of 2014). It was only after trials started to demonstrate the clinical promise of PD-1/PD-L1 inhibitors that efforts began to concentrate on developing a biomarker for likelihood of response. Efforts have mainly focused on PD-L1 (the ligand of PD-1), as logic dictated that there could be a correlation between levels of PD-L1 expression and effectiveness of treatment. (This is a similar approach to some currently marketed biomarker-associated drugs like the EGFR inhibitor Tarceva, which is approved for EGFR-mutation-positive NSCLC, or the HER2-targeted agent Herceptin, which is approved for HER2-overexpressing breast cancer.)  However, PD-L1, whether measured on tumor cells or tumor-infiltrating lymphocytes, and at different concentrations has not proven to be a satisfactory biomarker for PD-1 and PD-L1 inhibitors, as treatment responses are still observed in so-called PD-L1-negative patients. Despite this, several PD-1/PD-L1 inhibitors have been involved in trials focused on the PD-L1-positive population, and Keytruda holds FDA approval for PD-L1-positive NSCLC as a single agent.


MSI-H and dMMR – an indirect biomarker for Keytruda and other PD-1/PD-L1 inhibitors

MSI-H and dMMR are an indirect biomarker to predict response to PD-1 and PD-L1 inhibitors.

MSI’s are short, repetitive DNA sequence tracts scattered throughout the genome that are particularly prone to DNA replication errors. Changes in the number of MSI’s can occur due to inactivation of MMR genes (i.e. dMMR), leading to an accumulation of genetic mutations (a higher mutational load) that contributes to carcinogenesis (Bogaert J & Prenen H, 2014). The higher mutational load in patients with MSI-H or dMMR tumors is hypothesized to correspond to an increased presence of tumor-specific neoantigens that can be seen by the immune system (specifically, T-cells) (Schumacher TN & Schreiber RD, 2016). Tumors with higher neoantigen load therefore may be more susceptible to treatment with immune checkpoint inhibitors, such as PD-1 and PD-L1 inhibitors.

Keytruda is the best studied agent in the MSI-H and dMMR population. Its recent FDA approval in this setting is based on Phase II clinical data, however it is also the focus of an ongoing Phase III trial in first-line metastatic colorectal cancer with MSI-high or MMR deficiency (KEYNOTE-177; NCT02563002). Opdivo has also announced data from a Phase II trial (CheckMate-142; NCT02060188) in MSI-H colorectal cancer earlier in 2017.


The market potential of the MSI-H and dMMR population

While MSI and dMMR are relatively rare, this does not mean the potential market for targeting this population is negligible. MSI-high and dMMR is found in approximately 4% of colorectal cancers (Le DT, 2015; Goldstein J, 2014), which is a relatively high incidence oncology indication. MSI-H and dMMR are also found in gastrointestinal cancers and endometrial cancer, as well as more rarely in other high incidence tumor-types such as breast, prostate, bladder, and thyroid cancer. This gives Keytruda broad application across solid tumors. We know that drugs targeting niche populations are able to produce meaningful revenues; for example ALK-translocations are present in just 6% of NSCLC patients, yet this population has become a $600 million market with the forecasted potential to exceed $2 billion (Non-Small-Cell Lung Cancer Disease Landscape & Forecast, 2017, Decision Resources Group). Keytruda therefore has the potential to turn this new approval into a lucrative opportunity. A greater limit on Keytruda’s uptake in MSI-H and dMMR solid tumors will likely be the label stipulation that it is for “patients with no suitable alternative”. In some indications this may be reached quickly, after two or three lines of therapy, but in other indications with a plethora of drugs available this could be much longer. This could really restrict Keytruda’s potential sales in this setting, as later-lines of therapy tend to hold smaller drug-treatable populations.


In conclusion, the recent approval of Keytruda for MSI-H or dMMR solid tumors is likely to open a new chapter in oncology. While the introduction of personalized medicine has been written by indication-based biomarker-driven approvals in oncology, the true story may ultimately lie in prescribing according to biomarker, rather than disease.


Insights from this article are drawn from a range of DRG products and services, including:

  • Non-Small-Cell Lung Cancer Disease Landscape & Forecast, 2017, Decision Resources Group
  • Colorectal Cancer Disease Landscape & Forecast, 2016, Decision Resources Group
  • Immune Checkpoint Inhibitors Special Topics, 2017, Decision Resources Group



Bogeart J and Prenen H. Molecular genetics of colorectal cancer. Annals of Gastroenterology. 2014;27(1): 9-14.

Goldstein J, et al. Multicenter retrospective analysis of metastatic colorectal cancer (CRC) with high-level microsatellite instability (MSI-H). Annals of Oncology. 2014;25(5): 1032-1038.

Le DT, et al. Programmed death-1 blockade in mismatch repair deficient colorectal cancer. Journal of Clinical Oncology. 2016; 34:(suppl 4S; abstr 195).

Schumacher TN and Schreiber RD. Neoantigens in cancer immunotherapy. Science. 2015;348(6230):69-74.


How Glympse Bio oversubscribed their Series B funding amidst the pandemic

View Now