Purple Background
Articles & News

Case study: Learning from The Pembrolizumab Story

Inovia Bio
January 22, 2024

Three drugs both alike in dignity, in immunotherapy where we lay our scene:

Bad Shakespeare aside, I’d like to propose a simple thought experiment, what would happen If the same drug were developed by two different companies?  A good analogue for this situation is the interesting case study of the PD-(L)1 inhibitor class.

To claim that any of the approved PD-(L)1 inhibitors is arguably “better” from a scientific perspective is difficult however, if we look at the sales figures of Pembrolizumab, Atezolizumab and Nivolumab from 2019 we can see an interesting story.

It’s incredibly difficult not to notice the delta in sales between the three drugs.

The question then becomes why has Pembrolizumab vastly outperformed their competition – despite not only being second to market but also initially lagging in sales compared to BMS (Nivolumab) up to 2018?

What follows is an interesting brief overview of how the team at Merck were able to use agile strategic development to bring an asset originally destined for out-licensing to become one of the best-selling cancer therapies of all time & what drug development teams can learn from them.

The label:

The first hint comes from a broad exploration of the approved FDA labels of all three products.

A cursory glance shows that Pembrolizumab surpasses both Nivolumab and Atezolizumab in the number of approved indications, a factor that typically aligns with increased sales. Moreover, when comparing treatments for the same condition, Pembrolizumab demonstrates a more extensive label than its counterparts, Nivolumab and Atezolizumab, as exemplified by the Non-Small Cell Lung Cancer (NSCLC) indication statements provided below.

The question remains of how the team got here - especially considering that the development of Pembrolizumab was at one point at least 4 years behind Nivolumab?

A troubled start:

Merck gained access to Pembrolizumab through its 2006 acquisition of Schering-Plough. While originally treated with scepticism by senior executives at Schering, the early Pembrolizumab team were able to continue their pre-clinical work and eventually start to piece together an IND. However, post-acquisition a prioritisation exercise by the Merck initially ranked the potential of Pembrolizumab as low and lead to the eventual termination of all development work on the antibody, ultimately listing Pembrolizumab for out-licensing.

The CTLA-4 signal:

Despite an apparent term sheet already in place that valued Pembrolizumab at “next to nothing” things changed in 2010. A publication in the New England Journal reporting the results of a phase III randomised trial of Ipilimumab vs gp100 peptide vaccine, in previously treated patients with unresectable stage III or IV melanoma demonstrated a survival benefit of 10.1 months and a hazard ratio of 0.66 in the Ipilimumab only arm, gave credence to immune checkpoint inhibition as a viable therapeutic strategy. Early data also starting to emerge on Nivolumab that showed signs of efficacy.

Based on the emerging data the Merck team urgently re-prioritised the Pembrolizumab program and filed an IND in less than a year (In Dec 2010 to be precise) and began recruiting patients for what became the largest phase I study in the history of oncology in March 2011.

Adaptive agility:

The Phase I study (KEYNOTE-01/ NCT01295827) initially resembled a typical 3+3 dose escalation design to establish the safety, tolerability and ultimately the RP2D in patients with a diagnosis of either advanced melanoma, any type of carcinoma, progressive metastatic disease, or progressive locally advanced disease not amenable to local therapy, with a planned expansion cohort to investigate the efficacy of the drug at the MTD in participants with advanced melanoma or renal cell carcinoma.

Merck quickly responded to early signals from the study (two patients achieved complete response (one with melanoma and one with Merkel cell carcinoma), three had partial response (all with melanoma), and 15 had stable disease (including 10 different types of solid tumours, including melanoma and NSCLC) by amending the protocol within seven months to eliminate renal cell carcinoma from the first expansion cohort and focus on advanced melanoma patients with progressively locally advanced or metastatic disease.

What follows from this point is an exemplar of why adaptive trial designs are an important tool in the arsenal of drug development teams, a summary of the study and protocol amendments is shown below:

(Kang, et al 2017)

As opposed to following the traditional linear development strategy where each phase is a “distinct” element, (phase I safety, phase II dose finding / early efficacy and phase III registration grade large RCTs) the team at Merck took what is in essence the definition of an agile approach.

The development team utilised an adaptive design that allowed them establish the safety , optimise dosing, signal find and demonstrate efficacy simultaneously.

The approach of KEYNOTE-01 also allowed the team test various hypothesis in different indications and respond in a data driven way. This approach led to a time from IND to the first FDA approval being approximately 4 years as opposed to an approximate 8 year timeline had they had followed a linear development program.

(Kang, et al 2017)

There were two key decisions that the team took which were ultimately critical to cementing a competitive advantage over BMS.

Incorporating a companion diagnostic:

The decision to incorporate a biomarker early during Pembrolizumab’s development enabled the team to gather high-quality data on PD-L1 expression across various tumor types. This approach allowed for analysis of response rates at different expression levels, leading to a well-informed and strategic approach to indication selection and the co-development of a companion diagnostic.

The NSCLC cohorts from KEYNOTE-001 were instrumental in creating high-quality, pre-specified training and test datasets for PD-L1 expression, leading to the development and eventual FDA approval of the pharmDx PD-L1 assay.

Initially, the choice to focus on a biomarker-defined cohort was met with skepticism, as it appeared to narrow the patient segment and necessitate a shift in community practice. However, this strategy proved to be the critical factor in later obtaining approval for first-line NSCLC indication as the team had established which population was most likely to respond. In contrast, a competing study of Nivolumab by BMS, which also used PD-L1 as a biomarker but with a significantly lower threshold, failed to meet its primary endpoint in the same setting.

Regulatory strategy:

The Merck team leveraged the early efficacy signals observed in the advanced melanoma cohort to secure Breakthrough Therapy Designation (BTD) for Pembrolizumab from the FDA as well as a later BTD for NSCLC. This designation facilitated more frequent interactions with the agency, enabling in-depth discussions about trial designs and development strategies. This heightened level of regulatory collaboration allowed for rapid iteration of the team’s strategy, ensuring that the proposed clinical development plan was aligned with regulatory requirements and was sufficiently “de-risked”. This allowed to team to have the confidence to change the traditional development culture at Merck and adopt an agile approach to development.

The Further development of Pembrolizumab:

Pembrolizumab's progression is marked by a deep comprehension of patient populations. Merck's strategy involved a keen understanding of its mechanism of action and the identification of predictive biomarkers for response, starting with PD-L1 expression and later incorporating MSI-H/dMMR status. This approach enabled the expansion of Pembrolizumab into various indications across several settings, ensuring a high response probability in multiple treatment contexts. The key to Pembrolizumab's success lay in the consistent analysis and interpretation of data from the early stages of its development program. While there's a common belief that mimicking a more advanced competing program is a safer strategy, Pembrolizumab's journey highlights the significance of ongoing evidence generation and the value of evidence-based differentiation in drug development.

Lessons for drug development teams:

The story of the development of Pembrolizumab provides key insights for drug development teams.

1. Evidence generation strategy:

Despite Merck originally shelving Pembrolizumab the company’s ability to rapidly identify relevant external data from the scientific literature and other external sources allowed them to reprioritise the program.

2. Identifying opportunities:

Despite initially being behind Nivolumab, Merck were able to differentiate their strategy through a better understanding of patient populations and mechanisms of action.

3. Integrated CDP:

The integrated nature of the Merck clinical development plan allowed for multiple hypothesis testing and high-quality evidence generation which was ultimately supportive of accelerated approvals in shorter timelines.

4. Precision drug development:

By incorporating biomarker testing early, Merck was able to establish and test hypothesis of which patients were likely to respond to Pembrolizumab, this led to focused indication selection as well as smaller, quicker trials later in the development program.

5. Regulatory collaboration:

Merck’s use of the opportunities granted by the BTD to frequently engage with the FDA allowed the team to establish what would be acceptable to the regulator and quickly itterate strategy.

6. Adaptive designs:

The adaptive design utilised by the team allowed for simultaneous evaluation of safety, dosing optimisation, signal finding, and efficacy. While this approach required over 8 protocol amendments, it enabled quicker decision-making and shortened the overall development timeline.

The Inovia Bio Platform

The Only Tool Validated In Both Large Pharma & Early Biotech That Locates The Right Opportunity For An Asset

The Inovia Bio platform systematically identifies opportunities for both early and late-stage assets, highlighting strategic options for development teams with unprecedented precision and speed.


The InovaSight platform continuously integrates bioinformatics and strategic data to provide evidence-based indication prediction and prioritisation to drug development teams. InovaSight enables teams to rapidly develop target product profiles and clinical development plans that maximise the success of a program.


The InovaMine platform instantaneously integrates and analyses clinical, RWE, operational, commercial and biomarker data allowing development and strategic teams to not only respond to evidence generation in real time but to also identify opportunities only visible through a "systems-wide" lens.






For a free consultation on the topic, click on the link below

InovaSight takes the pain out of producing internal decks or hiring agencies for workshops.