PePS2 Results

Results of the PePS2 trial of pembrolizumab in performance status 2 non-small-cell lung cancer patients published in Lancet Respiratory Medicine.

Photo by Sara Bakhshi on Unsplash


PePS2 is a phase II clinical trial of pembrolizumab in non-small-cell lung cancer (NSCLC) patients with a performance status of 2. Trial results were recently published in Lancet Respiratory Medicine (Middleton et al. 2020). Previous trials in the KEYNOTE series (Garon et al. 2015; Herbst et al. 2016) had showed that pembrolizumab was efficacious in NSCLC patients.

The gap in the evidence

Performance status (PS) is a variable defined by the Eastern Cooperative Oncology Group to reflect cancer patients’ general well-being and activities of daily life. It is an ordered variable with five levels reflecting greater levels of disability:

  • 0 – patients are asymptomatic;
  • 1 – patients are symptomatic but completely ambulatory;
  • 2 – patients are symptomatic, spending up to 50% of day time in bed;
  • 3 – patients are symptomatic, spending at least 50% of day time in bed;
  • 4 – patients are bedbound.

We will say that a patient with a peformance status of 2 is PS2.

The KEYNOTE-001 trial (Garon et al. 2015) showed that pembrolizumab was efficacious in NSCLC patients. The subsequent KEYNOTE-010 trial (Herbst et al. 2016) showed that pembrolizumab was superior to docetaxel chemotherapy. However, it is doubtful that a PS2 patient would be strong enough to tolerate chemotherapy. Hence, to maintain equipoise in KEYNOTE-010, the trialists restricted randomisation to PS0 and PS1 patients. That trial was successful and pembrolizumab was licenced in NSCLC. However, as is typical, the treatment was licenced in the patient population that had been investigated in the trial. This covered just the PS0 and PS1 patients and crucially excluded PS2 patients. The KEYNOTE trials had showed that pembrolizumab was tolerable, particularly compared to the harsh chemotherapy alternatives. It is precisely in the PS2 population where a relatively tolerable but effective treatment was sorely needed. However, by virtue of their probable unsuitablility to chemotherapy, PS2 patients had been excluded from a promising new treatment.

The PePS2 trial sought to address this gap in the evidence.


In the KEYNOTE-001 trial, Garon et al. (2015) had introduced the programmed death ligand 1 tumour proportion score biomarker (PD-L1 TPS, or simply TPS), being the percentage of cells in a tumour biopsy that expressed PD-L1. Pembrolizumab is a PD-L1 blockade drug so it was reasonably hypothesised that the treatment would work best in those patients whose tumours expressed lots of PD-L1. The KEYNOTE-001 showed this to be precisely the case.

KEYNOTE-001 also showed that first-line patients (those patients who had received no previous anti-cancer therapy) achieved outcomes that appeared to be slightly different to patients receiving pembrolizumab as a subsequent-line therapy.

In PePS2, we sought to evaluate the effects of pembrolizumab on co-primary efficacy and toxicity outcomes whilst stratifying the analysis for these baseline covariates. I developed a statistical design for this purpose, described here (Brock et al. 2019).


On 19-March-2020, the results were published in Lancet Respiratory Medicine. Alessi and Awad (2020) provided an accompanying commentary of the data in that same journal.


Information on the PePS2 project is collated on the project page.


Alessi, Joao V, and Mark M Awad. 2020. “Immunotherapy in Lung Cancer: Effective for Patients with Poor Performance Status?” The Lancet Respiratory Medicine, March, S2213260020301077.
Brock, Kristian, Lucinda Billingham, Christina Yap, and Gary Middleton. 2019. “A Phase II Clinical Trial Design for Associated Co-Primary Efficacy and Toxicity Outcomes with Baseline Covariates.” In Bayesian Statistics and New Generations, edited by Raffaele Argiento, Daniele Durante, and Sara Wade, 125–33. Springer Proceedings in Mathematics & Statistics. Cham: Springer International Publishing.
Garon, Edward B, Naiyer a Rizvi, Rina Hui, Natasha Leighl, Ani S Balmanoukian, Joseph Paul Eder, Amita Patnaik, et al. 2015. “Pembrolizumab for the Treatment of Non-Small-Cell Lung Cancer.” The New England Journal of Medicine 372 (21): 2018–28.
Herbst, Roy S., Paul Baas, Dong Wan Kim, Enriqueta Felip, José L. Pérez-Gracia, Ji Youn Han, Julian Molina, et al. 2016. “Pembrolizumab Versus Docetaxel for Previously Treated, PD-L1-Positive, Advanced Non-Small-Cell Lung Cancer (KEYNOTE-010): A Randomised Controlled Trial.” The Lancet 387 (10027): 1540–50.
Middleton, Gary, Kristian Brock, Joshua Savage, Rhys Mant, Yvonne Summers, John Connibear, Riyaz Shah, et al. 2020. “Pembrolizumab in Patients with Non-Small-Cell Lung Cancer of Performance Status 2 (PePS2): A Single Arm, Phase 2 Trial.” The Lancet Respiratory Medicine 0 (0).
Kristian Brock
Statistical Consultant

I am a clinical trial methodology statistician that likes to use Bayesian statistics.