PD-1 Inhibitors Running Neck-and-Neck in Head and Neck

By: Haris Vikis, Ph.D. Associate Consultant, Clinical & Scientific Assessment, Kantar Health and Stephanie Hawthorne, Ph.D., Vice President, Clinical & Scientific Assessment, Kantar Health

Patients with recurrent or metastatic (R/M) head and neck squamous cell carcinoma (HNSCC) who are refractory to platinum-based therapy have a poor prognosis with a median overall survival of  six months or less.1,2 While Erbitux® (cetuximab, Eli Lilly/Bristol-Myers Squibb/Merck KGaA) and platinum-based regimens are most commonly used in locally advanced and metastatic disease, immunotherapies are about to take center stage, offering significantly improved therapeutic options based on today’s findings at the American Society for Clinical Oncology (ASCO) annual meeting. Within the last two years, the PD-1 checkpoint inhibitors Keytruda® (pembrolizumab, Merck & Co.) and Opdivo™ (nivolumab, Bristol-Myers Squibb/Ono Pharmaceuticals) have garnered approvals in several tumor types, including melanoma, non-small cell lung cancer, and more recently renal cell carcinoma and Hodgkin’s lymphoma for Opdivo. Data presented at ASCO 2016 now have single-agent Keytruda and Opdivo racing for approval in HNSCC.

Over the last 12 months Opdivo and Keytruda have each reported promising topline efficacy and safety data in the R/M setting for HNSCC.3-5 This prompted regulatory filing for accelerated approval for Keytruda in April 2016 based on data from the KEYNOTE-012 Phase I study.3,4 A U.S. Food and Drug Administration (FDA) action date of August 9, 2016, has been set, and it is anticipated to result in the first immunotherapy approval in HNSCC, likely subject to confirmation of the ongoing randomized Phase III KEYNOTE-040 trial. Hot on the heels of Keytruda is Opdivo, which recently reported positive data at the 2016 American Association of Cancer Research (AACR) annual meeting based on a planned interim analysis of the CHECKMATE-141 Phase III study.3 The FDA quickly granted Breakthrough Therapy designation, and a regulatory filing is highly anticipated in the near future. Given the high unmet needs within HNSCC and the enthusiasm for immunotherapy that has penetrated oncology over the last few years across multiple tumor types, these agents are likely to be approved and adopted quickly. Today at the ASCO 2016 annual meeting, four oral abstracts were presented6-9 detailing updated efficacy and safety data that strongly support the registration programs for both Keytruda and Opdivo, further prompting cross-trial comparative analyses.

Today’s presentation on CHECKMATE-141, a randomized (2:1) Phase III trial of 361 patients testing Opdivo versus investigator’s choice therapy in R/M platinum refractory patients,4 was largely a reprisal of data presented at AACR in April 2016. In the ASCO presentation, Opdivo demonstrated a median overall survival (mOS) of 7.5 months (HR: 0.70, P=0.01 when compared with the control arm). More benefit was observed in the PD-L1 ≥ 1% subgroup with a mOS of 8.7 months (HR: 0.55, 0.36-0.83) and an objective response rate (ORR) of 17%. Benefit in the PD-L1 ≤ 1% population exhibited an ORR of 12% and a mOS of 5.7 months (HR: 0.89, 0.54-1.45), which were similar to the control arm. However, as first described at AACR there is some divergence in the Kaplan-Meier curve at eight to nine months, suggesting some long-term benefit with Opdivo in this patient subgroup. Nevertheless, these data largely dismiss PD-L1 status as an effective biomarker of response, as all patients benefit regardless of PD-L1 status. The updated data presented today continue to support this. The same story holds true for HPV positivity as a biomarker of response, as both HPV-positive and HPV-negative patients continue to derive benefit. The Opdivo arm yielded superior safety. Treatment-related adverse events (TRAEs) totaled 59% (Grade 3-4 = 13%) versus 78% (35% Grade 3-4) in the control arm, and no new safety signals were reported. These data are generally consistent with the improved safety profiles in other tumor types and in the KEYNOTE trials discussed below. Finally, newly released data showed that Opdivo stabilized measures of patient-reported outcomes (PROs), while patients on investigator’s choice had declines in function and worsening of symptoms, adding another reason to support Opdivo as a new standard-of-care option in R/M HNSCC.

Merck has an aggressive development strategy for Keytruda in HNSCC, which was evident in the three abstracts reported on two Keytruda trials in the R/M setting. Long-term follow-up of expansion cohorts in HNSCC (n=192) as part of KEYNOTE-012 was presented at ASCO 2016.7 Updated data revealed an ORR that was largely maintained at 18%, with observations of significant durable responses as 85% of responses lasted six months or longer. The median OS was slightly reduced to eight months from previous reporting.3,4 A companion abstract reported in-depth biomarker analysis for PD-L1, which consistently showed increased ORR, OS and progression-free survival (PFS) in PD-L1-positive versus PD-L1-negative tumors.8 Interestingly, when tumor and inflammatory cells were included as part of a PD-L1 biomarker score, the ORR increased to 21% in PD-L1-positive versus only 6% in PD-L1 negative tumors, indicating a significantly increased ability to predict response, which may be an important consideration in the future. Biomarkers beyond PD-L1, including PD-L2 and an IFN-γ gene signature, were also strongly correlated with response, PFS and OS.

And finally newly released data from KEYNOTE-055, a large, single-arm Phase II study in platinum- and Erbitux-refractory patients, also were presented and largely confirmed the findings of KEYNOTE-012. This study was performed in a more heavily pretreated population (84% with two or more prior regimens) and reported an ORR of 17% in the total and PD-L1-positive population. TRAEs were reported at 60% (12% Grade 3-4) and agreed well with safety signals observed in the KEYNOTE-012 trial and CHECKMATE-141 trials.

How will data from these three trials be viewed in cross-trial comparisons? While the patient populations are largely similar (with the exception of the KEYNOTE-055, which was conducted in a more heavily pretreated population), the efficacy and safety data are remarkably similar between Keytruda and Opdivo (see table). Efficacy is consistently in the 18% range for ORR, two months for PFS and eight months for OS at the median. All measures of efficacy were independent of HPV status, and PD-L1 status was  largely predictive of those who may benefit more but could not exclude those patients who may benefit regardless even among the PD-L1–negative population.

Patient Setting

(≥2 prior lines)


prior plat


91% had prior plat


prior plat and cetuximab

mOS 7.5 mos

8.7 mo (PD-L1 ≥ 1%)

8 mos 8 mos
mPFS 2.0 mos 2.2 mos 2.1 mos
ORR 17% (PD-L1 ≥ 1%) 18% (PD-L1 ≥ 1%) 17% (PD-L1 ≥ 1%)
Any TRAEs 59% 64% 60%
Grade 3/4 TRAEs 13% 13% 12%


Based on the filing in April, Keytruda is set to become the first immunotherapy approved for HNSCC. While Opdivo has the virtues of a larger and randomized study, Keytruda now has confirmatory evidence from the Phase II KEYNOTE-055 study, and it remains to be seen whether the data presented at ASCO are included as part of the Keytruda regulatory filing. If the FDA follows suit as it did in melanoma and lung, it will grant Keytruda accelerated approval and await data from a randomized Phase III trial (KEYNOTE-040) before granting full approval. However, Keytruda will need to compete in this setting with Opdivo, which has comparable data in a randomized Phase III trial, and durvalumab (AstraZeneca), whose ongoing EAGLE Phase III trial has yet to report. As Merck did with Keytruda, AstraZeneca has also guided for possible accelerated filings based on Phase II data.

As single-agent therapies, there is little evidence to suggest a difference between Keytruda and Opdivo PD-1/-L1 inhibitors in terms of efficacy and safety. These agents may be able to distinguish themselves by the additional development strategies being employed ‒ durvalumab and Keytruda are also exploring use of the PD-L1 biomarker to select patients; Opdivo is focusing strictly on an all-comers population; and durvalumab is also being developed in combination with tremelimumab. While refinement of biomarker populations based on PD-L1, PD-L2 and IFN-γ gene signatures have been presented, little evidence to date suggests that PD-L1 is the appropriate biomarker to distinguish responders from non-responders. Success of one PD-1 inhibitor over the other may come down to physician preference.

It is clear that single-agent Keytruda and Opdivo significantly improve the treatment armamentarium in R/M HNSCC for a patient population with limited options. Responses are durable, and the agents are well tolerated with distinct and manageable toxicity profiles. While the race for the first immunotherapy in HNSCC may be neck-and-neck, future developmental activity continues and undoubtedly is moving toward evaluation of combination therapies and integration of immunotherapy into earlier lines of therapy.


  1. Machiels JP et al., Lancet Oncol. 2015 May;16(5):583-94
  2. Kushwaha VS et al., Cancer Biol Ther. 2015;16(2):346-51
  3. Seiwert, Abstract LBA 6008, ASCO 2015
  4. Chow, Abstract 2866, ESMO 2015
  5. Gillison, Abstract CT099, AACR 2016
  6. Ferris, Abstract 6009, ASCO 2016
  7. Mehra, Abstract 6012, ASCO 2016
  8. Bauml, Abstract 6011, ASCO 2016
  9. Chow, Abstract 6010, ASCO 2016
article register

Recent Posts

Recent Comments



Post a Comment

You must be logged in to post a comment