First-line treatment of head and neck cancer could soon be less EXTREME given KEYNOTE-048.

by Arnold DuBell | Oct 23, 2018
Co-authored by Greg Wolfe

Until recently, the only approved targeted agent for treatment of patients with squamous cell carcinoma of the head and neck (SCCHN) was Erbitux® (cetuximab, Merck KGaA / Eli Lilly).  Erbitux was FDA approved for locally-advanced disease in 2006 (EMA in 2006, Japan in 2012) based on results from a National Cancer Institute-sponsored trial that compared radiotherapy (RT) with or without Erbitux.[1] Erbitux subsequently received FDA approval for use in combination with a platinum agent plus 5-fluorouracil (5-FU), based on results of the EXTREME trial[2].  Additional targeted agents finally entered the SCCHN market when checkpoint inhibitors (CPIs), Keytruda® (pembrolizumab, Merck) and Opdivo® (nivolumab, Bristol-Myers Squibb / Ono Pharmaceuticals), were FDA approved (August 2016, Keytruda; November 2016, Opdivo) for use in recurrent or metastatic SCCHN patients who progressed following platinum-containing chemotherapy.  Keytruda’s accelerated approval was based on the Phase Ib KEYNOTE-012 trial[3], and Opdivo’s full approval was based on results from the CheckMate 141 trial[4].  While Opdivo is also approved in Japan (March 2017) and Europe (April 2017), Keytruda is not yet approved in these regions.  Interestingly, Merck originally hoped to use the KEYNOTE-040 trial, that compared Keytruda to single-agent therapies (methotrexate, docetaxel, or Erbitux) in biomarker-unselected platinum-refractory SCCHN patients, to convert its FDA accelerated approval to a full approval, but this trial failed to meet its primary endpoint of overall survival (HR 0.81, one-sided p=0.0204)[5].

With this history, it may have come as somewhat of a surprise when Merck announced via press release in July 2018 that the first-line KEYNOTE-048 trial met one of its primary endpoints as PD-L1+ patients treated with Keytruda monotherapy had an improved OS compared to those treated with the EXTREME regimen (Erbitux plus a platinum and 5-FU).  Data for other endpoints, including data for the combination of Keytruda with chemotherapy, were not announced at that time. Patients enrolled in the KEYNOTE-048 (n=825) were randomized to one of three arms: Keytruda monotherapy (200 mg IV day 1 q3w for up to two years), Keytruda in combination with a platinum and 5-FU, or the EXTREME regimen.  Patients were required to provide tissue for biomarker analysis; however, PD-L1 positivity was not an inclusion criterion.  PD-L1 positivity was measured using a combined proportion score (CPS), defined as the number of PD-L1 staining cells (tumor cells, lymphocytes, and macrophages) divided by the number of total viable tumor cells, multiplied by 100.  The first threshold for PD-L1+ in KEYNOTE-048 was a CPS ≥ 20; if positive, then a second threshold would be set at a threshold of CPS ≥ 1 and the statistical analysis accordingly modified.  Trial primary endpoints were overall survival (OS) and progression-free survival (PFS) in all patients, in patients with CPS ≥1, and in patients with CPS ≥20.

The eagerly awaited data from KEYNOTE-048 were finally presented Dr. Barbara Burtness at ESMO 2018[6].  In CPS ≥ 20 patients, administration of Keytruda monotherapy resulted in a statistically significant improvement in OS comparted to those who received EXTREME (14.9 months versus 10.7 months, HR 0.61, p =0.0007).  Keytruda monotherapy was also superior to the EXTREME regimen in the CPS ≥ 1 

patients, in terms of OS (12.3 months versus 10.3 months, HR 0.78, p =0.0086); although, Keytruda did not improve PFS in either PD-L1-defined subgroup [(CPS ≥20: 3.4 months versus 5.0 months, HR 0.99, p=0.5); (CPS ≥1: 3.2 months versus 5.0 months, HR 1.16, 95% CI: 0.96-1.39). OS Kaplan-Meier curves and the PFS curves showed that patients on the EXTREME regimen seemed to perform slightly better than Keytruda during the first 7 months, and then the curves diverged in favor of Keytruda.  Response rates help explain the phenomenon seen in the Kaplan-Meier curves. In the CPS ≥ 20 patients, patients that received the EXTREME regimen had a higher overall response rate than patients treated with Keytruda (36.1% versus 23.3%), although complete responses were higher with Keytruda (7.5% versus 3.3%).  That said, patients who responded to Keytruda enjoyed a much longer duration of response (median, mDOR; 20.9 months versus 4.2 months).  Similar data were presented for patients in the CPS ≥ 1 group where Keytruda monotherapy induced fewer, but more durable responses than the EXTREME regimen (ORR: 19.1% versus 34.9%; CR: 5.4% versus 2.7%; mDOR: 20.9 months versus 4.5 months).  Keytruda was better tolerated in these patients: 58.3% of patients treated with Keytruda versus 96.9% of patients on the EXTREME regimen experienced an adverse event of any grade, and this was reflected in a decreased rate of treatment discontinuation due to adverse events (4.7% versus 19.9%). More common toxicities for Keytruda were hyperthyroidism (approximately 17%), fatigue (approximately 14%), rash (approximately 7%), and pneumonitis (approximately 6%).

Dr. Burtness then presented data for the combination of Keytruda with chemotherapy (a platinum plus 5-FU).  In the total population of patients, regardless of PD-L1 expression, the combination of Keytruda plus chemotherapy was associated with a significant improvement in overall survival compared to the EXTREME regimen (13.0 months versus 10.7 months, HR 0.77, p=0.0034); however, PFS was not improved (4.9 months versus 5.1 months, HR 0.92, p=0.2).  Overall response rates were similar between the two arms (35.6% versus 36.3%) although the CR rate (6.0% versus 2.9%) and the mDOR (6.7 months versus 4.3 months) were improved with Keytruda. Frequency of toxicities (any grade) were similar between these two arms chemotherapy-containing arms (65.35 versus 96.9%).  Common toxicities for the combination of Keytruda and chemotherapy were anemia (approximately 48%), nausea (approximately 45%), neutropenia (approximately 33%), fatigue (approximately 30%), mucosal inflammation (approximately 29%), thrombocytopenia (approximately 28%), and vomiting (approximately 28%).

Data from KEYNOTE-048 were very promising and will likely support regulatory approval of Keytruda with or without chemotherapy for treatment of patients with SCCHN.  The question becomes: how will physicians might consider use of the regimens? One possible outcome was aptly provided by the discussant, Dr. Jean-Pascal Machiels.  He first noted the fact that for the first seven months the EXTREME regimen performed at least slightly better than Keytruda and suggested some physicians might select this regimen, with its higher response rate, for their patients with high disease burden.  At the end of his discussion, Dr. Machiels presented a possible treatment algorithm for treating first-line SCCHN patients should Keytruda gain approval in the first-line setting.  For patients with a CPS of 0, he recommended use of the EXTREME regimen.  Keytruda plus chemotherapy was recommended both for patients for whom no CPS status is available and for those patients with CPS scores between 1 and 20.  Keytruda monotherapy was preferred for most of those patients with a CPS score ≥ 20 (except for those with high disease burden) since the trial was not designed to evaluate what additive benefit chemotherapy provided.  The one concern Dr. Machiels mentioned for the use of the Keytruda-chemotherapy combination was sequencing: it is clear to him that platinum-based chemotherapy administration after Keytruda monotherapy would be appropriate, as would be administration of a CPI to patients that receive the EXTREME regimen as the first-line treatment.  However, there are yet no studies that address the question of what to offer patients that receive the Keytruda-chemotherapy combination in first-line.

Other CPIs are in phase III development, so Keytruda could soon have competition in this space.  AstraZeneca is evaluating the combination of their PD-L1 inhibitor, Imfinzi® (durvalumab), and tremelimumab, a CTLA-4 inhibitor, in the Phase III KESTREL trial.  AstraZeneca has guided for H1 2019 filing. If results from the KESTREL trial are positive, Keytruda may not enjoy a substantial first-to-market advantage.  Furthermore, the combination of Opdivo plus Yervoy® (ipilimumab, Bristol-Myers Squibb) is under investigation in the Phase III CheckMate-651 trial.  Patient accrual is ongoing in CheckMate-651, so Keytruda may have time entrench itself before this trial reads out.  Both competitors are evaluating combinations of a PD-1 / PD-L1 inhibitor plus a CTLA-4 inhibitor; therefore, some potential exists for these regimens to show a stronger efficacy benefit, and potentially, immune-related adverse events to consider. 

With positive data from KEYNOTE-048 that may well lead to approval of Keytruda monotherapy or a Keytruda-chemotherapy combination in first-line, and potential promise for combinations of Opdivo plus Yervoy, and Imfinzi plus tremelimumab, physicians may soon have additional treatment options that will improve outcomes for patients with SCCHN. 


  1.  Bonner JA et al.; NEJM, 354: 567, 2006
  2. Vermorken JB et al.; NEJM, 359:1116, 2008
  3. Seiwert TY et al.; Lancet Oncol, 17:956, 2016
  4. Ferris RL et al.; NEJM, 375:1856, 2016
  5. Cohen E et al.; Abstract LBA45_PR, ESMO 2017
  6. Burtness B et al.; Abstract LBA8_PR, ESMO 2018

Leave a comment

More than one Google Analytics scripts are registered. Please verify your pages and templates.