European Journal of Cancer 107 (2019) 142e152 Available online at www.sciencedirect.com ScienceDirect journal homepage: www.ejcancer.com Original Research Durvalumab for recurrent or metastatic head and neck squamous cell carcinoma: Results from a single-arm, phase II study in patients with 25% tumour cell PD-L1 expression who have progressed on platinum-based chemotherapy* Dan P. Zandberg a,*, Alain P. Algazi b, Antonio Jimeno c, James S. Good d, Jérôme Fayette e, Nathaniel Bouganim f, Neal E. Ready g, Paul M. Clement h, Caroline Even i, Raymond W. Jang j, Stuart Wong k, Ulrich Keilholz l, Jill Gilbert m, Moon Fenton n, Irene Braña o, Stephanie Henry p, Eva Remenar q, Zsuzsanna Papai r, Lillian L. Siu j, Anthony Jarkowski s,1, Jon M. Armstrong s, Kobby Asubonteng s, Jean Fan s, Giovanni Melillo s, Ricard Mesı́a t a University of Pittsburgh Medical Center, Hillman Cancer Center, Pittsburgh, PA, USA b University of California San Francisco Helen Diller Family Comprehensive Cancer Center, San Francisco, CA, USA c Division of Medical Oncology, University of Colorado School of Medicine, Aurora, CO, USA d Institute of Head and Neck Studies and Education, Queen Elizabeth Hospital, Birmingham, UK e Clinical Oncology, Cancer Center Centre Le´on Be´rard, University of Lyon, Lyon, France f Department of Oncology, McGill University Health Centre, Montréal, QC, Canada g Duke University Medical Center, Durham, NC, USA h Department of Oncology, Leuven Cancer Institute, KU Leuven, Belgium i Department of Head and Neck Oncology, Institut Gustave Roussy, Villejuif, France j Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada k Division of Hematology Oncology, Medical College of Wisconsin, Milwaukee, WI, USA l Charité Comprehensive Cancer Center, Berlin, Germany m Henry-Joyce Cancer Clinic, Nashville, TN, USA n The West Cancer Center, University of Tennessee Health Science Center, Memphis, TN, USA o Medical Oncology Department, Vall d’Hebron Institute of Oncology (VHIO), Vall d’Hebron University Hospital, Universitat Autonoma de Barcelona, Barcelona, Spain p Department of Oncology-Hematology, Radiotherapy, and Nuclear Medicine, CHU UCL Namur, Namur, Belgium q National Institute of Oncology (Országos Onkológiai Inte´zet), Budapest, Hungary * This study was supported by AstraZeneca. * Corresponding author: Division of Hematology/Oncology, University of Pittsburgh Medical Center Hillman Cancer Center, UPMC Cancer Pavilion, 5150 Centre Avenue Room 552, Pittsburgh, PA 15232, USA. E-mail address: zandbergdp@upmc.edu (D.P. Zandberg). 1 Now Bristol-Myers Squibb. https://doi.org/10.1016/j.ejca.2018.11.015 0959-8049/ª 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 143 r State Health, Center Higatian Defanse Forses, Budapest, Hungary s AstraZeneca, Gaithersburg, MD, USA t Medical Oncology Department, Catalan Institute of Oncology, University of Barcelona, IDIBELL, Barcelona, Spain Received 30 August 2018; received in revised form 9 November 2018; accepted 10 November 2018 Available online 18 December 2018 KEYWORDS Abstract Background: Patients with recurrent/metastatic head and neck squamous cell car- Antibodies; cinoma (R/M HNSCC) progressing on platinum-based chemotherapy have poor prognoses Monoclonal; and limited therapeutic options. Programmed cell death-1 (PD-1) and its ligand 1 (PD-L1) Immunotherapy; are frequently upregulated in HNSCC. The international, multi-institutional, single-arm, Head and neck cancer; phase II HAWK study (NCT02207530) evaluated durvalumab monotherapy, an anti-PD- HPV; L1 monoclonal antibody, in PD-L1-high patients with platinum-refractory R/M HNSCC. Recurrent or Patients and methods: Immunotherapy-naı̈ve patients with confirmed PD-L1-high tumour cell metastatic; expression (defined as patients with 25% of tumour cells expressing PD-L1 [TC  25%] using Survival rate the VENTANA PD-L1 [SP263] Assay) received durvalumab 10 mg/kg intravenously every 2 weeks for up to 12 months. The primary end-point was objective response rate; secondary end- points included progression-free survival (PFS) and overall survival (OS). Results: Among evaluable patients (n Z 111), objective response rate was 16.2% (95% con- fidence interval [CI], 9.9e24.4); 29.4% (95% CI, 15.1e47.5) for human papillomavirus (HPV)-positive patients and 10.9% (95% CI, 4.5e21.3) for HPV-negative patients. Median PFS and OS for treated patients (n Z 112) was 2.1 months (95% CI, 1.9e3.7) and 7.1 months (95% CI, 4.9e9.9); PFS and OS at 12 months were 14.6% (95% CI, 8.5e22.1) and 33.6% (95% CI, 24.8e42.7). Treatment-related adverse events were 57.1% (any grade) and 8.0% (grade 3); none led to death. At data cut-off, 24.1% of patients remained on treatment or in follow-up. Conclusion: Durvalumab demonstrated antitumour activity with acceptable safety in PD-L1- high patients with R/M HNSCC, supporting its ongoing evaluation in phase III trials in first- and second-line settings. In an ad hoc analysis, HPV-positive patients had a numerically high- er response rate and survival than HPV-negative patients. ª 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction for patients with previously treated R/M HNSCC with PD on or after a platinum-based therapy [13,14]. Recurrent/metastatic (R/M) head and neck squamous Durvalumab is a selective, high-affinity, engineered cell carcinoma (HNSCC) has poor prognosis and low human IgG1 monoclonal antibody (mAb) that blocks survival rates [1]. Until recently, patients with progres- PD-L1 binding to PD-1 and CD80, allowing T-cells to sive disease (PD) after first-line combination chemo- recognise and kill tumour cells (TC). Durvalumab has therapy and cetuximab were treated with shown antitumour activity in patients with HNSCC chemotherapeutic monotherapy, which yielded objective [15,16]. In a phase I/II study that included 62 patients response rates (ORRs) of 4e13% [2e5]. with R/M HNSCC, ORR with durvalumab was 11% in HNSCC tumours often exhibit T-cell infiltration and all patients, 18% in 22 patients with high PD-L1 expres- can be antigenic due to high mutation burden or virally sion (TC  25%) and 8% in 37 patients with PD-L1-low/ associated antigens, yet often escape immune elimina- negative expression (TC < 25%) [16]. Encouraging 6- and tion through inhibition of antitumour T-cell responses 12-month OS rates of 62% and 42%, respectively, were resulting from expression of checkpoint pathway com- observed in this pre-treated population [16]. ponents, such as programmed cell death-ligand 1 (PD- In this report, we present the safety and efficacy of L1) [6,7]. Targeting the programmed cell death-1 (PD- durvalumab from an international, multicentre study 1)/PD-L1 pathway has resulted in clinically meaningful (NCT02207530) in patients with R/M HNSCC with PD- activity and improved overall survival (OS) in patients L1-high expression (TC  25%) after progression after with previously treated R/M HNSCC [8e12]. In 2016, 2 only 1 platinum-containing regimen given in the R/M immuno-oncology agents targeting PD-1 were approved setting. 144 D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 2. Methods assessment according to local standard procedures and participating institutions’ individual standards; status 2.1. Study design may have been measured by p16 immunohistochemistry, fluorescence in situ hybridisation or polymerase chain In this single-arm, phase II trial, eligible patients reaction. received 10 mg/kg durvalumab by intravenous infusion Adverse events (AEs) were monitored every 2 weeks every 2 weeks for up to 12 months or until confirmed for the first 8 weeks and then every 4 weeks and graded PD, initiation of alternative cancer therapy, unaccept- according to National Cancer Institute CTCAE v4.03. able toxicity or consent withdrawal. The primary end- Safety monitoring continued for 90 d post-last dose. point was ORR using blinded independent central re- Treatment-related AEs (TRAEs) of special interest view (BICR) as measured by Response Evaluation (AESIs) were defined as AEs associated with potential Criteria in Solid Tumors (RECIST) v1.1 guidelines. For inflammatory or immune-related events. A suspected a complete or partial response, radiographic confirma- immune-mediated AE (imAE) was identified and tion was needed 4 weeks after first response. Second- defined as an AESI that required the use of systemic ary end-points were best objective response, duration of steroids or other immunosuppressants, and/or, for spe- response, time to response, disease control rate, cific endocrine events, endocrine therapy. All pneumo- progression-free survival (PFS) and OS; safety and nitis AEs were classified as suspected imAEs, regardless tolerability; and impact of treatment on symptoms and of concomitant steroid use. Patient-reported outcomes health-related quality of life (HRQoL) measures. (PRO) assessments are presented in the Appendix. 2.2. Patient population 2.4. Statistical methods Patients aged 18 years with histologically confirmed R/ The primary objective assessment was based on all M HNSCC (oral cavity, oropharynx, larynx or hypo- evaluable patients (evaluable analysis set), which pharynx) not amenable to therapy with curative intent included all treated patients who had baseline tumour (surgery or radiation therapy with or without chemo- assessments and measurable disease at baseline accord- therapy/biologic therapy) and with PD-L1-high ing to BICR and was set to determine whether the lower (TC  25%) expression were included. The PD-L1 limit of the 95% confidence interval (CI) of ORR was TC  25% cut-off was chosen for its ability to >13%. Secondary efficacy variables were analysed based discriminate between responders and non-responders in on evaluable analysis set and the full analysis set, which HNSCC [17,18]. Eligible patients had tumour progres- included all treated patients. Distributions of PFS and sion or recurrence during or after treatment with only 1 OS were estimated by KaplaneMeier method. platinum-based systemic regimen for R/M disease. Data underlying the findings described in this study Further eligibility criteria are provided in the Appendix. may be obtained in accordance with AstraZeneca’s data All patients provided written informed consent, and sharing policy described at https://astrazenecagroup any locally required authorisations were obtained from trials.pharmacm.com/ST/Submission/Disclosure. patient/legal representatives before any protocol-related procedures. 3. Results 2.3. Study assessments 3.1. Patient disposition and baseline characteristics Tumour response was assessed by computed tomogra- phy or magnetic resonance imaging every 8 weeks for The data cut-off date was 31st March 2017, approxi- the first 48 weeks and then every 12 weeks until mately 12 months after the last patient began treatment. confirmed PD. Tumour response (complete [CR], partial Of 112 patients treated, 111 were evaluable for efficacy [PR], stable disease [SD] or PD) was based on BICR assessment; 1 patient did not have measurable disease at according to RECIST v1.1. Patients with CR, PR or SD baseline according to BICR (Fig. A.1). At data cut-off, at 12 months entered follow-up. Upon progression at 21 patients (18.8%) had completed 12 months of treat- any time, asymptomatic patients or those without ment, and 27 (24.1%) remained on study on treatment or functional decline were permitted to restart durvalumab in follow-up. Ninety-one patients (81.3%) discontinued for up to 12 further months. treatment: 69.6% due to PD, 7.1% due to AEs and 4.5% PD-L1 expression levels of newly acquired or archival due to patient decision. tumour tissues (<3 years old) were assessed by immu- Demographic and disease characteristics of patients nohistochemistry with the VENTANA PD-L1 (SP263) are shown in Table 1. Median age of patients was 60 Assay (Ventana Medical Systems, Inc., AZ). years (range, 24e84 years), and 59.8% received prior Human papillomavirus (HPV) status was either cetuximab. Most patients were either current or former collected from historical medical records or based on (n Z 69; 61.6%) smokers. Among 99 patients evaluable D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 145 Table 1 leading to dose interruptions of durvalumab. One pa- Patient disposition and baseline characteristics. tient discontinued treatment due to grade 4 pneumo- Patient disposition N nitis, grade 3 hepatitis and grade 2 nephritis, all of which Patients screened, na 158 were considered related to durvalumab; this patient later Patients who failed screening, n 46 died due to PD. There were no TRAEs leading to death. Patients receiving durvalumab, n (%)b 112 (100) Most AEs were mild to moderate (grade 1 or 2). Alto- Evaluable patients, nc 111 Patients completing 12 months of treatment, n (%) 21 (18.8) gether, 64 patients (57.1%) reported at least 1 TRAE Patients who discontinued therapy, n (%) 91 (81.3) (Table 2). Grade 3/4 TRAEs occurred in 9 patients, of Progression, n (%) 78 (69.6) which 2.7% had increased gamma-glutamyltransferase. Baseline characteristicsb N Z 112 AESIs occurred in 39 patients: skin related (15.1%), Median age, years (range) 60 (24e84) endocrine related (12.5%), renal/hepatic related (10.7%), Male, n (%) 80 (71.4) ECOG PS, n (%)d 0 34 (30.4) Table 2 1 77 (68.8) Treatment-related adverse events.a Classification at study enrolment, n (%) Treatment-related adverse events Durvalumab (N Z 112) Locoregional recurrence 39 (34.8) Metastatic diseasee 73 (65.2) n (%) Primary tumour location Any treatment-related AE 64 (57.1) Oral cavity 47 (42.0) Treatment-related AEs 0 (0) Oropharynx 40 (35.7) leading to death Larynx 15 (13.4) Treatment-related AEs 1 (0.9) Hypopharynx 9 (8.0) leading to discontinuation Other 1 (0.9) Treatment-related AEs PD-L1 positive 112 (100) occurring in >5% of patients Smoking/nicotine status, n 102 Nausea 11 (9.8) Current, n (%) 10 (8.9) Fatigue 11 (9.8) Former, n (%) 59 (52.7) Hypothyroidism 10 (8.9) Never, n (%) 43 (38.4) Asthenia 9 (8.0) HPV/p16 status, nf 99 Pruritus 7 (6.3) Positive, n (%) 34 (34.3) Diarrhoea 6 (5.4) Oropharynx 20 (58.8) Decreased appetite 6 (5.4) Oral cavity 7 (20.6) Grade 3/4 treatment-related AEs 9 (8.0) Hypopharynx 2 (5.9) Gamma-glutamyltransferase 3 (2.7) Larynx 5 (14.7) increased Negative, n (%) 65 (65.7) Tumour pain 1 (0.9) Oropharynx 17 (26.2) Dehydration 1 (0.9) Oral cavity 33 (50.8) Hypokalaemia 1 (0.9) Hypopharynx 6 (9.2) Hypophosphatemia 1 (0.9) Larynx 8 (12.3) Pneumonitis 1 (0.9) Other 1 (1.5) Hepatitis 1 (0.9) Prior cetuximab, n (%) 67 (59.8) Chest pain 1 (0.9) ECOG PS, European Cooperative Oncology Group performance Alanine aminotransferase 1 (0.9) status; HPV, human papillomavirus; PD-L1, programmed cell death- increased ligand 1. Aspartate aminotransferase 1 (0.9) a Informed consent. increased b Full analysis set, n Z 112. Blood bilirubin increased 1 (0.9) c One patient did not have measurable disease at baseline according Treatment-related AEs of 39 (34.8) to blinded independent central review. special interest d One missing evaluation. Skin related e One patient had both locoregional recurrence and metastatic Rash 9 (8.0) disease. Dermatitis 8 (7.1) f HPV/p16 status was collected regardless of tumour site. Sites Endocrine related outside of oropharynx were tested. Hypothyroidism 11 (9.8) Hyperthyroidism 3 (2.7) Renal/hepatic related for HPV/p16 status, 34 (34.3%) were positive and 65 Select hepatic events 9 (8.0) Select renal events 3 (2.7) (65.7%) were negative. Oropharynx was the primary site Gastrointestinal related in most (59%) HPV-positive tumours. Diarrhoea 6 (5.4) Lung related 3.2. Safety Pneumonitis 4 (3.6) Other rare/miscellaneous 3 (2.7) Median duration of durvalumab treatment was 3.5 AE, adverse event. a months (range, 0.3e12.2). Nine patients had TRAEs As assessed by investigator. 146 D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 gastrointestinal related (5.4%), lung related (3.6%) or 4.4e20.9) among 65 patients with HPV-negative tu- other (2.7%) (Table 2). imAEs were observed in 18 pa- mours (Table 3). Among patients with HPV-positive tients treated with durvalumab; the majority were grade tumours, ORR with durvalumab was 30% (95% CI, 1/2 (hypothyroidism [11 patients], select hepatic events 11.9e54.3) for oropharyngeal primary site and 28.6% and hyperthyroidism [2 each], diarrhoea, select renal (95% CI, 8.4e58.1) for non-oropharyngeal site. events and rash [1 each]). Four patients had pneumo- Among patients with HPV-negative cancer, ORR was nitis, 1 of which was grade 3. 11.8% (95% CI, 1.5e36.4) for oropharyngeal site and 10.4% (95% CI, 3.5e22.7) for non-oropharyngeal site 3.3. Efficacy (Table 3). KaplaneMeier estimate of median PFS was 2.1 Eighteen patients treated with durvalumab had a months (95% CI, 1.9e3.7) (Fig. 1A). PFS at 6, 12 and 18 response (ORR, 16.2%; 95% CI, 9.9e24.4) (Table 3), months was 25.5% (95% CI, 17.6e34.1), 14.6% (95% CI, including 1 CR (0.9%) and 17 PR (15.3%). Three pa- 8.5e22.1) and 8.7% (95% CI, 2.7e19.2), respectively. tients (2.7%) had an unconfirmed CR/PR. Disease Median PFS in patients with HPV-positive and HPV- control rate (CR þ PR þ SD) at 24 weeks was 23.4%, negative tumours was 3.6 months (95% CI, 1.9e5.6) and 58 patients (52.3%) had PD during the first 12 versus 1.8 months (95% CI, 1.6e2.0), respectively. months of treatment. Median time to response was 2.0 At median follow-up of 6.1 months (range, 0.2e24.3), months (range, 1.6e9.2). At data cut-off, median median OS was 7.1 months (95% CI, 4.9e9.9) (Fig. 1B). duration of response was estimated to be 10.3 months OS at 12 and 18 months was 33.6% (95% CI, 24.8e42.7) and 10 of 18 patients (55.6%) had ongoing responses. and 23.0% (95% CI, 14.3e32.9), respectively. At data Decreases in size of target lesions occurred in 40.5% of cut-off, 27 patients (24.1%) remained on study treatment all patients (Fig. A.2). ORR was consistent across or in follow-up without progression. KaplaneMeier most subgroups evaluated, with the exception of HPV estimate of median OS in patients with HPV-positive status. and HPV-negative tumours was 10.2 months (95% CI, In an exploratory analysis testing HPV/p16 at all 7.2e16.3) versus 5.0 months (95% CI, 3.4e8.4), anatomical sites (i.e. not limited to oropharynx), ORR respectively (Fig. 1C). among 34 patients with HPV-positive tumours was 29.4% (95% CI, 15.1e47.5) and 10.8% (95% CI, 3.4. Patient-reported outcomes in symptoms and HRQoL Based on the European Organisation for Research and Table 3 Treatment of Cancer (EORTC) 30-item core QoL Best objective response rate with durvalumab treatment. questionnaire (QLQ-C30), PROs were improved over Best objective response Durvalumab baseline for several measures (Fig. 2A). Clinically (N Z 111) meaningful improvements, defined as an increase in ORR, n (%) 18 (16.2) baseline score of 10 at 2 consecutive assessments 14 95% CI 9.9e24.4 Complete response 1 (0.9) d apart, were observed for global health status/QoL Partial response 17 (15.3) (13.5%; 95% CI, 8.1e21.8), physical functioning (17.1%; Stable disease 7 (6.3) 95% CI, 10.1e27.6) and fatigue (21.3%; 95% CI, Unconfirmed complete/partial response 3 (2.7) 14.1e31.0) (Table A.1). To note, increased scores on the Progressive disease 58 (52.3) global health status and functioning scales indicate Median time to response, months (range) 2.0 (1.6e9.2) Median duration of response, months 10.3 better health status/function. Median time to deterio- ration was 3.7 months for global health status/QoL, 5.0 ORR by HPV/p16 status, n (%) HPV positive (n Z 34) 10 (29.4) months for physical functioning and 1.8 months for 95% CI 15.1e47.5 fatigue. Oropharynx (n Z 20) 6 (30.0) Head and neck cancerespecific symptoms evaluated Non-oropharynx (n Z 14) 4 (28.6) using the EORTC QoL questionnaire head and neck Oral cavity (n Z 7) 1 (14.3) cancer module (QLQ-H&N35) (Fig. 2B) showed clini- Larynx (n Z 5) 2 (40.0) Hypopharynx (n Z 2) 1 (50.0) cally meaningful improvements, defined as a decrease in HPV negative (n Z 65) 7 (10.8) baseline score of 10 at 2 consecutive assessments 14 95% CI 4.4e20.9 d apart, for mouth pain (24.6%; 95% CI, 16.0e36.0), Oropharynx (n Z 17) 2 (11.8) swallowing (19.4%; 95% CI, 11.4e30.9), taste and smell Non-oropharynx (n Z 48) 5 (10.4) (34.3%; 95% CI, 24.1e46.3) and speech (28.4%; 95% CI, Oral cavity (n Z 33) 3 (9.1) Hypopharynx (n Z 6) 1 (16.7) 19.7e39.0) (Table A.2). To note, decreased scores on Larynx (n Z 8) 1 (12.5) symptom scales represent symptom improvement. Me- Other (n Z 1) 0 (0) dian time to deterioration was 4.6 months for mouth CI, confidence interval; HPV, human papillomavirus; ORR, objective pain, 4.9 months for swallowing, 4.9 months for taste response rate. and smell and 2.7 months for speech. D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 147 Fig. 1. KaplaneMeier estimates of PFS (A) and OS (B). Exploratory analysis of KaplaneMeier estimates of OS by HPV status (C). HPV status was unknown for 13 patients (full analysis set; N Z 112). CI, confidence interval; HPV, human papillomavirus; OS, overall survival; PFS, progression-free survival. 148 D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 4. Discussion patients. There were no deaths due to TRAEs, and only 1 patient discontinued durvalumab because of a TRAE. 4.1. Conclusions Finally, improvements in PROs measured by both EORTC QLQ-C30 and QLQ-H&N35 scales were Durvalumab demonstrated clinically meaningful anti- observed with durvalumab treatment, suggesting tumour activity in patients with HNSCC with PD-L1- improved QoL for patients. high expression (TC  25%) who had progressed after first-line platinum-based therapy in the R/M setting. 4.2. Discussion ORR was 16.2% (95% CI, 9.9e24.4), with over half of patients (55.6%) maintaining their response at data cut- To our knowledge, this is the first large prospective off. A median OS of 7.1 months was observed, with study evaluating an anti-PD-L1 mAb in patients with R/ approximately one-third (33.6%) of patients surviving M HNSCC. While direct comparison to CHECK- 1 year. In an ad hoc analysis, HPV-positive patients MATE 141 or KEYNOTE-040 phase III studies is with HPV-positive tumours had a numerically higher limited based on differences in study design and inclu- ORR and OS than patients with HPV-negative sion criteria, including PD-L1 expression, it is important tumours. to note that, despite an ORR of 13.3% (95% CI, Durvalumab exhibited a manageable safety profile, 9.3e18.3) in CHECKMATE 141, nivolumab improved consistent with previous reports [16,19]. While most OS compared with standard-of-care therapy (hazard patients (57.1%) reported at least 1 TRAE, most were ratio, 0.70; 97.73% CI, 0.51e0.96, p Z 0.01) [10,12]. grade 1/2, with grade 3/4 TRAEs occurring in 8.0% of This is similar to other second-line trials in advanced Fig. 2. Patient-reported outcomes assessed by EORTC QLQ-C30 v3 (full analysis set; N Z 112) (A) and EORTC QLQ-H&N35 (full analysis set; N Z 112) (B). Error bars represent 95% CI for the mean change from baseline. CI, confidence interval; EORTC, European Organisation for Research and Treatment of Cancer; QLQ-C30, 30-item core QoL questionnaire; QLQ-H&N35, QoL questionnaire head and neck cancer module; QoL, quality of life. D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 149 Fig. 2. (continued). solid tumours, including bladder cancer, where, despite patients with HPV-positive tumours had higher ORR low response rates, treatment with single-agent anti-PD- and longer OS in KEYNOTE-012, where all patients 1 or PD-L1 mAbs led to significant OS improvement were PD-L1 positive (TC  1%) [9,11]. In an explor- compared with traditional chemotherapy [19]. The sta- atory subgroup analysis of PD-L1-negative patients tistical threshold of 95% CI for ORR was set at >13% in treated with nivolumab in CHECKMATE 141, patients this study. Although the lower limit for ORR was only with HPV-positive tumours had median OS of 10 9.9%, promising OS was observed with durvalumab in months versus 7.1 months for those with HPV-negative this platinum-pre-treated patient population. tumours [10]. It is possible that virally associated The ad hoc exploratory analysis showed that among tumour antigenicity in combination with intratumoural this PD-L1-high cohort, patients with HPV/p16-positive inflammation is predictive of response. Further research tumours (all tumour locations) treated with durvalumab is needed to determine whether HPV is independently experienced higher response rates and longer survival associated with improved treatment outcomes with anti- than patients with HPV-negative tumours. Conflicting PD-1 or anti-PD-L1 mAbs. results have been observed as to whether patients with In this study, non-oropharyngeal cancer patients HPV-positive tumours have improved efficacy with were included in the exploratory analysis of HPV status. blockade of the PD-1/PD-L1 pathway [9e11]; however, Interestingly, while HPV does not have a consistent or it is important to consider both PD-L1 expression and established role in oncogenesis or prognosis of non- HPV status together, given PD-L1 expression can be oropharyngeal sites, the response rate of non- predictive of efficacy. While analysis of KEYNOTE-055 oropharyngeal HPV-positive tumours was similar to with pembrolizumab showed similar ORR by HPV that of oropharyngeal HPV-positive tumours. Together, status, this analysis included both PD-L1-negative and all patients with HPV-positive tumours still had doubled PD-L1-positive patients [9]. Similar to our results, median OS compared with HPV-negative tumours. 150 D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 Reasons for this remain unclear, and further analysis is editing. AJ, PC, UK, SW, DZ, A Jarkowski, AA, GM, needed to determine whether HPV was biologically JA, JG, KA, LS, MF, NB, NR, RJ, SH, ZSP, JF, relevant in these non-oropharyngeal patients, or JGilbert and JFan reviewed the manuscript. whether the presence of HPV in non-oropharyngeal primary tumours has any immunologic effect that might Conflict of interest statement lead to better efficacy. Finally, HNSCC patients in whom toxicity from Dan P. Zandberg: Research Funding: AstraZeneca/ disease and treatment can often lead to long-term MedImmune (Inst), Bristol-Myers Squibb (Inst), Gli- impairment, disability and handicap, showed improve- knik (Inst), MacroGenics (Inst), Merck (Inst). ments in HRQoL assessments. The impact of durvalu- Alain P. Algazi: Consulting or Advisory Role: Array mab versus standard-of-care chemotherapy on patient Pharmaceuticals. Research Funding: Aceria (Inst), QoL will be assessed in 2 randomised phase III trials AstraZeneca/MedImmune (Inst), Bristol-Myers Squibb [20,21]. (Inst), Genentech (Inst), Incyte (Inst), Merck (Inst), This study has 2 design features that warrant careful Novartis (Inst), OncoSec (Inst). interpretation of its results. First, as a non-randomised Antonio Jimeno: Consulting or Advisory Role: Astra- single-arm study, it lacks a standard-of-care control Zeneca. Research Funding: AstraZeneca/MedImmune arm. Second, the patient population had only PD-L1- (Inst). high-expressing tumours, compared with earlier studies James S. Good: Honoraria: Bristol-Myers Squibb, of patients with tumours of PD-L1-high and PD-L1- BTG, Eisai, Merck, Sirtex. Consulting or Advisory Role: low/negative status; the absence of PD-L1-low/negative GenesisCare UK. Research Funding: AstraZeneca. tumours allowed focused analysis on this particular Travel, Accommodations, Expenses: BTG, GenesisCare subgroup of patients that may be more likely to respond UK, Sirtex. to anti-PD-L1 mAb therapy. A companion phase II Jérôme Fayette: Honoraria: AstraZeneca, Bristol- study, CONDOR, has assessed responses to durvalu- Myers Squibb, Merck. Consulting or Advisory Role: mab monotherapy in a PD-L1-low/negative HNSCC Bristol-Myers Squibb. Travel, Accommodations, Ex- population (NCT02319044) [22]. The variety of tech- penses: AstraZeneca, Bristol-Myers Squibb. nologies, assays and cut-off algorithms used to deter- Nathaniel Bouganim: No relationship to disclose. mine PD-L1 expression make comparisons across Neal E. Ready: Consulting or Advisory Role: AbbVie, different trials challenging. Another limitation is that AstraZeneca, Bristol-Myers Squibb, Celgene, Merck. HPV status was determined by different methods at the Paul M. Clement: Consulting or Advisory Role: Abb- discretion of each participating site. Vie (Inst), AstraZeneca (Inst), Bristol-Myers Squibb Overall, encouraging efficacy was observed with (Inst), Leo Pharma (Inst), MSD (Inst), Vifor Pharma single-agent durvalumab in PD-L1-high R/M HNSCC (Inst). Research Funding: AstraZeneca (Inst), MSD patients, most of whom had failed cetuximab in addition (Inst). Patents, Royalties, Other Intellectual Property: to platinum-based chemotherapy. Currently, 2 ongoing NIH-US government (Inst). phase III trials (EAGLE NCT02369874; KESTREL Caroline Even: Consulting or Advisory Role: Astra- NCT02551159) will evaluate durvalumab with or Zeneca, Bristol-Myers Squibb, Innate Pharma, Merck without tremelimumab in patients in the first- and Serono, MSD. Travel, Accommodations, Expenses: second-line settings for R/M HNSCC [20,21]. Merck Serono. Raymond W. Jang: Honoraria: Ipsen. Research Author contributions Funding: AstraZeneca (Inst), Boston Biomedical (Inst), Bristol-Myers Squibb (Inst), Lilly (Inst), Merck (Inst), DZ, A Jarkowski, GM, JA and LS provided study Novartis (Inst). concepts. DZ, A Jarkowski, GM, JA, KA, NR, JF and Stuart Wong: No relationship to disclose. JFan designed the study. PC, AJ, UK, SW, AA, JG, LS, Ulrich Keilholz: Honoraria: Amgen, AstraZeneca, MF, NB, NR, RJ, SH, ZSP, JF and JGilbert helped in Bristol-Myers Squibb, GSK, Merck, Pfizer. Consulting data acquisition. JA, KA and JFan assisted in quality or Advisory Role: AstraZeneca, Bristol-Myers Squibb, control of data and algorithms. PC, AJ, UK, SW, DZ, Merck, Pfizer. Speakers’ Bureau: Amgen, AstraZeneca, A Jarkowski, AA, GM, JA, JG, KA, LS, MF, NB, NR, Bristol-Myers Squibb, Merck, Pfizer. Research Funding: RJ, SH, ZSP, JF, JGilbert and JFan took part in data AstraZeneca, Merck, Pfizer. Travel, Accommodations, analysis and interpretation. JA and KA helped in sta- Expenses: AstraZeneca, Bristol-Myers Squibb, Merck. tistical analysis. PC, SW, DZ, A Jarkowski, AA, GM, Jill Gilbert: Honoraria: TRM Oncology. Consulting JA, KA, LS, NB, NR, ZSP and JF contributed to or Advisory Role: AstraZeneca, TRM Oncology. manuscript preparation. AJ, UK, SW, DZ, A Jarkow- Research Funding: AstraZeneca (Inst), Bristol-Myers ski, AA, GM, JA, JG, KA, LS, MF, NB, NR, RJ, SH, Squibb (Inst), Merck (Inst), Pfizer (Inst). Travel, ZSP, JF, JGilbert and JFan involved in manuscript Accommodations, Expenses: TRM Oncology. D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 151 Moon Fenton: Consulting or Advisory Role: compared with intravenous methotrexate for recurrent squamous AstraZeneca. cell carcinoma of the head and neck [corrected]. J Clin Oncol 2009;27:1864e71. Irene Braña: Spanish Ministry of Science e Health [5] Kushwaha VS, Gupta S, Husain N, Khan H, Negi MP, Jamal N, Research and Development Strategy Rio Hortega Grant. et al. Gefitinib, methotrexate and methotrexate plus 5-fluorouracil Stephanie Henry: No relationship to disclose. as palliative treatment in recurrent head and neck squamous cell Eva Remenar: No relationship to disclose. carcinoma. Cancer Biol Ther 2015;16:346e51. Zsuzsanna Papai: No relationship to disclose. [6] Chau NG, Li YY, Hung YP, Hammerman PS, Rodig SJ, Schoenfeld JD, et al. Analysis of immune infiltrates in a genomically Lillian L. Siu: Consulting or Advisory Role: AstraZe- characterized clinical cohort of head and neck squamous cell car- neca. Research Funding: AstraZeneca (Inst). cinoma (HNSCC) patients (pts). J Clin Oncol 2016;34. abstr 6052. Anthony Jarkowski: Employment: AstraZeneca, [7] Allen CT, Clavijo PE, Van Waes C, Chen Z. Anti-tumor immu- Bristol-Myers Squibb. Stock or Other Ownership: nity in head and neck cancer: understanding the evidence, how AstraZeneca. tumors escape and immunotherapeutic approaches. Cancers (Basel) 2015;7:2397e414. Jon M. Armstrong: Employment and Stock or Other [8] Chow LQ, Haddad R, Gupta S, Mahipal A, Mehra R, Tahara M, Ownership: AstraZeneca. et al. Antitumor activity of pembrolizumab in biomarker- Kobby Asubonteng: Employment and Stock or Other unselected patients with recurrent and/or metastatic head and Ownership: AstraZeneca. neck squamous cell carcinoma: results from the phase Ib Giovanni Melillo: Employment and Stock or Other KEYNOTE-012 expansion cohort. J Clin Oncol 2016;34:3838e45. [9] Bauml J, Seiwert TY, Pfister DG, Worden F, Liu SV, Gilbert J, Ownership: AstraZeneca. et al. Pembrolizumab for platinum- and cetuximab-refractory Ricard Mesı́a: Consulting or Advisory Role: AstraZe- head and neck cancer: results from a single-arm, phase II study. neca, Bayer, Bristol-Myers Squibb, Merck Serono, J Clin Oncol 2017;35:1542e9. Merck Sharp & Dohme. Speakers’ Bureau: AstraZe- [10] Ferris RL, Blumenschein Jr G, Fayette J, Guigay J, Colevas AD, neca, Bristol-Myers Squibb, Merck Serono. Licitra L, et al. Nivolumab for recurrent squamous-cell carcinoma of the head and neck. N Engl J Med 2016;375:1856e67. [11] Seiwert TY, Burtness B, Mehra R, Weiss J, Berger R, Eder JP, Acknowledgments et al. Safety and clinical activity of pembrolizumab for treatment of recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-012): an open-label, multicentre, phase 1b The authors would like to thank the patients, their trial. Lancet Oncol 2016;17:956e65. families and caregivers, and all investigators involved in [12] Cohen EE, Harrington KJ, Le Tourneau C, Dinis J, Licitra L, this study. The authors would also like to acknowledge Ahn M-J, et al. Pembrolizumab (pembro) vs standard of care the contribution of Sophie Wildsmith to biomarker (SOC) for recurrent or metastatic head and neck squamous cell carcinoma (R/M HNSCC): phase 3 KEYNOTE-040 trial. Ann analyses. Medical writing support, which was in accor- Oncol 2017;28. abstr LBA45_PR. dance with Good Publication Practice (GPP3) guide- [13] OPDIVO (nivolumab) [prescribing information]. Princeton, NJ: lines, was provided by Jubilee Stewart, PhD, of Bristol-Myers Squibb; 2018. PAREXEL (Hackensack, NJ, USA) and was funded by [14] KEYTRUDA (pembrolizumab) [prescribing information]. AstraZeneca. Whitehouse Station, NJ: Merck & Co, Inc.; 2018. [15] Segal NH, Hamid O, Hwu W, Massard C, Butler M, Antonia SJ, et al. A phase 1 multi-arm dose-expansion study of Appendix A. Supplementary data the anti-programmed cell death-ligand-1 (PD-L1) antibody MEDI4736: preliminary data. Ann Oncol 2014;25. iv361e72 Supplementary data to this article can be found online (abstr 1058PD). at https://doi.org/10.1016/j.ejca.2018.11.015. [16] Segal NH, Ou S-HI, Balmanoukian AS, Massarelli E, Brahmer JR, Weiss J, et al. Updated safety and efficacy of dur- valumab (MEDI4736), an anti-PD-L1 antibody, in patients from References a squamous cell carcinoma of the head and neck (SCCHN) expansion cohort (abstr 949O). Ann Oncol 2016;27:328e50. [1] Leon X, Hitt R, Constenla M, Rocca A, Stupp R, Kovacs AF, [17] Rebelatto MC, Midha A, Mistry A, Sabalos C, Schechter N, et al. A retrospective analysis of the outcome of patients with Li X, et al. Development of a programmed cell death ligand-1 recurrent and/or metastatic squamous cell carcinoma of the head immunohistochemical assay validated for analysis of non-small and neck refractory to a platinum-based chemotherapy. Clin cell lung cancer and head and neck squamous cell carcinoma. Oncol (R Coll Radiol) 2005;17:418e24. Diagn Pathol 2016;11:95. [2] Vermorken JB, Trigo J, Hitt R, Koralewski P, Diaz-Rubio E, [18] Wildsmith S, Scott M, Midha A, Barker C, Whiteley J, Rolland F, et al. Open-label, uncontrolled, multicenter phase II Ratcliffe M, et al. PD-L1 expression in patients screened for phase study to evaluate the efficacy and toxicity of cetuximab as a single 2 head and neck squamous cell carcinoma clinical studies agent in patients with recurrent and/or metastatic squamous cell (HAWK and CONDOR). Presented at: AACR annual meeting; carcinoma of the head and neck who failed to respond to April 14-17; 2018. Chicago, IL. Abstr 5530. platinum-based therapy. J Clin Oncol 2007;25:2171e7. [19] Massard C, Gordon MS, Sharma S, Rafii S, Wainberg ZA, [3] Zenda S, Onozawa Y, Boku N, Iida Y, Ebihara M, Onitsuka T. Luke J, et al. Safety and efficacy of durvalumab (MEDI4736), an Single-agent docetaxel in patients with platinum-refractory met- anti-programmed cell death ligand-1 immune checkpoint inhibi- astatic or recurrent squamous cell carcinoma of the head and neck tor, in patients with advanced urothelial bladder cancer. J Clin (SCCHN). Jpn J Clin Oncol 2007;37:477e81. Oncol 2016;34:3119e25. [4] Stewart JS, Cohen EE, Licitra L, Van Herpen CM, [20] ClinicalTrials.gov. Study of MEDI4736 monotherapy and in Khorprasert C, Soulieres D, et al. Phase III study of gefitinib combination with tremelimumab versus standard of care therapy 152 D.P. Zandberg et al. / European Journal of Cancer 107 (2019) 142e152 in patients with head and neck cancer (EAGLE). https://www. [22] Siu LL, Even C, Mesia R, Daste A, Krauss J, Saba NF, et al. A clinicaltrials.gov/ct2/show/NCT02369874?termZNCT02369874& randomized, open-label, multicenter, global phase 2 study of rankZ1. [Accessed October 25 2018]. durvalumab (D), tremelimumab (T), or D plus T, in patients with [21] ClinicalTrials.gov. Phase III open label study of MEDI PD-L1 low/negative recurrent or metastatic head and neck 4736 with/without tremelimumab versus standard of care (SOC) squamous cell carcinoma: CONDOR [abstract]. Presented at: in recurrent/metastatic head and neck cancer (KESTREL). Multidisciplinary head and neck cancer symposium; February https://www.clinicaltrials.gov/ct2/show/NCT02551159? 15e17; 2018. Scottsdale, AZ. termZNCT02551159&rankZ1. [Accessed October 25 2018].