The majority of the subcommittee recommended treatment with anti-PD-1 monotherapy (82%) for R/M HNSCC patients who are platinum-refractory including those who progressed within six months of platinum-based chemotherapy. Alternatively, if a clinical trial is available, the majority (94%) found this to be the preferred option, especially if it is a biomarker-based, hypothesis-driven clinical trial (59%).

For R/M, platinum-refractory patients with good performance status (ECOG PS = 0/1) the subcommittee recommends enrollment in a clinical trial as their first treatment choice followed by treatment with anti-PD-1 monotherapy as second choice, and radiotherapy, chemotherapy, or EGFR targeted therapy as third choice. For R/M, platinum-refractory patients with poor performance status (ECOG PS = 2) the subcommittee recommends anti-PD-1 monotherapy as their first choice of treatment, EGFR targeted therapy as second choice, and radio and/or chemotherapy as third choice. The committee noted several factors that may affect clinical decision making with respect to use of single agent PD-1 therapy for platinum-refractory R/M HNSCC. In the setting of rapidly growing disease, 59% recommended treatment with cytotoxic chemotherapy. For these patients, docetaxel was chosen more frequently than methotrexate, paclitaxel or carboplatin-5FU, or cetuximab. In appropriate patients who progress on IO monotherapy, cetuximab + platinum-based therapy may still be offered as a treatment alternative for fit patients in the second-line setting. We note that there is no prospective evidence to support the use of ICI in combination with cytotoxic chemotherapy in the platinum-refractory setting.

Additional predictive biomarkers for immunotherapy strive to characterize the functional state of immune cells in the tumor microenvironment. Towards this goal, immune gene expression profile (GEP) scores via analysis of extracted RNA, characterize the “T cell activation status” in the tumor microenvironment, and have been observed to predict anti-PD-1 efficacy across various tumor types, including HNSCC [14, 61–64]. A composite score based on interferon gamma and five interferon gamma-related genes (CXCL9, CXCL10, IDO1, HLA-DRA, and STAT1) significantly correlated with response rate as well as PFS in patients with R/M HNSCC treated with pembrolizumab in KEYNOTE-012. Specifically, patients with a score above the Youden index - a measure of the effectiveness of a diagnostic test [65] - displayed ORR of 40% compared to 5% for those who scored below the index (95% negative predictive value) [14]. Furthering these results, in a cohort of 258 patients with R/M HNSCC treated with pembrolizumab in KEYNOTE-012 and -055, a high GEP score based on 18 genes was significantly and independently associated with increased response, PFS, and OS [66]. High tumor mutational burden (TMB), which has been associated with better efficacy with treatment with ICIs in numerous solid tumors [67–69], was predictive of response and PFS but not OS in this analysis. Importantly, the predictive value of GEP and TMB were each independently associated with response, and responses were highest in patients that had both high GEP and TMB [66].

KEYNOTE-048 provided category 1 evidence for PD-L1 expression analysis by CPS ≥20 and ≥ 1, showing improved overall survival for such patients upon treatment with pembrolizumab monotherapy. The subsequent approval of pembrolizumab for use as a single agent for patients whose tumors express PD-L1 CPS ≥ 1 marks the first mandated biomarker testing of patients with R/M HNSCC tumors.

Ninety-four percent of the subcommittee defined positivity for PD-L1 as ≥1% TPS or ≥ 1 CPS by IHC staining. However, it is important to note that expression levels may differ depending on the antibody used and whether staining includes tumor alone (TPS) or tumor plus stroma (CPS). The majority of the subcommittee (81%) also agreed that the best use of biomarker testing when treating patients with HNSCC with immunotherapy is by combined positive score (CPS). The remaining 3 subcommittee members did not recommend any specific form of biomarker testing for patients with R/M HNSCC prior to administration of approved immunotherapies since no testing is currently recommended or required by the FDA (pending FDA approvals for treatment related to KN-048 data). One such subcommittee member noted that CPS has more predictive value in retrospective analysis and will become the biomarker of choice when pembrolizumab monotherapy is approved by the FDA.

While pembrolizumab received FDA approval in 2017 for the treatment of adult and pediatric patients with unresectable or metastatic solid tumors that have been identified as being microsatellite instability-high (MSI-H) [70], the frequency of MSI-H HNSCC tumors is relatively low, at about 1–3% [71, 72]. Given the low rate of MSI incidence in HNSCC, the subcommittee (88%) recommended against standard MSI testing, unless the patient is having a genome profile performed already which will provide such information. TMB analysis is not currently recommended by the FDA in HNSCC.

In KEYNOTE-012, patients were required to provide tissue tumor biopsy samples for analysis [14]. Patients were classified as having HPV+ disease if the primary location of their tumor was in the oropharynx and at least 70% of cells stained moderately or strongly positive for p16 by IHC [47]. Of the patients in the head and neck cohorts, the percentage of HPV+ patients was relatively small with 45 (23%) being HPV+ and 147 (77%) being HPV- [74]. When stratified by HPV status, response rates were higher in HPV+ patients compared to HPV- patients, with demonstrated ORRs of 24% (95% CI, 13–40%)and 16% (95% CI, 10–23), respectively [14, 47]. Response duration ranging above 60 weeks was observed among both HPV+ and HPV- patients [14]. For patients enrolled in KEYNOTE-012 irrespective of PD-L1 status, ORR was 32% (9/28 patients) and 14% (15/104 patients) among patients with HPV+ and HPV- disease, respectively [42].

Furthermore, the phase II HAWK trial examined the efficacy of durvalumab monotherapy vs. SOC chemotherapy in immunotherapy-naïve patients with R/M HNSCC with high tumor PD-L1 expression. In an exploratory analysis by HPV status, patients with HPV-positive oropharyngeal squamous cell carcinoma demonstrated an ORR of 30%, compared with 10.8% in those who were HPV- [54].

Patients enrolled in the CheckMate 141 study were analyzed post-hoc by HPV status as determined by p16 staining. Here, 63 (26%) patients were HPV+, 50 (21%) were HPV-, 127 (53%) were not tested [75]. Analyses revealed a benefit with nivolumab compared to SOC chemotherapy, irrespective of HPV status (HPV- patients, HR 0.73, 95% CI: 0.42–1.25; HPV+ patients, HR 0.56, 95% CI 0.32–0.99) [43]. A recent update confirmed a consistent benefit of nivolumab compared to SOC across both HPV+ and HPV- patients (HPV- patients, HR 0.59, 95% CI: 0.38–0.92; HPV+ patients, HR 0.60, 95% CI: 0.37–0.97) [48].

Overall, HPV status should not affect selection of patients with platinum-refractory R/M HNSCC for ICI therapy. Specifically, while 55.5% of the subcommittee stated that HPV status (based on p16 overexpression) should be included in treatment planning, 83% voted that it does not influence their decision to treat patients with R/M HNSCC with SOC immunotherapy. For their reasoning, the subcommittee noted a lack of strong data suggesting p16+ patients experience a distinct benefit and that data thus far indicate that both p16+ and p16- populations benefit from available checkpoint inhibitors.

Patients enrolled in the CheckMate 141 trial were treated until disease progression or an unacceptable level of toxicity occurred [43]. Patients were evaluated for response using the RECIST 1.1 criteria every six weeks starting nine weeks post-treatment initiation. CheckMate 141 described clinical outcomes by best OR for the anti-PD-1 and the SOC arms. The trial was stopped early due to a survival advantage for patients treated in the experimental arm with a statistically significant hazard ratio of 0.70 [43]. A report from the AACR 2017 annual meeting found that some patients treated beyond progression with nivolumab demonstrated clinical benefit. In this study, 139 of 240 patients (58%) randomized to nivolumab experienced disease progression according to RECIST v1.1 definitions. Among those who experienced PD, 41% were treated beyond progression for a mean duration of 2.0 months. Median OS was 12.7 months in this group (95% CI 9.7, NR). Moreover, after initial progression, 23% of patients treated beyond progression experienced a reduction in target lesion size, including 2 patients (5%) with over 30% reduction in tumor size [87].

During KEYNOTE-012 and KEYNOTE-040, pembrolizumab was administered for 24 months or until disease progression, unacceptable toxicity or investigator decision to halt treatment. In KEYNOTE-040, imaging occurred using RECIST v1.1 first at nine weeks and then every six weeks following. Modified RECIST methods were also used to account for distinct responses due to treatment with pembrolizumab. This version of RECIST allowed for continued treatment after initial radiographic progression until confirmation imaging at least four weeks later. PD in this trial was defined as ≥20% increase in the sum of diameters of target lesions and had to demonstrate an absolute increase of ≥5 mm. The appearance of at least one new lesion was also considered progression. Radiographic responses were analyzed in real time for confirmation of PD by RECIST v1.1. Survival follow-up occurred every 12 weeks [44]. Of note, CheckMate 153 assessed patients with advanced NSCLC who completed 1 year of nivolumab therapy and were subsequently randomized to continuous nivolumab or to stopping nivolumab, with the option to reinitiate therapy in the case of disease progression. This trial reported that continuous treatment with nivolumab until PD was associated with superior PFS compared with a 1-year fixed duration treatment (hazard ratio [HR], 0.43; 95% CI, 0.25–0.76) [88].

Moreover, data from R/M HNSCC patients treated in four French centers (Antoine Lacassagne Center, Nice; Léon Bérard Center, Lyon; Curie Institute, Paris; Gustave Roussy, Villejuif) between September 2012 and September 2015 were retrospectively collected to investigate tumor growth kinetics post-treatment with PD-1/PD-L1 inhibitors. Among the 64 R/M HNSCC patients identified, 34 were eligible for analysis. Images for pre-baseline and during immunotherapy analysis were retrospectively reviewed to assess the ORR and the PFS according to RECIST 1.1 and irRECIST. Patterns of recurrence included loco-regional recurrence in 14 patients, distant metastases in 11 patients, and occurrence of both in 9 patients. Hyperprogression was observed in ten patients (29%) total, including 9/23 patients with at least a loco-regional recurrence and only 1/11 patients with exclusively distant metastases. Hyperprogression was significantly correlated with a shorter PFS according to RECIST (P = 0.003) and irRECIST (P = 0.02), but not with OS (P = 0.77). No pseudo-progression events were reported [79]. Additionally, 1 out of 104 patients recruited to the KEYNOTE-012 trial was reported to have pseudoprogression [14].

While category 1 evidence does not exist here, all studies demonstrating efficacy of anti-PD1 have used RECIST v1.1 and this version continues to be used in most current immunotherapy clinical trials [89].

The subcommittee was split in their recommendation of RECIST (56%) versus irRECIST (44%) to assess response in patients with HNSCC being treated with immunotherapy. Moreover, the correct timeframe and interval for patient monitoring during and after immunotherapy treatment is not completely understood. Such monitoring is crucial to promptly identify and manage signs of progression, symptoms and adverse events. 53% of the subcommittee recommends a one-month timeframe for initial clinical follow-up for identification of signs of immune-related symptoms and AEs. For continued identification of signs of immune-related symptoms and AEs, the subcommittee recommends patients to be evaluated at least monthly, and sometimes more frequently in the setting of active AEs (71%). In monitoring patients for signs of response after initial follow-up, the majority of the subcommittee (65%) recommends patient evaluation (via radiographic imaging) every three months with SOC imaging to be adapted to patient disease status, response, and tolerability of the regimen.

In determining duration of treatment in the case of a patient experiencing a CR or near CR after treatment with anti-PD1 therapy, 53% of the subcommittee recommend continuing treatment for at least two years (up to indefinitely) or until the patient experiences disease progression or toxicity. Only 20% of the subcommittee recommended stopping immunotherapy at this point and monitoring until progression. For patients with HNSCC who enter disease remission on immunotherapy, the subcommittee was split on continuing treatment for a total duration of one (40%) or two years (40%) given no disease progression or toxicity.

Regarding efficacy metrics, the subcommittee unanimously agreed that describing study outcomes in terms of ORR and OS benefit after treatment with checkpoint blockade therapy in patients with HNSCC is sufficient for future therapeutic consideration. Forty-one percent of the subcommittee further clarified their responses by noting that they would favor OS over ORR. For initial assessment, the subcommittee recommends using either a CT (53%) or PET-CT (41%) scan following a baseline clinical exam of the patient. To best capture the dynamics of changing tumor size, the subcommittee recommends imaging, particularly utilizing a CT scan (44%).

Furthermore, the subcommittee (88%) agreed that it is not acceptable to treat beyond progression if a patient has symptomatic progression/clinical deterioration. In the event radiographic progression is observed early in treatment, and the patient is clinically stable, the majority of the subcommittee (76%) recommends continuing immunotherapy treatment until progression is confirmed on a second scan. Of note, subcommittee members stated that this recommendation to continue immunotherapy until a second scan confirms progression may be modified depending on clinical trial options available for 2nd or 3rd line treatment as well as the specific characteristics and kinetics of the patient’s disease such as PD-L1 expression, prior therapies, disease burden, or rapid progression with high symptom burden. As for treating patients with HNSCC with disease progression on or after treatment with a PD-1 inhibitor, 81% of the subcommittee recommended enrollment in a clinical trial and 78% recommended treating with palliative radiotherapy and/or chemotherapy. Of those who recommended chemotherapy, 57% specifically recommended treatment with a taxane.

With respect to tumor characteristics influencing treatment, 56% of the subcommittee agreed that anatomical site of the tumor is an important consideration. For instance, potential for airway obstruction, surgical resection or radiotherapy to the site may alter the course of treatment. Similarly, the committee was split (44% versus 44%, with 12% undecided) regarding the notion that patients with bulkier tumors were less likely to respond to immunotherapy.

As noted above, cancer treatment-response kinetics observed in immunotherapy patients may differ from those in patients treated with chemotherapy, radiotherapy, surgery and targeted therapies. This unique response can be seen in patients after checkpoint blockade in a number of ways, including pseudoprogression, hyperprogression, and delayed response. 94.1% of the subcommittee agreed that the term “pseudoprogression” should be avoided in a setting of worsening symptoms. 70.6% of the subcommittee recommended defining hyperprogression as “a rapid increase in tumor growth rate (minimum two-fold) compared to the expected growth rate.” Alternatively, other accepted definitions of hyperprogression found in the literature include: a “RECIST progression at the first evaluation and as a ≥2-fold increase of the tumor growth rate between the reference and the experimental periods” [90], “time-to-treatment failure (TTF) < two months, >50% increase in tumor burden compared to pre-immunotherapy imaging, and >two-fold increase in progression pace” [91], “a tumor growth kinetics ratio equal to or greater than two” [79].

In CheckMate 141, toxicities were assessed using the National Cancer Institute Common Terminology Criteria for Adverse events (NCI CTCAE) version 4.0 throughout the study and treatment was discontinued when appropriate [43]. The long-term follow-up conducted by this study demonstrates the lasting effects and safety profile of nivolumab regardless of PD-L1 status. AEs were measured at each treatment visit and for 100 days after final infusion. Potentially immunologic AEs were classified as select AEs [43]. A reduced rate of TRAEs (59% vs 78%, respectively) and grade 3/4 AEs (13% vs. 35%, respectively) were observed in patients who received nivolumab compared to patients who were treated with chemotherapy. Endocrine disorders, primarily hypothyroidism, occurred in 8% of patients. Among irAEs, gastrointestinal events were less common during nivolumab compared to SOC (6.8% vs. 14.4%). Other irAEs such as rash, pruritus and hypothyroidism, however, were higher in the nivolumab-treated group compared to SOC [43].

In KEYNOTE-012, laboratory safety tests were conducted within ten days of treatment initiation and within 72 h of each subsequent dose [14, 42]. Serious AEs were monitored for 90 days after the last dose of study treatment up to 34 months and non-serious AEs were monitored for 30 days post-last dose of study treatment up to 32 months [14, 42, 47]. Treatment was halted in the event of a grade 4 toxicity, or a grade 3 toxicity that did not resolve within 12 weeks of the last treatment dose. Overall, treatment with pembrolizumab was well-tolerated. In the expansion cohort of KEYNOTE-012, only 17% of patients receiving pembrolizumab experienced grade 3/4 adverse events, most commonly alanine aminotransferase and aspartate aminotransferase elevations, as well as hyponatremia. Twenty percent of these patients experienced an irAE, most of which were considered grade 1/2. Grade 3 irAEs included pneumonitis, diabetes mellitus, decubitus ulcer, colitis, and liver injury. Three patients (one each: G3 pneumonitis, G3 colitis, and G2 interstitial lung disease) discontinued treatment [42].

In KEYNOTE-040, AEs and laboratory irregularities were collected throughout treatment and for 30 days following treatment for low grade toxicities and 90 days following treatment for serious AEs. Toxicities were graded using the NCI CTCAE version 4.0. In this study, fewer patients treated with pembrolizumab experienced grade ≥ 3 TRAEs compared with SOC (13% vs 36%). The most common pembrolizumab-associated TRAE was hypothyroidism (13%) and fatigue with SOC (18%). Treatment-related deaths occurred in four patients who received pembrolizumab (unspecified cause, large intestine perforation, malignant neoplasm progression, and Stevens-Johnson syndrome) and two patients who received SOC (malignant neoplasm progression and pneumonia) [44].

Some investigators have raised concerns in clinical trial design that include the quality and completeness of irAE reporting, noting time to occurrence, and constraints due to limited follow up periods [101]. For instance, some studies have only required safety assessment up to 30 days after the last dose, whereas other studies explicitly exclude reporting of adverse events that occur more than 30 days after the last dose, or after starting another cancer treatment [101].

Of note, while most irAEs appear to occur during immunotherapy [1, 3–13, 15, 16, 19], there is growing evidence to suggest the existence of post-immunotherapy irAEs, which occur months or years after treatment discontinuation [16–18, 30, 102]. With an increasing number of neoadjuvant/adjuvant IO trials currently being conducted in the definitive/curative setting, it will be necessary to recognize this emerging clinical entity and perhaps adjust follow-up and reporting times.

The subcommittee discussed when to change clinical management of patients treated with IO therapies based on irAEs. The subcommittee felt that general management of head and neck cancer toxicity is aligned with the practical management of irAEs in other solid tumor types and provided recommendations concerning immune-related toxicities specific to head and neck cancer. For further detail into toxicity management strategies please refer to the NCCN Clinical Practice Guidelines in Oncology: Management of Immunotherapy-Related Toxicities (2019) [103].

The majority of the subcommittee (76%) recommends evaluating patients with HNSCC treated with checkpoint blockade for signs of adverse events at least once monthly during the course of treatment. For those patients who develop an irAE < grade 3, the majority of the subcommittee recommends continuing ICIs for grade 1 events with the exception of some neurologic, hematologic or cardiac toxicities. For grade 2 events, the subcommittee recommends stopping IO therapy and providing closely monitored outpatient treatment, including consideration of oral steroids. For irAE development ≥ grade 3, the majority of the subcommittee recommends admitting the patient to the hospital (79%), administering steroids (77%), and halting treatment (67%).

The majority of the subcommittee recommended routine monitoring of thyroid function (94%), neck and airway through imaging (62.5%), and AST/ALT levels (75%). Lipase evaluation was recommended by 44% of the subcommittee, while brain imaging was only recommended by 6% of members. The subcommittee was split on whether whole-body imaging is necessary during treatment.

Concerning thyroid function, the subcommittee recommends routine patient evaluation during treatment with anti-PD-1 agents. In patients that develop hypothyroidism, the majority of the subcommittee (75%) recommended continuing immunotherapy, providing levothyroxine for management, and evaluating thyroid function in two-month intervals.

Moreover, 67% of the subcommittee agreed that pneumonitis is not a greater concern in immunotherapy patients with HNSCC compared to other cancers. However, the subcommittee did suggest that some patients with HNSCC may be at a higher risk of developing pulmonary problems such as those already aspirating, or patients with previous radiation to the thorax.

In the event of bulky, progressing disease leading to organ dysfunction or compromise, 53% of the subcommittee recommended halting immunotherapy. Subcommittee members noted that evaluation of the underlying condition is necessary to determine whether compromising bulky disease is due to tumor progression versus inflammation from immunotherapy, or other cause. Medical oncologists should obtain surgical consultation early in their treatment course for patients with a tenuous airway or vascular encasement, even in the setting of clinical response to immunotherapy. Clinicians should be aware that delayed irAEs may occur months after discontinuation of immunotherapy and patients should be observed for such toxicities indefinitely [104–107].

Finally, there are occasional unique clinical scenarios. For instance, patients with some HNSCC- or immune-related comorbidities were not included in the clinical trials and the effect of steroid treatment on immune response and clinical activity of immunotherapy is not well understood. Thus, safety considerations for treating a patient with bulky disease, which may compromise the airway and/or potentiate vascular blowout, are not well characterized. Furthermore, complete response of large tumors after immunotherapy that result in exposure of major vessels is not well described and may necessitate reconstructive interventions not previously considered in the setting of metastatic or locoregionally advanced disease.

Patient inclusion criteria were similar in nivolumab and pembrolizumab clinical trials. All trials required an ECOG performance status of 0 or 1. Patient exclusion criteria for Checkmate 141 enrollment included brain metastases, active immunosuppression, or histologically confirmed R/M carcinoma of the nasopharynx, squamous cell carcinoma of unknown primary, and salivary gland or non-squamous histology [43].

Exclusion criteria for pembrolizumab trials KEYNOTE-012, KEYNOTE-040, KEYNOTE-028 and KEYNOTE-048 all included active autoimmune disease, receipt of systemic steroid therapy above a physiologic dose, CNS metastases, active infection requiring treatment (including HIV, hepatitis B and C), and known history of HIV [14]. Particular to KEYNOTE-048, participant ineligibility included a primary tumor site of nasopharynx (any histology).

A systematic review of 13 articles plus 4 meeting presentations was conducted to determine if ICI therapy was safe and efficacious in patients with HIV infection and advanced stage cancer, including NSCLC, melanoma and Kaposi sarcoma. This analysis was comprised of 73 patients from 13 articles (11 case reports and 2 case series) and 4 meeting abstracts. Sixty-two patients were treated with anti–PD-1 therapy, 6 with anti–CTLA-4 therapy, 4 with anti–PD-1/CTLA-4 therapy, and 1 with sequential ipilimumab and nivolumab therapy. Among 34 patients with known pre-treatment and post-treatment HIV loads, HIV remained suppressed in 26 of the 28 (93%) with undetectable HIV load. ORR was 30% for NSCLC, 27% for melanoma, and 63% for Kaposi sarcoma. Results demonstrated that ICI therapy was generally well tolerated, inducing grade 3 or higher immune-related adverse events identified in 6 of 70 patients. Further, this study found no association with adverse changes in HIV load or CD4 cell count [111]. Case reports have been presented demonstrating response to ipilimumab and/or nivolumab in HIV+ patients with advanced melanoma without any detriment [112]. Interim results from the CITN-12 clinical trial suggested that pembrolizumab was safe for HIV+ patients with cancer on anti-retroviral therapy [113].

Moreover, PD-1 has been shown to be upregulated in hepatitis C virus (HCV)-specific CD8+ cells, suggesting that although patients with chronic hepatitis have traditionally been excluded from clinical trials, anti–PD-1 therapy may have beneficial effects in patients with HCV infection. A case report details the use of anti-PD-1 to treat metastatic Merkel cell carcinoma (MCC) in a patient with untreated chronic HCV infection. Treatment resulted in a rapid antitumor response as well as a rapid decline in HCV RNA without apparent hepatocellular injury. Such results suggested that anti–PD-1 therapy may induce antitumor immune responses while also working to restore antiviral T-cell function and overcome viral immune escape. As such, there are several ongoing clinical trials of anti–PD-1 checkpoint blockade in patients with hepatocellular carcinomas which do allow the enrollment of patients with untreated HCV (NCT02658019, NCT02940496, and NCT02702414) [114].

The majority of subcommittee members agreed that the presence of recurrent and/or metastatic disease (89%), previous platinum therapy (78%) and patient performance status (56%), influence whether they would recommend immunotherapy for a specific patient. Additionally, 83% of the subcommittee noted that clinical trial eligibility and availability would also play a role in determining whether to administer FDA-approved immunotherapies. It should be noted that this vote was taken prior to approval of pembrolizumab in 1st line R/M disease (on June 10, 2019).

Overall, the subcommittee recognized that while the data regarding patients with autoimmune disease are sparse, the pool of patients considered eligible is increasing. For instance, the subcommittee agrees that age (89%), lung metastases (89%) or co-morbidities (75%) are not reasons to disqualify a patient from receiving anti-PD-1 immunotherapy, and that elderly patients actually tolerate immunotherapies better than cytotoxic therapies. Additionally, the subcommittee agrees (81%) that patients with autoimmune disease should not automatically be excluded but rather, the decision should be tailored to the specific disease. The subcommittee recommends that patients with controlled diseases such as Hepatitis C (75%) are generally suitable for ICI treatment, as are HIV+ patients (75%) with normal CD4+ T cell counts and who are on antiretroviral therapy. Moreover, a substantial minority of the subcommittee also agreed that significant burden and pace of disease requiring rapid tumor burden reduction (44%) and steroid dosing for any reason over 10 mg/day of prednisone or equivalent (38%) would be reasons not to give anti-PD-1 monotherapy to a platinum chemotherapy-refractory HNSCC patient.

cSCC is a relatively rare cancer of the head and neck but represents 20% of all non-melanoma skin cancers [115]. Advanced cSCC carries a poor prognosis and until this year there were no FDA-approved systemic therapies [116–118]. However, given a high TMB and disease-risk associated with immunosuppression [118], a phase I dose-escalation study (NCT02383212) and subsequent, phase 2 pivotal study (NCT02760498), examined the PD-1 inhibitor cemiplimab in patients with advanced cSCC. These studies resulted in a FDA approval for cemiplimab on September 28, 2018 for the treatment of patients with metastatic or locally-advanced cSCC who are not candidates for curative surgery or radiation. Approval was based on a clinically meaningful and durable ORR (47.2% [95% CI: 38–47]) of 108 total patients (75 with metastatic cSCC and 33 with LA cSCC) in these two open-label clinical trials at the median follow-up time of 8.9 months (4% CR and 44% PR) with 61% of responses reaching ≥ six months in the metastatic disease cohort of the phase 2 study [119].

Examining another rare cancer of the head and neck, KEYNOTE-028 included a cohort for patients with recurrent or metastatic PD-L1 positive NPC who had failed on prior standard therapy. ORR was 25.9% (95% CI: 11.1–46.3), with 26% of patients experiencing PR and 52% of patients with SD. Drug-related AEs occurred in 74% of patients (30% grade 3/4), most commonly pruritus (26%), fatigue (19%), and hypothyroidism (19%) [119]. Additionally, the phase 2 study, NCI-9742 (NCT02339558), evaluated nivolumab antitumor efficacy in heavily pretreated patients with R/M-NPC [120]. Of 44 patients, ORR was 20.5% (CR = 1; PR = 8), 1-year OS rate was 59% (95% CI, 44.3 to 78.5%), and 1-year PFS rate was 19.3% (95% CI, 10.1 to 37.2%) [120].

Atezolizumab (anti-PD-L1) has also shown promising efficacy for patients with NPC as demonstrated in the HNC cohort of the phase 1 PCD4989g clinical trial (NCT01375842) [121]. Of the 32 patients enrolled, objective responses by RECIST v1.1 occurred in 22% of patients, with median of PFS 2.6 months (range 0.5–48.4), and median OS 6.0 months (0.5–51.6+) [121].

Additionally, results from two single-arm, phase 1 trials examined camrelizumab, a PD-1 inhibitor, in the treatment of R/M-NPC. Safety and preliminary antitumor efficacy were reported for camrelizumab monotherapy in the second-line (NCT02721589) and in combination with gemcitabine and cisplatin in the first-line (NCT03121716) compared to SOC [122]. In the camrelizumab monotherapy cohort 34% (95% CI: 24–44) of evaluable patients experienced an ORR with a median follow-up of 9.9 months (IQR 8.1–11.7). Sixteen percent of patients experienced TRAEs of grade 3 or 4. For the combination study, 91% (95% CI: 72–97) of evaluable patients experienced an ORR with a median follow-up time of 10.2 months (IQR 9.7–10.8), and 87% of patients developed grade 3 or 4 TRAEs [122].

Additional rare cancers of the head and neck include salivary gland carcinoma (SGC) and thyroid cancer. KEYNOTE-028 reported that in patients with PD-L1-positive (≥1% of tumor or stroma cells) unresectable or metastatic salivary gland carcinoma, pembrolizumab demonstrated a RR of 12% and a manageable safety profile [123]. A phase I/II study (NCT02404441) characterized the safety and efficacy of spartalizumab, a monoclonal antibody (mAb) which binds PD-1, in patients with anaplastic thyroid cancer (ATC), an aggressive cancer with limited treatment options [124]. At the data cut-off date of Jan 23, 2018, the ORR by RECIST 1.1 (confirmed + unconfirmed PRs) was 5/30 (17%), with four CRs [124]. Furthermore, new data have prompted the phase 2 clinical trial (NCT03072160) which is investigating the use of pembrolizumab for treatment of TKI-naïve patients with recurrent or metastatic medullary thyroid carcinoma (MTC), for whom surgery is not a curative option.

Category 1 evidence led to the FDA approval of the checkpoint inhibitor cemiplimab for treatment of patients with advanced cSCC. Prescribing dose information is listed as 350 mg administered as an intravenous infusion over 30 min every three weeks, until disease progression or unacceptable toxicity.

The majority of the subcommittee agreed with the recommendation of providing immune checkpoint inhibition to patients with cutaneous tumors in the head and neck area. The subcommittee recognizes the need for more clinical data concerning immunotherapy efficacy in the treatment of rare HNC. Many subcommittee members noted that existing clinical trial data for these rare cases have been generally positive, and support clinical trial enrollment for these patients. However, in lieu of an available clinical trial, 43% of the subcommittee recommends treating with SOC or targeted therapy based on gene expression analyses.

The subcommittee agreed (76.5%) that patients with NPC are distinct from other HNSCC patients. Reasoning for separation is primarily due to etiological differences in disease, which results in differential SOC. Of the members that recommended separate consideration of NPC patients, 57% recommended clinical trial enrollment as the primary treatment option for recurrent and metastatic disease. Multiple clinical trials in patients with platinum refractory NPC have demonstrated single-agent activity of PD-1/PD-L1 inhibitors with response rates of ~ 20–25%. Therefore, where clinical trial enrollment is not feasible, patients with platinum-refractory NPC may derive clinical benefit from single-agent PD-1/PD-L1 checkpoint blockade.

The subcommittee recognizes the significant promise in ongoing clinical trials regarding increasing response in patients with HNSCC and providing durable, long-term outcomes. The subcommittee notes that overall outcomes in patients with HNSCC remain poor, and that clinical trials involving immune checkpoint modulating therapies are an excellent treatment option in most patients. As no combination strategies are currently approved in the IO-refractory disease setting, a majority of the subcommittee (94%) recommends enrolling a patient with R/M HNSCC into a clinical trial assessing a combination immunotherapeutic approach. In this sense, consensus was reached between all clinical members of the subcommittee to recommend combination therapy (notably chemotherapy + IO, once FDA-approved) for rapidly growing disease due to the need for an enhanced response rate. Additionally, in identifying what tumor characteristics influence treatment modality and sequence, the majority of the subcommittee noted growth rate as being the most important, however, many also noted the combined importance of tumor volume and tumor size. Other mentions of importance include existing immune conditions, tumor site and patient symptoms.

CheckMate 141 assessed exploratory quality of life measurements using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire–Core 30 module (EORTC QLQ-C30), the head-and-neck–specific module (QLQ-H&N35) and the European Quality of Life–5 Dimensions (EQ-5D-3 L) criteria. Scores for these modules range from 0 to 100, with higher scores indicating better functioning or well-being or higher symptom burden (although scales measuring symptom burden were reverse-scored to facilitate presentation). The proportion of patients reporting health problems was assessed using the three-level version of the EQ-5D-3 L questionnaire. Patients also completed the EQ-5D-3 L visual-analogue scale, for which scores range from 0 to 100 and higher scores indicate better perceived health status [43].

Assessments were obtained at the first two follow up visits post-treatment (~ 35 days and ~ 80 days, respectively). Patients who received nivolumab reported increased quality of life improvement (> 10 points) compared to patients who received chemotherapy. Evaluations noted that patients receiving nivolumab showed improvement in social function, fatigue, and cognitive ability. Nivolumab treatment also delayed time to clinical deterioration compared to chemotherapy, including delaying loss of sensory abilities, onset of pain, and difficulty of social engagement [43]. Altogether, patient-reported outcomes from CheckMate 141 revealed stabilization or slight improvement in quality of life measures such as social functioning and pain, while SOC resulted in a clinically meaningful worsening across many of the same measures [43].

Data on health-related quality of life (HRQoL) of pembrolizumab vs SOC for R/M HNSCC from the KEYNOTE-040 clinical trial were presented at ASCO 2018. KEYNOTE-040 included a HRQoL analysis using pre-specified questionnaires for all patients who received at least one dose of pembrolizumab. The questionnaires EORTC QLQ-C30, EORTC QLQ-H&N35, and EQ-5D were given to patients at baseline, followed by weeks three, six, nine, and every six weeks thereafter up to 12 months or termination of treatment and, finally, at a 30-day post-treatment safety visit [166]. In this trial, HRQoL was assessed by comparing the mean change from baseline to week 15 and time to deterioration (TTD) was defined as a ten point or greater decline from baseline. At week 15, global health status (GHS)/QOL scores remained stable for patients on pembrolizumab (least-squares [LS] mean, 0.39; 95% CI: − 3.00–3.78) but declined for those treated with SOC (LS mean − 5.86; 95% CI: − 9.68 – − 2.04), demonstrating a difference in LS mean of 6.25 points (95% CI: 1.32–11.18; nominal 2-sided p = 0.013) [166]. The largest variance from those treated with pembrolizumab was seen in the patients treated with docetaxel (LS mean - 10.23; 95% CI: 3.15–17.30). Overall these data support previous reports of a clinically meaningful benefit for patients treated with pembrolizumab compared to SOC [166].

Category 1 evidence from CheckMate 141 demonstrated that while patients in the SOC group reported clinically meaningful worsening of quality of life, as well as of pain, sensory problems, and social-contact problems, patients treated with nivolumab remained nearly stable or showed slight improvements with significant p values at both week 9 and week 15 for most comparisons [43].

Cancer patients are faced with an overwhelming amount of information regarding treatment options. The treating physician and staff have the opportunity and responsibility to provide adequate support and education for the patient. 93% of the clinicians on the subcommittee reported that they provide face-to-face counseling with their patients in addition to providing literature to educate patients on how immunotherapy works and its associated toxicities. 53% of the subcommittee also recommends meeting with patients and their family members during office visits to aid in information retention.

The subcommittee recommends that patients should be provided with literature in the doctor’s office (or online resources) to learn more fully about how immunotherapy works, what kinds of treatments and trials are available, and what their experience of treatment might be like, including toxicities. Clinical trials should be a standard part of a doctor’s discussion with the patient about their treatment options, especially for patients whose disease has recurred after first-line therapy.

As for direct patient monitoring during and after treatment, 47% of the subcommittee recommends having the patient use an electronic system to manage daily levels of pain, discomfort and depression. The subcommittee believes an electronic system would lead to validated, reproducible measures that would allow for a regular assessment of response and toxicities and would allow the patient to record and manage symptoms and level of discomfort. If using a patient scale specific for head and neck cancers, the subcommittee (44%) recommends the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC-HN35) over other options.

With fewer high-grade side effects, treatment with ICIs may provide stabilization or improvement in quality of life compared to SOC. However, immunotherapy is not without TRAEs that impact a patient’s quality of life and need to be monitored. Overall, the subcommittee was split as to which treatment management issues need more attention, noting the effect of head and neck tumors on nutrition and maintaining a patient’s quality of life as the most important. The subcommittee stated that basic health and nutrition are not to be neglected throughout a patient’s treatment, especially when HNSCC makes eating, drinking and breathing more challenging, and patients often require specific support.

40% of the subcommittee agreed that quality of life should be monitored every three months, and several subcommittee members recommend patient quality of life to be evaluated as often as the patient receives treatment (27%). The majority of the subcommittee reported treating depression in HNSCC patients with counseling and selective serotonin reuptake inhibitors (SSRIs) (57%). It was also suggested that doctors should pay close attention to depression in general appointments and should be sure to inquire into and monitor patients’ emotional well-being, specifically noting that it is easy to develop depression with the stress and fear associated with this diagnosis.