Treatment of metastatic and recurrent head and neck cancer

INTRODUCTION — The prognosis of patients with recurrent or metastatic head and neck squamous cell cancer is generally poor. The median survival in most series is 6 to 15 months depending on patient- and disease-related factors. Symptom-directed care plays an important role in the management of these patients. (See "Management and prevention of complications during initial treatment of head and neck cancer" and "The role of parenteral and enteral/oral nutritional support in patients with cancer", section on 'Head and neck cancer' and "Overview of managing common non-pain symptoms in palliative care".)

Systemic therapy is indicated for most patients with metastatic or advanced recurrent squamous cell carcinoma of the head and neck. The choice of systemic regimen is influenced by multiple clinical factors, including patient comorbidities, performance status, previous therapy, and pathologic features (ie, programmed death ligand 1 [PD-L1] expression status). Systemic treatment options include the following, administered as single-agent therapy or in combination:

●Immune checkpoint inhibitors (ICIs)

●Conventional cytotoxic chemotherapy

●Molecularly targeted agents (eg, epidermal growth factor receptor [EGFR] inhibitors)

The approach to systemic therapy for metastatic or advanced recurrent squamous cell carcinoma of the head and neck is reviewed here, along with the management of oligometastatic disease. The use of chemotherapy as a component of initial definitive therapy is discussed separately, as is the use of systemic chemotherapy in other malignancies of the head and neck region:

●(See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy".)

●(See "Treatment of recurrent and metastatic nasopharyngeal carcinoma".)

●(See "Malignant salivary gland tumors: Treatment of recurrent and metastatic disease".)

●(See "Head and neck sarcomas".)

GENERAL APPROACH — The general initial treatment approach to a patient with recurrent or metastatic squamous cell carcinoma of the head and neck is summarized in the algorithm (algorithm 1).

The treatment approach is influenced by prior therapies (surgery, radiation, chemotherapy), patient performance status and comorbidities, and tumor programmed death ligand 1 (PD-L1) expression status.

When assessing the trials described in this topic, it is important to remember that contemporary studies may be complicated by lead time bias that results from more sensitive diagnostic techniques, leading to earlier detection of asymptomatic metastases in patients receiving radiologic monitoring, and the inclusion of more patients with more favorable prognosis (ie, human papillomavirus [HPV] related metastatic tumors).

Clinical settings — Systemic therapy is chosen based on exposure to previous therapies and the distribution of disease:

●Oligometastatic disease (see 'Oligometastatic disease' below)

●Locally recurrent disease (see 'Locally recurrent disease' below)

●No prior systemic therapy (see 'Previously untreated patients' below)

●Previous systemic therapy for initial diagnosis of locoregional disease (see 'Previous systemic therapy for locoregional disease' below)

●Previous systemic therapy after diagnosis of recurrent or metastatic disease (see 'Subsequent therapy' below)

Prognostic factors — Patient-specific factors that influence prognosis need to be considered when choosing therapy. Factors associated with longer survival in patients include the following [1-5]:

●Ambulatory performance status (Eastern Cooperative Oncology Group [ECOG] 0 or 1 versus 2 (table 1))

●Prior response to chemotherapy

●Longer time since completion of definitive therapy

●HPV associated oropharyngeal cancers

●Tumor PD-L1 expression status based on combined positive score (CPS), a predictive marker for response to anti-programmed cell death protein 1 (PD-1)

In contrast, factors associated with a poor prognosis include the following:

●Weight loss

●Poor performance status

●Prior radiation therapy

●Active smoking

●Significant comorbidity

Increasing age is often associated with increasing comorbidity. However, evidence suggests that healthy older adult patients with metastatic and recurrent head and neck cancer have survival outcomes comparable to younger patients, but they may experience increased toxicity [6,7].

ACTIVE AGENTS — Conventional cytotoxic drugs, immune checkpoint inhibitors (ICIs), and molecularly targeted compounds all have activity in metastatic and recurrent head and neck cancer.

●ICIs – Pembrolizumab (table 2) is a checkpoint inhibitor of programmed cell death protein 1 (PD-1). It has shown activity, alone or with chemotherapy, in previously untreated patients with advanced squamous cell carcinoma of the head and neck. Both pembrolizumab and nivolumab (table 3), another PD-1 checkpoint inhibitor, have also shown clinically significant activity in patients who progressed on or after platinum-based regimens [8-11].

Further information regarding the use of immunotherapy for advanced squamous cell carcinoma of the head and neck in the first-line and second-line settings is discussed below. (See 'Pembrolizumab with or without platinum and fluorouracil' below and 'PD-1 inhibitor immunotherapy' below.)

●Platinum-based chemotherapy – The platinum agents cisplatin and carboplatin are used both as single agents and to form the backbone for most combination regimens [12-19].

Although carboplatin is often considered to be less systemically effective than cisplatin in head and neck cancer, there is little direct evidence. Carboplatin may be preferred in some cases since it is associated with less neurotoxicity, nephrotoxicity, ototoxicity, and nausea and vomiting compared with cisplatin, although carboplatin causes more myelosuppression [20-23]. In addition, carboplatin has much shorter infusion times than cisplatin. Adequately powered randomized trials directly comparing cisplatin with carboplatin as single agents and in multidrug regimens are lacking in this patient population.

•Of the three studies that have compared the combination of cisplatin or carboplatin and fluorouracil for induction chemotherapy in patients with locally advanced disease [20,21,24], only one showed a statistically increased response rate and survival with the use of a cisplatin-based rather than carboplatin-based regimen [21]. Quality of life and treatment-related toxicity are considerations for patients who do not have a curative option.

•Similarly, indirect evidence from the phase 3 trial assessing the role of cetuximab in combination with a platinum-based regimen supports the use of carboplatin [25]. In that trial, there was a similar improvement in progression-free survival (PFS) attributable to cetuximab in patients treated with cisplatin plus fluorouracil and those given carboplatin plus fluorouracil, and there was no apparent difference in results between patients treated with the two platinum regimens.

•A Southwest Oncology Group (SWOG) trial showed similar survival for a regimen of cisplatin plus fluorouracil versus carboplatin plus fluorouracil in the metastatic recurrent setting, with a nonstatistically significant improvement in response rate (32 versus 21 percent) and overall survival (OS; 7 versus 5 months) for the cisplatin-containing regimen [1].

●Other chemotherapy and targeted agents – In addition to platinum, taxanes (eg, docetaxel (table 4), paclitaxel) are the most active single-agent chemotherapy regimen. Other active agents include nanoparticle albumin-bound paclitaxel (nabpaclitaxel) [26-30], fluorouracil [2], and methotrexate [1,15,26]. Other agents that may possess activity but have been less extensively evaluated include etoposide [31], gefitinib [32-36], gemcitabine [37,38], capecitabine [39], pemetrexed [40-43], and bleomycin [3,16,44]. (See 'Combination chemotherapy' below.)

Cetuximab (table 5), a targeted monoclonal antibody to epidermal growth factor receptor (EGFR), has activity alone or in combination with chemotherapy [45]. (See 'Cetuximab with platinum-based chemotherapy' below and 'Cetuximab' below.)

OLIGOMETASTATIC DISEASE — Patients with head and neck cancer with one or a limited number of distant metastases at presentation or relapse may benefit from aggressive treatment that includes eradication of all known sites of disease.

The most extensive data come from small series of patients with oligometastatic disease in the lungs, in whom disease in the primary site and regional lymph nodes has been completely controlled [46]. In this setting, surgical metastasectomy can result in prolonged disease-free survival in a significant percentage of appropriately selected patients. Data on oligometastatic disease involving sites other than the lung are more limited. (See "Second primary malignancies in patients with head and neck cancers", section on 'Distinguishing second primary lung malignancies from distant metastasis' and "Surgical resection of pulmonary metastases: Benefits, indications, preoperative evaluation, and techniques".)

●In a series that included 97 patients with squamous cell carcinoma from primary sites other than the oral cavity, the five-year overall survival (OS) rate was 32 percent [47]. Similarly, in a series of 83 patients from one institution, the five-year survival was 34 percent for the 41 patients with squamous cell carcinoma [48].

●The primary site is an important factor influencing outcome. Patients with oligometastatic disease from a primary tumor of the oral cavity appear to have a significantly worse prognosis. As an example, in one series of 23 patients with a primary lesion of the oral tongue, 22 (96 percent) relapsed, with a median time to death of 10 months [49]. Similar poor results were seen in another series of 17 patients with oral cavity primary tumors, in which there were no five-year survivors [47].

Interpretation of these series is complicated by the possibility that some of these patients were surgically cured of one or more second primary squamous cell lung cancers rather than of a head and neck metastasis.

Other approaches to disease eradication, such as stereotactic body radiation therapy (SBRT), have also been used to treat one or a limited number of pulmonary metastases [50]. However, long-term follow-up data for SBRT are more limited compared with surgical resection. In one small randomized trial in patients with multiple metastatic sites, administering SBRT to a single metastatic site in combination with immunotherapy (nivolumab (table 3)) did not improve response rates in "off-target" lesions or survival, compared with nivolumab alone [51]. (See "Radiation therapy techniques in cancer treatment", section on 'Stereotactic radiation therapy techniques' and "Stereotactic body radiation therapy for lung tumors", section on 'Lung metastases'.)

No clinical trials have been conducted to assess the role of adjuvant chemotherapy following surgical resection of all metastatic disease sites. Although there may be benefit to adjuvant chemotherapy, there is no evidence to make a recommendation for its use in this situation.

LOCALLY RECURRENT DISEASE — Patients with locally recurrent head and neck cancer present a therapeutic challenge. Treatment options are limited, and morbidity can be substantial. Nevertheless, selected patients with recurrent disease may achieve long-term survival. The treatment of locally recurrent disease is discussed separately. (See "Treatment of locally recurrent squamous cell carcinoma of the head and neck".)

PREVIOUSLY UNTREATED PATIENTS

Selection of therapy — The choice of therapy for patients with advanced head and neck cancer who have not received prior systemic therapy is affected by the patient's performance status, symptoms related to locoregional disease burden, comorbidities, and programmed death ligand 1 (PD-L1) combined positive score (CPS) from biopsy specimens (algorithm 1). This includes patients who initially present with advanced disease and are not candidates for definitive therapy, as well as those who develop metastatic or incurable locally recurrent disease following local treatment (surgery and/or radiation therapy [RT]) without chemotherapy.

The KEYNOTE-048 trial established the use of pembrolizumab, with or without chemotherapy, as a first-line regimen for eligible patients with metastatic or recurrent squamous cell carcinoma of the head and neck [52-54]. The addition of pembrolizumab to a platinum and fluorouracil combination (table 6) improves overall survival (OS) compared with cetuximab plus a platinum and fluorouracil combination. For those with high PD-L1 expression (CPS ≥1), single-agent pembrolizumab (table 2) also improves OS compared with cetuximab plus a platinum and fluorouracil combination, although a subgroup analysis suggests this is mainly driven by those with a CPS ≥20 [53]. (See 'Pembrolizumab with or without platinum and fluorouracil' below.)

For patients with no prior systemic chemotherapy or immunotherapy and who are eligible for combination therapy based on performance status and comorbidities, we suggest the following:

●Patients with rapidly progressive disease may benefit from a more rapid tumor response. For such patients who are eligible for combination chemotherapy plus immunotherapy, we suggest the use of pembrolizumab with a platinum plus fluorouracil (table 6), regardless of CPS. Alternatively, some contributors substitute a taxane (paclitaxel or docetaxel) for the fluorouracil. (See 'Pembrolizumab with platinum plus taxane' below.)

However, for patients who are not eligible for combination therapy, pembrolizumab monotherapy (table 2) is an appropriate alternative. Some patients may also benefit from local therapy for symptom control prior to initiating systemic therapy.

●For those without rapidly progressive disease, our preferred approach is based on CPS:

•CPS ≥20 – We suggest single-agent pembrolizumab (table 2) rather than pembrolizumab plus chemotherapy.

•CPS ≥1 and <20 – For patients with good performance status (eg, Eastern Cooperative Oncology Group [ECOG] performance status ≤2 (table 1)), we suggest pembrolizumab with platinum-based chemotherapy rather than single-agent pembrolizumab. While chemotherapy regimens containing platinum plus fluorouracil (table 6) are preferred, some contributors alternatively substitute a taxane (paclitaxel or docetaxel) for the fluorouracil.

For patients with a lesser performance status where it is preferred to avoid the potential toxicities of systemic chemotherapy, single-agent pembrolizumab is a reasonable alternative.

•CPS <1 or unavailable – We suggest platinum-based chemotherapy, with or without pembrolizumab rather than other chemotherapy regimens.

For patients who are eligible for combination therapy but have a contraindication to immunotherapy, options include a doublet cytotoxic chemotherapy regimen, with or without concurrent or sequential cetuximab (table 5). Doublet cytotoxic chemotherapy regimens, generally a platinum-based regimen with either a taxane or fluorouracil, increase the objective response rate (ORR) compared with single-agent chemotherapy, although no improvement in OS has been demonstrated. The addition of cetuximab to cisplatin plus fluorouracil increases OS compared with cisplatin plus fluorouracil alone. (See 'Combination chemotherapy' below and 'Cetuximab with platinum-based chemotherapy' below.)

Patients not eligible for combination chemotherapy plus immunotherapy may be candidates for single-agent therapy. Options include a platinum (typically carboplatin), taxanes (eg, paclitaxel, docetaxel (table 4), nabpaclitaxel), fluorouracil, methotrexate, and cetuximab (table 5). (See 'Active agents' above.)

Patients with severe comorbidities or a poor performance status may be best treated with supportive care. (See 'Subsequent therapy' below.)

Pembrolizumab with or without platinum and fluorouracil — For those with metastatic or recurrent head and neck cancer without prior exposure to systemic therapy, the combination of pembrolizumab plus platinum-based chemotherapy (table 6) improves OS beyond that seen with cetuximab plus chemotherapy [52-55]. For those with high PD-L1 expression, single-agent pembrolizumab (table 2) also improves OS, compared with cetuximab plus chemotherapy, with less toxicity. Responses to pembrolizumab, either alone or in combination with chemotherapy, are more durable than those seen with cetuximab plus chemotherapy. The US Food and Drug Administration (FDA) has approved the use of pembrolizumab for this indication for all patients, along with cisplatin and fluorouracil and as a single agent for those whose tumors express PD-L1 CPS ≥1 (algorithm 1). When given as a single agent or in combination, pembrolizumab is administered for a maximum of two years.

The use of pembrolizumab (table 2) as second-line therapy is discussed separately. (See 'PD-1 inhibitor immunotherapy' below.)

In an open-label phase 3 trial (KEYNOTE-048, NCT02358031), 882 patients with recurrent or metastatic squamous cell carcinoma of the head and neck not curable by surgery and without prior systemic therapy in the metastatic setting were randomly assigned on a 1:1:1 basis to receive single-agent pembrolizumab versus pembrolizumab plus a platinum and fluorouracil combination versus cetuximab plus a platinum and fluorouracil combination [52-54]. PD-L1 status was assessed using CPS, which stains lymphocytes and other cells in addition to tumor cells. Results were evaluated for tumors with varying CPS scores and the entire study population regardless of CPS.

Pembrolizumab-containing combination therapy, in comparison with cetuximab-containing combination therapy, demonstrated the following:

●At median follow-up of 45 months, improved OS in the total study population (median 13 versus 11 months, four-year OS 19 versus 5 percent, hazard ratio [HR] 0.71, 95% CI 0.59-0.85) and in those with CPS ≥1 (median 14 versus 11 months, four-year OS 22 versus 4 percent, HR 0.64, 95% CI 0.53-0.78) [54].

●Similar ORRs (44 versus 38 percent for CPS ≥20, 37 versus 36 percent for CPS ≥1), but longer duration of response for pembrolizumab-based combination therapy [54].

●Similar rates of grade ≥3 treatment-related adverse events between the treatment groups (72 versus 69 percent) [54].

●In post-hoc analyses by CPS subgroup, at median follow-up of approximately 33 months [53] and 45 months [54]:

•Improved OS in patients with CPS ≥20 (median 15 versus 11 months, HR 0.62, 95% CI 0.46-0.84) [54].

•Improved OS in patients with CPS between 1 and 19 (median 13 versus 10 months, HR 0.71, 95% CI 0.54-0.94) [53].

•Similar OS in those with CPS <1 (11.3 versus 10.7 months, HR 1.21, 95% CI 0.76-1.94) [53].

•Similar progression-free survival (PFS) in patients with CPS ≥20 (median 6 versus 5 months, HR 0.76, 95% CI 0.57-1.02) [54], those with CPS between 1 and 19 (median 4.9 months each, HR 0.93, 95% CI 0.71-1.21), and those with CPS <1 (median 5 versus 6 months, HR 1.46, 95% CI 0.93-2.3) [53].

•Similar ORRs for patients with CPS between 1 and 19 (29 versus 33 percent) and those with CPS <1 (31 versus 39 percent) [53].

Single-agent pembrolizumab, in comparison with cetuximab-containing combination therapy, demonstrated the following:

●At median follow-up of 45 months, improved OS in the total study population (median 12 versus 11 months, four-year OS 15 versus 7 percent, HR 0.81, 95% CI 0.68-0.97) and in those with CPS ≥1 (median OS 12 versus 10 months, four-year OS 17 versus 6 percent, HR 0.74, 95% CI 0.61-0.89) [54].

●Lower PFS in the total population (median 2 versus 5 months, HR 1.29, 95% CI 1.09-1.53), with more early progression in the pembrolizumab arm [54].

●Lower ORRs in the total population (17 versus 36 percent), but longer duration of response in all populations [54].

●Fewer grade ≥3 treatment-related adverse events (17 versus 69 percent) [54].

●In post-hoc analyses by CPS subgroup, at median follow-up of approximately 33 months [53] and 45 months [54]:

•Improved OS in patients with CPS ≥20 (median 15 versus 11 months, HR 0.61, 95% CI 0.46-0.81) [54].

•Similar OS in patients with CPS between 1 and 19 (median 11 versus 10 months, HR 0.86, 95% CI 0.66-1.12) [53].

•Lower OS in patients with CPS <1, although the results were not statistically significant (median 8 versus 11 months, HR 1.51, 95% CI 0.96-2.37) [53].

•Similar PFS in patients with CPS ≥20 (median 3 versus 5 months, HR 0.96, 95% CI 0.74-1.26) [54], those with CPS between 1 and 19 (2 versus 5 months, HR 1.25, 95% CI 0.96-1.61), and lower PFS for those with CPS <1 (median 2 versus 6 months, HR 4.31, 95% CI 2.63-7.08) [53].

•Lower ORRs for patients with CPS between 1 and 19 (15 versus 34 percent) and those with CPS <1 (5 versus 42 percent) [53].

Pembrolizumab with platinum plus taxane — For patients treated with pembrolizumab plus chemotherapy, some contributors prefer a chemotherapy regimen that includes platinum plus a taxane (such as paclitaxel or docetaxel) rather than regimens that include platinum plus fluorouracil. Pembrolizumab plus carboplatin and paclitaxel is active in clinical studies of recurrent and metastatic head and neck cancer [56]. In addition, regimens that contain platinum with a taxane are similarly effective and better tolerated than regimens containing platinum plus fluorouracil [57]. (See 'Cetuximab with platinum-based chemotherapy' below.)

In a phase IV trial (KEYNOTE-B10), 92 patients with treatment-naïve recurrent or metastatic head and neck cancer were treated with up to six cycles of carboplatin plus paclitaxel in combination with pembrolizumab [56]. Pembrolizumab monotherapy was subsequently continued to complete up to a total of 35 cycles. In preliminary results, at median follow-up of eight months, the ORR for the entire study population was 48 percent, including a complete response rate of 5 percent. Median PFS and OS for all patients were 6 and 12 months, respectively. The grade ≥3 toxicity rate was 71 percent, including two deaths attributed to sepsis and hypersensitivity reaction.

Cetuximab with platinum-based chemotherapy — For patients who have a good performance status but are not eligible for immunotherapy, the combination of cetuximab with either a platinum and fluorouracil or a platinum plus a taxane are both options for initial therapy of advanced squamous cell carcinoma of the head and neck (algorithm 1). We prefer the use of cetuximab in combination with a platinum plus a taxane, which confers similar efficacy but less toxicity relative to cetuximab in combination with a platinum plus fluorouracil.

Cetuximab is a monoclonal antibody that targets the epidermal growth factor receptor (EGFR). The addition of cetuximab to a platinum-based doublet regimen increases OS compared with doublet chemotherapy alone. Cetuximab is approved for first-line use in combination with platinum-based chemotherapy. Docetaxel in combination with a platinum and cetuximab demonstrates a similar survival benefit and less toxicity relative to fluorouracil with a platinum and cetuximab [57].

Of note, cetuximab is administered either weekly or every two weeks, in combination with a platinum and fluorouracil. Clinicians may offer cetuximab on an every-two-week schedule as a convenient and alternative dosing to the weekly schedule, both of which have approval from the FDA in patients with squamous cell carcinoma of the head and neck [58]. If the every-two-week schedule of cetuximab (table 5) is chosen in this setting, our general approach is to reserve its use for the maintenance phase of therapy.

●Cetuximab with platinum and fluorouracil – In the phase III EXTREME trial, 442 patients with recurrent or metastatic squamous cell carcinoma of the head and neck were assigned to a platinum (cisplatin 100 mg/m² on day 1 or carboplatin area under the concentration x time curve [AUC] 5 day 1) plus fluorouracil (1000 mg/m² per day for four days) every three weeks, with or without cetuximab (400 mg/m² for the initial dose, followed by subsequent weekly 250 mg/m²) [25]. Overall, 39 percent of patients had received systemic cytotoxic chemotherapy at least six months prior to randomization as part of their initial definitive management. Chemotherapy was given for a maximum of six cycles, although cetuximab could be continued as maintenance therapy until disease progression or toxicity. Crossover to include cetuximab was not allowed in the chemotherapy arm of the trial.

Chemotherapy plus cetuximab significantly prolonged OS compared with chemotherapy alone (median 10.1 versus 7.4 months, HR for death 0.80, 95% CI 0.64-0.99). Significant improvements were also seen in the PFS and objective response rates (median 5.6 versus 3.3 months and 36 versus 20 percent, respectively). In a retrospective analysis of the trial, the improvements seen with cetuximab appeared to be similar in those whose tumor was human papillomavirus (HPV) related compared with those whose tumor was HPV negative [59].

Crossover was not allowed. Thus, this trial does not exclude the possibility that sequential use of cisplatin plus fluorouracil followed by cetuximab at the time of failure would be equivalent in survival with the three-drug regimen.

The overall incidence of severe (grade 3 or 4) adverse events with the cetuximab regimen was not significantly increased compared with chemotherapy alone (82 versus 76 percent). However, treatment with cetuximab resulted in a significant increase in the incidence of severe hypomagnesemia (5 versus 1 percent), severe skin reactions (9 versus <1 percent), and sepsis (4 versus <1 percent). Severe infusion-related reactions to cetuximab occurred in 3 percent of cases. (See "Infusion-related reactions to therapeutic monoclonal antibodies used for cancer therapy" and "Cutaneous adverse events of molecularly targeted therapy and other biologic agents used for cancer therapy".)

●Cetuximab with platinum and taxane – In an open-label phase II trial, 541 patients with recurrent or metastatic disease were randomly assigned to six cycles of cisplatin and fluorouracil plus cetuximab versus four cycles of cisplatin and docetaxel plus cetuximab [57]. Both regimens were followed by maintenance cetuximab.

At a median follow-up of 34 months (among the 539 patients included in the intention-to-treat analysis), relative to the fluorouracil-based regimen, the docetaxel-based regimen demonstrated similar OS (median 14.5 versus 13.4 months; HR 0.91, 95% CI 0.75-1.10), decreased toxicity (grade ≥3 adverse events 81 versus 93 percent), less delays in administration (10 versus 27 percent), less substitution of carboplatin for cisplatin (6 versus 25 percent), and a higher percentage of patients initiating maintenance cetuximab (72 versus 52 percent).

Combination chemotherapy — A number of combination chemotherapy regimens without cetuximab have been evaluated in patients with advanced head and neck cancer, most of which include a platinum compound (algorithm 1). Although these regimens significantly improve response rates compared with a single agent, an improvement in OS has not been demonstrated compared with single-agent chemotherapy.

Platinum and fluorouracil — Cisplatin (100 mg/m² intravenous, day 1) and fluorouracil (1000 mg/m² per day, continuous infusion, for four days) is a standard combination chemotherapy regimen for use in the palliative setting. In phase III trials, this regimen produced response rates around 30 percent, which is significantly better than with single-agent cisplatin or methotrexate [1,2,16,60].

As an example, a Southwest Oncology Group (SWOG) trial randomly assigned 261 patients with advanced head and neck cancer to one of three regimens: cisplatin plus fluorouracil, carboplatin plus fluorouracil, or single-agent methotrexate [1]. The overall response rate was greater for the two combination regimens compared with methotrexate (32 and 21 versus 10 percent, respectively). However, the median survival was similar with all three regimens (6.6, 5.0, and 5.6 months, respectively). The percentage of patients with any grade 3 or 4 toxicity was greatest for those treated with the cisplatin regimen compared with the carboplatin regimen or single-agent methotrexate (33 versus 26 and 16 percent, respectively).

Because of the toxicity of this regimen and the need for prolonged infusion, we often prefer to use a platin plus taxane combination or single agents as first-line treatment in the palliative setting.

Platinum plus taxane — Both cisplatin and carboplatin have been combined with either paclitaxel or docetaxel in various combination regimens. None of these regimens have demonstrated superiority compared with cisplatin plus fluorouracil.

The most extensive data on the role of a platinum-taxane regimen come from a phase III trial conducted by the ECOG, in which 204 patients were randomly assigned to cisplatin plus fluorouracil or cisplatin plus paclitaxel [60]. With a median follow-up of eight months, there were no statistically significant differences in OS (median survival 8.7 versus 8.1 months and one-year survival rate 41 versus 32 percent) or objective response rate (27 versus 26 percent). Gastrointestinal and hematologic toxicities were more common in patients receiving cisplatin plus fluorouracil.

Combinations of docetaxel (75 to 100 mg/m²) plus cisplatin (70 to 75 mg/m²) given every three weeks have been evaluated in phase II studies that included both previously treated and chemotherapy-naïve patients. Overall response rates ranged from 32 to 55 percent, with median response durations around five months [61-64]. These higher response rates were associated with severe side effects in up to 70 percent of patients. Weekly therapy with this combination may have similar efficacy and a more favorable toxicity profile [65]. Docetaxel plus cisplatin has not been compared with cisplatin plus fluorouracil in a randomized phase III trial.

The substitution of carboplatin for cisplatin simplifies administration and is associated with less nephrotoxicity, neurotoxicity, ototoxicity, and emesis. Carboplatin has also been combined with either paclitaxel or docetaxel in phase II trials [66-68]. These combinations have not been adequately assessed in randomized trials.

Non-platinum-based regimens — Other combinations, such as gemcitabine plus paclitaxel and paclitaxel plus cetuximab, may be useful in patients who are not candidates for platinum-based chemotherapy.

●Gemcitabine plus paclitaxel – In a SWOG study in 57 patients with metastatic or recurrent head and neck cancer, the combination of gemcitabine (3000 mg/m²) plus paclitaxel (150 mg/m²) on days 1 and 15 of 28-day cycles was associated with a 28 percent objective response rate [69]. Median PFS and OS were four and eight months, respectively. However, it is not clear if this is superior to single-agent taxane therapy.

●Paclitaxel plus cetuximab – In an open-label phase II trial, the combination of weekly paclitaxel and cetuximab resulted in a response rate of 54 percent and median PFS and OS of approximately four and eight months [70]. The efficacy of this regimen is consistent with other doublet regimens.

Regimens not used — Certain regimens are not offered as initial therapy for recurrent or metastatic head and neck cancer either due to a lack of OS benefit in randomized trials and/or significant toxicity.

Chemotherapy regimens that we do not offer include cisplatin plus pemetrexed [43] and various three- or four-drug chemotherapy regimens such as paclitaxel, cisplatin, and fluorouracil [71-73]; docetaxel, cisplatin, and fluorouracil [74,75]; methotrexate, vinblastine, doxorubicin, and cisplatin [76]; and paclitaxel, ifosfamide, and cisplatin or carboplatin [77,78].

Targeted therapies that are not used include panitumumab or bevacizumab, as the addition of these agents to chemotherapy failed to improve OS in separate randomized trials [79,80]. These agents should not be substituted for cetuximab.

We also do not offer nivolumab plus ipilimumab, as this combination of immune checkpoint inhibitors (ICIs) failed to improve OS over chemotherapy with cetuximab plus platinum-based chemotherapy in a randomized phase III trial (CheckMate 651) [81]. Similarly, in a separate randomized phase II trial (CheckMate 714), initial therapy with nivolumab plus ipilimumab did not improve objective response rates over single-agent nivolumab [82].

PREVIOUS SYSTEMIC THERAPY FOR LOCOREGIONAL DISEASE

Choice of therapy — Concurrent and/or induction chemotherapy is frequently used in combination with radiation therapy (RT) for definitive therapy in patients who present with locoregional disease. Although such treatment has the potential to cure some patients, many relapse, either with recurrent disease in the head and neck region or with distant metastases. (See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy".)

Some patients who relapse may be candidates for salvage therapy with curative intent, using either surgery or reirradiation. However, most patients who recur ultimately require palliative systemic therapy. The optimal regimen for these patients has not been defined. (See "Treatment of locally recurrent squamous cell carcinoma of the head and neck".)

Key factors that need to be considered in choosing a treatment include the following:

●Original chemotherapy exposure and drugs (either concurrent or induction)

●Response to the initial chemotherapy or chemoradiation regimen

●Interval between initial treatment and disease progression

●Patient performance status and comorbidities

●Exposure to immunotherapy

Patients with progression after initial definitive therapy that included concurrent and/or induction chemotherapy in combination with RT can be treated with immunotherapy-based regimens, in some cases similarly to previously untreated patients (algorithm 1).

●Short period of disease control – For most patients with a relatively short period of disease control (less than six months) from initial definitive treatment that included platinum-based chemotherapy, we offer single-agent immunotherapy rather than single-agent or combination therapy with a platinum, taxane, cetuximab (table 5), or methotrexate. Chemotherapy is unlikely to offer sustained benefit in this population due to a high likelihood of chemotherapy resistance. Clinical trial options should also be considered. (See 'PD-1 inhibitor immunotherapy' below.)

●Prolonged period of disease control – For patients with a relatively prolonged period of disease control (greater than or equal to six months) following initial definitive treatment that included platinum-based chemotherapy, we offer pembrolizumab with or without platinum-based chemotherapy. Carboplatin is the preferred platinum for patients who have already received cisplatin. The approach is similar to previously untreated patients.

The KEYNOTE-048 trial, which evaluated these options in previously untreated patients, allowed enrollment of patients who had received definitive locoregional therapy prior to recurrence, as long as they had not received prior systemic therapy in the recurrent/metastatic setting. This is described in more detail separately. (See 'Pembrolizumab with or without platinum and fluorouracil' above.)

SUBSEQUENT THERAPY

Choice of therapy — Many patients who receive systemic therapy for metastatic or recurrent head and neck cancer ultimately relapse. The choice of subsequent therapy is evolving. There is no agreed-upon standard of care, and experts differ in their preferred approach. Additionally, as immunotherapy regimens become standard in the first-line setting, optimal second-line regimens become less clear. As such, clinical trials are often appropriate.

Options outside of a clinical trial include immunotherapy, cytotoxic chemotherapy, and targeted agents in various combinations. The choice of treatment regimen is influenced by prior treatment history, performance status and comorbidities, and the potential toxicities of a proposed regimen. Symptom-directed care also plays an important role in the management of these patients.

We suggest the following:

●For patients with progression following initial treatment with systemic platinum-based chemotherapy (with or without cetuximab) and who are eligible for immunotherapy, we offer immunotherapy with a programmed death protein 1 (PD-1) antibody (pembrolizumab (table 2) or nivolumab (table 3)). (See 'PD-1 inhibitor immunotherapy' below.)

●For those patients with progression following initial treatment with systemic platinum-based chemotherapy (with or without cetuximab) and who are ineligible for immunotherapy, options include single-agent cetuximab (table 5) if not previously received; taxanes (table 4) if not previously received; a platinum for patients with disease progression more than six months after prior platinum-based chemotherapy (preferably carboplatin if prior cisplatin exposure); fluorouracil; or methotrexate. (See 'Cytotoxic chemotherapy' below.)

●For patients with progression on immunotherapy alone and no prior exposure to cetuximab, we suggest the combination of a platinum plus a taxane or a platinum plus fluorouracil, with or without the use of cetuximab [83]. Although the addition of cetuximab increases overall survival (OS) when added to a platinum plus fluorouracil as first-line therapy in recurrent/metastatic squamous cell carcinoma of the head and neck, there is no evidence that this is more effective than using cetuximab sequentially as second-line therapy. (See 'Cetuximab with platinum-based chemotherapy' above.)

●For patients with progression on combination immunotherapy and platinum plus fluorouracil-based chemotherapy, we suggest either cetuximab alone (table 5) or cetuximab in combination with taxane-based chemotherapy [54]. Since these patients have progressed on chemotherapy, their tumors are likely to be chemotherapy resistant; therefore, the use of cetuximab may be warranted. (See 'Cetuximab' below.)

●Patients with severe comorbidities and/or a poor performance status may be best treated with supportive care and without initiating systemic therapy.

PD-1 inhibitor immunotherapy — Both nivolumab (table 3) and pembrolizumab (table 2) target programmed death protein 1 (PD-1) and have important clinical activity in patients with head and neck squamous cell carcinoma; both are approved by the US Food and Drug Administration (FDA) for patients who have previously been treated with platinum-based chemotherapy for metastatic or recurrent squamous cell carcinoma the head and neck. The toxicity associated with these agents is similar to that observed with their use in other indications. (See "Principles of cancer immunotherapy" and "Toxicities associated with immune checkpoint inhibitors".)

Pembrolizumab — Pembrolizumab (table 2) is a PD-1 inhibitor that has demonstrated activity in previously treated patients. In prospective trials, objective responses have been reported in 15 to 18 percent, with the majority of responses lasting more than six months. It has demonstrated a modest survival benefit, with less toxicity when compared with methotrexate, docetaxel, or cetuximab (table 5). The use of pembrolizumab in previously untreated patients is discussed separately. (See 'Pembrolizumab with or without platinum and fluorouracil' above.)

In an open-label, phase 3 clinical trial (KEYNOTE-040), 495 patients who had failed prior platinum therapy were randomly assigned to pembrolizumab or standard of care (investigator's choice of methotrexate, docetaxel, or cetuximab (table 5)) [10]. Immune checkpoint inhibitors (ICIs) were allowed at the time of progression in any treatment group and were more commonly used in those assigned to the standard of care group (13 versus 5 percent). The following were reported with a median follow-up over seven months:

●Patients treated with pembrolizumab had fewer adverse events (63 versus 84 percent) but more immune-related toxicity (eg, hypothyroidism, pneumonitis).

●Although not statistically significant, more patients assigned to pembrolizumab had tumor responses (15 versus 10 percent). These responses were more durable (median duration of response 18 versus 5 months).

●For the group as a whole, pembrolizumab modestly improved OS (12-month survival 37 versus 27 percent, median survival 8.4 versus 6.9 months, hazard ratio [HR] 0.8, 95% CI 0.65-0.98).

●In a prespecified exploratory analysis, the survival benefit was demonstrated in the 26 percent of patients who had programmed death ligand 1 (PD-L1) expression ≥50 percent (HR 0.53, 95% CI 0.35-0.81), whereas it was not statistically significant in those with PD-L1 expression <50 percent (HR 0.93, 95% CI 0.73-1.17).

These results support the use of pembrolizumab in this population, especially among those with high PD-L1 expression. Further data regarding the activity of pembrolizumab come from two large, nonrandomized studies in previously treated patients:

●In the KEYNOTE-012 study, 192 patients were treated with pembrolizumab on one of two schedules: 10 mg/kg every two weeks or a fixed dose of 200 mg every three weeks [8,11,84,85]. This was a heavily pretreated population with 45 percent having received at least three prior therapies. Objective responses were observed in 34 patients (18 percent). The response rate was similar in human papillomavirus (HPV)-positive and HPV-negative patients. Responses lasted ≥6 months in 85 percent (range 2+ to 30+ months), and OS at 12 months was 38 percent.

●In the KEYNOTE-055 study, 171 patients were treated with pembrolizumab 200 mg every three weeks [86]. All patients had platinum- and cetuximab-refractory head and neck cancer. The objective response rate was 16 percent, and the median duration of response was eight months. Response rates were similar in patients who were HPV positive and HPV negative, and in the various PD-L1 subgroups.

In both trials, the toxicity was similar to that reported with pembrolizumab in other indications, except for a higher incidence of hypothyroidism (15 percent). (See "Toxicities associated with immune checkpoint inhibitors".)

Nivolumab — Nivolumab (table 3) is a PD-1 inhibitor that has demonstrated activity in previously treated patients, including those with prior cetuximab (table 5) exposure. In one prospective trial, objective responses were seen in 13 percent, and nivolumab demonstrated a modest survival benefit with less toxicity when compared with methotrexate, docetaxel, or cetuximab (table 5).

In the Checkmate-141 phase 3 trial, 361 patients with platinum-refractory, recurrent or metastatic disease were randomly assigned to either nivolumab (3 mg/kg every two weeks) or a single-agent investigator's choice of therapy (methotrexate, docetaxel, or cetuximab (table 5)) [9,87-90].

●With a minimum follow-up of 11.4 months, OS for the entire study population was longer in patients treated with nivolumab (median 7.7 versus 5.1 months, one-year survival rate 34 versus 19.7 percent, HR 0.71, 95% CI 0.55-0.9). The objective response rate was also increased with nivolumab (13.3 versus 5.8 percent).

●In a prespecified exploratory analysis, OS was increased with nivolumab in patients with PD-L1 expression ≥1 percent (8.7 versus 4.6 months, HR 0.55, 95% CI 0.36-0.83). OS was not significantly increased in patients with PD-L1 expression <1 percent (HR 0.89, 95% CI 0.54-1.45).

●In a post hoc exploratory analysis, OS was increased in patients with HPV-positive tumors treated with nivolumab (median 9.1 versus 4.4 months, HR 0.56, 95% CI 0.32-0.99). The difference in OS was not statistically significant in those with HPV-negative tumors (median 7.5 versus 5.8 months, HR 0.73, 95% CI 0.42-1.25).

Nivolumab was not routinely available for those patients randomized to chemotherapy, and routine use of nivolumab after second-line chemotherapy could negate or minimize the survival benefit of immediate nivolumab.

●In a subsequent post hoc exploratory analysis, nivolumab improved OS among those without prior cetuximab exposure (median 8.2 versus 4.9 months, HR 0.52, 95% CI 0.35-0.77) [90]. For those with prior cetuximab exposure, OS with nivolumab was similar to that seen with chemotherapy (median OS 7.1 versus 5.1 months, HR 0.84, 95% CI 0.62-1.15). Among those treated with nivolumab, estimated 12-month OS was similar, regardless of previous cetuximab exposure (39 and 31 percent for those without or with cetuximab exposure, respectively).

●Patient-reported outcomes on a quality of life questionnaire found that there was no clinically meaningful deterioration associated with treatment with nivolumab, in contrast to chemotherapy, where quality of life significantly decreased in 8 of 15 domains [87].

Based on the results of this phase 3 trial, nivolumab (table 3) was approved by the FDA [91].

Other agents

●Durvalumab with or without tremelimumab – Durvalumab, a PD-L1 inhibitor, does not have an established role as subsequent therapy in patients with recurrent or metastatic disease. While durvalumab showed initial clinical activity in a phase II study [92], it failed to improve OS compared with standard chemotherapy in randomized phase III trials, either as a single agent or in combination with the CTLA-4 inhibitor tremelimumab [93,94].

Cytotoxic chemotherapy — For patients who are candidates for further lines of cytotoxic chemotherapy, the choice of agent is based on prior treatment history and the overall condition of the patient. Patients who receive ICIs as part of initial therapy may demonstrate increased tumor sensitivity to subsequent-line chemotherapy [54,95]. For patients who progress through initial chemotherapy alone, objective responses to subsequent lines of single-agent cytotoxic chemotherapy are uncommon, particularly when contemporary response criteria are applied. There is no evidence that single-agent chemotherapy prolongs OS.

●Methotrexate – The limited activity of single-agent methotrexate in this setting is illustrated by a phase III trial that compared weekly intravenous methotrexate with gefitinib, an epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, in a heavily pretreated population [36]. Among the 152 patients treated with methotrexate, the overall response rate was 4 percent, and the median OS was seven months [36].

●Platinum and taxanes – Other single agents, such as cisplatin or docetaxel (table 4), have produced higher response rates than methotrexate in randomized trials, but this has generally been associated with increased toxicity.

One randomized trial (KEYNOTE-048) suggested a treatment benefit with taxanes in patients initially treated with ICI-based regimens [54]. However, these agents have not been shown to affect OS in patients who progress on chemotherapy alone [14-16,26].

●Gemcitabine – Gemcitabine has also shown evidence of activity in several observational studies and may represent a single-agent option for some previously treated patients [37,38].

Toxicity profiles for different chemotherapy agents vary widely and may be an important consideration when choosing a regimen. As an example, single-agent methotrexate causes more mucositis than cisplatin, but less nausea or vomiting [14]. When weekly docetaxel was compared with methotrexate, docetaxel was associated with grade 3 or 4 neutropenia (13 percent), febrile neutropenia (1 percent), anemia (19 percent), mucositis (9 percent), and ungual (nail) toxicity (pain and ungual loss, 9 percent), whereas methotrexate was associated with grade 3 or 4 anemia (15 percent) and mucositis (5 percent) [26].

Cetuximab — Cetuximab (table 5) is a monoclonal antibody that targets the epidermal growth factor receptor (EGFR), which is overexpressed in most squamous head and neck cancers. The activity of cetuximab in patients refractory to platinum-based chemotherapy has been demonstrated in several multicenter phase II trials, all of which observed an objective response rate of approximately 10 percent [45,96,97]. It appears to be effective, with approximately the same rate of benefit regardless of prior therapy. (See "Treatment protocols for squamous cell carcinoma of the head and neck".)

Cetuximab is administered as monotherapy either weekly or every two weeks (table 5). Clinicians may offer cetuximab on an every-two-week schedule as a convenient and alternative dosing to the weekly schedule, both of which have approval from the FDA in squamous cell carcinoma of the head and neck [58]. The standard schedule for cetuximab is 250 mg/m² weekly, and further dose escalation does not appear to be beneficial. Another study in 35 patients evaluated cetuximab at a dose of 500 mg/m² every two weeks; treatment was well tolerated and had similar efficacy (response rate 11 percent) [98]. Further dose escalation (750 mg/m²) on the every-two-week schedule did not result in any improvement in efficacy in another 26 patients (response rate 8 percent).

The activity of cetuximab is illustrated by a series of 103 patients with progressive disease following two to six cycles of platinum-based therapy [45]. Cetuximab was given with an initial dose of 400 mg/m², followed by 250 mg/m² weekly. Overall, 17 patients (13 percent) had an objective response, including 8 of 79 (10 percent) who had had progression with the cisplatin chemotherapy and 9 of 51 (18 percent) who had had stable disease. The median survival from the initiation of cetuximab therapy was 7.5 months.

INVESTIGATIONAL APPROACHES

Farnesyltransferase inhibitors — In patients with heavily pretreated HRAS mutant squamous cell carcinoma of the head and neck, the farnesyltransferase inhibitor tipifarnib demonstrated objective response rates of up to 56 percent in phase II trials [99]. These data are encouraging, and further studies are necessary before incorporating this approach into routine clinical practice. The use of farnesyltransferase inhibitors in head and neck and salivary gland tumors is discussed separately. (See "Head and neck squamous cell carcinogenesis: Molecular and genetic alterations", section on 'Ras-mediated pathway: Raf/MEK/ERK and PI3K' and "Malignant salivary gland tumors: Treatment of recurrent and metastatic disease", section on 'Investigational agents'.)

Other agents

●Cetuximab-based combination regimens – Various cetuximab-based combination regimens are being investigated in treatment-refractory recurrent or metastatic head and neck cancer. These include pembrolizumab plus cetuximab [100], nivolumab plus cetuximab [101], and ficlatuzumab plus cetuximab [102].

●Metronomic chemotherapy – Metronomic chemotherapy uses frequent, lower doses of chemotherapy over long periods of time to target tumor angiogenesis and minimize toxicity. Metronomic chemotherapy has been evaluated as initial and subsequent-line therapy in resource-limited areas [103,104]. This approach remains investigational.

The addition of low-dose nivolumab to metronomic chemotherapy (methotrexate with erlotinib and celecoxib) improved OS in a randomized phase III trial of patients with recurrent or newly diagnosed advanced head and neck cancer [104]. Initial reports also suggest that oral metronomic chemotherapy is active and well tolerated in patients with platinum-resistant disease [103], although this approach has not been evaluated in those receiving prior immunotherapy.

●Palbociclib – The clinical role of palbociclib, a cyclin-dependent kinase (CDK) inhibitor specific to CDK4 and CDK6, remains investigational in patients with recurrent or metastatic head and neck squamous cell carcinoma. Palbociclib has been evaluated in combination with cetuximab in patients with HPV negative disease resistant to platinum and/or cetuximab [105-107]. While initial trials suggested that palbociclib plus cetuximab could potentially reverse previous cetuximab resistance [106], a randomized phase II trial (PALATINUS) demonstrated that this combination did not improve OS in patients with platinum-resistant disease [107,108].

●Small molecule tyrosine kinase inhibitors – A number of small molecule tyrosine kinase inhibitors have been tested in patients with advanced head and neck cancer, including afatinib, gefitinib, sunitinib, erlotinib, lapatinib, and tivantinib. No clear clinical role has yet been established for this approach [32,34-36,109-112].

Large phase III trials have compared two EGFR inhibitors, gefitinib and afatinib, both alone and in combination with docetaxel, with standard chemotherapy [36,110,111,113]. None of these have demonstrated a significant increase in OS.

SPECIAL POPULATIONS

High tumor mutational burden — The management of patients with recurrent or metastatic head and neck cancer whose tumors express high levels of tumor mutational burden is discussed separately. (See "Tissue-agnostic cancer therapy: DNA mismatch repair deficiency, tumor mutational burden, and response to immune checkpoint blockade in solid tumors".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Head and neck cancer".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Throat cancer (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Treatment approach – The management of patients with no prior systemic therapy for recurrent or metastatic squamous cell carcinoma of the head and neck is influenced by prior initial treatment for locoregional disease, performance status, comorbidities, tissue programmed death ligand 1 (PD-L1) combined positive score (CPS), and related prognostic factors. (See 'General approach' above and 'Previously untreated patients' above.)

Multiple active systemic agents are available, including chemotherapy, immune checkpoint inhibitors (ICIs), and targeted therapy such as cetuximab (table 5). (See 'Active agents' above.)

●Overall survival benefit for immunotherapy plus chemotherapy – Immunotherapy with the programmed cell death protein 1 (PD-1) inhibitor pembrolizumab plus chemotherapy (table 6) improves overall survival (OS) compared with cetuximab plus chemotherapy. For those with a high CPS, single-agent pembrolizumab (table 2) also improves OS compared with cetuximab plus chemotherapy. (See 'Pembrolizumab with or without platinum and fluorouracil' above.).

●No prior systemic therapy, CPS <1 or unavailable – For patients with no prior systemic therapy for recurrent or metastatic disease, and who have a PD-L1 CPS of <1 or unavailable, we suggest platinum-based chemotherapy, with or without pembrolizumab, rather than other chemotherapy regimens (Grade 2C) (algorithm 1). (See 'Selection of therapy' above.)

•For those in whom pembrolizumab is not used, concurrent or sequential cetuximab (table 5) with combination chemotherapy (platinum with either fluorouracil or taxane) is a reasonable alternative. (See 'Cetuximab with platinum-based chemotherapy' above.)

•For those who are not eligible for combination chemotherapy and pembrolizumab, options include single-agent therapy with a platinum (typically carboplatin), taxanes (eg, paclitaxel, docetaxel (table 4), nabpaclitaxel), fluorouracil, methotrexate and cetuximab (table 5). (See 'Active agents' above.)

●No prior systemic therapy, CPS ≥1 – For patients with no prior systemic therapy for recurrent or metastatic disease who have PD-L1 CPS ≥1, we recommend pembrolizumab-based therapy rather than cetuximab plus chemotherapy or chemotherapy alone (Grade 1B) (algorithm 1). For those who are eligible for combination therapy but have a contraindication to pembrolizumab, options include a doublet cytotoxic chemotherapy regimen, with or without cetuximab. (See 'Selection of therapy' above.)

When using pembrolizumab-based therapy, our approach is dependent on the rapidity of tumor progression and PD-L1 CPS.

•Patients with rapidly progressive disease – For patients with rapidly progressive disease who are eligible for combination chemotherapy plus immunotherapy, we suggest pembrolizumab with a platinum plus fluorouracil (table 6) rather than pembrolizumab alone (Grade 2C), regardless of CPS. Alternatively, some contributors substitute a taxane (paclitaxel or docetaxel) for the fluorouracil. (See 'Pembrolizumab with or without platinum and fluorouracil' above and 'Pembrolizumab with platinum plus taxane' above.)

However, for patients who are not eligible for combination therapy, pembrolizumab monotherapy (table 2) is an appropriate alternative.

•Patients without rapidly progressive disease – For most patients without rapidly progressive disease, our preferred regimen depends on the tissue CPS, as follows (see 'Pembrolizumab with or without platinum and fluorouracil' above and 'Pembrolizumab with platinum plus taxane' above):

-CPS ≥20 – We suggest single-agent pembrolizumab (table 2) rather than pembrolizumab plus chemotherapy (Grade 2C).

-CPS ≥1 and <20 – For patients with a good performance status (eg, Eastern Cooperative Oncology Group [ECOG] performance status ≤2 (table 1)), we suggest pembrolizumab with platinum-based chemotherapy (table 6) rather than single-agent pembrolizumab (Grade 2C). While chemotherapy regimens containing platinum plus fluorouracil are preferred, some contributors alternatively substitute a taxane (paclitaxel or docetaxel) for the fluorouracil.

For those with a lesser performance status where it is preferred to avoid the potential toxicities of systemic chemotherapy, single-agent pembrolizumab is a reasonable alternative (table 2).

●Prior systemic therapy for locoregional disease – For patients with no prior systemic therapy for recurrent or metastatic disease who previously received initial systemic therapy for locoregional disease with a platinum-based regimen (with or without cetuximab) and are eligible for combination therapy, treatment suggestions are based on duration of disease control. (See 'Previous systemic therapy for locoregional disease' above.)

•Short period of disease control – For most patients with a relatively short period of disease control (<6 months), we suggest immunotherapy with a PD-1 antibody (ie, nivolumab (table 3) or pembrolizumab (table 2)) rather than single-agent or combination therapy with a platinum, taxane, cetuximab (table 5), or methotrexate (Grade 2B). Further chemotherapy is unlikely to offer sustained benefit in this population due to a high likelihood of chemotherapy resistance. Inclusion in a formal clinical trial is an appropriate alternative. (See 'PD-1 inhibitor immunotherapy' above.)

•Prolonged period of disease control – For patients with a relatively prolonged period of disease control (≥6 months), we offer pembrolizumab, with or without platinum-based chemotherapy, similar to the approach for previously untreated patients. Carboplatin is the preferred platinum for patients who have already received cisplatin. (See 'Pembrolizumab with or without platinum and fluorouracil' above.)

●Subsequent therapy – For patients who relapse after receiving prior systemic therapy for metastatic or recurrent disease, the choice of subsequent regimens is based on previous systemic therapy exposure(s), performance status, and comorbidities (see 'Subsequent therapy' above):

•For patients with progression following initial treatment with systemic platinum-based chemotherapy (with or without cetuximab) and who are eligible for immunotherapy, we suggest immunotherapy with a PD-1 antibody (pembrolizumab (table 2) or nivolumab (table 3)) rather than further lines of chemotherapy (Grade 2B). (See 'PD-1 inhibitor immunotherapy' above.)

•For patients with progression following initial treatment with systemic platinum-based chemotherapy (with or without cetuximab) and who are ineligible for immunotherapy, options include single-agent cetuximab if not previously received (table 5); taxanes (table 4) if not previously received; a platinum for patients with disease progression more than six months after prior platinum-based chemotherapy (preferably carboplatin if previous cisplatin exposure); fluorouracil; or methotrexate. (See 'Active agents' above and 'Cytotoxic chemotherapy' above and 'Cetuximab' above.)

•For patients with progression on immunotherapy alone and no prior exposure to cetuximab, options include combinations of a platinum plus a taxane or a platinum plus fluorouracil, with or without the use of cetuximab. (See 'Cetuximab with platinum-based chemotherapy' above and 'Platinum and fluorouracil' above and 'Platinum plus taxane' above.)

•For patients with progression on combination immunotherapy and platinum plus fluorouracil-based chemotherapy, options include cetuximab alone (table 5) or cetuximab in combination taxane-based chemotherapy. (See 'Cetuximab' above and 'Cetuximab with platinum-based chemotherapy' above.)

●Clinical trials – If available, patients may be enrolled in clinical trials at any time during their treatment course.

●Indications for supportive care – Patients with severe comorbidities and/or a poor performance status may be best treated with supportive care, without initiating systemic therapy. (See "Management and prevention of complications during initial treatment of head and neck cancer" and "The role of parenteral and enteral/oral nutritional support in patients with cancer", section on 'Head and neck cancer' and "Overview of managing common non-pain symptoms in palliative care".)