Salvage of Locoregionally Recurrent Head and Neck Cancer: An NRG Oncology Working Group Review

Abstract

Despite decades of progress in the treatment of head and neck cancer, recurrence and second primary cancers continue to occur. The management of non-metastatic, locoregionally recurrent or second primary cancers is a complex multidisciplinary challenge that often occurs without guidance from robust clinical trials. In 2023, the NRG Oncology cooperative group created the Recurrent and Metastatic Working Group with the express directive to investigate how resources could optimally address the key questions for the recurrent and metastatic populations. Here, we present our view of the state of the science and present considerations for future investigations.

Introduction

Recent progress in the management of head and neck cancer has led to improved cancer outcomes and reduced sequelae compared to prior decades. Unfortunately, despite these improvements, locoregionally recurrent disease remains a vexing problem necessitating continued use of salvage treatment strategies.

The most common location of recurrence for mucosal squamous carcinoma of the head and neck is locoregional, the majority of which occur within the first two years of definitive treatment.(1,2) Furthermore, an additional 10-20% of long-term survivors will develop a second primary cancer in the aerodigestive tract. The management of locoregionally recurrent or second primary (SP), non-metastatic head and neck cancer is complex, necessitating a multidisciplinary team approach for treatment selection. The heterogeneous nature of locoregional recurrence makes clinical trials challenging in this setting, and therefore treatment paradigms are often poorly defined. Multi-disciplinary teams are therefore left to design institutional paradigms, tailored to specific situations. With such a varied landscape and a lack of robust prospective clinical trial data, many providers may disagree on the optimal utilization and delivery of salvage therapies.

Considering the diverse landscape, researchers also often disagree about the key clinical trial questions to address. In 2023, the Head and Neck Core Committee of the NRG Oncology cooperative group created the Recurrent and Metastatic Working Group (RMHNWG) with the express directive to further investigate how cooperative group resources could optimally be focused to address the key questions in the recurrent and metastatic populations, to inform clinical trial design. Here, we present our view of the state of the science, including a review of NRG Oncology (and other cooperative groups’) previous efforts, and present considerations for future investigations. In this manuscript, we focus specifically on the management of locoregionally recurrent or second primary (RSP) mucosal squamous cell carcinomas without evidence of distant metastases.

Salvage Surgery

When considering mucosal squamous cancers of the head and neck that have recurred after previous surgery or radiation without evidence of distant metastases, it is universally accepted that patients who are candidates for salvage surgery should be considered for such. There are a few exceptions, such as non-operative salvage of a recurrent glottic larynx cancer initially treated only with resection, but in general salvage surgery remains the cornerstone of treatment for locoregionally recurrent or second primary cancers. Typically, most guidelines suggested sequencing of salvage surgery followed by adjuvant therapy as indicated by pathologic risk factors with re-irradiation, chemotherapy and/or checkpoint inhibitors, as discussed below.

Fortunately, there has been significant improvement of outcomes after salvage surgery over the past decades. Patient selection for any salvage has evolved with an improved understanding of prognostic factors.(3) The addition of transoral robotic surgery (TORS) and other endoscopic techniques provide minimally invasive options for small recurrences of the larynx and pharynx.(4–6) For larger recurrences, free-flap reconstruction is often required, and improved reconstruction techniques have continued to enhance cosmetic and functional outcomes. In the setting of previous radiation, free flap reconstruction provides an added benefit of providing non-irradiated tissue into the bed, which may facilitate safety of re-irradiation.

Considering the key focus of this manuscript is to expound on future cooperative group efforts, more emphasis on enumerating the many nuances of surgical salvage is beyond the intent of this paper. We do expect many lessons to come from the recently activated NRG HN015 (NCT07195734) which is discussed below and will inform the committee on salvage surgical techniques and standardization efforts. The working group acknowledges that improvements in salvage surgical outcomes have greatly facilitated the possibility of trials such as HN015. The task force therefore felt that the greatest opportunities for the resectable population may lie in studying peri-operative adjuvant or neoadjuvant strategies to improve cancer-specific outcomes.

Key Question: Is there a role for anti-PD-1 therapies perioperative to salvage surgery, and what is the optimal timing?

Systemic therapy for recurrent head and neck cancer has evolved dramatically since the introduction of anti-PD-1 “checkpoint” inhibitors. In recurrent or metastatic trials where it is assumed that the disease extent was not easily amenable to surgical salvage (or effective localized radiation), the addition of immunotherapy has produced superior outcomes over previous chemotherapy-based regimens. For example, the EXTREME trial established a median survival of 10 months and response rate of 36% for patients with recurrent or metastatic head and neck cancers (HNC) receiving combination chemotherapy with cetuximab.(7) Anti-PD-1 therapies have improved on these outcomes, with KEYNOTE-048 establishing a median survival of 13 months and response rate of 36% for patients with recurrent or metastatic HNC receiving combination chemotherapy with pembrolizumab.(8) Of note, prospective clinical trials pertinent to the discussion throughout this manuscript are summarized for the reader’s convenience in Table S1.

Platinum-refractory recurrent HNC represents a subset with a poorer prognosis compared to those who have not yet received platinum-based therapy (as with EXTREME). Nevertheless, in CheckMate 141, immunotherapy with nivolumab was superior to chemotherapy, achieving median survival of 7.5 months (compared to 5.1 months) and response rate of 13.3% (versus 5.8%).(9) With significant survival advantages over conventional chemotherapy regimens, NCCN guidelines have recommended the use of immunotherapy as the preferred first-line systemic therapy option in the recurrent or metastatic settings.(10)

Given the progress seen in the recurrent or metastatic setting where definitive local therapies are not an option, investigators have considered whether immunotherapy-based perioperative strategies might improve outcomes for patients eligible for salvage surgery. This strategy is enticing as neoadjuvant therapy has the potential to mitigate technical challenges of resectability common in the salvage scenario.(11) Recently, data from KEYNOTE-689 demonstrated a significant EFS benefit to perioperative pembrolizumab over standard of care in the de novo (untreated) locally advanced setting (3-year EFS 45.9% vs 59.8%, HR 0.66, 95% CI 0.49-0.88).(12) Similarly, preliminary data from NIVOPOSTOP, a randomized phase III trial that delivered nivolumab following surgical resection and chemoradiation in patients with positive margins or extracapsular extension demonstrated that nivolumab significantly improved the primary endpoint of DFS (3-year 52.5% vs 63.1%, HR 0.76, 95% CI 0.60-0.98).(13) Furthermore, in a phase II study combining PD-1 and CTLA-4 checkpoint inhibitors prior to resection of previously untreated locally-advanced oral cavity cancers, the combined arm reported 33% major pathologic response rates without surgical delays.(14) These perioperative/adjuvant approaches could be similarly applied to the salvage setting, where historical outcomes have been disappointing.(12,15–18)

One salvage study by the Dana-Farber group investigated neoadjuvant nivolumab and lirilumab prior to salvage surgery and demonstrated a 43% pathologic response rate (14% major pathologic response rate, 0% complete response).(19) This response rate was considered favorable given the expected <20% rate for single-agent immunotherapy in untreated cancers.(20),(21) Given the possible benefit in the salvage population and paucity of data, this strategy is worthy of further investigation.(22)

An alternative to neoadjuvant therapy is to deliver anti-PD-1 therapies in the adjuvant setting, after a salvage surgery. Such an approach allows for pathologic evaluation prior to delivery of therapy and may facilitate informed patient selection. This was one of the six cohorts included in the randomized phase II PATHWay study, which enrolled patients considered to have a predicted >40% risk of recurrence. Patients were then randomized to observation (standard of care) vs. pembrolizumab. The preliminary findings among all six cohorts combined demonstrated a statistically significant improvement in 2-year progression-free survival (PFS) from 33% to 54% with pembrolizumab.(23) Unfortunately, given the multiple cohorts included in the PATHWay study, this data is under-powered to draw firm conclusions specific to the adjuvant treatment after salvage surgery. This approach was also studied in a single-institution phase II trial of nivolumab after salvage surgery, which met its primary endpoint of a 2-year disease-free survival of 71%, improved compared to historical controls.(24) The ongoing ECOG EA3191 study, which includes patients who have undergone salvage surgery, randomizes patients to adjuvant chemotherapy concurrent with re-irradiation (ReRT) vs. pembrolizumab alone; this study will further clarify the role of adjuvant PD1 and is discussed in more detail below.

One possibility which may boost response rates to neoadjuvant immunotherapy is the addition of SBRT, which may be synergistic. This has only been tested in de novo HPV-negative patients but deserves discussion as a possible strategy which could be applied to the salvage setting. On this phase I/Ib trial, neoadjuvant durvalumab was delivered prior to surgical resection, and robust correlative analyses of immune cell infiltration into the tumor microenvironment were performed.(25) At the maximum tolerated dose of 24 Gy in 3 fractions, an impressive major pathologic (MPR) or complete response (CR) rate of 89% (95% CI 57.1%-100%) was observed. This surprisingly high rate of MPR and CR was attributed to the synergistic effect of hypofractionated radiotherapy given the lower results observed with other drug-only regimens and the increases in effector T-cells and antigen presentation post-treatment. While the study is impressive, the generalizability and extent of technical eligibility of patients require further testing, and outcomes in patients who require postoperative radiation should be assessed at long-term follow up. Overall, this intriguing strategy requires further confirmatory study but certainly deserves attention.

Taken together, the working group supports the utilization of cooperative group resources to investigate the role of novel systemic therapies and immunotherapies to improve outcomes of salvage surgeries. At present, the group is currently in the process of activating a phase II trial in this space to address this key question, the NRG Oncology trial NRG-HN015 (NCT07195734).

Key Question: Defining Appropriate Biomarkers to Guide Salvage Therapy

The committee remains very interested in novel biomarkers that may guide treatment decisions or otherwise improve outcomes after salvage therapy. The possibilities here are vast, but examples of emerging strategies include the use of circulating tumor DNA (or its relatives)(26), artificial-intelligence based assays(27), genomic assays(28) or novel imaging techniques (F18-FMISO PET)(29). In the re-irradiation setting, early reports suggest a prognostic value toward ctDNA.(30) The utility of F18-FMISO PET, however, may be impacted by the initial course of radiation and may not be of significant value in patients with locoregionally recurrent cancers.(31) The other proposed biomarkers have not been specifically investigated in the locoregionally recurrent setting to our knowledge.

Biomarkers can be classified as integral or integrated. Randomization to undergo salvage surgery based on ctDNA results, for instance, would be an example of a schema that directly tests the utility of an integral biomarker. Including ctDNA as a correlative in patients undergoing salvage surgery would be an example of testing an integrated biomarker. While the former “direct” strategy can provide the strongest validation, it also requires the most robust pilot data, which remains in short supply for this population. At this time, the committee would support concepts investigating either approach, although incorporation of integrated biomarkers seems most likely at present.

Adjuvant Radiation Therapy

For patients who have not previously received radiation therapy (RT), adjuvant RT after salvage surgery is typically delivered based on the same principles used for de novo disease. In other words, if there are risk factors such as margin positivity, extracapsular extension (ECE), lymphovascular space invasion (LVSI), depth of invasion (DOI), perineural invasion (PNI), pT3-4 AJCC classification, and node-positive disease then adjuvant RT is indicated. Less certain risk factors are “worst pattern of invasion” (WPOI) or close margins. It is unclear how these factors should be adjusted in the locoregionally recurrent and resected setting, considering recurrence itself is a negative prognostic factor.(32) While a trial clarifying the value of first-time adjuvant RT in the locoregionally recurrent setting would be academically interesting, enthusiasm amongst the task force is limited, viewing this as a impact and lower priority question.

For patients who have previously received RT overlapping the resected area, the role of adjuvant ReRT is a high-priority question and deserves investigation as described below.

Key Question: What is the benefit of modern ReRT in the adjuvant salvage setting?

There is very limited randomized data to support the role of ReRT in any setting, but one “positive” trial of ReRT exists in the adjuvant setting. The GORTEC cooperative group trial evaluated ReRT of 60 Gy over 11 weeks (using a 1 week on, 1 week off schedule) combined with 5-FU and hydroxyurea chemotherapy versus observation in 130 patients who underwent salvage surgery with macroscopic complete resection. The trial demonstrated a DFS benefit to adjuvant ReRT with a HR of 1.68 (95% CI 1.13-2.5, p=0.01), but this study used drugs, radiation techniques and schedules that have since fallen out of favor.(33) With advances in RT technologies and systemic therapies, the benefit of adjuvant ReRT in the modern era is speculated to be greater than previously demonstrated with likely less toxicity, though this is not yet proven.

The phase II ECOG EA3191 trial currently randomizes patients between ReRT with chemotherapy or adjuvant pembrolizumab without re-irradiation. A third arm including ReRT with pembrolizumab was terminated to facilitate enrollment. Activated in January 2021, this trial is moving forward with accrual of 84 of a planned 188 patients enrolled as of June 2025.(34)

Patient selection for adjuvant ReRT is challenging and unclear. In general, patients with a high functional status (KPS >80, ECOG 0-1), lower competing risks, and lower disease burden who are felt to be at a high risk of recurrence could be considered candidates for re-treatment. Clinical trials in this space will need to carefully define the most appropriate candidates. EA3191 has appropriately chosen the highest risk patients with extracapsular extension (ECE) or positive margins, a clear and simple criteria.

Target volumes have become more conformal with modern ReRT techniques. For instance, the EA3191 study treats the pre-operative volume + 7 mm including the postoperative bed but excluding anatomic boundaries.(34) The optimal volumes of elective nodal irradiation or other margin expansions remains unclear, but currently there is little enthusiasm among the task force to investigate this highly technical and nuanced question.

Dose and fractionation of ReRT in the adjuvant setting remains controversial. Daily fractionation is the subject of EA3191, but hyperfractionation has experienced a resurgence of interest since the publication of a randomized trial of patients with recurrent nasopharynx cancer.(35) This study was a phase III randomized trial from China of locally advanced recurrent nasopharynx cancers which compared hyperfractionation of 65 Gy in 54 fractions twice daily with once daily 60 Gy in 27 fractions, both with IMRT. This study found not only a 23% lower incidence of late grade 3+ toxicity in the hyperfractionation group, but also an improved 3-year overall survival of 74.6% in the hyperfractionated versus 55.0% in the daily fractionated arm. Extrapolating these results to the adjuvant setting, many are reconsidering the role of hyperfractionation.

Though hyperfractionated ReRT appears to be regaining interest, the logistical burden is significant, and other institutions have studied the role of extreme hypofractionation known as stereotactic body radiotherapy (SBRT) or stereotactic ablative body radiotherapy (SABR). These courses are typically delivered in 5 or fewer fractions, at 6-8.8 Gy per fraction, using highly conformal techniques. The logistical advantages are obvious compared to >50 twice daily fractions, but such a schedule may also carry radiobiologic benefits by overcoming accelerated repopulation and sublethal repair. The data to support postoperative/adjuvant ReRT remains limited, however, with one retrospective review of 28 patients forming the majority of the literature on the topic.(36)

Proton therapy is a rapidly maturing technology of interest which could decrease the volume of normal tissues twice-irradiated. Proton units are often capable of ultra-fast dose rates (FLASH) which may improve normal tissue sparing based on animal and preclinical data.(37) Further development of proton therapy, or other particle therapies, in the ReRT setting is of great interest to the task force. With the expansion of proton centers nationally, clinical trials may be more feasible in the current era compared to even a decade prior. Currently, however, adjuvant ReRT with proton therapy has remained clinically challenging and associated with osteoradionecrosis rates up to nearly 20%.(38) Future studies will need to consider these pragmatic points and define the techniques that can maximize the physical benefits of proton therapy.

Concurrent therapy with adjuvant ReRT remains under-investigated in the modern era. Off protocol, investigators have delivered a variety of radiosensitizers, or given the lack of benefit, often opted for radiation alone.(39) Previous platinum therapy complicates subsequent eligibility for systemic therapies and trials of adjuvant ReRT will need to remain flexible and pragmatic, likely allowing numerous reasonable choices such as cetuximab, carboplatin/paclitaxel, cisplatin, or radiation alone. The working group is concerned that concurrent therapy with adjuvant re-irradiation would be a difficult subtopic to tackle in the current environment and therefore ranked it a lower priority than establishing a benefit to adjuvant ReRT in general.

Unresectable Locoregional Recurrence

For many patients who have locoregional recurrence, there are technical, functional, or medical comorbidities that will preclude resection and favor a non-operative approach.

For patients with unresectable disease who have not previously received RT, or who received RT to another non-overlapping site, definitive-intent RT with neoadjuvant and/or concurrent systemic therapy is the treatment of choice, extrapolating from the treatment of de novo disease. The working group feels that given the continued option of first-time definitive RT, this is not a high-priority population to study in cooperative group trials at this time.

Recurrence after Previous Radiation Therapy

Patients who previously received RT and then experience a locoregional recurrence or SP within the RT field are a unique and unfortunately common population. If these tumors are deemed unresectable for one of the previously aforementioned reasons, the next steps are not well defined by the literature. NCCN, for instance, allows for systemic therapy, ReRT, clinical trial, or best supportive care in this scenario; a remarkably wide range of options.(10)

Chemoimmunotherapy is one very viable and commonly deployed standard options in this situation. The KEYNOTE-048 trial included patients with non-metastatic recurrences, but these were a minority of patients enrolled, at approximately 29% of patients in the trial. The median survival of 13.0 months achieved with pembrolizumab + chemotherapy in the trial as a whole is favorable compared to older studies of definitive ReRT, and therefore many clinicians feel that systemic therapy is a valid standard of care for patients with non-metastatic locoregional recurrences within a previously irradiated field.

Beyond KEYNOTE-048, one phase III trial known as LEAP-010 investigated the addition of lenvatinib to pembrolizumab in patients with recurrent or metastatic head and neck squamous carcinoma with a CPS score of ≥1.(40) The study is currently available in abstract form, so it is not yet published what fraction of patients were locoregionally in this study. Unfortunately, this appropriately powered trial did not demonstrate a survival benefit.

Furthermore, there are other novel emerging systemic agents which could eventually be shown to improve outcomes in this population. BC101, known as ficerafusp alfa, is a bifunctional EGFR antibody fused to a TGFβ immunomodulator, which was well tolerated with a 44% response rate on phase I trial in recurrent and metastatic patients. A phase II/III study was recently activated.(41) Similarly, petosemtamab is a bispecific antibody targeting EGFR and LGR5 which has now led to a phase III trial.(42) Danvatirsen, a novel STAT3 inhibitor, has shown encouraging response rates in combination with durvalumab and is now under phase II investigation in the PEMDA-HN trial.(43,44) Each of these novel therapies, among others, is an encouraging development which could improve outcomes for recurrent patients treated with systemic therapy. Thus, the future question related to anti-PD-1 and other emerging systemic strategies will be whether and how to integrate them with local therapy, and conversely whether novel agents can act as radiosensitizers to enhance the response to ReRT.

Key Question: What is the benefit of definitive re-irradiation in the modern era?

ReRT is one standard strategy to address unresectable recurrences in a previously-irradiated field. ReRT gained popularity over 25 years ago and remains an appropriate option in guideline statements for this specific population.(10,45) The RTOG has a legacy of trials in this space, with RTOG 9911 and RTOG 9610 as some of the most oft-quoted trials establishing ReRT as an acceptable option.(46,47) These trials delivered hyperfractionated split-course 3-dimensional radiation therapy (3DCRT) with concomitant chemotherapy and demonstrated median survivals of 12 and 8.5 months, respectively. Grade 5 toxicity rates were approximately 8% in both studies.

Notwithstanding these clinical trial data, practitioners remain divided regarding the appropriateness of definitive-intent, large-volume, fractionated ReRT. Some feel that its utility is limited and futile, and the burden of several weeks of treatment is often difficult for this patient population, especially those who have continued severe normal tissue sequelae from prior RT. On the other hand, in patients with adequate performance status, ReRT is delivered fairly routinely in many practices as an assumed standard of care. The absence of large, prospective, randomized clinical trial data in this space makes resolving this divide challenging.

Recent data from a single-arm multi-institution prospective trial investigated modern ReRT with nivolumab prior to, during and after full-dose ReRT. ReRT was delivered to 60-66 Gy in 30-33 fractions (investigator discretion) over 6-6.5 weeks with nivolumab, 240 mg, intravenously 2 weeks prior and every 2 weeks for 5 cycles during IMRT, then nivolumab, 480 mg, intravenously every 4 weeks for a total nivolumab duration of 52 weeks. With a median follow-up of 24.5 months (95% CI, 19.0-25.0), the estimated 1-year PFS was 61.7% (95% CI, 49.2%-77.4%). Patient reported outcomes (PROs) were favorable at a median follow-up of 24.5 months, with no grade 5 toxicities reported. This modern-era trial has fueled enthusiasm for the re-evaluation of ReRT. In addition, there is the promise of even further improvements using modern-era advanced RT technologies. For instance, a large single institution study using reported a 71.8% response rate and nearly 2-year median survival for patients treated with definitive proton ReRT.(48)

Meanwhile, SBRT has emerged as another possible novel technological option.(49–53) Retrospective data has supported SBRT as a viable option which is logistically favorable compared to hyperfractionated approaches, which can be burdensome.(54) RTOG Foundation 3507 is investigating the role of SBRT for smaller local or locoregional recurrences, given with or without pembrolizumab.

Randomized data clearly demonstrating a benefit to either long-course, fractionated or short-course, stereotactic definitive ReRT continues to elude our clinical community. The legacy RTOG 0421 phase III trial of chemotherapy with or without ReRT failed to accrue, which has limited enthusiasm for developing future trials directly testing this question. A head-to-head randomization to systemic therapy alone vs. a proposed ReRT treatment would likely be complex, with patient and provider biases impacting enrollment. Next-generation trials will need to overcome these challenges in order to provide a definitive answer to this question.

Meanwhile, improvements in systemic therapy are often cited as a rationale to avoid ReRT. Proponents of this approach suggest the improved overall survival in KEYNOTE-048 to 13 months, with the benefit of avoiding toxicity associated with ReRT. Critics, however, note that the median survival of 21 months (95% CI 19-31 months) demonstrated by Saba et al. seems superior to the median survival in KEYNOTE-048.(55) Given the population differences in these trials which confer different prognoses, it is impossible to state with certainty if ReRT has value, but in aggregate, the data appears to support a valid hypothesis for further investigation.

The task force feels the need to quantify the value of full-dose, modern ReRT is should be one of the highest priority questions in the locoregionally recurrent or second-primary population. The Re-irradiation Collaborative Oncology Group (ReCOG) is an initiative studying these questions in greater depth. ReCOG has prioritized developing better technical frameworks for ReRT and challenged the community to join in advancements of data sharing, composite dosimetry, dose analyses, modeling, and radiobiology, with the goal of facilitating clinical trial development.(56)

Working with the ReCOG network will be a necessary step in the development of definitive ReRT. Standardization of nomenclature, methods of composite dose analysis, and more precise radiobiologic modeling will be necessary to maximize the benefit and improve the safety and tolerability or ReRT. For instance, adaptive replanning based on mid-course response may help reduce risks of radiation necrosis and other treatment effects. How to measure mid-treatment response, whether radiographically or by other biomarkers (such as circulating or cell-free DNA) remains an open question worthy of further study.

Alternative radiation therapies used in ReRT could include brachytherapy or particle therapy with proton or even carbon ions.(57) There is significant single-institution data using these modalities for ReRT, which were reviewed in detail in a recent systematic review and Appropriate Use Criteria document.(45) Brachytherapy, however, has declined in utilization over the years due its complexity and increasingly competitive external beam solutions. Further study of brachytherapy or carbon ion therapy by cooperative groups is unlikely to be feasible on a large scale in the near term.

However, access to proton therapy continues to expand with increasingly affordable single-room solutions and a growing number of centers nationally and internationally. NRG Oncology has allowed proton therapy on various head and neck cancer protocols with success. The working group is optimistic that future protocols investigating definitive ReRT can and should seek to develop proton therapy for this patient population. Carbon ion therapy will be commented on in more depth below in the “uncommon histology” discussion.

Uncommon Histologies

So far, we have discussed the management of squamous cell carcinomas (SCC) arising from a mucosal origin. There are a multitude of other histologies that deserve attention, such as cancers of the major salivary glands which are unique and diverse. Management of these cases in the recurrent setting has only rarely been studied. Surgical principles resemble those used for squamous cell carcinomas, and thus nearly all recurrences of salivary gland histology are recommended to undergo salvage surgery at the time of locoregional recurrence when possible. ReRT, however, is extrapolated from squamous cell carcinoma data and often treatments of various salivary gland histologies are included together in single-institution reports treated as one disease. This is generally an inaccurate grouping given the extended natural history of some salivary gland histologies. Unfortunately, these data remain currently very unclear. Systemic therapy is often offered when feasible, particularly if targeted therapy is available (such as HER2 directed therapy for salivary duct carcinoma), but again these data remain limited.

Given the rarity of these situations, the working group’s enthusiasm for such trials is guarded. There may be a role for small phase I-II protocols to demonstrate feasibility of unique approaches. Alternatively, prospective observational databases could be constructed that allow for meaningful contributions outside the scope of a cooperative group. Regardless of the study design, structured compilation of additional data would be helpful in these rare situations.

Rare and radioresistant histologies in the head and neck are a potential application for carbon ion therapy, whose increased linear energy transfer (LET) may result in increased local control rates for these difficult-to-treat tumors. Single-institution data for ReRT with carbon ions is scarce but promising.(58,59) While the number of carbon ion centers is currently very limited, the working group remains optimistic that through international partnerships there could be coordination of efforts and feasible prospective development of this promising technology.

Resource Utilization

Modern clinical trials are complex and expensive to execute. Public funding for these efforts is becoming increasingly difficult, and investigators should be thoughtful as to its utilization to maximize population benefits.

Whether salvage therapy warrants investigation at all is a question worth discussing. Some may argue that all available resources should be invested in the initial course of therapy, so as to avoid salvage therapy entirely. Members of this working group find this view to be optimistic but somewhat impractical. Recurrences (and second primaries) will always occur in subsets of patients, even though we hope these subsets will become smaller over time. When recurrences do emerge, multidisciplinary teams need robust data to guide meaningful treatment, without harming patients with overly aggressive subsequent therapy. While many of the questions raised in this summary cannot be answered through a phase III trial, even small trials could make a significant impact in influencing clinical care decisions. This highlights the role for alternative trial designs, such as basket or pragmatic trials, which may help answer nuanced questions while saving resources.

While cooperative groups are able to study novel pharmaceutical agents, industry partnership is crucial and necessary to these endeavors and not always available. However, in the locoregionally recurrent space, very novel agents may not be needed given that there are very basic questions that remain unanswered. Furthermore, there are many diseases arising from the head and neck which are exceedingly rare, for which a very wide network of enrolling, highly dedicated institutions would be needed. This could be an important role the cooperative groups could fill, and we anticipate industry partnerships could be developed for these scenarios.

Our group sought to summarize the current state of the science and outline opportunities for meaningful publicly funded investigations on the topic of locoregionally recurrent head and neck cancer. We hoped to clarify the perspectives on the various major questions and facilitate trial development in this space over the coming years, and Figure 1 lists these highest priorities identified.

Figure 1:

Summary of Suggested Priorities

Conclusion

This review articulates select key questions in the management of non-metastatic, locoregionally recurrent cancers of the head and neck. This summary will inform discussion surrounding development of clinical trial concepts within the NRG Oncology environment moving forward.

Data Sharing Statement:

There are no new data associated with this article.