Reirradiation for locally recurrent head and neck cancer

INTRODUCTION — Despite advances in the treatment of head and neck cancer, 15 to 50 percent of patients will develop recurrent disease [1-6]. Survivors also are at risk to develop second primary tumors, the incidence of which is estimated at 8 to 22 percent, with approximately one-third occurring in the head and neck. (See "Second primary malignancies in patients with head and neck cancers".)

Therapeutic options are limited for patients who present with locally recurrent disease or a second primary tumor in a previously irradiated field. Surgical salvage with a curative-intent resection is the preferred option for those with limited-volume disease [7-9]. Reirradiation is an alternative for patients who are not candidates for surgical salvage [9,10].

Reirradiation with or without the addition of chemotherapy may hold promise for long-term survival for appropriately selected patients. Indications for reirradiation may include:

●Patients who undergo surgical salvage but are found to have high-risk features.

●Patients who are medically suitable for curative-intent interventions but are not amenable to curative-intent resection.

●Patients who are not candidates for curative-intent interventions but may benefit from palliative treatment.

General principles of treatment for locally recurrent and second primary head and neck cancer are discussed separately. Recurrence of nasopharynx cancer has a markedly better prognosis than other head and neck sites, and its management is also discussed separately:

●(See "Treatment of locally recurrent squamous cell carcinoma of the head and neck".)

●(See "Second primary malignancies in patients with head and neck cancers".)

●(See "Treatment of early and locoregionally advanced nasopharyngeal carcinoma".)

GENERAL PRINCIPLES

Patient selection — Patient selection is critical since reirradiation, with or without chemotherapy, is associated with considerable acute and late toxicity (algorithm 1) [9,11,12], although more modern radiation techniques may have lowered toxicity risks compared with historic controls [13]. Patient selection is based on tumor characteristics as well as an assessment of the patients' prognosis and their ability and willingness to tolerate the anticipated toxicity of these regimens. Patients who recur within the high-dose radiation area less than six months after the first course of radiation therapy (RT) generally have radiation-resistant disease and are not usually considered candidates for a second course of RT.

Tissues of the head and neck such as skin, nerves, blood vessels, and spinal cord normally receive significant radiation doses during the initial course of RT. Reirradiation exposes these tissues to further radiation and thereby incurs a risk of severe complications such as carotid artery stenosis or rupture, osteoradionecrosis, pharyngocutaneous fistulas, nonhealing skin ulcers, and spinal cord damage. This risk is likely reduced in the contemporary era given that most initial radiation courses are now delivered with conformal intensity-modulated radiation therapy (IMRT) sparing full-dose irradiation of tissues in close proximity. (See "Management and prevention of complications during initial treatment of head and neck cancer" and "Management of late complications of head and neck cancer and its treatment".)

Radioresistance — The development of a tumor recurrence in a previously irradiated field suggests the existence of a radioresistant clone that would limit the therapeutic effect of reirradiation. The initial disease-free interval plays an important role in the decision of whether or not to reirradiate. Patients with tumor recurrence within six months after the first course of RT are generally thought to have radiation-resistant disease and therefore to be unlikely to benefit from a second course of RT.

The problem of radioresistance may be particularly relevant to patients who develop local recurrence after initial treatment with concurrent chemoradiotherapy [14]. Strategies to overcome radioresistance include the addition of radiation sensitizers, such as chemotherapy and targeted agents, and dose escalation or intensification, with techniques such as hyperfractionation, accelerated fractionation, conformal RT, and brachytherapy [15-19]. Ideally, patients should be evaluated for a systemic agent that is different from that originally used for concurrent therapy.

Systemic regimens that incorporate checkpoint inhibitor immunotherapy are an appropriate treatment option for patients with advanced or metastatic disease previously treated with chemoradiotherapy. (See "Treatment of metastatic and recurrent head and neck cancer", section on 'Previous systemic therapy for locoregional disease'.)

The addition of immunotherapy to reirradiation in the treatment of locally recurrent disease requires further investigation. (See "Methods to overcome radiation resistance in head and neck cancer", section on 'Checkpoint inhibitor immunotherapy' and 'Novel/targeted agents' below.)

Treatment volume — A thorough analysis of the initial treatment volumes and dose distributions relative to the location of the locoregional recurrence or second primary tumor should be made when reirradiation is considered. Determination of the dose received at the site of recurrence may demonstrate geographic tumor miss and is necessary for treatment planning since a geographic or marginal miss may not represent the same radioresistant disease as a true in-field recurrence, even with a disease-free interval of less than a year.

Reirradiation treatment volumes for patients treated with IMRT (typically with concurrent chemotherapy) are usually limited to a 1 to 2 cm margin around the gross tumor or the surgical site, with even smaller margins when tumors are near critical structures such as the skull base/brainstem, spinal cord, optic nerve, or optic chiasm [16]. This approach was validated in a retrospective review of 66 reirradiated patients, which demonstrated that 96 percent of locoregional recurrences following reirradiation occurred within the gross tumor volume and only 4 percent occurred in untreated regional sites [20]. The cumulative dose to the spinal cord from all courses of treatment is usually limited to 50 Gy, with the reirradiation dose to the spinal cord restricted to 10 percent of the prescribed reirradiation dose [16,21]. Treatment volumes for patients reirradiated with stereotactic body radiation therapy (SBRT) are often limited to small deposits of gross disease and a <1 cm margin for patient setup uncertainty.

Dose — When employing standard, non-SBRT dosing schemes, doses ≥50 Gy to the tumor are necessary to yield substantial response rates and delay disease recurrence [22]. Many investigators and cooperative groups use 60 Gy as a target reirradiation dose, delivered through various reirradiation protocols [15,23]. Some large analyses demonstrate that patients treated to a higher radiation dose have improved survival and tumor control outcomes [14,24]. There is no consensus regarding total dose or use of any particular dose fractionation scheme. For patients with gross disease, we typically prescribe 70 to 75 Gy, and for those treated following gross total resection, we prescribe 60 to 66 Gy.

Modality — Three-dimensional conformal planning and advanced delivery techniques such as IMRT, brachytherapy, intraoperative radiation therapy (IORT), stereotactic radiosurgery, and SBRT have been used to deliver high doses to tumors while minimizing the radiation dose to surrounding normal structures. Proton therapy or carbon ion therapy may facilitate dose escalation and/or normal tissue sparing [25], but data on its efficacy and safety for recurrent head and neck cancer are limited [25-31]. (See "General principles of radiation therapy for head and neck cancer", section on 'Charged particle radiation'.)

IMRT is the external beam radiotherapy technique usually employed for head and neck reirradiation. This is based on observational studies demonstrating encouraging two-year rates of overall survival (up to 58 percent) and locoregional control (up to 64 percent) [32-34]. In addition, SBRT (alone or in combination with chemotherapy) achieves one-year locoregional disease control in approximately one-half of retreated patients [35,36]. (See 'External beam radiation therapy' below and 'Stereotactic body radiation therapy' below.)

Investigations are ongoing to further evaluate the integration of newer radiation techniques for the use of reirradiation. (See "General principles of radiation therapy for head and neck cancer".)

Elective neck treatment — In the case of local tumor recurrence, the need for elective irradiation to the grossly uninvolved cervical lymphatics is controversial [15]. In our practice, we do not routinely treat elective nodal regions since the primary pattern of recurrence following reirradiation is in the reirradiated site [20]. Larger radiation volumes also significantly increase the risk of severe late toxicity.

Some patients may be evaluated for elective neck irradiation if those nodal areas are outside of the previous high-dose radiation field, more often in cases of limited prior radiation and second primary tumors with a historic propensity for nodal metastases. However, this must be carefully considered as neck irradiation can significantly increase the risk of side effects. In a multi-institutional retrospective study of 505 patients, elective nodal radiation was associated with greater risks of acute toxicity and did not improve overall survival [24].

POSTOPERATIVE REIRRADIATION — We suggest postoperative reirradiation for completely surgically resected patients with high-risk pathologic features [15,16,37]. We offer the same concurrent chemotherapy protocols that are used for unresectable recurrences. (See 'Reirradiation with concurrent chemotherapy' below.)

The addition of postoperative reirradiation to surgical salvage may improve upon the results of using surgery alone. Retrospective reviews and phase I/II trials established the feasibility of postoperative reirradiation and suggested promising outcomes [38,39]. The Groupe d'Etude des Tumours Tete et Cou (GETTEC) in conjunction with Groupe d'Oncologie Radiotherapie Tete et Cou (GORTEC) conducted a prospective, randomized trial to confirm these results [40]. After a macroscopic complete resection, 130 patients were randomized to reirradiation (60 Gy in 12 weeks) in combination with chemotherapy (fluorouracil plus hydroxyurea) or no postoperative therapy. Postoperative reirradiation was associated with an improvement in progression-free survival (PFS; hazard ratio [HR] 1.6, 95% CI 1.1-2.4) but not overall survival.

Subsequently, many have focused attention on those patients at highest risk of second failure after surgical salvage: those who have undergone macroscopic complete resections but were found to have high risk pathologic features, including positive margins, perineural invasion, lymphovascular invasion, or extranodal extension. As examples, two nonrandomized trials reported a three-year overall survival of 44 percent with postoperative reirradiation alone and a four-year overall survival of 43 percent with postoperative reirradiation with concurrent chemotherapy in these high-risk patients; however, these results may reflect patient selection bias towards those with a more favorable prognosis [41,42].

A systematic review of 16 studies with 919 patients, including 522 who underwent reirradiation, reported wide ranges of local control (21 to 100 percent), two-year overall survival (24 to 81 percent), severe mucositis and/or dysphagia/pharyngitis (11 to 52 percent), late fibrosis (2 to 44 percent), and pharynx dysfunction (2 to 70 percent). The authors concluded that, while interpretation of the data is limited due to the heterogeneity of the patients and treatment characteristics, postoperative reirradiation with highly conformal techniques to doses >50 Gy should be considered for selected patients with high-risk features (such as extracapsular extension or residual disease) [43].

REIRRADIATION FOR UNRESECTABLE RECURRENCES — Although retrospective reviews and phase I/II studies have demonstrated that reirradiation with or without concurrent chemotherapy is feasible and that some patients may achieve long-term survival, it is not clear that reirradiation, with its accompanying toxicity, is superior to either systemic therapy alone or best supportive care.

For appropriately selected patients, we suggest reirradiation with concurrent chemotherapy over systemic therapy or reirradiation alone. The data supporting this approach are derived from small, nonrandomized studies. Randomized trials designed to address this question were terminated because of poor accrual [10]. Differences in patient populations due to selection bias influence outcomes and toxicity rates and limit comparisons between studies. As a general principle and when feasible, patients who have undergone previous concurrent chemoradiation should receive a different systemic agent than what was administered originally, given the high likelihood of resistance to the original cytotoxic drug.

Surgical debulking prior to reirradiation — For patients with an unresectable recurrence, surgical debulking (ie, an incomplete R1 or R2 resection) may be offered to select patients to palliate tumor-related symptoms. In contrast, experts differ in their use of surgical debulking prior to reirradiation in patients who do not require palliation.

Surgical salvage is the standard of care for patients with resectable disease, but the role of debulking surgery prior to reirradiation is less clear. Surgical debulking prior to reirradiation may allow better sparing of normal tissues and better facilitate definitive doses of radiation therapy (RT). However, it is unclear whether debulking improves disease control and overall survival.

Retrospective data suggest that, even in patients with an unresectable recurrence, surgical debulking prior to reirradiation improves disease control and/or survival [22,38,44]. In contrast, a prospective, nonrandomized phase II trial failed to show a difference in survival with the addition of surgery [45]. In this study, 35 patients with recurrent or second primary head and neck cancer were treated with two cycles of chemotherapy followed by local therapy (either surgery followed by chemotherapy [pemetrexed plus gemcitabine] and reirradiation or chemotherapy and reirradiation alone). Surgical candidacy was determined prior to the initial chemotherapy. In the final analysis of this study, survival was similar between those who underwent surgery (including incomplete debulking procedures) and those who did not [45].

Debulking surgery can provide immediate palliation of symptoms. However, surgery may disfigure the patient, and it could delay or even preclude the start of planned reirradiation (with or without chemotherapy) due to surgical complications.

Debulking surgery may also delay or preclude the more appropriate use of reirradiation due to postoperative complications. Therefore, clinicians offering this approach must select patients carefully, inform them of the limited data supporting this approach, and discuss the potential toxicities of multimodality therapy.

Reirradiation alone — Depending on the population examined, reasonable survival rates have been reported with primary reirradiation without chemotherapy [46-57]. However, many of these trials were done before concomitant chemotherapy and RT became the standard treatment for advanced head and neck cancer. More commonly, however, when RT to the head and neck is employed, chemotherapy is given concurrently to overcome radioresistance and improve outcomes.

External beam radiation therapy — Intensity-modulated radiation therapy (IMRT) is the external beam radiotherapy (EBRT) technique usually employed for reirradiation of recurrent head and neck cancer, as this approach improves local control with less treatment-related toxicity [32,33,44]. With EBRT alone, three-year overall survival rates are approximately between 13 and 22 percent [47,57]. (See 'Modality' above.)

Brachytherapy — Interstitial brachytherapy alone or in combination with EBRT has yielded five-year locoregional control rates as high as 50 to 70 percent and five-year overall survival of 11 to 30 percent [51,53,54,58]. Sites amenable to this approach include the oral cavity and oropharynx.

Brachytherapy is a reasonable treatment option for patients with recurrent disease at centers with appropriate expertise. As an example, one center has used interstitial brachytherapy for recurrent head and neck cancer since the mid-1970s, initially with low-dose-rate brachytherapy and then transitioning to high-dose-rate brachytherapy [54]. In the most recent report of their experience with high-dose-rate brachytherapy delivered at doses of 18 to 48 Gy in 3 to 4 Gy, twice daily fractions, the two-year disease-specific and overall survival rates were 45 and 37 percent, respectively. (See "General principles of radiation therapy for head and neck cancer".)

Intraoperative radiation therapy — Intraoperative radiation therapy (IORT; with electrons or high-dose-rate brachytherapy) combined with debulking surgery allows visualization of the target and avoidance of uninvolved structures. Two-year local control rates and overall survival of 60 and 55 percent have been reported [55,59]. While local control rates with IORT are acceptable with completely resected disease, high local failure rates may preclude its use for gross residual disease.

Stereotactic body radiation therapy — Stereotactic body radiation therapy (SBRT), which uses anatomic references (ie, skull or spine), surgically placed fiducials, external markers, and/or a rigid stereotactic frame to more accurately target radiation treatment, is emerging as an option for reirradiation of head and neck cancer [60]. Case series have reported good local control and toxicity comparable to conformal techniques.

SBRT allows for the delivery of "ablative doses" of hypofractioned radiation (ie, high dose per fraction) generally in one to five treatments by utilizing technology to optimize targeting accuracy, which facilitates dose escalation and normal tissue sparing [61]. (See "Radiation therapy techniques in cancer treatment", section on 'Stereotactic radiation therapy techniques'.)

SBRT should only be offered to those patients with relatively low-volume recurrent disease confined to a discrete focus or foci where SBRT can be administered in a relatively safe manner. When SBRT is used to treat head and neck cancer recurrences, the recurrence must be well visualized on imaging to ensure accurate delivery of ablative doses of radiation. In a retrospective study of 85 patients treated with head and neck SBRT reirradiation (70 percent of whom also received induction, concurrent, and/or adjuvant systemic therapy), the two-year overall survival and local control rates were 24 and 28 percent, respectively [62].

A large, single-institution report found that patients with an isolated neck recurrence had the lowest risk of late toxicity, while those with laryngeal and hypopharyngeal cancer had a 50 percent risk of grade 3 or greater toxicity following SBRT, significantly higher than at other sites (6 to 20 percent). Interestingly, SBRT dose >44 Gy was independently associated with both acute and late toxicity, while the RT treatment platform was not [63]. In another study, concurrent systemic therapy was associated with greater regional and distant control, and higher grade ≥3 toxicity [64]. While phase I data exist, the optimal dose and fractionation schedule of SBRT for head and neck reirradiation to maximize efficacy and minimize toxicity risk is unknown [65].

A pooled analysis conducted by the American Association of Physicists in Medicine Hypofractionated Treatment Effects in the Clinic (HyTEC) working group analyzed 300 cases in eight publications; from their data modeling, 35 to 45 Gy (in five fractions) was associated with greater tumor control probability and overall survival compared with doses <30 Gy [66].

There do not appear to be survival differences between head and neck reirradiation with SBRT and conventionally fractionated IMRT, albeit based upon retrospective data [13].

Reirradiation with concurrent chemotherapy — Many institutions have reported relatively promising results from single-institution phase I/II trials of reirradiation with concurrent chemotherapy [34]. These results have been repeated in the multi-institutional cooperative group setting, and in one study, the benefit of reirradiation with concurrent chemotherapy over observation was demonstrated [40,67,68].

Until additional data are published, we favor reirradiation with concurrent chemotherapy over either chemotherapy alone or RT alone. Although there is no standard concurrent chemotherapy regimen for use with reirradiation, patients should be treated with established regimens, which are discussed below, or encouraged to participate in clinical trials. The chemotherapy regimen should avoid drugs to which the patient was previously exposed.

Patient selection is critical as these regimens are associated with considerable toxicity. Patient selection is based upon tumor characteristics as well as an assessment of the patients' prognosis and their ability and willingness to tolerate the anticipated toxicity of these regimens. (See 'Toxicity' below and 'Prognostic factors' below.)

FHX/FHX2 regimens — One of the best studied combinations is concomitant fluorouracil (FU) and hydroxyurea with reirradiation (FHX) [22,23,40,67]. Investigators at the University of Chicago have reported on 45 patients enrolled in four prospective phase I-II trials [69]. Cisplatin was added to the chemotherapy regimen in three of the four trials. Five-year locoregional control, PFS, and overall survival were 14, 15, and 20 percent, respectively.

For the original FHX protocol, RT was delivered on alternating weeks (five-day treatment cycle followed by a nine-day break). The regimen was modified in subsequent studies to twice-daily, week-on/week-off RT (FHX2).

Radiation Therapy Oncology Group (RTOG) 9610, which included 81 patients, tested FHX2, but with a lower FU dose, in the cooperative group setting. The two-year and five-year survival rates were 15 and 4 percent, respectively [67,70].

In a pooled analysis of 115 patients treated with FHX/FHX2, FHX plus cisplatin, TFHX2 (including paclitaxel), and TFGX2 (including paclitaxel and gemcitabine) regimens at the University of Chicago, the three-year locoregional control rates, PFS, and overall survival were 51, 33, and 22 percent, respectively [22].

Platinum combinations — Promising single institution results with cisplatin and paclitaxel concurrent with hyperfractionated RT have been replicated in a cooperative group phase II trial, RTOG 9911 [68,71]. Hematopoietic growth factor, G-CSF, was added to the protocol to reduce the acute toxicity and treatment delays experienced during the initial studies. With 99 patients treated and 74 percent completing therapy, two-year PFS and overall survival rates were 16 and 26 percent, respectively. These survival rates are superior to the results from RTOG 9610 and exceed those generally seen with chemotherapy alone. However, they do come at the cost of apparent increased toxicity, and caution is warranted in comparing nonrandomized trials.

This regimen of cisplatin and paclitaxel was selected as the concurrent chemotherapy for the reirradiation arm in RTOG 0421; however, this trial has closed due to poor accrual [72].

Other promising platinum regimens include carboplatin/paclitaxel [73], docetaxel/cisplatin [74], cisplatin/FU [75], and carboplatin/pemetrexed, which have been used in combination with induction chemotherapy [76]. Cisplatin alone concurrent with reirradiation is also used [77].

Induction chemotherapy — Induction chemotherapy followed by surgery and reirradiation has been investigated based upon results in the first-line therapy of advanced head and neck cancer. (See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy", section on 'Induction chemotherapy'.)

Induction with pemetrexed and gemcitabine, followed by resection in patients amenable to surgery, and then reirradiation with concurrent carboplatin and pemetrexed yielded a one-year survival rate of 42 percent [76]. Poor response to induction chemotherapy predicted worse survival, a finding that may be used in subsequent studies to select patients for reirradiation.

Novel/targeted agents — New agents, alone or in combination with cytotoxic chemotherapy, are being added to reirradiation as a means of overcoming radioresistance.

Treatments that have been investigated include checkpoint inhibitor immunotherapy (eg, nivolumab [78], pembrolizumab plus SBRT [79]); single-agent cetuximab with either SBRT [36,80] or conventionally fractionated reirradiation [81-83]; cetuximab plus cisplatin [84]; and intratumoral injection of TNFerade (a replication defective adenoviral vector that carries a human tumor necrosis factor [TNF] gene linked to the radiation inducible promoter, early growth response protein [EGR-1]) in conjunction with FHX2 [85].

Other targeted agents have either not demonstrated efficacy, such as antiangiogenesis agents (ie, bevacizumab) [86,87] or are not feasible due to toxicity, such as bortezomib [88].

Isolated neck recurrence — Surgical salvage of an isolated neck recurrence is a standard curative-intent approach, and specific surgical issues that pertain to the irradiated neck are discussed elsewhere. (See "Treatment of locally recurrent squamous cell carcinoma of the head and neck".)

Interstitial brachytherapy and IORT are ideal for treatment of an isolated neck recurrence since the radiation dose can be selectively delivered to the neck. However, scant data exist. In one series of 22 patients treated with salvage neck dissection and low dose rate brachytherapy, the two-year regional control and overall survival rates were 67 and 57 percent [89]. Others, after appreciating that reirradiation of the neck with brachytherapy resulted in necrosis and fibrosis in 42 percent of patients, modified the technique to include skin resurfacing with a myocutaneous graft. Subsequently, this complication rate was lowered to 11 percent [90].

PALLIATIVE REIRRADIATION — Although data are limited, palliative irradiation is an option for patients whose tumor is significantly impacting quality of life and who are not candidates for an aggressive course of reirradiation. In general, tumor shrinkage with palliative reirradiation is expected to be greater than the response to chemotherapy. Lower doses and larger fraction sizes are delivered to patients receiving palliative reirradiation. The lower dose should result in a lower risk of acute toxicity, and the increased risk of late complications with larger fraction sizes is generally not relevant.

The "Quad Shot" approach to palliative radiation therapy (RT) delivers short, cyclical courses to maximize clinical response and minimize toxicity [91]. Each cycle consists of twice-daily hypofractionated RT administered over two days in four-week intervals for a total of two or more cycles depending on treatment response. This approach allows mucosal stem cells to repopulate before the next cycle [91].

While the best data regarding the Quad Shot approach come from an uncontrolled prospective trial in patients without prior RT exposure [91], we offer this approach to patients with prior RT exposure who are not candidates for more aggressive treatment, as the limited observational data in the setting of reirradiation are promising [92,93].

In two retrospective studies including a total of 101 patients with incurable recurrent or metastatic head and neck cancer treated with the Quad Shot approach (53 patients with prior RT), responses were seen in 65 to 73 percent, and there was a low rate of grade ≥3 toxicities [92-94]. Of note, previous RT exposure did not appear to be negatively predictive of palliative response, as responses were seen just as frequently in those with prior RT exposure [93].

Patients with multifocal locoregional recurrences, metastatic disease, and short interval between initial treatment and recurrence have such a limited survival that the likelihood of severe treatment-related toxicity usually outweighs any palliative benefit from reirradiation [47,94,95]. Systemic therapy for metastatic and recurrent head and neck cancer is discussed separately. (See "Treatment of metastatic and recurrent head and neck cancer".)

TOXICITY — Reirradiation is a treatment option for some patients, but toxicity is a very important consideration [96]. Acute and long-term complications, including treatment-related fatalities from bleeding, tissue necrosis, and infection, are substantial.

Treatment-related fatalities with irradiation approach 10 percent, and the incidence of acute and late grade ≥3 toxicity is up to 50 percent [67,68,97], although the true risk of severe toxicities is likely higher. As such, it is critical to select patients judiciously and provide robust supportive care during and after therapy. Nomograms are available that can predict severe late toxicity after reirradiation, using patient, disease, and treatment-related factors [97]. (See "Management and prevention of complications during initial treatment of head and neck cancer" and "Management of late complications of head and neck cancer and its treatment".)

Reirradiation alone — Even with reirradiation alone (ie, without concurrent chemotherapy), toxicity is high. Acute complications in this setting are mainly mucositis and pain. Reirradiation may also result in multiple severe toxicities within the same patient. A series of 41 patients, 78 percent treated with intensity-modulated radiation therapy (IMRT), reported radiation-related toxicity and disease-related problems occurring in 81 percent of patients with grade 3 or 4 toxicity in 68 percent [47].

The risk of long-term toxicity from reirradiation is high but can be limited with reductions in the size of the target volume with conformal delivery techniques and by adopting split-course schedules with smaller fraction sizes. With these modifications, the incidence of radiation necrosis is generally less than 25 percent [37]. Severe late complications, each with an incidence as high as 5 percent, include mucosal ulceration/necrosis, neck fibrosis, trismus, orocutaneous fistula, carotid blow out, hemorrhage, chondronecrosis/osteoradionecrosis, feeding tube dependency, and death [37].

In a pooled analysis of 39 studies with approximately 3800 patients undergoing head and neck reirradiation, the rate of any grade ≥3 late toxicity was 29 percent; the most common toxicities included radionecrosis, dysphagia requiring feeding tube placement, and trismus. Additionally, the estimated rate of fatal carotid blowout was 4 percent [96]. (See 'Carotid artery injury' below.)

When stereotactic body radiation therapy (SBRT) is used in the reirradiation setting, grade 3 or greater late toxicity is more common when the larynx/hypopharynx sites are targeted. Additionally, these later effects are more common after doses >44 Gy in five fractions [63].

Reirradiation with chemotherapy — The addition of chemotherapy to radiation therapy has added to the toxicity in most reports. Evidence suggests that acute toxicity with reirradiation and concurrent chemotherapy is similar to the toxicity associated with an initial course of concurrent chemoradiation [40,67,68,70].

The following two trials illustrate the scope and high rates of toxicity with reirradiation and concurrent chemotherapy:

●Radiation Therapy Oncology Group (RTOG) 9610 – Acute toxicity, occurring within 90 days of the start of treatment, resulted in death in 6 of 79 treated patients (8 percent), two due to hemorrhage and four due to neutropenia [67]. Grade 3 to 5 overall, nonhematologic, and hematologic acute toxicity occurred in 63, 49, and 37 percent of patients, respectively.

●RTOG 9911 – The total treatment-related fatality rate was 11 percent, with 5 percent occurring in the acute period [68]. Cause of death included neutropenic sepsis, dehydration and shock, pneumonitis, and stroke. Grade ≥3 overall, nonhematologic, and hematologic acute toxicity occurred in 78, 77, and 45 percent of patients, respectively. Of these, grade ≥3 neutropenia occurred in 24 percent, infection/neutropenic fever in 15 percent, mucositis in 14 percent, and GI toxicity in 48 percent.

Differentiating the toxicity of reirradiation from that of previous treatment can be difficult. In RTOG 9610, the cumulative incidence of grade 3 or greater toxicity in patients surviving more than one year was 12 percent [67,70]. In RTOG 9911, an estimated 85 and 32 percent experienced overall and nonhematologic grade 3 to 5 toxicity over the first two years of follow-up. There were three late fatalities in RTOG 9911, two due to carotid hemorrhages and one from oral-cutaneous fistula and soft tissue necrosis. Late toxicities reported in these trials and others include cervical fibrosis (31 to 48 percent), osteoradionecrosis (5 to 16 percent), trismus (9 to 24 percent), and carotid hemorrhage (2 to 5 percent) [16]. Although long-term spinal cord and brainstem toxicity has been rare, this may be due to selection of patients with tumors not in close proximity to the spinal cord [22,67,68,70].

Carotid artery injury — Carotid artery hemorrhage and rupture are particularly dreaded complications of reirradiation [37,98]. Patients at particular risk for this complication include those with long-term tracheostomies, active infections, pharyngocutaneous fistulas, poor wound healing, prior neck dissection, and tumor adherent to the carotid fascia [99].

In a literature review of 27 series that included 1554 patients managed with reirradiation, the rate of carotid blowout was 2.6 percent and appeared to be higher in patients treated with accelerated fractionation regimens [100]. In another study of 412 patients, the rate of carotid bleeding was 1.2 percent with a median time between courses of 2.4 years (range 0.3 to 11.7 years) among these patients [13,101].

One report (Hypofractionated Treatment Effects in the Clinic [HyTEC]) modeled carotid events after reirradiation with SBRT. In order to minimize risks of carotid or other bleeding events, this report recommended delivery of treatment on nonconsecutive days, restricting the reirradiated major vessel volume exceeding 20 to 30 Gy to be as small as possible, and considering alternative (ie, non-SBRT) RT approaches when circumferential carotid radiation is unavoidable [102]. (See "Management of late complications of head and neck cancer and its treatment", section on 'Carotid artery injury'.)

Stabilization of the carotid, by means of a surgical flap reconstruction or an endovascular stent, prior to reirradiation in patients at risk has been proposed but is untested [8,103].

PROGNOSTIC FACTORS — Several factors have been identified as prognostic of both locoregional control and survival for patients undergoing reirradiation for recurrent and/or second primary tumors:

●Second primary tumors fare better than recurrent tumors [46,70].

●Surgical debulking of the recurrent tumor is associated with improved outcomes [14,22,33,38,40].

●Extensive recurrent disease is associated with worse outcomes [11,47,95].

●A longer time interval to recurrence affords a better prognosis [11,14,16,67,104].

●Laryngeal and nasopharyngeal tumors have a more favorable prognosis as compared with other sites [33,56,105-107].

●Recurrences in the neck fare poorly as compared with primary site recurrences [77,101,108].

●Higher doses of radiation therapy yield better outcomes [11,13,14,22,33,38,75].

●Multi-agent chemotherapy has been associated not only with improved survival but also decreased distant metastases [22].

●Medical comorbidities and laryngeal/pharyngeal organ dysfunction have a markedly negative impact on survival [101,109].

●Prior treatment with concomitant chemotherapy and radiation has a worse survival [14].

In a recursive partitioning analysis of 412 patients, the interval between courses (>2 versus ≤2 years), resection of recurrence (among those with >2 years between courses), and organ dysfunction (among those with ≤2 years between courses) allowed grouping of patients into three groups with two-year overall survival rates of 62, 40, and 17 percent [101].

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".)

SUMMARY AND RECOMMENDATIONS

●Treatment approaches – Patients presenting with recurrent head and neck cancer in a previously irradiated field present a therapeutic challenge. (See 'Introduction' above.)

•Salvage surgery, if feasible, represents the preferred treatment option.

•Reirradiation, with or without chemotherapy, has become an accepted alternative to chemotherapy or supportive care and may offer long-term survival for selected patients.

●Patient selection – Because of the toxicity associated with reirradiation for recurrent head and neck cancer or a second primary tumor, patient selection is critical. Factors to be considered include the dose of prior radiation therapy (RT) and the radiation fields, the time since prior treatment, the extent and location of recurrent tumor, and the overall patient condition and comorbidities. (See 'Patient selection' above.)

●Resected disease with high-risk pathologic features – For patients with high-risk pathologic features after complete surgical resection of a recurrence or second primary, we suggest postoperative reirradiation (Grade 2C). (See 'Postoperative Reirradiation' above.)

●Resected disease without high-risk pathologic features – For patients without high-risk features after complete surgical resection of a recurrence or second primary, we do not suggest postoperative reirradiation (Grade 2B). (See 'Postoperative Reirradiation' above.)

●Unresectable, locally recurrent disease – For selected patients with unresectable, locally recurrent head and neck cancer without distant metastases, we suggest reirradiation with concurrent chemotherapy (Grade 2C). Stereotactic body radiation therapy (SBRT) is an emerging treatment option. (See 'Reirradiation with concurrent chemotherapy' above.)

•Surgical debulking (ie, an incomplete R1 or R2 resection) may be offered to select patients to palliate tumor-related symptoms. Experts differ in their use of surgical debulking prior to reirradiation in patients who do not require palliation. (See 'Surgical debulking prior to reirradiation' above.)

•There is no standard concurrent chemotherapy regimen. Patients should be treated with established regimens (eg, the FHX regimen [fluorouracil and hydroxyurea with reirradiation] or a paclitaxel-platinum combination) or encouraged to participate in clinical trials. (See 'Reirradiation with concurrent chemotherapy' above.)

●Palliative reirradiation – For some patients who are not candidates for curative-intent treatment but who have symptomatic local disease or are likely to experience symptoms from local progression, palliative reirradiation may provide symptom relief. (See 'Palliative reirradiation' above.)

●Toxicity – The toxicity from reirradiation is considerable. Treatment-related fatalities approach 10 percent, and the incidence of acute and late grade ≥3 toxicity is up to 50 percent. As such, it is critical to select patients judiciously and provide robust supportive care during and after therapy. (See 'Toxicity' above and "Management and prevention of complications during initial treatment of head and neck cancer" and "Management of late complications of head and neck cancer and its treatment".)