Current radiotherapy for recurrent head and neck cancer in the modern era: a state-of-the-art review

Head and neck cancer (HNC) is a broad term, including epithelial malignancies that occur in the paranasal sinuses, nasal cavity, oral cavity, pharynx, and larynx. Almost all these malignancies are head and neck squamous cell carcinoma (HNSCC). Approximately two-thirds of HNSCC patients have the advanced-stage disease with regional lymph nodes. The initial appearance of distant metastasis is uncommon only affecting approximately 10% of the patients [1]. HNC is the seventh most common cancer in the world, which is typically diagnosed in elderly patients associated with large amounts of tobacco and alcohol use. Besides, cases of human papillomavirus (HPV)-associated oropharyngeal cancer, which is mainly caused by HPV-16, are increasing in recent years, and this type of HNC is associated with a better prognosis than HPV-negative oropharyngeal cancer [2].

Despite the advancement of modern HNC treatment modalities, cancer recurrence is still a major problem, with a locoregional recurrence rate of 15–50% [3]. Of the HNSCC asymptomatic recurrences after definitive radiotherapy and chemotherapy, 93% are local or regional, and they mainly occur within the first 2 years after the initial treatment [4, 5]. Salvage surgery may be a curative option for patients with resectable locoregional recurrence [1]. For recurrent HNC (rHNC), only 15–30% of patients are indicative for surgery and the 5-year survival rate is 16–36% [6]. In the majority of cases, salvage surgery is not feasible, or it is only possible with severe complications and limited success rates [7]. When surgical treatment is not possible, the prognosis of patients with recurrent HNSCC (rHNSCC) is unsatisfactory. The median survival of untreated patients, receiving palliative chemotherapy, and conventional radiotherapy is 3–5, 6–9, and 9–14 months, respectively [8].

In the past, palliative chemotherapy was the main choice for patients with inoperable rHNC who had received high-dose radiotherapy. With the emergence of new technologies, the re-irradiation of recurrent tumors has markedly attracted clinicians’ attention [9]. Considering the high incidence of re-irradiation toxicities, for rHNC, especially in cases that have previously undergone radiotherapy, it is highly necessary to adopt a radiotherapy program with a high conformability and an accurate dose distribution, to reduce adverse reactions to normal tissues.

In recent years, several studies have shown that current radiotherapy technologies can effectively treat rHNC and result in a better prognosis for rHNC patients. These studies assessed the survival and local tumor control of patients, and adopt Common Terminology Criteria for Adverse Events (CTCAE), Radiation Therapy Oncology Group (RTOG) morbidity scoring criteria, or European Organization for Research and Treatment of Cancer (EORTC) Late Radiation Morbidity Score to evaluate adverse effects. In the present review, the efficacy and safety of intensity-modulated radiation therapy (IMRT), stereotactic body radiation therapy (SBRT), high-dose-rate brachytherapy (HDR-BRT), and low-dose-rate brachytherapy (LDR-BRT) in the treatment of rHNC patients were investigated.

Although the role of re-irradiation is controversial due to concerns about the high incidence of severe chronic toxicity, it is still a potentially curative treatment option for patients with locoregionally recurrent tumors. The clinical application of IMRT can provide effective biological doses for more conformal areas to improve tumor control while minimizing treatment-related toxicities [10‐12].

Lee et al. reviewed the efficacy of re-irradiation using IMRT for recurrent or second primary HNC (RSPHNC). From 2007 to 2018, a total of 17 studies were included in this review, involving 1635 patients. Except for a study with a median dose of 49 Gy, the re-irradiation dose ranged from 59.4 to 70 Gy, which did not significantly vary among different studies. The 2-year local control (LC) and overall survival (OS) rates were 52% (95% confidence interval [CI], 46%-57%) and 46% (95% CI, 41%-50%) with re-irradiation using IMRT, respectively. The pooled rates of late grade ≥ 3 and grade 5 toxicities were 26% (95% CI, 20%-32%) and 3.1% (95% CI, 2%-5%), respectively. In the subgroup analysis, the salvage surgery rate (< 42% vs. ≥ 42%) affected the 2-year LC rate (45.9% vs. 58.5%, P = 0.011) [13]. Another retrospective cohort study conducted in 2020 showed primary subsite (non-larynx/hypopharynx/oral cavity), recurrent tumor size (< 3 cm), the interval between radiotherapy (RT) courses (≥ 24 months), and salvage surgery were found to be associated with a longer OS, while the interval between RT courses (≥ 24 months) and salvage surgery were noted as positive prognostic factors of LC. The authors suggested that a longer interval from the previous RT course was associated with better LC and less aggressive recurrent disease. However, this study did not determine the prognostic impact of HPV on patients, as HPV status was only available in few patients with oropharyngeal cancer [14]. Caudell et al. conducted a multi-institution retrospective cohort study to investigate the effects of selective treatment volume, dose, and fraction on outcomes and toxicity. Their results indicated that dose ≥ 66 Gy may be correlated with the improved prognosis of patients undergoing definitive re-IMRT. Postoperatively, after treating gross disease, the dose of 50–66 Gy was found sufficient. Although the 2-year OS was 60% for patients with HPV+ recurrence or second primary oropharyngeal cancer compared with 39.5% for patients with HPV-negative cancer, this result did not reach statistical significance. Moreover, hyperfractionation and elective neck irradiation had no significant benefits, and they may increase toxicity [15].

In conclusion, IMRT is advantageous for the treatment of rHNC patients, especially because of the high conformity of the target volume and the optimized sparing of previously irradiated organs at risk (OARs) [16].

SBRT is a form of modern conformal high-precision external beam radiotherapy (EBRT) that can provide various radiobiological benefits in diverse clinical conditions. SBRT can deliver dose in multiple small beams (> 150 per treatment) with less skin toxicity and a smaller irradiated volume, and it may allow the implementation of a hypofractionated scheme [17]. The SBRT aims to provide highly precise and ablative doses of radiation for the radiotherapy of the target area. For the therapy of recurrent HNSCC, SBRT can increase the dose to 50 Gy in 5 divided doses [3]. Therefore, SBRT possesses the advantages of shorter treatment duration and avoids interruption of systemic treatment. SBRT can generate a steep dose gradient between target tissues and surrounding healthy tissues, thereby reducing the overall radiation dose to critical organs at proximity to the target volume [18, 19]. The results of recent studies on the treatment of rHNC with SBRT are summarized in Table 1 [3, 8, 9, 17, 20‐23].

In recent years, numerous studies have evaluated the safety and outcomes of SBRT in patients with HNC who had previously received radiotherapy. Rwigema et al. enrolled 85 rHNSCC patients (23 patients with distant metastasis at the time of treatment) who received SBRT at a dose of 15–44 (mean, 35) Gy, and 33% of patients received cetuximab concurrently. The 1- and 2-year LC and OS rates were 51.2% and 30.7%, and 48.5% and 16.1%, respectively. The median survival was 11.5 (range, 3–51) months. For patients without distant metastasis, the 1- and 2-year OS rates were 61.9% and 23.4%, respectively. The treatment was well-tolerated, and there were no grade 4 or 5 treatment-related toxicities. Moreover, within 6 months of the median follow-up, those patients who received SBRT < 35 Gy had significantly lower LC than those with ≥ 35 Gy, with a similar incidence rate of toxicity [8].

Rwigema et al. evaluated 96 patients with unresectable, previously irradiated rHNSCC, who received SBRT at the relapse sites, and 39 patients (40.6%) who received cetuximab concurrently in a retrospective cohort study. The 1-, 2-, and 3-year locoregional control (LRC) rates for a dose of 40–50 Gy were 69.4%, 57.8%, and 41.1%, respectively, and those of 15–36 Gy were 51.9%, 31.7%, and 15.9%, respectively. The incidence rates of grade 1, 2, and 3 acute toxicities were 37.5%, 17.7%, and 5.2%, respectively, while those of grade 1, 2, and 3 long-term complications were 16.7%, 9.3%, and 3.1%, respectively. This study demonstrated that it is feasible to increase the SBRT dose to 50 Gy in 5 fractions. Moreover, a higher SBRT dose was associated with a significantly higher LRC rate. Compared with a small gross tumor volume (GTV ≤ 25 cm³), a large tumor volume (GTV > 25 cm³) requires a higher SBRT dose to achieve the best response rate [3].

Cengiz et al. enrolled 46 rHNC patients, and 30 patients of them were histopathologically diagnosed with squamous cell carcinoma. The median dose of SBRT was 30 Gy (range, 18–35 Gy) in 5 fractions. Of the 37 patients who were evaluated for the treatment response, 10 (27%), 11 (29.8%), and 10 (27%) patients achieved complete response (CR), partial response (PR), and stable disease (SD), respectively. Besides, 31 (83.8%) patients achieved local disease control. The median OS was 11.93 months, and the median progression-free survival (PFS) was 10.5 months. The 1-year PFS and OS rates were 41% and 46%, respectively. Long-term complications of grade 2 or greater were observed in 6 (13.3%) patients. During the follow-up, 8 (17.3%) patients developed carotid blow-out syndrome, and 7 (15.2%) patients died of carotid artery hemorrhage. This fatal syndrome only occurs in patients who had tumors around the carotid artery and received all the prescribed doses [9].

Comet et al. prospectively enrolled 40 patients who had inoperable recurrent, or new primary HNC in previously irradiated areas. All patients received SBRT at a dose of 36 Gy in 6 fractions, 15 patients received concomitant cetuximab, and 1 patient received concomitant cisplatin. The median follow-up was 25.6 months, of whom 34 patients could be assessed for tumor response. The median OS was 13.6 months, and the response rate was 79.4% (15 and 12 patients achieved CR and PR, respectively). In addition, grade 3 toxicity occurred in 4 patients [17].

Yamazaki et al. assessed the prognosis of 107 patients with rHNC after re-irradiation with SBRT using a CyberKnife. The 2-year OS rate was 35%. Important prognostic factors for a longer OS were the primary site (nasopharynx), absence of ulceration, and the planning target volume (PTV) ≤ 40 cm³. This study recorded 22 serious toxicities, including 11 patients with carotid blow-out syndrome (CBOS), 9 of whom died. Since CBOS was the only fatal toxicity found after re-irradiation, ulceration affected OS through CBOS. In addition, unlike the alternate-day treatment schedules of other studies, this study used daily treatment, which may have increased toxicity [21].

Stanisce et al. evaluated the relevant outcomes of stereotactic body radiotherapy treatment for recurrent, previously irradiated HNC. This study enrolled 25 patients who were treated with CyberKnife. In total, 11 patients (44%) received concurrent cetuximab chemotherapy during re-irradiation. The median survival of all patients was 7.5 (range, 1.5–47.0) months, and the median survival of 20 (80%) patients who received curative treatment was 8.3 months. Besides, 1-year survival of the entire population was 32%. The 1- and 2-year survival rates of the curative sub-cohort were 40% and 20%, respectively. There were 8 (32%) and 7 (28%) patients with local and locoregional failure, respectively. The incidence rate of grade 3 acute toxicity was 4%, while that of grade 3 late toxicity was 6% [23].

To compare the efficacy of SBRT and IMRT, Vargo et al. enrolled 414 unresectable recurrent or second primary squamous cell carcinoma of the head and neck patients, of whom 217 received IMRT and 197 received SBRT. Of the IMRT patients, 84% received systemic therapy. The unadjusted 2-year OS rate for IMRT and SBRT groups was 35.4% and 16.3%, respectively (P < 0.01). For recursive partitioning analysis (RPA) class III patients, the OS in IMRT and SBRT groups was similar. In all class II patients, IMRT was associated with an improved OS (P < 0.001). Further subgroup analysis showed that when small tumors received SBRT with a dose of ≥ 35 Gy, the OS was comparable. For large tumors, treatment with IMRT was associated with an improved OS (rT0-rT2 tumors were characterized as “small” tumors, and rT3-rT4 tumors were characterized as “large” tumors). The incidence rate of acute grade ≥ 4 toxicity in the IMRT group was greater than that in the SBRT group, and there was no significant difference in the incidence rate of late toxicity. In addition to the RPA class, younger age, second primary tumor, and small tumor volume were independently associated with the improvement of OS in the SBRT group [22].

In general, SBRT can be used to treat patients with rHNC, especially those who have previously received radiotherapy, and is considered inappropriate for re-irradiation by conventional methods. The 2-year LRC rate was 31.7–64%, the 1- and 2-year OS rate were 32–58.9% and 16–35%, and the median OS was 7.5–14.4 months for rHNC patients who were treated with SBRT. Moreover, higher SBRT dose was associated with better LRC, and younger age, the primary site (nasopharynx), absence of ulceration, second primary tumor, and small tumor volume were prognostic factors for OS. In the majority of patients, SBRT effectively alleviated the disease with less toxicity.

In HNC, the conventional median survival of the patients receiving platinum (Pt)-based chemotherapy is about 6 months, and the recurrent rate after radical treatment can be as high as 30–50% [24]. In addition, recurrence mainly (80%) occurs in volumes that were previously exposed to high doses [25]. Therefore, due to the risks of toxicity, extreme morbidity, and mortality, re-irradiation with external beam is not possible in many cases. In addition, after previous treatment, less-defined anatomical location can hinder radiation [7]. Hence, the therapeutic advantages of brachytherapy for rHNC should be highlighted. Compared with EBRT, brachytherapy can deliver a high total dose directly to the tumor, and the rapid dose fall-off above PTV can protect surrounding normal tissues [6].

Brachytherapy is a type of radiotherapy, in which radionuclide sources are used to deliver radiation doses at a distance of up to a few centimeters by surface, intracavitary, intraluminal, or interstitial application. Brachytherapy, alone or in combination with EBRT, plays an important role in the treatment of diverse types of cancer. HDR-BRT uses radionuclides, such as Iridium-192, to irradiate a designated target point or volume at dose rates of 20 cGy per min (12 Gy per hour) or more. HDR-BRT is appropriate for the treatment of malignant or benign tumors, where the treatment volume or target point is defined and accessible [26]. HDR-BRT uses a remote after-loading source (most commonly Iridium-192) to deliver the dose through previously placed catheters or applicators. Decades of experience in HDR-BRT confirm its efficacy and safety [27]. Recent studies on the treatment of rHNC with HDR-BRT are summarized in Table 2 [6, 7, 24, 28‐31].

Brachytherapy refers to the use of radionuclides to treat malignant tumors or benign diseases through radioactive sources placed close to or into tumor or treatment site [33]. LDR-BRT is accomplished through permanent implants, in which the radioactive sources are permanently placed into cancerous tissues. At the designated point, LDR-BRT is delivered at a dose rate of 4–200 cGy per hour [33].

In the treatment of HNC, it is important to maximize LC and minimize morbidity [34]. Permanent interstitial ¹²⁵I seed implantation is one of the most promising brachytherapy techniques [35]. Permanent implantation of ¹²⁵I seeds into tumors aims to provide high doses of radiation to the tumor, and the radiation outside the implanted volume falls very sharply, which can minimize the damage to the peripheral neurovascular structures and overlying skin [36]. Figure 1 showed the procedure of CT-guided ¹²⁵I seed implantation and dose-volume histograms of gross tumor volume.

This technique possesses several advantages: (1) it is minimally invasive, and requires a short overall treatment duration, (2) dose distribution can be accurately predicted, (3) continuous radiation increases the possibility of destroying malignant cells during the cell cycle, (4) continuous LDR radiation from a low-energy source is effective against hypoxic components, which could be found in rapidly dividing cell populations, and (5) low incidence of acute adverse effects [37‐39].

Additionally, LDR-BRT takes the advantage of the radiobiological properties of tumor cells (mainly redistributed in the cell cycle) and healthy tissues (DNA damage repair)[27]. Numerous studies have confirmed the efficacy and safety of LDR-BRT (Table 3) [34‐41].

In 2010, Jiang et al. assessed the feasibility, efficacy, and morbidity of CT/ultrasonography-guided permanent percutaneous ¹²⁵I seed implantation in the treatment of rHNSCC. They enrolled 25 patients who received CT/ultrasonography-guided permanent percutaneous ¹²⁵I seed implantation, and the median actuarial D90 (the dose to 90% of the target volume) of the implanted ¹²⁵I seeds was 130 Gy. The median follow-up was 8 months. The median local DFS was 12 months, and the 1- and 2-year LC rates were 48.7% and 39.9%, respectively. The median LC for nodal recurrence and primary recurrence was 12 and 16 months, respectively, with no significant difference. The 1- and 2-year survival rates were 42.5% and 28.3%, respectively. No patient had grade 4 or 5 toxicity [37].

A number of scholars studied 14 patients with rHNC who received CT-guided ¹²⁵I seed implantation, and the post-plan showed that the median actuarial D90 of ¹²⁵I seeds was 157.5 Gy. The median local control was 18 months, and the median survival was 20 months. Regarding complications, grade 1 skin reaction was found in 1 patient, 1 patient experienced grade 1 mucosal reaction, and 1 patient developed ulceration with tumor progression after 11 months. Among all patients, 28.6% (4/14), 7.1% (1/14), and 7.1% (1/14) of patients died of local recurrence, metastasis, and liver cirrhosis, respectively [38].

In 2011, Jiang et al. recruited 29 patients with rHNC who received ultrasonography-guided permanent percutaneous ¹²⁵I seed implantation. The median actuarial D90 of ¹²⁵I seeds was 130 Gy. The median local control was 16 months, while the median survival was 13 months. Among 25 patients, 5 and 7 patients died of local recurrence and metastasis, respectively. Besides, 2 patients had recurrence at 3 and 8 months after implantation, and subsequently died of pneumonia; 1 patient died of heart disease, and 1 patient developed ulceration as cancer progressed [39].

Additionally, a study evaluated 81 lesions of 64 patients, which were permanently implanted with ¹²⁵I seeds under ultrasound guidance. According to the results, 27% and 53% of patients achieved CR and PR, respectively. The 1-, 3-, and 5-year tumor control rates were 75.2%, 73.0%, and 69.1%, respectively, while the median survival was 20 months. Severe complication was grade 4 skin ulceration in two patients with cervical lymph node recurrence who had previously received radiation therapy and had recurrent lesions invading the subcutaneous tissues prior to seeds implantation. Moreover, the 5-year LC rate of cervical lymph node recurrence was higher than that of the recurrence or residual lesions of primary HNC. D90 ≥ 130 Gy was noted as a positive prognostic factor for local tumor control, and location of recurrent lesions and time-to-progression (TTP, from the start of implantation to progression of the disease) were prognostic factors for survival. In addition, the advantages of ultrasound-guided seeds implantation include: (1) real-time guidance; (2) convenience and quick; (3) reproducible; (4) no additional radiation dose exposure, and disadvantages include: low image resolution and only 2D images can be obtained. Therefore, ultrasound-guided interstitial permanent ¹²⁵I seeds implantation is a preferred option for patients with cervical lymph node recurrences or metastases from head and neck, superficial maxillofacial, and base of the tongue cancers, while tumor invasion into the skin is a contraindication [35].

In 2019, Ji et al. evaluated the efficacy and prognostic factors of CT-guided radioactive ¹²⁵I seed implantation in the treatment of rHNC after EBRT. In total, 101 patients with rHNC underwent ¹²⁵I seed implantation under CT guidance, and the median D90 was 117 Gy. The median LC was 10 months, and the median survival was 15 months. In addition, nonsquamous cell carcinoma, D90 ≥ 120 Gy, lesion volume ≤ 20 cm³, and short-term efficacy (CR + PR) were correlated with better LC. High Karnofsky performance status (KPS) and lesion volume ≤ 20 cm³ were independent factors associated with survival [40].

Chen et al. evaluated the efficacy and safety of radioactive ¹²⁵I seed implantation under the guidance of CT as a salvage treatment for 25 patients with locally recurrent head and neck soft tissue sarcoma (rHNSTS) after surgery and EBRT. The median D90 was 152 (range, 106–179) Gy. When ¹²⁵I seeds were implanted, the objective response rate (ORR) was 76.0%. The 1-, 3-, and 5-year LPFS rates were 65.6%, 34.4%, and 22.9%, respectively, and the median LPFS was 16.0 months. The 1-, 3-, and 5-year OS rates were 70.8%, 46.6%, and 34.0%, respectively, and the median OS was 28.0 months. In addition, recurrent T stage and histological grade were found as prognostic factors for LPFS, while histological grade was noted as a predictor of OS [41]. The procedure and therapeutic efficacy of CT-guided ¹²⁵I seed implantation in a case of locally recurrent embryonal rhabdomyosarcoma were shown in Fig. 2.

Permanent interstitial ¹²⁵I seed implantation can provide targeted radiation to ensure that tumor cells are continuously killed for several months, and it also avoids high morbidity associated with EBRT or surgery. In recent studies, the 1-, 2-, 3-, 5-year LC ranged from 40.6% to 75.2%, 27.5% to 49.9%, 17.4% to 73.0%, 26.6% to 69.1%, and the 1-, 2-, 3-, 5-year OS ranged from 42.5% to 70.8%, 18.2% to 39%, 15.5% to 46.6%, 15.5% to 39%. The median LC was 10–24 months, and the median OS was 11–28 months. D90, tumor histological type, lesion volume, and short-term efficacy were prognostic factors for LC, and location of recurrent lesions, TTP, KPS, lesion volume and histological grade were associated with OS. Through LDR-BRT treatment, patients can achieve a better LC, and a lower incidence of toxicity can be attained.

The pioneering studies of the Radiation Therapy Oncology Group (RTOG) 9610 and 9911 have shown that conventional hypofractionated re-irradiation plus systemic therapy could achieve a 2-year survival rate of 15–26% in rHNSCC patients. However, the conventional re-irradiation resulted in severe (grade ≥ 3) acute and late toxicities with incidence rates of 63–78% and 22–37%, respectively. The median OS in these trials was only slightly higher than that of chemotherapy alone [22]. With the development of radiotherapy, a great number of scholars have investigated the therapeutic effects of the combination of radiotherapy and systemic therapy on the treatment of patients with rHNC (Table 4)[42‐53].

With the development of radiotherapy technology, its effects on patients with rHNC have been further improved. In particular, for patients who cannot undergo surgery, radiotherapy for recurrence is associated with a better prognosis. After salvage surgery or inoperable recurrence, patients can achieve long-term survival via radiotherapy. As the novel technologies of irradiation, IMRT, SBRT, HDR-BRT, and LDR-BRT have shown different characteristics. The comparison of different radiotherapy techniques for the irradiation of rHNC was shown in Table 5. Oncologists should pay further attention to the important role of radiotherapy in the treatment of rHNC. Besides, additional clinical research on the application of radiotherapy for diverse types of cancer is required for more reliable medical evidence, so that more patients can receive effective radiotherapy programs.