Proton Radiotherapy for Recurrent or Metastatic Head and Neck Cancers with Palliative Quad Shot.

Purpose
Some patients with previously treated, unresectable, recurrent or metastatic head and neck malignancies are not amenable to curative-intent treatment. Here, we investigated the quad-shot (RTOG 8502) regimen of hypofractionated proton radiotherapy (RT) in that patient population.


Materials and Methods
From 2013 to 2015, 26 patients with recurrent or metastatic cancers were treated with palliative proton RT to the head and neck with quad shot (3.7 Gy twice daily for 2 days). Patient characteristics and survival data were reviewed.


Results
Seventeen (65%) patients received ≥ 3 quad-shot cycles and 23 (88%) had prior head and neck RT. Overall palliative response was 73% (n = 19). The most common presenting symptom was pain (50%; n = 13), which improved in 85% (n = 22) of all patients. The overall grade-1 acute-toxicity rate was 58% (n = 15), and no acute grade 3 to 5 toxicities were observed.


Conclusions
The proton quad-shot regimen demonstrates favorable palliative response and toxicity profile, even in patients that received prior RT.


Introduction
Appropriate management of previously treated, unresectable, recurrent or metastatic head and neck malignancies remain a clinical challenge [1]. Salvage, full-dose external beam radiation (EBRT) with or without systemic therapy may offer a meaningful survival benefit (10%−48% at 2 years) at the expense of increased risk and degree of toxicity [2,3]. However, many patients are not amenable to curative-intent treatment because of tumor location, extent of disease, poor performance status, and comorbidities. Palliative radiotherapy (RT) is commonly used to control symptoms, improve quality of life (QOL), and possibly, prolong survival in these circumstances [2,4,5].
The optimal, palliative RT regimen has not been determined. In recent years, short-course hypofractionated RT has been considered more suitable than protracted, conventionalfractionation RT because it provides equivalent symptomatic improvement, tumor response, and survival outcomes while shortening overall treatment time and minimizing effects [5][6][7]. Although a variety of different hypofractionation schemes have been used clinically (from 2.5-8 Gy per fraction, to a total dose of 20-48 Gy), the quad shot regimen is believed to offer the best combination of efficacy, toxicity, and patient convenience [4,8].
The RTOG 8502 regimen, colloquially referred to as quad shot, consists of 3.7 Gy twice daily for 2 consecutive days at 4-week intervals and was originally used for treatment of pelvic malignancies [9]. Early studies demonstrated overall tumor responses of 53% to 77%, with a 53% to 85% improvement in symptom palliation in patients with end-stage head and neck cancers [10][11][12]. We previously evaluated the combination of quad shot with intensitymodulated RT (IMRT) and demonstrated a comparable palliative response [5], with only 5% of patients experiencing grade 3 to 4 acute toxicities. This regimen is now our current institutional standard for patients requiring palliative RT for incurable cancers of the head and neck [5].
However, there remains a significant population of patients with advanced head and neck cancers who suffer from tumor-related symptoms but are no longer candidates for photonbased EBRT. Some tumors recur in a previously heavily irradiated region, limiting additional radiation because of the risk to adjacent healthy structures, such as the spinal cord, brain stem, and temporal lobe. Patients with locally invasive disease near radiosensitive organs at risk (OAR), including the optic nerve and optic chiasm, and patients unable to tolerate additional cycles of palliative RT after the first IMRT quad shot cycle because of treatment toxicities, are poor candidates for photon-based EBRT [5].
intermediate doses to surrounding nontarget organs is unavoidable. In contrast to photonbased EBRT, proton therapy deposits most of the radiation energy within the target area rather than the surrounding healthy tissues outside the target volume. Proton therapy is preferable for patients with recurrent tumors in previously irradiated areas and tumors near radiation-sensitive structures because of its physical properties, which allow for a sharp drop-off in dose in the surrounding healthy tissues, including radiosensitive OARs (Figure).
To our knowledge, this is the first study evaluating the use of proton RT in palliation of head and neck cancers. We investigate palliative response, treatment toxicity, and survival of patients with incurable, recurrent or metastatic head and neck malignancies treated with palliative proton RT with quad shot.

Patients
This retrospective study was independently reviewed and approved by the institutional review board. In September 2013-15, 26 patients with incurable, recurrent or metastatic head and neck cancers were treated at our center with ≥ 1 cycle of palliative, hypofractionated proton RT to the head and neck with the quad shot regimen. Patients were determined ineligible for curative treatment by a multidisciplinary team generally because of the extent of their disease, medical comorbidities, and distant metastatic burden. Furthermore, 16 patients (62%) were considered poor candidates for further photon-based EBRT because of tumor location or previous local RT history.

Palliative Proton Radiotherapy
The palliative proton quad shot regimen prescribes 3.7 Gy relative biological effectiveness (RBE) per fraction, assuming an RBE value of 1.1. Each treatment cycle consisted of twicedaily fractions administered over 2 consecutive days to a total of 14.8 Gy (RBE) per cycle. Each cycle was repeated every 21 to 25 days in the absence of in-field disease progression or significant acute toxicity, as previously reported by Paris et al [10]. An additional rest period of 1 to 3 weeks between RT cycles was used in instances of acute toxicities or other concerns. Treatment volumes were reviewed before each cycle and replanned to account for tumor volume reduction for patients with significant response to therapy. The palliative radiation included both the symptomatic lesion, either primary site or regional disease, and any additional regional disease that could be safely treated.
Patients were treated with chemotherapy with cytotoxic or targeted agents at the discretion of the attending radiation oncologist and medical oncologist.

Radiotherapy Technique
Computed tomography (CT) imaging was used for patient simulation before each cycle with a thermoplastic 5-point head and neck mask. If available, positron emission tomography/CT or magnetic resonance imaging studies were fused to the CT simulation scan for target volume delineation. Gross tumor volume included symptomatic gross disease and other concerning large-volume disease and was identified through diagnostic imaging and clinical examination, then contoured onto the CT simulation images by the radiation oncologist. A patient-specific clinical target volume (CTV) was added to cover adjacent areas at high risk for microscopic tumor spread at physician discretion. The planning target volume was created based on the CTV with an additional 3 to 5-mm margin dependent upon available imaging guidance during treatment and setup uncertainty. This was reduced to as low as 1 mm when near critical healthy tissues. In the distal beam direction, the plan was generated to cover the CTV, and the plan was evaluated for a range uncertainty of + 2.5% + 2 mm. A smear radius of 7 mm, equal to the lateral setup uncertainty, was used.
For patients with prior local RT, the spinal cord and brainstem were prioritized as the OAR, with 70 Gy as the maximum allowable limit. A constraint guideline of a total maximum point dose of 60 Gy in 2-Gy equivalents from all treatments was used. The RT was delivered with uniform scanning proton beams, with beam-specific apertures and compensators.

Treatment Response and Toxicity Evaluation
Palliative response was defined as subjective relief of presenting symptom(s) or objective reduction of the gross tumor, as determined though physical examination or radiographic tumor response. Evaluation of the objective tumor response was performed 4 to 12 weeks after the last cycle of RT. Positron emission tomography response criteria in solid tumors and Response Evaluation Criteria in Solid Tumor (version 1.1) were used. Toxicity was scored for all patients, with toxicity events according to the Common Terminology Criteria for Adverse Events (version 4.03), with acute toxicity defined as an adverse event occurring within 3 months of palliative RT completion.

Statistical Methods and Definitions
Follow-up was maintained until death or date of last contact. Overall survival (OS) was measured from start of palliative proton RT until the date of death or last follow-up. Progression-free survival (PFS) was calculated from the start date of palliative proton RT until the date of in-field tumor progression or death. The OS and PFS rates were determined by the Kaplan-Meier technique. A Spearman q test was used to examine correlation between palliative response and clinical factors.

Palliative Treatment.
Seventeen patients (65%) completed ≥ 3 cycles of quad shot (Table 2). Seven patients (27%) received treatment for the primary site of their disease, and the remainder (n = 19; 73%) received palliation for symptomatic, regional disease. Of the remaining 9 patients who did not receive ≥3 cycles of quad shot, 4 patients (44%) only received 1 RT cycle, one half because of local tumor growth, and the remaining half because of progression of distant metastatic lesions. The other 5 patients (56%) received only 2 cycles of proton palliative RT, 2 because of local tumor progression, 1 because of logistical concerns, and 2 converted to systemic palliation or supportive care because of concurrent distant metastases. Two of the 26 patients (8%) were previously treated with 1-2 quad shot cycles by IMRT.
Chemotherapy administration was at the discretion of the treating medical oncologist and was based on factors that included performance status and presence of systemic disease. Twelve patients (46%) received systemic chemotherapy as part of their palliative regimen. Chemotherapy was administered concurrently in 5 of the 26 patients (19%) patients and induction-concurrent adjuvant was administered in 7 patients (27%). Seven of those 12 patients (58%) received a targeted agent (5 received cetuximab and 1 each received pembrolizumab and cabozantinib). The remaining 5 patients received cytotoxic chemotherapy, with 2 receiving methotrexate and 3 with a combination of carboplatin and paclitaxel.
Using Spearman's rho rank correlation coefficient, palliative response was significantly correlated with increasing number of Quad Shot cycles (0.720; P < 0.001) but not with KPS, prior RT, prior surgery, palliative chemotherapy, histology or T/N stage (Table 3).
On univariate analysis, age and a greater number of Quad Shot cycles were significantly associated with both PFS and OS (Table 4).

Toxicity.
No grade 3 to 5 acute toxicities were observed in our study. Grade 2 acute toxicities occurred in 2 patients (8%); 1 patient with recurrent soft-palate disease experienced mucositis, and 1 patient with recurrent cutaneous cancer had dermatitis. The rate of overall grade 1 acute toxicity was 58% (n = 15), with the most common adverse events being fatigue (8 patients; 31%) and dermatitis (6 patients; 23%). See Table 2 for a summary of the most common acute toxicities observed. Grade 3 late fibrosis was observed in 1 patient 5 months after their third quad shot cycle. The patient had a recurrent squamous cell carcinoma of the lip, which was heavily pretreated with 2 courses of radical photon-based EBRT.

Discussion
Incurable head and neck cancer has an extremely poor prognosis and is often accompanied by a heavy symptomatic burden because of constriction of the trachea and esophagus. The estimated median survival is 7-10 months even after active palliative treatment [1]. The median OS in our study was 9.1 months; however, our study demonstrated improved symptom response with the proton quad shot regimen. Although palliative systemic therapy can achieve acceptable tumor response rates ranging from 12% to 48% with less late toxicity [13], palliative RT remains an important option for patients with recurrent head and neck cancers who have local tumor progression or bulky tumor mass despite multiple lines of systemic therapy. Furthermore, proton palliative RT can benefit those who cannot receive photon-based palliative RT because of tumor location or prior local radiation, and it provides benefit to patients who have had previous radiation through decreased cumulative dose to sensitive surrounding tissue [14]. The proton quad shot regimen offers an alternative option for symptom palliation for patients not amenable to palliative photon RT.
To our knowledge, this is the first study reporting on proton therapy for palliation of head and neck malignancies. Treatment with the proton quad shot RT regimen in the current study achieved a palliative response rate of 73% (n = 19 patients); among whom, 63% (n = 12) were considered ineligible for additional photon-based RT. The overall tumor response rate (complete response + partial response) was 61% (n = 16 patients), which is comparable to other studies with palliative, short-course, hypofractionated palliative RT to treat recurrent head and neck cancers [4,5,8]. However, the toxicity rate among our cohort was quite low. Nonspecific grade 1 fatigue was the most commonly observed acute toxicity found in 31% of patients (n = 8), followed by grade 1 dermatitis in 23% of patients (n = 6). Grade 2 acute toxicities were observed in 2 patients (8%). The case of grade 2 mucositis was expected for the patient with a soft-palate recurrent tumor, and grade 2 dermatitis was observed in another patient with cutaneous disease.
The lower rates of toxicity observed may be due to a variety of factors. First, proton therapy offers the unique advantage of delivering a negligible radiation dose to healthy tissues behind the target volume. In our previous report on the palliative quad shot regimen study with IMRT to treat similar patients [5], the incidence of grade 2 to 3 acute toxicity was 33%, as compared with 8% (n = 2 patients) of grade 2 acute toxicity and an absence of grade 3 acute toxicities observed in the current study. It is not surprising that grade 1 dermatitis was one of the most commonly observed toxicities considering the entrance of the high-energy proton beam through the skin. Even so, the skin dosage could be further minimized through pencil beam scanning with intensity modulation, which may decrease both the acute and late skin and superficial tissue toxicities [15]. Secondly, the quad shot regimen provided a much longer interval (4 weeks) for normal tissue repair after each RT cycle, which may reduce acute and late toxicities. Other studies [16][17][18][19] of hypofractionated, palliative RT with a total dose of 30 to 50 Gy in 5 to 16 fractions for 2 to 5 weeks demonstrated a decent palliative response but with grade 3 acute mucositis rates as high as 18% to 66%. Another study examined the ''IHF2SQ regimen (HypoFractionnée 2 Séances Quotidiennes)'' of 3 Gy twice a day for 2 fractions, twice a week every other week for weeks 1 to 7 with concurrent, platinum-based chemotherapy for advanced head and neck cancers [20]. The reported grade 3 acute toxicity was only 4%, which is comparable to the 5% observed in our previous photon-based quad shot study with IMRT [5], whereas the median PFS was only 3.8 months [20]. Although acute toxicity evaluation was not standardized in this study, we believe the adverse events observed are representative of patients' palliative symptom relief.
Three patients (12%) presented with late grade 3 to 4 toxicity, which is higher than expected for the enrolled irradiation-naïve patients. In comparison, no severe late toxicities were noted in numerous other studies of the quad shot regimen conducted at different institutions, independent of 2-dimensional RT or IMRT use [5,10,11,17]. One patient in the prior studies with recurrent squamous cell carcinoma of the lip had late grade 3 fibrosis, likely from the cumulative effect of 2 courses of RT.
The quad shot regimen is currently the only hypofractionated RT method that allows target revision for each cycle of palliative RT to account for tumor shrinkage. The adaptive nature of this regimen allows for decreased healthy tissue dose [16][17][18][19]. Further RT is generally not offered for patients who did not benefit from previous cycles of the quad shot regimen because of tumor progression or other considerations, to avoid unnecessary side effects or ineffectual treatment. Therefore, the use of the proton quad shot regimen provides an alternative for patients not amenable to photon-based EBRT.
Although the proton quad shot regimen is beneficial for palliation, the expense of proton therapy cannot be overlooked. Although significantly more than photon-based RT, the cost of proton therapy is likely to decrease as the number of operational proton centers increases. Notwithstanding cost, the proton quad shot regimen balances effective palliation with toxicity, particularly in patients with previous courses of RT or lesions near critical structures. Additional work is needed to better delineate the value of proton therapy in this setting.
Our study reports on our institutional experience with incurable, recurrent or metastatic head and neck cancers. It is limited by its retrospective nature and the lack of a formal QOL assessment or standardized reporting of adverse events, including limitations of capturing of toxicity between cycles. However, the improvement in QOL can be inferred from the high rates of palliative and tumor response, as well as the notably low incidence and severity of toxicities. In conclusion, for patients with heavily treated, incurable, recurrent or metastatic head and neck cancer, proton hypofractionated palliative RT with the quad shot regimen resulted in a 73% rate (19 of 26 patients) of palliative response with minimal acute and late toxicities. The quad shot regimen by proton therapy is safe and effective and serves as a feasible alternative for patients not amenable to photon-based RT. Prospective studies will be needed to address which patient population and clinical situations would benefit most from proton hypofractionated RT.

Figure.
Quad shot proton treatment plan demonstrating optic nerve and cord sparing because of the sharp dose falloff of protons in a patient with verrucous carcinoma of the right alveolar ridge, with local recurrence in the right maxilla. Presenting symptoms and palliative response by number of completed quad shot cycles and acute toxicities.