A Systematic Review on Re-irradiation with Charged Particle Beam Therapy in the Management of Locally Recurrent Skull Base and Head and Neck Tumors

Purpose To evaluate the clinical outcomes and treatment related toxicities of charged particle-based re-irradiation (reRT; protons and carbon ions) for the definitive management of recurrent or second primary skull base and head and neck tumors. Materials and Methods The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were applied for the conduct of this systematic review. Published work in English language evaluating the role of definitive charged particle therapies in the clinical setting of reRT for recurrent or second primary skull base and head and neck tumors were eligible for this analysis. Results A total of 26 original studies (15 protons, 10 carbon ions, and 1 helium/neon studies) involving a total of 1,118 patients (437 with protons, 670 with carbon ions, and 11 with helium/neon) treated with curative-intent charged particle reRT were included in this systematic review. All studies were retrospective in nature, and the majority of them (n=23, 88 %) were reported as single institution experiences (87% for protons, and 90% for carbon ion-based studies). The median proton therapy reRT dose was 64.5 Gy (RBE 1.1) (range, 50.0 – 75.6 Gy ), while the median carbon ion reRT dose was 53.8 Gy (RBE 2.5 – 3.0) (range, 44.8 – 60 Gy ). Induction and/or concurrent chemotherapy was administered to 232 (53%) of the patients that received a course of proton reRT, and 122 (18%) for carbon ion reRT patients. ReRT with protons achieved 2-year local control rates ranging from 50% to 86%, and 41% to 92% for carbon ion reRT. The 2-year overall survival rates for proton and carbon ion reRT ranged from 33% to 80%, and 50% to 86% respectively. Late ≥ G3 toxicities ranged from 0% to 37%, with brain necrosis, ototoxicity, visual deficits, and bleeding as the most common complications. Grade 5 toxicities for all treated patients occurred in 1.4% (n= 16/1118) with fatal bleeding as the leading cause. Conclusions Based on current data, curative intent skull base and head and neck reRT with charged particle radiotherapy is feasible and safe in well-selected cases, associated with comparable or potentially improved local control and toxicity rates compared to historical reRT studies using photon radiotherapy. Prospective multi-institutional studies reporting oncologic outcomes, toxicity, and dosimetric treatment planning data are warranted to further validate these findings and to improve the understanding of the clinical benefits of charged particle radiotherapy in the reRT setting.


Introduction
identified through other sources (Fig. 1). Publications of the same study population from the same institution or group of investigators and/or series that used charged particle reRT as a palliative treatment option were excluded. After these reports were removed, the remaining 106 eligible items were screened based on the previously discussed search criteria, and a total of 56 records were further excluded. In addition, articles without any well specified clinical endpoints (ie. local control, survival, or side effects) of charged particle radiotherapy were also excluded. Of the 50 remaining publications, review articles, abstracts, letters to the editor, commentaries, and studies with 5 reRT patients were excluded. Thus, 26 original studies were found to have sufficient focus and relevance to be incorporated and analyzed in this systematic review. A meta-analysis was not performed due to the heterogeneity of the reported reRT head and neck series, and the lack of consistent statistical power and value in this setting.

Results
A total of 26 original published studies (15 protons, 10 carbon ions, and 1 helium/neon ion particle radiotherapy) involving a total of 1118 patients (437 with protons, 670 with carbon ions, and 11 patients with helium/neon ions) who were treated with curative-intent charged particle reRT were included in this comprehensive systematic review [10 -35]. According to head and neck subsite, charged particle reRT with either protons or carbon ions was most commonly used for recurrent sinonasal, nasopharyngeal and salivary gland tumors, with squamous cell and adenoid cystic cell carcinoma as the most frequent histologies. Table 1 and 2 summarizes the different reRT studies using either proton, carbon ion or helium/neon ion therapy for                the management of recurrent head and neck malignancies. All studies were retrospective in nature, and the majority (23 of 26 studies, 88 %) belonged to single institutional experiences (13 of 15 studies, or 87% for protons, and 9 of 10 studies, or 90% for carbon ions; the only helium/neon study also involved one institution). Geographically 10 (67%) of the 15 proton series were reported by the United States, 3 (20%) from Asia, and the remaining 2 studies (13%) from Europe. With respect to carbon ion radiotherapy, all studies were implemented either in Asia 5/10 (50%) or in Europe 5/10 (50%). The study series that used charged particle therapy with Helium and Neon was carried out in the Lawrence Berkeley Laboratory in the USA. The included studies reported on patients treated between 1981 -2018, with fifty percent (n¼13/26) of studies (n¼7/15 protons, and n¼ 6/10 carbon ions) occurring in the last ten years, suggesting an increased availability and interest in the use of charged particle radiotherapy in the reRT setting. Median age for the cohort of patients treated with proton therapy reRT was 57.9 years (range, 36.0 -68.0 years), and 50.3 years (range, 38.0 -60.0 years) for the carbon ion reRT group. These were evaluated on a per-study basis, as also for the RT doses and time intervals between RT courses reported below. The median previous RT dose was 64.5 Gy (range, 55.0 -75. With a median follow up of 23 months, reRT achieved 2-year local control (LC) rates ranging from 50% to 86% for proton reRT, and 41% to 92% for carbon ion reRT with a median follow up of 19 months. The 2-year overall survival rates for proton and carbon ion reRT ranged from 33% to 80% and 50% to 86% respectively. By head and neck subsite, the LC rates for sinonasal carcinomas ranged from 59 to 77%; nasopharyngeal carcinoma, 50% to 80%; and for salivary gland tumors, 60% to 86% with proton-beam reRT. For carbon-ion reRT, the LC rates for sinonasal carcinomas ranged from 41% to 67%; nasopharyngeal carcinoma, 45 to 85%; and 41% to 92% for salivary gland tumors. With respect to OS, the rates for sinonasal, nasopharyngeal carcinomas, and salivary gland tumors were 57% to 76%, 33% to 74%, and 57% to 80% with proton beam reRT, and 60% to 70%, 59 to 68%, and 64 to 82% with carbon-ion reRT.
In regard to dosimetric analysis, only 3 (20%) of the 15 proton reRT series, and none of the carbon ion reRT series, reported a detailed analysis of the employed constraints and delivered doses for the organs at risk. There was noted to be significant variability across series in the size of the tumor volumes treated, the margins employed for gross tumor volume (GTV) to the clinical target volume (CTV) expansion, and CTV to the planning target volume (PTV) expansion (Tables 1 and 2). In all series, elective nodal irradiation (ENI) was not routinely performed. With respect to associated treatment toxicities, the rates for acute grade 3 toxicities ranged from 1% to 35%, with dysphagia, mucositis and radiation dermatitis being the most frequent, and the rates for late grade 3 toxicities ranged up to 37% for protons and up to 35% for carbons, with brain necrosis, ototoxicity, visual deficits, and bleeding were most commonly reported. There was a total of 16 cases of grade 5 reported toxicities for all treated patients (n¼ 16/1118, 1.4%) with fatal bleeding as the leading cause (Tables 1 and 2).

Discussion
Locoregional recurrences remain a common pattern of failure, morbidity and death in head and neck cancer patients [36 -38]. Even with multimodality therapy, patients typically have poor oncologic outcomes, with increased severe treatment related toxicities [39 -41].
Management of patients suffering from recurrent head and neck cancer is typically very challenging with no single treatment algorithm appropriate for all patients. Recurrent head and neck cancers are a heterogeneous group of patients, involving a number of different histologies and disease subsites. When evaluating a patient for reRT, a multidisciplinary evaluation that considers patient age, baseline comorbidities, performance status, histology, tumor biology, anatomic location, prior treatment constraints/toxicities, time interval since prior RT course, organ dysfunction (tracheostomy, feeding tube dependency), and patient goals and expectations is essential to determine which interventions may be the most appropriate [42]. To our knowledge this is the first comprehensive systematic review of the current existing data on the use of charged particle reRT for the definitive management of recurrent or secondary skull base and head and neck malignancies.
Published literature on the use of photon radiotherapy reRT for recurrent or second primary skull base and head and neck cancers have shown significant variability in the reported outcomes and toxicities, depending on photon therapy modality used (3D-conformal, IMRT, brachytherapy, IORT, SBRT), delivered dose and additional therapies employed, with 2-yr LC rates ranging from 19% -67% and 2-yr OS rates ranging from 11% -81% and with grade 3 toxicities up to 59%, and in some series risk of grade 5 toxicity (often secondary to carotid rupture) in up to 24% [7, 39 -41, 43 -54]. Our analysis demonstrated 2-year local control rates in the range of 50% to 86% for proton, and 41% to 92% for carbon ion reRT. The 2-yr OS rates for proton and carbon ion reRT ranged from 33% to 80% and 50% to 86% respectively. In regards to treatment related late grade 3 toxicities this ranged between 0% to 37% overall, with only sixteen grade 5 reported toxicities of the 1118 analyzed patients for a 1.4% rate (Tables 1 and 2). Table 3 summarizes reported clinical outcomes by reRT modality [7, 39 -41, 43 -54]. Although no conclusive comparisons can be drawn from these data given inherent differences in patient selection among the studies, results are suggestive that potentially more favorable LC and toxicity outcomes could be realized with charged particle therapy in properly selected patients. Of note, SBRT reRT data which do not include concurrent systemic therapy, demonstrate lowest rates of severe late toxicity likely due to smaller treatment target volumes in these studies, yet with largely similar LC compared to other photon studies. A proposition to utilize proton SBRT with concurrent immunotherapy to enhance LC, while limiting severe toxicity is currently under investigation [ClinicalTrials.gov Identifier: NCT03539198]. More advanced charged particle delivery techniques such as intensity-modulated proton beam therapy (IMPT), and intensity-modulated carbon ion therapy may further improve the therapeutic window for reirradiation. This may enable treatment of larger recurrent tumors with higher doses which would otherwise be more difficult to achieve with photon-based SBRT reRT approaches, due to greater collateral dose to organs-atrisk when treating larger treatment volumes [55].
This systematic review is not exempt from several limitations including most of the evaluated series were retrospective from single institutions, with significant variability in patient selection, recurrent disease sites, histologies, treatment technique, doses and fractionation employed, and the reported toxicities and outcomes. In addition, there were also statistical limitations and biases of the analysis due to the inherent heterogeneity of the reRT reports and lack of availability critical data within some of the reRT series, including prior RT doses, median interval times between RT courses and not having well defined study endpoints that could impact on the interpretation of the outcomes. There are no randomized data comparing outcomes with photons (IMRT, SBRT) versus charged particle reRT (protons, carbon ions). With the continuous increase in availability of centers with capability to deliver charged particle radiotherapy we can anticipate more data will emerge, and help to further elucidate the potential clinical benefits of these treatment modalities.
Efforts should continue to be made to design clinical trials that can collect robust data on the use of charged particle reRT, and would advance the management of these patients resulting in a better understanding on the selection of patient candidates for this treatment paradigm. A published patient selection RPA classification may facilitate patient stratification in future studies using charged particle reRT, to inform best study design and treatment strategies [56]. Despite the complexity in the management of these malignancies, the current accumulated information on the use of charged particle reRT for recurrent or second primary skull base and head and neck cancers is encouraging and may advocate the potential clinical advantages of the use of charged particle in this setting.

Conclusions
Based on the current available data, curative intent head and neck reRT with charged particle radiotherapy is feasible and well tolerated in the majority of patients, with the potential to improve oncologic and toxicity outcomes in well-selected cases.
Prospective studies of patients reporting in more depth oncologic outcomes and dosimetric treatment planning data are necessary to further validate these findings.