Clinical trials for treating recurrent head and neck cancer with boron neutron capture therapy using the Tsing-Hua Open Pool Reactor

Head and neck (HN) cancer has a high frequency worldwide [1], but in Taiwan, China, squamous cell carcinoma (SCC) of the oral cavity is an endemic disease due to the habit of chewing betel nuts. In 2014, more than 7000 individuals were diagnosed with HN cancer [2]. Despite progress in multimodality treatment for this disease, locoregional recurrence is common and poses a therapeutic challenge, especially for patients who already received high-dose photon irradiation [3–5]. Consequently, there is a pressing need for other treatment options for recurrent HN cancer patients.

Boron neutron capture therapy (BNCT) is a type of particle radiotherapy based on the nuclear capture and fission reactions that occur when non-radioactive boron (¹⁰B) is irradiated with thermal neutrons (< 1 eV) to yield high-energy alpha-particles (⁴He) and recoiling lithium (⁷Li) nuclei. The path-lengths of these particles are approximately one cell diameter making BNCT a potentially ideal way to selectively destroy malignant cells and spare normal tissues if a sufficiently high concentration of ¹⁰B is localized in tumor cells [6]. Based on a literature review, high response rates (60–83%) and low toxicity rates were reported after BNCT for recurrent HN cancer in both single- and two-fraction treatment protocols [7–9]. As reported a more recent Finnish study, only one had grade 4 toxicity among 29 evaluated patients who received BNCT. The incidence of grade 3 toxicity ranged from 20% to 54%. BNCT for recurrent HN cancer patients was initiated in Taiwan, China in 2010 based on the Japanese and Finnish studies [7–9].

The National Tsing-Hua University has a 2-MW open pool research reactor [THOR]; manufactured by General Atomics, San Diego, CA, USA, which is the only epithermal neutron source for BNCT research in Taiwan, China. As a first step, we designed neutron beam and developed a treatment planning system (TPS) for clinical applications [10–12]. In 2004, the construction of our BNCT treatment facility was completed. The beam provided epithermal neutron fluxes > 1.0 × 10⁹ n/cm²/s with very low fast neutron and gamma ray contamination (Fig. 1). Later, the beam quality was measured and these data were imported into the TPS (THORplan), which provided 3-dimensional dose displays on computed tomographic (CT) images, and generated dose volume histograms for both the tumor and normal structures [11, 13] (Fig. 2). The purpose of this review is to summarize our clinical experience in treating patients with locally recurrent HN cancer at THOR.

From the experience of our first clinical BNCT trial and other studies [7, 9], re-recurrences near re-irradiated sites were common, despite initially good responses. We hypothesized that by irradiating a larger field around the recurrent GTV with photons, we could obtain better local control IG-IMRT had been used alone as salvage treatment for recurrent HN cancer with acceptable toxicity [4]. Therefore, we developed a new protocol in 2013 combining IG-IMRT with BNCT. The inclusion and exclusion criteria were similar to our previous two-fraction BNCT protocol, except that a longer interval (at least 6 months) was necessary between previous RT and BNCT. BNCT treatment planning was done as prescribed for the first fraction in the previous protocol. For IG-IMRT planning, a selected margin (3–5 mm) around the GTV was added to generate the clinical target volume (CTV), and an additional 3-mm margin was added around the CTV to generate the planning target volume. In two patients, lymphatic drainage areas adjacent to the GTVs were also included in the CTV. Additionally, special attention was paid to limit the dose to critical organs such as the carotid artery, spinal cord, brain stem and mandible. Patients were followed up similar to our previous protocol. After approval of the new protocol in 2014, we recruited 9 eligible patients, 4 of whom had squamous cell carcinomas of the oropharynx, 3 has squamous cell carcinomas of the oral cavity, one mucoepidermoid carcinoma of the parotid gland, and one osteogenic sarcoma of the mandible. The median radiation dose before BNCT was 66 Gy (range 60–102 Gy).

Seven patients were treated with combined therapy, one with BNCT alone due to severe acute toxicity, and another due to old age and the risk of carotid bleeding. The median BNCT doses were, 19.4 Gy-Eq and IG-IMRT: 45 Gy (in 25 fractions). The median follow-up time was 11.7 months (range 4.8–25.9 months). Regarding acute toxicities, almost all had low-grade oral mucositis, radiation dermatitis and alopecia, which were similar to that observed in the previous two-fraction BNCT protocol [14]. Regarding grade 3 toxicity, dysphagia and tumor pain were seen in two patients, infection in one, and facial edema in one. One patient with a recurrent oral cancer showed grade 4 oral bleeding. He was successfully treated by carotid artery embolization. Another patient had grade 4 dyspnea following facial edema and his symptoms subsided following tracheostomy. Three patients showed CR, as evidenced by PET scans 3 months after combined treatment, and another 3 showed PR by MRI. The other 3 had stable disease. The 1-year overall survival rate for all patients was 56%, and one patient was disease-free 25.9 months after combined treatment.

With chemotherapy alone, most patients who are diagnosed with inoperable, locoregionally recurrent HN cancer at previously irradiated sites die of the disease within a few months [3]. Re-irradiation with BNCT with limited toxicity may constitute successful salvage therapy for these patients. In reality, further local recurrence after BNCT is a major cause of treatment failure. There are many possible explanations for local failure following BNCT, including insufficient uptake and non-homogeneous BPA distribution within the tumor, insufficient depth penetration by epithermal neutrons, and an insufficient radiation dose or coverage of the CTV. By combining image-guided fractionated photon therapy with BNCT, we hoped that the second trial would decrease the recurrence rate following BNCT without significantly increasing toxicity. Theoretically, local control should be improved when a larger volume around the recurrent tumor bed (i.e., the CTV) is further treated with photon therapy through IMRT, because the BNCT dose to the adjacent normal tissue would be low. Thus, it may be appropriate to choose an “adequate” margin surrounding the GTV based on previous RT fields, dosages, nearby critical organs, and the risk of re-recurrence. Because the patient numbers of both trials were small and they had disparate primary sites, histopathology, clinical staging, and accumulated radiation doses before BNCT, it was difficult to compare the efficacy and toxicity of the two trials. However, from the dose volume histograms of BNCT and IMRT plans, the dose to normal tissues was often lower with BNCT than with IMRT in the same patient. We hypothesized that better local control with BNCT with combined IMRT, covering a larger tissue volume, at the expense of more severe toxicity, compared to a second application of BNCT. However, it is still too early to draw any conclusions regarding the second trial. Nevertheless, it remains very challenging to treat recurrent HN cancer patients with further radiation, even with a target type of radiotherapy such as BNCT. Recently, promising results from some clinical trials testing systemic treatments, such as molecularly-targeted therapy and immunotherapy for recurrent HN cancer have been reported [20, 21]. In the future, multimodality treatment combining BNCT with new therapeutic approaches such as the use of pulsed ultrasound [22] should be evaluated to avoid excessive normal tissue injury and obtain more durable tumor responses.