Int. J. Radiation Oncology Biol. Phys., Vol. 82, No. 1, pp. e67–e75, 2012Copyright � 2012 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/$ - see front matter

jrobp.2010.09.057

doi:10.1016/j.i

CLINICAL INVESTIGATION Head and Neck Cancer

BORON NEUTRON CAPTURE THERAPY IN THE TREATMENT OF LOCALLYRECURRED HEAD-AND-NECK CANCER: FINAL ANALYSIS OFA PHASE I/II TRIAL

LEENA KANKAANRANTA, M.D.,* TIINA SEPP€AL€A, PH.D.,yz HANNA KOIVUNORO, M.SCI.,yz

KAUKO SAARILAHTI, M.D.,* TIMO ATULA, M.D.,x JUHANI COLLAN, M.D.,* EERO SALLI, PH.D.,k

MIKA KORTESNIEMI, PH.D.,k JOUNI UUSI-SIMOLA, PH.D.,yk PETTERI V€ALIM€AKI, M.SCI.,yz

ANTTI M€AKITIE, M.D.,x MARKO SEPP€ANEN, M.D.,{ HEIKKI MINN, M.D.,# HANNU REVITZER, M.SCI.,**MAURI KOURI, M.D.,* PETRI KOTILUOTO, PH.D.,yy TOM SEREN, LIC. TECH,yy IIRO AUTERINEN, M.SCI.,yy

SAULI SAVOLAINEN, PH.D.,yk AND HEIKKI JOENSUU, M.D.*

*Department of Oncology, xDepartment of Otorhinolaryngology, and kHelsinki and Uusimaa Hospital District Medical ImagingCenter, Helsinki University Central Hospital, Helsinki, Finland; yDepartment of Physics, University of Helsinki, Helsinki, Finland;zBoneca Corporation, Helsinki, Finland; {Turku PET Centre, Turku University Hospital, Turku, Finland; #Department of Oncology,Turku University Central Hospital, Turku, Finland; **Aalto University School of Science and Technology, Esopo, Finland; and yyVTT

Technical Research Centre of Finland, Espoo, Finland

Reprincology, HFIN-0002(+358)-9-Suppor

of FinlandConflic

S., S.S., aBoneca Cture therasation fro

Purpose: To investigate the efficacy and safety of boron neutron capture therapy (BNCT) in the treatment of in-operable head-and-neck cancers that recur locally after conventional photon radiation therapy.Methods and Materials: In this prospective, single-center Phase I/II study, 30 patients with inoperable, locally re-curred head-and-neck cancer (29 carcinomas and 1 sarcoma) were treated with BNCT. Prior treatments consistedof surgery and conventionally fractionated photon irradiation to a cumulative dose of 50 to 98 Gy administeredwith or without concomitant chemotherapy. Tumor responses were assessed by use of the RECIST (Response Eval-uation Criteria in Solid Tumors) and adverse effects by use of the National Cancer Institute common terminologycriteria version 3.0. Intravenously administered L-boronophenylalanine–fructose (400 mg/kg) was administeredas the boron carrier. Each patient was scheduled to be treated twice with BNCT.Results: Twenty-six patients received BNCT twice; four were treated once. Of the 29 evaluable patients, 22 (76%) re-sponded to BNCT, 6 (21%) had tumor growth stabilization for 5.1 and 20.3 months, and 1 (3%) progressed. The me-dian progression-free survival timewas 7.5months (95%confidence interval, 5.4–9.6months). Two-year progression-free survival and overall survival were 20%and 30%, respectively, and 27%of the patients survived for 2 years with-out locoregional recurrence. The most common acute Grade 3 adverse effects were mucositis (54% of patients), oralpain (54%), and fatigue (32%).Three patientswere diagnosedwith osteoradionecrosis (eachGrade 3) andonepatientwith soft-tissue necrosis (Grade 4). Late Grade 3 xerostomia was present in 3 of the 15 evaluable patients (20%).Conclusions: Most patients who have inoperable, locally advanced head-and-neck carcinoma that has recurred ata previously irradiated site respond to boronophenylalanine-mediated BNCT, but cancer recurrence after BNCTremains frequent. Toxicity was acceptable. Further research on novel modifications of the method iswarranted. � 2012 Elsevier Inc.

Boron neutron capture therapy, Radiotherapy, Head-and-neck cancer, Boronophenylalanine.

INTRODUCTION

Prognosis in head-and-neck cancer patients whose diseaserecurs after surgery and radiotherapy remains poor. Salvage

t requests to: Heikki Joensuu, M.D., Department of On-elsinki University Central Hospital, Haartmaninkatu 4,9 Helsinki, Finland. Tel: (+358)-9-47173208; Fax:47174202; E-mail: heikki.joensuu@hus.fited by grants from theAcademy of Finland, Cancer Society, and Helsinki University Central Hospital research funds.t of interest: L.K., T.S., H.K., J.C., M. Kortesniemi., J.U.-nd M. Kouri have received financial compensation fromorporation for the work done to treat boron neutron cap-py patients. S.S. and H.J. have received financial compen-m Boneca Corporation for consultations. The study

e67

surgery leads to durable disease control in 15% to 30% of thepatients (1–3). When cancer is considered unresectable, thedisease is almost uniformly fatal with a median survival time

sponsor, Boneca Corporation, had no role in designing the study,in performing data analysis or interpretation of the data, or in writ-ing of the report. Trial nurses employed by the sponsor helped theinvestigators in data collection. The corresponding author had fullaccess to the data and had the final responsibility for the decision tosubmit the manuscript for publication.Acknowledgment—The authors thank the study nurses for profes-sional help in conducting the study.Received July 26, 2010, and in revised form Sept 13, 2010.

Accepted for publication Sept 28, 2010.

e68 I. J. Radiation Oncology d Biology d Physics Volume 82, Number 1, 2012

of only a few months (3, 4). Palliative chemotherapy isregarded as a standard treatment, but full-dose, repeat irradi-ation is an option for selected patients and may result indurable control in approximately 10% of such patients(4–7). However, reirradiation with or without concomitantchemotherapy is associated with a risk for severe organinjury due to a high accumulated radiation dose in up to80% of the patients and treatment-related death in up to15% (4–8). Efficient, well-tolerated treatment options arelacking, and patients with unresectable recurrent head-and-neck carcinoma pose a difficult therapeutic problem.

In this study we evaluated boron neutron capture therapy(BNCT) as treatment of inoperable head-and-neck cancersthat have recurred locally after surgery and prior conven-tional radiotherapy or chemoradiation therapy. BNCT isbased on the neutron capture reaction that occurs when non-radioactive boron (10B) is irradiated with neutrons of low(thermal) energy. This causes a nuclear decay yieldinghigh-energy a particles and recoiling lithium (7Li) nuclei(9). Because a and 7Li have short range in tissue and theyproduce dense ionization along their tracks, most radiationeffect is local and occurs within the cells that contain boron.Therefore, the success of BNCT depends on a selective up-take of sufficient amounts of 10B into cancer cells comparedwith normal tissues. This can be achieved by use of boroncarriers that are preferentially taken up by cancer, such asa derivative of phenylalanine, L-boronophenylalanine (L-BPA). After administration of a boron carrier compound(usually by an intravenous infusion), the tumor site is irradi-ated with neutrons. At present, the source of neutrons isa nuclear reactor.

BNCT might be of particular clinical value in the treat-ment of recurrent irradiated tumors, where the radiationdose to the normal tissues needs to be kept low whilea high dose is administered to the tumor. Hypothetically,this can be achieved by use of BNCT, because head-and-neck carcinomas accumulate L-BPA (10). BNCT is usuallyadministered only once or twice, and a high single dose ofirradiation is delivered to the target volume within approxi-mately 1 h instead of a time period of 6 to 7 weeks when con-ventionally fractionated photon radiation therapy is used.Results from a few small series (11–13) and case reports(14, 15) suggest that BNCT is effective in the treatment oflocally recurrent head-and-neck cancer, and that mosttumors respond to BNCT.

This study evaluates the efficacy and tolerability of boro-nophenylalanine (BPA)–based BNCT, administered twice atan interval of fewweeks, in the treatment of inoperable head-and-neck cancer that has recurred locally despite conven-tional radiotherapy. We previously reported the results ofPart I (Phase I) of the study based on 12 patients and amedianfollow-up time of 14 months (16). Of the 12 patients treated,10 (83%) responded to BNCT despite prior photon radiationtherapy in history, and adverse effects related to BNCTwereconsidered acceptable. These results were judged favorableenough to allow expansion of the study to a Phase II trial(Part II). Here we report the final analysis of the study based

on 30 patients and a median follow-up of 31 months. To ourknowledge, this is the first registered Phase II trial address-ing BNCT in the treatment of head-and-neck cancer.

METHODS AND MATERIALS

Study design and objectivesThirty patients with inoperable, locally recurred head-and-neck

cancer were accrued to this Phase I/II, prospective, noncompara-tive, open-label, single-center study (identifier NCT00114790 athttp://www.clinicaltrials.gov). The primary objectives were re-sponse to BPA-mediated BNCT, treatment safety, response oftumor-related symptoms, and tumor response when evaluatedwith fluorine 18 (18F)–labeled L-boronophenylanaline (18F-L-BPA) via positron emission tomography (PET) whenever available,and the secondary objectives were progression-free survival (PFS)and overall survival.

PatientsThe study participants were required to have histologically ver-

ified head-and-neck cancer that had been irradiated with conven-tional radiotherapy, and locally recurred cancer was consideredinoperable by head-and-neck surgeons. Whenever 18F-fluoro-BPAPET was performed to estimate tumor BPA uptake, we requiredthat at least 2.5 times more 18F-fluoro-L-BPA accumulated in thetumor compared with corresponding contralateral normal tissue.Other exclusion criteria included cancer outside of the locoregionalregion,World Health Organization Performance Status greater than3, ongoing systemic anticancer therapy, a time interval of less than3 months from prior radiotherapy, and inability to lie in a cast for 30to 60 minutes. Our institutional ethics committee approved the trialprotocol (HUS429/E6/03). The patients provided written informedconsent before initiation of the study procedures.

Positron emission tomographyPETwas carried out with either a GEAdvance or a GEDiscovery

STE scanner (GE Medical Systems, Milwaukee, WI) operated inthree-dimensional (3D) mode by use of intravenously administered18F-fluoro-L-BPA at activity of 160 to 240 MB as the tracer (17).Tumor–to–normal tissue ratios were measured from static emissionscans obtained 20 to 40 minutes after 18F-fluoro-L-BPA injection.

Radiation dose planningThe head-and-neck region was imaged with computed tomogra-

phy (CT), contrast-enhanced T1-weighted magnetic resonance im-aging (MRI), and PET, after which the images were co-registered(aligned). A 3D model was constructed via CT. MRI and PETwere used to define the target volume. The clinical target volume(CTV) included the gross tumor volume (GTV) with a margin ofapproximately 1.5 cm. No further margin was added to the planningtarget volume (PTV) around the CTV. All macroscopic tumorswere included in the CTV, but no attempt was made to irradiate sub-clinical metastases (16). The 3D Monte Carlo software packageSERA (Simulation Environment for Radiotherapy Applications;Idaho National Laboratory, Idaho Falls, ID) was used for dose plan-ning (18). The doses to the tumor, target volume, and normal tissueswere computed individually as a function of the blood average bo-ron concentration during irradiation assuming a tumor–to–wholeblood boron concentration ratio of 3.5:1. Tissue compositions asdefined by the International Commission on Radiation Units &Measurements were used (19).

BNCT for head-and-neck cancer d L. KANKAANRANTA et al. e69

Both physical and radiobiologically weighted doses were used indose planning. The weighted total dose (DW) was defined as thesum of the physical dose components (Di) multiplied by weightingfactors (wi) of each dose component in a tissue (20, 21). Whenweighted doses are being reported, the annotation ‘‘(W)’’ isadded after the dose for clarity. The mucosal membrane absorbedphysical dose was selected as the dose-limiting factor and was lim-ited to 6 Gy or less for each BNCT treatment. The mean PTV nor-mal tissue dose was limited to less than 4 Gy (W) per treatment inthe first 3 study participants, to less than 7 Gy (W) in the following 4patients, and to less than 10 Gy (W) in subsequent patients.

Table 1. Characteristics of patients and tumors

CharacteristicNo. of

patients (%)Median (minimum,

maximum)

Patient characteristicsAge (y) 61.0 (37, 79)Gender

Male 14 (47)Female 16 (53)

WHO Performance Status0 0 (0)1 14 (47)2 13 (43)3 3 (10)

Prior cumulative photon RTdose (Gy)

60 (50, 98)

Administration of BNCTTwo BNCT treatments were scheduled to be administered at 3- to

5-week intervals. L-BPA was purchased from Interpharma Praha(Prague, Czech Republic), Ultrafine (Manchester, United Kingdom),or Syntagon (S€odert€alje, Sweden). L-BPAwas complexed with fruc-tose to form L-BPA-fructose (L-BPA-F) to increase solubility, and400mg/kg of L-BPA-Fwas then administered intravenously at a con-centration of 30 g/L over a period of 2 h before neutron irradiation(20). Neutron irradiations were given at the Finnish BNCT facility(Otaniemi, Espoo, Finland) by use of the FiR 1 250-kW TrigaMark II nuclear research reactor (General Atomics, San Diego,CA) with beam properties well-suited for clinical BNCT (22). Neu-tron irradiation was given from 2 portals, each with a median beamtime of 18.6 minutes. Blood samples for boron concentration mea-surements were collected periodically and analyzed by use of induc-tively coupled plasma–atomic emission spectrometry (20, 23).

Concomitant chemotherapyduring photon RT

10 (33)

Time from prior photon RT toBNCT (mo)

16 (3,177)

Tumor characteristicsSite

Oral cavity 11 (37)Nasopharynx 9 (30)Nose/nasal cavity 3 (10)Oropharynx 2 (7)Other* 5 (17)y

Histologic typeSquamous cell carcinoma 24 (80)Adenoid cystic carcinoma 4 (13)Otherz 2 (7)

Tumor diameter (cm) 4.4 (1.5, 8.4)

Other therapyWe administered 10 mg of cetirizine hydrocloride orally prior to

L-BPA-F infusion and 10 to 15 mg/d of dexamethasone after com-pletion of L-BPA-F infusion to alleviate radiation-associatededema. The dexamethasone dose was tapered down within 3 to 4weeks. All patients were hospitalized for a few days after BNCTto ensure safe recovery from BNCT and to manage acute toxicity.None of the patients received systemic cancer therapy during or

immediately after BNCT. Nine patients received 1 to 7 cycles(median, 3) of cancer chemotherapy after cancer had progressedfollowing BNCT. The remaining patients (n = 21) who had diseaseprogression received best supportive care, which included a thirdBNCT treatment in 1 case.

Fig. 1. Patient accrual. BNCT = boron neutron capture therapy.

Evaluation of response and adverse effectsThe patients were followed up at 4- to 12-week intervals after

BNCT, and CT or MRI was performed at 1, 3, 6, and 12 monthsafter neutron irradiation. Adverse effects were evaluated accordingto the National Cancer Institute common terminology criteria ver-sion 3.0, and treatment response was evaluated by use of the RE-CIST (Response Evaluation Criteria in Solid Tumors) (24).Theminimum duration of stabilized disease was defined as 3 monthsand was calculated from the date of the first BNCT.

T category at recurrencerT0 3 (10)x

rT1 2 (7)rT2 7 (23)rT3 4 (13)rT4 14 (47)

N category at recurrencerN0 22 (73)rN1 0 (0)rN2 8 (27)rN3 0 (0)

Abbreviations:WHO = World Health Organization; RT = radia-tion therapy; BNCT = boron neutron capture therapy; T category =tumor category of TNM classification; N category = node categoryof TNM classification.* Other sites consisted of lacrimal gland (n = 1), parotid gland/

adjacent skin (n = 1), larynx (n = 1), and site unspecified (n = 2).y The percentages do not add up to 100% because of rounding.z Includes 1 myofibroblastic sarcoma and 1 transitional cell car-

cinoma.x All 3 patients had rN2 neck disease.

Table 2. Treatment administered

ParameterMedian (minimum,

maximum)

Time between BNCT treatments (d) 41 (32, 92)Average weighted GTV dose [(Gy) (W)]

First BNCT 23 (14, 37)Second BNCT 22 (15, 30)

Average weighted PTV dose [(Gy) (W)]First BNCT 22 (15, 33)Second BNCT 21 (15, 28)

Average tumor physical dose (Gy)First BNCT 7 (5, 11)Second BNCT 6 (5, 9)

Maximum mucosa physical dose (Gy)First BNCT 5 (3, 6)Second BNCT 5 (3, 6)

Abbreviations: BNCT = boron neutron capture therapy; GTV =gross tumor volume; PTV = planning target volume.

e70 I. J. Radiation Oncology d Biology d Physics Volume 82, Number 1, 2012

Statistical analysisThe initial target number of accrued patients was 12, but because

the treatment administered turned out to be relatively well-toleratedand 10 (83%) of the first 12 patients responded to BNCT (16), thestudy was amended in December 2005 to allow a maximum of 30participants. Assuming a response rate of 0.83 (83%), 30 patientswould allow determination of the response rate with a 95% confi-dence interval ranging from 0.70 to 0.96.Efficacy analyses were based on the intention-to-treat principle.

The safety population included patients who received at least 1 doseof L-BPA-F and 1 neutron irradiation. A Kaplan-Meier estimate forsurvival was calculated from the date of the first BNCT to the dateof death, and for PFS, it was calculated to the date of cancerprogression either locally or distally or to death whenever death oc-curred before progression. Patients alive at the time of the analysiscutoff date (December 31, 2009) were censored. Locoregional re-currence–free survival was calculated from the date of the firstBNCT to the date of local progression or to death, whichever oc-curred first, and patients whowere alive without locoregional recur-rence were censored. Survival between groups was compared byuse of the log-rank test. All p values are 2-sided.

RESULTS

Patient accrual and characteristicsThirty patients were entered into the study between De-

cember 1, 2003, and September 23, 2008. During this timeperiod, 25 further patients with inoperable, locoregionally

Fig. 2. (A) Recurrent adenoid cystic carcinoma of lacrimal glanMRI views. (B) Partial response after BNCT in transverse (B1

recurred head-and-neck carcinoma were treated withBNCToutside of the study (21 [84%] of whom were ineligi-ble for the trial) (Fig. 1). Of the study patients, 4 (13%) re-ceived BNCT only once because of cancer progression (n= 1), oliguria (n = 1), or patient request (n = 2). The mediantime interval between the two treatments was 41 days (range,32–91 days).

d before BNCT shown in transverse (A1) and frontal (A2)) and frontal (B2) views. The arrows point at the tumor.

BNCT for head-and-neck cancer d L. KANKAANRANTA et al. e71

The patient and tumor characteristics are shown in Table1. The median time from prior photon irradiation to thedate of first BNCT treatment was 16 months (range, 3–177months), and the median cumulative photon radiotherapydose was 60 Gy (range, 50–98 Gy). Of the recurrent tumors,22 (73%) were staged as rT3, rT4, or rN2 disease; 7as rT2N0M0 cancer; and 1 nasopharyngeal cancer asrT1N0M0 disease. The mean of the longest tumor diametersbefore BNCTwas 4.4 cm (range, 1.5–8.4 cm), the mean sizeof the GTV at irradiation planning was 99 cm3 (range,13–517 cm3), and the mean size of the initial PTV was257 cm3 (range, 88–987 cm3). Fifteen patients underwent18F-fluoro-L-BPA PET before BNCT. The tumor-to-background ratios of the standardized uptake values in the13 fully evaluable cases ranged from 2.5 to 9.0 (median,4.1), suggesting approximately 4-fold BPA accumulationin the tumors compared with the corresponding normaltissues.

Radiation dose deliveredThe median average blood boron concentration was 19.6

mg/g (range, 13.0–26.5 mg/g) at the time of irradiation of thefirst portal and 16.4 mg/g (range, 10.9–23.0 mg/g) when thesecond portal was irradiated. The median calculated averagetumor dose delivered to the GTV during the first scheduledBNCT was 23 Gy (W) [range, 14–37 Gy (W)], and during

Fig. 3. Recurrent squamous cell carcinoma of jaw. (A) Before Bat the tumor.

the second BNCT, it was 22 Gy (W) [range, 15–30 Gy(W)] (Table 2). A mean of 91% (SD, 1.7%) of the tumordoses resulted from the boron neutron capture reactions,5% (1.3%) from gamma dose, 3% (0.5%) from nitrogen neu-tron capture reactions, and 1% (0.3%) from fast neutrondose. The highest normal tissue doses accumulated in thebuccal mucosa, the maxillary sinus mucosa, and the oropha-ryngeal mucosa and were, on average, 11 Gy (W), 9 Gy (W),and 9 Gy (W), respectively, as calculated per 1 BNCT.

Treatment efficacyOne patient who died 2months after BNCTwas not evalu-

able for response. Of the 29 evaluable patients, 13 (45%;95% confidence interval [CI], 27–63%) achieved a completeclinical response and 9 (31%; 95% CI, 14–48%) achieveda partial response, resulting in an objective response rateof 76% (22 of 29; 95% CI, 60–92%) (Figs. 2 and 3). Ofthe remaining 7 patients, 6 had disease stabilization fora median of 8.5 months (range, 5.1–20.3 months) and 1(3%) progressed. Tumor pretreatment 18F-fluoro-L-BPAuptake could be compared with the post-treatment uptakemeasured before the second BNCT in 7 cases; uptakedecreased in 6 (86%; median decrease, 58%; range,30–100%) and remained unaltered in 1.

The median follow-up time of the study participants whowere alive was 31 months (range, 24–57 months). Three

NCT. (B) Partial response after BNCT. The arrows point

Fig. 4. (A) Locoregional failure-free survival, (B) progression-freesurvival, and (C) overall survival after BNCT. The patients cen-sored are indicated with a bar. Two- and four-year survival figuresare provided. CI = confidence interval.

e72 I. J. Radiation Oncology d Biology d Physics Volume 82, Number 1, 2012

patients were alive without cancer with follow-up times of25, 34, and 57 months. Most (n = 20 [74%]) of the remain-ing 27 patients had first cancer recurrence or persistence ata locoregional site, 3 were first diagnosed with distant me-tastases, 2 had local and distant metastases diagnosed at thesame time, and 2 died from an intercurrent cause (myocar-

dial infarction in 1 and pneumonia in 1) while free fromcancer.

The median time of locoregional local control was 7.9months (95%CI, 3.5–12.3 months), and the 2- and 4-year lo-cal control rates were 27% (95% CI, 13–41%) and 16%(95% CI, 0–33%), respectively (Fig. 4). The median PFStime was 7.5 months (95% CI, 5.4–9.6 months), and 20%(95% CI, 13–27%) and 8% (95% CI, 0–18%) of the patientswere alive without cancer progression 2 and 4 years afterBNCT, respectively. Of the 30 patients, 4 (13%) were aliveat the time of analysis. The median overall survival timewas 13.0 months (95% CI, 8.2–17.8 months), and the 2-and 4-year survival rates were 30% (95% CI, 16–44%) and18% (95% CI, 4–33%), respectively. PFS or overall survivaldid not differ between patients diagnosed with recurred na-sopharyngeal cancer (n = 9) and the rest of the patients(p = 0.89 and p = 0.78, respectively).

BNCT resulted in subjective improvement of cancer-related symptoms in a substantial proportion of patients.Of the 23 patients who had tumor pain at the time ofBNCT, 13 (57%) reported relief of symptoms, in additionto 9 (60%) of the 15 patients who had dysphagia, 6 (86%)of the 7 patients who had diplopia, and 5 (31%) of the 16 pa-tients who had trismus.

Treatment safetyGrade 3 oral mucositis was recorded in 16 (53%) of the 30

patients, Grade 3 oral pain in 16 (53%), and Grade 3 fatiguein 10 (33%) (Table 3). One patient was diagnosed withGrade 4 pneumonia within the first 3 months from the treat-ment. The median weight was 61 kg (range, 43–91 kg) at thetime of the first BNCT and 61 kg (range, 45–89 kg) at thesecond BNCT.

The most common late adverse effects were fatigue (pres-ent in 12 [71%] of the 17 evaluable patients) and xerostomia(10 [67%] of 15 patients) (Table 4). Of the 15 evaluable pa-tients, 3 (20%) were diagnosed with osteoradionecrosis,each classified as Grade 3 in severity. One patient hadpost-traumatic life-threatening soft-tissue necrosis (Grade4), where the prior therapies likely contributed. Pneumoniawas frequent in this patient population (diagnosed in 6[35%] of 17 evaluable patients).

DISCUSSION

Many patients diagnosed with inoperable, irradiated, lo-cally recurred head-and-neck cancer have no overt distantmetastases at the time of cancer recurrence and might stillbe cured if an effective salvage therapy were available.Our results suggest that the majority of such patients respondto L-BPA–based BNCT, and most of the remaining patientsachieve disease stabilization lasting for several months. The4-year locoregional recurrence–free survival rate was 16%,indicating that some of the responses were durable.

These efficacy results obtained in a heavily pretreated pa-tient population may not be inferior to those achieved withfull-dose conventional reirradiation administered with or

Table 4. Late* adverse effects recorded

Adverse effect

Severity of adverseeffect [n (%)] Not

evaluable(n)None Grade 1 or 2 Grade 3

Fatigue 5 (29) 9 (53) 3 (18) 13Xerostomia 5 (33) 7 (47) 3 (20) 15Dysgeusia 7 (44) 9 (56) 0 (0) 14Trismus 8 (50) 8 (50) 0 (0) 14Dysphagia 9 (56) 4 (25) 3 (19) 14Cataract 9 (56) 4 (25) 3 (19) 14Alopecia 10 (59) 7 (41) 0 (0) 13Fibrosis/induration 11 (65) 6 (35) 0 (0) 13Pneumonia 11 (65) 0 (0) 6 (35) 13Oral mucositis 11 (69) 5 (31) 0 (0) 14Dry eyes 11 (69) 2 (13) 3 (19)y 14Tumor pain 12 (71) 5 (29) 0 (0) 13Fibrosis–deep tissue 12 (71) 5 (29) 0 (0) 13Pharyngeal mucositis 11 (73) 4 (27) 0 (0) 15Oral pain 13 (76) 4 (24) 0 (0) 13Myositis 13 (76) 4 (24) 0 (0) 13Osteoradionecrosis 12 (80) 0 (0) 3 (20) 15Otitis 13 (81) 3 (19) 0 (0) 14Vomiting 13 (81) 2 (13) 1 (6) 14Periodontal disease 13 (81) 2 (13) 1 (6) 14Dermatitis 14 (82) 3 (18) 0 (0) 13Ulceration 14 (88) 1 (6) 1 (6) 14Nausea 14 (88) 2 (13) 0 (0)y 14Impaired hearing 14 (88) 2 (13) 0 (0)y 14Ocular surface disease 14 (88) 2 (13) 0 (0)y 14Soft-tissue necrosis 14 (88) 1 (6) 1 (6)z 14Keratitis 14 (88) 2 (13) 0 (0)y 14Septicemia 15 (88) 0 (0) 2 (12)x 13Burn 15 (88) 2 (12) 0 (0) 13Diplopia 15 (94) 0 (0) 1 (6) 14

* Adverse effects recorded after the first 3 months of follow-upas calculated from the date of the first BNCT and at least 1 monthafter the second BNCT.

y The percentages do not add up to 100% because of rounding.z One post-traumatic soft-tissue necrosis that may be BNCT re-

lated was classified as Grade 4 toxicity.x One septicemia was Grade 3 in severity, and one was Grade 4.

Table 3. Acute* adverse effects recorded

Adverse effect

Severity of adverse effect [n (%)]

None Grade 1 or 2 Grade 3

Oral mucositis 2 (7) 12 (40) 16 (53)Alopecia 3 (10) 27 (90) 0 (0)Fatigue 7 (23) 13 (43) 10 (33)y

Oral pain 8 (27) 6 (20) 16 (53)Pharyngeal mucositis 10 (33) 15 (50) 5 (17)Dysphagia 14 (47) 11 (37) 5 (17)y

Dysgeusia 15 (50) 15 (50) 0 (0)Trismus 16 (53) 11 (37) 3 (10)Periodontal disease 17 (57) 13 (43) 0 (0)Xerostomia 19 (63) 10 (33) 1 (3)y

Nausea 21 (70) 6 (20) 3 (10)Ocular surface disease 21 (70) 9 (30) 0 (0)Tumor pain 22 (73) 3 (10) 5 (17)Myositis 22 (73) 7 (23) 1 (3)y

Vomiting 23 (77) 7 (23) 0 (0)Fibrosis/induration 24 (80) 6 (20) 0 (0)Otitis 24 (80) 6 (20) 0 (0)Dry eyes 26 (87) 4 (13) 0 (0)Pneumonia 26 (87) 1 (3) 3 (10)Keratitis 27 (90) 3 (10) 0 (0)Osteoradionecrosis 27 (90) 2 (7) 1 (3)Burn 27 (90) 3 (10) 0 (0)Ulceration 27 (90) 2 (7) 1 (3)Dermatitis 28 (93) 2 (7) 0 (0)Diplopia 28 (93) 1 (3) 1 (3)y

Impaired hearing 28 (93) 2 (7) 0 (0)Soft-tissue necrosis 28 (93) 1 (3) 1 (3)y

Septicemia 29 (97) 0 (0) 1 (3)z

Fibrosis–deep tissue 29 (97) 1 (3) 0 (0)Cataract 29 (97) 1 (3) 0 (0)Oliguriax 29 (97) 0 (0) 1 (3)

* Adverse effects recorded within the first 3 months of follow-upas calculated from the date of the first BNCTor within 1month afterthe second BNCT.

y The percentages do not add up to 100% because of rounding.z A single patient had one Grade 3 and one Grade 4 septicemia.x Considered to be related to L-BPA infusion.

BNCT for head-and-neck cancer d L. KANKAANRANTA et al. e73

without chemotherapy, but comparisons between studies aredifficult, notably because of patient selection. We treateda substantial proportion of the patients considered forBNCT outside of the trial while the study was open for ac-crual, because not all patients fulfilled the study inclusioncriteria. The results obtained with BNCT among the 25 pa-tients treated outside of the study are not the topic of this re-port, but they appeared to be in line with the trial results. Ofthe 25 patients, 2 (8%) died within 1 month after BNCT, but3 (16%) of the 19 patients whose follow-up time exceeds 2years are alive without cancer recurrence (data not shown)and 1 (5%) patient has survived more than 5 years withoutcancer recurrence.

Most cancer recurrences after BNCTwere local. Systemicchemotherapy or targeted therapy administered concomi-tantly with radiotherapy for head-and-neck cancer increasesefficacy and improves survival (25, 26), and such therapymight also improve the efficacy of BNCT. We have,therefore, initiated a prospective clinical trial wherecetuximab, a monoclonal antibody directed at the

epidermal growth factor receptor, is given immediatelyafter BNCT (ClinicalTrials.gov identifier NCT00927147).Other potential approaches that might improve the efficacyof BNCT include sequential administration of BNCT withintensity-modulated radiotherapy, concomitant administra-tion of platinum compounds, intra-arterial or prolonged in-fusion of the boron carrier (27), and novel boron carriercompounds (9). BNCT warrants further study also amonga patient population with unresectable head-and-neck canceras first-line treatment, perhaps as administered before con-ventional chemoradiation therapy.

BNCT-related adverse effects resembled those of conven-tional radiotherapy. As reirradiation using photons, reirradia-tion using BNCT was associated with substantial toxicity.The life-threatening toxicities consisted of 1 post-traumaticsoft-tissue necrosis and 3 cases of Grade 3 osteoradionecro-sis. In addition, 2 patients had life-threatening pneumonia,which may have been treatment related. We administeredBNCT twice and after prior conventional radiotherapy,

e74 I. J. Radiation Oncology d Biology d Physics Volume 82, Number 1, 2012

which may have increased adverse effects. Yet, toxicity ofBNCT might be moderate compared with reirradiation. Forexample, in one report 16.5% of the patients treated with re-irradiation plus chemotherapy died of treatment-related tox-icity, 11% required surgery for osteoradionecrosis of themandible, and 57% were dependent on gastrostomy tubefeedings during follow-up (7). Similarly, a randomized studythat compared salvage surgery followed by reirradiation plusconcomitant chemotherapywith salvage surgery alone foundreirradiation to be associated with significantly improved lo-coregional control and disease-free survival whereas overallsurvival did not improve, which was in part because of morefrequent treatment-related deaths and second primary can-cers in the radiotherapy arm (6). However, little is knownabout long-term toxicity of BNCT.

No randomized clinical trials addressing BNCT have beencarried out, and the prior Phase II trials have addressedBNCT in the treatment of glioblastoma (20, 28–32).Because sufficient numbers of epithermal neutrons for

BNCT can at present only be obtained from a nuclearreactor, availability of BNCT is currently limited to a fewcenters worldwide. Compact accelerator-based neutron gen-erators that might be compatible with commercial produc-tion and installment in hospital environments are underdevelopment and may allow more widespread use ofBNCT in the future.

We conclude that BNCT is effective in the treatment of lo-cally recurred, inoperable and previously irradiated head-and-neck cancer and that its tolerability can be regarded asacceptable. Most patients either respond to treatment orhave disease stabilization for several months. Althoughsome responses are durable, cancer progression is commonand usually occurs first at a locoregional site. Comparisonof BNCT to reirradiation with conventional radiotherapyin a randomized study is warranted. Potential approachesto improve BNCT efficacy include administration ofBNCT with systemic cancer therapy or in sequence withconventional radiotherapy.

REFERENCES

1. Wong LY, Wei WI, Lam LK, et al. Salvage of recurrent headand neck squamous cell carcinoma after primary curative sur-gery. Head Neck 2003;25:953–959.

2. Gilbert H, Kagan AR. Recurrence patterns in squamous cellcarcinoma of the oral cavity, pharynx, and larynx. J Surg Oncol1974;6:357–380.

3. Tan HK, Giger R, Auperin A, et al. Salvage surgery after con-comitant chemoradiation in head and neck squamous cell car-cinomas—Stratification for postsalvage survival. Head Neck2010;32:139–147.

4. Tanvetyanon T, Padhya T,McCaffrey J, et al. Prognostic factorsfor survival after salvage reirradiation of head and neck cancer.J Clin Oncol 2009;27:1983–1991.

5. Spencer SA, Harris J, Wheeler RH, et al. RTOG 96-10: Reirra-diation with concurrent hydroxyurea and 5-fluorouracil in pa-tients with squamous cell cancer of the head and neck. Int JRadiat Oncol Biol Phys 2001;51:1299–1304.

6. Janot F, de Raucourt D, Benhamou E, et al. Randomized trial ofpostoperative reirradiation combined with chemotherapy aftersalvage surgery compared with salvage surgery alone in headand neck carcinoma. J Clin Oncol 2008;26:5518–5523.

7. Salama JK, Vokes EE, Chmura SJ, et al. Long-term outcome ofconcurrent chemotherapy and reirradiation for recurrent andsecond primary head-and-neck squamous cell carcinoma. IntJ Radiat Oncol Biol Phys 2006;64:382–391.

8. De Crevoisier R, Bourhis J, Domenge C, et al. Full-dose reirra-diation for unresectable head and neck carcinoma: Experienceat the Gustave-Roussy Institute in a series of 169 patients. JClin Oncol 1998;16:3556–3562.

9. Barth RF, Coderre JA, Vicente GH, et al. Boron neutron capturetherapy of cancer: Current status and future prospects. ClinCancer Res 2005;11:3987–4002.

10. Wittig A, Collette L, Appelman K, et al. EORTC trial 11001:Distribution of two 10B-compounds in patients with squamouscell carcinoma of head and neck, a translational research/phase1 trial. J Cell Mol Med 2009;13:1653–1665.

11. Kato I, Fujita Y, Maruhashi A, et al. Effectiveness of boronneutron capture therapy for recurrent head and neck malignan-cies. Appl Radiat Isot 2009;67(Suppl.):S37–S42.

12. Kimura Y, Ariyoshi Y, Shimahara M, et al. Boron neutron cap-ture therapy for recurrent oral cancer and metastasis of cervicallymph node. Appl Radiat Isot 2009;67(Suppl.):S47–S49.

13. Fuwa N, Suzuki M, Sakurai Y, et al. Treatment results of boronneutron capture therapy using intra-arterial administration ofboron compounds for recurrent head and neck cancer. Br J Ra-diol 2008;81:749–752.

14. Kouri M, Kankaanranta L, Sepp€al€a T, et al. Undifferentiated si-nonasal carcinoma may respond to single-fraction boron neu-tron capture therapy. Radiother Oncol 2004;72:83–85.

15. Aihara T, Hiratsuka J, Morita N, et al. First clinical case of bo-ron neutron capture therapy for head and neck malignant.Headand Neck 2006;28:850–855.

16. Kankaanranta L, Sepp€al€a T, Koivunoro H, et al. Boron neutroncapture therapy in the treatment of locally recurred head andneck cancer. Int J Radiat Oncol Biol Phys 2007;69:475–482.

17. Havu-Auren K, Kiiski J, Lehti€o K, et al. Uptake of 4-borono-2-[(18)F]fluoro-L-phenylalanine in sporadic and neurofibroma-tosis 2-related schwannoma and meningioma studied withPET. Eur J Nucl Med Mol Imaging 2007;34:87–92.

18. Nigg DW. Computational dosimetry and treatment planningconsiderations for neutron capture therapy. J Neurooncol2003;62:75–86.

19. International Commission on Radiation Units & Measure-ments. ICRU report 46: Photon, electron, proton, and neutroninteraction data for body tissues. Bethesda, MD: InternationalCommission on Radiation Units & Measurements; 1992.

20. Joensuu H, Kankaanranta L, Sepp€al€a T, et al. Boron neutroncapture therapy of brain tumors: Clinical trials at the Finnishfacility using boronophenylalanine. J Neurooncol 2003;62:123–134.

21. Coderre JA, Morris GM. Review: The radiation biology of bo-ron neutron capture therapy. Rad Res 1999;151:1–18.

22. Wheeler FJ, Nigg DW, Capala J, et al. Boron neutron capturetherapy (BNCT): Implications of neutron beam and boroncompound characteristics. Med Phys 1999;26:1237–1244.

23. Ryyn€anen PM, Kortesniemi M, Coderre JA, et al. Models forestimation of the (10)B concentration after BPA-fructose com-plex infusion in patients during epithermal neutron irradiationin BNCT. Int J Radiat Oncol Biol Phys 2000;48:1145–1154.

24. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelinesto evaluate the response to treatment in solid tumors. EuropeanOrganization for Research and Treatment of Cancer, NationalCancer Institute of the United States, National Cancer Instituteof Canada. J Natl Cancer Inst 2000;92:205–216.

BNCT for head-and-neck cancer d L. KANKAANRANTA et al. e75

25. Pignon JP, le Maıtre A, Maillard E, et al. Meta-analysis of che-motherapy in head and neck cancer (MACH-NC): An updateon 93 randomised trials and 17,346 patients. Radiother Oncol2009;92:4–14.

26. Vermorken JB, Mesia R, Rivera F, et al. Platinum-based che-motherapy plus cetuximab in head and neck cancer. N Engl JMed 2008;359:1116–1127.

27. SuzukiM, SakuraiY,NagataK, et al. Impact of intra-arterial ad-ministration of boron compounds on dose-volume histogramsin boron neutron capture therapy for recurrent head-and-necktumors. Int J Radiat Oncol Biol Phys 2006;66:1523–1527.

28. Nakagawa Y, Pooh K, Kobayashi T, et al. Clinical review of theJapanese experience with boron neutron capture therapy anda proposed strategy using epithermal neutron beams. J Neuro-oncol 2003;62:87–99.

29. Diaz AZ. Assessment of the results from the phase I/II boronneutron capture therapy trials at the Brookhaven National Lab-oratory from a clinician’s point of view. J Neurooncol 2003;62:101–109.

30. Busse PM, Harling OK, Palmer MR, et al. A critical examina-tion of the results from the Harvard-MIT NCT program phase Iclinical trial of neutron capture therapy for intracranial disease.J Neurooncol 2003;62:111–121.

31. Henriksson R, Capala J, Michanek A, et al. Boron neutron cap-ture therapy (BNCT) for glioblastoma multiforme: A phase IIstudy evaluating a prolonged high-dose of boronophenylala-nine (BPA). Radiother Oncol 2008;88:183–191.

32. Miyatake S, Kawabata S, Yokoyama K, et al. Survival benefitof boron neutron capture therapy for recurrent malignant glio-mas. J Neurooncol 2009;91:199–206.