CLINICAL INVESTIGATIONHead and Neck
INTRAOPERATIVE RADIATION THERAPY FOR RECURRENT
HEAD-AND-NECK CANCER: THE UCSF EXPERIENCE
ALLEN M. CHEN, M.D.,* M. KARA BUCCI, M.D.,* MARK I. SINGER, M.D.,†JOAQUIN GARCIA, M.D.,‡
MICHAEL J. KAPLAN, M.D.,§ALBERT S. CHAN, C.M.D., R.T.T.,* AND THEODORE L. PHILLIPS, M.D.*
Departments of *Radiation Oncology,†Otolaryngology, Head and Neck Surgery, and‡Pathology, University of California,
San Francisco, Comprehensive Cancer Center, San Francisco, CA;§Department of Otolaryngology, Head and Neck Surgery,
Stanford University School of Medicine, Stanford, CA
Purpose: To review a single-institutional experience with the use of intraoperative radiation therapy (IORT) for
recurrent head-and-neck cancer.
Methods and Materials: Between 1991 and 2004, 137 patients were treated with gross total resection and IORT
for recurrence or persistence of locoregional cancer of the head and neck. One hundred and thirteen patients
(83%) had previously received external beam radiation as a component of definitive therapy. Ninety-four patients
(69%) had squamous cell histology. Final surgical margins were microscopically positive in 56 patients (41%).
IORT was delivered using either a modified linear accelerator or a mobile electron unit and was administered
as a single fraction to a median dose of 15 Gy (range, 10–18 Gy). Median follow-up among surviving patients was
41 months (range, 3–122 months).
Results: The 1-year, 2-year, and 3-year estimates of in-field control after salvage surgery and IORT were 70%,
64%, and 61%, respectively. Positive margins at the time of IORT predicted for in-field failure (p ? 0.001). The
3-year rates of locoregional control, distant metastasis-free survival, and overall survival were 51%, 46%, and
36%, respectively. There were no perioperative fatalities. Complications included wound infection (4 patients),
orocutaneous fistula (2 patients), flap necrosis (1 patient), trismus (1 patient), and neuropathy (1 patient).
Conclusions: Intraoperative RT results in effective disease control with acceptable toxicity and should be
considered for selected patients with recurrent or persistent cancers of the head and neck.© 2007 Elsevier Inc.
Intraoperative radiation, Head and neck, Cancer.
Despite progress in surgical and reconstructive techniques
as well as advances in radiation therapy delivery methods, a
significant proportion of patients with head-and-neck cancer
develop locoregional recurrence or experience disease per-
sistence after completion of definitive therapy (1–4). Al-
though the competing risk of developing distant metastasis
is fairly high, up to 60% of this population will die as a
direct consequence of uncontrolled tumor growth at the
primary site (5). In addition, locoregional recurrences can
potentially impact patient quality of life, not only serving as
a source of functional impairment, but also as a morbid
reminder of disease presence. Nevertheless, the optimal
treatment for locally recurrent or persistent cancer of the
head and neck remains to be defined, largely because of the
heterogeneity of this population with respect to disease-
related and patient-related factors such as previous treat-
ment, site of tumor recurrence, disease extent, and perfor-
mance status. Although surgical resection generally forms
the mainstay of treatment, effective salvage therapy is often
precluded by anatomic inaccessibility, prior radiation ther-
apy, and the risk of postsurgical complications (6). Indeed,
published rates of successful surgical salvage range from
15% to 65% (7–9). Understandably, most physicians view
the opportunity for salvage as a balance between the modest
prospects of disease cure and preservation of quality of life.
In an attempt to improve the rate of successful salvage,
we have been performing intraoperative radiation therapy
(IORT) at the University of California, San Francisco
(UCSF) in conjunction with salvage surgery for locally
recurrent and persistent cancer of the head and neck since
1991. The advantage of IORT in this particular setting is
that normal tissue surrounding the tumor bed can be man-
ually retracted and directly shielded. This visualization al-
lows treatment with radiation even after prior delivery of
full courses of conventional external beam radiation ther-
apy. In addition, there are purported radiobiologic advan-
tages, which may result in enhanced tumor control (10). The
Reprint requests to: Allen M. Chen, M.D., Coastal Radiation
Oncology, Inc., 100 Casa St., Suite C, San Luis Obispo, CA
93405. E-mail: email@example.com
Theodore L. Phillips, M.D., is a member of the board of direc-
tors for Intraop Corporation, Santa Clara, CA.
Conflict of interest: none.
Received June 15, 2006, and in revised form Aug 16, 2006.
Accepted for publication Aug 17, 2006.
Int. J. Radiation Oncology Biol. Phys., Vol. 67, No. 1, pp. 122–129, 2007
Copyright © 2007 Elsevier Inc.
Printed in the USA. All rights reserved
0360-3016/07/$–see front matter
purpose of this analysis is to review a long-term, single-
institution experience with the use of IORT focusing on
clinical outcome and safety.
METHODS AND MATERIALS
Between March 1991 and December 2004, 137 patients were
treated with IORT at the time of surgery for head-and-neck cancer
at UCSF. The following patients were excluded from the analysis:
23 patients treated with IORT as a component of primary surgery
for cancer at the time of initial diagnosis, 8 patients treated with
IORT for second or third locoregional recurrence, 4 patients with
gross residual disease after subtotal resection, 3 patients who
received less than the planned dose because of technical failure,
and 2 patients with metastatic disease at the time of IORT. The
remaining 137 patients who had gross total resection and IORT at
the time of salvage surgery for histologically diagnosed recurrent
or persistent head-and-neck cancer formed the primary population
of this study. All patients had been previously treated with defin-
itive therapy for their initial disease consisting of surgical resection
with or without adjuvant therapy in 108 patients (79%) and exter-
nal beam radiation therapy with or without chemotherapy in an
additional 29 patients (21%). A total of 113 patients (83%) had
previously undergone a full-course of external beam radiation
therapy either in the definitive (31 patients) or postoperative (82
patients) setting. Among these patients, conventional fractionation
with once-daily treatment was used in 89 patients (79%) and
hyperfractionation with twice-daily treatment in 24 patients (21%).
No patient had previously received interstitial radiation. The me-
dian total external beam radiation dose previously administered to
the IORT site was 64 Gy (range, 45–72 Gy). Median age was 57
(range, 16–93 years). Ninety-two patients (67%) were male. Ra-
cial distribution was as follows: 72 Caucasian (53%), 30 Asian
(22%), 19 Hispanic (14%), and 16 African American (12%).
The median duration from the date of primary surgery or the last
date of definitive radiation therapy for initial disease to salvage
surgery and IORT was 13 months (range 4–107 months) with 12
patients (9%) undergoing salvage therapy more than 5 years from
initial treatment. Preoperative evaluation before salvage therapy
included patient history, physical examination, and basic blood
work in all patients. Axial imaging with computed tomography of
the head and neck was performed in all patients as part of their
workup for recurrent or persistent disease. One hundred and twelve
patients (82%) had additional imaging with magnetic resonance
imaging. Positron emission tomography was performed in 22
patients (16%). Informed consent for IORT was obtained in all
patients at the time of consultation in the Department of Radiation
Oncology before surgery.
The technique used for delivery of IORT depended on the time
of salvage treatment. Between March 1991 and December 1997,
45 patients (33%) were treated using a modified linear accelerator
(Clinac 20, Varian Radiation Division, Palo Alto, CA) located in
the Department of Radiation Oncology with various electron en-
ergies (6, 9, 12, 15, and 18 MeV) by means of a variety of cones
made of 3-mm thick methyl methacrylate (Lucite, Plexiglas) with
internal diameters ranging from 2.5 to 9.5 cm. Patients treated with
this unit were transported from the operating room under general
anesthesia and positioned on the radiation therapy couch immedi-
ately before the delivery of IORT. In January 1997, a specialized
electron machine with mobile capability (Mobetron, Intraop, Santa
Clara, CA) was installed in a dedicated operating room suite at
UCSF. Beginning in December 1997, all patients receiving IORT
were treated using this machine with various electron energies (4,
6, 9, and 12 MeV) by means of stainless steel cones with internal
diameters ranging from 3 to 10 cm. Some had beveled ends,
enabling adequate apposition of the cone to sloping surfaces in the
surgical bed. Ninety-two patients (67%) were treated with the
Mobetron electron IORT unit for recurrent or persistent head-and-
Maximal resection of all visible disease was attempted before
delivery of IORT to the tumor bed through the surgical incision
site. The operative procedure varied depending on the discretion of
the surgeon and the type of previous surgery, with consideration
given to the extensiveness, histology, and location of the recurrent
tumor. Frozen sections were routinely performed intraoperatively
to ensure maximum clearance of tumor. The area at highest risk for
tumor recurrence was then delineated by the surgeon and radiation
oncologist. The appropriate applicator and cone, selected at the
discretion of the treating radiation oncologist, were then manually
positioned over the target area and attached to the surgical table
using a Bookwalter clamp. Gauze packing, retractors, or pliable
1–2 mm thick lead shields were used to displace or protect normal
tissues near or inside the treatment field. The applicator was
aligned to the linear accelerator by moving the table under the
guidance of a laser docking system. Gantry rotation was performed
to minimize air gaps and to reduce obliquity. Before delivery of
IORT, the treatment field was suctioned to prevent any accumu-
lated fluid or blood products from acting as a bolus. All attempts
were made to isolate or shield critical nerves and vessels from the
In general, the treatment cone was chosen so that it encom-
passed the target volume with a 1–2 cm margin of normal
tissue. Bolus was routinely used to ensure a full surface dose.
The entire target volume was encompassed with a single IORT
field in 110 (80%) patients. The remaining 27 patients had more
than one IORT field treated. A total of 191 fields were treated
in 137 patients. The median cone size was 5 cm (range, 2.5–10
cm). The IORT dose was prescribed to the median 90% isodose
line (range, 80–100%) to encompass the area at risk for micro-
scopic disease. Electron energy was chosen on the basis of the
thickness of the tumor bed taking into consideration the pres-
ence of underlying critical structures. The most commonly
prescribed energy was 6 MeV, which was used in 71% of the
treated fields. The median IORT dose was 15 Gy (range, 10–18
Gy) and selected at the discretion of the treating radiation
oncologist. No definitive policy existed at UCSF regarding the
choice of IORT dose during the span of this analysis, but 91%
were treated to a dose of 15 Gy. Thirty-five patients, 11 of
whom had radiation as part of their previous definitive treat-
ment, proceeded to receive additional post-IORT external beam
radiation therapy to a median dose of 54 Gy (range, 40–71 Gy)
using conventional once-daily fractionation. For these patients,
the median total radiation dose delivered from the time of
salvage surgery was 69 Gy (range, 55–84 Gy). Ninety-nine
patients (72%) subsequently received chemotherapy after sal-
vage surgery and IORT.
Follow-up and statistical analysis
After the completion of surgery, patients were hospitalized for
postsurgical monitoring. The median hospitalization time was 6
123 IORT for recurrent head-and-neck cancer ● A. M. CHEN et al.
days (range, 1–11 days). Patients were asked to return to the clinic
at 3-month intervals for 2 to 3 years and then yearly. Follow-up
imaging was dependent on the treating physician, but in general
was performed every 3 months. Patient follow-up was reported to
the date last seen in clinic or to the date of death. In some cases,
referring physicians were contacted to obtain information regard-
ing patient health status. The median follow-up duration was 18
months for all patients (range, 1–122 months) and 41 months
among surviving patients (range, 3–122 months).
The endpoints analyzed were overall survival, IORT in-field
control, locoregional control, and distant metastasis-free survival.
All events were measured from the date of salvage surgery and
IORT for the locally recurrent or persistent disease. Subsequent
in-field recurrence was defined as radiographic, clinical, or patho-
logic evidence of recurrent disease in the salvaged site treated with
IORT. All other events were considered out-of-field. Regional
failure was recorded separately if there was evidence of an enlarg-
ing cervical or supraclavicular mass distinct from the primary site.
For patients who received IORT to two or more sites at the time of
salvage surgery, an in-field recurrence was recorded if disease
failure occurred at either site. All events were scored until the
patient’s death. The incidence and severity of complications were
assessed at each follow-up visit. Actuarial rates of overall survival,
IORT in-field control, locoregional control, and distant metastasis-
free survival were calculated by the Kaplan-Meier method, with
comparisons among groups performed with two-sided log–rank
tests (11). All p values reported were two-sided with p ? 0.05 used
to denote statistical significance.
All patients had clinical, radiographic, or pathologic ev-
idence of recurrent or persistent disease at the original
tumor site or regional lymph nodes before salvage surgery
and IORT. Eighty-seven patients (64%) were treated with
IORT for isolated local failures, 39 patients (28%) for
isolated regional failures, and 11 patients (8%) for both
local and regional failures. Primary tumor sites were: 24%
oropharynx; 23% oral cavity; 12% paranasal sinus; 12%
parotid gland; 7% hypopharynx; 6% submandibular gland;
6% skin; 3% nasopharynx/nasal cavity; 3% larynx; 2% ear;
and 2% unknown. The most common histologic subtype
was squamous cell carcinoma, occurring in 94 patients
(69%). Other subtypes included adenoid cystic carcinoma
(18 patients), mucoepidermoid carcinoma (9 patients), ade-
nocarcinoma (7 patients), sarcoma (4 patients), melanoma
(3 patients), and acinic cell carcinoma (2 patients).
Table 1 outlines the pathologic findings of the 137 pa-
tients who underwent gross total resection at the time of
salvage surgery and were treated with IORT for locally
recurrent or persistent cancers of the head and neck. The
median size of the recurrent tumor at the time of pathologic
analysis was 2.2 cm (range, 0.0–9.5 cm). Three patients had
no evidence of tumor in the resection specimen at the time
of IORT, despite having positive biopsies immediately be-
fore salvage surgery. In all other patients, histology was
consistent with the known primary tumor. Final surgical
margins were microscopically positive in 56 patients (41%)
and negative in 81 patients (59%). Perineural invasion was
present in the pathology specimen at salvage surgery for 41
patients (30%). Thirty-four patients (25%) and 16 patients
(12%) had pathologic evidence of muscle and bony involve-
Sixty-two patients were alive at the time of this analysis.
Overall survival for the entire patient population at 1, 2, and
3 years was 68%, 52%, and 36%, respectively. Patients who
received IORT for recurrent or persistent disease at the
primary site had significantly better survival than those who
were treated with IORT for disease involving the neck. As
illustrated in Fig. 1, the 3-year overall survival was 44% for
patients treated at the primary site only compared with 19%
for those treated at the neck (p ? 0.001). Median survival
was 20 months (range, 2–122 months) and 12 months
(range, 1–51 months), respectively. None of the other clin-
ical or disease characteristics analyzed, including age at
IORT (?50 vs. ?50 years, p ? 0.34), initial treatment
modality (surgery vs. radiation therapy, p ? 0.74), time
interval from primary treatment to IORT (?12 months vs.
?12 months, p ? 0.11), histology (squamous vs. nonsqua-
mous cell carcinoma, p ? 0.09), pathologic tumor size at
salvage (?2 cm vs. ?2 cm, p ? 0.14), margin status (p ?
0.17), perineural invasion (p ? 0.23), muscle involvement
(p ? 0.42), bony invasion (p ? 0.21), use of external beam
radiation after IORT (p ? 0.84), and IORT delivery method
Table 1. Pathologic findings at the time of salvage surgery
Tumor size (cm)
Abbreviations: IORT ? intraoperative radiation therapy; cm ?
124 I. J. Radiation Oncology ● Biology ● Physics Volume 67, Number 1, 2007
(modified linear accelerator vs. mobile electron unit, p ?
0.55) was predictive of overall survival.
IORT in-field control
Thirty-five patients experienced a subsequent recurrence
at the operative site after completion of salvage surgery and
IORT, 19 of which were isolated first events. Twenty-three
of the in-field recurrences (66%) were biopsy-proven, with
the remaining 12 recurrences (34%) classified based on
clinical or radiographic findings. The median time to second
recurrence after IORT was 6 months (range, 1–54 months),
with 28 in-field recurrences occurring within 1 year from
the time of IORT. For the entire patient population, the
1-year, 2-year, and 3-year estimates of in-field control were
76%, 69%, and 67%, respectively. In-field control after
salvage surgery and IORT is illustrated in Fig. 2.
The only parameter predictive of in-field recurrence was
positive microscopic margins at the time of salvage surgery
and IORT. As depicted in Fig. 3, the 1-year, 2-year, and
3-year rates of IORT in-field control were 87%, 82%, and
82%, respectively, for patients treated with negative surgi-
cal margins compared to 65%, 53%, and 48%, respectively,
for those with positive margins (p ? 0.002). None of the
remaining parameters analyzed, including age at diagnosis,
initial treatment modality, time from primary treatment to
IORT, site of IORT, histology, pathologic tumor size, per-
ineural invasion, muscle involvement, bony invasion, and
use of post-IORT external beam radiation was predictive of
in-field control (p ? 0.05, for all). The 3-year rate of in-field
control for patients treated with the linear accelerator was
64% compared with 69% for patients who underwent IORT
with the mobile electron unit (p ? 0.39).
Among the 87 patients who underwent IORT to disease at
the primary site, 15 patients developed regional recurrences
at out-of-field sites at a median of 10 months (range, 3–74
months) after completion of salvage therapy. Nine of these
were isolated recurrences, and the remaining 6 recurrences
occurred concurrently or subsequent to the development of
distant metastases. Among the 50 patients who received
IORT to disease sites in the neck, 15 experienced a subse-
quent second recurrence at the treated site at a median of 3
months (range, 1–20 months) after salvage therapy. Six of
these were isolated recurrences, and two occurred concur-
rently or subsequent to the development of disease failure at
the primary site. The remaining seven recurrences occurred
concurrently or subsequent to the development of distant
metastases. As illustrated in Fig. 4, the 1-year, 2-year, and
3-year estimates of locoregional control for the entire pa-
tient population were 68%, 61%, and 51%, respectively.
Fig. 1. Overall survival for all patients treated with intraoperative
radiation therapy according to treatment site (primary vs. neck).
Fig. 2. Intraoperative radiation therapy in-field control for the
entire patient population.
Fig. 3. Intraoperative radiation therapy (IORT) in-field control for
all patients treated with IORT according to surgical margin status.
125 IORT for recurrent head-and-neck cancer ● A. M. CHEN et al.
As shown in Fig. 5, the 1-year, 2-year, and 3-year distant
metastasis-free survival rates were 78%, 67%, and 46%.
Patients treated with IORT to the primary tumor site had a
3-year distant metastasis-free survival of 61% compared
with 30% for those who were treated with IORT to disease
sites in the neck (p ? 0.001). None of the other clinical or
disease characteristics analyzed, including age, initial treat-
ment modality, time from primary treatment to IORT, his-
tology, pathologic tumor size, margin status, perineural
invasion, muscle involvement, bony invasion, use of post-
IORT external beam radiation, and IORT delivery method
was predictive of distant metastasis-free survival (p ? 0.05,
There were no perioperative fatalities. Four patients had
superficial wound infections without tissue breakdown re-
quiring intravenous antibiotics during the immediate post-
operative period. Two patients, both of whom received
IORT doses of 15 Gy for locally recurrent cancers of the
oral cavity, developed orocutaneous fistulas at approxi-
mately 1 and 3 months after salvage surgery that required
subsequent surgical reconstruction. Both of these patients
were previously treated for primary squamous cell carci-
noma involving the hard palate with surgery and postoper-
ative radiation therapy to doses of 54 Gy and 60 Gy,
respectively. Neither received post-IORT radiation therapy.
Another patient, treated with an IORT dose of 15 Gy for
recurrent squamous cell carcinoma of the base of tongue
after previously undergoing surgery with postoperative ra-
diation therapy to 60 Gy, experienced wound dehiscence
with flap necrosis requiring surgical debridement. An addi-
tional patient, treated with an IORT dose of 15 Gy for
recurrent mucoepidermoid carcinoma of the parotid gland
initially managed with surgery alone, developed trismus 4
months after completion of IORT that resolved over the
course of 1 year with occupational therapy. Last, 1 patient
treated with an IORT dose of 15 Gy for adenoid cystic
carcinoma of the submandibular gland, initially managed
with surgery and postoperative radiation to a dose of 64 Gy,
developed sudden-onset facial pain secondary to neuropathy
at approximately 1 month after salvage treatment. This
patient’s symptom was managed medically with narcotics
and gradually resolved over the course of 2 years. No
patient developed clinical evidence of osteoradionecrosis,
bone fracture, brain necrosis, or carotid artery hemorrhage.
Prognosis after recurrence or persistence of locoregional
disease after completion of definitive therapy for head-and-
neck cancer is relatively poor, but varies based on such
factors as histology, tumor site, disease extent, and perfor-
mance status (12–15). Depending on the specific selection
criteria used, the proportion of patients eligible for salvage
surgery after experiencing recurrent or persistent locore-
gional disease ranges from 34% to 75% (16–18). However,
even among patients with seemingly resectable lesions,
surgery can be technically challenging due to the difficulties
of operating in previously manipulated or irradiated tissue
as well as the proximity of tumor to critical structures such
as the carotid artery, skull base, esophagus, and trachea.
Moreover, the low cure rates associated with surgical sal-
vage have led some authors to question whether the modest
benefits outweigh the potentially increased morbidity (19,
20). Similarly, salvage therapy using reirradiation is con-
troversial. In a large study of 169 patients, De Crevoisier
demonstrated that full-dose reirradiation of head-and-neck
sites resulted in unacceptably high rates of late toxicity,
including mucosal necrosis in 21%, osteoradionecrosis in
Fig. 4. Locoregional control for the entire patient population.
Fig. 5. Distant metastasis-free survival for the entire patient pop-
126 I. J. Radiation Oncology ● Biology ● Physics Volume 67, Number 1, 2007
8%, and 5 deaths from carotid hemorrhage (21). Although
the results of more recent studies analyzing reirradiation
using altered fractionation and brachytherapy have been some-
what more promising, the role of reirradiation after previous
full-course radiation remains investigational (22–25).
The primary advantage of IORT is that it can deliver a
high dose of radiation therapy to microscopic disease while
minimizing the risk of reirradiation to dose-limiting normal
tissue. By enhancing local control without increasing tox-
icity, this approach may theoretically lead to an improve-
ment in the therapeutic ratio. Encouraging results from
Phase III trials involving the pancreas, stomach, and retro-
peritoneum have led investigators to hypothesize that this
technique may be of considerable utility for other sites
where tumor infiltration may make complete resection dif-
ficult, such as the head and neck (26–28). This present
series demonstrates that although long-term survival is rel-
atively limited for patients who present in this setting,
effective disease control can be obtained with this multimo-
dality approach in appropriately selected patients.
The results of our analysis compare favorably with those
from other single-institutional experiences. Rate et al. dem-
onstrated a 2-year local control rate of 61% among 47
patients treated with IORT for recurrent head-and-neck can-
cer (29). However, Nag et al. reported a 2-year local control
rate of only 13% among 38 patients treated with IORT for
recurrent head-and-neck cancer (30). Most recently, Pin-
heiro et al. reported 5-year in-field control rates of 41% and
52% for 34 patients with squamous cell carcinoma and 10
with nonsquamous cell carcinoma of the head and neck,
respectively, treated with IORT at the Mayo Clinic (31).
Patient selection criteria and heterogeneity with respect to
disease characteristics are most likely the major factors
responsible for the wide variation in published rates of
disease control. For instance, patients with laryngeal can-
cers are more amenable to successful salvage therapy com-
pared with tumors arising from other sites, and patients
presenting with neck recurrences generally have a worse
prognosis than those with recurrences at the primary site
(18). Similarly, those who present with a first local recur-
rence have a better prognosis than those who are treated for
a second or third recurrence. Imbalances in clinical param-
eters such as performance status, weight loss, and the ex-
tensiveness of presalvage workup between series could also
lead to conflicting results between series.
Differences in therapeutic approaches and technique may
have also contributed to the observed discrepancies in out-
come. Studies reporting poorer local control often included
patients with gross residual disease after surgery or who
were treated with palliative intent (29–32). It is particularly
notable that the vast majority of centers reporting on their
experiences with IORT for recurrent head-and-neck cancer
only considered its use for “high-risk” patients with positive
or close margins after salvage surgery. Indeed, the propor-
tion of patients treated with positive margins was unspeci-
fied in some series and was as high as 100% in others
(29–33). This is in contrast to the present series in which
only 41% of the patients were treated with positive micro-
scopic margins and none with gross disease. Moreover,
there are no studies, to our knowledge, that have analyzed
the effect of IORT dose on local control, although some
authors have speculated that higher doses may be required
in the presence of microscopic residual disease (30). Simi-
larly, some authors suggest that the addition of adjuvant
external beam radiation therapy or chemotherapy after sal-
vage surgery and IORT in selected patients may enhance
local control, although we were unable to detect any statis-
tically significant benefit in the relatively small number of
patients treated in this manner in our series (10). Last,
variations in the methods of reporting clinical outcomes
with IORT must be considered in interpreting results and
drawing conclusions. Although most reports have defined
local control as in-field control within the IORT port, others
included both failures within and outside the IORT port in
determining local control (29–34).
A limitation of this report was that as a result of its
retrospective nature, we were unable to differentiate be-
tween true recurrent cancers and new primary tumors of the
head and neck among the patients undergoing salvage ther-
apy. This may have been relevant because patients present-
ing with the latter tend to have improved prognosis com-
pared with those with the former. Based on the concept of
“field cancerization” originally proposed by Slaughter et al.
in 1953, it is well known that patients who experience and
survive head-and-neck cancer are at an elevated risk for
developing a second primary tumor in the same anatomic
field during the ensuing years (35). That the overwhelming
majority of patients (91%) in the present series were treated
with salvage surgery and IORT more than 5 years after
initial treatment for their primary head-and-neck cancers
suggest, however, that the incidence of second primary
tumors in our population was low.
Although several studies have established that the role of
IORT in the setting of gross residual disease is extremely
limited because of the exceedingly high rates of local re-
currence, the significance of microscopically positive mar-
gins has been somewhat controversial, both for the treat-
ment of head-and-neck cancers and for disease at other sites
(10). Similar to other series, we were able to demonstrate a
statistically significant difference in local control between
patients treated with negative and positive microscopic mar-
gins with IORT. Toita et al. also reported significantly
improved in-field control rates of 82% and 55% for close
and positive margins, respectively, which translated into a
significantly improved overall survival (36). Garrett et al.
treated 28 patients with advanced or recurrent squamous
cell carcinoma and reported local control rates of 87% for
close margins and 75% for microscopic margins (37). Most
recently, Schleicher et al. showed that rates of local recur-
rence were 60% and 50% for patients treated with micro-
scopically positive and negative margins, respectively (38).
Although patients with positive microscopic margins but
without gross disease after salvage resection can be success-
fully treated with IORT, our findings strongly suggest that
127 IORT for recurrent head-and-neck cancer ● A. M. CHEN et al.
negative margins should be attained to ensure favorable
The relatively low rate of observed complications with
long follow-up is particularly reassuring. In contrast, Haller
et al. demonstrated a major complication rate of 16%,
including wound infections, cerebrospinal fluid leak, and
fistula, among 53 patients who underwent IORT in conjunc-
tion with surgery for advanced head-and-neck cancer (39).
Similarly, Toita reported a 2-year complication rate of
32.8% among 25 patients treated with IORT including 5
patients with osteoradionecrosis and 3 with carotid artery
hemorrhage (36). The variation in observed complication
rates can likely be accounted for by differences in IORT
dose. Studies reporting high incidence of complications
routinely used higher doses compared with the 15 Gy used
at our institution. For instance, Freeman et al. reported 3
cases of osteoradionecrosis in 3 patients who received IORT
doses of 100 Gy, 50 Gy, and 20 Gy, respectively (34).
Because of the limited IORT dose range employed in the
present series, we were unable to evaluate the effect of dose
on clinical outcome or complications. Differences in pre-
scription points may have also contributed to observed
discrepancies in complications. Some groups routinely pre-
scribe to maximum depth without accounting for individual
patient anatomy. It is worth noting that at our institution,
IORT is sometimes aborted if patient anatomy prevents
delivery of a uniform dose to tumor bed. Furthermore, it is
important to recognize that all of the observed complica-
tions cannot be clearly attributed to IORT because the
surgical complication rate for salvage surgery is high, es-
pecially in previously irradiated tissue, even when IORT is
Although randomized data are lacking and comparisons
of observed outcomes after salvage surgery with or without
IORT for head-and-neck cancer among retrospective series
are limited by the heterogeneity of patient populations, our
results compare favorably to historical controls. Goodwin
published a meta-analysis of 12 studies and reported a
5-year overall survival rate of 36% for patients treated with
surgery alone for locally recurrent head-and-neck cancer
(18). However, eligibility criteria for surgical resection for
recurrent or persistent disease vary between institutions, and
it is important to note that reported outcomes are con-
founded by selection criteria inherent in these studies. At
our institution, treatment decisions are typically made with
multidisciplinary input, considering disease control, quality
of life, and cosmetic factors. Even so, surgical criteria and
operative approaches lack uniformity and continue to vary
based on physician preferences. In conclusion, our results,
spanning 15 years, demonstrate that although overall sur-
vival after recurrence or persistence of locoregional head-
and-neck cancer after definitive therapy is relatively limited,
effective disease control, with acceptable side effects, can
be achieved in appropriately selected patients using salvage
surgery and IORT.
1. Vokes EE, Weichselbaum RR, Lippman EM, et al. Head and
neck cancer. N Engl J Med 1993;328:184–194.
2. Taneja C, Allen H, Koness RJ, et al. Changing patterns of
failure of head and neck cancer. Arch Otolaryngol Head Neck
3. Fu KK, Pajak TF, Trotti A, et al. A Radiation Therapy
Oncology Group (RTOG) phase III randomized study to com-
pare hyperfractionation and two variants of accelerated frac-
tionation to standard fractionation radiotherapy for head and
neck squamous cell carcinomas: First report of RTOG 9003.
Int J Radiat Oncol Biol Phys 2000;48:7–16.
4. Calais G, Alfonsi M, Bardet E, et al. Randomized trial of
radiation therapy versus concomitant chemotherapy and radi-
ation therapy for advanced-stage oropharynx carcinoma. J Nat
Cancer Inst 1999;91:2081–2086.
5. Honk WK, Bromer RH, Amato DA, et al. Patterns of relapse
in locally advanced head and neck cancer patients who
achieved complete remission after combined modality treat-
ment. Cancer 1985;45:1242–1245.
6. Ridge JA. Squamous cancer of the head and neck: Surgical
treatment of local and regional recurrences. Semin Oncol
7. Schwartz GJ, Mehta RH, Wenig BL, et al. Salvage treatment
for recurrent squamous cell carcinoma of the oral cavity. Head
8. Gleich LL, Ryzenman J, Gluckman JL, et al. Recurrent ad-
vanced (T3 or T4) head and neck squamous cell carcinoma.
Arch Otolaryngol Head Neck Surg 2004;130:35–38.
9. Jones KR, Lodge-Rigal RD, Reddick RL, et al. Prognostic
factors in the recurrence of stage I and II squamous cell cancer
of the oral cavity. Arch Otolaryngol Head Neck Surg
10. Carter YM, Jablons DM, DuBois JB, et al. Intraoperative
radiation therapy in the multimodality approach to upper aero-
digestive tract cancer. Surg Oncol Clin N Am 2003;12:1043–
11. Kaplan EL, Meier P. Nonparametric estimation from incom-
plete observations. J Am Stat Assoc 1958;53:547–581.
12. Davidson J, Keane T, Brown D, et al. Surgical salvage after
radiotherapy for advanced laryngopharyngeal carcinoma. Arch
Otolaryngol Head Neck Surg 1997;123:420–424.
13. Marcial VA, Hanley JA, Ydrach A, et al. Tolerance of surgery
after radical radiotherapy of carcinoma of the oropharynx.
14. Parsons JT, Mendenhall WM, Stringer SP, et al. Salvage
surgery following radiation failure in squamous cell carci-
noma of the supraglottic larynx. Int J Radiat Oncol Biol Phys
15. Argiris A, Li Y, Forastiere A. Prognostic factors and long-
term survivorship in patients with recurrent or metastatic
carcinoma of the head and neck. Cancer 2004;101:2222–
16. Taussky D, Dulguerov P, Allal AS. Salvage surgery after
radical accelerated radiotherapy with concomitant boost tech-
nique for head and neck carcinomas. Head Neck 2005;27:182–
17. Wong LY, Wei WI, Lam LK, et al. Salvage of recurrent head
and neck squamous cell carcinoma after primary curative
surgery. Head Neck 2003;25:953–959.
18. Goodwin WJ. Salvage surgery for patients with recurrent
squamous cell carcinoma of the upper aerodigestive tract:
128I. J. Radiation Oncology ● Biology ● Physics Volume 67, Number 1, 2007
when do the ends justify the means? Laryngoscope 2000;
19. Arnold DJ, Goodwin WJ, Weed DT, et al. Treatment of
recurrent and advanced stage squamous cell carcinoma of the
head and neck. Semin Radiat Oncol 2004;14:190–195.
20. De Crevoisier R, Domenge C, Wibault P, et al. Full-dose
reirradiation for unresectable head and neck carcinoma: Ex-
perience at the Gustave-Roussy Institute in a series of 169
patients. J Clin Oncol 1998;16:3556–3562.
21. Machtay M, Rosenthal DI, Chalian AA, et al. Pilot study of
postoperative reirradiation, chemotherapy, and amifostine af-
ter surgical salvage for recurrent head-and-neck cancer. Int J
Radiat Oncol Biol Phys 2004;59:72–77.
22. Puthawala A, Syed AM, Gamie S, et al. Interstitial low-dose-
rate brachytherapy as a salvage treatment for recurrent head-
and-neck cancers: Long-term results. Int J Radiat Oncol Biol
23. Hepel JT, Syed AM, Puthawala A, et al. Salvage high-dose
rate (HDR) brachytherapy for recurrent head-and-neck cancer.
Int J Radiat Oncol Biol Phys 2005;62:1444–1450.
24. Milano MT, Vokes EE, Salama JK, Stenson KM, et al. Twice-
daily reirradiation for recurrent and second primary head-and-
neck cancer with gemcitabine, paclitaxel, and 5-fluorouracil
chemotherapy. Int J Radiat Oncol Biol Phys 2005;61:1096–
25. Zelefsky MJ, Zimberg SH, Raben A, et al. Brachytherapy for
locally advanced and recurrent lymph node metastases.
J Brachytherapy Int 1998;14:123–130.
26. Sindelar WF, Kinsella WT. Randomized trial of intraoperative
radiotherapy in unresectable carcinoma of the pancreas. Int J
Radiat Oncol Biol Phys 1986;12:148.
27. Sindelar WF, Kinsella TJ, Tepper JE. Randomized trial of
intraoperative radiotherapy in locally advanced gastric cancer.
Proc Am Soc Clin Oncol 1987;6:91.
28. Sindelar WF, Kinsella TJ, Chen PW, et al. Intraoperative
radiotherapy in retroperitoneal soft tissue sarcomas. Final re-
sults of a prospective randomized clinical trial. Arch Surg
29. Rate WR, Garrett P, Hamaker R, et al. Intraoperative radiation
therapy for recurrent head and neck cancer. Cancer 1991;67:
30. Nag S, Schuller DE, Martinez-Monge R, et al. Intraoperative
electron beam radiotherapy for previously irradiated advanced
head and neck malignancies. Int J Radiat Oncol Biol Phys
31. Pinheiro AD, Foote RL, McCaffrey TV, et al. Intraoperative
radiotherapy for head and neck skull base cancer. Head Neck
32. Spaeth J, Andreopoulos D, Unger T, et al. Intraoperative
radiotherapy—5 years of experience in the palliative treatment
of recurrent and advanced head and neck cancers. Oncology
33. Vikram B, Strong EW, Shah JP, et al. Intraoperative radio-
therapy in patients with recurrent head and neck cancer. Am J
34. Freeman SB, Hamaker RC, Singer MI, et al. Intraoperative
radiotherapy of head and neck cancer. Arch Otolaryngol Head
Neck Surg 1990;116:165–168.
35. Slaughter DP, Southwick HW, Smejkal W. “Field canceriza-
tion” in oral stratified squamous epithelium: Clinical implica-
tions of multicentric origin. Cancer 1953;6:963–968.
36. Toita T, Nakano M, Takizawa Y, et al. Intraoperative radia-
tion therapy for head and neck cancer. Int J Radiat Oncol Biol
37. Garrett P, Pugh N, Ross D, et al. Intraoperative radiation
therapy for advanced or recurrent head and neck cancer. Int J
Radiat Oncol Biol Phys 1987;13:785–788.
38. Schleicher UM, Phonias C, Spaeth J, et al. Intraoperative
radiotherapy for pre-irradiated head and neck cancer. Ra-
diother Oncol 2001;58:77–81.
39. Haller JR, Mountain RE, Schuller DE, et al. Mortality and
morbidity with intraoperative radiotherapy for head and neck
cancer. Am J Otolaryngol 1996;17:308–310.
129 IORT for recurrent head-and-neck cancer ● A. M. CHEN et al.