]
Re-irradiation is a potentially curative treatment
option for some patients with unresectable disease.
However, increased risk of severe or life-threatening
treatment-related toxicity and tumor radioresistance
pose challenges to re-irradiation. Because locoregional
progression is the most common cause of death in patients
with HNC, obtaining local control may effect the survival of patients with locoregionally failed disease.
In addition, local tumor progression may affect morbidity
due to disfigurement in appearance, uncontrollable
pain, cancer bleeding, infection, and impairment of
speech and swallowing, thus resulting in a poor quality
of life. Patients with recurrent or second primary HNC
having a history of irradiation comprise a challenging
heterogeneous group. Published data include diverse
recurrent or second primary tumors in the extent and
location, prior radiotherapy (RT) parameters, elapsed
time since prior treatment, and extent and severity of
normal tissue sequelae. Data in current literature on
acute and late normal tissue recovery from prior treatment
are not available. Lack of data pertaining to re-irradiation
tolerance poses significant challenges or even statistileads
to fear pertaining to meeting these patients daily
in the clinic. Recently, high precision RT, including intensity-
modulated radiotherapy (IMRT), has demonstrated
the ability to reduce toxicity and improve disease
control. Novel systemic agents and radiotherapy
techniques, including stereotactic body RT (SBRT) and
proton therapy, are also being actively explored.[]
Salvage surgery is the standard for patients suitable
for surgery, but it can be successfully performed only in
25% of patients. In a Phase II multicenter randomized
trial (Radiation Therapy Oncology Group [RTOG] 99-
01), the outcome of chemotherapy treatment with postsurgical
re-irradiation in recurrent disease was evaluated.
[] Overall, 130 patients who underwent salvage surgery
were randomized to the observation and chemoradiotherapy
arms. Local control and disease-free survival
were increased in the chemoradiotherapy arm.[,]
Chang et al showed that age, Charlson comorbidity
index score, clinical stage at first diagnosis, and recurrence-
free interval were significant independent prognostic
factors for overall survival (OS) of patients with
recurrent head-and-neck squamous cell carcinoma
(HNSCC). Regardless of the recurrence stage or site,
salvage surgery is the recommended procedure. Re-
RT alone and concurrent chemoradiotherapy are more
suitable for inoperable recurrent HNSCCs.[]
The use of SBRT with re-irradiation in recurrent
disease has demonstrated the advantages of better protection
of organs at risk and higher doses in the target
volume. The duration of treatment is shortened in patients
with poor prognosis due to hypofractionation.
Unger et al reported the outcome of 65 patients treated
with median 30 Gy (21-35) SBRT in 2?5 fractions (33
patients received concomitant chemotherapy). Median
OS for all patients was 12 months, and 2-year OS and
locoregional control (LRC) rates were 41% and 30%,
respectively. Complete response, partial response, and
progressive disease rates were 54%, 27% and 20%, respectively.
Higher total dose, surgical resection, and
nasopharynx site were significantly associated with
improved LRC; surgical resection and non-squamous
histology were associated with improved OS.[] In a
retrospective analysis of 46 patients in whom Cengiz
et al performed a median 30 Gy (18?35 Gy) SBRT in
1?5 fractions, the OS at 1 year was 46%. Complete response,
partial response, and stable disease rates were
27%, 30%, and 27%, respectively.[] Treatment responses
were similar to those observed in other studies,
but the late grade 4 toxicity rate was high. Of the 8
patients with late carotid blowout, 7 died due to carotid
hemorrhage.[]
The 1-year OS rate was 47% for a Phase II multicenter
study in which Lartigau et al used concomitant
cetuximab with 36 Gy SBRT in 6 fractions between 11
and 12 days for the treatment of patients with recurrent
or newly diagnosed HNC. Complete response, partial
response, and stable disease rates were 49%, 20%, and
23%, respectively. Grade 3/4 toxicity was seen in 32%
of patients. Only one patient died because of the choice
of patients with the carotid artery wrapped around a
fewer than third of the carotid artery, resulting in a carotid
blowout.[] Seventy patients in the Pittsburgh
series were retrospectively matched and analyzed. Of
these, 35 patients received concurrent cetuximab treatment
with SBRT and the remaining patients received
only SBRT. In the cetuximab group, OS was higher
(24.5 vs. 14.8 months). Grade 3/4 toxicity was not different
between the two groups.[] Patient selection
criteria, differences in tumor histology, radiotherapy
fractionation, and dose differences make it difficult to
compare studies. In conclusion, phase III multicenter
trials are needed to demonstrate whether the 2-year
median survival, particularly Heron et al.?s study, is reproducible.[]
A multi-institutional study validated the recursive
partitioning analysis (RPA) classification for patients
with unresected recurrences treated with SBRT and
compared outcomes of patients with unresected disease
treated with IMRT.[] Authors have analyzed
412 patients from 7 institutions in North America.
RPA identified 3 prognostic subgroups: class I included
patients >2 years from initial RT with resected tumors;
class II included patients >2 years from initial RT with
unresected tumors or those ?2 years without feeding
tube or tracheostomy dependence; and the remaining
patients formed class III. The authors highlighted
the differences between the cohorts, a reflection of the
selection bias inherent in retrospective studies where
treatment is typically selected based on the baseline
characteristics. SBRT-treated patients had a shorter interval
between RT courses (1.2 years vs. 3.1 years) and
were more likely to have prior chemotherapy (64% vs.
46%) than IMRT-treated patients. After adjustment
analysis, differences in OS and locoregional failure
(LRF) between IMRT and SBRT were no longer statistically
significant. In a second attempt to minimize
bias introduced by baseline differences, a subgroup
analysis by RPA class was performed. In the poorprognosis
class III group, the investigators observed
a statistically insignificant difference in 2-year OS
with IMRT than with SBRT (16.2% vs. 3.6%). Among
class II patients, IMRT was associated with a statistically significantly better 2-year OS rate compared with
SBRT (39.1% vs. 18.6%, P < .001). This difference was
attenuated substantially when further stratified by tumor
size and SBRT dose. Specifically, patients with
small tumors (?25 cm3 or rT0-2) treated with ?35 Gy
of SBRT had a 2-year OS rate closer to that of patients
treated with IMRT (38.5% vs. 50%). For patients with
larger tumors treated with SBRT, the OS difference remained
significantly worse than that in those treated
with IMRT irrespective of the SBRT dose ( 28.2% with
IMRT vs. 9.1% with SBRT <35 Gy or 8.8% with SBRT
?35 Gy). The rate of LRF mirrored that of OS, which
is no surprise because LRF is frequently the ultimate
cause of death for recurrent HNC. In terms of toxicity
trade-off, the rate of acute grade 4 or 5 toxicity was
low, but such toxicity was more common with IMRT
(5.1% vs 0.5%), and rates of grade >3 toxicity beyond
90 days were comparable after adjustment for competing
risks of recurrence or death (12.4% with IMRT vs.
11.6% with SBRT). Although the toxicity of IMRT and
SBRT re-irradiation appeared similar, the study demonstrated
differences in outcome between IMRT and
SBRT, particularly among class II patients with larger
tumors, which need to be evaluated in a prospective
setting. Patterns-of-failure analysis after SBRT would
be helpful to determine whether failures are marginal
to the treated gross tumor volume, particularly among
larger tumors that may require a larger clinical target
volume margin. Besides the standard clinical examination
and imaging, other methods are needed to help
clinicians decide when a tumor can be treated with a
minimal margin and when a larger margin is needed
to account for subclinical spread. In addition, dose and
fractionation may play a role in the observed OS and
LRF differences between SBRT and IMRT.[]
Carotid blowout syndrome and spinal cord myelopathy
are rare and are late lethal side effects of
re-irradiation.[] For re-irradiation of the full cord
cross-section at 2 Gy per day after prior conventionally
fractionated treatment, cord tolerance appears
to increase at least by 25% 6 months after the initial
course of RT based on animal and human studies. For
partial cord irradiation as part of spine radiosurgery, a
maximum cord dose of 13 Gy in a single fraction or 20
Gy in 3 fractions appears associated with a <1% risk of
injury.[] Carotid blowout syndrome was observed in
patients who were treated more than once in the study
by Yazici et al. and for whom the maximum carotid artery
dose was below 34 Gy.[] However, Gebhardt et
al. reported no carotid blowout syndrome below 47.6
Gy maximum point dose to carotid artery.[]
