VOLUME 27 䡠 NUMBER 12 䡠 APRIL 20 2009

JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T

Prognostic Factors for Survival After Salvage Reirradiation

From the Head and Neck and Thoracic
Programs, and the Statistic Core, H.
Lee Moffitt Cancer Center and
Research Institute, Tampa, FL.
Submitted September 11, 2008;
accepted December 30, 2008;
published online ahead of print at
www.jco.org on March 16, 2009.
Authors’ disclosures of potential con-
flicts of interest and author contribu-
tions are found at the end of this
article.
Corresponding author: Tawee
Tanvetyanon MD, 12902 Magnolia Dr,
Tampa FL 33612; e-mail: tanvett@
moffitt.org.
The Appendix is included in the
full-text version of this article,
available online at www.jco.org.
It is not included in the PDF version
(via Adobe® Reader®).
© 2009 by American Society of Clinical
Oncology
0732-183X/09/2712-1983/$20.00
DOI: 10.1200/JCO.2008.20.0691
of Head and Neck Cancer
Tawee Tanvetyanon, Tapan Padhya, Judith McCaffrey, Weiwei Zhu, David Boulware, Ronald DeConti,
and Andrea Trotti
A B S T R A C T
Purpose
Patients who develop recurrent or new primary head and neck cancer in a previously irradiated site
have poor prognosis. Reirradiation is a treatment option, although it is associated with substantial
toxicities. We investigated potential prognostic factors, including comorbidity and pre-existing
organ dysfunction, for survival after reirradiation.
Methods
Institutional electronic records of patients treated with reirradiation between January 1998 and
2008 were reviewed. Comorbidity was assessed by Charlson index and Adult Comorbidity
Evaluation-27 (ACE-27) grading. Organ dysfunction was defined as feeding tube dependency,
functioning tracheostomy, or soft tissue defect.
Results
There were 103 patients, including 46 patients who underwent salvage surgery before reirradia-
tion. Median progression-free and overall survivals were 12.1 months (95% CI, 9.7 to 16.6) and
19.3 months (95% CI, 13.9 to 29.9), respectively. Significant comorbidity was present in 36% of
patients by Charlson index and 24% by ACE-27. Baseline organ dysfunction was present in 37%
of patients. Median overall survivals were 5.5 months among those with both organ dysfunction
and comorbidity per Charlson index, and 4.9 months per ACE-27, compared with 59.6 and 44.2
months, respectively, among the patients with neither organ dysfunction nor comorbidity
(P ⬍ .001 and ⬍ .001). Other independent prognostic factors were interval from previous radiation,
recurrent tumor stage, tumor bulk at reirradiation, and reirradiation dose. A nomogram to predict
the probability of death within 24 months after reirradiation was developed (concordance
index ⫽ 0.75).
Conclusion
Comorbidity and pre-existing organ dysfunction are among several important prognostic factors for
patients undergoing reirradiation. For those with both comorbidity and organ dysfunction,
reirradiation largely serves as a palliative therapy.

J Clin Oncol 27:1983-1991. © 2009 by American Society of Clinical Oncology

In recent years, the practice of full-dose, re-

INTRODUCTION
Prognosis is poor when a recurrent or new pri-
mary head and neck cancer develops in an area
previously treated with radiation. In the absence
of distant metastatic disease, salvage surgery pro-
vides a durable disease control in approximately
15% of such patients.1 Unfortunately many have
an unresectable tumor, and those patients with
advanced recurrent stage, positive surgical mar-
gins, or adverse pathologic features still have poor
outcomes after salvage surgery.2,3 For patients
with an unresectable tumor, however, the disease
is uniformly fatal. Palliative chemotherapy is the
standard treatment, providing a typical median
survival of about 6 months.4
peated radiation (reirradiation) to the area has
been reported as feasible.5-9 Theoretically, this
practice may cause serious organ injury due to
high accumulative radiation dose and may be in-
effective for an apparently radiation therapy–
resistant tumor. Nevertheless, with the use of
reirradiation, a durable control of disease has
been reported in approximately 10% of patients
with an unresected tumor and 20% of those with a
resected tumor.5,6 Reirradiation, however, is asso-
ciated with considerable toxicities. In a prospec-
tive phase II study, the cumulative incidence of
serious (acute plus late) toxicities was 80% at 6
months after treatment initiation.7 Treatment-
related death occurred in approximately 8% of

© 2009 by American Society of Clinical Oncology 1983

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Tanvetyanon et al

patients, including carotid hemorrhage which occurred in 2% to

Table 1. Patient, Tumor, and Treatment Characteristics (N ⫽ 103)
5% of patients.5,7-9
Because of these toxicities, there is a strong rationale to refine Characteristic No. %
patient selection to undergo reirradiation. Some disease-related fac- Patient
Male sex 80 77.7
tors, such as greater volume of tumor and shorter time from previous Median age, years 62.3
radiation9,10 confer a poor prognosis, while some treatment-related Range 34.4-87.3
factors, such as the use of intensity-modulated radiation therapy or Previous cancer occurrence
higher dose of reirradiation confer a good prognosis.8,11-12 To date, Once 73 70.9
⬎ once 30 29.1
however, there is very limited information with regard to the use of Median duration since previous radiation, 35.9
patient-related factors as prognostic factors. months
Comorbidity burden affects survival in many cancers including Range 2.0-504.0
Having received chemotherapy during previous 29 28.2
head and neck cancer.13 However, as a competitive cause of death, radiation
comorbidity plays a major role among young patients or those with Median previous radiation dosage, Gy 64
early stages of disease.14 Its impact on survival is generally lessened or Range 10-76.8
Tumor
sometimes totally obliterated for patients with short survival.15 In New primary tumor 33 32.0
addition, for head and neck cancer patients who need salvage therapy, Site of disease
there is another unique type of morbidity— organ dysfunction. More Nasopharynx 4 3.9
or less organ dysfunction, such as feeding tube dependency, tracheos- Oral cavity 26 25.2
Oropharynx 37 35.9
tomy, or soft tissue defect, is often present as a result of disease or Larynx 12 11.7
previous treatment. This impairs quality of life, which in turn influ- Other 24 23.3
ences survival.16 In this article, we report our institutional experience Recurrent T stage
T0 14 13.6
with reirradiation and evaluate multiple potential prognostic factors,
T1 12 11.6
including comorbidity and organ dysfunction, for survival after sal- T2 27 26.2
vage reirradiation. T3 7 6.8
T4 43 41.8
Recurrent N stage
N0 69 67
METHODS N1 12 11.7
N2 17 16.5
N3 5 4.8
Patients
Recurrent stage group
After approval by the institutional review board, electronic medical
I 10 9.7
records of all patients with head and neck cancer who underwent radiation II 24 23.3
between January 1998 and 2008 were reviewed. Patients were included if they III 12 11.7
received reirradiation to the area significantly overlapping with the previously IVA 52 50.5
irradiated area. This database contained five patients whose outcomes were IVB 5 4.8
previously reported7; 13 patients were treated at outside institutions. Histology
Squamous cell carcinoma 98 95.2
Data Collection Other 5 4.8
Staging was determined based on American Joint Committee on Cancer, Tumor grade
sixth edition, before reirradiation.17 Patients with no measurable disease at Well differentiated/low grade 17 16.5
primary or nodal site at recurrence were classified as T0 or N0. Isolated N1 Moderately differentiated/intermediate grade 51 49.5
disease was grouped as stage III; N2, IVA; and N3, IVB. Tumor bulk was a sum Poorly differentiated/high grade 29 28.2
Unknown 6 5.8
of unidimensional measurements of diameter at all disease sites before reirra-
Median tumor bulk at reirradiation, cm 3.0ⴱ
diation and was recorded as zero for patients without measurable disease. New Range 0.5-12.0
primary cancers were those arising from a different anatomic head and neck Treatment
site or occurring later than 5 years of previous disease. Close margins were Having undergone salvage surgery before 46 47
defined as 0.1 mm or smaller. Adverse events and their severity were graded reirradiation
according to National Cancer Institute Common Toxicity Criteria, version Median re-irradiation dose, Gy 60
3.0. To avoid accounting for morbidities resulting from progressive disease or Range .60-75.6
further therapy, adverse events were censored on disease progression. Deaths Having received special radiation technique 53 53.5
Three-dimensional planning 4 4.0
were verified by the National Social Security Death Index. Response evaluation
Intensity modulated radiation therapy 48 48.0
was performed among patients with measurable disease before reirradiation Proton therapy 1 1.0
based on Response Evaluation Criteria in Solid Tumors,18 typically at 7 to 9 Having received concurrent chemotherapy 72 69.9
weeks after completion of reirradiation. Patients who began reirradiation, but with reirradiation
were unable to complete the planned therapy were included in analysis. Single-agent treatment
Cisplatin 40 38.8
Comorbidity and Organ Dysfunction Carboplatin 12 11.7
To understand a possible differential effect of burden versus severity Cetuximab 9 8.7
of comorbidity on survival, comorbidity was assessed by two measures— Taxane-based regimen 10 9.7
Charlson index and Adult Comorbidity Evaluation-27 (ACE-27) grading. Hydroxyurea 1 1.0
Both have shown reliability and validity in head and neck cancer.19,20 Charlson ⴱ
Data from 55 patients with measurable disease.
index measures the burden of comorbidity using a weighted index of 19
clinical conditions. ACE-27 grading measures the severity of comorbidity
based on 26 disease systems. For Charlson index, the definition of moderate

1984 © 2009 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Prognostic Factors for Reirradiation of Head and Neck Cancer

renal failure was a calculated glomerular filtration rate lower than 50 mL;
peptic ulcer disease was a history of bleeding or ulcer diagnosed by endoscopy
with active disease requiring medical treatment; and chronic obstructive lung
disease was a radiographic change plus symptom or need for chronic medical
treatment. Organ dysfunction was defined as feeding tube dependency, func-
tioning tracheostomy, or soft tissue defect including uncovered open
wound of skin or mucosa, fistula, or osteonecrosis. Prophylactic feeding
tube placement before initiation of reirradiation was not considered as or-
gan dysfunction.
Statistical Methods
Correlation between Charlson index and ACE-27 was examined using
Spearman’s coefficient. Overall survival was calculated from the first date of
reirradiation initiation until death or last follow-up using Kaplan-Meier’s
method. Progression-free survival was measured until last follow-up, progres-
sion of disease, or death. Follow-up rule was passive; all patients were censored
on May 15, 2008. The log-rank test was used to examine the difference in
survival between groups. For categoric variables, test for heterogeneity of risk
between groups and trend test were performed.
Multivariable regression analyses were performed using Cox propor-
tional hazards modeling (backward elimination with a significance level to stay
P ⫽ .15), and this model formed the basis for the survival prediction model.
Covariates were selected from clinically relevant survival predictors, including
comorbidity, radiation dose, use of special radiation technique, use of concur-
rent chemotherapy, use of salvage surgery, presence of organ dysfunction,
interval since last radiation, tumor bulk, T stage, stage group, and previous
recurrence number. Two models were examined separately based on Charlson
index or ACE-27. Analyses were performed using SAS statistical software
version 9.1 (SAS Institute, Cary, NC).
To construct a prognostic nomogram, independent predictors of sur-
vival were included. To make the nomogram useful for prospective patients,
reirradiation dose was excluded. Discrimination was evaluated using a concor-
dance index (C index), measuring the probability that, given a pair of ran-
domly selected patients, the model correctly predicted which patient would
experience an event first. The C index can range between 0.5 (random chance)
and 1.0 (perfectly discriminating model). The calibration compared predicted
survival with actual survival and was done by grouping patients according to
predicted survival and then plotting as actual versus predicted survival. All
statistical analyses were performed using R software program (http://www.r-
project.org/).21
RESULTS
Patients, Treatment, and Outcome
There were 103 patients, including 46 patients who had salvage
surgery before reirradiation (Table 1). Isolated neck disease was ob-
served in 14 patients; seven of whom underwent salvage surgery
before reirradiation. The most common indication for postopera-
tive reirradiation was positive or close surgical margins, in 31 pa-
tients (67.4%). Those with a negative surgical margin had at least one
of the following features: extracapsular extension of nodal disease
(three patients), perineural invasion (two patients), multiple lymph
node involvement (four patients), lymphovascular invasion (two pa-
tients), and soft tissue invasion (10 patients). Reirradiation was typi-
cally given as 2 Gy fraction daily using linear accelerator. No significant
change in the radiation technique occurred during the study period. In
five patients, twice daily treatment with hyperfractionation was used.7
Concurrent chemotherapy with cisplatin was typically adminis-
tered as 75 to 100 mg/m2 every 3 weeks, carboplatin as area under
the time-concentration curve was 2, weekly, and cetuximab as 400
mg/m2 loading dose followed by 250 mg/m2 weekly until completion
of reirradiation.
Tumor response was radiographically assessable in 57 pa-
tients. Complete response was observed in 67%; partial response in
11%; stable disease in 4%; and progressive disease in 19%. The
minimal follow-ups for patients censored for progression-free and
overall survival were both 0.8 months. To date, 71 patients have
had progression of disease or died, with a median progression-free
survival of 12.1 months (95% CI, 9.7 to 16.6 months). Deaths have
occurred in 59 patients, resulting in a median overall survival of
19.3 months (95% CI, 13.9 to 29.9 months). At 1 and 2 years, there
were 34 and 53 cumulative deaths, corresponding to survival rates
of 64.8% (95% CI, 54.4 to 73.5%) and 40.0% (95% CI, 29.4 to
50.5), respectively.

Table 2. Frequency of Comorbid Disease Among Patients With Significant Comorbidity (37 patients per Charlson index and 25 patients per ACE-27)
Frequency Frequency

Disease per Charlson Index No. % Disease per ACE-27 (grade ⱖ 2)
Myocardial infarction 9 15
Congestive heart failure 3 5
Peripheral vascular disease 7 12
Cerebrovascular disease 3 5
Dementia 0 0
Hemiplegia 0 0
Chronic obstructive pulmonary disease 13 22
Connective tissue disease 4 6
Ulcer disease 3 5
Mild liver disease 2 3
Diabetes 6 10
Diabetes with end organ damage 2 3
Moderate or severe renal insufficiency 5 8
Leukemia 1 2
Lymphoma 1 2
Active solid tumor 1 2
AIDS 0 0
Total frequency 60 100
No. %
Cardiovascular disease 11 41
Neurological disease 0 0
Psychiatric disease 0 0
Substance abuse problem 3 11
Respiratory disease 4 15
Rheumatologic disease 2 7
Gastrointestinal disease 0 0
Endocrine disease 3 11
Morbid obesity 0 0
Renal disease 1 4
Malignant disease 3 11
Immunologic disease 0 0
27 100

Abbreviation: ACE-27, Adult Comorbidity Evaluation-27.

www.jco.org © 2009 by American Society of Clinical Oncology 1985

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Tanvetyanon et al

Acute toxicities of grade 3 or higher occurred in 45 patients
(43.7%). Of these, gastrointestinal toxicities (mucositis and dyspha-
gia) occurred in 41 patients (39.8%); skin toxicities in three patients
(2.9%); and hematologic toxicity in one patient (1%). Hospitalization
during or near the period of reirradiation was documented in seven
patients (6.8%), mostly due to dehydration or infection. Late toxicities
of grade 3 or higher occurred in 49 patients (47.5%). Of these, gastro-
intestinal toxicity (feeding tube dependent beyond 6 months after
reirradiation or until last follow-up or censored) was documented in
28 patients; trismus in nine patients; laryngeal toxicity in eight pa-
tients; skin toxicity in six patients; mandibular fracture or necrosis in
six patients; and mucocutaneous fistula in five patients. Treatment-
related deaths were verified in two patients (2%), including one pa-
tient with active rheumatoid arthritis who developed edema of face
and airway. One patient developed grade 4 carotid rupture 1 year
after reirradiation.
Comorbidity and Organ Dysfunction
By Charlson index, 37 patients (36%) had significant comorbid-
ity (Charlson index ⱖ 1) and by ACE-27 grading, 25 patients (24%)
had significant comorbidity (ACE-27 ⱖ 2; Tables 2 and 3). We found
a modest, but significant, correlation between the two measurements.
Using Spearman’s correlation coefficient, in which 1.0 denotes per-
fectly positive correlation and ⫺1.0 denotes perfectly negative corre-
lation, the coefficient was 0.6 (P ⬍ .001). Organ dysfunction was
present in 38 patients (36.9%). Feeding tube dependency before reir-
radiation was observed in 25 patients (24.3%). Functioning tracheos-
tomy was present in 16 patients (15.5%); soft tissue defect, fistula, or
osteonecrosis was present in four patients (3.9%). There were seven
patients (6.8%) with multiple organ dysfunctions.
Impact of Comorbidity and Organ Dysfunction
on Survival
Both Charlson and ACE-27 comorbidity indices exhibited an
ability to define prognosis. As frequency or severity of comorbidity
increased, survival progressively decreased. By Charlson index, pa-
tients with no comorbidity had a median overall survival of 22.6
months versus 12.8 months among those with at least one comorbid-
ity (P ⫽ .002); the median progression-free survival was 13.1 versus
10.9 months, respectively (P ⫽ .035). By ACE-27 grading, patients
with no or mild comorbidity had a median overall survival of 21.2
versus 12.8 months among those with comorbidity of moderate or
severe grade (P ⫽ .0005); the median progression-free survival was
14.4 versus 7.0 months, respectively (P ⫽ .005). Organ dysfunction
also had a negative effect on survival. Patients with no organ dysfunc-
tion had a median overall and progression-free survival rates of 30.7
and 16.5 months, compared with 11.3 and 7.3 months among those
with organ dysfunction (P ⫽ .0002 and .0116, respectively).
When classifying patients into four groups based on comorbidity
and organ dysfunction, in the cohort with both comorbidity and
organ dysfunction, no long-term survival was observed. Based on
Charlson comorbidity index, the median overall and progression-free
survival rates of patients with neither comorbidity nor organ dysfunc-
tion (n ⫽ 41) were 59.6 months and 16.5 months, compared with 5.5
months and 4.8 months respectively, among those with both comor-
bidity and organ dysfunction (n ⫽ 13; P ⬍ .001, ⬍ .001; Figs 1A, 1C).
Based on ACE-27, the median overall and progression-free survival
rates of patients with neither organ dysfunction nor significant co-
morbidity (n ⫽ 50) were 44.2 and 18.9 months, compared with 4.9
and 4.9 months, respectively, among those with both organ dysfunc-
tion and significant comorbidity (n ⫽ 10; P ⬍ .001, ⬍ .001; Figs
1B, 1D).
Multivariable Analysis of Prognostic Factors
First, we performed a univariable analysis on relevant variables
that may be prognostic (Table 4). It appeared that stage (especially T
stage), tumor bulk at reirradiation, reirradiation dose, organ dysfunc-
tion, and comorbidity burden were prognostic. For organ dysfunc-
tion, when considering each type separately, only feeding tube
dependency reached the level of statistical significance. When consid-
ering common comorbid diseases separately, renal disease was a sig-
nificant predictor of survival, although in other diseases, negative
survival trends were observed.
Second, we performed a multivariable analysis of nonduplicative
variables that were statistically significant or approaching significance
in the univariable analysis. We found that the time interval between
previous radiation and reirradiation became a significant survival
predictor. Comorbidity (by either Charlson index or ACE-27), radia-
tion dose, organ dysfunction, stage group, T stage, and tumor bulk
before reirradiation remained independent survival predictors. The
use of concurrent chemotherapy was not found to be associated with a
better survival. However, among patients who received chemothera-
py, there were greater proportions of those with organ dysfunction,
recurrent advanced tumor stage, or gross residual tumor than those
who did not receive chemotherapy.
Development of Prognostic Nomogram
Significant prognostic factors were fitted into a nomogram (Fig
2), based on their contribution to the accuracy of prediction. Since
survival at 24 months after salvage treatment for recurrent head and

Table 3. Frequency of Patients Classified by Charlson Index and Adult Comorbidity Evaluation-27 Grade
Charlson Index Total Patients

Adult Comorbidity Evaluation-27 Grade 0 1 2 3 4 5 No. %

0 32 0 0 1 0 0 33 32
1 29 11 2 2 1 0 45 44
2 4 7 5 1 2 0 19 18
3 1 0 2 2 0 1 6 6
Total patients, No. 66 18 9 6 3 1 103
% 64 17 9 6 3 1 100

1986 © 2009 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Prognostic Factors for Reirradiation of Head and Neck Cancer

A B
1.0 Both comorbidity and 1.0 Both comorbidity and
Progression-Free Survival (probability)

Progression-Free Survival (probability)

organ dysfunction organ dysfunction
No comorbidity; with No comorbidity; with
organ dysfunction organ dysfunction
0.8 0.8
With comorbidity; no With comorbidity; no
organ dysfunction organ dysfunction
No comorbidity or No comorbidity or
0.6 organ dysfunction 0.6 organ dysfunction
censored censored

0.4 0.4

0.2 0.2

Fig 1. Survival classified by comorbidity

and organ dysfunction. (A) Impact of
0 20 40 60 80 100 0 20 40 60 80 100 Charlson Index and organ dysfunction on
progression-free survival (P ⬍ .001) (B)
Time (months) Time (months) Impact of Adult Comorbidity Evaluation-27
(ACE-27) grading and organ dysfunction
C D on progression-free survival (P ⬍ .001). (C)
1.0 Both comorbidity and
1.0 Both comorbidity and Impact of Charlson Index and organ dys-
organ dysfunction organ dysfunction function on overall survival (P ⬍ .001). (D)
No comorbidity; with No comorbidity; with Impact of ACE-27 grading and organ dys-
organ dysfunction organ dysfunction function on overall survival (P ⬍ .001).
0verall Survival (probability)

0verall Survival (probability)

0.8 With comorbidity; no 0.8 With comorbidity; no

organ dysfunction organ dysfunction
No comorbidity or No comorbidity or
organ dysfunction organ dysfunction
0.6 0.6
censored censored

0.4 0.4

0.2 0.2

0 20 40 60 80 100 0 20 40 60 80 100

Time (months) Time (months)

neck cancers is often considered a successful outcome of therapy,22 we ment between predicted and observed outcomes (Appendix Figs
designed a nomogram to predict the probability of death within 24 A2A and A2B, online only).
months after reirradiation. The final variables included in our nomo-
gram were comorbidity, organ dysfunction, isolated neck recurrence,
DISCUSSION
tumor bulk before reirradiation, and time interval since last radiation.
For comorbidity information, we chose Charlson index because of its In this single institutional experience, representing one of the largest
simplicity and ease of use (nomogram based on ACE-27 available in reirradiation series, we evaluated multiple potential prognostic fac-
the online-only Appendix Fig A1). T stage did not fit in the model as tors, including comorbidity and organ dysfunction. The observed
categoric variable because patients with T3 had a worse survival than comorbidity burdens were comparable with or slightly lower than
those with T4, perhaps by chance due to small sample. Nonetheless, those typically reported in studies of treatment-naïve patients with
there was a considerable difference in survival among patients with T0 head and neck cancer.13,23 The prevalence of baseline organ dysfunc-
compared with others. The median survival of patients with isolated tion was about one third. We found that comorbidity and organ
neck recurrence (T0) was not reached, compared with 17.3 months dysfunction were independent predictors of survival. If both predic-
among the rest (P ⫽ .013). This information was integrated into the tors were present, no long-term survivors were seen. This information
nomogram. Its C index and bootstrap-corrected C index were 0.75 can be integrated with other disease-related factors to form a predic-
and 0.73, respectively. The calibration curve showed good agree- tive model. Based on the proposed nomogram, for most patients who

www.jco.org © 2009 by American Society of Clinical Oncology 1987

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Tanvetyanon et al

Table 4. Univariable and Multivariable Analyses for Prognostic Factors

Characteristic No. Hazard Ratio 95% CI Univariable P Multivariable P ⴱ
Patient
Sex
Female 23 1.00 Reference .23
Male 80 1.52 0.77 to 2.99
Age
Per year increase 103 0.99 0.97 to 1.02 .52
Interval since last radiation
Per month increase 103 0.99 0.99 to 1.00 .09 .008
Previous recurrence number
Once 73 1.00 Reference .31
⬎ once 30 1.35 0.76 to 2.39
Organ dysfunction
No 65 1.00 Reference .0004 .0005
Yes 38 2.63 1.55 to 4.47
Feeding tube dependence
No 78 1.00 Reference .002
Yes 25 2.41 1.38 to 4.21
Tracheostomy
No 87 1.00 Reference .07
Yes 16 1.82 0.96 to 3.45
Soft tissue defect
No 99 1.00 Reference .15
Yes 4 2.11 0.76 to 5.84
Charlson comorbidity index
Per index increase 103 1.34 1.09 to 1.63 .005 .016
0 66 1.00 Reference .002
1 or more 37 2.29 1.34 to 3.91
ACE-27 comorbidity grade
Per grade increase 103 1.66 1.22 to 2.25 .001 .032
0 or 1 78 1.00 Reference .001
2 or more 25 2.68 1.51 to 4.78
Having diabetes
No 95 1.00 Reference .12
Yes 8 2.11 0.83 to 5.39
Having heart failure or infarction history
No 91 1.00 Reference .08
Yes 12 1.95 0.91 to 4.14
Having peripheral vascular disease
No 96 1.00 Reference .30
Yes 7 0.47 0.12 to 1.95
Having chronic obstructive lung disease
No 90 1.00 Reference .10
Yes 13 1.79 0.89 to 3.59
Having renal insufficiency
No 98 1.00 Reference .03
Yes 5 3.04 1.09 to 8.48
Tumor
Tumor status
Recurrent disease 70 1.00 Reference .64
New primary 33 1.13 0.67 to 1.92
Recurrent T stage
T0 14 1.00 Reference .01 .0007
T1 12 1.82 0.41 to 8.20 .0002 by trend test
T2 27 3.13 0.92 to 10.66
T3 7 6.34 1.39 to 28.78
T4 43 5.53 1.66 to 18.38
Recurrent N stage
N0 69 1.00 Reference .54
N1 12 0.55 0.22 to 1.40 .84 by trend test
N2 17 1.04 0.49 to 2.15
N3 5 1.36 0.48 to 3.81
(continued on following page)

1988 © 2009 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Prognostic Factors for Reirradiation of Head and Neck Cancer

Table 4. Univariable and Multivariable Analyses for Prognostic Factors (continued)

Characteristic No. Hazard Ratio 95% CI Univariable P Multivariable P ⴱ
Recurrent stage group
I 10 1.00 Reference .17
II 24 4.09 0.94 to 17.89 .058 by trend test
III 12 2.51 0.49 to 12.95
IVA 52 4.49 1.07 to 18.82
IVB 5 5.28 0.96 to 28.92
Tumor bulk after salvage surgery
Per cm increase 103 1.12 1.02 to 1.23 .01 ⬍ .0001
Tumor differentiation
Per grade increase 97 0.92 0.75 to 1.13 .43
Treatment
Having had salvage surgery
No 57 1.00 Reference .11 NS
Yes 46 0.65 0.38 to 1.10
Reirradiation dose, Gy
Per Gy increase 101 1.00 1.00 to 1.00 ⬍ .0001 .002
ⱕ 50 38 1.00 Reference .003
⬎ 50 63 0.45 0.26 to 0.76
Special radiation technique
No 46 1.00 Reference .14 NS
Yes 53 0.67 0.39 to 1.14
Concurrent chemotherapy
No 30 1.00 Reference .12 NS
Yes 72 1.61 0.88 to 2.97

Abbreviations: ACE-27, Adult Comorbidity Evaluation-27; NS, not statistically significant.

ⴱ
Only variables analyzed in multivariable proportional hazard models shown.

have both organ dysfunction and comorbidity and who do not have
an isolated neck recurrence, reirradiation will result in a negligible
chance of survival at 2 years.
Several reasons may explain the profound negative effect of co-
morbidity and organ dysfunction on survival. First, both comorbidity
and organ dysfunction may increase the risk of death from cancer-
related cause due to suboptimal therapy or poor treatment tolerance.
For instance, among patients with significant renal impairment, cis-
platin may not be used or be fully tolerated. Second, both risk factors
may also increase the risk of death from other medical causes. For
example, toxicity from reirradiation can exacerbate comorbid disease
or further increase the severity of existing organ dysfunction, resulting
in deaths. Third, organ dysfunction—which is largely a late toxicity of
previous radiation—may also be a marker of aggressive disease biol-
ogy: More aggressive diseases necessitate more extensive previous
treatments, which in turn, result in more organ dysfunctions. Given
the large impact on survival, it appears that comorbidity and organ
dysfunction not only pose as a competing cause of death, but also
interact negatively with reirradiation.
Nevertheless, for patients with poor prognostic factors, in-
cluding those having both comorbidity and organ dysfunction,
reirradiation should still be kept as a palliative option. In this
setting, only a prospective randomized trial can delineate the ben-
efit and risk of reirradiation as opposed to other supportive mea-
sures alone. Indeed, for such patients who are symptomatic from
tumor, treatment options are few (including chemotherapy, best
supportive care, or if chemotherapy has failed, best supportive care
only). For patients with poor prognostic factors, but without
symptomatic or gross disease, we believe that reirradiation should
be deferred because its role is mainly palliative. In contrast, for
patients with good prognostic factors, reirradiation, whenever fea-
sible, should be strongly considered because it has a curative role in
a significant proportion of patients.
Limitations of our study include the inability to quantify other
types of organ dysfunction, such as trismus, xerostomia, and chronic
pain, and to understand whether these are also prognostic. We were
not able to fully determine the degree of overlap between reirradiation
and previous radiation fields, which may also be prognostic. Greater
degree of overlap results in greater chance of exceeding the radiation
tolerance of nearby vital structures, impeding optimal reirradiation.
Finally, our proposed nomogram will need external validation. Its C
index of 0.75 indicates an imperfect discrimination. This predictive
model is based on cases from a tertiary referral centers with long-term
experiences in reirradiation.
Based on our experience, although we did not detect a significant
effect of intensity-modulated radiation therapy (IMRT) on survival,
we recommend IMRT (as oppose to two- or three-dimensional radi-
ation technique) for most patients undergoing reirradiation. IMRT
allows a tight dose distribution, limits injury to adjacent organs, and
may improve reirradiation outcomes.8,24 We also recommend a radi-
ation dose of 50 to 70 Gy, preferably with concurrent systemic therapy.
Several studies have shown that reirradiation dose strongly correlates
with survival,8,9,11 and in first-line setting, concurrent systemic ther-
apy and radiation are more efficacious than radiation alone.12,25 In
addition, advanced radiation techniques , such as tomotherapy or
proton-beam therapy, may facilitate treatment near the base of skull.

www.jco.org © 2009 by American Society of Clinical Oncology 1989

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Tanvetyanon et al

0 10 20 30 40 50 60 70 80 90 100
Points

Present
Comorbidity
Absent

Present
Organ Dysfunction
Absent

No
Isolated Neck Recurrence
Yes

Tumor Bulk (cm)

0 1 2 3 4 5 6 7 8 9 10 11 12

Time Interval (months)

550 500 450 400 350 300 250 200 150 100 50 0

Total Points
0 20 40 60 80 100 120 140 160 180 200 220 240 260

Linear Predictor
-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0 0.5 1.0 1.5 2.0 2.5 3.0

24−Month Survival Probability

0.9 0.8 0.7 0.4 0.2 0.1

Fig 2. Nomogram predicting probability of death within 24 months after re-irradiation. Comorbidity, any comorbid disease based on Charlson index (congestive heart
failure, history of myocardial infarction, peripheral vascular disease, cerebrovascular disease, dementia, hemiplegia, chronic obstructive pulmonary disease, connective
tissue disease, moderate or severe renal insufficiency, diabetes, ulcer disease, liver disease, leukemia, lymphoma, active solid tumor, or AIDS); Organ dysfunction,
feeding tube dependency, functioning tracheostomy, or soft tissue defect (fistula, uncovered wound, osteonecrosis); isolated neck recurrence, tumor present in the
neck without measurable mucosal tumor at salvage surgery (if applicable) or re-irradiation; Tumor bulk, sum of the maximal unidimensional diameter of tumor at
the neck plus mucosal lesion at the time of re-irradiation; Time interval, interval between completion of previous radiation and initiation of re-irradiation. To read the
nomogram, obtain the value of each variable and draw a straight line up until this intersects the line labeled as ⬙points⬙. That value at the point of intersection denotes the
number of points incurred. By repeating this process for each factor, a points score for each variable is obtained and accumulated. Finally, locate the value of the total points
on the horizontal line labeled as ⬙total points⬙ and draw a straight line down, the estimated probability of survival at 24 months is indicated, ranging from 0 to 1.0.

Finally, for treatment near carotid artery, we recommend a Doppler

AUTHOR CONTRIBUTIONS
ultrasound before reirradiation. In our practice, patients with signifi-
cant stenosis are considered for an appropriate vascular intervention Conception and design: Tawee Tanvetyanon, Andrea Trotti
before reirradiation. Financial support: Tawee Tanvetyanon
In summary, several prognostic factors, including comorbidity Administrative support: Tawee Tanvetyanon, Tapan Padhya, Judith
and organ dysfunction, should be considered when making treatment McCaffrey, Ronald DeConti, Andrea Trotti
decision on reirradiation. For patients with favorable risk factors, Provision of study materials or patients: Tawee Tanvetyanon, Tapan
reirradiation should be strongly considered as it offers the promise of Padhya, Judith McCaffrey, Ronald DeConti, Andrea Trotti
Collection and assembly of data: Tawee Tanvetyanon, Tapan Padhya,
long-term survival. For patients with poor risk factors, however, such
Andrea Trotti
promise is unlikely. Data analysis and interpretation: Tawee Tanvetyanon, Weiwei Zhu,
David Boulware, Ronald DeConti, Andrea Trotti
Manuscript writing: Tawee Tanvetyanon, Tapan Padhya, Weiwei Zhu,
AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS David Boulware, Ronald DeConti, Andrea Trotti
OF INTEREST Final approval of manuscript: Tawee Tanvetyanon, Tapan Padhya,
Judith McCaffrey, Weiwei Zhu, David Boulware, Ronald DeConti,
The author(s) indicated no potential conflicts of interest. Andrea Trotti

REFERENCES
1. Gilbert H, Kagan AR: Recurrence patterns in
squamous cell carcinoma of the oral cavity, pharynx,
and larynx. J Surg Oncol 6:357-380, 1974
2. Goodwin WJ Jr: Salvage surgery for patients
with recurrent squamous cell carcinoma of the up-
per aerodigestive tract: When do the ends justify the
means? Laryngoscope 110:1-18, 2000 (suppl 93)
3. Jones AS, Bin Hanafi Z, Nadapalan V, et al: Do
positive resection margins after ablative surgery for
head and neck cancer adversely affect prognosis? A
study of 352 patients with recurrent carcinoma
following radiotherapy treated by salvage surgery.
Br J Cancer 74:128-132, 1996
4. Forastiere AA, Metch B, Schuller DE, et al:
Randomized comparison of cisplatin plus fluoroura-
cil and carboplatin plus fluorouracil versus metho-
trexate in advanced squamous-cell carcinoma of the
head and neck: A Southwest Oncology Group study.
J Clin Oncol 10:1245-1251, 1992

1990 © 2009 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222

Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
 Prognostic Factors for Reirradiation of Head and Neck Cancer
5. Spencer SA, Harris J, Wheeler RH, et al:
RTOG 96-10: reirradiation with concurrent hydroxyu-
rea and 5-fluorouracil in patients with squamous cell
cancer of the head and neck. Int J Radiat Oncol Biol
Phys 51:1299-1304, 2001
6. Janot F, de Raucourt D, Benhamou E, et al:
Randomized trial of postoperative reirradiation com-
bined with chemotherapy after salvage surgery
compared with salvage surgery alone in head and
neck carcinoma. J Clin Oncol 26:5518-5523, 2008
7. Langer CJ, Harris J, Horwitz EM, et al: Phase
II study of low-dose paclitaxel and cisplatin in com-
bination with split-course concomitant twice-daily
reirradiation in recurrent squamous cell carcinoma of
the head and neck: Results of Radiation Therapy
Oncology Group Protocol 9911. J Clin Oncol 25:
4800-4805, 2007
8. Salama JK, Vokes EE, Chmura SJ, et al:
Long-term outcome of concurrent chemotherapy
and reirradiation for recurrent and second primary
head-and-neck squamous cell carcinoma. Int J Ra-
diat Oncol Biol Phys 64:382-391, 2006
9. De Crevoisier R, Bourhis J, Domenge C, et al:
Full-dose reirradiation for unresectable head and
neck carcinoma: Experience at the Gustave-Roussy
Institute in a series of 169 patients. J Clin Oncol
16:3556-3562, 1998
10. Hwang JM, Fu KK, Phillips TL: Results and
prognostic factors in the retreatment of locally re-
current nasopharyngeal carcinoma. Int J Radiat On-
col Biol Phys 41:1099-1111, 1998
www.jco.org
Downloaded from ascopubs.org by 139.47.23.222 on December 4, 2017 from 139.047.023.222
Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
11. Lee N, Chan K, Bekelman JE, et al: Salvage
re-irradiation for recurrent head and neck cancer. Int
J Radiat Oncol Biol Phys 68:731-740, 2007
12. Sulman EP, Schwartz DL, Le TT, et al: IMRT
reirradiation of head and neck cancer-disease con-
trol and morbidity outcomes. Int J Radiat Oncol Biol
Phys 73:399-409, 2009
13. Piccirillo JF, Spitznagel EL Jr, Vermani N, et al:
Comparison of comorbidity indices for patients with
head and neck cancer. Med Care 42:482-486, 2004
14. Alho OP, Hannula K, Luokkala A, et al: Differ-
ential prognostic impact of comorbidity in head and
neck cancer. Head Neck 29:913-918, 2007
15. Read WL, Tierney RM, Page NC, et al: Differ-
ential prognostic impact of comorbidity. J Clin Oncol
22:3099-3103, 2004
16. Karvonen-Gutierrez CA, Ronis DL, Fowler KE,
et al: Quality of life scores predict survival among
patients with head and neck cancer. J Clin Oncol
26:2754-2760, 2008
17. Greene FL, Page DL, Fleming ID, et al: AJCC
Cancer Staging Manual (ed 6). New York, NY,
Springer, 2002
18. Therasse P, Arbuck SG, Eisenhauer EA, et al:
New guidelines to evaluate the response to treat-
ment in solid tumors. J Natl Cancer Inst 92:205-216,
2000
19. Paleri V, Wight RG: A cross-comparison of
retrospective notes extraction and combined notes
extraction and patient interview in the completion
of a comorbidity index (ACE-27) in a cohort of
■ ■ ■
United Kingdom patients with head and neck
cancer. J Laryngol Otol 116:937-941, 2002
20. Paleri V, Wight RG: Applicability of the adult
comorbidity evaluation - 27 and the Charlson in-
dexes to assess comorbidity by notes extraction in a
cohort of United Kingdom patients with head and
neck cancer: A retrospective study. J Laryngol Otol
116:200-205, 2002
21. Harrell FE Jr: Regression Modeling Strategies
With Applications to Linear Models, Logistic Regres-
sion, and Survival Analysis. New York, NY, Springer
Verlag, 2001
22. Kim AJ, Suh JD, Sercarz JA, et al: Salvage
surgery with free flap reconstruction: Factors affect-
ing outcome after treatment of recurrent head and
neck squamous carcinoma. Laryngoscope 117:
1019-1023, 2007
23. Argiris A, Brockstein BE, Haraf DJ, et al:
Competing causes of death and second primary
tumors in patients with locoregionally advanced
head and neck cancer treated with chemoradiother-
apy. Clin Cancer Res 10:1956-1962, 2004
24. Sulman EP, Schwartz DL, Le TT, et al: IMRT
reirradiation of head and neck cancer-disease con-
trol and morbidity outcomes. Int J Radiat Oncol Biol
Phys 2008 Jun 13 [Epub ahead of print]
25. Bonner JA, Harari PM, Giralt J, et al: Radio-
therapy plus cetuximab for squamous-cell carcinoma
of the head and neck. N Engl J Med 354:567-578,
2006
© 2009 by American Society of Clinical Oncology 1991