Int. J. Radiation Oncology Biol. Phys., Vol. 69, No. 4, pp. 1032–1041, 2007Copyright � 2007 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/07/$–see front matter

doi:10.1016/j.ijrobp.2007.05.017

CLINICAL INVESTIGATION Head and Neck

ARE WE INFLUENCING OUTCOME IN OROPHARYNX CANCER WITH INTENSITY-MODULATED RADIOTHERAPY? AN INTER-ERA COMPARISON

C. WESLEY HODGE, M.D.,* SØREN M. BENTZEN, PH.D.,* GORDON WONG, M.D.,*

KAREN L. PALAZZI-CHURAS, M.P.H.,*y PEG A. WIEDERHOLT, R.N.,*y VINAI GONDI, M.S.,*

GREGORY M. RICHARDS, M.D.,* GREGORY K. HARTIG, M.D.,y AND PAUL M. HARARI, M.D.*

Departments of *Human Oncology, and ySurgery-Otolaryngology, University of Wisconsin Hospital and Clinics, Madison, WI

Purpose: To analyze the outcome in all oropharynx cancer patients treated at the University of Wisconsin during1995–2005 and highlight the methodologic challenge in comparing outcome after intensity-modulated radiother-apy (IMRT) with that of historical controls.Methods and Materials: Outcomes were compared in 195 oropharynx cancer patients after definitive radiotherapywith curative intent in the pre-IMRT era (pre-IMRT, n = 105), after IMRT (IMRT+, n = 52) or after non-IMRTtechniques during the IMRT era (IMRT�, n = 38).Results: With a median follow-up of 30.4 months, the 3-year overall survival rate in IMRT+, IMRT�, and pre-IMRT patients was 88.2%, 81.1%, and 67.7%, respectively; and for locoregional control was 96.1%, 78.1%,and 81.1%. Patients from the IMRT era more frequently received concurrent chemotherapy (67% vs. 6%, p <0.001) and underwent adjuvant neck dissection (52% vs. 29%, p = 0.002). Patients with T3–4 disease and bilateralneck disease were significantly less likely to receive IMRT. Cox regression analysis identified IMRT as a significantprognostic factor (p = 0.04); however, after including T stage in the model, IMRT lost independent significance (p =0.2). Analysis of a potential effect of IMRTon Grade 3+ mucositis or skin reaction was also hampered by the changein other treatment characteristics.Conclusions: Outcomes in oropharynx cancer have improved at our institution since the introduction ofIMRT. However, multiple factors have contributed to this improvement, and presentation of IMRT outcomeswithout the full context of historical and contemporary controls may yield data that overstate outcome afterIMRT. � 2007 Elsevier Inc.

Oropharynx, Cancer, Intensity-modulated radiotherapy, Radiotherapy, Head and neck.

INTRODUCTION

Squamous cell carcinoma of the oropharynx is an aggressive

malignancy with a high propensity for locoregional tumor in-

filtration. The rich lymphatic network of the oropharynx pre-

disposes tumors in this area to manifest early regional lymph

node involvement. The achievement of locoregional control

is central to successful outcome in oropharynx cancer and re-

lates directly to overall survival (1). Historically, surgery and

radiation have served as the dominant treatment modalities

for these tumors. More recently, radiotherapy (RT) or chemo-

radiation approaches have found increased favor in light of

the excellent locoregional control rates, as well as the func-

tional and cosmetic challenges that can accompany definitive

surgery for oropharynx cancer.

10

Intensity-modulated RT (IMRT) has emerged as a power-

ful new radiation technique within the past decade (2, 3). For

oropharynx patients, IMRT offers the potential benefit of di-

minished radiation dose to several head-and-neck normal tis-

sue structures. Thus, in contrast to chemoradiation, in which

the aim is to enhance locoregional tumor control, the intro-

duction of IMRT is driven by a reduced-toxicity rationale.

The crucial question is whether this sparing of normal tissue

effects can be achieved without compromising locoregional

tumor control probability. Early reports from an increasing

number of single-institution studies demonstrate excellent

outcomes in the treatment of oropharynx cancer with

IMRT, with 2-year locoregional control rates in excess of

85%, leading many investigators to conclude that these

Reprint requests to: C. Wesley Hodge, M.D., Department of Ra-diotherapy, University of Wisconsin, 600 Highland Ave., K4/B29,Madison, WI 53792. Tel: (608) 263-8500; Fax: (608) 263-9167;E-mail: hodge@humonc.wisc.edu

Presented at the 48th Annual Meeting of the American Societyfor Therapeutic Radiology and Oncology (ASTRO), November5–9, 2006, Philadelphia, PA.

Acknowledgments—The authors thank the entire staff of the Univer-sity of Wisconsin Department of Radiotherapy for their support andfor the world-class level of compassionate cancer care theyprovide to patients on a daily basis.

Conflict of interest: none.Received Feb 13, 2007, and in revised form May 1, 2007.

Accepted for publication May 2, 2007.

32

IMRT inter-era comparison d C. W. HODGE et al. 1033

control rates are better than or, at the very least, not inferior to

what has been achieved with conventional radiation dose dis-

tributions (4–7). The aim of the present study was to illustrate

that even the latter claim of noninferiority may not be

straightforward to support on the basis of single-group out-

come reporting.

At the University of Wisconsin, we began using IMRT

for selected oropharynx cancer patients in November 2001.

We herein report the results of our IMRT experience, using

our entire contemporary oropharynx cancer patient cohort

(dating back to 1990) as a control, in an effort to provide

a more comprehensive context in which to examine the recent

improvement in outcome and the potential contribution of

IMRT to this improvement.

METHODS AND MATERIALS

Patient and staging evaluationBetween July 1990 and November 2005, 195 patients with histo-

logically confirmed cancer of the oropharynx received definitive

radiation with curative intent at the University of Wisconsin. Two

patients presented with lymphoepithelioma, and the remainder dem-

onstrated squamous cell carcinoma histology. Pretreatment evalua-

tion included a comprehensive history and physical examination,

including either direct or fiberoptic visualization of the oral cavity,

oropharynx, nasopharynx, hypopharynx, and larynx, dental evalua-

tion, radiographic imaging of the head and neck, primarily consist-

ing of computed tomography (CT), magnetic resonance imaging

(MRI) for selected patients, radiographic imaging of the chest (X-

ray or CT), complete blood counts, and positron emission tomogra-

phy (PET) scans for selected patients once this technology became

readily available. Patients were staged according to the 1998 Amer-

ican Joint Committee on Cancer staging classification. Patient

demographics, including tumor stage distribution, are detailed in

Table 1. All patients received RT as the primary treatment modality.

ChemotherapySixty-five patients (33%) received concurrent chemotherapy. The

chemotherapy regimens used were as follows: low-dose weekly

cisplatin (30 mg/m2) in 52 patients (27%), high-dose cisplatin (100

mg/m2) administered every 3 weeks in 11 patients (6%), and cetux-

imab (400 mg/m2 initial dose, followed by weekly doses of 250

mg/m2) in 2 patients (1%). Of the patients who received chemother-

apy, 53 (82%) completed their full regimen as initially prescribed.

SurgerySeventy-seven patients underwent post-RT neck dissection

(Table 1), and 14 patients underwent pre-RT neck dissection, pre-

dominantly at outside facilities before referral to the University of

Wisconsin. During 1990–2005 at our institution, the decision to

perform a postradiation neck dissection has been primarily based

on the initial nodal staging. All patients with N2–3 neck disease

are strongly considered for adjuvant neck dissection at 8–10 weeks

after completion of radiation. Until very recently, we have not used

imaging studies as a primary factor to guide consideration regarding

the use of postradiation neck dissection.

IMRT technical considerationsAll patients receiving IMRT underwent a planning CT scan with

i.v. contrast in our department. Patients were scanned in the supine

position and fitted with customized Aquaplast masks (Aquaplast,

Wyckoff, NJ) to immobilize the head and neck. Additionally, pa-

tients who were to receive linear accelerator–based IMRT were fit-

ted with a custom maxillofacial bite tray with embedded fiducial

markers. These fiducials provide real-time information to optical po-

sition sensors mounted within the treatment room regarding patient

translation and rotation relative to the designated isocenter during

treatment, enabling highly precise localization and delivery of radi-

ation treatments, as previously described (8). The planning CT was

then obtained with 2.5-mm intervals between the acquired axial

images.

Computed tomography–based treatment planning was performed

on the Pinnacle3 treatment-planning system (Philips Medical Sys-

tems, Andover, MA). In selected cases, fusion with PET imaging

was used to help delineate gross tumor volume (GTV) and nodal dis-

ease, which was subsequently contoured on the planning CT scan.

The clinical target volume (CTV), defined as all areas at risk for mi-

croscopic disease, was also contoured. The CTV typically encom-

passed the GTV plus a variable margin, as well as the draining

locoregional lymphatics. In patients with unilateral node-positive

disease, the CTV was further subdivided into a high-risk CTV

(CTV1), which contained the GTV, tumor margin, and ipsilateral

lymphatics, and a low-risk CTV (CTV2), which contained the

Table 1. Patient characteristics

IMRT era

Factor Pre-IMRT era IMRT� IMRT +

GenderMale 75 (71) 36 (95) 38 (73)Female 30 (29) 2 (5) 14 (27)

Tumor locationTonsil 50 (48) 20 (53) 26 (50)Base of tongue 44 (42) 18 (47) 22 (42)Soft palate 11 (10) 0 4 (8)

T stageTis 0 0 1 (2)T1 27 (26) 4 (11) 13 (25)T2 43 (41) 18 (47) 31 (60)T3 23 (22) 8 (21) 6 (11)T4 12 (11) 8 (21) 1 (2)

N stageN0 28 (27) 1 (2) 8 (15)N1 12 (11) 2 (5) 9 (17)N2a 9 (9) 3 (8) 6 (12)N2b 35 (33) 14 (37) 21 (40)N2c 14 (13) 11 (29) 4 (8)N3 7 (7) 7 (18) 4 (8)

AJCC stage0 0 0 1 (2)I 5 (5) 0 1 (2)II 18 (17) 0 5 (10)III 14 (13) 2 (5) 10 (19)IV 68 (65) 36 (95) 35 (67)

Fractionation schemeOnce daily 26 (25) 21 (55) 52 (100)Twice daily 40 (38) 0 0Once/twice daily hybrid 39 (37) 17 (45) 0

Concurrent chemotherapy 6 (6) 32 (84) 28 (54)Neck dissection 39 (37) 25 (66) 27 (52)

Abbreviations: IMRT = intensity-modulated radiotherapy;AJCC = American Joint Committee on Cancer.

Values are number (percentage).

1034 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

contralateral lymphatics. Primary cervical lymph node stations

(Levels II–V) were comprehensively treated as a standard approach,

although the most superior jugular nodes in an N0 neck, contralat-

eral to the primary tumor, were not routinely included to afford im-

proved parotid gland sparing. This technique reflects data from

several institutions demonstrating a low risk of contralateral high

neck failure for this cohort of patients (9, 10). A planning target vol-

ume (PTV) was generated to account for setup error and organ mo-

tion. Differential doses were assigned on the basis of disease risk;

gross disease was encompassed by the PTV70 (GTV + 3 mm),

high-risk subclinical disease by the PTV60 (CTV1 + 3 mm), and

low-risk subclinical disease by the PTV54 (CTV2 + 3 mm). Normal

anatomic structures, including the parotid glands, spinal cord, brain-

stem, mandible, and optic nerves were additionally contoured. In-

verse treatment planning was then performed with specifications

that 100% of the respective PTVs receive the prescribed dose.

Dose constraints to normal structures were as follows: contralateral

parotid #22.5 Gy to 30% volume, spinal cord maximum dose #35

Gy, brainstem maximum dose #50 Gy, lens maximum dose #2 Gy.

Dose–volume histograms were generated and reviewed by the treat-

ing physician to ensure compliance with these constraints. Examples

of a typical treatment plan and dose–volume histogram are shown in

Fig. 1. An isodose line was selected at the discretion of the treating

physician to provide adequate dosimetric coverage to the various

areas at risk. Treatments were delivered using 6-MV photons with

a median dose of 70 Gy (range, 65.1–70.4 Gy).

Conventional RT technical considerationsTreatment planning for patients who received conventional three-

dimensional conformal RT (3D-CRT) used the same positioning

and immobilization procedures outlined above. Patients did not

require a maxillofacial bite tray, however, with the exception of

selected patients who used a mouthpiece for tongue or dental manip-

ulation. All 3D-CRT treatments comprehensively covered the pri-

mary tumor bed and draining cervical lymphatics, including the

supraclavicular lymph nodes by means of a separate anterior supra-

clavicular field. At our institution the supraclavicular field is half-

beam blocked superiorly, whereas the lateral beams are allowed to

Fig. 1. Isodose distribution and dose–volume histogram of an oropharynx intensity-modulated radiotherapy plan for apatient with T2 N2b M0 squamous cell carcinoma of the left base of tongue. PTV = planning target volume.

IMRT inter-era comparison d C. W. HODGE et al. 1035

diverge across the field junction with the supraclavicular field,

thereby ensuring optimal beam flatness near the lateral isocenter

across the dominant primary tumor bed in the oropharynx. Appro-

priate spinal cord blocking is used. Virtually all of the 3D-CRT

plans used a ‘‘shrinking field technique’’ with clinically appropriate

off-cord field reductions and final boosts to the regions of gross

tumor involvement. Photon beam energies varied from 4 to 6 MV,

and electron beams were used (most commonly 6–9 MV) to boost

dose to bilateral level V nodal stations as clinically appropriate.

Two base-of-tongue patients received a boost to the primary tumor

using a 192Ir implant. The median dose to the primary tumor in the

conventional 3D RT cohort was 72 Gy (range, 60–78 Gy), with

once-daily fractionation used in 47 patients (33%), twice-daily frac-

tionation in 56 patients (39%), and hybrid once/twice daily regi-

mens in 40 patients (28%).

Follow-upAll patients were evaluated with head-and-neck examination by

the treating physician at least once per week during treatment. Post-

treatment evaluations were performed monthly for 6 months, then

every 2 to 3 months for the first 2 years, biannually through 5 years,

and annually thereafter. Clinical examination included indirect

mirror examination or fiberoptic nasopharyngoscopy visualizing

all head-and-neck mucosal surfaces, and careful evaluation of the

cervical and supraclavicular nodal areas. Radiographic follow-up

commonly included a CT scan within the first 6 months after treat-

ment, followed by annual chest X-rays. Additional imaging studies

were ordered as clinically indicated. Chronic toxicity was assessed

during each follow-up visit and assigned a score according to Radi-

ation Therapy Oncology Group adverse events criteria.

Data review and statistical methodsAfter detailed review of all medical records data, a computerized

head-and-neck database was created, which contained comprehen-

sive data regarding patient demographics, diagnosis, treatment,

and outcome. After institutional review board approval and certifi-

cation of Health Insurance Portability and Accountability Act com-

pliance for this database, the above-mentioned patient cohort was

retrospectively reviewed. The first patient treated with an IMRT

plan began treatment on November 15, 2001. Our historical control

cohort (designated pre-IMRT) consists of patients who initiated

treatment before this date, back to July 1990. Patients who initiated

conventional RT treatments on or after this date are designated as

IMRT�. Patients treated with IMRT are designated IMRT+. Time

to failure was designated from the time of clinical or radiographic

recurrence, using Day 1 of RT as time zero.

Statistical analysis was performed on SPSS v. 14.0 (SPSS, Chi-

cago, IL), and includes 2- and 3-year Kaplan-Meier estimates of

overall survival, cause-specific survival, and locoregional control.

Comparisons of demographic and treatment-related factors both

within the IMRT era and between treatment eras were made using

the Pearson chi-square, Fisher’s exact, and Mann-Whitney tests

where appropriate. Univariate and multivariate analysis of possible

prognostic factors were performed using the Cox proportional

hazards model.

RESULTS

Treatment outcomesMedian follow-up was 30.6 months (range, 3.0–165.8

months) for all patients, 33.9 months (same range) for all

living patients, 23.8 months (range, 3.0–52.9 months) for

IMRT-era patients, and 54.1 months (range, 3.3–165.8

months) for pre-IMRT patients. Nine patients (5%) were

lost to follow-up. Sixty-six patients (34%) have died.

Twenty-nine patients died as a result of their oropharynx can-

cer or complications thereof. There were 16 deaths from sec-

ond primary malignancies, and the remaining deaths were

attributable to intercurrent disease (21 patients). The 2- and

3-year overall survival rates were 94.5% and 88.2%, respec-

tively, for IMRT+ patients, 81.1% and 81.1% for IMRT� pa-

tients, and 78.6% and 67.7% for pre-IMRT patients (Fig. 2).

The 2- and 3-year cause-specific survival rates were 97.7%

and 97.7%, respectively, for IMRT+ patients, 83.5% and

83.5% for IMRT� patients, and 87.7% and 79.7% for pre-

IMRT patients (Fig. 3). The 2- and 3-year locoregional control

rates were 96.1% and 96.1%, respectively, for IMRT+ pa-

tients, 78.1% and 78.1% for IMRT� patients, and 82.3%

and 81.1% for pre-IMRT patients (Fig. 4). Five-year overall

survival, cause-specific survival, and locoregional control

rates for the pre-IMRT patients were 54.6%, 73.5%, and

78.5%, respectively.

Concurrent chemoradiation has become a standard-of-care

treatment approach for advanced oropharynx cancer patients

on the basis of results of randomized trials and subsequent

meta-analyses (11–15). Before these trials, however, the

incorporation of chemotherapy in definitive head-and-neck

cancer management was controversial. For this reason, the

vast majority of patients in this series who received concur-

rent chemoradiation did so during the more recent IMRT

era, with 60 patients (67%) receiving concurrent cisplatin-

based treatment, as opposed to 6 patients (6%) in the pre-

IMRT era (p < 0.001, Fisher’s exact test). Similarly, the

rate of posttreatment selective neck dissection has increased

somewhat during the IMRT era in our series, reflecting an

increased percentage of N2–3 patients. Forty-six patients

(51%) proceeded to selective neck dissection after radiation

in the IMRT era, compared with 31 patients (30%) in the

pre-IMRT era (p = 0.002, Mann-Whitney test).

Patterns of failureThere were 37 failures within the entire oropharynx patient

cohort. The contributions of local, regional, and distant dis-

ease to these failures are detailed in Fig. 5. The median

time to failure at any site for the entire study population

was 8.6 months (range, 2.0–45.9 months), and median time

to locoregional failure was 6.8 months (range, 2–45.9

months). Only 4 failures at any site were observed after 2

years (3 locoregional, 1 distant), and no failures occurred af-

ter 4 years. There were 2 locoregional failures in the cohort of

patients who received IMRT, detected 3.9 and 6.9 months af-

ter completion of chemoradiation and radiation alone, respec-

tively. In IMRT-era patients who received conventionally

delivered radiation, there were 7 locoregional failures, with

a median time to recurrence of 4.4 months (range, 2.0–26.6

months). There were 21 locoregional failures in the pre-

IMRT group, with a median time to locoregional failure of

8.6 months (range, 2.3–45.9 months). Twenty-nine patients

1036 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

Fig. 2. Overall survival. IMRT = intensity-modulated radiotherapy.

received some form of salvage or palliative treatment after

recurrence. Of the 28 patients treated with salvage therapy,

4 remain alive with no evidence of disease recurrence, all

of whom received surgery as the primary salvage treatment.

Acute toxicityThere were no treatment-related fatalities. Hospitalization

was required in 35 patients, 15 (14.3%) in the pre-IMRT

era, 11 (29%) in IMRT� patients, and 9 (17%) in

IMRT+ patients. Eighteen of the patients who were hospi-

talized for acute toxicity were receiving concurrent chemo-

therapy, comprising 51% of all patients hospitalized and

27% of the chemotherapy cohort. The incidence of percuta-

neous endoscopic gastrostomy tube placement was higher

in the IMRT-era group (52% vs. 33%, p = 0.006), although

insignificant between IMRT+ and IMRT� patients within

Fig. 3. Cause-specific survival. IMRT = intensity-modulated radiotherapy.

IMRT inter-era comparison d C. W. HODGE et al. 1037

Fig. 4. Locoregional control. IMRT = intensity-modulated radiotherapy.

the IMRT era (46% IMRT+ vs. 61% IMRT�, p = 0.1),

again reflecting the evolution of standard clinical practice

at this institution. Using the Mann-Whitney test, it was

determined that there was no significant difference in the

incidence of Grade 3 mucositis between treatment eras

(65% IMRT vs. 65% pre-IMRT, p = 1.0), or within the

IMRT era (58% IMRT+ vs. 75% IMRT�, p = 0.2).

Chronic toxicityReported crude incidence of moderate xerostomia was

67% in the pre-IMRT group, 63% in the IMRT� group,

and 56% in the IMRT+ group. Mann-Whitney analysis

shows no significant difference between treatment eras (p =

0.3) and no difference between patients who received

IMRT vs. conventional 3D-CRT (p = 0.4). Overall, there

were 9 reported cases of Grade 3 or higher chronic toxicity

(5%), with 4 cases of osteoradionecrosis, 2 of progressive

dental caries prompting extraction, and 1 each of facial

edema, swallow dysfunction, and tracheal fibrosis. These

toxicities had a median time to occurrence of 11.3 months

(range, 8.6–98 months). There were no reported Grade 4

chronic toxicities. There was one Grade 3 toxicity reported

within the IMRT+ cohort (2%), two in the IMRT� group

(5%), and six in the pre-IMRT group (6%).

Prognostic factorsUnivariate and multivariate analysis was performed in an

attempt to identify factors within the three treatment cohorts

that were of prognostic significance. The use of IMRT was

a significant positive prognostic factor in univariate analysis

(p = 0.02). Other factors that had positive prognostic signif-

icance on univariate analysis included smoking history and T

stage. Tumor site, American Joint Committee on Cancer

stage, use of concurrent chemotherapy, treatment era, and

history of alcohol abuse were among the factors not reaching

statistical significance in univariate analysis.

Multivariate analysis was conducted to further quantify the

potential impact of receiving IMRT. This analysis was re-

stricted to patients treated in the IMRT era. In a Cox model

with IMRT+/IMRT� as the only covariate, this factor was

statistically significant (p = 0.03), with a ratio of hazard rates

of 0.21 (95% confidence interval 0.04–0.998). However,

when T stage was entered into the model, this factor assumed

dominant prognostic significance (p = 0.003) and IMRT lost

statistical significance (p = 0.2). The ratio of hazard rates for

IMRT+/IMRT� after adjusting for T stage was estimated at

Local

DistantRegional

7(19%)

9(24%)

02(5%)

16(43%)

0 3(8%)

Fig. 5. Patterns of recurrence for entire oropharynx cohort,1990–2005.

1038 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

0.37 but with a wide 95% confidence interval (0.07–1.97).

This low statistical power to arrive at a precise estimate of

prognosis after IMRT vs. conventional RT can best be illus-

trated by applying the hazard rate estimate to an assumed

80% 3-year survival after treatment with conventional ther-

apy. Under this assumption, we estimate that the true change

in prognosis resulting from IMRT with 95% probability lies

between a 16-percentage-point disadvantage and an 18-per-

centage-point advantage. Thus the present study, despite be-

ing of comparable size to other published IMRT experiences

in head and neck squamous cell carcinoma, cannot provide

a prognostic estimate for IMRT vs. conventional delivery

that has clinically useful precision.

DISCUSSION

The use of definitive radiation as a primary treatment strat-

egy for oropharynx cancer is well established. Although the

long-term sequelae of radiation are by no means trivial, this

treatment approach commonly preserves good cosmetic

and functional outcome, with tumor control rates similar or

superior to those found in published surgical series (16,

17). The use of IMRT, although offering several potential do-

simetric advantages over conventionally delivered radiation

(18), remains a relatively new technique in clinical practice.

The impact of IMRT on preservation of salivary function

has been increasingly well described in the recent literature

(19–23). Nevertheless, no prospective trial has been con-

ducted identifying a benefit with regard to tumor control or

survival. The recently completed Radiation Therapy Oncol-

ogy Group trial 0022 examining the use of IMRT for patients

with intermediate-stage cancers of the oropharynx may pro-

vide valuable information regarding the feasibility of exe-

cuting of a multi-institution head-and-neck IMRT trial,

although this is a straightforward single-arm Phase II study

(24). The vast majority of the published data at the time of

this writing originate from single-institution retrospective

studies.

With respect to locoregional control, cause-specific sur-

vival, and overall survival for oropharynx cancer patients,

the results of the present study compare favorably to those

reported by other institutions (Table 2) (4–7, 10, 25–33, 47,

48). De Arruda et al. (6) recently reported on 52 patients

treated with IMRT at Memorial Sloan-Kettering Cancer Cen-

ter, demonstrating overall survival of 98%, local control of

98%, and regional control of 88% at 2 years. Huang et al.(4) also report outcomes that exceed those typically observed

in patients treated with 3D-CRT, with 2-year overall survival,

disease-free survival, and locoregional control rates of 89%,

91%, and 89%, respectively. Eisbruch et al. (10) demon-

strated overall survival, disease-free survival, and locore-

gional control rates of 77%, 75%, and 82%, respectively, at

3 years.

As we analyze overall outcomes from the current orophar-

ynx cancer series at the University of Wisconsin, we identify

several potential contributing factors to the excellent out-

come in IMRT patients. One notable finding is the imbalance

in our IMRT+ and IMRT� patients with regard to T stage.

From the entire cohort of 52 IMRT+ patients, only 1 had

T4 disease, supporting the notion that patient selection may

contribute to the improved outcome seen in our IMRT+

patients. Best clinical judgment has guided the selection of pa-

tients thought most likely to derive benefit from IMRT tech-

niques without risk of compromise in tumor target dosing.

In some cases, the desired parotid-sparing benefit of IMRT

treatment cannot be successfully (and safely) achieved in pa-

tients with bulky primary tumors and/or bilateral upper cervi-

cal disease. Indeed, within the IMRT era it was significantly

less likely for our patients with bilateral N2 or N3 neck disease

to be treated with IMRT techniques. In these cases, it is our

view that comprehensive treatment (despite a higher level

of salivary toxicity) is preferable to highly conformal RT

with the potential increased risk of geographic failure. Our

selection of IMRT candidates confirms that those patients

who were not deemed appropriate IMRT candidates were fre-

quently, and by definition, at higher risk by virtue of advanced

T or N stage. Despite this, we observed no significant differ-

ence in overall survival or locoregional control between our

IMRT� and pre-IMRT cohorts.

The comprehensive treatment of oropharynx cancer has

undergone gradual evolution in the 15 years encompassed

by this review. Concurrent chemotherapy has become a com-

mon standard of care, in large part on the basis of the dem-

onstration of a modest survival advantage in meta-analyses

of randomized clinical trials (13–15). The significantly in-

creased percentage of patients in the IMRT era who received

concomitant chemotherapy is, therefore, likely to convey

some survival benefit, regardless of the radiation technique

used. Similarly, postradiation selective neck dissection in

patients with >N1 neck disease has been increasingly used

at our center over the time span of this series, reflecting

both more aggressive surgical management and more com-

mon N2–3 neck presentations. This is quite possibly a result

of stage migration from increasingly sophisticated imaging

capabilities (e.g., the increasing use of PET imaging in the

initial staging) to appropriately identify patients with clini-

cally occult nodal metastases. Although the role of adjuvant

neck dissection in this setting remains controversial, with

some institutions favoring a more conservative approach

(34, 35), the ultimate rate of neck control using this ap-

proach is excellent in several institutional series, including

our own (36–38). This issue may also contribute to the

locoregional control in our IMRT-era patients, who were

more likely to proceed to a planned selective neck dissection

after radiation or chemoradiation. No significant difference

in the prevalence of acute toxicities was observed between

treatment eras, although the lack of difference despite

more aggressive treatment may suggest some quality-of-

life benefit.

Several issues in the present study are characteristic of sim-

ilar reports in the literature. First and foremost, any compar-

ison of a contemporary cohort with a historical one creates an

imbalance in follow-up. This in fact underlies a premise of

the present report; namely that single-institution reviews of

IMRT inter-era comparison d C. W. HODGE et al. 1039

Table 2. Published outcome results with IMRT in oropharynx cancer

Institution (year) (reference) N Interval (y)Overall

survival (%)Cause-specificsurvival (%) Locoregional control (%)

U. of Michigan (2000) (27) 58* 2 NS NS 79Washington U. (2001) (5) 12 2 100 80y 88Heidelberg (2003) (32) 48z 2 92 NS 93 localUCSF (2003) (4) 41 2 89 91y 89U. Michigan (2004) (10) 80 3 77 75y 82 (94 OP)MDACC (2004) (47) 80 3 NS NS 94MCV (2004) (28, 29) 20 2 57 71 76 local, 67 regionalWashington U. (2004) (26) 52x 2 NS NS 79MDACC (2004) (25) 31 4 NS 66y 78U. of Iowa (2005) (7) 56k 3 82 NS 92 (98 OP)Zurich (2006) (33) 53k 2 87 NS 88 local, 93 regionalU. of Florida (2006) (30, 48) 64 2 89 94 90U. of Chicago (2006) (31) 69 3 69 NS 94MSKCC (2006) (6) 48 2 98 NS 98 local, 88 regionalU. of Wisconsin (2006) (present series) 52 3 88 97 96

Abbreviations: IMRT = intensity-modulated radiotherapy; U. = University; NS = not specified; UCSF = University of California, SanFrancisco; OP = oropharynx; MDACC = M. D. Anderson Cancer Center; MCV = Medical College of Virginia; MSKCC = Memorial Sloan-Kettering Cancer Center.

N reflects the cohort of patients who received definitive IMRT.* Includes 33 OP patients, separate results not reported.y Disease-free survival rate.z Includes 9 OP patients, separate results not reported.x Includes 28 OP, separate results not reported.k Includes an unspecified number of patients treated with IMRT postoperatively.

contemporary clinical results compared with historical con-

trols may inadvertently ignore several confounding factors.

Although the median follow-up of 23.8 months for our

IMRT-era patients is within the range of other reported series

(18–47 months) (4–7), it is considerably less than the 54.1

months median follow-up of our historical comparison. Be-

cause the clinical reports of outcome using IMRT are in their

relative infancy, more mature data may attenuate these

results somewhat. Despite this, it is noteworthy that in our

series nearly 90% of all failures occurred within the first

2 years, which is concordant with locoregional control data ob-

served in several other large, single-institution series (39–44).

The toxicity data from the present study are in general

agreement with reports from other institutional series (16,

20). Quantitative assessment has been performed by several

institutions (19, 22, 45, 46), and we are currently preparing

a comprehensive quality-of-life evaluation for this patient

population. Our failure to identify an independent outcome

benefit with the use of IMRT should not be interpreted as

an implication that IMRT offers no advantage over compre-

hensive RT, because we have herein analyzed outcome inde-

pendent of toxicity data. Obviously, the potential to offer

equivalent tumor control with reduced toxicity is a significant

cornerstone of the IMRT approach. Our results suggest

excellent outcomes with regard to tumor control using

IMRT. However, critical evaluation of head-and-neck cancer

patients on an individual basis is a likely component of this

success, and universal application of IMRT absent specific

tumor distribution considerations may inadvertently place

some patients at unnecessary risk of treatment failure.

CONCLUSIONS

Locoregional tumor control and survival outcome for oro-

pharynx cancer patients has shown improvement at our insti-

tution since the introduction of IMRT. However, the relative

contribution of radiation technique, concurrent chemother-

apy, adjuvant neck dissection, and stage migration secondary

to enhanced imaging techniques to this improvement remains

unclear. Indeed, since the introduction of IMRT at our center,

there has been clear selection of patients judged most likely to

derive clinical benefit to diminish normal tissue toxicity

while maintaining full tumor target coverage. Patients treated

with non-IMRT techniques in the IMRT era did not fare

worse than historical controls, despite an increased preva-

lence of adverse prognostic factors. In multivariate analysis,

when adjusted for T stage, IMRT has neither positive nor

negative prognostic importance. However, the current sam-

ple size does not allow us to conclude that IMRT and con-

ventional delivery are equivalent with any useful statistical

precision. Indeed, from a radiobiologic perspective, we

would not have anticipated an improved tumor outcome

with the use of IMRT over prior radiation schedule tech-

niques at our center. This is not proof that IMRT does not

provide adequate locoregional therapy; the point of our

analysis is simply that the current published experience is

inconclusive regarding this important question. This review

suggests that, despite superb head-and-neck IMRT results

presented by many institutions, the presentation of IMRT

outcomes without the full context of historical and contempo-

rary controls may inadvertently overstate the benefit of this

technique on tumor outcome.

1040 I. J. Radiation Oncology d Biology d Physics Volume 69, Number 4, 2007

REFERENCES

1. Wadsley JC, Bentzen SM. Investigation of relationship betweenchange in locoregional control and change in overall survivalin randomized controlled trials of modified radiotherapy inhead-and-neck cancer. Int J Radiat Oncol Biol Phys 2004;60:1405–1409.

2. Hong TS, Ritter MA, Tome WA, et al. Intensity-modulatedradiation therapy: Emerging cancer treatment technology. BrJ Cancer 2005;92:1819–1824.

3. Hong TS, Tome WA, Harari PM. Intensity-modulated radiationtherapy in the management of head and neck cancer. Curr OpinOncol 2005;17:231–235.

4. Huang K, Lee N, Xia P, et al. Intensity-modulated radiotherapyin the treatment of oropharyngeal carcinoma: A single institu-tional experience [Abstract]. Int J Radiat Oncol Biol Phys2003;57(Suppl. 2):302.

5. Chao KS, Majhail N, Huang CJ, et al. Intensity-modulated radi-ation therapy reduces late salivary toxicity without compromis-ing tumor control in patients with oropharyngeal carcinoma: Acomparison with conventional techniques. Radiother Oncol2001;61:275–280.

6. de Arruda FF, Puri DR, Zhung J, et al. Intensity-modulated ra-diation therapy for the treatment of oropharyngeal carcinoma:The Memorial Sloan-Kettering Cancer Center experience. IntJ Radiat Oncol Biol Phys 2006;64:363–373.

7. Yao M, Dornfeld KJ, Buatti JM, et al. Intensity-modulated radi-ation treatment for head-and-neck squamous cell carcinoma—the University of Iowa experience. Int J Radiat Oncol BiolPhys 2005;63:410–421.

8. Hong TS, Tome WA, Chappell RJ, et al. The impact of dailysetup variations on head-and-neck intensity-modulated radia-tion therapy. Int J Radiat Oncol Biol Phys 2005;61:779–788.

9. Astreinidou E, Dehnad H, Terhaard CH, et al. Level II lymphnodes and radiation-induced xerostomia. Int J Radiat OncolBiol Phys 2004;58:124–131.

10. Eisbruch A, Marsh LH, Dawson LA, et al. Recurrences nearbase of skull after IMRT for head-and-neck cancer: Implicationsfor target delineation in high neck and for parotid gland sparing.Int J Radiat Oncol Biol Phys 2004;59:28–42.

11. Adelstein DJ, Li Y, Adams GL, et al. An intergroup phase IIIcomparison of standard radiation therapy and two schedulesof concurrent chemoradiotherapy in patients with unresectablesquamous cell head and neck cancer. J Clin Oncol 2003;21:92–98.

12. Calais G, Alfonsi M, Bardet E, et al. Randomized trial of radi-ation therapy versus concomitant chemotherapy and radiationtherapy for advanced-stage oropharynx carcinoma. J NatlCancer Inst 1999;91:2081–2086.

13. El-Sayed S, Nelson N. Adjuvant and adjunctive chemotherapyin the management of squamous cell carcinoma of the head andneck region. A meta-analysis of prospective and randomizedtrials. J Clin Oncol 1996;14:838–847.

14. Munro AJ. An overview of randomised controlled trials ofadjuvant chemotherapy in head and neck cancer. Br J Cancer1995;71:83–91.

15. Pignon JP, Bourhis J, Domenge C, et al. Chemotherapy addedto locoregional treatment for head and neck squamous-cell carcinoma: Three meta-analyses of updated individualdata. MACH-NC Collaborative Group. Meta-Analysis of Che-motherapy on Head and Neck Cancer. Lancet 2000;355:949–955.

16. Mendenhall WM, Stringer SP, Amdur RJ, et al. Is radiationtherapy a preferred alternative to surgery for squamous cell car-cinoma of the base of tongue? J Clin Oncol 2000;18:35–42.

17. Parsons JT, Mendenhall WM, Stringer SP, et al. Squamous cellcarcinoma of the oropharynx: Surgery, radiation therapy, orboth. Cancer 2002;94:2967–2980.

18. Puri DR, Chou W, Lee N. Intensity-modulated radiation therapyin head and neck cancers: Dosimetric advantages and update ofclinical results. Am J Clin Oncol 2005;28:415–423.

19. Chao KS, Deasy JO, Markman J, et al. A prospective studyof salivary function sparing in patients with head-and-neckcancers receiving intensity-modulated or three-dimensionalradiation therapy: Initial results. Int J Radiat Oncol BiolPhys 2001;49:907–916.

20. Lin A, Kim HM, Terrell JE, et al. Quality of life after parotid-sparing IMRT for head-and-neck cancer: A prospective longitu-dinal study. Int J Radiat Oncol Biol Phys 2003;57:61–70.

21. Pacholke HD, Amdur RJ, Morris CG, et al. Late xerostomia af-ter intensity-modulated radiation therapy versus conventionalradiotherapy. Am J Clin Oncol 2005;28:351–358.

22. Parliament MB, Scrimger RA, Anderson SG, et al. Preservation oforal health-related quality of life and salivary flow rates after in-verse-planned intensity-modulated radiotherapy (IMRT) for head-and-neck cancer. Int J Radiat Oncol Biol Phys 2004;58:663–673.

23. Saarilahti K, Kouri M, Collan J, et al. Intensity modulated radio-therapy for head and neck cancer: Evidence for preserved sali-vary gland function. Radiother Oncol 2005;74:251–258.

24. Eisbruch A, Harris J, Garden AS, et al. Phase II multi-institu-tional study of IMRT for oropharyngeal cancer (RTOG 00-22): Early results [Abstract]. Int J Radiat Oncol Biol Phys2006;66(Suppl. 3):46–47.

25. Chao KS, Ozyigit G, Blanco AI, et al. Intensity-modulated ra-diation therapy for oropharyngeal carcinoma: Impact of tumorvolume. Int J Radiat Oncol Biol Phys 2004;59:43–50.

26. Chao KS, Ozyigit G, Tran BN, et al. Patterns of failure in patientsreceiving definitive and postoperative IMRT for head-and-neckcancer. Int J Radiat Oncol Biol Phys 2003;55:312–321.

27. Dawson LA, Anzai Y, Marsh L, et al. Patterns of local-regionalrecurrence following parotid-sparing conformal and segmentalintensity-modulated radiotherapy for head and neck cancer.Int J Radiat Oncol Biol Phys 2000;46:1117–1126.

28. Lauve A, Morris M, Schmidt-Ullrich R, et al. Simultaneous in-tegrated boost intensity-modulated radiotherapy for locally ad-vanced head-and-neck squamous cell carcinomas: II—Clinicalresults. Int J Radiat Oncol Biol Phys 2004;60:374–387.

29. Lauve A, Morris M, Schmidt-Ullrich R, et al. A phase I trial us-ing a parotid-sparing, accelerated intensity-modulated radio-therapy (IMRT) regimen to treat locally advanced head andneck squamous cell carcinoma [Abstract]. Int J Radiat OncolBiol Phys 2003;57(Suppl. 2):302.

30. Mendenhall WM, Amdur RJ, Palta JR. Intensity-modulated ra-diotherapy in the standard management of head and neck can-cer: Promises and pitfalls. J Clin Oncol 2006;24:2618–2623.

31. Milano MT, Vokes EE, Kao J, et al. Intensity-modulated radi-ation therapy in advanced head and neck patients treated withintensive chemoradiotherapy: Preliminary experience andfuture directions. Int J Oncol 2006;28:1141–1151.

32. Munter MW, Thilmann C, Hof H, et al. Stereotactic intensitymodulated radiation therapy and inverse treatment planningfor tumors of the head and neck region: Clinical implementationof the step and shoot approach and first clinical results. Radio-ther Oncol 2003;66:313–321.

33. Studer G, Huguenin PU, Davis JB, et al. IMRT using simulta-neously integrated boost (SIB) in head and neck cancer patients.Radiat Oncol 2006;1:7.

34. Narayan K, Crane CH, Kleid S, et al. Planned neck dissection asan adjunct to the management of patients with advanced neckdisease treated with definitive radiotherapy: For some or forall? Head Neck 1999;21:606–613.

35. Peters LJ, Weber RS, Morrison WH, et al. Neck surgery in pa-tients with primary oropharyngeal cancer treated by radiother-apy. Head Neck 1996;18:552–559.

IMRT inter-era comparison d C. W. HODGE et al. 1041

36. Boyd TS, Harari PM, Tannehill SP, et al. Planned postradio-therapy neck dissection in patients with advanced head andneck cancer. Head Neck 1998;20:132–137.

37. Doweck I, Robbins KT, Mendenhall WM, et al. Neck level-specific nodal metastases in oropharyngeal cancer: Is therea role for selective neck dissection after definitive radiation ther-apy? Head Neck 2003;25:960–967.

38. Sewall G, Palazzi-Churas K, Richards G, et al. Planned post-radiotherapy neck dissection: Rationale and clinical outcomes.Laryngoscope 2007;117:121–128.

39. Bataini JP, Asselain B, Jaulerry C, et al. A multivariate primary tu-mour control analysis in 465 patients treated by radical radiotherapyfor cancer of the tonsillar region: Clinical and treatment parametersas prognostic factors. Radiother Oncol 1989;14:265–277.

40. Bentzen SM, Johansen LV, Overgaard J, et al. Clinical radiobi-ology of squamous cell carcinoma of the oropharynx. Int JRadiat Oncol Biol Phys 1991;20:1197–1206.

41. Brunin F, Mosseri V, Jaulerry C, et al. Cancer of the base of thetongue: Past and future. Head Neck 1999;21:751–759.

42. Mendenhall WM, Morris CG, Amdur RJ, et al. Definitive radio-therapy for tonsillar squamous cell carcinoma. Am J Clin Oncol2006;29:290–297.

43. Mendenhall WM, Morris CG, Amdur RJ, et al. Definitive radio-therapy for squamous cell carcinoma of the base of tongue. Am JClin Oncol 2006;29:32–39.

44. Perez CA, Patel MM, Chao KS, et al. Carcinoma of the tonsillarfossa: Prognostic factors and long-term therapy outcome. Int JRadiat Oncol Biol Phys 1998;42:1077–1084.

45. Eisbruch A, Dawson LA, Kim HM, et al. Conformal and inten-sity modulated irradiation of head and neck cancer: The poten-tial for improved target irradiation, salivary gland function, andquality of life. Acta Otorhinolaryngol Belg 1999;53:271–275.

46. Roesink JM, Moerland MA, Battermann JJ, et al. Quantitativedose-volume response analysis of changes in parotid glandfunction after radiotherapy in the head-and-neck region. Int JRadiat Oncol Biol Phys 2001;51:938–946.

47. Garden AS, Morrison WH, Rosenthal DI, et al. Intensity mod-ulated radiation therapy (IMRT) for metastatic cervical adenop-athy from oropharynx carcinoma [Abstract]. Int J Radiat OncolBiol Phys 2004;60(Suppl. 1):318.

48. Schoenfeld G, Amdur RJ, Morris CG, et al. Patterns of failureand toxicity of intensity-modulated radiation therapy for headand neck cancer [Abstract]. Int J Radiat Oncol Biol Phys2006;66(Suppl. 3):445.