C

R(c(hectmtswr(s1o

RP

Int. J. Radiation Oncology Biol. Phys., Vol. 66, No. 4, pp. 1017–1021, 2006Copyright © 2006 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/06/$–see front matter

doi:10.1016/j.ijrobp.2006.06.033

LINICAL INVESTIGATION Head and Neck

DEFINITIVE RADIOTHERAPY FOR HEAD-AND-NECK CANCER WITHRADIOGRAPHICALLY POSITIVE RETROPHARYNGEAL NODES:

INCOMPLETE RADIOGRAPHIC RESPONSE DOES NOT NECESSARILYINDICATE FAILURE

STANLEY L. LIAUW, M.D.,* ANTHONY A. MANCUSO, M.D.,† CHRISTOPHER G. MORRIS, M.S.,*ROBERT J. AMDUR, M.D.,* AND WILLIAM M. MENDENHALL, M.D.*

Departments of *Radiation Oncology, and †Radiology, University of Florida College of Medicine, Gainesville, FL

Purpose: Our aim was to report the control rate of radiographically positive retropharyngeal (RP) nodes withradiation therapy (RT) and to correlate posttreatment imaging with clinical outcome.Methods and Materials: Sixteen patients treated with definitive RT for head-and-neck cancer had radiograph-ically positive RP nodes (size >1 cm in largest axial dimension, or presence of focal enhancement, lucency, orcalcification), and both pre-RT and post-RT image sets available for review. An additional 21 patients withunconfirmed radiographically positive RP nodes had post–RT imaging, which consisted of computed tomography(CT) at a median of 4 weeks after completing RT. Patients with positive post-RT RP nodes underwentobservation with serial imaging.Results: Of 16 patients with pre-RT and post-RT images available for review, 9 (56%) had a radiographiccomplete response, and of 21 patients with unconfirmed positive RP nodes with post-RT images available forreview, 14 (67%) had a radiographic complete response. In all, 14 patients with incomplete response on post–RTimaging experienced control of their disease with no further therapy, and no RP node or neck failures were notedduring a median follow-up of 2.8 years. Six patients with positive post-RT RP nodes had serial imaging availablefor review, and none demonstrated radiographic progression of disease.Conclusions: Radiographic response at 4 weeks may not accurately reflect long-term locoregional control, as RPnodes may continue to resolve over time. The highest index of suspicion should be reserved for patients withprogressive changes in size, focal lucency, or focal enhancement on serial imaging after RT. © 2006 Elsevier Inc.

Head-and-neck cancer, Radiotherapy, Computed tomography, Retropharyngeal nodes.

(paptrrtnco

rhdc

0

INTRODUCTION

etropharyngeal (RP) nodes, first described by Rouviére1), lie in the retropharyngeal space medial to the internalarotid arteries at the level of vertebral bodies C1 to C32–5). Given the potential to harbor disease in the case ofead-and-neck cancer (6, 7), RP nodes are important tovaluate with imaging before definitive treatment, espe-ially because they are inaccessible to physical examina-ion. There are limited published data on the influence ofetastatic RP nodes on clinical outcome and reports from

he modern era are conflicting. The largest reported seriesuggests that patients treated with radiation therapy (RT)ho have metastatic RP nodes have a higher rate of neck

elapse and distant metastasis and a lower overall survival8), but this association is not supported by smaller surgicaleries in which RP nodes at risk are primarily dissected (9,0). Complicating this discrepancy is the uncertain abilityf imaging to accurately identify pathologic involvement

Reprint requests to: William M. Mendenhall, M.D., Department ofadiation Oncology, University of Florida Health Science Center,

.O. Box 100385, Gainesville, FL 32610-0385. Tel: (352) 265- A

1017

11), as radiographically positive RP nodes are not alwaysathologically involved. Regardless of these issues, it isccepted that RP nodes at risk for involvement should beroperly evaluated and treated in patients receiving defini-ive RT (12). There are no published data that address theesponse of involved RP nodes to RT, or the significance ofadiographically abnormal RP nodes post-RT. The goal ofhis study is to characterize the radiographic response of RPodes in patients treated with definitive RT for head-and-ancer, and correlate posttreatment imaging with clinicalutcome.

METHODS AND MATERIALS

This was an institutional review board–approved, retrospectiveeview of 550 consecutive patients with lymph node–positiveead-and-neck cancer treated at the University of Florida withefinitive RT with curative intent between 1990 and 2002. Patientharacteristics and treatment techniques for this set of patients are

287; Fax: (352) 265-0759; E-mail: mendewil@shands.ufl.eduReceived May 2, 2006, and in revised form June 14, 2006.

ccepted for publication June 18, 2006.

du1wfiutRrwdptacabshdppnaericgcARrc

pmdohptgt

irRttre

R

a1ahciho5(

rcppco

Sp

nrRtmo

Fpwt

Fsf0

1018 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 4, 2006

iscussed in detail elsewhere (13). RT was generally prescribedsing lateral, opposed head-and-neck fields to a dose of 74.4 Gy at.2 Gy/fraction twice daily. Radiographically positive RP nodesere covered by extending the superior border of the treatmentelds to include the retropharyngeal space anterior to C1 to C3,sing the superior landmark of the auditory meatus on conven-ional planning. Of 550 patients, 88 had radiographically positiveP nodes at initial presentation as dictated in tumor board notes or

adiographic reports in each patient’s chart. Retropharyngeal nodesere assessed at 3-mm intervals using a scanning plane perpen-icular to vertebral bodies C1 to C4, according to an institutionalrotocol for all head-and-neck cancer patients staged with con-rast-enhanced computed tomography (CT). Pre-RT images werevailable for recall from radiology archives for 27 patients. Soft-opy images were then reviewed jointly by a radiation oncologistnd a neuroradiologist on a computer workstation with outcomeslinded during review. RP nodes were characterized on the basis ofize (largest axial dimension) or focal abnormality (lucency, en-ancement, or calcification) and graded on a five-point scale: 0 �efinitely normal; 1 � probably normal; 2 � indeterminate; 3 �robably abnormal; and 4 � definitely abnormal. Radiographicallyositive RP nodes were defined as any retropharyngeal lymphode �1 cm or with any Grade 3 or 4 focal abnormality. Examplesre included in Figs. 1 to 3. Of 27 patients, 11 were subsequentlyxcluded from this study because of an inability to identify aadiographically positive RP node with these criteria (n � 4),nvolvement of a high-level II lymph node lateral to the internalarotid arteries rather than an RP node (n � 3), difficulty in distin-uishing an RP node from the primary disease (n � 2), lack of i.v.ontrast (n � 1), or no post–RT imaging available for review (n � 1).

total of 16 patients had 17 confirmed radiographically positiveP nodes and subsequent post–RT imaging (median 28 days,

ange, 15 to 42 days after completing RT). Patient and treatmentharacteristics for this group are described in Table 1.

Computed tomography (CT) was used to assess RP nodes atresentation and after RT in all cases. One patient also hadagnetic resonance imaging at presentation to better define the

isease extending into the nasopharynx, and this was the only setf images available for review on this patient. Images were onand for an additional 21 patients who did not have availablere-RT images, leaving a total of 37 patients with CT data afterreatment of a presumed radiographically positive RP node. As aeneral rule, patients with radiographically positive RP nodes afterreatment were observed with serial imaging with the understand-

ig. 1. Patient with a T3N2B base-of-tongue cancer who had are-treatment scan showing a 1.8-cm right retropharyngeal nodeith a definite focal lucency. Imaging 28 days after radiation

fherapy (RT) shows a retropharyngeal node of �1 cm.

ng that more aggressive therapy (salvage dissection or stereotacticadiosurgery) would be considered for progressive disease in theP node if the patient had the disease controlled elsewhere. The

iming of serial imaging was not strictly defined at the outset ofhe study period, but tended to be at every 3 to 4 months untiladiographic resolution, declining index of suspicion for dis-ase, or progression of disease elsewhere.

RESULTS

adiographic responseCharacteristics of radiographically positive RP nodes,

long with post-RT response, are described in Table 2. Of6 patients with radiographically positive RP nodes as wells pre-RT and post-RT images available for review, 9 (56%)ad a complete radiographic response on post-RT CT. In-omplete response was usually a result of a focal abnormal-ty (n � 6) rather than large size alone, as only 2 patientsad post-RT size �1 cm. Most focal abnormalities detectedn post-RT CT were identical on pretreatment imaging (n �), although 2 patients developed a new focal abnormalitycalcification) after RT.

Review of an additional 21 sets of post-RT CT scans withadiographically positive RP nodes by report demonstratedomplete radiographic response in 14 patients (67%). Of 7atients with incomplete response, 2 had nodes �1 cmost-RT but none were radiographically positive by sizeriteria alone. All patients had a focal lucency with (n � 1)r without (n � 6) calcification.

erial imaging of radiographically abnormalost-RT RP nodesSix patients with radiographically positive post-RT RP

odes had at least one subsequent CT scan available foreview. No patient demonstrated progressive disease in theP nodes. Two patients with focal calcification had persis-

ent calcification on serial imaging up to 4 months and 8onths. Three patients with focal lucency had resolution

f lucency by 2 months, 3 months, and 1 year with further

ig. 2. Patient with a T2N2B tonsil cancer who had a pretreatmentcan showing a 1.7-cm left retropharyngeal node with a definiteocal lucency. Imaging 29 days after radiation therapy shows a.9-cm retropharyngeal node with residual definite focal lucency.

ollow-up. One patient with a post-RT RP node of 1.2 cm

hl

C

nspd

w(pwld

creo(uctgntpsp

ptrptcrtpRim(mtie

RT co

P

T

N

F

C

P

h

1019Response of radiographically positive RP nodes ● S. L. LIAUW et al.

ad regression to normal size on the only available fol-ow-up study 2 years later.

linical outcomeAll 14 patients with incomplete response in the RP

odes after RT were observed rather than subjected to aalvage procedure for presumed disease. None of theseatients with post-RT radiographic abnormalities hadisease recurrence in the retropharyngeal nodes or neck,

Fig. 3. Patient with a T3N2B tonsil cancer had a pretreadefinite focal lucency and calcification. Imaging 24 daysnode with residual calcification; imaging 8 months after

Table 1. Characteristics of 16 patients with confirmedradiographically involved retropharyngeal (RP) nodes and

post–radiation therapy (RT) imaging

Characteristic Number (%)

rimary siteOropharynx 14 (88%)

Tonsil 9Base of tongue 4Soft palate 1

Hypopharynx 1 (6%)Posterior pharyngeal wall 1

Multiple synchronous primaries 1 (6%)stageT2 4 (25%)T3 4 (25%)T4 8 (50%)stageN1 1 (6%)N2A 0 (0%)N2B 6 (38%)N2C 4 (25%)N3 5 (31%)

ractionationConventional (once-daily) 5 (31%)Hyperfractionation 11 (69%)

hemotherapy 10 (63%)Neoadjuvant 2Concurrent 7Neoadjuvant and concurrent 1

retreatment RP node abnormality*Size �1 cm 11 (69%)Focal lucency 6 (38%)Focal enhancement 2 (13%)Focal calcification 2 (13%)

* Numbers do not add up to 16 because each positive node can

iave multiple abnormalities.

ith a median follow-up of 2.8 years in all patientsrange, 0.4 –7.9 years) and 5 years in living patients. Twoatients experienced failure with disease recurrence else-here; of these, 1 had distant disease 1 month after RT and

ocal disease 4 months later, and the other with distantisease alone 11 months after RT.

DISCUSSION

Involvement of retropharyngeal nodes with head-and-neckancer is highest for cancers arising in the nasopharynx, pha-yngeal wall, soft palate, tonsil, hypopharynx, and cervicalsophagus (7). Despite the possibility that imaging can lead toverdiagnosis of pathologic involvement (11), surgical series9–11, 14) still report rates of involvement greater than 20%pon dissecting retropharyngeal nodes for patients with certainancers of the head and neck. RT may be a preferred modalityo treat retropharyngeal nodes at risk at initial presentationiven the difficulty of retropharyngeal node dissection as theodes may be adherent to pharyngeal musculature and difficulto access near the base of skull. Furthermore, treatment withrimary RT with or without chemotherapy can potentiallypare the patient from surgery and the associated morbidity ofrimary resection or neck dissection (15, 16).

This study shows that the control rate of radiographicallyositive RP nodes is high with primary RT as long asreatment volumes provide adequate coverage of the area atisk. A total of 37 patients who had imaging of the retro-haryngeal nodes before and/or after RT were controlled inhe RP nodes and neck. The high rate of locoregionalontrol did not correlate with the relatively low completeadiographic response rate 4 weeks after RT (56% in pa-ients with confirmed disease on pretreatment scan, 67% inatients with unconfirmed initial radiographically positiveT nodes). The results suggest that radiographic abnormal-

ties including large size or focal lucency in the RP nodesay require more time after RT to completely resolve

although calcifications may persist for �1 year after treat-ent). Given this finding, it may be best to defer salvage

reatment for incompletely responding RP nodes until serialmaging shows progression, unless clinical suspicion for dis-ase is high. Risks of delaying salvage treatment include lim-

scan showing a 1.5-cm left retropharyngeal node with aradiation therapy (RT) shows a 0.9-cm retropharyngealntinues to show residual calcification.

tmentafter

ting the chance of curing the disease and making surgical

dRm4

iftcofipifnmdhacn(srl

tamscrfcfniipct

icrmcb

TTTTTTTTT

TTTTTTT

pt

imag

1020 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 4, 2006

issection more difficult as radiation fibrosis progresses afterT. For patients with neck disease who are treated with pri-ary RT, we evaluate the neck for post-RT neck dissectionweeks after RT to avoid these risks.It may be surprising that none of the 14 patients with an

ncomplete radiographic response experienced locoregionalailure, but radiographic abnormalities are likely to overcallhe true rate of pathologic involvement. Morrissey et al.orrelated radiographically positive RP nodes (size �8 mmr with focal lucency or enhancement) with pathologicndings on RP node dissection in 26 patients and found aositive predictive value of 33% (11). This number is sim-lar to the positive predictive values of 31% to 43% that weound on correlating radiographically abnormal cervicaleck nodes (size �1.5 cm or with focal lucency, enhance-ent, or calcification) with pathologic findings on neck

issection in 193 hemi-necks (13). Clinical control rates areigh in part because involved RP nodes are usually �2 cm,nd with adequate doses of radiation should be highlyurable; control rates of clinically involved, 2-cm neckodes treated with RT alone can be up to 90% or greater17). Because of the lack of RP node failure in this smalleries, it is not possible to define which, if any, post-RTadiographic abnormality (e.g., size, focal abnormality) isikely to be clinically important.

Table 2. Response of radiographically involved retro

Imaging cha

Stage and primary site Pre-RT study

2N3 tonsil 1.0 cm G4 luc4N2B tonsil 1.0 cm G3 luc2N3 base of tongue 0.5 cm G4 enh4N3 tonsil 1.1 cm3N2C tonsil 1.2 cm4N3 base of tongue 1.1 cm4N2B pharyngeal wall 1.4 cm3N2B base of tongue 1.8 cm G4 luc4N2C tonsil 1.7 cm (R)

1.3 cm (L)4N2B soft palate 2.6 cm with carotid encasement*4N2C base of tongue 0.9 cm G3 enh2N2B tonsil 1.7 cm G4 luc2N3 multiple synchronous 1.7 cm G4 luc3N2B tonsil 1.5 cm G4 calc and luc3N1 tonsil 1.1 cm4N2C tonsil 0.8 cm G4 calc (R)

Abbreviations: calc � focal calcification; enh � focal enhanceositive retropharyngeal nodes); luc � focal lucency; R � rightherapy.

* Only case in which imaging was done by magnetic resonance

There should be no question that adequate coverage of e

REFEREN

1938. p 44 –56.

he retropharyngeal nodes at risk is an important consider-tion when treating patients with head-and-neck cancer pri-ary RT (12). A previous report from this institution de-

cribes the outcome of 619 patients with head-and-neckancer treated with curative-intent RT, of whom 8% hadadiographically positive RP nodes (size �1 cm or withocal abnormality) (8). Multivariate analysis indicated de-reased neck control, distant control, and overall survivalor all primary head-and-neck sites analyzed, except theasopharynx, in patients with positive RP nodes. Althought is possible that the use of concurrent chemotherapy mayncrease the efficacy of RT and thereby slightly diminish therognostic value of radiographically positive RP nodes,overage of RP nodes should continue to be common prac-ice when at risk of involvement.

CONCLUSION

Radiographically positive RP nodes are highly controlledn patients treated with definitive RT for head-and-neckancer. Radiographic response at 4 weeks may not accu-ately reflect long-term locoregional control, and RP nodesay continue to regress over time. The most appropriate

riteria to diagnose positive RP nodes with CT after RT maye a progressive change in size, focal lucency, or focal

geal (RP) nodes to radiation therapy in 16 patients

stics

Post-RT study Response

egative – not measurable Completeegative – not measurable Completeegative – 0.5 cm Completeegative – 0.5 cm Completeegative – 0.6 cm Completeegative – 0.6 cm Completeegative – 0.7 cm Completeegative – 0.8 cm Completeegative – 0.9 cm (R) Completeegative – 0.9 cm (L)2 cm Incomplete by size6 cm G3 enh Incomplete by enh9 cm G4 luc Incomplete by luc6 cm G4 luc Incomplete by size and luc9 cm G4 calc Incomplete by calc8 cm G4 calc Incomplete by new calc8 cm G4 calc (R);9 cm G4 calc (L)

Incomplete by calc (and new calc)

G3 � Grade 3, G4 � Grade 4; L � left (in cases of bilaterallyes of bilaterally positive retropharyngeal nodes); RT � radiation

ing and not computed tomography.

nhancement on serial imaging.

CES

1. Rouviére H. Anatomy of the human lymphatic system.Tobias MJ, translator. Ann Arbor, MI: Edwards Brothers,

2. Mancuso AA, Harnsberger HR, Muraki AS, et al. Computedtomography of cervical and retropharyngeal lymph nodes:

pharyn

racteri

NNNNNNNNNN1.0.0.1.0.0.0.0.

ment;(in cas

Normal anatomy, variants of normal, and applications in stag-

1

1

1

1

1

1

1

1

1021Response of radiographically positive RP nodes ● S. L. LIAUW et al.

ing head-and-neck cancer. Part I: normal anatomy. Radiology1983;148:709–714.

3. Mancuso AA, Harnsberger HR, Muraki AS, et al. Computedtomography of cervical and retropharyngeal lymph nodes:Normal anatomy, variants of normal, and applications in stag-ing head-and-neck cancer. Part II: pathology. Radiology 1983;148:715–723.

4. Gregoire V, Coche E, Cosnard G, et al. Selection and delin-eation of lymph node target volumes in head and neck con-formal radiotherapy. Proposal for standardizing terminologyand procedure based on the surgical experience. RadiotherOncol 2000;56:135–150.

5. Chong VF, Fan YF. Radiology of the retropharyngeal space.Clin Radiol 2000;55:740–748.

6. Ballantyne AJ. Significance of retropharyngeal nodes in can-cer of the head and neck. Am J Surg 1964;108:500–504.

7. Ferlito A, Shaha AR, Rinaldo A. Retropharyngeal lymph nodemetastasis from cancer of the head and neck. Acta Otolaryngol2002;122:556–560.

8. McLaughlin MP, Mendenhall WM, Mancuso AA, et al. Ret-ropharyngeal adenopathy as a predictor of outcome in squa-mous cell carcinoma of the head and neck. Head Neck 1995;17:190–198.

9. Amatsu M, Mohri M, Kinishi M. Significance of retropha-ryngeal node dissection at radical surgery for carcinoma ofthe hypopharynx and cervical esophagus. Laryngoscope2001;111:1099 –1103.

0. Gross ND, Ellingson TW, Wax MK, et al. Impact of retro-

pharyngeal lymph node metastasis in head and neck squamous

cell carcinoma. Arch Otolaryngol Head Neck Surg2004;130:169–173.

1. Morrissey DD, Talbot JM, Cohen JI, et al. Accuracy ofcomputed tomography in determining the presence or absenceof metastatic retropharyngeal adenopathy. Arch OtolaryngolHead Neck Surg 2000;126:1478–1481.

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

3. Liauw SL, Mancuso AA, Amdur RJ. Postradiotherapy neckdissection for lymph node-positive head-and-neck cancer: Theuse of computed tomography to manage the neck. J Clin Oncol2006;24:1421–1427.

4. Hasegawa Y, Matsuura H. Retropharyngeal node dissection incancer of the oropharynx and hypopharynx. Head Neck 1994;16:173–180.

5. Department of Veterans Affairs Laryngeal Cancer StudyGroup. Induction chemotherapy plus radiation compared withsurgery plus radiation in patients with advanced laryngealcancer. N Engl J Med 1991;324:1685–1690.

6. Forastiere AA, Goepfert H, Maor M, et al. Concurrent che-motherapy and radiotherapy for organ preservation in ad-vanced laryngeal cancer. N Engl J Med 2003;349:2091–2098.

7. Mendenhall WM, Million RR, Cassisi NJ. Squamous cellcarcinoma of the head and neck treated with radiation therapy:The role of neck dissection for clinically positive neck nodes.

Int J Radiat Oncol Biol Phys 1986;12:733–740.