Int. J. Radiation Oncology Biol. Phys., Vol. 53, No. 5, pp.

1174 –1184, 2002

Copyright © 2002 Elsevier Science Inc.
Printed in the USA. All rights reserved
0360-3016/02/$–see front matter

PII S0360-3016(02)02881-X

CLINICAL INVESTIGATION Head and Neck

DETERMINATION AND DELINEATION OF NODAL TARGET VOLUMES FOR

HEAD-AND-NECK CANCER BASED ON PATTERNS OF FAILURE IN
PATIENTS RECEIVING DEFINITIVE AND POSTOPERATIVE IMRT

K. S. CLIFFORD CHAO, M.D.,* FRANZ J. WIPPOLD, II, M.D.,† GOKHAN OZYIGIT, M.D.,*
BINH N. TRAN, M.D.,* AND JAMES F. DEMPSEY, PH.D.‡
Departments of *Radiation Oncology and †Radiology, Washington University Medical School, St. Louis, MO; ‡Department of
Radiation Oncology, University of Florida, Miami, FL

Purpose: We present the guidelines for target volume determination and delineation of head-and-neck lymph
nodes based on the analysis of the patterns of nodal failure in patients treated with intensity-modulated
radiotherapy (IMRT).
Methods and Materials: Data pertaining to the natural course of nodal metastasis for each head-and-neck cancer
subsite were reviewed. A system was established to provide guidance for nodal target volume determination and
delineation. Following these guidelines, 126 patients (52 definitive, 74 postoperative) were treated between
February 1997 and December 2000 with IMRT for head-and-neck cancer. The median follow-up was 26 months
(range 12–55), and the patterns of nodal failure were analyzed.
Results: These guidelines define the nodal target volume based on the location of the primary tumor and the
probability of microscopic metastasis to the ipsilateral and contralateral (Level I–V) nodal regions. Following
these guidelines, persistent or recurrent nodal disease was found in 6 (12%) of 52 patients receiving definitive
IMRT, and 7 (9%) of 74 patients receiving postoperative IMRT had failure in the nodal region.
Conclusion: On the basis of our clinical experience in implementing inverse-planning IMRT for head-and-neck
cancer, we present guidelines using a simplified, but clinically relevant, method for nodal target volume
determination and delineation. The intention was to provide a foundation that enables different institutions to
exchange clinical experiences in head-and-neck IMRT. These guidelines will be subject to future refinement when
the clinical experience in head-and-neck IMRT advances. © 2002 Elsevier Science Inc.

Target definition, IMRT, Imaging-based, Head-and-neck cancer.

INTRODUCTION these emerging technologies requires training and expertise

available at only a handful of academic institutions.
Recent advances in radiotherapy (RT) planning and delivery Recent articles have reviewed the normal anatomy of the
technology have made imaging-based RT a reality. Prom-
neck (1– 4). Although these articles provide information on
ising treatment results of head-and-neck and prostate con-
the patterns of lymph node metastasis for carcinoma of the
formal RT or intensity-modulated RT (IMRT), as well as
head and neck, these materials were mainly focused on the
the approval of a special IMRT billing code by the Health
delineation of lymph node regions or levels in the normal
Care Financing Administration, have provided significant
incentives for the radiation oncology community to incor- neck. However, in clinical practice, we often encounter
porate this imaging-based technology into daily clinical significant distortion of normal head-and-neck anatomy ow-
practice. Protecting critical normal tissue without compro- ing to the existence of a large mass in the neck or a previous
mising tumor target coverage requires extensive knowledge operative dissection that limits the application of these
of the patterns of tumor extension and spread and the ability articles’ findings in clinical practice. In this paper, we
to delineate accurately both the tumor target volumes and present guidelines, based on our clinical experience, to
the normal structures. Use of CT or MRI has created a assist radiation oncologists in determining the target vol-
virtual operative suite in which crucial anatomic informa- umes in clinically and radiologically positive (N⫹) or neg-
tion can be explored noninvasively on a computer screen. ative (N⫺) lymph nodes in the neck for definitive IMRT.
Nevertheless, familiarity and appropriate application of We also provide guidelines for target volume delineation in

Reprint requests to: K. S. Clifford Chao, M.D., Department of Acknowledgments—The authors are in debt to Elaine Pirkey for
Radiation Oncology, Box 97, M. D. Anderson Cancer Center, preparing the manuscript, and also thank Drs. Walter Bosch and
1515 Holcombe Boulevard, Houston, TX 77030. Tel: (713) 792- James Purdy for their support and encouragement for this study.
3400; Fax: (713) 745-6994; E-mail: cchao@mdanderson.org Received Jun 8, 2001, and in revised form Mar 22, 2002.
This work was supported in part by NIH Grant CA-89198. Accepted for publication Apr 3, 2002.

1174
 Nodal target definition for head-and-neck IMRT ● K. S. C. CHAO et al. 1175

Table 1. Classification and definition of neck nodes

Robbins’ classification

Level Terminology Definition (surgical/anatomic landmarks)

Ia Submental group Contains submental and submandibular triangles bounded by posterior belly of
Ib Submandibular group digastric muscle, hyoid bone inferiorly, and body of mandible superiorly
II Upper internal jugular group Contains upper internal jugular lymph nodes and extends from level of hyoid
bone inferiorly to skull base superiorly
III Middle internal jugular group Contains middle internal jugular lymph nodes from hyoid bone superiorly to
cricothyroid membrane inferiorly
IV Lower internal jugular group Contains lower internal jugular lymph nodes from cricothyroid membrane
superiorly to clavicle inferiorly
V Spinal accessory group Contains lymph nodes in posterior triangle bounded by anterior border of
trapezius posteriorly, posterior border of sternocleidomastoid muscle
anteriorly, and clavicle inferiorly (for descriptive purposes, Level V may be
further subdivided into upper, middle, or lower levels corresponding to the
superior and inferior planes that define Levels II, III, and IV)
VI Anterior compartment group Contains lymph nodes of anterior compartment from hyoid bone superiorly to
suprasternal notch inferiorly; on each side, the lateral border is formed by
the medial border of the carotid sheath
VII Upper mediastinal group Contains lymph nodes inferior to suprasternal notch in upper mediastinum

Other groups include retropharyngeal, buccinator (facial), intraparotid, preauricular, postauricular, suboccipital.

postoperative settings when the pathologic examination re- disease. The pathologic nodal status of these 74 patients was
veals either the presence or absence of extracapsular exten- N0 in 18 patients, N1 in 13 patients, N2 in 37 patients (10
sion (ECE⫹ or ECE⫺, respectively). N2a, 19 N2b, 8 N2c), and N3 in 6 patients. ECE was present
in 32 patients. Chemotherapy was not routinely given in the
postoperative setting; however, 5 patients with extensive
METHODS AND MATERIALS nodal involvement in the lower neck received adjuvant
Patient and tumor characteristics cisplatin-based chemotherapy at the treating physician’s
Between February 1997 and December 2000, 165 head- discretion. If no residual disease was present after definitive
and-neck cancer patients were treated using an IMRT tech- IMRT, patients were given no additional therapy and re-
nique. Patients included in this parotid-sparing IMRT pro- mained in routine follow-up. In patients with recurrent or
tocol were those with histologically confirmed head-and- persistent disease, salvage therapy was individually deter-
neck cancer who required either definitive or postoperative mined. In general, when patients presented without distant
RT. Thirty-nine patients receiving either palliative reirra- metastasis or were medically suitable, surgery was the pri-
diation or IMRT as a boost were excluded from this anal- mary mode of salvation therapy. Otherwise, chemotherapy
ysis. The remaining 126 eligible patients (30 females, 96 or supportive measures were instituted.
males) without evidence of distant metastasis at presenta-
tion formed the body of this analysis. The median age was Determination of clinical target volumes
56 years (range 13– 84). The primary tumor was located at We used similar terminology to those proposed by the
the nasopharynx in 12 patients, paranasal sinuses or nasal Robbins classification (5) for the determination of clinical
cavity in 9 patients, oral cavity in 15 patients, oropharynx in target volumes (CTVs) (Table 1). The determination of
63 patients, supraglottic larynx in 8 patients, hypopharynx CTVs was based on the incidence and location of metastatic
in 8 patients, and other regions of the head and neck in 3 neck nodes from various head-and-neck subsites, which
patients and was unknown in 9 patients. Fifty-two patients were gathered from the published literature and are summa-
received definitive IMRT, 4 with Stage II disease, 9 with rized in Table 2. As shown in Table 2, the distribution of
Stage III disease, and 39 with Stage IV disease. The clinical/ nodal metastasis to different lymph node levels in the head-
radiologic nodal status of these 52 patients was N0 in 12 and-neck region varies by primary tumor subsite. The num-
patients, N1 in 9 patients, N2 in 24 patients (9 N2a, 11 N2b, ber under each column represents the percentage of lymph
4 N2c), and N3 in 7 patients. Among them, 17 refused node metastasis in patients with squamous cell carcinoma
chemotherapy and were treated with RT alone, and the arising from various head-and-neck subsites. Because met-
remaining 35 were treated with concurrent chemotherapy astatic nodes may manifest in more than one nodal level at
per an intramural protocol. Chemotherapy was a cisplatin- presentation, the summation of the percentage from all
based regimen in all cases. Seventy-four patients received nodal levels (regions) may exceed 100%, especially in the
postoperative IMRT: 5 with Stage I disease, 4 with Stage II N⫹ group.
disease, 17 with Stage III disease, and 39 with Stage IV Treatment of the contralateral neck remains controver-
1176 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 5, 2002

Table 2. Incidence and distribution of metastatic disease in clinically negative and positive neck nodes

Radiologically
Pathologic nodal metastasis (%)
enlarged
retropharyngeal
nodes (%) Level I Level II Level III Level IV Level V

Clinical presentation N⫺ N⫹ N⫺ N⫹ N⫺ N⫹ N⫺ N⫹ N⫺ N⫹ N⫺ N⫹

Nasopharynx 40 86 — — — — — — — — — —
Oral cavity
Oral tongue — — 14 39 19 73 16 27 3 11 0 0
Floor of mouth — — 16 72 12 51 7 29 2 11 0 5
Aveolar ridge and
RMT — — 25 38 19 84 6 25 5 10 1 4
Oropharynx
Base of tongue 0 6 4 19 30 89 22 22 7 10 0 18
Tonsil 4 12 0 8 19 74 14 31 9 14 5 12
Hypopharynx
Pharyngeal wall 16 21 0 11 9 84 18 72 0 40 0 20
Pyriform sinus 0 9 0 2 15 77 8 57 0 23 0 22
Larynx
Supraglottic larynx 0 4 6 2 18 70 18 48 9 17 2 16
Glottic larynx — — 0 9 21 42 29 71 7 24 7 2

Abbreviations: N ⫽ node; RMT ⫽ retromolar trigone.

Data from McLaughlin et al., 1995 (22), Candela et al., 1990 (23), Shah et al., 1990 (24), Bataini et al., 1985 (25), Byers et al., 1988
(26), and Lindberg, 1972 (27).

sial, because very few data are available on the patterns of pharyngeal wall, and pyriform sinus. The probability of
pathologic node distribution in the contralateral neck and contralateral nodal metastasis can be predicted with better
also treatment likely results from clinical judgment rather accuracy if these tumor characteristics are taken into ac-
than scientific evidence. However, contralateral nodal re- count. Table 4 shows the tumor factors of oral cavity
gions should be included, especially in tumors that tend to carcinoma that could influence the incidence of contralateral
spread to the contralateral neck nodes or in tumor arising nodal metastasis. On the other hand, tumors arising from the
from or invading to a midline structure such as the soft true vocal cord, paranasal sinuses, and middle ear have a
palate, base of tongue, posterior pharyngeal wall, or naso- low risk of lymph node metastasis, and only the ipsilateral
pharynx. For example, nodal metastasis exists in 85–90% of neck needs to be included in the IMRT field (1). If a tumor
patients with nasopharyngeal carcinoma, and about 50% of arises from buccal mucosa and retromolar trigone, which
them have bilateral disease; therefore, both sides of the neck has a lower chance of contralateral neck node metastasis,
need to be treated (6). Table 3 shows the incidence of especially when the primary tumor size is small and no
macroscopic or microscopic bilateral nodal metastases
present with ⬎30% of tumor residing in the base of tongue, Table 4. Factors influencing contralateral lymph node metastasis
in oral cancer

Table 3. Incidence of contralateral or bilateral neck node RR of contralateral

metastasis by primary tumor site Variable LN metastasis 95% CL

cN⫹, cN⫹, cN⫺, Tumor site

bilateral contralateral pN⫹, Tongue 1.0 Reference
(%) only (%) bilateral (%) FOM 1.5 0.9–2.6
RMT 0.3 0.1–1.1
Oral tongue 12 — 33 Distance from midline (cm)
FOM 27 — 21 ⬎1 1.0 Reference
BOT 37 — 55 Cross ⬍1 2.8 1.1–7.5
Tonsil 16 2 — Cross ⬎1 12.7 5.6–29.1
Pharyngeal wall 50 — 37 Tumor stage
Pyriform sinus 49 6 59 T1 1.0 Reference
Supraglottis 39 2 26 T2–3 2.2 0.7–5.5
Glottic larynx — — 15 T4 5.8 2.0–16.3

Abbreviations: BOT ⫽ base of tongue; FOM ⫽ floor of mouth. Abbreviations: RR ⫽ relative risk; LN ⫽ lymph node; CI ⫽
Data from Northrop et al., 1972 (28), Bataini et al., 1985 (25), confidence interval; RMT ⫽ retromolar trigone; FOM ⫽ floor of
Byers et al., 1988 (26), Woolgar, 1999 (29), and Buckley and mouth.
MacLennan, 2000 (30). Modified from Kowalski and Medina, 1998 (31), with permission.
 Nodal target definition for head-and-neck IMRT ● K. S. C. CHAO et al. 1177

Table 5. Target volume specification for definitive and postoperative IMRT

Target Definitive IMRT High-risk postoperative IMRT Intermediate-risk postoperative IMRT

CTV1 Gross tumor and adjacent Surgical bed with soft tissue involvement Surgical bed without soft tissue involvement
soft tissue/nodal regions or nodal region with extracapsular or nodal region without extracapsular
extension extension
CTV2 Elective nodal regions* Elective nodal regions* Elective nodal regions*

* Suggested guidelines for elective nodal regions shown in Table 6.

Abbreviations: IMRT ⫽ intensity-modulated radiotherapy; CTV ⫽ clinical target volume.

involvement of the ipsilateral neck node is evident, the signed on the basis of specific soft tissue landmarks for
contralateral neck may not need to be treated. surgical procedures and are not easily seen on CT and MRI
Having the tumor characteristics and clinical/pathologic slices; we implemented modified guidelines for the delin-
information in mind, the following guidelines, using a sim- eation of the various node levels in the neck (Table 7). The
plified, but clinically relevant, method for target volume recent recommendations proposed by several authors on the
determination and delineation, were implemented. The locations of the surgical neck compartments were also in
CTV1 for postoperative IMRT patients encompasses the good agreement with our system (1, 3, 9). Using our guide-
preoperative gross tumor volume (GTV) plus a 2-cm mar- lines, all margins of each specific neck node level could be
gin, including the resection bed with soft-tissue invasion by demarcated on axial CT sections.
the tumor or ECE by metastatic neck nodes truncating air Figures 1 and 2 depict the nodal target volume delineation at
and uninvolved bones. The surgical bed was determined by different levels of the neck in patients receiving definitive and
preoperative CT imaging, surgical defects, or postoperative postoperative IMRT. We operatively demarcated the nodal
changes seen on the postoperative CT scan. CTV1 for GTV in Fig. 1 to provide readers visual assistance in under-
definitive IMRT patients encompasses the gross tumor and standing the location of gross nodal disease and the corre-
region adjacent to the gross tumor but not directly involved sponding design of CTV1 and CTV2. Sparing salivary gland
by tumor based on clinical findings and CT or MRI imaging. function is important in preserving the quality of life of pa-
Radiologically or clinically involved neck nodes were also tients. The literature has shown a dose response of parotid
included in CTV1 with 2-cm margins truncating air and gland function after RT (10–13). We elected not to spare the
bone. deep lobe of the parotid gland to prevent marginal failure in the
CTV2 for both definitive and postoperative IMRT groups parapharyngeal space. Therefore, only the superficial lobes of
primarily includes the clinically/radiologically or patholog- parotid glands were demarcated in Fig. 1A. Figure 2 differen-
ically uninvolved cervical lymph nodes, deemed as elective tiates ECE⫹ and ECE⫺ necks in patients receiving postoper-
nodal regions or prophylactically treated neck. Our concept ative IMRT. In the ECE⫹ neck, soft tissues are included more
of selective neck treatment was similar to the recently generously and the CTV1 needs to extend close to the skin
published recommendations proposed by Gregoire et al. (1). surface, especially in the region or level of the ECE⫹
These target volume specifications were integrated with the nodes specified by the pathologic examination (Fig.
published clinical data shown in Table 2. On the basis of 2A,C,E). When postoperative IMRT is needed for the
these historical data, we proposed that a treatment of the N0 ECE⫺ neck, the target volume should avoid the skin
neck is warranted if the probability of occult cervical me- surface to decrease the acute dermal toxicity. In our
tastasis is ⬎5%. The target volume (CTV1 and CTV2) experience, sparing 2–3 mm of dermal structures in the
specifications for various head-and-neck tumor subsites for target volume design (Fig. 2B,D,F) results in much better
definitive and postoperative IMRT are summarized in Ta- radiation tolerance, fewer treatment breaks, and no com-
bles 5 and 6. promise in locoregional control. Although the lower neck
To facilitate clinical throughput, at our institution, IMRT was treated with conventional techniques in most pa-
was applied to the upper neck for salivary gland sparing. tients, similar principles were applied to depict the CTV
The lower neck was treated with a conventional AP lower delineation in the lower neck region for the reader’s
neck port if indicated. The standard superior border for the reference in Figs. 1 and 2. Also, it needs to be empha-
lower neck field was at the level of the thyroid notch. A sized that margins for organ motion or patient setup error
similar approach has also been implemented elsewhere (7). were not included in delineating the target volume, be-
In patients with tumor or metastatic lymph node extending cause they need to be determined by individual institu-
below this level, the junction line was adjusted to avoid tions implementing a head-and-neck IMRT program. Us-
bisecting gross disease. ing a reinforced thermoplastic mask for immobilization,
our prior study indicated that a 3-mm margin was needed
Delineation of CTV for IMRT plan computation to count for patient setup
Because the definition of neck node level and anatomic uncertainty (8, 14).
boundaries described in the Robbins classification was de- Table 8 summarizes the corresponding dose prescrip-
1178 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 5, 2002

Table 6. Clinical target volume determination for head-and-neck IMRT

Tumor site Clinical presentation CTV1 CTV2

Oral cavity
Buccal T1–2N0 P IN (I–III)
RMT T3–4N* P ⫹ IN (I–III) CN (I–III)
N2c P ⫹ IN ⫹ CN (I–V)
Oral tongue T1–2N0 P IN ⫹ CN (I–IV)
T3–4N* P ⫹ IN (I–IV) CN (I–IV)
N2c P ⫹ IN ⫹ CN (I–V)
FOM T1–2N0 P IN ⫹ CN (I–III)
T3–4N* P ⫹ IN (I–III) CN (I–III)
N2c P ⫹ IN ⫹ CN (I–V)
Oropharynx
BOT T1–2N0 P IN ⫹ CN (II–IV, RPLN)
T3–4N* P ⫹ IN (II–IV, RPLN) CN (I–V, RPLN)
N2c P ⫹ IN ⫹ CN (I–V, RPLN)
Tonsil T1–2N0 P IN ⫾ CN (II–IV, RPLN)
T3–4N* P ⫹ IN (II–IV, RPLN) CN (I–V, RPLN)
N2c P ⫹ IN ⫹ CN (I–V, RPLN)
Hypopharynx T1–2N0 P IN ⫹ CN (II–IV)
T3–4N* P ⫹ IN (II–IV, RPLN) CN (II–IV)
N2c P ⫹ IN ⫹ CN (I–V, RPLN)
Larynx† T1–2N0 P IN ⫹ CN (II–IV)
T3–4N* P ⫹ IN (II–IV) CN (II–IV)
N2c P ⫹ IN ⫹ CN (I–V)
Nasopharynx T1–2N0 P IN ⫹ CN (I–V, RPLN)
T3–4N* P ⫹ IN (I–V, RPLN) CN (I–V, RPLN)
N2c P ⫹ IN ⫹ CN (I–V, RPLN)

Abbreviations: IMRT ⫽ intensity-modulated radiotherapy; CTV ⫽ clinical target volume; P ⫽ gross tumor with margins for definition
IMRT or surgical bed for postoperative IMRT; IN ⫽ ipsilateral neck nodes (level); CN ⫽ contralateral neck nodes (level); N* ⫽ N1–3
except N2c; RMT ⫽ retromolar trigone; BOT ⫽ base of tongue; FOM ⫽ floor of mouth; RPLN ⫽ retropharyngeal lymph nodes.
†
T1–2 carcinoma of the true vocal cord excluded.

tions for CTV1 and CTV2 (8). CTV1 was considered a patients was 70.23 ⫾ 3.44 Gy to CTV1 and 60.15 ⫾ 2.87
higher risk volume, and a higher dose was given to this Gy to CTV2. The mean dose to CTV1 and CTV2 in 74
target volume. Using these guidelines, the radiation dose postoperative cases was 65.05 ⫾ 4.21 Gy and 57.78 ⫾
(mean ⫾ standard deviation) for 52 definitive IMRT 5.58 Gy, respectively.

Table 7. Recommendations for radiologic boundaries of neck node regions

Level Recommended boundary

I Cranial: Mylohyoid muscle; Caudal: Hyoid bone

Anterior: Symphisis menti; Posterior: Posterior edge of submandibular gland
Lateral: Medial edge of mandible; Medial: Lateral edge of anterior belly of digastric muscle
II Cranial: Cranial base; Caudal: Bottom edge of hyoid bone
Anterior: Posterior margin of submandibular gland; Posterior: Posterior edge of SCM
Lateral: Medial edge of SCM; Medial: Medial edge of vessel bundle, paraspinal muscle
III Cranial: Cranial: Bottom edge of hyoid bone; Caudal: Lower margin of cricoid cartilage
Anterior: Posterlateral edge of sternohyoid muscle; Posterior: Posterior edge of SCM
Lateral: Medial edge of SCM; Medial: Medial edge of vessel bundle, paraspinal muscle
IV Cranial: Lower margin of cricoid cartilage; Caudal: Cranial border of clavicle
Anterior: Posterolateral edge of SCM muscle; Posterior: Anterior edge of paraspinal muscle
Lateral: Lateral border of SCM; Medial: Medial border of vessel bundle, lateral border of thyroid
V Cranial: Cranial base; Caudal: Cranial border of clavicle
Anterior: Posterior edge of SCM; Posterior: Anterior edge of trapezius muscle
Lateral: Platysma muscle, skin; Medial: Paraspinal muscle
Retropharyngeal Cranial: Base of skull; Caudal: Cranial edge of body of hyoid bone
Anterior: Levator veli palatini; Posterior: Prevertebral muscles
Lateral: Medial edge of vessel bundle; Medial: Midline

Abbreviations: SCM ⫽ sternocleidomastoid muscle; Vessel bundle ⫽ internal carotid artery and internal jugular vein.
 Nodal target definition for head-and-neck IMRT ● K. S. C. CHAO et al. 1179

Fig. 1. Axial-enhanced CT scans at level of (A) pterygoid plates, (B) mandible, (C) submandibular gland, (D) hyoid
bone, (E) thyroid cartilage, and (F) cricoid cartilage in a patient with metastatic head-and-neck cancer. CTVs with the
presence of metastatic lymphadenopathy compared with those without radiologic evidence of metastatic neck nodes.
Gross tumor operatively demarcated to provide readers with visual assistance in understanding location of gross nodal
disease and corresponding target volume. CTV1, red line; CTV2, blue line. Ib ⫽ level Ib node; II ⫽ level II node; III ⫽
level III node; V ⫽ level V node; N⫹ ⫽ positive nodes; N⫺ ⫽ negative nodes; NR ⫽ nodes of Rouviere; GTV ⫽
grossly enlarged lymph node.

Analysis of nodal failure after IMRT treatment of definitive IMRT. To analyze the patterns of failure,
The median follow-up was 26 months (range 12–55). recurrent or persistent disease was defined on CT or MRI
Persistent disease was defined as histopathologically or by surgical/pathologic findings. Disease in patients for
proven residual disease within 6 months after completion whom radiologic imaging was available at the time of
1180 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 5, 2002

Fig. 2. Axial enhanced postoperative CT scans at level of (A,B) mandible, (C,D) thyroid cartilage, and (E,F) thyroid
gland in a patient with metastatic head-and-neck cancer. CTVs with presence of ECE (A,C,E) compared with volumes
without ECE (B,D,F). CTV1, red line; CTV2, blue line. GTV ⫽ grossly enlarged lymph node.

recurrence was co-registered with the treatment planning works, Matlab) to analyze treatment failures and catego-
CT dataset using a commercial virtual simulation work- rize the failures as (1) “in-field,” if ⬎95% of disease
station (Marconi, VoxelQ). Dose–volume histograms of volume was within either CTV1 or CTV2, (2) “margin-
failures within the IMRT field (excluding low neck re- al,” if 20 –95% of the disease volume was within CTV1
currence) were calculated on an 8 mm3 isotropic voxel or CTV2; and (3) “out-field,” if ⬍20% of the disease
grid using commercial data analysis software (Math- volume was within either CTV1 or CTV2.
 Nodal target definition for head-and-neck IMRT ● K. S. C. CHAO et al. 1181

Table 8. IMRT clinical target volume and normal tissue dose specification with biologic equivalent dose correction for head-and-neck
cancer—Washington University guideline

IMRT (Gy)

Conventional Intermediate-risk
technique High-risk postoperative postoperative
Target volume (Gy) Definitive (35 fractions) (33 fractions) (30 fractions)

CTV1 66–70/2 70/2 66/2 60/2

CTV2 50–54/2 56/1.66 54/1.64 52/1.73

Normal tissue tolerance for IMRT prescription: Optic nerve and optic chiasm 55 Gy, retina 45 Gy, brainstem 50 –55 Gy, spinal cord
45– 48 Gy, parotid gland 20 –30 Gy, mandible 70 Gy.
Abbreviations: IMRT ⫽ intensity-modulated radiotherapy; CTV ⫽ clinical target volume.

RESULTS tumors showed evidence of lymphovascular space invasion.

One recurred within CTV2 and another recurred in the
Patterns of failure analysis
lower neck.
Using these guidelines for nodal target volume determi-
nation and delineation, persistent or recurrent nodal disease
was found in 6 (12%) of 52 patients receiving definitive DISCUSSION
IMRT. Four failures were “in-field” to CTV1 and two were
in the lower neck outside the IMRT volume. Seven (9%) of In 1948, Rouviere (15) described the anatomic details of
74 patients receiving postoperative IMRT had failure in the the cervical lymphatic network. On the basis of this descrip-
nodal region. One failure was “marginal” to CTV1 but tion, the TNM atlas proposed a terminology that divides the
“in-field” to CTV2. Two failures were marginal to CTV2. head-and-neck lymph nodes into 12 groups according to
Two failures were in the lower neck and outside the IMRT their relationship with the adjacent muscles, vessels, and
field. Another two failures were in CTV1 but one had also nerves (16). In 1991, a Committee for Head and Neck
lower neck failure that was outside the IMRT field. The Surgery and Oncology of the American Academy for Oto-
detailed clinical characteristics are summarized in Table 9. laryngology–Head and Neck Surgery postulated a classifi-
Three of 6 nodal failures in the definitive IMRT group were cation (Robbins’ classification) that divides the neck into 6
salvaged with definitive surgery. Two of 7 patients with levels or 8 nodal groups for those lymph nodes routinely
failure after postoperative IMRT were also salvaged with removed during neck dissection (5). Lymph nodes that are
surgery. The remaining patients with neck failure were not routinely dissected, such as the retropharyngeal, parotid,
given either supportive treatment or chemotherapy, because buccal, or occipital nodes, are not included in Robbins’
they were not suitable for salvage surgery. There were 18 classification. Both systems describe the boundaries of the
patients with pathologic N0 neck. Nevertheless, the neck node region on the basis of anatomic structures, such as the
was prophylactically treated because 10 (63%) of 18 tumors major blood vessels, muscles, nerves, bones and cartilage,
were locally advanced (T3 or T4) and the remaining 8 and the radiologic boundaries of these nodal levels have

Table 9. Clinical characteristics and patterns of nodal failure of 13 patients treated with head-and-neck IMRT

Pt. No. ECE status T N AJCC Aim Site of failure Location

1 NA 4 1 IV Def R Level II LN CTV1

2 NA 4 2C IV Def Bilateral neck CTV1
3 NA 4 3 IV Def R Level II LN CTV1
4 NA 3 3 IV Def Lower neck Lower neck
5 NA 4 1 IV Def L Level II LN CTV1
6 NA 2 2A IV Def Lower neck Lower neck
7 ⫺ 2 2 IV Postop R Level III LN Lower neck
8 ⫺ 3 0 IV Postop L Level II LN CTV2
9 ⫹ 3 2C IV Postop R Level IB LN CTV2
10 ⫹ 3 2B IV Postop L Level II LN CTV2
11 ⫹ 4 2B IV Postop Thyroid cartilage L CTV1 and
Level IA LN Lower neck
12 ⫹ X 2C — Postop R Level V Neck Posterior neck
13 ⫺ 2 0 II Postop L Level III/IV LN Lower neck

Abbreviations: IMRT ⫽ intensity-modulated radiotherapy; Postop ⫽ postoperative; Def ⫽ definitive; L ⫽ left; R ⫽ right; LN ⫽ lymph
node; ECE ⫽ extracapsular extension; NA ⫽ not applicable.
1182 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 5, 2002

Table 10. Sensitivity and specificity of CT and MRI in detecting clinically negative but pathologically positive neck nodes

Author Modality Patients (n) Sensitivity (%) Specificity (%)

Stern et al., 1990 (32) CT 53 40 92

Friedman et al., 1990 (33) CT 68 68 90
MRI 16 80 82
Moreau et al., 1990 (34) CT 32 50 86
van den Brekel et al., 1993 (35) CT 49 78 86
MRI 55 88 83
Righi et al., 1997 (36) CT 32 50 86

been recently summarized (1– 4, 9). We added the retropha- will continue to advance our understanding of tumor exten-
ryngeal nodal group to Robbins’ classification to assist sion and assist in the delineation of the true GTV; however,
readers in a better understanding of the nodal target volume it is an evolving area of research and beyond the scope of
determination and delineation for head-and-neck IMRT. We this report.
also introduced our institutional recommendations for the Differing from previously published atlases of target def-
radiologic boundaries of these nodal levels in Table 8. inition of the normal neck (1– 4), we depict examples of
either patients with grossly enlarged neck nodes for defin-
Target volume determination based on clinical/pathologic itive IMRT or those who have undergone neck dissection
data and require postoperative irradiation because of the pres-
A thorough understanding of the natural course of tumor ence of various pathologic risk factors. One significant
spread ensures the delineation of the CTV that represents pathologic factor to take into consideration for target vol-
the region potentially containing microscopic disease, as ume delineation is the presence of ECE. The probability of
defined in the International Commission on Radiation Units tumor extending outside the nodal capsule increases as a
and Measurements Reports 50 and 62 (17, 18). Ideally, if an function of tumor size. When metastatic nodal disease ex-
imaging modality can provide sufficient information on pands and ruptures the capsule of cervical lymph nodes, the
whether certain nodal regions contain micrometastasis, the incidence of local recurrence increases. Huang et al. (20)
accurate determination of nodal target volume for head-and- demonstrated a higher tumor recurrence in patients with
neck IMRT will be possible. Unfortunately, neither physical ECE, and postoperative RT improved locoregional control.
examination nor radiologic imaging techniques used in clin- Peters et al. (21) defined the resected neck into high and
ical practice are proficient to detect microscopic disease. low-risk groups to which different radiation doses were
Sako et al. (19) found that to be clinically detectable the recommended. Therefore, when delineating the target vol-
submandibular nodes must measure ⱖ0.5 cm in size. Sim- ume in the postoperative neck, inclusion of generous soft
ilarly, a deep cervical node located adjacent to muscles must tissue margins around the tumor bed is imperative. Should
exceed 1 cm in diameter to be clinically palpable. Notably, the information regarding which lymph node levels contain-
the incidence of occult nodal metastasis ranged from 25% to ing ECE⫹ nodes not be available pathologically, a preop-
60% in the 1950 –1960s. Even with advances in morphol- erative imaging study (CT or MRI) can assist in determining
ogy-based imaging techniques, such as CT and MRI, deter- which regions require a more generous soft tissue margin
mination of nodal metastasis based on the size of the lymph for CTV1 delineation.
node still underestimates between 12% and 60% of micro- The dilemma comes when definitive IMRT is used to
metastasis (Table 10). Detection of micrometastasis by treat an undissected neck, and no pathologic information is
functional imaging is an evolving area of research and may available to determine whether metastatic disease has ex-
not be applicable clinically in the immediate future. There- tended outside the lymph node capsules. In this case, we
fore, the current clinical practice to determine the target look to surgical pathologic experience for guidance. Table
volume for IMRT relies on historical information from 11 summarizes the incidence of ECE in various sizes of
surgical pathologic experiences. On the basis of the natural lymph nodes that contain metastatic disease. When the
course of nodal spreading from the primary tumor in each nodal size is as small as 1 cm, 17– 40% may have broken
head-and-neck subsite and the analysis of nodal failure in through the capsule. When the size of the metastatic node
126 patients treated with IMRT, we found that the guide- exceeds 3 cm, ⬎75% have ECE. This information is perti-
lines for nodal target volume determination and delineation nent to target volume design, because additional soft tissue
as outlined in Tables 5 through 7 are satisfactory. margins around the whole nodal level in which grossly
enlarged nodes reside need to be included in CTV1, which
Issues pertaining to target volume delineation usually provides at least 2-cm margins around the gross
We acknowledge that accurate delineation of the GTV disease, truncating air and bone (Fig. 1). The CTVs we used
based on the true extension of viable gross tumor is crucial. for the most part were generous and likely contributive to
The investigations to excel functional imaging techniques the high control rates. Because head-and-neck IMRT is still
 Nodal target definition for head-and-neck IMRT ● K. S. C. CHAO et al. 1183

Table 11. Incidence of extracapsular extension of metastatic clinical experience and knowledge continue to advance,
neck node by size these guidelines will be adjusted accordingly.
Node size (cm)

Author ⬍1 1–3 ⬎3 CONCLUSION

Annyas et al., 1979 (37) 23 53 74 Derived from surgical pathologic information in the lit-
Johnson et al., 1981 (38) — 65 75 erature and validated with our clinical IMRT experience, we
Carter et al., 1987 (39) 17 83 95 present guidelines to assist in determining and delineating
Hirabayashi et al., 1991 (40) 43 — 81 the nodal target volume for patients receiving definitive or
Data presented as the percentage.
postoperative head-and-neck IMRT. The intention was to
provide a foundation that enables different institutions to
exchange clinical experiences in head-and-neck IMRT.
in its infancy, we elected to be generous in the target volume These guidelines will be refined further as clinical experi-
delineation to avoid undesirable marginal failure. As the ence in head-and-neck IMRT advances.

REFERENCES
1. Gregoire V, Coche E, Cosnard G. Selection and delineation of cinoma: A comparison with conventional techniques. Ra-
lymph node target volumes in head and neck conformal ra- diother Oncol 2001:61:275–280.
diotherapy: Proposal for standardizing terminology and pro- 14. Low DA, Chao KS, Mutic S, et al. Quality assurance of serial
cedure based on the surgical experience. Radiother Oncol tomotherapy for head and neck patient treatments. Int J Radiat
2000:56:135–150. Oncol Biol Phys 1998;42:681– 692.
2. Martinez-Monge, Fernandes PS, Gupta N. Cross-sectional 15. Rouviere H. Anatomie humaine descriptive et topographique.
nodal atlas: A tool for the definition of clinical target volumes 6 ed. Paris: Masson et Cie; 1948.
in three-dimensional radiation therapy planning. Radiology 16. Spiessl B, Beahrs OH, Hermanek P. Illustrated guide to the
1999;211:815– 828. TNM/pTNM classification of malignant tumours. 2nd ed. Ber-
3. Nowak PJ, Wijers OB, Lagerwaard FJ, et al. A three-dimen- lin: Springer; 1992.
sional CT-based target definition for elective irradiation of the 17. International Commission on Radiation Units and Measure-
neck. Int J Radiat Oncol Biol Phys 1999;45:33–39. ments. ICRU report 50: Prescribing, recording, and reporting
4. Wijers OB, Levendag PC, Tan T, et al. A simplified CT-based photon beam therapy (supplement to ICRU report 50). Wash-
definition of the lymph node levels in the node negative neck. ington, DC: ICRU; 1999.
Radiother Oncol 1999;52:35– 42. 18. International Commission on Radiation Units and Measure-
5. Robbins KT, Medina JE, Wolfe GT, et al. Standardizing neck ments Report. ICRU 50: Prescribing, recording, and reporting
dissection terminology: Official report of the Academy’s photon beam therapy. Washington, DC: ICRU; 1993.
Committee for Head and Neck Surgery and Oncology. Arch 19. Sako K, Bradier RN, Marchotta FC, et al. Feasibility of
Otolaryngol Head Neck Surg 1991;117:601– 605. palpation in the diagnosis of metastases to cervical nodes.
6. Fletcher GH, Million RR. Malignant tumor in nasopharynx. Surg Gynecol Obstet 1964;118:989 –990.
AJR Am J Roentgenol Radium Ther Nucl Med 1965;93:44 –55. 20. Huang DT, Johnson CR, Schmidt-Ullrich R, et al. Postoper-
7. Hunt MA, Zelefsky MJ, Wolden S, et al. Treatment planning ative radiotherapy in head and neck carcinoma with extracap-
and delivery of intensity-modulated radiation therapy for pri- sular lymph node extension and/or positive resection margins:
mary nasopharynx cancer. Int J Radiat Oncol Biol Phys 2001: A comparative study. Int J Radiat Oncol Biol Phys 1992;23:
49:623– 632. 737–742.
8. Chao KS, Low DA, Perez CA, et al. Intensity-modulated 21. Peters LJ, Goepfert H, Ang KK, et al. Evaluation of the dose
radiation therapy in head, and neck cancers: The Mallinckrodt for postoperative radiation therapy of head and neck cancer:
experience. Int J Cancer 2000:90:92–103. First report of a prospective randomized trial. Int J Radiat
9. Som PM, Curtin HD, Mancuso AA. An imaging-based clas- Oncol Biol Phys 1993;26:3–11.
sification for the cervical nodes designed as an adjunct to 22. McLaughlin MP, Mendenhall WM, Mancuso AA, et al. Ret-
recent clinically based nodal classifications. Arch Otolaryngol ropharyngeal adenopathy as a predictor of outcome in squa-
Head Neck Surg 1999;125:388 –396. mous cell carcinoma of the head and neck. Head Neck 1995;
10. Emami B, Lyman J, Brown A, et al. Tolerance of normal 17:190 –198.
tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys 23. Candela FC, Kothari K, Shah JP. Patterns of cervical node
1991;21:109 –122. metastases from squamous carcinoma of the oropharynx and
11. Eisbruch A, Ten Haken RK, Kim HM, et al. Dose, volume, hypopharynx. Head Neck 1990;12:197–203.
and function relationships in parotid salivary glands following 24. Shah JP, Candela FC, Poddar AK. The patterns of cervical
conformal, and intensity-modulated irradiation of head and lymph node metastases from squamous carcinoma of the oral
neck cancer. Int J Radiat Oncol Biol Phys 1999;45:577–587. cavity. Cancer 1990;66:109 –113.
12. Chao KS, Deasy JO, Markman J, et al. A prospective study of 25. Bataini JP, Bernier J, Brugere J, et al. Natural history of neck
salivary function sparing in patients with head-and-neck can- disease in patients with squamous cell carcinoma of orophar-
cers receiving intensity-modulated or three-dimensional radi- ynx and pharyngolarynx. Radiother Oncol 1985;3:245–255.
ation therapy: Initial results. Int J Radiat Oncol Biol Phys 26. Byers RM, Wolf PF, Ballantyne AJ. Rationale for elective
2001:49:907–916. modified neck dissection. Head Neck Surg 1988;10:160 –167.
13. Chao KS, Majhail N, Huang C, et al. Intensity-modulated 27. Lindberg R. Distribution of cervical lymph node metastases
radiation therapy reduces late salivary toxicity without com- from squamous cell carcinoma of the upper respiratory and
promising tumor control in patients with oropharyngeal car- digestive tracts. Cancer 1972;29:1446 –1449.
1184 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 5, 2002

28. Northrop M, Fletcher GH, Jesse RH, et al. Evolution of neck imaging techniques and ultrasound-guided aspiration cytology
disease in patients with primary squamous cell carcinoma of for the assessment of neck node metastases: A prospective
the oral tongue, floor of mouth, and palatine arch, and clini- comparative study. Eur Arch Otorhinolaryngol 1993;250:11–
cally positive neck nodes neither fixed nor bilateral. Cancer 17.
1972;29:23–30. 36. Righi PD, Kopecky KK, Caldemeyer KS, et al. Comparison of
29. Woolgar JA. Histological distribution of cervical lymph node ultrasound-fine needle aspiration and computed tomography
metastases from intraoral/oropharyngeal squamous cell carci- in patients undergoing elective neck dissection. Head Neck
nomas. Br J Oral Maxillofac Surg 1999;37:175–180. 1997;19:604 – 610.
30. Buckley JG, MacLennan K. Cervical node metastases in la- 37. Annyas AA, Snow GB, van Slooten EA, et al. Prognostic
ryngeal and hypopharyngeal cancer: A prospective analysis of factors of neck node metastasis: Their impact on planning a
prevalence and distribution. Head Neck 2000:22:380 –385.
treatment regimen. Read before the American Society of Head
31. Kowalski LP, Medina JE. Nodal metastases: Predictive fac-
and Neck Surgeons. Los Angeles, 1979.
tors. Otolaryngol Clin North Am 1998;31:621– 637.
32. Stern WB, Silver CE, Zeifer BA, et al. Computed tomography 38. Johnson JT, Barnes EL, Myers EN, et al. The extracapsular
of the clinically negative neck. Head Neck 1990;12:109 –113. spread of tumors in cervical node metastasis. Arch Otolaryn-
33. Friedman M, Mafee MF, Pacella BL Jr, et al. Rationale for gol 1981;107:725–729.
elective neck dissection in 1990. Laryngoscope 1990;100:54 – 39. Carter RL, Bliss JM, Soo KC, et al. Radical neck dissections
59. for squamous carcinomas: Pathological findings and their clin-
34. Moreau P, Goffart Y, Collignon J. Computed tomography of ical implications with particular reference to transcapsular
metastatic cervical lymph nodes: A clinical, computed tomo- spread. Int J Radiat Oncol Biol Phys 1987;13:825– 832.
graphic, pathologic correlative study. Arch Otolaryngol Head 40. Hirabayashi H, Koshii K, Uno K, et al. Extracapsular spread
Neck Surg 1990;116:1190 –1193. of squamous cell carcinoma in neck lymph nodes: Prognostic
35. van den Brekel MW, Castelijns JA, Stel HV, et al. Modern factor of laryngeal cancer. Laryngoscope 1991;101:502–506.