Customize your JAMA Network experience by selecting one or more topics from the list below.
Hypothesis: Selective neck dissection (SND) is adequate treatment for node-positive neck disease.
Selective neck dissection, or the selective removal of nodal groups at risk for harboring cervical metastases, is an extension of the concept of the functional neck dissection first introduced in the 1960s. Until 1963, radical neck dissection (RND) was the standard surgical treatment of the cervical lymphatic nodes. In 1963 Suarez proposed a conservative, functional approach to neck dissection that was popularized by Bocca et al.1 These authors described the removal of cervical lymph nodes from levels I through V with their enveloping fascial sheath while sparing nonlymphatic anatomic structures adjacent to, but not containing, lymph nodes, such as the accessory nerve, internal jugular vein, and sternocleidomastoid muscle.1 The resulting modified RND (MRND) has been shown to avoid the cosmetic and functional morbidity associated with the classic RND without compromising oncologic safety or efficacy as a staging and therapeutic procedure.
Christine G. Gourin, MD
Various modifications of the functional or MRND have been proposed for the elective treatment of the clinically negative neck (cN0) when surgical treatment of the primary tumor is planned and the risk of occult metastases is greater than 20%.2 Selective neck dissections are based on the observation that tumors of the upper aerodigestive tract drain to predictable nodal levels according to the site of the primary tumor, and that a comprehensive neck dissection encompassing all 5 nodal levels may therefore be unwarranted.3 The incidence of occult metastases in the cN0 neck exceeds 20% for tumors arising from the oral cavity, oropharynx, hypopharynx, and supraglottic larynx.4 The presence of cervical metastases is associated with a poorer prognosis. Accurate detection of occult metastases is difficult, as approximately 50% of occult metastases are less than 5 mm and missed on physical and radiologic examination.5 Extracapsular spread may be present in as many as 50% of patients with occult metastases.6 If neck dissection is performed therapeutically when neck disease becomes clinically apparent, survival and surgical salvage rates are significantly decreased.6 An appropriately selected SND, which removes lymph nodes only at the levels likely to harbor metastatic disease, provides important prognostic information and, in some circumstances, may be therapeutic.
In a review of 1119 RNDs performed electively for cN0 neck disease and therapeutically for clinically node-positive neck disease, Shah7 found that occult metastases were present in 33% of elective neck dissections and that level V was never involved in the absence of involvement of other nodal levels. Levels I, II, and III were at greatest risk for nodal metastases from squamous cell carcinoma of the oral cavity, and levels II, III, and IV were at greatest risk for metastases from carcinomas of the oropharynx, hypopharynx, and larynx. These findings support the use of supraomohyoid neck dissection (removing lymph nodes from levels I through III) for oral cavity tumors in cN0 patients and anterolateral neck dissection (removing lymph nodes from levels II through IV) in patients with cN0 tumors of the oropharynx, hypopharynx, and larynx.7 Byers et al8 have shown that “skip metastases” to level IV occur in 16% of patients with squamous cell carcinoma of the oral tongue and advocate removal of level IV nodes as part of SND for oral tongue cancer.
The regional recurrence rate following SND alone in patients with histologically negative nodes (pN0) is approximately 5%,2,3,9-16 which compares favorably with the rate in patients undergoing MRND for clinically and histologically negative nodal disease.1,17 It is generally accepted that SND is as effective as comprehensive neck dissection in staging the cN0 neck and is adequate treatment for the pN0 neck. The therapeutic efficacy of SND in patients with pathologic node-positive neck disease, however, is more controversial. Data suggest that SND alone may be adequate treatment in patients found to have a single histologically positive lymph node without extracapsular spread (pN1). The regional recurrence rate ranges from approximately 4% to 10% in this situation, and does not appear to be significantly affected by the addition of postoperative radiation therapy (XRT).2,9,11,12,14,16,18 Because these results compare favorably with those in patients from a comparable cohort treated with MRND, the following question arises: Is there is a therapeutic role for SND in the treatment of the clinically positive neck?
The application of SND in the management of selected node-positive neck disease was first described by Byers.9 In a review of 234 supraomohyoid neck dissections and 297 anterolateral neck dissections performed for cN0 or cN1 disease, Byers reported a 5% recurrence rate for pN0 disease and a 10% recurrence rate for pN1 disease.9 In these 2 pathologic groups, no significant difference in regional recurrence was found with the addition of postoperative XRT. Patients with multiple histologically positive nodes or extracapsular spread had a 24% recurrence rate after surgery alone, and a 15% recurrence rate with the administration of postoperative XRT.2 These findings compared favorably with those from a cohort treated with MRND and similar indications for postoperative XRT. Byers concluded that SND alone was adequate treatment for patients with pN0 or pN1 disease, but the addition of postoperative XRT significantly improved both survival and recurrence rates when multiple nodes were positive or extracapsular spread was present.9
Pitman et al4 have demonstrated that SND is as efficacious as MRND in staging the cN0 neck, with no recurrences noted outside the field of dissection. In a follow-up study evaluating 300 SNDs performed for cN0 disease, the recurrence rates were similar in patients with pN0 disease (3%) and in patients with pathologic node-positive disease limited to fewer than 3 lymph nodes without extracapsular spread and treated without postoperative XRT (4%).16 Only 0.7% of recurrences occurred outside the dissected field. The application of these findings to patients with multiple levels of involvement was limited by small patient numbers.
Several retrospective studies have reported the outcome of SND when applied in the treatment of the clinically and histologically positive neck disease. Ambrosch et al14 reported their results with the use of SND in 283 patients with oral cavity, oropharyngeal, laryngeal, and hypopharyngeal primary cancers that were clinically staged N1 or N2. Patients with pN1 disease had a regional recurrence rate of 5% independent of the administration of postoperative XRT. In patients with pN2 disease (more than 1 histologically positive lymph node), the recurrence rate was 7% with postoperative XRT and 24% without XRT. Most recurrences (85%) occurred within the dissected field. The authors concluded that both the rate and site of recurrence were similar to those noted with MRND when similar indications for postoperative XRT were applied, and they did not support the use of a more comprehensive neck dissection.
Chepeha et al19 reported that SND in 52 patients with clinical evidence of metastases smaller than 3 cm (N1, N2B, and N2C) with XRT administered postoperatively when more than 2 positive nodes or extracapsular spread were present, resulted in regional control in 94% of patients. One third of recurrences, which were successfully salvaged, occurred out of the field of dissection in level V. They concluded that a more complete neck dissection would not have provided measurable therapeutic benefit. In a similar patient population with similar indications for postoperative XRT, Traynor et al20 found a regional recurrence rate of 4% with no recurrence out of the field of dissection and concluded that the indications for SND could be extended to include patients with nonfixed nodal disease smaller than 3 cm without clinical evidence of extracapsular spread.
Pellitteri et al11 reported the use of SND in patients with cN1, cN2B, and cN3 disease, with postoperative XRT administered to patients with histologically positive nodes at multiple levels or with extracapsular spread. The regional recurrence rate in patients with cN2 or cN3 disease was 10%, with only 1 recurrence occurring out of the field of dissection in level V. The recurrence rate in patients with multiple histologically positive nodes was 13% and did not differ significantly from the 7% rate observed in patients with pN0 or pN1 disease. This is the only study that included patients with cN3 disease, although it accounted for only 4 patients with that staging.
A multi-institutional experience with the use of SND in 129 patients with clinically and histologically positive nodal disease was recently reported by Andersen et al.21 Patients with cN1, cN2A, cN2B, and cN2C disease were included as well as 1 patient with cN3 disease. Postoperative XRT was administered for histologic evidence of multiple levels of nodal involvement or the presence of extracapsular spread. The authors reported a recurrence rate of 7%, with no recurrences outside the dissected field. The presence of extracapsular spread was significantly associated with recurrence.
These data suggest that SND is adequate treatment for pN1 disease and that it yields regional control rates for more advanced pathologic node-positive disease similar to those of more extensive comprehensive neck dissections when similar indications for postoperative XRT are applied.
No truly randomized prospective study has been performed to determine the efficacy of SND compared with MRND in the treatment of patients with histologically positive lymph nodes. Most studies supporting the use of SND in the histologically positive neck are retrospective, and therefore the therapeutic equivalence of SND and MRND in this setting is unproven. Leemans and Snow22 have combined several large studies reporting the results of elective SND and elective MRND and calculated a statistically significant lower recurrence rate in pathologic node-positive necks treated with MRND (4%) compared with SND (11%).
Several studies suggest that SND is efficacious in the treatment of pathologic node-positive neck disease only if postoperative XRT is administered for any histologically positive node. A follow-up study by Byers et al10 describing 363 more recently treated patients with cN0, cN1, and cN2B neck disease found a recurrence rate of 36% in pN1 necks treated without XRT, much higher than the 6% rate that they observed with postoperative XRT. The recurrence rate in patients with pN2B disease was 14% without postoperative XRT and 8% with postoperative XRT. While there were only 11 patients with pN1 disease who did not receive postoperative XRT, the authors concluded that postoperative XRT was advisable in patients with any nodal involvement. All regional failures following supraomohyoid neck dissection in pN1 necks occurred within the dissected field, suggesting that a more comprehensive neck dissection would not have benefited these patients but that removing the internal jugular vein might have more thoroughly cleared nodal disease.
Spiro et al13 reported the use of supraomohyoid neck dissection in 287 patients with oral cavity and oropharyngeal carcinoma and cN0 or cN1 disease. All patients with histologically positive nodes received postoperative XRT. The regional recurrence rate was 5% for pN0 and 7% for pathologic node-positive necks with cN0 disease. In patients with clinical evidence of metastases (cN1), the recurrence rate was 6% if nodes were positive histologically. Only 1 recurrence occurred outside of the dissected field, in a parapharyngeal node, suggesting that a more comprehensive neck dissection would not have benefited these patients. They suggested that SND had a therapeutic role in patients with limited neck disease when combined with postoperative XRT.
Two randomized prospective studies comparing the results of SND with those of MRND in patients with cN0 tumors were performed by the Brazilian Head and Neck Cancer Study Group and evaluated 148 patients with oral carcinoma and 132 patients with laryngeal cancer.15,23 Postoperative XRT was administered if any histologically positive lymph node was found. Patients were not truly randomized, as both studies were performed with the caveat that patients in their SND arm underwent frozen section analysis of any suspicious lymph node, with conversion to MRND if frozen section evaluation revealed the node to be histologically positive. Interestingly, intraoperative frozen section evaluation identified only 20% of patients with histologically positive lymph nodes, and therefore conversion to MRND was not performed in most of the pathologic node-positive patients assigned to the SND group. No significant difference was found in the recurrence rates of patients treated with SND (8% for oral cavity and 3% for laryngeal carcinoma) compared with the recurrence rates of those treated with MRND (9% for oral cavity and 3% for laryngeal carcinoma), with similar application of postoperative XRT for pathologic node-positive disease.
In a series of 164 patients with oral cavity carcinoma stage cN1 or cN2A with similar indications for postoperative XRT (any histologically positive node), Kowalski and Carvalho12 found a recurrence rate of 4% for pN0 disease and 12% for pathologic node-positive disease after supraomohyoid neck dissection. No patient had level V involvement, and only 1 patient (0.6%), in whom other dissected levels were involved, also had level IV involvement. The authors concluded that patients with pathologic node-positive disease had a higher rate of recurrence, which appeared to be independent of the type of neck dissection performed and may instead be related to tumor characteristics and pathologic stage of disease—a conclusion supporting the use of postoperative XRT.
In patients treated with RND, MRND, or SND for oral carcinoma, Kowalski24 reported that survival after neck recurrence of disease was significantly influenced by the type of initial neck dissection and previous postoperative XRT. The presence of “skip metastases,” as reported by Byers et al,8 raises the concern that SND may result in a sampling error when used in the clinically positive neck. In clinically node-positive necks treated with RND, Shah7 found a 15% incidence of histologically positive nodes in level IV in patients with oral cavity carcinoma and in patients with oropharyngeal or hypopharyngeal cancer, a 13% incidence of involvement in level I, and a 10% incidence of involvement of level V. These levels would not be included routinely in SNDs for these primary tumor sites. All patients had histologically positive nodes at other levels that would have been included in the SND, but these findings raise the possibility that SND may inadequately stage patients with clinically node-positive disease.
When the indications for postoperative XRT are based on the number of involved lymph nodes, SND may underestimate the true extent of disease in pathologic node-positive patients and result in undertreatment. The use of postoperative XRT appears to have a significant therapeutic effect in patients with involvement of multiple nodes. The indications for postoperative XRT are not consistent from institution to institution: some surgeons advocate XRT if 1 lymph node is found to contain metastatic disease,12-15 whereas others recommend XRT if more than 1 node is positive,3,11,20 and still others for involvement of more than 2 nodes or more than 1 level.4,16,19 If criteria for the administration of postoperative XRT are based on the number of positive nodes, then the possibility of sampling error and ensuing undertreatment of the neck are real. The only consistently uniform application of XRT is in the presence of extracapsular spread.
Extracapsular spread is associated with a higher recurrence rate, a significantly diminished disease-free interval, and a poorer prognosis despite the use of postoperative XRT.6,16,25 Microscopic extracapsular spread is as poor a prognostic factor as macroscopic extracapsular spread, and it is associated with significantly decreased recurrence-free survival.26 The incidence of extracapsular spread increases with increasing nodal size and is as high as 75% in histologically positive nodes larger than 3 cm.25 The greater incidence of extracapsular spread seen with more extensive disease raises the concern that patients with cN2A and cN3 disease may not be adequately treated by SND.12 There is a lack of data regarding the use of SND in patients with lymph nodes larger than 3 cm, as such patients are typically treated with RND or MRND followed by XRT.
Selective neck dissection is efficacious in the treatment of patients with limited neck disease (smaller than 3 cm). In patients with pN1 disease without extracapsular spread, SND appears to be adequate treatment. Selective neck dissection is inadequate for treatment of multiple positive nodes or extracapsular spread without the administration of postoperative XRT. If one is going to use the results of SND to determine whether to administer adjuvant XRT, a randomized prospective study is needed to determine if SND is adequate treatment for patients with limited clinically N-positive neck disease, or if it may result in a sampling error that would affect subsequent adjuvant therapy and outcome. In patients with advanced neck disease or advanced primary tumors, the risk of nodal involvement in areas not typically addressed with neck dissection, such as the retropharyngeal nodes or parapharyngeal nodes, warrants the use of postoperative XRT.
There is a lack of data to support the use of SND in the case of clinically advanced neck disease. The performance of SND requires technical skill and clinical judgment, as all lymphatic nodes at risk of harboring metastatic disease must be removed. The presence of nodal fixation, disease greater than 3 cm, or clinical evidence of extracapsular spread violate the rationale behind SND: as fascial compartments and lymphatic flow patterns may be disrupted, the procedure may result in incomplete removal of disease and nodal levels at risk. Under these conditions, the MRND and the RND, which removes all echelons of nodal drainage, must remain standards of care until there is scientific data to support the use of SND in this circumstance.
Correspondence: Christine G. Gourin, MD, Department of Otolarynogology–Head and Neck Surgery, Medical College of Georgia, 1120 15th St, BP 4109, Augusta, GA 30912 (email@example.com).
Submitted for Publication: April 15, 2004; accepted July 8, 2004.
This article was corrected on 5/4/06.
Gourin CG. Is Selective Neck Dissection Adequate Treatment for Node-Positive Disease? Arch Otolaryngol Head Neck Surg. 2004;130(12):1431–1434. doi:10.1001/archotol.130.12.1431
Browse and subscribe to JAMA Network podcasts!
Create a personal account or sign in to: