A, Local control. B, Regional control. C, Cause-specific survival.
Basheeth N, O’Leary G, Sheahan P. Elective Neck Dissection for N0 Neck During Salvage Total LaryngectomyFindings, Complications, and Oncological Outcome. JAMA Otolaryngol Head Neck Surg. 2013;139(8):790-796. doi:10.1001/jamaoto.2013.3995
Optimum management of the clinically negative neck in the presence of primary site recurrence of laryngeal or hypopharyngeal cancer remains unclear.
To examine the incidence of occult cervical nodal metastatic disease in patients undergoing salvage laryngectomy with necks clinically staged as N0 at the time of recurrence and to define the role of elective neck dissection in this population with regard to risk of complications and oncologic outcomes.
Retrospective study of patients treated from 1996 through 2011.
Academic teaching hospital.
All patients undergoing salvage total laryngectomy for squamous cell carcinoma of larynx or hypopharynx after failed initial radiotherapy or chemoradiotherapy with radiographically N0 neck at time of recurrence.
Main Outcomes and Measures
Primary outcome measure was incidence of cervical nodal metastases detected by means of pathological examination of elective neck dissection specimens. Secondary outcome measures were incidence of major postoperative complications and regional cancer control.
Forty-five patients were included. Thirty-eight underwent elective unilateral (20) or bilateral (18) ND at the time of laryngectomy. Three patients (8%) had occult metastases (3 of 56 dissected heminecks [5%]). The incidence of major wound complications was significantly greater in patients undergoing bilateral ND (12 of 18 [67%]) than in patients undergoing unilateral or no ND (8 of 27 [30%]) (P = .03). There was no significant difference in regional control according to whether bilateral, unilateral, or no ND was performed.
Conclusions and Relevance
The incidence of occult metastatic disease in N0 necks in patients undergoing salvage surgery after radiotherapy is low. Neck dissection in this population does not seem to have a significant impact on regional cancer control. The need for elective ND, particularly bilateral ND, should be balanced against possible increased risk of morbidity in this group.
The optimum management for cancers of the larynx and hypopharynx has evolved dramatically over recent years. The development of organ preservation treatment protocols has allowed many advanced cancers to be treated nonsurgically.1,2 Unfortunately, local recurrences still occur. However, in many cases, these recurrent cancers are suitable for salvage treatment by means of total laryngectomy with or without pharyngectomy.3,4
The optimum management of the neck in patients who have received radiotherapy is controversial. Recent data have suggested that in cases in which a complete tumor response is confirmed by means of imaging, posttreatment neck dissection (ND) is not required.5- 7 However, these studies are confined to cases without simultaneous primary site recurrence, and nearly all contain a preponderance of patients with oropharyngeal cancer. Thus, the optimum management of the clinically negative neck in the presence of primary site recurrence of laryngeal or hypopharyngeal cancer remains unclear.
Elective ND has been shown to be highly effective in the treatment of the clinically negative neck in the primary setting.8 However, an increased complication rate has been reported for ND performed after radiotherapy or chemoradiotherapy.9,10 Thus, the need for elective ND in patients who have undergone radiotherapy should be weighed against the possible increased risk of morbidity for the patient.
The purpose of the present study was 2-fold: First, we wished to investigate the incidence of occult cervical nodal metastatic disease in patients undergoing salvage total laryngectomy at our institution. Second, we wished to examine the role of elective ND in the management of patients with clinically negative necks undergoing salvage surgery. Specifically, we wished to investigate whether the performance of ND was associated with any increased risk of major complications and whether it had any impact on subsequent locoregional tumor control.
The present study comprised a retrospective review of 94 consecutive patients undergoing laryngectomy or pharyngolaryngectomy at our institution during a 15-year period. Cases were identified by means of review of the operating theater logbook for the years covered by the study period (1996-2011). Inclusion criteria were as follows: cases of salvage laryngectomy or pharyngolaryngectomy performed with curative intent after failed radiotherapy or chemoradiotherapy for squamous cell carcinoma of the larynx or hypopharynx; eligible patients had radiographically N0 necks at the time of recurrence.
Case notes of patients meeting inclusion criteria were reviewed, and information was obtained regarding the following: initial TNM staging, initial treatment, recurrence TNM staging, details of preoperative imaging, details of salvage surgery performed, results of pathologic analysis from the salvage surgery, postoperative complications, and postsalvage surgery locoregional control. Selective ND was defined as removal of neck levels II and III or II, III, and IV, with preservation of level V and the sternomastoid muscle (although in most cases the manubrial head of the sternomastoid was divided to facilitate postoperative stoma and tracheoesophageal prosthesis management). Modified radical ND was defined as removal of neck levels II, III, IV, and V including the sternomastoid muscle.
Postoperative major wound complications were defined as wound hematoma requiring return to the operating room; wound infection or collection requiring opening and drainage in the operating room or on the ward; and pharyngocutaneous fistula, defined as leakage of saliva from the pharynx onto the skin or into the stoma.
Statistical analysis was performed using Prism, version 6.0 (GraphPad). Survival data were analyzed using Kaplan-Meier curves. Survival curves were compared using the Mantel-Cox method. Analysis of 2 × 2 contingency tables was performed using the Fisher exact test.
Outcomes were compared according to extent of ND performed (bilateral vs unilateral or no ND). Because of the small number of patients not undergoing any ND at the time of salvage surgery (7), these patients were included with patients undergoing unilateral ND for the purpose of analysis of differences in postoperative complications. Permission to perform the study was obtained from the Cork Clinical Research Ethics Committee.
The final study population consisted of 45 patients (36 men and 9 women) with a mean (range) age of 64.3 (42-78) years. Forty-one patients had carcinoma of the larynx, and 4 had carcinoma of the hypopharynx. The primary tumor subsite in the larynx was the supraglottis in 11, glottis in 18, transglottic in 9, and subglottis in 3. Forty patients underwent total laryngectomy with partial pharyngectomy; 4 underwent total pharyngolaryngectomy; and 1 underwent total pharyngolaryngoesophagectomy. Nineteen patients also received pectoralis major myofascial or myocutaneous flaps, and 1 underwent stomach pull-up.
Initial treatment was with radiation only in 28 cases and with chemoradiation in 17 cases. All patients received conventional radiotherapy, with median radiation dose of 66 Gy. Chemotherapy was generally given as weekly cisplatin or carboplatin, plus 5-fluorouracil.
Imaging performed prior to salvage laryngectomy included computed tomography (CT) (all 45 patients), magnetic resonance imaging (MRI) (35 patients), and positron emission tomography (PET) (12 patients).
Elective ND was performed on 38 patients at the time of laryngectomy (20 unilateral ND and 18 bilateral ND) (Table 1). Of 56 heminecks dissected, 40 underwent selective ND and 16 underwent modified radical ND. Seven patients did not undergo any ND (Table 1).
Among the 38 patients undergoing elective ND, 3 (8%) had pathological evidence of metastatic cancer, representing 5% of dissected heminecks (3 of 56). One of these was in a 60-year-old man with a hypopharyngeal primary tumor that was initially staged T4N0 and treated with radiation alone. The second was in a 62-year-old man with a subglottic primary tumor, initially staged T4N0 and treated with chemoradiation. The third was in a 56-year-old man with a transglottic primary tumor, initially staged T4N0 and treated with chemoradiation. All 3 of these patients had undergone preoperative imaging by means of CT and MRI scanning. None had undergone PET scan.
Postoperatively, 20 patients (44%) developed a major complication (Table 2). Complications included 17 pharyngocutaneous fistulae (38%), 5 hematomas (11%), and 3 wound infections (7%). Four patients developed more than 1 complication. Table 2 shows the incidence of postoperative complications according to whether no ND, unilateral ND, or bilateral ND was performed. The overall incidence of postoperative major complications was significantly higher in patients undergoing bilateral ND (67%) than in patients undergoing unilateral or no ND (30%) (P = .03).
Table 3 shows the incidence of complications according to the type of ND performed, as well as use of pectoralis major flap, and initial treatment with chemoradiation vs radiotherapy alone. The only factor significant for postoperative major complications was extent of ND (unilateral or none vs bilateral). The median (range) length of follow-up for the entire cohort was 24 (4-180) months and for surviving patients was 34.5 (4-168) months. Fourteen patients developed local recurrence. Of these, 5 developed simultaneous local and regional recurrence. In 2 additional patients (1 who had undergone bilateral ND and 1 who had not undergone any ND), there was extensive local recurrence extending into the lateral necks, which made it difficult to exclude simultaneous regional recurrence. No patient developed isolated regional recurrence.
Local control rates according to extent of ND are shown in Figure, A, and regional control rates, in Figure, B. Because of the absence of any cases of isolated regional recurrence, regional control is given without exclusion of cases of local recurrence. Cause-specific survival is shown in Figure, C.
Local recurrence developed in 2 of 7 patients not undergoing ND, in 4 of 20 patients undergoing unilateral elective ND, and in 8 of 18 patients undergoing bilateral elective ND. There was no significant difference in local control rates between patients not undergoing ND, undergoing unilateral elective ND, and undergoing bilateral elective ND (P = .51) (Figure, A).
Simultaneous local and regional recurrence developed in 1 patient not undergoing ND and in 4 patients undergoing bilateral ND. There was a trend toward a poorer regional control rate in patients undergoing bilateral ND, but this was not significant (P = .08) (Figure, B).
There was a trend seen toward poorer cause-specific survival among patients undergoing bilateral ND; however, this was not statistically significant (P = .09) (Figure, C).
The findings of the present study would suggest the following: (1) pathological metastases in cervical nodes are uncommon in patients undergoing salvage surgery for laryngeal or hypopharyngeal cancer whose tumors are staged rN0 at the time of recurrence; (2) postoperative major wound complications seem to be increased in patients undergoing concomitant bilateral ND with salvage laryngectomy, although there does not seem to be any increased risk of complications in patients undergoing unilateral ND compared with no ND; and (3) there does not seem to be any significant difference in local control, regional control, or cause-specific survival according to how the rN0 neck is managed in the salvage setting, although the number of patients not undergoing any ND in our study was low. Taken together, these findings would suggest that the need for and benefit of elective ND in patients undergoing salvage laryngectomy without radiographic evidence of nodal disease should be measured against possible risk of increased morbidity, particularly in cases in which a bilateral ND is being considered.
The incidence of pathological metastatic disease in elective ND specimens in patients with clinically negative necks at the time of salvage laryngectomy in the present study was 8% of patients, or 5% of heminecks. This is comparable to reported rates of metastatic neck disease in elective ND specimens of patients undergoing salvage surgery for recurrent head and neck cancer in the literature of between 3% and 19%.11- 17Table 4 gives a summary of previously published studies reporting rate of metastases in elective NDs in salvage laryngectomy. All of these studies are confined to patients undergoing salvage laryngectomy, with the exception of that by Dagan et al,16 which also included patients with oropharynx cancer; however, two-thirds of patients in that study had cancer of the larynx or hypopharynx.
There are several possible reasons for the differences in reported rates of occult metastatic neck disease at the time of salvage laryngectomy. These include differences in study inclusion criteria, differences in radiation or chemoradiation protocols, differences in imaging modalities used to restage the neck at the time of recurrence, and differences in pathologic interpretation of the viability of tumor cells (in the case of surgical procedures performed soon after the completion of radiotherapy).
In the present study, all 45 patients underwent CT scanning prior to salvage laryngectomy, and more than three-quarters (35 of 45) also underwent MRI. It is possible that this widespread use of imaging may have been an important factor that contributed to our low rate of occult metastatic disease. Use of imaging procedures has not always been detailed in previous studies, including some of the studies that reported higher rates of pathological metastatic disease.11,15 However, CT, in particular, has been shown to have a high sensitivity and negative predictive value in the assessment of the postradiotherapy neck.18 It is possible that the accuracy of preoperative staging might be enhanced by increased use of PET-CT scanning. The use of PET-CT has been shown to have a high negative predictive value in the assessment of cervical nodes after primary chemoradiotherapy.19 It is notable that none of the 3 patients found to have occult metastases in the present study had undergone this modality. However, the total number of patients in our series undergoing PET scanning was quite low (12). Thus, additional data regarding the accuracy of PET-CT for patients with confirmed primary site recurrence are awaited.
Of note, most of the series listed in Table 4 had N0 neck at time of initial presentation as an inclusion criterion.12- 16 We did not exclude patients with N+ necks at time of initial presentation, although the number of such cases in our series (3) was small. Thus, our findings of a low incidence of metastatic disease in salvage elective ND may not be applicable to patients with confirmed metastatic neck disease at initial presentation. There is a body of evidence to suggest that adjuvant elective ND is not necessary for patients with necks initially staged N1 to N2 that show a complete response to radiotherapy or chemoradiotherapy on imaging.5- 7 However, most of these studies include only patients who also have a complete response at the primary site, and nearly all also contain a preponderance of patients with oropharyngeal cancer. In contrast, Amit et al11 reported patients with N+ necks at time of initial presentation whose necks were rendered N0 by radiation or chemoradiation and who underwent elective ND at the time of salvage laryngectomy to have a significantly increased risk of occult metastases.
One concern regarding the performance of elective ND in the salvage setting is risk of subjecting the patient to additional toxic effects or complications. Postradiotherapy ND has been reported by several authors to be associated with an increased rate of complications,9,10 whereas concomitant ND at the time of salvage total laryngectomy has been reported to be a risk factor for pharyngocutaneous fistula formation.20 In their paper, Dagan et al16 reported a higher incidence of toxic effects including pharyngocutaneous fistula and a 5% mortality rate in patients who underwent elective ND at the time of salvage surgery compared with patients not undergoing ND, and Yao et al15 also reported a trend toward increased incidence of pharyngocutaneous fistula in the ND group. Bohannon et al17 also reported a higher complication rate in patients undergoing ND. Of note, most of the patients undergoing ND in the study by Bohannon et al17 underwent bilateral ND.
In our series, we found a significantly increased incidence of major wound complications in patients undergoing bilateral ND compared with those undergoing unilateral ND or not undergoing ND. The number of patients not undergoing ND was too low to make a comparison between these patients and patients undergoing unilateral ND; however, the complication rate in patients undergoing no ND was almost identical to that seen in the unilateral ND group, suggesting that unilateral ND alone probably does not add substantially to postoperative morbidity.
There are some possible explanations why the complication rate may be increased in patients undergoing bilateral ND. One is that such patients may also be more likely to have more advanced primary tumors, requiring more extensive surgery at the primary site and more complex reconstruction, thus increasing risks of leakage and other problems. However, patients in the present series undergoing unilateral and bilateral ND seem to have been well matched with regard to primary tumor site and stage (Table 1). Another possible explanation is that performance of bilateral ND may aggravate disruption of cervical vascularity and lymphatic drainage and thus impair healing of the neopharynx, increasing the risk of fistula formation. Finally, adding bilateral NDs to an already difficult and lengthy procedure may increase the risk of surgical staff fatigue.
Interestingly, we did not find any difference in risk of major wound complications between patients who underwent selective vs modified radical or radical ND. However, our data did not include assessment of shoulder function postoperatively. Selective ND has been shown to confer significant benefits from a shoulder function point of view21,22 and recently has been shown to be efficacious in the treatment of the postradiotherapy neck.23 For these reasons, it has become our preferred form of ND in recent years for N0 or low-volume N+ cases.
There were no cases of isolated regional recurrence in our series. For this reason, for the purpose of analysis of the effect of extent of ND on regional control, we included cases that had simultaneous primary site and regional recurrence. We did not find any difference in local control, regional control, or cancer-specific survival according to management of the neck in the salvage setting. If anything, there was a trend toward poorer outcomes in patients undergoing bilateral ND. This is also in keeping with the findings of Dagan et al,16 Yao et al,15 and Bohannon et al.17
There are some limitations to the present study. The most important of these are its retrospective nature, which precludes the exclusion of bias regarding how the neck was managed; as well as the small number of patients not undergoing any ND, which precluded meaningful statistical evaluation of differences in oncological outcome and postoperative complications between patients undergoing and not undergoing ND. Indications for performance of unilateral vs selective ND, as well as type of ND, were not always documented. Furthermore, although our findings suggest that cervical nodal metastatic disease is uncommon in patients undergoing salvage laryngectomy with radiographically negative necks, they do not exclude the possibility that there may be a subset of patients at increased risk of harboring metastases, although the small number of pN+ cases in our series precluded meaningful evaluation of such risk factors. By contrast, the universal use of CT scanning and widespread use of MRI likely facilitated a high degree of accuracy and consistency in preoperative recurrence staging of the neck.
The findings of the present study would suggest that cervical nodal metastatic disease is uncommon in elective ND specimens of patients undergoing salvage total laryngectomy in whom preoperative imaging results are negative for nodal disease and that performance of elective ND does not seem to have any significant impact on postoperative regional cancer control. By contrast, whereas the performance of unilateral ND does not seem to incur any significant increase in operative morbidity, there seems to be a significant increase in postoperative major complications in cases in which bilateral ND is performed. In cases in which preoperative imaging results are suspicious for regional nodal disease or in which exposure of vessels in the neck is going to be required for simultaneous free flap reconstruction of the pharynx, ND is clearly indicated. However, our data would suggest that in cases of radiographically negative necks in patients not undergoing free flap reconstruction, the need for bilateral elective ND should be weighed against this possible increased risk of morbidity.
Corresponding Author: Patrick Sheahan, MD, FRCSI (ORL-HNS), Department of Otolaryngology–Head and Neck Surgery, South Infirmary Victoria University Hospital, Old Blackrock Rd, Cork City, County Cork, Ireland (firstname.lastname@example.org).
Submitted for Publication: November 7, 2012; final revision received December 16, 2012; accepted January 24, 2013.
Author Contributions: Dr Sheahan had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition of data: Basheeth, Sheahan.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Sheahan.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Sheahan.
Administrative, technical, or material support: All authors.
Study supervision: All authors.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented as a poster at the 2013 American Head and Neck Society meeting; April 10-11, 2013; Orlando, Florida.