Abstract
Purpose
To investigate the difference in survival and complication outcomes between patients with a clinically and radiologically N0 neck who received an elective neck dissection at the time of salvage total laryngectomy compared to those who had salvage total laryngectomy alone.
Materials and methods
A retrospective chart review was performed on 125 salvage total laryngectomy patients who were clinically and radiologically N0 preoperatively. Performance of an elective neck dissection and other factors were tested for associations with various postoperative complications, disease-free survival, and overall survival.
Results
Ninety-eight patients underwent elective neck dissection, of which ten had positive nodal pathology. Elective neck dissection was not significantly associated with complications or survival outcomes. Positive nodal disease was associated with worse disease-free and overall survival on multivariate analysis.
Conclusions
In patients with clinically and radiologically N0 necks undergoing salvage total laryngectomy, an elective neck dissection can provide prognostic information but does not appear to be significantly associated with increased complications or improved survival.
1
Introduction
Routine elective neck dissection (END) for the clinically and radiologically negative (N0) neck in the setting of recurrent laryngeal squamous cell carcinoma after previous radiotherapy or chemoradiotherapy remains controversial. Over the past five years, the incidence of occult nodal disease after salvage total laryngectomy (STL) with END reported in the literature ranges from 0% to 28.3% . Many studies have failed to demonstrate a survival benefit with END in this setting , while others have noted a benefit for some patients . A number of studies have found a higher complication rate after END while others suggested no significant difference . The present study represents to the best of our knowledge the largest retrospective review to date with the aim of elucidating the impact of performing an elective neck dissection during STL on survival and postoperative complication outcomes. Furthermore, we attempt to identify other factors that may be associated with increased complications, occult disease, or mortality in N0 patients undergoing salvage total laryngectomy.
2
Materials and methods
The current study is a retrospective chart review performed with permission from the institutional review board at our institution. One hundred twenty-five patients who underwent salvage total laryngectomy (STL) with clinically and radiologically negative necks at presentation of recurrence were identified from an existing billing database of 424 patients treated with total laryngectomy at our institution between 2000 and 2012. Inclusion criteria were that patients must have had a salvage total laryngectomy performed after previous radiotherapy or chemoradiotherapy for squamous cell carcinoma of the larynx and they must have an N0 neck at time of recurrence as determined by physical exam and imaging with CT, MRI, PET, or PET/CT scan. For the purpose of this study, “recurrent” disease included tumor recurrence, persistence, or second primary in the larynx. Patients with laryngectomy for non-squamous cell carcinoma or nonmalignant disease, laryngeal metastasis from another primary, previous neck dissection, or documented evidence of nodal disease at any time prior to STL were excluded.
Extensive pre- and post-operative demographic and clinical data were collected for each patient. Comprehensive review of all postoperative notes up to one year after surgery was performed in search for reported complications. Review of clinic notes up to the most recent available follow-up was performed in search for information regarding recurrent disease and survival status. Publically available death records were also checked for information beyond the follow up available in the medical records.
Statistical analyses were performed using SAS version 9.3 (Cary, NC). Associations between age and each of the binary outcomes were explored with the Student’s t -test. Each factor was assessed for potential bivariate associations with each complication under consideration using the chi-square or Fisher’s exact tests as appropriate in order to determine which factors may be predictive of increased complications or occult nodal disease. Multivariable analysis was subsequently performed using logistic regression to investigate the associations with binary outcomes including: overall complications, fistula formation, and wound complications (defined as history of infection or dehiscence). The following patient factors were included as predictors in the logistic models: age, T-stage at recurrence, site of recurrence, neck dissection status, performance of a pectoralis flap, and initial treatment modality. For oncologic outcomes, disease-free survival and overall survival were estimated using Kaplan–Meier curves. Log-rank tests were used for bivariate comparisons for these variables: neck dissection, nodal disease status, recurrence site, recurrence T stage, initial treatment modality, and pathologic T stage. Cox proportional hazards regression models were fit for both survival outcomes to further investigate adjusted associations with nodal disease status controlling for age, recurrence site, initial treatment modality, and pathologic T stage. All statistical significance in the current study was based on an alpha level of 0.05.
2
Materials and methods
The current study is a retrospective chart review performed with permission from the institutional review board at our institution. One hundred twenty-five patients who underwent salvage total laryngectomy (STL) with clinically and radiologically negative necks at presentation of recurrence were identified from an existing billing database of 424 patients treated with total laryngectomy at our institution between 2000 and 2012. Inclusion criteria were that patients must have had a salvage total laryngectomy performed after previous radiotherapy or chemoradiotherapy for squamous cell carcinoma of the larynx and they must have an N0 neck at time of recurrence as determined by physical exam and imaging with CT, MRI, PET, or PET/CT scan. For the purpose of this study, “recurrent” disease included tumor recurrence, persistence, or second primary in the larynx. Patients with laryngectomy for non-squamous cell carcinoma or nonmalignant disease, laryngeal metastasis from another primary, previous neck dissection, or documented evidence of nodal disease at any time prior to STL were excluded.
Extensive pre- and post-operative demographic and clinical data were collected for each patient. Comprehensive review of all postoperative notes up to one year after surgery was performed in search for reported complications. Review of clinic notes up to the most recent available follow-up was performed in search for information regarding recurrent disease and survival status. Publically available death records were also checked for information beyond the follow up available in the medical records.
Statistical analyses were performed using SAS version 9.3 (Cary, NC). Associations between age and each of the binary outcomes were explored with the Student’s t -test. Each factor was assessed for potential bivariate associations with each complication under consideration using the chi-square or Fisher’s exact tests as appropriate in order to determine which factors may be predictive of increased complications or occult nodal disease. Multivariable analysis was subsequently performed using logistic regression to investigate the associations with binary outcomes including: overall complications, fistula formation, and wound complications (defined as history of infection or dehiscence). The following patient factors were included as predictors in the logistic models: age, T-stage at recurrence, site of recurrence, neck dissection status, performance of a pectoralis flap, and initial treatment modality. For oncologic outcomes, disease-free survival and overall survival were estimated using Kaplan–Meier curves. Log-rank tests were used for bivariate comparisons for these variables: neck dissection, nodal disease status, recurrence site, recurrence T stage, initial treatment modality, and pathologic T stage. Cox proportional hazards regression models were fit for both survival outcomes to further investigate adjusted associations with nodal disease status controlling for age, recurrence site, initial treatment modality, and pathologic T stage. All statistical significance in the current study was based on an alpha level of 0.05.
3
Results
3.1
Patient characteristics
One hundred and twenty-five patients met the inclusion criteria out of the 424 laryngectomy patients identified. Median age was 70 years (ranging from 38 to 90 years). Initial therapy included radiation therapy in 104 (83.2%) and chemoradiation in 20 (16.0%) patients. Site of recurrence was the glottis in 90 patients (72.0%), the supraglottis in 17 patients (13.6%), and 18 patients had a transglottic tumor (14.4%). Ninety-eight patients (78.4%) underwent neck dissection and 27 (21.6%) did not have neck dissection. Within the neck dissection group, 47 (48.0%) underwent unilateral neck dissections while 51 (52.0%) had bilateral neck dissection. Patient and tumor characteristics are summarized in Table 1 .
| All patients, n = 125 | No neck dissection, n = 27 | Elective neck dissection, n = 98 | P value | |
|---|---|---|---|---|
| Mean age | 69.0 | 73.1 | 67.8 | 0.0238* |
| Standard deviation | 10.7 | 9.4 | 10.8 | |
| Median age | 70 | 74 | 70 | |
| Range | 38-90 | 54-90 | 38-90 | |
| Sex | ||||
| Male | 118 (94%) | 24 | 94 | 0.1712 |
| Female | 7 (6%) | 3 | 4 | |
| Initial T stage | – | |||
| T1 | 55 (44%) | 14 | 41 | |
| T2 | 35 (28%) | 10 | 25 | |
| T3 | 10 (8%) | 1 | 9 | |
| T4 | 4 (3%) | 0 | 4 | |
| Unknown | 21 (17%) | 2 | 19 | |
| Initial treatment modality † | ||||
| Radiation | 104 (83%) | 24 | 80 | 0.5606 |
| Chemoradiation | 20 (16%) | 3 | 17 | |
| Recurrence site | 0.4764 | |||
| Supraglottic | 17 (14%) | 3 | 14 | |
| Glottic | 90 (72%) | 22 | 68 | |
| Transglottic | 18 (14%) | 2 | 16 | |
| Preoperative imaging modality † | ||||
| CT | 84 (67%) | 21 | 63 | – |
| PET | 11 (9%) | 2 | 9 | |
| PET/CT | 27 (22%) | 4 | 23 | |
| MRI | 2 (2%) | 0 | 2 | |
| Recurrence T stage | 0.0085* | |||
| rT1 | 5 (4%) | 3 | 2 | |
| rT2 | 24 (19%) | 8 | 16 | |
| rT3 | 44 (35%) | 10 | 34 | |
| rT4 | 52 (42%) | 5 | 47 | |
| Pectoralis flap | 25 (20%) | 0 | 25 | – |
| Microvascular flap | 7 (6%) | 0 | 7 | – |
| Pathologic T stage | 0.0029* | |||
| T1 | 9 (7%) | 5 | 4 | |
| T2 | 22 (18%) | 7 | 15 | |
| T3 | 38 (30%) | 9 | 28 | |
| T4 | 56 (45%) | 5 | 51 |
3.2
Postoperative complications
Overall, 54 patients (43.2%) had postoperative complications. The most common complication was pharyngocutaneous fistula (PCF), which occurred in 34 (27.2%) patients. Fourteen patients (11.2%) and 20 patients (16.0%) developed wound infection and wound dehiscence respectively, with a total of 26 patients (20.8%) having at least one of these wound complications. There were two (1.6%) deaths within thirty days of surgery. Four (3.2%) patients developed carotid blowout within five months postoperatively, and three of them died of the blowout. Two blowout patients received bilateral ND, one received unilateral ND, and the fourth did not have a neck dissection.
Within the END group, 41 of 98 patients (41.8%) developed postoperative complications as compared to 13 of 27 (48.1%) no-END patients, and this difference was not significant (p = 0.558). However, there was a significantly higher proportion of complications in the patients with a history of chemoradiation as compared to those who were radiated (65% vs. 39%, p = 0.0346). The difference in fistula formation was not significant (p = 0.169), but patients with chemoradiation were observed to have significantly higher proportions of wound complications as compared to those who were radiated (40% vs. 17%, p = 0.0341). However, after multivariable logistic regression analysis, chemoradiation was no longer significantly predictive of increased odds for overall complications (OR = 0.99, 95% CI = 0.95–1.02) nor wound complications (OR = 2.6, 95% CI = 0.77–8.9). Complication demographics comparing no-END to END are summarized in Table 2, and Table 3 demonstrates associations between patient characteristics and complications.
| All patients, n = 125 | No neck dissection, n = 27 | Elective neck dissection, n = 98 | P value | |
|---|---|---|---|---|
| Total complications | 54 (43%) | 13 (48%) | 41 (42%) | 0.5577 |
| Salivary fistula/leak | 34 (27%) | 9 (33%) | 25 (26%) | 0.4186 |
| Wound infection | 14 (11%) | 2 (7.4%) | 12 (12%) | 0.7325 |
| Wound dehiscence | 20 (16%) | 2 (7.4%) | 18 (18%) | 0.2396 |
| Chyle leak | 0 | 0 | 0 | – |
| Hematoma | 5 (4.0%) | 2 (7.4%) | 3 (3.0%) | 0.2946 |
| Revision procedure | 20 (16%) | 5 (19%) | 15 (15%) | 0.7675 |
| Flap failure (out of 33 flaps) | 3 (2.4%) | 0 | 3 (3.1%) | – |
| Medical complications | 12 (9.6%) | 2 (7.4%) | 10 (10.2%) | 1 |
| Carotid blowout | 4 (3.2%) | 1 (3.7%) | 3 (3.1%) | 1 |
| Hospital stay | 0.6653 | |||
| < 7 | 50 (40%) | 9 (33%) | 41 (42%) | |
| 7–14 | 60 (48%) | 15 (56%) | 45 (46%) | |
| 14–28 | 15 (12%) | 3 (11%) | 12 (12%) | |
| Two or more complications | 25 (20%) | 6 (22%) | 19 (19%) | 0.7444 |
| Overall complications | P value | Fistula formation | P value | Wound complications | P value | |
|---|---|---|---|---|---|---|
| Treatment modality | 0.0346* | 0.169 | 0.0341* | |||
| Radiation | 41/104 (39%) | 26 (25%) | 18 (17%) | |||
| Chemoradiation | 13/20 (65%) | 8 (40%) | 8 (40%) | |||
| Site | 0.6231 | 0.6227 | 0.828 | |||
| Supraglottic | 7/17 (41%) | 5 (29%) | 4 (24%) | |||
| Glottic | 41/90 (46%) | 26 (29%) | 18 (20%) | |||
| Transglottic | 6/18 (33%) | 3 (17%) | 4 (22%) | |||
| Recurrent T stage | 0.556 | 0.0905 | 0.4906 | |||
| rT1/rT2 | 10/29 (34%) | 5 (17%) | 4 (14%) | |||
| rT3 | 20/44 (45%) | 17 (39%) | 9 (20%) | |||
| rT4 | 24/52 (46%) | 12 (23%) | 13 (25%) | |||
| Microvascular flap | 0.4641 | 0.3886 | 0.448 | |||
| Yes | 4/7 (57%) | 3 (43%) | 2 (29%) | |||
| No | 50/118 (42%) | 31 (26%) | 24 (20%) | |||
| Positive nodal disease on pathology | 0.3123 | 0.1175 | 0.6952 | |||
| Yes | 6/10 (60%) | 5 (50%) | 3 (30%) | |||
| No | 35/88 (40%) | 20 (23%) | 20 (23%) | |||
| Pathologic T stage | 0.3638 | 0.2328 | 0.1048 | |||
| T1/T2 | 11/31 (35%) | 5 (16%) | 3 (10%) | |||
| T3 | 15/38 (39%) | 13 (34%) | 7 (18%) | |||
| T4 | 28/56 (50%) | 16 (29%) | 16 (29%) | |||
| Laterality of neck dissection (n = 98) | 0.4954 | 0.6462 | 0.3326 | |||
| Bilateral | 23/51 (45%) | 14 (27%) | 14 (27%) | |||
| Unilateral | 18/47 (38%) | 11 (23%) | 9 (19%) |
3.3
Occult nodal disease
Among the 98 patients undergoing elective neck dissection, 10 (10.2%) had pathological evidence of nodal disease. Table 4 describes characteristics of these patients. No preoperative patient characteristic was significantly predictive of occult nodal disease.
| Initial treatment modality | Initial T stage | Recurrence site | Recurrence T stage | Preoperative imaging modality | Laterality of neck dissection | Status at last follow up | Overall survival in months | |
|---|---|---|---|---|---|---|---|---|
| Patient 1 | XRT | T1 | Transglottic | 4 | CT | Unilateral | Deceased | 5.4 |
| Patient 2 | XRT | T1 | Glottic | 2 | CT | Unilateral | Deceased | 12.5 |
| Patient 3 | XRT | T2 | Glottic | 4 | CT | Unilateral | Deceased | 4.4 |
| Patient 4 | CHMXRT | T4 | Glottic | 4 | PET/CT | Unilateral | Deceased | 66.3 |
| Patient 5 | XRT | UN | Glottic | 3 | CT | Unilateral | Deceased | 39.1 |
| Patient 6 | XRT | T2 | Supraglottic | 3 | MRI | Bilateral | Deceased | 31.4 |
| Patient 7 | CHMXRT | UN | Glottic | 4 | PET/CT | Unilateral | Deceased | 39.6 |
| Patient 8 | XRT | T1 | Glottic | 3 | CT | Unilateral | Alive, with disease | 16 (last follow-up) |
| Patient 9 | XRT | UN | Glottic | 4 | PET | Bilateral | Deceased | 5.3 |
| Patient 10 | XRT | T1 | Glottic | 3 | CT | Bilateral | Deceased | 5.9 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
