Introduction
Early glottic cancer can be treated with surgery or radiotherapy (RT), with no significant difference in oncological outcome. However, RT appears to ensure better voice and swallowing results.Reference Higgins, Shah, Ogaick and Enepekides1 Nevertheless, the recurrence rate after primary RT is 5–13 per cent for tumour–node–metastasis (TNM) stage T1 and 25–30 per cent for T2 carcinomas.Reference Motamed, Laccourreye and Bradley2, Reference Cellai, Frata, Magrini, Paiar, Barca and Fondelli3 Surgery is the mainstay of salvage treatment in patients with recurrence after RT failure,Reference Agra, Ferlito, Takes, Silver, Olsen and Stoeckli4, Reference Sher, Haddad, Norris, Posner, Wirth and Goguen5 but management of recurrent clinically diagnosed node-negative (N0) neck tumours remains controversial and, at the time of salvage laryngeal surgery, surgeons must decide whether or not to perform elective neck dissection. Previous reports indicate an incidence of 10–28 per cent occult neck metastasis in persistent and recurrent laryngeal disease after primary RT.Reference Yuen, Wei and Wong6–Reference Zbären, Christe, Caversaccio, Stauffer and Thoeny9 Accurate restaging with scrupulous reassessment of the nodal status cannot completely rule out the presence of occult neck disease.Reference Spiro, Gallo and Shah10–Reference Gallo, Deganello, Scala and De Campora12 Small nodal metastases are still a diagnostic challenge despite continuing improvements in imaging techniques. Computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography and ultrasound-guided fine needle aspiration cytology are all generally used in clinical practice to stage head and neck cancers, and can be useful for detecting cervical lymph node metastases. However, their diagnostic accuracy for sub-centimetre metastases and micrometastases is limited. Positron emission tomography has a low specificity and sensitivity for nodal metastases of diameter less than 6 mm.Reference Brouwer, de Bree, Comans, Castelijns, Hoekstra and Leemans13, Reference Phillips, Smith, Parvathaneni and Laramore14 Recently, CT perfusion scans have been shown to depict nodal metastases of less than 1 cm in diameter.Reference Trojanowska, Trojanowski, Bisdas, Staśkiewicz, Drop and Klatka15 In untreated patients, cervical lymphatic tumour spread follows a predictable path;Reference Spiro, Gallo and Shah10 this knowledge has led to improvements in the surgical management of clinically staged N0 laryngeal cancer.Reference Deganello, Gitti, Meccariello, Parrinello, Mannelli and Gallo11, Reference Gallo, Deganello, Scala and De Campora12 Nowadays, it is widely accepted that the incidence and location of occult metastases are related to tumour stage and site. Selective neck dissection of levels II–IV (formerly known as ‘jugular node dissection’Reference Spiro, Gallo and Shah10) is considered adequate in patients with clinically staged N0 laryngeal cancer, while modified radical neck dissection is now considered obsolete.Reference Deganello, Gitti, Meccariello, Parrinello, Mannelli and Gallo11, Reference Gallo, Deganello, Scala and De Campora12 Furthermore, to minimise post-operative morbidity, super-selective dissection of sublevels IIa–III has recently been proposed because of the rare involvement of sublevel IIb and level IV in clinically staged N0 laryngeal carcinomas.Reference Ferlito, Silver and Rinaldo16, Reference Santoro, Franchi, Gallo, Burali and de' Campora17 In a previously irradiated neck, however, the usual regional lymphatic pathway for both the larynx and neck may have been extensively modified by radiation-induced sclerosis of the lymphatic vessels. Hence, a post-RT recurrent laryngeal tumour could spread to lymph nodes beyond the original predicted pathway. To date, there are no consistent data describing the prevalence of neck metastases and the incidence and location of occult neck metastases following RT failure for initial early glottic cancer.
This study aimed to investigate the appropriateness of elective neck dissection during salvage laryngeal surgery in patients with recurrent clinically staged N0 tumours following RT failure for initial early glottic cancer. We also aimed to identify the subpopulation at the highest risk of occult neck disease and the neck levels at the highest risk of regional tumour spread.
Materials and methods
We assessed all patients from 1995 to 2005 with local recurrence of early laryngeal cancers (T1–2N0) following elective RT, who had undergone salvage surgery.
Inclusion criteria were no previous treatment, primary curative RT for early glottic carcinoma (T1–2N0), biopsy-proven laryngeal carcinoma after radiation, and salvage surgery. Our case series comprised 107 male patients and 3 female patients, with a median age of 63.2 years (range 47–83 years). Tumours had been initially staged as T1N0 in 75 patients and as T2N0 in 35 patients. The total dose varied from 60 to 70 Gy: 69 patients (62.7 per cent) received a mean dose of 65.0 ± 3.0 Gy to the larynx alone, while 41 patients (37.3 per cent) were irradiated with a mean dose of 64.3 ± 2.3 Gy to the larynx and a mean dose of 52.8 ± 5.0 Gy to the neck. Follow-up time was calculated from the date of salvage surgery to the date of the last follow up or death. Follow up ranged from 1 to 200 months, with a mean of 45.42 ± 44.36 months. Head and neck examinations were scheduled every month for the first year, every two months for the second year, every three months for the third year, and every six months thereafter until the fifth year. Magnetic resonance imaging and/or CT scans of the head and neck were performed twice a year for the first three years, and then annually. Patients with suspicious lesions underwent panendoscopy with biopsy and imaging. Persistent and recurrent laryngeal cancers were originally classified according to the current TNM Classification of Malignant Tumours; for this study, each tumour was re-evaluated and restaged according to the seventh edition of the TNM staging system, using information in the clinical records.Reference Edge, Byrd, Compton, Fritz, Greene and Trotti18 Radiological criteria to define a N0 neck tumour were based on the characterisation of lymph nodes detected by a contrast-enhanced CT scan: diameter less than 10 mm, no central necrosis and no contrast enhancement of the lymph node capsule.Reference Deganello, Gitti, Meccariello, Parrinello, Mannelli and Gallo11 No specific indications for elective neck dissection were used in this study. Therefore, elective neck procedures were performed according to the surgeon's preference. The extent of neck dissection ranged from selective neck dissection of levels II–IV to comprehensive modified radical neck dissection. Modified radical neck dissection was performed in the earlier cases prior to implementation of an institute policy recommending selective neck dissection (levels II–IV).Reference Spiro, Gallo and Shah10
Descriptive statistics were compared using Student's t-test for continuous variables and Fisher's exact test for categorical variables. The Kaplan–Meier method was used for survival analysis. All analyses were performed using STATA version 10.0 software (StataCorp, College Station, Texas, USA). A p value of less than 0.05 was considered statistically significant.
Results
After RT, the median time to local recurrence was 23.5 ± 33.2 months (range 1–216 months, 95 per cent confidence interval (CI) 17.25 to 29.8). The most common salvage procedure was total laryngectomy (90 patients, 81.8 per cent), while only 20 patients (18.2 per cent) underwent salvage partial laryngectomy without elective or therapeutic neck dissection. Restaging showed 34 recurrent T1, 36 recurrent T2, 29 recurrent T3 and 11 recurrent T4a tumours; 6 patients also had clinically staged neck metastases (1 recurrent T4aN1, 2 recurrent T2N1, 1 recurrent T3N1, 1 recurrent T2N2b and 1 T2N2c). Neck level II was involved in four cases (three with N1 and one with N2btumours), and neck levels III and IV were involved in only two cases; all of these patients underwent radical neck dissection or modified radical neck dissection. Of the remaining 104 patients, 97 were managed using a ‘wait-and-see’ protocol for neck relapse and 7 patients with recurrent N0 tumours underwent elective neck dissection: 2 with ipsilateral selective neck dissection (levels II–IV), 2 with bilateral selective neck dissection (levels II–IV) and 3 with ipsilateral modified radical neck dissection. The characteristics of both groups at original diagnosis are shown in Table I; tumour stages were similar in both patient groups. No occult lymph node metastases were documented in the seven cases of elective neck dissection.
Table I Patient characteristics
Post-operative medical and surgical complications are summarised in Table II. There was a higher rate of post-operative complications in the elective neck dissection group than in the ‘wait-and-see’ group (71.4 per cent vs 26.8 per cent, respectively, p = 0.02, Fisher exact test), resulting from a higher rate of pharyngocutaneous fistula development (57.2 per cent vs 13.4 per cent, respectively). Furthermore, 2 (2.1 per cent) patients in the ‘wait-and-see’ group died as a result of peri-operative myocardial stroke. The median follow up after salvage surgery was 45.4 ± 44.4 months (range 1–200 months, 95 per cent CI 37.04 to 53.8), while the median time to a second local or locoregional relapse was 36.5 ± 36.3 months (range 1–200 months, 95 per cent CI 29.6 to 43.4). The incidence of local, regional and distant failure is shown in Table III. In the ‘wait-and-see’ group, three patients (one with a recurrent T3 tumour and two with recurrent T2 tumours) developed pathologically staged N1 or N2b neck metastasis without local recurrence; in all cases, level II was the only level involved. Of note, post-operative local recurrences were significantly higher in the neck dissection group (n = 5, 71.5 per cent) than in the ‘wait-and-see’ group (n = 27, 27.8 per cent, p = 0.028). Distant metastases were detected in two patients in the ‘wait-and-see’ group (2.1 per cent). At the end of the study, the overall survival rate was 42.9 per cent in neck dissection patients, and 62.0 per cent in the ‘wait-and-see’ group (shown in Figure 1). No significant differences in overall survival were observed between patients with recurrent node-positive (N+) tumours receiving therapeutic neck dissection at the time of salvage laryngectomy and those with recurrent N0 tumours receiving elective neck dissection or salvage laryngectomy alone (p = 0.071; Figure 2).
Fig. 1 Kaplan–Meier curve showing overall survival of patients who underwent simultaneous elective neck dissection and salvage laryngeal surgery (NP) and those who underwent only salvage laryngeal surgery (WP) for recurrent early glottic cancer. (p = 0.7)
Fig. 2 Kaplan–Meier curve showing overall survival after curative radiation therapy for node-negative (N0) vs node-positive (N+) tumours. (p = 0.071)
Table II Post-operative complications for patients undergoing salvage surgery
*n = 7. †n = 97.
Table III Sites of recurrence after salvage surgery in patients who died from the disease
*One patient had multiple metastatic sites (neck + hypopharynx); †percentage of total group; ‡percentage of patients with local tumours.
Multivariate regression analysis demonstrated that an initial T1 stage tumour and age less than 60 years correlate with a higher risk of neck failure after curative RT (Table IV; p = 0.04 and p = 0.038, respectively). Indeed, Fisher's exact test confirmed that a younger age was significantly associated with a risk of neck metastasis following local recurrence (p < 0.001).
Table IV Multivariate analysis of neck failure risk after curative radiation therapy
SEM = standard error of the mean; CI = confidence interval; T1 = tumour–node–metastasis stage T1
Discussion
Radiotherapy (RT) is often indicated for early glottic cancers to preserve organ function and post-treatment voice quality compared with surgery. Unfortunately, RT may be unsuccessful and some patients will experience recurrence or persistent laryngeal cancer. The Union for International Cancer Control and American Joint Committee on Cancer staging criteria for laryngeal recurrences are identical to those for untreated primary tumours, except for addition of the prefix ‘recurrent’ applied to the TNM categories. Regional recurrence after curative RT correlates with poor prognosis, independent of the initial tumour stage.Reference Lacy and Piccirillo19, Reference Leon, Lopez, Garcia, Viza, Gich and Quer20 Our data confirmed this finding: we showed that prognosis after salvage laryngectomy was more strongly associated with TNM restaging than with initial staging.
Previous studies reported a rate of occult neck metastases of 4–28 per cent in supraglottic and glottic cancer patients treated with RT alone.Reference Yuen, Wei and Wong6, Reference Yao, Roebuck, Holsinger and Myers7, Reference Bohannon, Desmond, Clemons, Magnuson, Carrol and Rosenthal21 This is especially true for advanced laryngeal carcinomas after failure of chemoradiation protocols; however, there are few data for early glottic cancers that recur after RT. In fact, at the time of first diagnosis, these patients have a very low risk of occult neck disease. After RT failure, both the risk of cancer spreading to other laryngeal sites and frequent progression or upstaging of persistent or recurrent lesions may lead to a higher rate of occult neck disease. We analysed only cases of persistent or recurrent cancers after primary RT that initially presented as early glottic laryngeal carcinomas. To the best of our knowledge, no other studies have considered this specific patient cohort. Thus, there is very little information on the correct management of recurrent N0 neck tumours when salvage laryngectomy is indicated.
In our case series, 6 patients who underwent therapeutic neck dissection after primary RT had recurrent N+ tumours (5.5 per cent; 4 with N1 and 2 with N2 tumours) and 3 out of 97 patients in the ‘wait-and-see’ group had neck recurrence at follow up after salvage laryngeal surgery. Pathological examination showed that no neck dissection patient had occult neck disease and none experienced neck treatment failure during follow up. Thus, in our series, neck disease was detected in only nine patients (8.7 per cent; six with recurrent N+ tumours at the end of primary RT and three with recurrent N+ tumours within the ‘wait-and-see’ group at follow up). There was a higher rate of neck metastasis development in patients aged less than 60 years (p < 0.001), and an increased risk in those who originally presented with a T2 glottic carcinoma or at a younger age (p = 0.04 and p = 0.038, respectively). Furthermore, the overall neck recurrence rate was 2.9 per cent after salvage laryngeal surgery. This low incidence of neck disease may be related to radiation-induced neck sclerosis sealing lymphatic vessels and thus hampering cancer lymphatic embolisation. Nevertheless, extension of the irradiation field to include the neck (larynx plus neck levels II–IV) did not completely protect from regional failure development, although it seemed to delay it. In fact, all six patients with recurrent N+ tumours received RT to the larynx alone, while the radiation field was extended to the neck in all three recurrent N0 patients.
• Surgery is the main salvage treatment for laryngeal cancer patients with recurrence after radiotherapy
• The management of clinically negative cervical lymph nodes in persistent or recurrent laryngeal cancers after primary radiotherapy is controversial
• For initial early glottic cancer recurring after radiotherapy without clinical or radiological lymph node involvement, salvage laryngeal surgery alone may be appropriate regardless of restaging
• Elective neck dissection is not recommended because of the increased risk of post-operative complication with no survival advantage
The five-year survival rate was 57.14 per cent in the small group of neck dissection patients and 64.6 per cent in the ‘wait-and-see’ group; however, the difference was not statistically significant (p = 0.7). Nevertheless, this result is consistent with the demonstrated impact of lymph node metastases on head and neck cancer prognosis. Several recent studies concluded that neck dissection is indicated for recurrent advanced stage laryngeal cancer.Reference Yao, Roebuck, Holsinger and Myers7, Reference Wax and Touma22 However, these studies failed to demonstrate a survival advantage in patients who underwent neck dissection. Some authors have advocated bilateral neck dissection for all advanced stage (T3 and T4) glottic recurrences and for all supraglottic recurrences regardless of stage.Reference Wax and Touma22 In our series, there were 38 recurrent T3–4 cancers, and 8 of these had supraglottic involvement in the recurrence. Furthermore, we documented involvement of the supraglottic larynx in 12 T2 recurrences. Among these 50 patients, 4 underwent elective neck dissection with no evidence of lymph node metastasis at pathological examination; in the remaining 46 patients, only 1 showed neck relapse at follow up. Furthermore, neck dissection did not lead to an overall survival benefit in our case series, and neck dissection did not provide additional protection from local recurrence (71.5 per cent vs 27.8 per cent, p = 0.028). Nevertheless, no regional failure was recorded in neck dissection patients, while only three patients in the ‘wait-and-see’ group experienced neck failure (0.0 per cent vs 3.1 per cent, p = 1.0). However, these data are not consistent with other reports.Reference Bohannon, Desmond, Clemons, Magnuson, Carrol and Rosenthal21, Reference Wax and Touma22 Conversely, neck dissection during salvage laryngeal cancer was associated with an increased rate of post-operative complications (71.4 per cent in the neck dissection group vs 26.8 per cent in the ‘wait-and-see’ group, p = 0.02), especially pharyngocutaneous fistulas (Table II; 57.2 per cent), as previously reported.Reference Wax and Touma22–Reference Gallo, Deganello, Gitti, Santoro, Senesi and Scala24 These discrepancies suggest a selection bias toward treating RT-recurrent laryngeal carcinoma patients with salvage laryngeal surgery and concomitant neck dissection, and a possible negative impact of uni- or bilateral neck procedures on devascularised irradiated soft neck tissue, associated with a higher risk of locoregional complications, especially pharyngocutaneous fistulas.
Conclusion
Patients who experience recurrent or persistent laryngeal cancer after primary RT for early glottic cancer without clinical and radiological involvement of cervical lymph nodes benefit from exclusive salvage laryngeal surgery. The prevalence of neck disease in these patients is less than 10 per cent, and the incidence of occult neck disease is less than 5 per cent. This is also true for recurrent T3–4 lesions and for extension to the supraglottic or hypoglottic areas. Moreover, elective neck dissection is not recommended in these cases because it exposes patients to an increased risk of post-operative complications without a definite survival advantage.
