Introduction
Radical radiotherapy (RT) or combined chemoradiotherapy (CRT) are organ preserving treatment modalities for locally advanced head and neck cancer.Reference Paleri, Urbano, Mehanna, Repanos, Lancaster and Roques1 Recurrent or persistent nodal metastasis after RT or CRT is common and remains a challenge to successful treatment of these cancers. A study done by ven der Putten et al. showed that 129 (23.9 per cent) patients out of a total of 540 patients developed regional recurrence or residues after CRT, with 68 of them who were thought to have unresectable nodal disease.Reference van der Putten, van den Broek, de Bree, van den Brekel, Balm and Hoebers2 Although most would agree that salvage neck dissections (ND) are required in these circumstances, the extent to which the surgery should be performed lacks evidence. Until recently, dissection of all five neck levels was advocated as the gold standard treatment.Reference Murthy, Kundu, Budrukkar, Gupta, Laskar and Krishnatry3,Reference Lavertu, Adelstein, Saxton, Secic, Wanamaker and Eliachar4 However, studies on selective ND have reported equal or superior effectiveness with less morbidity when compared to the more radical approach.Reference van der Putten, van den Broek, de Bree, van den Brekel, Balm and Hoebers2,Reference Stenson, Haraf, Pelzer, Recant, Kies and Weichselbaum5-Reference Mukhija, Gupta, Jacobson, Eloy and Genden12 The rationale behind this approach is that nodal metastases in head and neck cancers tend to spread in a predictable patternReference Lindberg13,Reference Li, Wei, Guo, Yuen and Lam14 and that the RT or CRT would obliterate the majority of lymph node micrometastases.Reference Robbins, Doweck, Samant and Vieira15
Besides selective ND, a few reports on superselective neck dissection (SSND) were published to propose an even more limited ND as a salvage treatment option.Reference Robbins, Doweck, Samant and Vieira15–Reference Robbins, Dhiwakar, Vieira, Rao and Malone17 Although there is no universally agreed definition of SSND, these studies define SSND as removal of all node-bearing tissue of one to two adjacent lymph node levels.Reference Robbins, Shannon and Vieira16 A recent study by Okano et al. showed that an even targeted single level ND could be considered to lower complication rates while maintaining oncological outcomes.Reference Okano, Hayashi, Matsuura, Shinozaki and Tomioka18 The study evaluated the outcomes of patients who underwent ND of the clinically abnormal levels only without removal of adjacent uninvolved neck levels. The three-year survival rate and disease-specific survival rate were reported as 59 per cent and 66 per cent, respectively, which is in keeping with or superior to studies on selective NDReference van der Putten, van den Broek, de Bree, van den Brekel, Balm and Hoebers2,Reference Dhiwakar, Robbins, Vieira, Rao and Malone9
To our knowledge, there is no literature in the United Kingdom that investigates the extent of salvage ND. Therefore, we aim to investigate if selective ND can be safely performed in patients with recurrent or residual nodal metastasis while maintaining acceptable oncological and post-operative outcomes as determined by complications, survival rates and regional control.
Materials and methods
From January 2016 until December 2018, a total of 266 patients with suspected recurrent head and neck cancer were referred to the West of Scotland Head and Neck Cancer Multidisciplinary Team meeting. Only patients with newly diagnosed squamous cell carcinoma (SCC) of the head and neck who were subsequently treated with CRT or RT followed by salvage ND were included in this analysis (n = 53). Exclusion criteria were bone tumours, skin tumours, patients who received surgery only for the original treatment and patients without salvage ND. Of those patients, all 53 patients were subsequently analysed in our study.
All patients were staged according to the TNM 8 AJCC/UICC staging system.Reference Zanoni, Patel and Shah19 RT was carried out 5 days a week with 80 per cent (41/51) of patients receiving 65 Gy in 30 fractions and the total dose ranged from 55 Gy to 66 Gy. The CRT regimen consists of the above regimen with concomitant two cycles of cisplatin (15 patients) or various cycles of cetuximab (4 patients). The date of completion of initial treatment was recorded to identify the time it took for recurrent or persistent neck disease to be detected clinically or radiologically in addition to subsequent survival following salvage ND.
ND was performed if there was equivocal or residual response on post-treatment positron emission tomography–computed tomography (PET-CT) imaging. The types of salvage ND performed were based on recommendation by the multidisciplinary team meetings and the surgical procedures occurred at variable intervals after CRT or RT. The type of ND was defined by the Neck Dissection Classification Update Revisions proposed by the American Head and Neck Society and the American Academy of Otolaryngology-Head and Neck SurgeryReference Robbins, Clayman, Levine, Medina, Sessions and Shaha20 with the addition of superselective ND (SSND) as defined above. The nodal yield and number of pathologic lymph nodes were recorded.
Post-operative complication rates were recorded in our study including haemorrhage, infection, cranial nerve injury, osteoradionecrosis, Horner's syndrome and admission for pneumonia. We calculated the regional control rate and survival rate on the Kaplan–Meier plot. Values of p less than 0.05 were considered statistically significant for all results of this study.
Local Caldicott Guardian approval was granted, and after consultation with the online HRA Tool,21 formal ethical review was not required.
Results and analysis
Patient demographics
Fifty-three patients were included. The median age was 59 years (SD 11.1). Most patients were male (44, 83 per cent). Twenty-nine (55 per cent) of the study population were current smokers, and 7 (13 per cent) had never smoked (Table 1).
Table 1. Patient demographics, types of treatment regime and nature of recurrence
The most common primary site was larynx (24, 45.3 per cent) followed by oropharynx (13, 24.5 per cent), oral cavity (8, 15.1 per cent), hypopharynx (5, 9.4 per cent), unknown primary (2, 3.8 per cent) and nasopharynx (1, 1.9 per cent). The most common pre-treatment T and N classification in our study population was T2 (22, 41.5 per cent) and N0 (28, 52.8 per cent), respectively (Table 2). Twenty-four patients had initial classification of N1 or above, and 13 (54.2 per cent) of them displayed evidence of extracapsular spread (ECS) on imaging. Human papilloma virus (HPV)/p16 status was recorded in 29 patients, 12 (41.4 per cent) with positive and 17 (58.6 per cent) with negative HPV/p16 status. HPV/p16 status was available in all patients with oropharyngeal cancer, 11 (84.6 per cent) with positive HPV/p16 status. The primary treatment was RT in 27 (50.9 per cent), and 18 (34.0 per cent) patients underwent CRT. The remainder received either surgery followed by RT (7, 13.2 per cent) or surgery followed by CRT (1, 1.9 per cent).
Table 2. T and N staging of tumour
Regarding the nature of recurrence, 10 (18.9 per cent) had equivocal response of nodes on PET-CT following primary treatment, and 38 (71.7 per cent) patients presented with failure of either primary site or nodes. Recurrence was identified through surveillance in five (9.4 per cent) patients.
BMI = body mass index; CRT = combined chemoradiotherapy; FNA = fine needle aspiration; PET-CT = positron emission tomography–computed tomography imaging; RT = radical radiotherapy
Salvage ND data
Of the 53 neck dissections performed, 62 per cent33 of the patients had tumour resection of the primary site in combination with the ND. A total of 42 per centReference Dhiwakar, Robbins, Rao, Vieira and Malone22 of the ND were bilateral ND. In our study, dissection of neck levels II–IV was the most prevalent procedure (20, 37.7 per cent) (Table 3).
Table 3. Nodal levels dissected and corresponding pathological features
Pathologic features
Pathologic examination of the dissected lymph nodes confirmed the presence of positive SCC in 43.4 per centReference Goguen, Posner, Tishler, Wirth, Norris and Annino23 of the patients. 20 (37.7 per cent) patients had evidence of ECS or extranodal extension (ENE) in the resected specimen. The median number of nodes dissected per patient was 25, which ranged from 2 to 107 nodes. In patients with an equivocal nodal response on PET-CT, the rate of ND with viable SCC was 70 per cent (7/10 patients).
Post-operative complications
The most common complications after salvage ND were wound infection (13, 24.5 per cent), admission after surgery with pneumonia (12, 22.6 per cent), evidence of aspiration on videofluoroscopy (6, 11.3 per cent) and haemorrhage (5, 9.4 per cent). Post-operative infection, injuries to the accessory (15.4 per cent) and vagal (7.7 per cent) nerves were more common in patients with dissection of levels I–V than patients with dissection of levels II–IV (Table 4).
Table 4. Post-operative complication rates comparing dissection of levels I–V and levels II–IV
In patients who did not have viable SCC in their dissections (n = 30), the complication rate was 43.3 per cent (13/30). For patients with cancer identified in their dissections, the complication rate was 69.6 per cent (16/23).
Outcomes
The three-year overall survival rate was 50.9 per cent. Overall survival was not significantly different for patients with levels II–IV dissection and patients with levels I–V dissection (p = 0.52) (Figure 1). The three-year survival rate of HPV-positive patients (78.6 per cent) was better than HPV-negative patients (52.9 per cent) (p = 0.13). Patients with positive lymph nodes on pathologic examination experienced worse three-year survival outcome (21.7 per cent) than those without positive lymph nodes (56.7 per cent) (p = 0.01). Patients with evidence of ECS or ENE had a poor survival rate (20.0 per cent) compared with patients with pathological lymph nodes without evidence of ECS or ENE (100.0 per cent) (p = 0.03).
Figure 1. Overall survival following salvage ND according to two groups of different levels of ND
Discussion
Extent of ND
Our study clearly demonstrated that selective ND produces less post-operative complications than radical ND while maintaining three-year survival outcome as the traditional gold standard treatment. Dissection of levels II–IV had a three-year survival rate of 60 per cent, which is comparable to that seen of other studies by van der Putten et al.,Reference van der Putten, van den Broek, de Bree, van den Brekel, Balm and Hoebers2 Dhiwakar et al.Reference Dhiwakar, Robbins, Vieira, Rao and Malone9,Reference Dhiwakar, Robbins, Rao, Vieira and Malone22 and Okano et al.Reference Okano, Hayashi, Matsuura, Shinozaki and Tomioka18 (Table 5).
Table 5. Comparison of three-year overall survival rates between different studies on selective ND
When comparing between dissection of levels II–IV and levels I–V only without resection of primary site, our results showed that post-operative infection, accessory nerve injury and vagal nerve injury were more common in patients with radical ND than selective ND.
• Salvage neck dissection for squamous cell carcinoma is performed for residual or recurrent nodal disease after chemoradiotherapy or radical radiotherapy for locally advanced head and neck cancer
• Selective ND produces less post-operative complications than radical ND
• No significant difference exists in overall three-year survival rate between patients with levels II–IV dissection and that of levels I-V dissection (p = 0.84)
• More than half (56.6 per cent) of the dissections did not demonstrate the presence of viable squamous cell carcinoma (SCC) in lymph nodes
• Positive human papilloma virus (HPV) status, absence of pathologic lymph nodes and absence of extracapsular spread (ECS) or extranodal extension (ENE) were associated with an increased three-year survival after salvage neck dissection
Pathologic features
More than half (56.6 per cent, 30/53) of the dissections did not demonstrate the presence of viable SCC in lymph nodes. A total of 73.9 per cent (17/23) of patients had pathological lymph nodes in levels II and III only, which suggests most nodal metastasis involves these two neck levels first before spreading to other neck levels. Interestingly, several studies on post-RT or post-CRT ND also revealed a large proportion of dissections with no proven evidence of nodal metastases after histopathological examination.Reference van der Putten, van den Broek, de Bree, van den Brekel, Balm and Hoebers2,Reference Stenson, Haraf, Pelzer, Recant, Kies and Weichselbaum5,Reference Hoch, Bohne, Franke, Wilhelm and Teymoortash6,Reference Doweck, Robbins, Mendenhall, Hinerman, Morris and Amdur8–Reference Mukhija, Gupta, Jacobson, Eloy and Genden12,Reference Robbins, Doweck, Samant and Vieira15,Reference Robbins, Shannon and Vieira16,Reference Goguen, Posner, Tishler, Wirth, Norris and Annino23,Reference Boyd, Harari, Tannehill, Voytovich, Hartig and Ford24
Of the 28 patients with an original staging of N0, 5 (17.9 per cent) of them had positive lymph nodes after salvage ND. Only 1 of these patients (20.0 per cent) survived over three years compared to 14 patients (14/23, 60.9 per cent) who were N0 on original staging and without positive lymph nodes after salvage ND.
Nodal yield has been shown in studies to impact survival rates in treatment naïve patients.Reference Ebrahimi, Zhang, Gao and Clark25–Reference Merz, Timmesfeld, Stuck and Wiegand27 Ebrahimi et al.Reference Ebrahimi, Zhang, Gao and Clark25 and de Kort et al.Reference de Kort, Maas, Van Es and Willems26 concluded that the dissection should include at least 18 lymph nodes, whereas Merz et al.Reference Merz, Timmesfeld, Stuck and Wiegand27 had suggested that the removal of at least 15 lymph nodes was enough to improve survival. In our study, the median number of nodes dissected per patient was 25. Ebrahimi et al.Reference Ebrahimi, Zhang, Gao and Clark25 showed that there was no survival difference with increased number of dissected levels although mean nodal yields were increased in patients with more levels dissected.
Factors that affect survival rate
Our study showed that positive HPV status, absence of pathologic lymph nodes and absence of ECS or ENE were associated with an increased three-year survival after salvage ND. Positive HPV status, in particular, has been shown to be a positive prognostic factor in head and neck cancers.Reference Clark, Holmes, O'Connell, Harris, Seikaly and Biron28–Reference Ang, Harris, Wheeler, Weber, Rosenthal and Nguyen-Tan30
Limitations
Due to the specific nature of this study, we had a small sample size of 53 patients over three years which limits the value of statistical analysis. As we have a relatively small sample size, we included patients who underwent primary site resection and salvage ND simultaneously, which could affect the reliability of our analysis on the extent of neck dissection. Another limitation is that we included patients with various initial treatments, such as RT, CRT, primary surgery followed by RT and primary surgery followed by CRT. Our study was conducted retrospectively, which limited data collection.
Conclusion
In conclusion, our study demonstrated that selective ND reduces post-operative complications while maintaining similar survival rate when compared with more extensive dissections.
Acknowledgements
We thank Joanna Murnane for her assistance in data collection for this paper.
Funding
The authors did not receive financial support from any organization for the submitted work.
Competing interests
The authors declare no conflict of interest for the submitted work.
