Impact of modified radical neck dissections on the number of retrieved nodes, recurrence and survival

Neck dissection is an integral part of head and neck tumors. Throughout its history, it has undergone changes looking for an improvement in functional outcome without loss of oncologic efficiency.


INTRODUCTION
Crille described radical neck dissection in 1906 as a procedure for en bloc removal of the sternocleidomastoid muscle, the internal jugular vein, the accessory nerve, and lymphatic and fatty tissues. The purpose of this procedure was to assure that most of the lymph nodes were removed, thereby maximizing its therapeutic effect and improving staging. 1 Martin later defended this idea by arguing in favor of removing the abovementioned three non-lymphatic structures as a method for achieving full neck lymphadenectomy. 2 Although full lymphadenectomy is achievable, classic radical neck dissection causes significant sequelae due to the removal of non-lymphatic structures. Thus, 40 years ago, efforts were made to preserve these structures -in particular the accessory nerve -when not affected by the tumor; these procedures became known as modified radical neck dissections. 3 This trend continued, and methods were found to preserve the internal jugular vein and the sternocleidomastoid muscle, which resulted in the so called functional neck dissection. 4,5 Notwithstanding the increasing popularity of these modified radical neck dissection procedures, a major issue is the impact of these modifications on surgical efficacy. We therefore investigated the number of recovered lymph nodes in each technique and the impact on regional recurrence and survival to test this effect.

MATERIAL AND METHOD
This retrospective study investigated patients who had undergone neck dissection with the removal of lymphatic levels I to V, with or without removal of the accessory nerve, the internal jugular vein, or the sternocleidomastoid muscle. Patients with tumors of the lower mouth and oropharynx diagnosed histologically as squamous cell carcinoma were included. Patients who underwent extended neck dissections, preoperative radiotherapy or chemotherapy, or prior surgery were not included. For the statistical analysis patients were grouped according to the number of structures preserved into three groups: radical neck dissection (RND), neck dissection preserving the accessory nerve (MRND + XI), and neck dissection preserving the accessory nerve and the internal jugular vein (MRND IJV+XI). The AJCC 2002 6 stating method was applied to all patients.
The Stata 10.1 (StataCorp, Texas, USA) software for MacOS X was used for the statistical analysis. Values are presented as the mean and standard deviation (SD). The analysis of variance (ANOVA) with Scheffé's test for multiple comparisons among groups was applied. The Kaplan-Meier and Cox methods were applied for assessing survival and regional recurrences; the hazard ratio (HR) value was indicated, and the confidence interval (CI) was 95%. Significant p values were p<0.05.

RESULTS
There were 481 patients that fit the inclusion criteria. There were 424 (88.15%) males and 57 (11.85%) females. The mean age at the time of treatment was 55.97 years (SD 10.0 years). The most common primary site was the oral tongue (187 cases; 38.88%), followed by the floor of the mouth (126 cases; 29.19%), the retromolar trigone (69 cases; 14.35%), and the tonsils and lower gingival margin (44 cases; 9.15%).  Table 2 shows the distribution of neck dissections according to each group and laterality.  (Table  3); this test revealed that significantly more lymph nodes were recovered in patients undergoing modified neck dissections.
Follow-up ranged from 0.3 to 322.23 months (mean -60.79 months, SD -22.39 months). Local recurrence occurred in 122 cases; 74 of these were neck recurrences, and 46 were distance recurrences; 250 patients had recurrences in at least one site. Statistically significant factors for neck recurrences were age (p=0.037), staging of the neck (p=0.006), primary tumor thickness (p=0.002), and postoperative radiotherapy (p=0.046). In this model the type of neck dissection had no impact on regional recurrence   (p=0.878, Table 4). Figure 1 shows the neck recurrence curves according to the type of neck dissection. Harrell's agreement test revealed a higher predictive ability of the model without including the type of neck dissection (Table  5). Statistically significant factors for disease-free survival were the T stage (p=0.001), the N stage (p=0.001), the presence of lymphatic vascular embolization (p=0.019), and tumor thickness (p=0.008). The type of neck dissection had no significant impact (p=0.185, Table 6). Survival curves on Figure 2 show a similar behavior in both groups. Harrell's agreement test showed that the model without the type of neck dissection was superior to the model with the type of neck dissection (Table 7).

DISCUSSION
The extent of neck dissection has been an important topic for debates. Based on surgical and autopsy data, removal of lymphatic and fatty tissues without removing non-lymphatic structures was found to be oncologically sound. Neck lymph nodes are located within fibro-adipose tissue next to nerves and blood vessels; there are no lymph nodes in the muscle aponeurotic fascia. 4 Bocca and Pignataro have confirmed these findings. 5 Resection of the accessory nerve has been routinely avoided in selected cases since the 1950s because of the associated morbidity. It is preserved whenever there is a clear cleavage plan between the nerve and compromised lymph nodes. 7 A criticism of modified radical neck dissections is the decrease in recovered lymph nodes relative to the number of preserved non-lymphatic structures. According to Busaba et al., classical radical neck dissection had a significantly higher number of recovered lymph nodes compared to the variations of this procedure. 8 These authors suggested that this could have a negative impact on the prognosis of patients. However, Siddiquee et al. contested this finding by showing that classical and modified radical neck dissection had comparable oncological efficacy. 9 This confirmed previously published papers that had also demonstrated a similar oncological efficacy of modified neck dissection, which has lower morbidity, especially in shoulder function. 10

CONCLUSION
Our series showed that the modified radical neck dissection has a similar oncological efficacy to the classical radical neck dissection, and that contrary to previous arguments in the literature, the number of recovered lymph nodes is higher than that in classical neck dissection. There is thus no loss in staging on pathology and the prognosis of patients, or in their survival and disease-free interval.