Quantitative analysis of lymph nodes in neck dissection specimens . Morphologic study

PURPOSE: To quantify the amount of lymph nodes harvested in modified radical neck dissection. METHODS: Cross-sectional anatomical study conducted in 28 non-preserved cadavers. RESULTS: The mean number of lymph nodes found in each nodal level of the 56 modified radical neck dissections performed were: level IA – 1.5 (95% CI: 1.1 – 1.8), level IB – 2.5 (95% CI: 2.1 – 2.9), level IIA – 7.2 (95% CI: 6.0 – 8.5), IIB level – 6.5 (95% CI: 5.5 – 7.4), level III – 6.6 (95% CI: 5.7 – 7.4), level IV – 8.6 (95% CI: 7.1 – 10.1), level V – 11 (95% CI: 9.2 – 12.7), totalizing 43.8 lymph nodes (95% CI: 40.3 – 47.4). CONCLUSION: The results defined a parameter in relation to the minimum recommended nodal yield in a modified radical neck dissection, as well as the number of lymph nodes in each level of this dissection, performed in clinical practice.


Introduction
Head and neck cancers have great potential for regional spread by metastasis to lymph nodes (LNs).Except for the presence of distant metastases, the most adverse prognostic factor in head and neck squamous cell carcinoma (HNSCC) is regional nodal involvement 1 .Treatment of this disease requires the knowledge of this dissemination mechanism, the distribution of LNs in specific anatomical regions and appropriate surgical techniques for the approach, which is called neck dissection (ND).
The number of LNs expected to be retrieved in a specimen of modified radical neck dissection (MRND) is uncertain 2 , with variable references in the literature (mean of 31-55 LNs), depending on the processing method by the pathologist and the surgeon's ability for the complete removal of these nodes located within the anatomic boundaries of the neck nodal levels [3][4][5][6][7][8][9] .Pathological assessment (macroscopic and microscopic) of the ND specimen becomes extremely important, providing data as the total number of LN recovered (whether metastatic or not), the size of the metastasis, the presence of extracapsular spread or micrometastasis.Pathological status of the neck (pN at TNM classification) is one of the determining factors in the indication of adjuvant therapy 6 .
This study aims to quantify the number of lymph nodes retrieved in modified radical neck dissection specimens performed in a series of non-preserved cadavers, and to verify if there is any relation between the number of lymph nodes and demographic and anthropometric data of the individuals.

Methods
The authors declare no conflict of interest and all the cadavers were treated with the highest respect.The research protocol was approved by the Institutional's Research Ethics Committees (protocol number: 312/2013).

Sample estimation
Mean estimation modelling was chosen for sample size calculation 10 , and it was based on the study of Agrama et al. 3 that summarized the results of six studies, totalizing 5065 valid patients and establishing as weighted mean 35.2 LNs in a radical ND and weighted standard deviation of 5.9 LNs.Thus, using this study as reference and adopting standard deviation of 6 LNs and maximum error of the estimative of 1.75 LNs, with an overall significance level of 0.05, the sample size calculated was 45 ND.Estimating a possible loss of data of 25%, the final number of 56 specimens (28 cadavers) was assumed to be the goal.

Anatomic study
The cadavers were placed in supine position with the head in neutral position and arms parallel to the body.A subscapular cushion was used for neck extension.Demographic (sex, age and race) and anthropometric (weight, height and body mass index -BMI) data were recorded.
Exclusion criteria were: previous cervical manipulation, the presence of head and neck neoplasm and any anatomic abnormality on the neck.
A wide cervicothoracic incision was performed on each cadaver and the steps of modified radical neck dissection were reproduced, with the removal of fibrofatty tissue containing the deep cervical lymph nodes, in a level-by-level basis, according to the conventional classification and anatomical limits 11 (Figure 1).As a small change in the technique, the sternocleidomastoid muscle was sectioned in its upper third (on the junction level with the spinal accessory nerve) to provide full access to the jugular LNs.The final aspect of the dissection and the retrieved bilateral cervical nodal levels are shown in Figure 2.  The harvested tissue of each level and sublevel, dissected within the limits recommended by the Committee for Neck Dissection Classification of the American Head and Neck Society (AHNS) 12 , was delivered to the pathologist and dissected manually.Palpable LNs were placed in cassettes and separately preserved in 10% formalin before pathological examination.The remaining tissue was sectioned into 3-5 mm pieces and paraffin embedded.Paraffin block was cut in 4µm histological sections and stained with hematoxylin and eosin for further evaluation of microscopic LNs.Microscopic analysis were accessed by the same pathologist for histological confirmation of the manually dissected LNs.To be considered a LN, microscopically the structure should be described as an encapsulated lymphoid tissue aggregate with a marginal sinus 2,11 .

Statistical analysis
The values of the qualitative variables were described using absolute and relative frequencies.The distribution of the discrete quantitative variables was defined as non-parametric according to Kolmogorov-Smirnov test.These variables were described by mean and standard deviation and also by range, median, interquartile range and 95% confidence interval.Mann-Whitney test was performed for the comparison of quantitative data between groups.Correlation between quantitative variables was performed using Spearman´s correlation test.All statistical analysis were performed using SPSS ® version 17.0 software (SPSS ® Inc; Illinois, USA) and a significance level of 5% was adopted (P≤0.05).

Results
The sample consisted of 28 cadavers (56 specimens of modified radical neck dissection) and 2381 LNs were analyzed.
Eighteen (64.3%) were male and the mean age was 67. and level V had the highest mean of LNs (11.9 LNs) compared to the other nodal levels (Table 1).11LNs), compared to the other levels (Table 3).The comparative analysis of both sides of the neck (Table 4) showed no statistically significant difference, except for the number of LNs in level IV that was significantly higher in the left side (7.0 ± 4.9 LNs in right side vs. 10.2 ± 5.9 LNs in left side, P=0.026 -Mann-Whitney´s test).

Discussion
This study identified the expected number of retrieved lymph nodes from each level of a modified radical neck dissection.
There was also association of the nodal yield, in some specific situations, with sex, height, age, weight and body mass index of the individual.There are many studies accessing this issue, but showing a wide range in the mean number of lymph nodes identified in the specimens of MRND.Studies performed on cadaver model, which are not frequent in the literature, have shown lower nodal yields comparing to those found in this study.The confidence interval of the mean number of the LNs was calculated in order to improve external validity of the findings.To our knowledge, no similar approach has been published yet.
Only two studies published on cadaver model were found in the literature.Friedman et al. 6 , in an effort to provide reference for surgeons and pathologists regarding expected yield from selective ND, found the range of 6 to 44 in the number of LNs in specimens of RND in 20 cadavers.Norling et al. 18 , as a result of the dissection of 9 cadavers, identified a range of 8 to 24 LNs, but only six specimens contained the five nodal levels.
Both studies were performed with a small sample and without the subdivision of levels I and II (IA and IB, IIA and IIB, respectively) recommended by the ND classification of the Committee for Head and Neck Surgery and Oncology the AAO-HNS 11 .
Compared to these previous studies in cadavers, we found a higher mean number of LNs retrieved from each nodal level, as well as in the whole specimen of the ND, with level V showing most significant difference.The nodal yield identified was also higher than those found in most studies with casuistics of patients undergoing ND as part of the treatment of HNSCC, except for the data from Kohler and Kowalski 15 and Liao et al. 16 , who found a higher mean number.Although considering the possibility of hypertrophy and remodeling of cervical LNs, induced by the pathological state related to the presence of the primary tumor 6,18 , used by the authors of the cadaver studies to justify the lower number of LNs found, the present study reported the identification of a number of LNs comparable to those shown in clinical studies.
The results and their relationship to demographic and anthropometric data showed some statistically significant associations, but some correlation coefficients were low.The number of LNs in level IV, in the left side NDs, was higher in males (mean of 11.9 vs. 7.2 LNs for females) and had a positive correlation with height.This finding may be related to the position and anatomical variations of the omohyoid muscle 19,20 , which is considered the upper limit of level IV, while the clavicle is its lower limit.Accordingly, as the measure of the distance between this muscle and the clavicle is quite variable, it may influence on the amount of LNs recovered at this level 19 .However, studies directed to this anatomical region 19,21 found no significant differences in measures of these distances related to sex, age and height.
The number of LNs on level V had positive correlation with weight and BMI of individuals, confirming findings of a previous study of lymph node content in patients undergoing ND diagnosed with melanoma 22 .
The total number of LN was negatively correlated with age, which was also described by Amat et al. 17 and Norling et al. 18 , who attributed this relationship to the natural decrease in immune response and lymph node reactivity with aging.
Given the importance of ND in clinical and surgical practice of the treatment of HNSCC, and the wide variation of data available in the literature regarding the nodal yield, the main contribution of the present report was to calculate, with statistical accuracy, the mean number of LNs in the cervical nodal levels, as well as to estimate the confidence interval of this mean.Hence, a meaningful recommendation for a minimum number of LNs to be identified in the pathology specimen of a neck dissection could be established, suggesting a simple and practical quality control tool.

Conclusion
The results defined a parameter in relation to the minimum recommended nodal yield in a modified radical neck dissection, as well as the number of lymph nodes in each level of this dissection, performed in clinical practice.

FIGURE 1 -
FIGURE 1 -Radical modified neck dissection performed in the left side neck of a non-preserved cadaver; (A) Leve IA (within the two anterior bellies of the digastric muscles) after the specimen removed; (B) Level IB (within the two bellies of the digastric muscle and the mandible) with the submandibular gland and lymph nodes retracted; (C) Levels II.III and IV (jugular chain of lymph nodes) with the specimen removed and the sternocleidomastoid muscle folded; (D) Level V (posterior triangle) showing the spinal accessory nerve.

FIGURE 2 -
FIGURE 2 -Final aspect of the dissection (A) and the nodal levels of both neck sides removed (B).
4 ± 10.3 years.Fifteen subjects were classified as Caucasian (53.6%), 12 Afro-descendants (42.9%) and one Asian (3.6%), with mean weight of 59.6 ± 13.3 kg, mean height of 164 ± 9.1 cm and mean body mass index (BMI) of 22.0 ± 3.9 kg/m 2 .The analysis of the nodal yield in the right side of the neck showed a mean number of 44.4 LNs (range: 26 -99 LNs)

TABLE 1 -
Descriptive of the right side of the neck.
CI. Confidence interval; SD.Standard deviationOn the left side, the mean number of retrieved LNs was 43.3 (range: 19 -64 LNs), with level IV and V showing the highest mean LNs (10.2 and 10 LNs respectively) in comparison to other levels (Table2).

TABLE 2 -
Descriptive of the left side of the neck.
(range 19 -99 LNs), with the highest mean of LNs in level V (

TABLE 3 -
Descriptive of both sides of the neck.
CI. Confidence interval; SD.Standard deviation

TABLE 4 -
Comparison of the number of lymph nodes recovered in each level in both sides of the neck.
decreases with the increasing of age.Other associations did not present any significant difference (data not shown).