Abstract
Background/Aim: Lingual lymph node (LLN) metastasis from tongue cancer occurs at four subsites. However, subsite-related prognosis is unknown. This study aimed to analyze the association between LLN metastases and disease-specific survival (DSS) with respect to these four anatomic subsites. Patients and Methods: Patients with tongue cancer treated between January 2010 and April 2018 at our institute were reviewed. The four subgroups of LLNs were median, anterior lateral, posterior lateral, and parahyoid. DSS was evaluated. Results: LLN metastases occurred in 16 of the 128 cases; six and 10 cases were identified during initial and salvage therapy, respectively. Zero, four, three, and nine cases were median, anterior lateral, posterior lateral, and parahyoid LLN metastases, respectively. The 5-year DSS of patients with LLN metastasis was significantly poor on univariate analysis; parahyoid LLN metastasis showed the worst prognosis. Multivariate analysis indicated that only advanced nodal stage and lymphovascular invasion were significant survival factors. Conclusion: Parahyoid LLNs may require the most caution in tongue cancer. The significance of LLN metastases alone for survival was not confirmed on multivariate analysis.
Lingual lymph nodes (LLNs) are intervening lymph nodes that appear inconsistently in the lymphatic vessels of the tongue. In 1938, Rouvière (1) classified LLNs into median LLNs located in the lingual septum (between the genioglossus muscles) and lateral LLNs located along the marginal lymphatic trunks. Clinical LLN metastases in tongue cancer occur at four sites. LLN metastasis to the lingual septum and sublingual space was reported in 1985 (2), and LLN metastasis along the lingual artery, around the great horn of the hyoid bone, was reported in 2009 (3). Additionally, LLN metastasis located between the hyoglossus muscle and the superficial layer of the deep cervical fascia was reported in 2021 (4). However, the anatomical understanding of LLNs is still unclear, which may be partially due to the ambiguous definition of the lateral LLNs. The debate implicates the need for further discussion regarding the anatomical region and the threshold for routine dissection of these rather rare potential metastases. Recently, LLN metastasis has been reported as an independent negative prognostic factor (5-8); this finding supports the use of the therapeutic strategy of radical resection, known as compartmental resection, proposed by Calabrese et al. (9, 10).
In this report, the frequency and prognosis of LLN metastasis at the four subsites are discussed, reflecting on the clinicopathological features of LLN metastases and the anatomical debate regarding LLNs and the cervical fascia. Moreover, this study aimed to analyze the association between LLN metastases and disease-specific survival (DSS) with respect to these four anatomic subsites.
Patients and Methods
The study cohort included patients with squamous cell carcinoma of the tongue who were initially treated at our institution between January 2010 and April 2018. Clinical data were obtained from medical records and were retrospectively analyzed. In the report, the clinical classification of LLNs consisting of four groups was applied: 1) median LLNs located in the lingual septum; 2) anterior lateral LLNs located in the sublingual space; 3) posterior lateral LLNs located between the hyoglossus muscle and the superficial layer of the deep cervical fascia; and 4) parahyoid LLNs located along the lingual artery around the great horn of the hyoid bone (Figure 1) (4, 11, 12). When an LLN was identified presurgically, it was classified based on radiologic findings; on incidental identification during or after resection, it was classified based on pathological findings.
Schematic representation of lingual lymph node (LLN) subsites. 1) Median LLN located in the lingual septum; 2) anterior lateral LLN located in the sublingual space; 3) posterior lateral LLN located between the hyoglossus muscle and the superficial layer of the deep cervical fascia; and 4) parahyoid LLN located along the lingual artery, around the great horn of the hyoid bone. Black star: Hyoglossus muscle; MHM: mylohyoid muscle; SLG: sublingual gland; SMG: submandibular gland, covered by the two layers of the deep cervical fascia; LA: lingual artery; XII: hypoglossal nerve; HB: hyoid bone.
Differences in the distribution of the baseline characteristics between patients with and without LLN metastases were evaluated using Wilcoxon rank-sum test and Fisher’s exact test. The Kaplan-Meier method was used to estimate median survival, and the crude differences between the two groups were analyzed using a log-rank test. We conducted multivariable Cox proportional hazards models controlling for the following clinical characteristics, which are possible confounders of LLN metastases and survival: age, depth of invasion (DOI) of the primary lesion, nodal metastases, pathological differentiation, lymphovascular invasion, and LLN metastases.
We used propensity score analyses to balance (13) measurable confounders between patients with LLN metastases and those with other metastases. Multivariate logistic regression was used to predict LLN metastases based on the following confounding covariates: DOI, nodal metastases, pathological differentiation, and lymphovascular invasion. Each patient was then assigned an estimated propensity score, which was the individual predicted probability of developing LLN metastases (5-8). Cox regression was also performed by applying propensity scores to adjust for group differences in three alternative ways: 1) regression adjustment; 2) propensity score matching, which paired patients with LLN metastases and others in a 1:1 ratio; and 3) use of the propensity score to create stabilized weights, defined as the inverse probability of treatment weighting (IPTW) (14, 15). All statistical tests were two-sided, and 95% confidence intervals were calculated. All statistical analyses were conducted using EZR (Saitama Medical Center, Jichi Medical University, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) (16).
This study was approved by the ethics board of our institution (405-02022) and was performed in accordance with the Declaration of Helsinki and Ethical Guidelines for Medical Research Concerning Humans of the Ministry of Education, Culture, Sports, Science, and Technology, Japan. Written informed consent was obtained for the publication of patient data. The eighth edition of the Union for International Cancer Control’s TNM classification was used in this study.
Results
The characteristics of the patients enrolled in the study are shown in Table I. Overall, 16 (12.5%) cases of LLN metastases in 128 patients with tongue cancer were initially treated at our institution between January 2010 and April 2018 (Table I); all patients had stage M0. Patients were followed-up for at least 2 years or until their death. There were 89 men and 39 women, with ages ranging from 21 to 91 years (median, 65 years). Most of the basic characteristics between patients with and without LLN metastases showed significant differences in univariate analyses.
Baseline characteristics of the patients.
Figure 2 shows the unadjusted Kaplan-Meier survival curves of the two groups (patients with and without LLN metastases). The median DSS was not reached in patients without LLN metastases, compared with 42 months for patients with LLN metastases. The unadjusted 5-year DSS probabilities were 49.9% (95%CI=20.1%-74.0%) and 85.5% (95%CI=77.0%-91.0%) for patients with and without LLN metastases, respectively. LLN metastases were related with significantly worse survival (p=0.007).
Kaplan-Meier survival curves of the groups with no lingual lymph node (LLN) metastasis and the group with LLN metastasis.
Among the 16 patients with LLN metastases, six and ten cases were identified at the time of initial and salvage therapy, respectively. Regarding patients with LLN metastases that were identified at the time of initial therapy, all of the LLNs were excised at the time of initial therapy and were resected in a continuous fashion with the primary lesion. On the other hand, LLN metastases that were identified at the time of salvage therapy in all of the 10 patients were identified as recurrences after the primary resection. Additionally, the mean latency interval from initial therapy to identification of LLN metastasis was 9.6 months (range=1-33 months). Based on the classification of lingual nodes, zero, four, three, and nine patients had median, anterior lateral, posterior lateral, and parahyoid LLN metastases, respectively (Table II). Patients with parahyoid LLN metastases and those with LLN metastases treated with salvage therapy tended to show the worst survival (Figure 3 and Figure 4).
Anatomical subsites and treatment timing of lingual lymph node (LLN) metastases.
Kaplan-Meier survival curves of the groups with no lingual lymph node (LLN) metastasis, anterior lateral LLN metastasis, posterior lateral LLN metastasis, and parahyoid LLN metastasis.
Kaplan-Meier survival curves of patients with lingual lymph node (LLN) metastasis comparing the groups in which LLN metastasis was treated during initial therapy and salvage therapy.
Controlling for clinical characteristics in adjusted Cox proportional hazards models, there was no significant difference in the 5-year DSS between patients with and without LLN metastases (HR=1.37; 95%CI=0.48-3.92); advanced nodal stage (HR=3.41; 95%CI=1.09-10.7) and lymphovascular invasion (HR=5.26; 95%CI=1.16-23.7) were significant covariates (Table III). The results of the propensity score analyses were inconsistent. Of the three propensity score-adjusted models, the regression-adjusted model and matching model did not demonstrate evidence of a negative impact of LLN metastases on DSS, whereas IPTW depicted LLN metastases as a significant negative confounder for survival (Table IV).
Multivariate analyses of factors related to 5-year disease-specific survival in the total cohort.
Effect of lingual lymph node (LLN) metastases on hazard ratios for disease-specific survival.
Discussion
The main findings were as follows: 1) The incidence of LLN metastases was 12.5%, and the parahyoid LLN was the most common metastatic site of squamous cell carcinoma of the tongue; 2) parahyoid LLN metastases showed the worst survival on univariate analysis; and 3) the 5-year DSS in patients with LLN metastases was significantly poor on univariate analysis, but this was not confirmed by multivariate analysis.
When considering the clinical significance of LLN metastases, one cannot ignore the fact that the definition and frequency of LLNs in the healthy population are still under debate. Anatomical studies have reported that the incidence of LLNs ranges from 8.6% to 30.2%. In contrast, previously reported incidences of median and lateral LLN metastasis were 0.7-3.0% and 1.4-14.3%, respectively (12). Inconsistency in the literature is evident, and this may be because the definition of the lateral LLNs has been somewhat ambiguous. A recent anatomical study emphasized the need to define the anatomical border of the LLNs according to the deep cervical fascia (11); a better understanding of the difference between the lateral LLNs and level IB lymph nodes is advocated. Our study is unique in that it evaluated the frequency and prognostic significance of LLN metastasis in concordance with the new classification based on recently improved anatomical understanding (4).
Patients with parahyoid LLN metastases and those with LLN metastases treated with salvage therapy tended to show the worst survival on univariate analysis; the correlation between the two factors was not evaluable due to the small sample size. Our findings are in accordance with the previous report from Ando et al. that despite all efforts at salvage treatment, all but one of the 14 patients with parahyoid LLN metastasis died due to uncontrolled neck disease (3). From an anatomical standpoint, parahyoid LLNs are different in that they do not necessarily bond to the lingual architecture and thus are at a higher risk of being left behind after resection. To ablate these nodes, wide resection may be considered. It has been described that wide resection of the entire ipsilateral lingual architecture continuously with the tissue of the neck positively contributes to survival (9, 10). Further study is needed to determine its indication; this is because the majority of patients do not have LLNs in the first place (11, 12, 17) and routine dissection of LLNs has not been reported to improve survival (3, 6). There may be a need to further discuss the order of precedence to inspect and dissect LLN subsites, as indicated in the present study.
The 5-year DSS in patients with LLN metastases was significantly poor in the univariate analysis, but this was not confirmed by multivariate analysis. In recent years, an increasing number of studies regarding LLNs have been reported. Retrospective studies with large cohorts have claimed that LLN metastases are strongly correlated with other regional cervical lymph node metastases and that LLN metastases have a significant negative impact on survival (5, 6). Although we agree with the recent consensus, the present study indicates that their definite survival impact may not be significant on its own. We believe that the interpretation of LLN metastases must be clarified in future studies to determine whether LLN metastases need to be thoroughly dissected as possible sentinel nodes or if it is merely a rare phenomenon representing the malignant nature of the disease.
A limitation of our study is that our analysis was validated on a rather small cohort compared to recent studies. The incidence of LLN metastasis is low and an accumulation of further evidence is warranted. Additionally, further studies regarding predictive factors of late cervical lymph node metastasis, including LLN, may be desirable (18).
In the present study, parahyoid LLN metastasis was the most frequent subsite and tended to exhibit the worst survival rate. The survival impact of LLN metastases alone was not confirmed in the present study. The incidence of LLN metastases is still debated, and the threshold for ablating LLN warrants further discussion.
Acknowledgements
The Authors would like to thank Dr. Misa Iijma for her conception of and advice on the article.
Footnotes
Authors’ Contributions
Both Authors contributed equally to the conception and writing of the work.
Conflicts of Interest
The Authors have no conflicts of interest to disclose in relation to this study.
- Received January 4, 2023.
- Revision received February 20, 2023.
- Accepted February 21, 2023.
- Copyright © 2023 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.