Abstract
Introduction: Surgical management of level V in clinically node positive (cN+) oral squamous cell carcinomas (OSCCs) is controversial. The objectives of the study were to identify predictors of level V metastases in cN+ OSCC. Methods: This retrospective study is based on institutional data of operated cN+ OSCC between April 2018 and December 2022. Clinical and pathological parameters were subjected to univariate analysis. Significant parameters in univariate analysis were further subjected to multivariate analysis. A p value of less than 0.05 was considered statistically significant. Results: None of cN1 or pN1 patients had a level V metastasis. No skip metastasis to level-V was noticed. Total number of positive lymph nodes, lymph node ratio (LNR), extranodal extension, pN classification, and the presence of level II and III metastases were found to be significant predictors for level V metastases. The post hoc analysis suggested that ≥5 positive nodes, LNR >0.1, and pN3 status were independent risk factors for level V metastases. Conclusion: Selective neck dissection for N+ OSCC is feasible in the N1 neck, preferably where nodal metastases is limited to level-I only. Patients with a bulky nodal disease, particularly those with N3 neck, ≥5 positive nodes, ENE, and metastatic lymph nodes in levels II and III should be offered comprehensive neck dissection.
Introduction
Surgical management of level V nodal basin in oral squamous cell carcinoma (OSCC) is controversial. Level V dissection has been associated with an increased post-operative shoulder morbidity due to excessive dissection around spinal accessory nerve. Although good quality evidence exists regarding preservation of shoulder function by avoiding level IIB dissection [1, 2], addressing level V where the nerve courses through immediately after exiting level IIB is not well studied. The majority of regional metastases are confined to level I to III; especially level IB and IIA [3‒5]. Hence, there is an increasing trend of neck dissection comprising level I to IV in clinically node positive (cN+) OSCC [6‒14]. However, a large prospective trial from India has shown level IIA and III positivity to be predictors of level V metastases [4]. Neck recurrences after initial treatment are usually advanced and seldom surgically salvageable [15, 16]. Moreover, lower neck nodal involvement affects the overall survival significantly [17]. So, careful selection of patients is an absolute necessity before omitting level V dissection in cN+ OSCC. The present study attempted to identify predictors of level V metastases beyond level II and III nodal positivity with an aim to develop individualized recommendations regarding the extent of neck dissection in cN+ oral cancers.
Materials and Methods
Study Design and Participants
A retrospective analysis from a prospectively maintained institutional database was done. The study was in accordance with the guidelines set by the Declaration of Helsinki and International Council for Harmonisation – Good Clinical Practice. The study protocol was reviewed and approved by the Institutional Review Board (IRB) and Institutional Ethics Committee (IEC) (EC/NEW/Inst/2022/UA/0180), Approval No.: AIIMS/IEC/23/402. Since this was a retrospective study conducted ensuring absolute confidentiality of patient details, informed consent was waived by the IEC. Patients with cN+ OSCC, as determined by clinical examination, and contrast-enhanced magnetic resonance imaging for tongue and floor of mouth (FOM) primaries and contrast-enhanced computed tomography for non-tongue/FOM primaries, who underwent surgery with comprehensive neck dissection (CND) (level I–V) from 1st April 2018 to 31st December 2022 were included. Patients who were operated on for residual or recurrent disease had a pathological N0 status, and no level-wise nodal description in final histopathology were excluded from the final analysis. Tumors were staged according to the AJCC 8th TNM staging system. A consistent protocol-based treatment strategy was followed during the study period. Bilateral neck dissections were performed for tumors crossing midline or clinical evidence of bilateral neck nodes.
Data Definitions and Categorization
Datasets included patient characteristics (age, sex, ECOG performance status, addiction to tobacco, and alcohol), clinical details (subsite, clinical T and N classification, neck dissection), and histopathological parameters (tumor volume, histological grade, depth of invasion, lymphovascular invasion, perineural invasion, worst pattern of invasion, bone invasion, lymph nodal yield (LNY), total number of positive lymph nodes, lymph node ratio [LNR], extra-nodal extension, level-wise lymph nodal distribution). LNR was defined as the number of positive nodes divided by LNY. The total number of positive lymph nodes and LNR was further categorized into three subgroups according to a previously published study (Table 1) [18].
Basic characteristics of the patients
Characteristics . | Number (%) . | Characteristics . | Number (%) . |
---|---|---|---|
Age | Mean 46.42 (±11.54) | cT classification | |
Median 45 (26–71) | 1 | 2 (1.4) | |
2 | 13 (9.2) | ||
0–30 years | 13 (9.1) | 3 | 31 (21.8) |
31–60 years | 111 (78.2) | 4 | 96 (67.6) |
61 and above years | 18 (12.7) | ||
Sex | cN classification | ||
Male | 128 (90.1) | 1 | 25 (17.6) |
Female | 14 (9.9) | 2 | 81 (57.0) |
3 | 36 (25.4) | ||
Smokeless tobacco | C/L neck dissection | ||
Yes | 106 (74.7) | CND | 13 (9.1) |
No | 31 (21.8) | SND (level I–III) | 22 (15.5) |
Not known | 5 (3.5) | SND (level I–IV) | 16 (11.3) |
Smoking tobacco | pT classification | ||
Yes | 50 (35.2) | 1 | 5 (3.5) |
No | 87 (61.3) | 2 | 21 (14.8) |
Not known | 5 (3.5) | 3 | 36 (25.4) |
4 | 80 (56.3) | ||
Alcohol | pN classification | ||
Yes | 57 (40.1) | 1 | 36 (25.4) |
No | 80 (56.4) | 2 | 70 (49.2) |
Not known | 5 (3.5) | 3 | 36 (25.4) |
ECOG PS | Total positive LNs | ||
0 | 20 (14.1) | 1–2 | 77 (54.2) |
1 | 116 (81.7) | 3–4 | 33 (23.2) |
2 | 6 (4.2) | ≥5 | 32 (22.5) |
Subsites | LNR | ||
Buccal mucosa and RMT | 72 (50.7) | <0.1 | 93 (65.5) |
Tongue and FOM | 43 (30.3) | 0.1–0.4 | 44 (31.0) |
Upper alveolus and palate | 5 (3.5) | >0.4 | 5 (3.5) |
Lower alveolus | 18 (12.7) | ||
Lip | 4 (2.8) |
Characteristics . | Number (%) . | Characteristics . | Number (%) . |
---|---|---|---|
Age | Mean 46.42 (±11.54) | cT classification | |
Median 45 (26–71) | 1 | 2 (1.4) | |
2 | 13 (9.2) | ||
0–30 years | 13 (9.1) | 3 | 31 (21.8) |
31–60 years | 111 (78.2) | 4 | 96 (67.6) |
61 and above years | 18 (12.7) | ||
Sex | cN classification | ||
Male | 128 (90.1) | 1 | 25 (17.6) |
Female | 14 (9.9) | 2 | 81 (57.0) |
3 | 36 (25.4) | ||
Smokeless tobacco | C/L neck dissection | ||
Yes | 106 (74.7) | CND | 13 (9.1) |
No | 31 (21.8) | SND (level I–III) | 22 (15.5) |
Not known | 5 (3.5) | SND (level I–IV) | 16 (11.3) |
Smoking tobacco | pT classification | ||
Yes | 50 (35.2) | 1 | 5 (3.5) |
No | 87 (61.3) | 2 | 21 (14.8) |
Not known | 5 (3.5) | 3 | 36 (25.4) |
4 | 80 (56.3) | ||
Alcohol | pN classification | ||
Yes | 57 (40.1) | 1 | 36 (25.4) |
No | 80 (56.4) | 2 | 70 (49.2) |
Not known | 5 (3.5) | 3 | 36 (25.4) |
ECOG PS | Total positive LNs | ||
0 | 20 (14.1) | 1–2 | 77 (54.2) |
1 | 116 (81.7) | 3–4 | 33 (23.2) |
2 | 6 (4.2) | ≥5 | 32 (22.5) |
Subsites | LNR | ||
Buccal mucosa and RMT | 72 (50.7) | <0.1 | 93 (65.5) |
Tongue and FOM | 43 (30.3) | 0.1–0.4 | 44 (31.0) |
Upper alveolus and palate | 5 (3.5) | >0.4 | 5 (3.5) |
Lower alveolus | 18 (12.7) | ||
Lip | 4 (2.8) |
ECOG PS, Eastern Cooperative Oncology Group performance status; RMT, retromolar trigone; FOM, floor of mouth; C/L, contralateral; CND, comprehensive neck dissection; SND, selective neck dissection; LN, lymph node; LNR, lymph node ratio.
Statistical Analyses
Results were analyzed using IBM SPSS for Windows, version 26.0 (Armonk, NY, USA). Descriptive statistics were expressed in mean (±SD), median (IQR), and proportions, as applicable. Clinical and pathological parameters were subjected to univariate ANOVA. Significant parameters in univariate analysis were further subjected to multivariate regression analysis. Post hoc ANOVA with Bonferroni correction was used for pairwise comparisons between subcategories of the total number of positive lymph nodes, LNR, and pN classification. A p value of less than 0.05 was considered statistically significant.
Results
Four hundred thirty-eight (n = 438) patients with cN+ OSCC underwent surgery during the given period. Among these, 151 patients with cN1 neck underwent selective neck dissection (SND) (level I–III and level I–IV), and 37 patients were operated on for residual or recurrent disease. Final histopathology confirmed pN0 status for 103 patients, and level-wise nodal description was absent for 5 patients. One hundred and forty-two (n = 142) patients were included in the final analysis. Thirteen (n = 13) patients underwent bilateral CND resulting in 155 analyzable neck dissection specimens (Fig. 1).
The mean age of the study population was 46.4 (±11.5) years. Majority of the patients were males, had addiction to tobacco and alcohol, primaries involving buccal mucosa and retromolar trigone, and presented at a locally advanced stage (cT3/4, cN2/3). The median LNY was 42 (29.5) [19]. The basic characteristics are highlighted in Table 1.
In our study population, 15 out of 142 (10.6%) pN+ and 15 out of 245 (6.1%) cN+ OSCC had evidence level V metastases. A total of 40% (n = 6) of them had oral tongue and FOM primaries, followed by lower alveolus (n = 4, 26.7%). None of cN1 or pN1 patients had a level V metastasis. No skip or isolated metastasis to level V was noticed. A total of 86.7% (n = 13) and 66.7% (n = 10) of level V metastases had associated level II and III metastases, respectively (Table 2).
Characteristics of patients with level V metastases
Characteristics . | Number (%) . | Characteristics . | Number (%) . |
---|---|---|---|
Subsites | Total positive LNs | ||
Buccal mucosa and RMT | 3 (20.0) | 3–4 | 4 (26.7) |
Tongue and FOM | 6 (40.0) | ≥5 | 11 (73.3) |
Upper alveolus and palate | 1 (6.7) | ||
Lower alveolus | 4 (26.7) | ||
Lip | 1 (6.7) | ||
cT classification | LNR | ||
2 | 3 (20.0) | <0.1 | 3 (20.0) |
3 | 3 (20.0) | 0.1–0.4 | 9 (60.0) |
4 | 9 (60.0) | >0.4 | 3 (20.0) |
cN classification | pT classification | ||
2 | 10 (66.7) | 1 | 1 (6.7) |
3 | 5 (33.3) | 2 | 2 (13.3) |
3 | 1 (6.7) | ||
4 | 11 (73.3) | ||
Histological grade | pN classification | ||
Well differentiated | 5 (33.3) | 2 | 5 (33.3) |
Moderately differentiated | 8 (53.4) | 3 | 10 (66.7) |
Poorly differentiated | 2 (13.3) | ||
DOI | Level II metastases | ||
≤5 mm | 1 (6.7) | Present | 13 (86.7) |
>5 mm, ≤10 mm | 5 (33.3) | Absent | 2 (13.3) |
>10 mm | 9 (60.0) | ||
LVI | Level III metastases | ||
Present | 6 (40.0) | Present | 10 (66.7) |
Absent | 9 (60.0) | Absent | 5 (33.3) |
PNI | |||
Present | 5 (33.3) | ||
Absent | 10 (66.7) |
Characteristics . | Number (%) . | Characteristics . | Number (%) . |
---|---|---|---|
Subsites | Total positive LNs | ||
Buccal mucosa and RMT | 3 (20.0) | 3–4 | 4 (26.7) |
Tongue and FOM | 6 (40.0) | ≥5 | 11 (73.3) |
Upper alveolus and palate | 1 (6.7) | ||
Lower alveolus | 4 (26.7) | ||
Lip | 1 (6.7) | ||
cT classification | LNR | ||
2 | 3 (20.0) | <0.1 | 3 (20.0) |
3 | 3 (20.0) | 0.1–0.4 | 9 (60.0) |
4 | 9 (60.0) | >0.4 | 3 (20.0) |
cN classification | pT classification | ||
2 | 10 (66.7) | 1 | 1 (6.7) |
3 | 5 (33.3) | 2 | 2 (13.3) |
3 | 1 (6.7) | ||
4 | 11 (73.3) | ||
Histological grade | pN classification | ||
Well differentiated | 5 (33.3) | 2 | 5 (33.3) |
Moderately differentiated | 8 (53.4) | 3 | 10 (66.7) |
Poorly differentiated | 2 (13.3) | ||
DOI | Level II metastases | ||
≤5 mm | 1 (6.7) | Present | 13 (86.7) |
>5 mm, ≤10 mm | 5 (33.3) | Absent | 2 (13.3) |
>10 mm | 9 (60.0) | ||
LVI | Level III metastases | ||
Present | 6 (40.0) | Present | 10 (66.7) |
Absent | 9 (60.0) | Absent | 5 (33.3) |
PNI | |||
Present | 5 (33.3) | ||
Absent | 10 (66.7) |
DOI, depth of invasion; LVI, lymphovascular invasion; PNI, perineural invasion; FOM, floor of mouth; RMT, retromolar trigone.
Tumor volume, histological grade, total number of positive lymph nodes, LNR, LNY, the presence of extra-nodal extension (ENE), pN classification, and the presence of level II and III metastases were significant factors associated with level V metastases in univariate analysis. In multivariate analysis, total number of positive lymph nodes, LNR, ENE, pN classification, and the presence of level II and III metastases were found to be significant predictors for level V metastases (Table 3). The post hoc analysis suggested that ≥5 positive nodes, LNR >0.1, and pN3 status were independent risk factors for level V metastases (Table 4).
Analysis of clinical and pathological parameters as predictors of level V metastases
Characteristics . | punivariate . | pmultivariate . | Characteristics . | punivariate . | pmultivariate . |
---|---|---|---|---|---|
Smokeless tobacco | 0.69 | - | Bone invasion | 0.09 | - |
Smoking tobacco | 0.93 | - | LVI | 0.08 | - |
Alcohol | 0.47 | - | PNI | 0.15 | - |
Subsite | 0.21 | - | Total positive LNs | <0.001* | <0.001* |
cT classification | 0.54 | - | LNR | <0.001* | <0.001* |
cN classification | 0.64 | - | ENE | <0.001* | <0.001* |
Tumor volume (mm3) | 0.001* | 0.08 | pT classification | 0.25 | - |
Histological grade | 0.007* | 0.19 | pN classification | <0.001* | <0.001* |
DOI | 0.13 | - | Level II metastases | 0.01* | 0.04* |
WPOI | 1 | - | Level III metastases | <0.001* | <0.001* |
Characteristics . | punivariate . | pmultivariate . | Characteristics . | punivariate . | pmultivariate . |
---|---|---|---|---|---|
Smokeless tobacco | 0.69 | - | Bone invasion | 0.09 | - |
Smoking tobacco | 0.93 | - | LVI | 0.08 | - |
Alcohol | 0.47 | - | PNI | 0.15 | - |
Subsite | 0.21 | - | Total positive LNs | <0.001* | <0.001* |
cT classification | 0.54 | - | LNR | <0.001* | <0.001* |
cN classification | 0.64 | - | ENE | <0.001* | <0.001* |
Tumor volume (mm3) | 0.001* | 0.08 | pT classification | 0.25 | - |
Histological grade | 0.007* | 0.19 | pN classification | <0.001* | <0.001* |
DOI | 0.13 | - | Level II metastases | 0.01* | 0.04* |
WPOI | 1 | - | Level III metastases | <0.001* | <0.001* |
WPOI, worst pattern of invasion; ENE, extra-nodal extension.
*Asterisks for significant values.
Post hoc analysis for significant predictors of level V metastases
Characteristics . | Comparison . | p value (Bonferroni) . | 95% confidence interval . | |
---|---|---|---|---|
lower bound . | upper bound . | |||
Total positive LNs | 1–2 vs. 3–4 | 0.07 | −0.25 | 0.01 |
1–2 vs. ≥5 | <0.001* | −0.51 | −0.24 | |
3–4 vs. ≥5 | <0.001* | −0.41 | −0.10 | |
LNR | <0.1 vs. 0.1–0.4 | <0.001* | −0.31 | −0.08 |
<0.1 vs. >0.4 | <0.001* | −0.85 | −0.28 | |
0.1–0.4 vs. >0.4 | 0.01* | −0.67 | −0.08 | |
pN classification | pN1 vs. pN2 | 0.53 | −0.20 | 0.06 |
pN1 vs. pN3 | <0.001* | −0.45 | −0.16 | |
pN2 vs. pN3 | <0.001* | −0.36 | −0.11 |
Characteristics . | Comparison . | p value (Bonferroni) . | 95% confidence interval . | |
---|---|---|---|---|
lower bound . | upper bound . | |||
Total positive LNs | 1–2 vs. 3–4 | 0.07 | −0.25 | 0.01 |
1–2 vs. ≥5 | <0.001* | −0.51 | −0.24 | |
3–4 vs. ≥5 | <0.001* | −0.41 | −0.10 | |
LNR | <0.1 vs. 0.1–0.4 | <0.001* | −0.31 | −0.08 |
<0.1 vs. >0.4 | <0.001* | −0.85 | −0.28 | |
0.1–0.4 vs. >0.4 | 0.01* | −0.67 | −0.08 | |
pN classification | pN1 vs. pN2 | 0.53 | −0.20 | 0.06 |
pN1 vs. pN3 | <0.001* | −0.45 | −0.16 | |
pN2 vs. pN3 | <0.001* | −0.36 | −0.11 |
*Asterisks for significant values.
Discussion
Involvement of neck nodes is considered the single most important prognostic factor in patients with head and neck squamous cell carcinomas [20]. An appropriate elective neck dissection is crucial for improved regional control and survival in these patients [16]. Though CND has been the standard of care for cN+ OSCC, it is also associated with increased shoulder morbidity due to excessive tissue handling around spinal accessory nerve, particularly during level V dissection. A study on the distribution of regional metastases in OSCC by Shah et al. [3] laid the foundation for the researchers to explore less radical approaches for the neck [3]. Subsequently, a series of studies were published favoring feasibility of SND in cN+ OSCC [6‒14]. SND offered an equivalent regional control and survival rates when compared to CND in selected cN+ patients.
As the treatment of cancer is becoming more individualized over time, patient-centric predictive algorithms or nomograms are slowly being inducted into clinical practice rather than generalized guideline-based treatment. The incidence of level V metastases in OSCC is less than 5% and is always associated with metastases in preceding nodal stations [4, 12]. Hence, whether to address level V during neck dissection requires a more individualized approach rather than overall cN status. In a large prospective study, Pantvaidya et al. [4] identified metastases at level IIA and III as predictors of level V metastases and recommended CND in this subgroup of node positive oral cancer patients [4]. Moreover, factors predicting overall lymph nodal metastases in OSCC do not predict level V metastases [12, 21]. In another prospective study, Maharaj et al. [12] showed ENE at other levels was the only predictor of level V metastases in multivariate analysis [12]. Level V metastases were present in 6.1% of cN+ and 10.6% of pN+ necks in the present study cohort. The higher incidence can be explained by exclusion of 151 cN1 patients who underwent SND from the final analysis as none of the cN1 (n = 25) or pN1 (n = 36) patients in this study showed level V metastases. If the above subset of patients along with the findings is included in the final analysis, the incidence in our study would have dropped to 3.8% (15 out of 396) as reported in the literature [3, 4, 12]. The study by Parikh et al. [22] also reported less than 1% incidence of level V metastases in cN1 OSCC [22]. Two-thirds of our patients with level V metastases had primaries in tongue, FOM, and lower alveolus, though none of the primary sites attained statistical significance. Level V metastases were always associated with metastases in any of the preceding nodal levels in the present study as well.
The regional failure rate after SND in N+ OSCC, though reported to be statistically equivalent to CND, varied from 11% to 20% [6‒11]. Nodal relapses are notorious for advanced stage of presentation and ENE [15, 16], further emphasizing the importance of careful selection of patients for SND in N+ neck [23]. In our study, ≥5 positive nodes, LNR >0.1, ENE at any nodal level, pN3 status, and level II and III metastases were found to be independent predictors of level V metastases in multivariate analysis. Liang et al. [13] in his systematic review and meta-analysis has not shown any significant difference in regional recurrence, disease-specific and overall mortality between SND and CND [13]. But the results of this study must be interpreted with caution because of the following reasons. First, a significant weightage of the pooled estimate has been fetched from a single study [8]. Second, this highest weighted study did not include N2c and N3 in its cohort. Third, the length of follow-up, inclusion criteria (cN classification), and the extent of SND were inconsistent across the included studies.
The strength of our study lies in the fact that a single study provides sufficient information about the predictors of level V metastases; enough to aid the surgeon tailoring the extent of neck dissection in N+ oral cancer patients. The major limitation is that all the independent predictors of level V metastases are pathological, restricting its feasibility in the clinical setting. Being retrospective in nature, the authors could not verify the precision of clinical data and eliminate inter-observer variability. However, all the significant predictive factors can be reliably extrapolated to clinically bulky nodal disease, multiple palpable nodes, extra-nodal extension, and the presence of suspicious nodes at level II and/or III. Though clinicopathological parameters may vary between primary subsites [24], overall treatment principle remains the same.
The present study can generate level-2b evidence which is, at best, hypothesis generating. Prospective randomized trials comparing SND and CND in cN+ oral cancer stratified according to cN classification, though would have been ideal to generate level-1 evidence, may not be practically feasible due to ethical concerns. The only registered randomized trial in this regard (CTRI/2017/09/009920) is currently underway and does not include N2c and N3 subset of the patients. In the absence of such trials, the results of the present study should serve as a guide to customize extent of neck dissection in cN+ OSCC.
Acknowledgments
The authors thank Dr. Amit Kumar, Department of Otolaryngology – Head and Neck Surgery, All India Institute of Medical Sciences, New Delhi, India, for providing language help and proofreading the article.
Statement of Ethics
This study protocol was reviewed and approved by the Institutional Review Board and Institutional Ethics Committee (DHR Reg. No. EC/NEW/Inst/2022/UA/0180), Approval No. AIIMS/IEC/23/402. Since this was a retrospective institutional chart review ensuring absolute confidentiality of the patient details, informed consent was waived by the IEC.
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This study was not supported by any sponsor or funder.
Author Contributions
K.S.M.: conceptualization, data curation, formal analysis, methodology, project administration, resources, software, supervision, visualization, and roles/writing – original draft, review, and editing. V.S.K.: conceptualization, data curation, formal analysis, visualization, and roles/writing – original draft. A.V.: data curation, resources, visualization, and roles/writing – original draft. T.A., A.P., and A.U.: data curation, methodology, visualization, and roles/writing – original draft. P.K., D.D.M., and A.S.: data curation, visualization, and roles/writing – review and editing. A.M., R.P., N.R., and N.S.D.: data curation and resources. A.B. and M.P.: methodology, resources, supervision, visualization, and roles/writing – review and editing. M.M.: project administration, resources, supervision, and roles/writing – review and editing.
Additional Information
Conference proceedings: The abstract was presented in the 7th Manipal Research Colloquium 2023.
Data Availability Statement
The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from the corresponding author (K.S.M.) ([email protected]; [email protected]) upon reasonable request.