The Depth of Invasion and Level IV Cervical Node Metastasis in Patients with Clinically N0 Tongue Cancer

Abstract

Background

The accurate indication for level IV dissection is crucial for preventing complications such as phrenic nerve damage and chylous fistulas in clinically N0 tongue cancer. Although the depth of invasion is an established independent risk factor for occult lymph node metastasis in tongue cancer, its relationship with level IV metastasis has not been evaluated. This study investigated the relationship between the depth of invasion and level IV nodal metastasis in clinically N0 tongue cancer.

Methods

We retrospectively investigated clinical N0 patients who underwent glossectomy and level I–IV neck dissection. We examined lymph node metastasis, risk factors, and the relationship between depth of invasion and metastasis.

Results

Our study included 58 patients, and no patient had isolated level IV metastasis. Additionally, there was no level IV metastasis in well-differentiated tumors. Tumor size, depth of invasion, differentiation, and perineural invasion were significantly associated with level IV neck metastasis. We found a critical tumor size of 2.5 cm and depth of invasion of 8 mm for level IV neck metastasis.

Conclusion

Based on our findings, we recommend that level IV dissection should be considered for poorly differentiated tumors, tumors greater than 2.5 cm in size, and those deeper than 8 mm. This study highlights the importance of depth of invasion as a prognostic factor for predicting level IV metastasis and suggests that our findings can be used to prevent unnecessary level IV dissections that may lead to complications in tongue cancer surgery.

Introduction

Oral cavity cancers are a significant health problem. More than 370,000 cases occur worldwide each year, resulting in 177,757 deaths [1]. Squamous cell carcinoma of the tongue (TSCC) is the most common malignant tumor in the oral cavity. The incidence of TSCC is increasing annually [2]. Its clinical course is aggressive, and it has a poorer prognosis than do buccal mucosa and gingival cancer [3]. Smoking and alcohol consumption are important risk factors for developing this disease [4].

Surgery is the most important treatment for tongue carcinoma patients without distant metastasis. Adjuvant therapies are considered for advanced disease. According to the American Joint Committee on Cancer (AJCC) TNM classification, preoperative staging must be established for adequate treatment. The crucial criteria for staging are lymph node involvement and tumor size. The depth of invasion (DOI) was also included in the latest staging system published in 2017 [5]. The relationship between DOI and neck metastasis is well known. However, the critical value of the DOI for metastasis has been identified in various studies, but no studies have investigated the relationship between the DOI and level IV metastasis. Predicting the level IV metastasis prevent the unnecessary wider dissections can result in complications such as chylous fistula and phrenic nerve injury in patients with clinical N0 tongue cancer. This study included only clinical N0 tongue cancer patients who underwent level I–IV neck dissection, and the relationship between the DOI and level IV metastasis was investigated.

Materials and Methods

This study was approved by the Institutional Clinical Research Ethics Committee and included 58 clinical N0 tongue cancer patients who underwent glossectomy and level I–IV neck dissection surgery. The treatment decisions and follow-ups of all patients were made by the tumor board of the head and neck team, which consisted of radiologists, medical and radiation oncologists, pathologists and head and neck surgeons. Bilateral neck dissection was performed in seven patients whose tumors crossed the midline. We retrospectively investigated clinically N0 patients who underwent glossectomy and level I–IV neck dissection. Patients who were followed for at least 24 months after surgery were included in the study. Patients who had previously undergone chemotherapy/radiotherapy, who had distant metastases, who had undergone level I–III neck dissection, or who had clinical N + disease were excluded. Demographic information, TNM classification, surgical treatment, final histopathological results, and clinical course data were retrospectively reviewed. Surgical specimens were subdivided into levels immediately after the operation and sent separately for histopathological evaluation. The number and level of metastatic lymph nodes were recorded. The histopathology slides were reviewed by an experienced head and neck pathologist. The depth of invasion was measured from the horizon of the basement membrane of the adjacent squamous mucosa to the deepest point of the tumor in a perpendicular direction by using a Nikon Eclipse Ni-U microscope and Nikon NIS-Elements imaging software. The relationship between the DOI and lymph node metastasis was investigated. The DOI of tumors was evaluated for a cutoff value to predict neck metastasis and level IV neck metastasis.

Statistical Analysis

Histograms and Q‒Q plots were examined, and the Shapiro‒Wilk test was applied to assess the normality of the data. To compare the differences between groups, two-sided independent samples t tests or Mann‒Whitney U tests were used for continuous variables, and Pearson’s χ² or Fisher χ² tests were applied for categorical variables. ROC analyses were applied to determine the ability of tumor size and DOI markers to predict neck and level IV lymph node metastasis. For each marker, cutoff values were determined using the Youden index. For each marker, the sensitivity, specificity, and positive and negative predictive values were calculated with 95% confidence intervals using these cutoff values. Analysis was conducted using Turcosa Cloud (Turcosa Ltd., Turkey) statistical software or R 4.2.0 (www.r-project.org). A p value less than 5% was considered to indicate statistical significance.

Results

Fifty-eight patients admitted between 2005 and 2020 whose files and clinical information could be accessed were included in this study. Thirty-seven patients (63.8%) were female, and 21 (36.2%) were male. The median age was 56.66 ± 15.73 years. The tumors were located on the right side in 28 (48.3%) patients, on the left side in 26 (44.8%), and in the middle of the tongue in 4 (6.8%). Three of the tumors were located laterally but crossed the midline. Fifty-one (87.9%) patients underwent unilateral neck dissection, and 7 (12.1%) patients underwent bilateral neck dissection. Sixty-five neck dissections from 58 patients were included in the study. There were ten patients with T1 tumors, thirty patients with T2 tumors, fifteen patients with T3 tumors, and three patients with T4 tumors in the study. There were multiple metastases in 9 patients. The median number of lymph nodes dissected was 33 (15–49). Most of the patients had involvement levels of I and II. Six patients had isolated level II metastases, and two had isolated level III metastasis (Table 1). No patients had isolated level IV metastasis. Two patients had bilateral neck involvement. The distribution of histopathologically positive lymph nodes in our series is shown in Table 1. Perineural invasion was present in 32.7% of the patients, lymphovascular invasion was present in 18.9% of the patients.

Table 1.

Table presenting the descriptive statistics

Variables	Descriptive statistics
Age (Years)	56.66 ± 15.73
Tumor size (cm)a	2.5(1.5–3.3)
DOI (cm)a	0.7(0.5–1.0)
Gender
Female	37(63.8%)
Male	21(36.2%)
Differentiation
Well differentiated	35 (60,34%)
Moderately differentiated	16(27.58%)
Poorly differentiated	7(12.06%)
Primary tumor
Right side	28(48.3%)
Left side	26(44.8%)
Midline	4(6.8%)
Surgery
Partial glossectomy	51(87.9%)
Hemiglossectomy	6(10.3%)
Total glossectomy	1(1.7%)
Neck dissection
Unilateral neck dissection	51(87.9%)
Bilateral neck dissection	7(12.1%)
Neck metastasis
Absent	46(70.8%)
Present	19(29.2%)
Pathology level IV
Absent	59(90.8%)
Present	6(9.2%)
Distribution of pathologically positive nodes
Distribution of lymph nodes	Neck
II	6
III	2
I + II	4
I + III	1
I + IV	1
III + IV	2
I + III + IV	1
I + II + III + IV	2
	19

^aData are expressed as mean ± standard deviation, median (1st quartile–3rd quarter), and n(%)

There was no relationship between neck metastasis and age or sex. There was no level IV metastasis in the case of a well-differentiated tumor (Table 2). Tumor size, depth of invasion, differentiation, and perineural invasion were significantly associated with level IV metastasis (Fig. 1) (p values: 0.010, 0.002, 0.009, and 0.006, respectively). Moreover, only DOI and perineural invasion (PNI) were associated with neck metastasis (p = 0.007/0.016) (Table 2). Level IV metastasis was present in six of the dissected necks, and none were isolated to level IV. The cutoff DOI was 8 mm for level IV metastases. The critical size for level IV metastasis was 2.5 cm. The test sensitivity was 100%, and the selectivity was 69.23%. For metastatic lymph nodes, the cutoff value for DOI was also 8 mm. The test sensitivity was 64.71%, and the selectivity was 82.93% (Fig. 2).

Table 2.

The comparison results of the variables according to all regions of the neck and level IV

Variables	Level IV metastasis			p
	No (n = 52)	Yes (n = 6)
Tumor size (cm)a	2.4(1.5–3.0)	3.8(2.9–4.1)		0.010
DOI (cm)a	0.7(0.5–0.9)	1.2(1.0–1.8)		0.002
Age	56.67 ± 15.31	56.50 ± 20.72		0.173
Gender
Female	34(65.45%)	3(50.0%)		0.657
Male	18(34.6%)	3(50.0%)
Surgery
Partial glossectomy	46(88.5%)	5(83.3%)		0.823
Hemiglossectomy	5(9.6%)	1(16.7%)
Total glossectomy	1(1.9%)	0(0.0%)
Type of neck dissection
Unilateral neck dissection	46(88.5%)	5(83.3%)		0.591
Bilateral neck dissection	6(11.5%)	1(16.7%)
Differentiation
Well differentiated	35(67.30%)	0(0.0%)		0.009
Moderately differentiated	12(23.07%)	4(66.7%)
Poorly differentiated	5(9.61%)	2(33.3%)
PNI
Present	14 (26.92%)	5(83.3%)		0.006
Absent	38(73.07%)	1(16.7%)
LVI
Present	9(17.30%)	2(33.3%)		0.290
Absent	43 (82.69%)	4(66.7%)
	Neck metastasis
	No (n = 41)		Yes (n = 17)
Tumor size (cm)a	2.5(1.5–3.0)		2.6(1.9–3.8)	0.325
DOI (cm)a	0.6(0.5–0.9)		0.9(0.7–1.2)	0.007
Age	57.20 ± 14.76		55.35 ± 18.27	0.688
Gender
Female	26(63.4%)		11(64.7%)	0.926
Male	15(36.6%)		6(35.3%)
Surgery
Partial glossectomy	36(87.8%)		15(88.2%)	0.999
Hemiglossectomy	4(9.8%)		2(11.8%)
Total glossectomy	1(2.4%)		0(0.0%)
Type of neck dissection
Unilateral neck dissection	36(87.8%)		15(88.2%)	0.660
Bilateral neck dissection	5(12.2%)		2(11.8%)
Differentiation
Well differentiated	28(68.29%)		7(41.2%)	0.075
Moderately differentiated	10(24.39%)		6(35.3%)
Poorly differentiated	3(7.31%)		4(23.5%)
PNI
Present	9(21.95%)		10(58.8%)	0.016
Absent	32(78.04%)		7(41.2%)
LVI
Present	8(19.51%)		3(17.6%)	0.296
Absent	33(80.48%)		14(82.4%)

^aData are expressed as median (1st quartile–3rd quarter) and n (%)

Fig. 1.

Histopathological samples of the depth of invasion, well-differentiated SCC, poorly differentiated SCC, and perineural invasion

Fig. 2.

The performance of distinguishing tumor depth and tumor size for total neck metastasis and level IV metastasis

Seventeen (29.3%) of the patients experienced recurrence. Eleven patients had locoregional disease, and six had both locoregional and distant disease. Fifteen patients died due to illness. Level IV regional recurrence was not observed.

Discussion

The rich blood supply of the tongue, the presence of lymphatic flow, and the absence of fascia in the muscles are the factors that facilitate the spread of the disease [3, 6, 7]. Due to rich lymphatic flow, the occult metastasis rate reaches 45% in clinically N0 TSCC [8, 9]. This rate shows that neck treatment is one of the most critical issues [10]. Although some authors prefer a wait-and-see policy for early-stage tumors, most yield prophylactic neck dissection (PND) [11–13]. Authors who prefer PND have claimed that surgery is necessary for a cure, proper staging, and adjuvant therapy [14, 15].

Cervical lymph node metastasis in early-stage tongue cancer usually occurs at levels I–III [16]. The occurrence of skip metastases to level IV indicates that this region should be included in prophylactic neck dissection. However, the necessity of prophylactic level IV neck dissection has recently become controversial. While some clinicians argue that this region should be included [17–20], others have stated that level I–III neck dissection is sufficient due to the low incidence of recurrence and metastasis in level IV patients [21, 22]. Including level IV in PND may increase the duration of the procedure and postsurgical morbidity.

Many studies on the relationships between lymph node metastasis and sex, age, histopathology, tumor size, and DOI have been published. Studies show that young people have a worse prognosis than older people [23]. The incidence of tongue SCC is increasing in the young population [2, 24]. Hilly et al. reported that younger people with TSCC had a greater rate of regional metastases than older adults did [25]. In that study, the authors also claim that the recurrence of the disease is more aggressive in young patients. However, recent studies do not support this argument and claim that cancer-specific survival and disease prognosis are similar between elderly and young people [26–28]. There was also no relationship between sex and nodal metastasis according to various studies [29, 30]. Our study revealed no relationship between neck metastasis and age or sex.

According to the WHO Classification of Head and Neck Tumors, prominent squamous pearls, intercellular bridges, eosinophilic cytoplasm, dyskeratotic cells, and the absence of hyperchromatic nuclei are associated with good differentiation. The lack of these features of squamous cells is considered poorly differentiated and is associated with increased cervical lymph node metastasis [31, 32]. This study revealed a relationship between poorly differentiated tumors and level IV metastasis. Studies focused on the entire neck rather than solely on the level IV region. Additionally, none of the well-differentiated tumors exhibited level IV metastasis in our study.

The lymphatic system protects our body from infections and prevents the invasion of tumor cells. However, if tumor cells invade lymph vessels, they can spread throughout the body via vessels. Lymphovascular invasion and PNI are poor prognostic factors for oral cavity cancer [29, 33]. The lymph node metastasis ratio is greater in patients with LVI and PNI [29, 34]. Our study demonstrated that the PNI was also significantly related to level IV metastasis.

Tumor size was used to establish the TNM stage. Most of the studies explained that tumor size is one of the most important criteria for prognosis [35]. One study reported that for tumors larger than 3 cm, the risk of metastatic lymph nodes increases [36]. Our study also revealed that the critical size for level IV metastasis was 2.5 cm.

Tumor thickness has also been shown to be an important prognostic factor for nodal metastasis and local recurrence [35]. In a study by Rastogi et al. [37], the authors investigated the relationship between tumor thickness and skip metastasis. They found that the tumor cutoff value for skip metastasis was 3 mm. According to a meta-analysis by Huang SH et al. [38], tumor thickness strongly predicts lymph node metastasis in patients with oral cavity cancer. This meta-analysis, which included sixteen studies, investigated the cutoff value for nodal metastasis and found that the cutoff point for tumor thickness was 4 mm. In a study by Yuen et al. [35], selective RT was recommended for patients with tongue cancer with tumors thicker than 9 mm. Although recent studies have shown that tumor thickness is correlated with DOI, DOI is more important for metastasis [39]. While previous studies have indicated that a tumor DOI greater than 3 mm increases the risk of neck metastasis [40], recent studies have shown that a DOI of 7.25 mm is the most significant predictive factor for occult disease [41]. Terada et al. showed that for the rate of occult lymph node metastasis per DOI, there were two peaks at 5 to 6 mm and 9 to 10 mm [42]. Some studies have suggested that elective neck dissection should be considered for tongue cancer patients if the invasion depth is greater than 4 mm [43, 44]. In our study, 17 (29.3%) patients who were clinically N0 had positive lymph node metastasis, and level IV metastasis was observed in 6 patients. In addition, the cutoff value for the DOI of the evaluated tumor was 8 mm in both the metastatic group and the level IV metastasis group (Table 3).

Table 3.

Table showing the ROC (receiver operating characteristic) analysis

	Variables	Roc statistics		Diagnostic statistics
		AUC	p	Sensitivity (95% CI)	Specificity (95% CI)	Positive predictive value (%95 CI)	Negative predictive value (%95 CI)
Neck metastasisis	Tumor size > 2	58.20	0.329	70.59(44.1–89.6)	48.78(32.9–64.9)	36.36(20.40–54.88)	80.00(59.30–93.17)
	DOI > 0.8	72.50	0.004	64.71(38.4–85.7)	82.93(67.9–92.8)	50.00(28.22–71.78)	83.33(67.19–93.63)
Level 4 metastasis	Tumor size > 2.5	81.40	0.004	100.00(54.07–100.00)	63.46(48.96–76.38)	24.00(45.13–93.56)	100.00(89.42–100.00)
	DOI > 0.8	87.00	< 0.001	100.00(54.07–100.00)	69.23(54.90–81.28)	27.27(10.73–50.22)	100.00(90.26–100.00)

AUC area under the curve, CL confidence interval

A study by Byers et al. suggested the addition of level IV dissection to supraomohyoid neck dissection and suggested that skip metastasis is more common in aggressive SCCs. However, in the Byers study, clinical N0 and N + patients were included, and 15.8% of these patients were evaluated as having skip metastases [17]. Crean et al. [18]. advocated extending supraomohyoid neck dissection in their study. In this study, 49 patients with oral cavity cancer were included. Of these patients, 14 had tongue cancer, and 5 had level IV skip metastases. Only 3 of these patients had isolated level IV metastases. In addition, the number of T4 patients in the present study was greater than that in our study.

According to Shah et al. [45], 192 of 501 patients with oral cavity cancer had N0 disease, and radical neck dissection was performed. Only 6 of these patients (3%) were level IV positive. Li et al. [46] studied 384 patients with upper aerodigestive SCC. Of these patients, 153 had oral cavity cancer, and 5 (3.2%) had level IV metastases.

These studies show that isolated level IV metastasis is rare. In the patients in our study, there was also no isolated level IV metastasis. Many studies have shown that MRI is useful for predicting the DOI [47, 48]. The use of frozen sections is also helpful for determining the DOI perioperatively [49, 50]. Imaging systems and intraoperative frozen sections can be used to determine the DOI. These findings might help surgeons decide whether to perform level IV dissection. Thus, complication rates can be reduced.

Conclusion

Determining the depth of invasion preoperatively or perioperatively could help predict level IV metastasis. Predicting the occult level IV metastases prevent the unfavorable complications and long operation times. Our study revealed that poorly differentiated tumors larger than 2.5 cm and deeper than 8 mm with perineural invasion are at high risk for level IV metastasis.

Author contributions

İ.Y and A.O.Ç. wrote the main manuscript text. K.D. prepared figures. A.V. and S.Ç. prepared tables. V.Ç analyzed the data. S.D. and M.G. evaluated Magnetic Resonance imaging. All authors reviewed the manuscript.

Funding

This study was not supported by any funding.

Data Availability

The data that support the findings of this study are available on request from the corresponding author.

Declarations

Conflict of interest

There are no financial or non-financial interests that are directly or indirectly related to the work submitted for publication.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Approval for the study was granted by the Institutional Clinical Research Ethics Committee (Decision no: 2021/257).

Informed Consent

Informed consent was obtained from all indıvidual participants included in the study.

Consent for Publication

Consent for publication was obtained forevery indıvidual person’s data included in the study.