The clinical value of Delphian and pre-tracheal lymph nodes in predicting lateral lymph nodes metastasis of papillary thyroid carcinoma

Abstract

Background

Occult lymph node metastasis of papillary thyroid carcinoma is common. However, whether undergoing prophylactic lateral lymph node dissections is still controversial. This cross-sectional study with large cohort of patients aims to investigate the clinical value of Delphian and pre-tracheal lymph node in predicting lateral lymph node metastasis of papillary thyroid carcinoma.

Materials and methods

A retrospective analysis was conducted on 865 papillary thyroid carcinoma patients with Delphian and pre-tracheal lymph node data who underwent thyroidectomy plus central and lateral lymph node dissection. Data on clinicopathological characteristics were collected. Subsequently, a predictive model was established based on the results of the univariate and multivariate analyses.

Results

The rates of Delphian and pre-tracheal lymph node metastasis and lateral lymph node metastasis were 54.7% and 39.1%, respectively. Having ≥ 3 or 1–2 Delphian and pre-tracheal lymph node metastasis dramatically increased the risk of lateral lymph node metastasis (OR = 8.5, 95% CI 5.3–13.4 and OR = 3.9, 95% CI 2.7–5.7, respectively). The upper tumour had a 3.7 times higher risk of lateral lymph node metastasis than other locations. Patients ≤ 42 years or tumour size >8 mm had a higher risk of lateral lymph node metastasis.

Conclusions

Delphian and pre-tracheal lymph node metastasis was associated positively with the risk of lateral lymph node metastasis. For patients without clinical lateral lymph node metastasis, the Delphian and pre-tracheal lymph node could be considered to harvest as the first step in a thyroidectomy to facilitate further conduct of the operation.

Introduction

With an anticipated 3.0% annual growth rate, 5,86,202 incident cases of thyroid cancer were reported worldwide in 2020, placing it 11th among all malignant tumours [1]. Papillary thyroid carcinoma (PTC) makes up about 95% of all thyroid cancers. It is distinguished by a tendency for early metastasis, with reported rates of lymph node metastasis (LNM) ranging from 20% to 90% at the time of diagnosis [2–4].

According to previous studies, occult lateral lymph node metastasis (LLNM) can reach 64% [4]. Fortunately, occult LLNM rarely impacts PTC patients’ overall survival, so prophylactic lateral lymph node dissection (LLND) is not recommended by the American Thyroid Association (ATA) guidelines [5]. But there is no denying that LLNM is associated with a shorter disease-free survival in PTC [6, 7]. Moreover, although lymph node recurrence does not immediately affect patients’ life prognoses, it significantly has adverse effects on the quality of life experienced by the patients [8, 9]. Therefore, the Chinese and Japanese guidelines suggest that for patients with intermediate and high-risk factors, LLND could be considered to reduce the risk of recurrence and secondary surgery [10, 11].

Consequently, increasing the accuracy of identifying metastatic lymph nodes at the lateral level is very important. Numerous prior clinical investigations have shown that patients with Delphian lymph node (DLN) metastasis were significantly likely to have LLNM [12, 13] and that DLN metastasis was linked to a poor prognosis [14, 15]. However, the clinical detection rate of DLN, which ranges from 23% to 69.4% [16, 17], is low, which restricts the application of DLN in clinical settings. Between the thymus and the isthmus, the pre-tracheal lymph nodes (PLN) are made up of six to eight lymph nodes. PLN, being a single subgroup of central compartments, is also commonly regarded as the initial stage of LNM. Due to their shallow placement, DLN and PLN are easily accessed, and dissection might not raise the operation time and the incidence of postoperative complications.

Thus, this retrospective study aimed to explore the significance of DLN and PLN, and investigate the value of DLN and PLN in predicting LLNM in PTC, to guide clinicians estimate the risk of LLNM and develop an individualized treatment plan for patients.

Materials and methods

Patients

The local institutional ethics committee approved this retrospective cross-sectional study verbally and informed consent was not required because there was no violation of patient privacy or interest in having personal information disclosed. Data were gathered from 3916 consecutive patients who underwent thyroid surgery at the Department of Breast and Thyroid Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China, between January 2013 and December 2018. Adolescents, distant metastases, clinical LLNM (cN1b), absence of central lymph node dissection (CLND) and LLND, history of reoperation or other head and neck surgery, unclear DLN and PLN data, and insufficient clinicopathological data were among the factors that precluded patients from participation. Ultimately, this study included 865 eligible individuals (Figure 1).

Figure 1.

Papillary thyroid carcinoma patients exclusion flowchart.

Surgical procedure

At our center, a unilateral lobectomy plus an ipsilateral central lymph node dissection (CLND) was the basic surgical approach. We routinely performed CLN intraoperative frozen section biopsy and if central lymph node metastasis (CLNM) was confirmed in intro-operative finding, and combined with the patient’s wishes and other risk factors, prophylactic LLND was considered. For cN1b patients, therapeutic LLND was executed. All resected specimens were sent to the Department of Pathology for examination. The patients or family members were informed of all surgical procedures and all agreed.

Demographic and clinicopathological data

Data on the following clinicopathological variables were collected: sex, age, size, location, multifocality, laterality, capsule invasion, Hashimoto’s thyroiditis (HT) and LNM. cN1b was considered if the physical or imaging examination indicated LLNM. The largest lesion diameter and location were recorded as the size and location. tumour size and location were described in the preoperative ultrasonography report. Micro-extra-thyroid-extension (mETE) and gross-extra-thyroid-extension (gETE) were recorded in intraoperative or postoperative findings. Multifocality and HT were recorded in both preoperative and intraoperative data. LNM was confirmed by postoperative examination.

Statistical analysis

The Mann–Whitney U test was performed to compare differences in nonnormally distributed continuous variables. The independent-sample t-test was used to analyze continuous variables that are normally distributed. Categorical variables were compared using the Chi-square test. Statistical significance was set at p < 0.05. Variables that showed a significant association were included in the multivariable logistic regression model. Based on this model, a nomogram was developed to predict the probability of LLNM, and the predictive ability of the nomogram was measured using the receiver operating characteristic (ROC) curve, and concordance index (C-index). All statistical analyses were performed using SPSS (version 25.0; SPSS Inc., 96 Chicago, IL, United States) and graphs were generated using R software version 4.3.1.

Results

Clinicopathologic characteristics

The majority of the 865 patients were women (630, 72.8%), and the ratio of male to female was 1:2.7. The patients’ ages ranged from 18 to 83 years old (mean = 42.2 years; median = 40 years). Tumour size ranged from 3 to 64 mm, with a median size of 10 mm. The median value of the metastatic number of DLN and PLN was 1 (IQR 0, 2). Table 1 and Supplementary Tables 1 and 2 provide additional information.

Table 1.

Patient characteristics.

Characteristics	Data, N (%)
Total	865 (100)
Sex
Male	235 (27.2)
Female	630 (72.8)
Age years, median (IQR)	42.0 (33.0–50.5)
BMI, median (IQR)	22.9 (20.9–25.4)
Size mm, median (IQR)	10 (8–15)
cN
cN0	745 (86.1)
cN1a	120 (13.9)
Location
Upper	232 (26.8)
Other	633 (73.2)
mETE
No	701 (81.0)
Yes	164 (19.0)
gETE
No	786 (90.9)
Yes	79 (9.1)
Laterality
Left	335 (38.7)
Right	381 (44.0)
Bilateral	149 (17.2)
Multifocality
No	725 (83.8)
Yes	140 (16.2)
HT
No	668 (77.2)
Yes	196 (22.7)
DLN and PLN
Negative	392 (45.3)
Positive	472 (54.7)
Metastatic number of DLN and PLN, Median (IOR)	1 (0–2)
Harvested number of DLN and PLN, Median (IOR)	6 (4–8)
Metastatic rate of DLN and PLN, Median (IOR)	0.13 (0–0.41)
LLNM
No	527 (60.9)
Yes	338 (39.1)

Abbreviations: BMI, body mass index; cN, clinical lymph node; cN0, clinical lymph node-negative; cN1a, clinical lymph node positive in central compartment; mETE, micro-extra-thyroid-extension; gETE, gross-extra-thyroid-extension; DLN, Delphian lymph nodes; PLN, pre-tracheal lymph nodes; LLNM, lateral lymph node metastasis.

Univariate analysis

Among the 865 patients, there were 338 (39.1%) patients with LLNM. The median metastatic number of the DLN and PLN in patients without and with LLNM were 0 and 2, respectively. Dividing the number of DLN and PLN metastasis into three groups (0, 1–2, and ≥3), the proportion of LLNM particularly is higher in the DLN and PLN metastasis ≥ 3 groups. Furthermore, LLNM was significant with age (≤42 years n = 203/453, 44.8%; >42 years n = 135/412, 32.8%), size (≤10 mm, n = 110/444, 24.8%; >10 mm, n = 228/421, 54.2%), cN (cN0 n = 281/745, 37.7%; cN1a n = 57/120, 47.5%), tumour location (upper n = 133/232, 57.3%; other n = 205/633, 32.4%), gETE (no n = 298/786, 37.9%; yes n = 40/79, 50.6%), laterality (left n = 105/335, 31.3%; right n = 156/381, 40.9%; bilateral n = 77/149, 51.7%), multifocality (no n = 270/725, 37.2%; yes n = 68/140, 48.6%) (all p < 0.05) (Table 2). Conversely, no significant association was discovered between LLNM and sex, BMI, mETE, and HT (Table 2).

Table 2.

Lateral lymph node metastasis characteristics.

Characteristics	LLNM, N (%)		p-Value
	Negative	Positive
	527 (60.9)	338 (39.1)
Sex			0.057
Male	131 (55.7)	104 (44.3)
Female	396 (62.9)	234 (37.1)
Age years, median (IOR)	43.0 (34.0, 52.0)	40.0 (31.0, 48.0)	<0.001
≤42 Years	250 (55.2)	203 (44.8)	<0.001
>42 Years	277 (67.2)	135 (32.8)
BMI, median (IOR)	23.0 (20.9, 25.4)	22.9 (20.9, 25.2)	0.654
Size mm, median (IOR)	9.0 (7.0, 13.0)	13.0 (10.0, 19.0)	<0.001
≤8 mm	236 (80.3)	58 (19.7)	<0.001
9–10 mm	98 (65.3)	52 (34.7)
11–15 mm	106 (49.8)	107 (50.2)
>15 mm	87 (41.8)	121 (58.2)
cN			0.042
cN0	464 (62.3)	281 (37.7)
cN1a	63 (52.5)	57 (47.5)
Location			< 0.001
Upper	99 (42.7)	133 (57.3)
Other	428 (67.6)	205 (32.4)
mETE			0.486
No	431 (61.5)	270 (38.5)
Yes	96 (58.5)	68 (41.5)
gETE			0.027
No	488 (62.1)	298 (37.9)
Yes	39 (49.4)	40 (50.6)
Laterality			< 0.001
Left	230 (68.7)	105 (31.3)
Right	225 (59.1)	156 (40.9)
Bilateral	72 (48.3)	77 (51.7)
Multifocality			0.012
No	455 (62.8)	270 (37.2)
Yes	72 (51.4)	68 (48.6)
HT			0.083
No	418 (62.5)	251 (37.5)
Yes	109 (55.6)	87 (44.4)
DLN and PLN metastasis, Median (IOR)
0	0 (0, 1)	2 (1, 3)	< 0.001
1–2	317 (80.9)	75 (19.1)	< 0.001
≥3	150 (51.2)	143 (48.8)
	60 (33.3)	120 (66.7)

Abbreviations: BMI, body mass index; cN, clinical lymph node; cN0, clinical lymph node-negative; cN1a, clinical lymph node positive in central compartment; mETE, micro-extra-thyroid-extension; gETE, gross-extra-thyroid-extension; DLN, Delphian lymph nodes; PLN, pre-tracheal lymph nodes; LLNM, lateral lymph node metastasis.

Multivariate analysis

In the multivariate analysis, the probability of LLNM was increased in patients with ≥3 or 1–2 DLN and PLN metastasis, compared with those without DLN and PLN metastasis. The upper tumour had a 3.7-fold increased risk than the other locations (OR = 3.7, 95% CI 2.5–5.3, p < 0.001). In addition, younger patients indicated 1.5 times higher odds of LLNM (OR = 1.5, 95% CI 1.1–2.1, p = 0.013). compared with tumour size ≤8 mm, patients with larger tumour size had a higher risk of LLNM. More details are shown in Table 3 and Figure 2. No important differences in LLNM were found between the cN, gETE, laterality, and multifocality (p > 0.05) (Table 3).

Table 3.

Risk of lateral lymph node metastasis adjusted for other variables.

Variables	OR	95% CI	P-value
DLN and PLN metastasis
0–Ref	–	–	–
1–2	3.9	2.7–5.7	<0.001
≥3	8.5	5.3–13.4	<0.001
Age
≤42 years	1.5	–2.1	0.013
>42 years–Ref	–	–	–
Size
≤8 mm – Ref	–	–	–
9–10 mm	2.0	1.2–3.3	0.005
11–15 mm	3.4	2.2–5.3	< 0.001
>15 mm	4.8	3.0–7.5	< 0.001
cN
cN0–Ref	–	–	–
cN1a	0.7	0.5–1.8	0.201
Location
Upper	3.7	2.5–5.3	<0.001
Other–Ref	–	–	–
gETE
No–Ref	–	–	–
Yes	1.4	0.8–2.4	0.231
Laterality
Left	0.1	0.4–1.0	0.655
Right	0.8	0.6–1.5	0.935
Bilateral–Ref	–	–	–
Multifocality
No–Ref	–	–	–
Yes	1.4	0.9–2.2	0.149

Abbreviations: DLN, Delphian lymph nodes; PLN, pre-tracheal lymph nodes; cN, clinical lymph node; cN0, clinical lymph node-negative; cN1a, clinical lymph node positive in central compartment; gETE, gross-extra-thyroid-extension.

Figure 2.

Forest plots of risk factors for lateral lymph node metastasis.

Prediction mode and nomogram

Based on multivariate logistic regression analysis, we developed a prediction model for LLNM.

$$\begin{array}{l}\text{Model}:\text{Patient\hspace{0.17em}Age}+\text{Tumour\hspace{0.17em}Size}+\\ \text{Tumour\hspace{0.17em}Location}+\text{DLN\hspace{0.17em}and\hspace{0.17em}PL\hspace{0.17em}N\hspace{0.17em}metastasis}\end{array}$$

A nomogram incorporating the above four predictors is shown in Figure 3. The ROC curve is illustrated in Figure 4, and the AUC was 0.806 (95% CI, 0.777–0.835) and 0.800 by bootstrapping validation, which demonstrated good discriminative ability.

Figure 3.

Nomogram for prediction of lateral lymph node metastasis.

Figure 4.

ROC curves of the logistic regression model.

Discussion

Accurate evaluation of LLN status is essential for PTC patients’ individualized treatment. Based on 865 PTC patients in this study, we found that the risk of LLNM was significantly increased by DLN and PLN metastasis. Our study indicates that age ≤42 years, larger tumour size, upper lobe tumour, and DLN and PLN metastasis were all positively correlated with LLNM risk. Together, our data signify that readily available preoperative and intraoperative clinicopathological characteristics can be utilized to assess the risk of LLNM and facilitate surgical decision-making with individualized lymphadenectomy for patients.

The DLN receives lymphatic drainage from the larynx and the thyroid gland then flows towards the CLN and LLN [18]. Previous investigations have demonstrated that DLN metastasis is a predictor of further disease in the CLN and LLN [14, 16, 19]. However, we found that the detection rate of DLN is not particularly high in our daily clinical work, which limits the clinical application value. As a subgroup of CLN, PLN is located in the fascia between the thyroid gland and the thyroid isthmus, with a shallow position and larger number, which is easy to extract during operation, which is also of great significance in PTC. Nevertheless, PLN is less well studied in PTC. In Qing study, PLN metastasis was significantly associated with contralateral CLNM [20]. Moreover, Wang et al. found that patients with positive PLN had a higher risk of more than 5 CLNM, revealing the significant importance of PLN [21].

As a result, differentiated from earlier research, we first combined DLN and PLN to predict the metastasis risk of the LLN, committed to improving the accuracy of the comprehensive assessment of LLNM. Based on the results obtained, we think that for patients with cN0 or cN1a, harvesting the DLN and PLN can be considered as the first step in a thyroidectomy to facilitate further conduct of the operation.

Tumours located in the upper pole are prone to LLNM is the point of view in most studies [22, 23]. This is closely related to the pattern of lymphatic drainage pathways, and lymphatic drainage occurs in the upper direction when the tumour is located in the upper third [24]. According to our analysis, the risk of LLNM was 3.7 times higher in the upper tumour than in the other places, consistent with earlier research [25]. In addition, a primary tumour located in the upper portion of the thyroid is an independent risk factor for skip metastasis [26, 27]. Therefore, when managing the upper tumour, the lymph node in the lateral compartment should be more vigilant by the surgeon.

Much evidence has demonstrated that a large tumour size is related to the LLNM and leads to advanced tumour staging [22, 28]. Nevertheless, Lee et al. showed that there was no difference in the rate of LLNM in micro-PTC and PTC [29]. Our findings showed as the diameter of the tumour increases, the risk of LLNM increases.

Younger patients with PTC are more likely to develop LNM, but the correlation between age and LLNM is controversial. Although some studies suggest no significant association between age and LLN, many studies suggest that younger age is a risk factor for LLNM [25, 30]. In the present study, younger age is an independent predictor of LLNM (OR = 1.5, 95% CI 1.1–2.1). Further research is needed to explore their correlation.

Based on the results of multivariate regression analysis, we established a nomogram of the possibility of LLNM with the C-index was 0.806 and 0.800 by bootstrapping validation. Objectively, the model lacks external validation and some preoperative thyroid-related laboratory indicators can also be included further.

There were several potential limitations to this study. The main limitation is that this is a cross-sectional study, lacking follow-up data to verify the recurrence rate and clinical benefit. Furthermore, we are currently examining more long-term follow-up data. Besides, this was a single-center retrospective study with a lack of prospective validation. In addition, more information on the DLN and PLN, such as the size of the metastatic lymph nodes, needs to be further studied.

Conclusions

In conclusion, our retrospective study has found that DLN and PLN metastasis, large tumour size, tumour in the upper pole, and age ≤42 years may increase the risk of LLNM. In clinical practice, when patients with these risk factors, thyroid surgeons should be cautious to evaluate the LLN status to make individualized lymphadenectomy for each patient.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Authors contributions

All authors have made substantial contributions to this work and have approved the final version of the manuscript. Concept and design: Xinliang Su, Tao Xu, Chun Huang; Acquisition of data: Xinliang Su, Chun Huang, Jing Zhou; Statistical analysis: Chun Huang, Jing Zhou; Data interpretation: Chun Huang, Jing Zhou, Yuchen Zhuang; Authorship of the original draft: Chun Huang; Review and editing: Xinliang Su, Tao Xu; Funding: Xinliang Su.

Disclosure statement

No potential conflict of interest was reported by the authors.

Ethics approval and consent to participate

All procedures performed in studies involving human participants were following 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. The study was approved by The Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (approval code: 2022-K501) and informed consent was waived because there was no patient interest or privacy breach in the form of disclosure of personal information.

Data availability statement

The datasets used and/or analyzed for the development of this manuscript are available from the corresponding author upon reasonable request.