Obesity Induces Different Regional Patterns of Lymph Node Involvement: Assay on 8,979 Thyroid Cancers

Introduction Lymph node (LN) metastasis is the rst station of extra-gland metastasis of thyroid cancer. LN status can inuence clinical decisions and the prognosis of patients. The aim of our study was to explore the relationship between obesity and regional patterns of LN involvement in papillary thyroid carcinoma (PTC). Materials and methods This study retrospectively analyzed the data from 8,979 thyroid cancer patients. The rate of LN metastasis, the number of LN metastasis, the maximum diameter of positive LN, the number of LN dissections, and the LN ratios (LNR) were compared between normal-weight and obese patients. Results LN metastasis was more common in the obese group than normal-weight patients with PTC (52.6 vs. 42.0%, P <0.001). The number of LN metastasis and the maximum diameter of positive LNs were also higher in obese patients (2.41 vs. 1.79, P <0.001; 2.00 vs. 1.60, P =0.007, respectively). The number of lateral neck LN dissections was higher in the obese group (14.37 vs. 12.10, P <0.001), there was no signicant difference in the central LNs. The LNR was higher in the central LNs of obese patients (0.18 vs. 0.14, P <0.001), yet there was no difference in the lateral LNs. Conclusions Obesity was associated with poor prognoses with PTC, which was related to the LNs. There was an inverse regional difference (central and lateral LNs) between obesity and the number of LN dissections and LNR, suggesting that caution was needed when performing central neck dissection in obese patients.


Introduction
The prevalence of obesity is approximately 40% worldwide, affecting more than 2 billion adults (1).
Obesity has been identi ed as an independent risk factor for many cancers. Some studies have reported that nearly 40% of cancers may be attributed to obesity (2), and there is strong evidence to suggest that obesity is related to cancers of the esophagus, liver, pancreas, gallbladder, ovary, thyroid, kidneys, and plasma cells (3). Although obesity has been linked to an increased risk of diabetes and coronary artery disease, the impact of obesity on the incidence rates, risk factors, morbidity, and mortality of thyroid cancer requires further exploration (4). As thyroid cancer is the most common type of endocrine tumor, understanding how body mass index (BMI) impacts this disease has vital public health implications (5).
Obesity also affects the diagnostic assessment of patients negatively. Deglise et al. found that obese women were less likely to have undergone ultrasound (OR = 0.5) or MRI (OR = 0.3) and were at an increased risk of prolonged hospital stays (OR = 4.7) in the clinic (6).
A signi cant association has been established between elevated BMIs and increased papillary thyroid carcinoma (PTC) incidence rates (7). In a previous study, we identi ed obesity as a risk factor for thyroid cancer, speci cally when tumor sizes were larger than 1 cm with multifocality and extrathyroidal extensions (8)(9)(10). However, the connection between obesity and lymph node (LN) metastasis of thyroid cancer has not been thoroughly investigated. One of the most important prognostic factors in PTC is LN status (11). LN metastasis is the rst station of extra-gland metastasis of thyroid cancer (11). Hence, LN metastasis is an excellent predictor of the prognosis of patients with thyroid cancer. LN status can in uence important clinical decisions, such as therapeutic options (11). The American Thyroid Association (ATA) guidelines use the number of metastatic LNs and the maximum diameter of positive LNs as important indicators for predicting the risk of recurrence (11). Compared to patients with less than ve metastatic LNs, the recurrence rates are much higher for patients with more than ve metastatic LNs (19% vs. 8%) (12,13). In terms of maximum positive LN diameter, recurrence rates are signi cantly higher in patients with LNs smaller than 3 cm (27% vs. 5%) (13).
In this study, we focused on the relationship between obesity and the patterns of LN involvement in PTC, with an emphasis on the number of metastatic LNs, the maximum diameter of positive LNs, the lymph node ratio (LNR), the number of dissected LNs, and the LN skip metastasis.

Materials And Methods
Setting and data collection

Participants
This study is a retrospective analysis of patients with operable PTC. All the patients eligible for this analysis received central neck node dissection. Patients were aged ≥ 18 years, pathologically-con rmed to have PTC, and normal weight or obese (according to WHO standards). The exclusion criteria included: non-PTC patients, other thyroid cancer subtypes, different types of cancer, and patients requiring reoperations (Fig. 1).

Treatments
Oncological treatment, which ranged from surgery to radioactive iodine (RAI), was standardized for all patients and in accordance with the multidisciplinary tumor board consensus. The histological subtype was assessed according to the WHO classi cation. According to the Chinese guidelines for diagnosis and treatment of differentiated thyroid, all patients with thyroid cancer routinely underwent prophylactic central neck dissection (CND) (14). For those LNs suggested to be malignant by ultrasonography, ne needle aspiration cytology (FNAC) was used to con rm the diagnosis. Patients with cervical lymph node metastasis were con rmed by preoperative FNAC or intraoperative frozen pathological examinations.
These patients underwent therapeutic cervical lymph node dissection. Prophylactic cervical lymph node dissection was not recommended. The upper bound of the range of CND is the lower hyoid bone, the lower bound is the superior sternum fossa, the outside is the common carotid artery, and the inside is the inside of the trachea. The lateral LN dissection minimum ranged from the IIa, III, IV, and Vb area, while other areas were treated according to the results of the FNAC.

De nitions
Body-mass index. According to the WHO-BMI standard, BMI is between 18.5 ≤ BMI < 25 kg/m 2 , which was de ned as normal-weight. Obesity is de ned as BMI ≥ 30 kg/m 2 (15). Height and weight measurements used to calculate the BMI were retrieved from electronic registration databases. BMI was calculated as weight divided by height squared. Upon the rst admission, the demographics and clinical data, including height and weight, were recorded.
Pattern of metastatic LNs. The pattern of metastatic LNs included the rate of metastatic LNs, the number of metastatic LNs, the number of dissected LNs, the lymph node ratio (LNR), the maximum diameter of positive LNs, and the rate of LN skip metastasis.
Maximum diameter of positive LNs. The maximal tumor diameter of the largest metastatic LN using the concept of micro-metastases in breast cancer (16).
Lymph node ratio (LNR). LNR was de ned as the number of nodes involved by the tumor divided by the total number of resected lymph nodes during the surgical treatment.
LN skip metastasis. De ned as a lateral lymph node metastasis without central lymph node involvement.

Pathological examination
Specimens removed during the operation were submitted for histological analysis to determine the presence and size of metastatic LNs. For this study, an experienced pathologist reviewed the pathological slides and measured the properties of each LN.

Outcomes and covariates
We used BMI as an index and further correlated it with the presence and pattern of LN metastases. Patients were identi ed as LN positive if they had a pathologic LN status of pN1a or pN1b. Based on prior studies, we included important predictors of LN status in multivariable models (17). For example, information regarding the following thyroid tumor characteristics was obtained from medical records: age, weight, height, tumor size (mm), multifocality, extrathyroidal extension, rate of metastatic LNs (%), number of metastatic LNs, number of LN dissections, LNR (%), maximum diameter of positive LNs (mm), and LN skip metastasis (%).

Statistical analysis
All data were collected using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). Continuous variables were expressed as mean (standard deviation), and categorical variables were expressed as a percentage (frequency). Continuous variables were analyzed by the t-test or one-way analysis of variance (ANOVA). Categorical variables were examined using the χ 2 -test or Fischer's exact test. P < 0.05 was considered statistically signi cant. Statistical analysis was performed using SPSS 22.0 software (IBM, Chicago, IL, USA).

Baseline characteristics of patients
Our database included 14,795 patients with PTC, of which 8,979 patients were included in the nal analysis (Fig. 1). The male-to-female ratio is 1:5. The average age is 42.52 ± 9.38 years. The average BMI is 23.26 ± 3.34 kg/m 2 . Obesity accounted for 9.4% of patients.

Baseline characteristics of LNs
In total, 110,580 LNs were dissected, of which 16,589 LNs were metastases (15.0%). The rate of metastatic LNs was 43.0% (3858/8979). The mean number of LN metastases was 1.85 ± 3.73. The average number of LN dissected was 12.32 ± 11.86. The LNR was 0.14 ± 0.23%, and the maximum diameter of positive LNs was 1.64 ± 3.99 mm. The rate of LN skip metastasis is 4.1% (367/8979), as shown in Table 1. Abbreviation: BMI = body mass index, TSH = thyroid stimulating hormone, FT3 = free triiodothyronine, FT4 = free thyroxine, LN = lymph node, LNR = lymph node ratio. * P < 0.05, ** P < 0.01 Impact of BMI on the rate of metastatic LNs As shown in Table 2, the rate of metastatic LNs in obese patients with PTCs was signi cantly higher than that of normal-weight patients (52.6 vs. 42.0%, P < 0.001). The rates of central metastatic LNs and lateral neck metastatic LNs were signi cantly higher in obese patients than normal-weight patients (47.6 vs. 38.0%, P < 0.001; 22.2 vs. 16.9%, P < 0.001). The rates of lateral neck metastatic LNs on both the left and right sides were higher than normal-weight patients (11.1 vs. 8.6%, P = 0.015; 13.7 vs. 9.8%, P = 0.001).

Impact of BMI on the size of positive LNs
As shown in Fig. 2A, the maximum diameter of positive LNs in obese patients was higher than normalweight patients (2.00 vs. 1.60 mm, P = 0.007). There was no signi cant difference in the rate of LN skip metastasis between the two groups (Fig. 2B).

Impact of BMI on the number of dissected LNs
The total number of LN dissections in the obese patients with PTC was signi cantly higher than the normal-weight patients (14.37 vs. 12.10, P < 0.001). The obese patients have a higher total number of lateral neck LN dissections (8.51 vs. 6.36, P < 0.001). This difference exists on the left and right sides of the neck (Table 4). However, there was no difference in the number of CNDs (P = 0.466).

Impact of BMI on the LNR
The LNR in obese patients was signi cantly higher than normal-weight patients (0.18 vs. 0.14, P < 0.001) ( Table 5). While the difference was statistically signi cant in the central LNs (0.79 vs. 0.58, P < 0.001), but there was no statistically signi cant difference in the lateral neck LNs (P = 0.067) ( Table 5).

Discussion
This study is a retrospective analysis of 8,979 patients with PTC, with a focus on the correlation between LN status and obesity. Obesity not only increased the rate of metastatic LNs, but also increased the number of metastatic LNs and the maximum diameter of positive LNs, which are two indicators of a poor prognosis. We analyzed two indicators that have been overlooked in previous studies, including the number of LN dissections and the LNR. These results showed opposite regional differences between obesity, the number of LN dissections, and the LNR (Fig. 3).

Effect on the rate of metastatic LNs
Previous studies on the relationship between obesity and metastatic LNs in thyroid cancer have been controversial (  (8,23,24). One of the possible mechanisms is that serum leptin levels are higher in obese patients with PTC. In vitro studies have indicated that leptin promotes invasion and migration of thyroid cancer cell lines (29,30). However, Kim, Gasior, Grani, Tresallet, and others have found that obesity is not associated with neck metastatic LNs in patients with thyroid cancer (19,20,26,28).
Simultaneously, Paes et al. found that obesity is negatively correlated with metastatic LNs (18). This nding may be due to differences in ethnicity. In the current study, we found that the rate of metastatic LNs in obese patients with PTC was signi cantly higher than normal-weight patients (52.6% vs. 42.0, P < 0.001). This effect exists in both the central and lateral neck regions (Fig. 3).

Number of metastatic LNs and the sizes of positive LNs
The 2015 ATA guidelines suggest that more than ve metastatic LNs yields an intermediate risk of recurrence (31). Previously, Leboulleux et al. found a recurrence rate of 3% with less than ve metastases, while 6 to 10 metastases was associated with a recurrence rate of 7% and > 10 metastases with 21% (32). Previous studies have given minimal attention to the relationship between the number of metastatic LNs and obesity. The current study revealed a positive correlation between obesity and metastatic LNs (2.41 vs. 1.79, P < 0.001), which was re ected in the central and lateral neck LNs (Fig. 3).
The maximum diameter of positive LNs is another indicator of poor prognoses. The ATA guidelines classify the maximum diameter of positive LN between 0.2 and 3.0 cm as the intermediate risk of recurrence (31). In a previous study, the rate of locoregional recurrence was 5% for patients with maximum positive LN diameters smaller than 0.2 cm (12). When the maximum diameter of positive LNs were > 3 cm, the rate of locoregional recurrence increased to 27% (13). In the current study, obese patients with PTC had larger maximum positive LN diameters (2.00 vs. 1.60, P = 0.007), suggesting that obesity may lead to poorer prognoses. Previous studies have primarily shown that obesity increases the rate of LN metastasis. This paper con rmed that obesity not only increases the rate of metastatic LNs, but also increases the number of metastatic LNs and the maximum diameter of positive LNs.

Number of LN dissections
Neck LNs are commonly wrapped in adipose tissue, and obese patients have more adipose tissue. Only a few studies have focused on whether the increased adipose tissue in obese patients can affect the dissection of neck LNs. This study found that obese patients had more dissectible LNs, yet this association only exists in the lateral neck area (14.37 vs. 12.10, P < 0.001), which seems to be more bene cial for obese patients. However, the number of CNDs is not related to obesity. There is a regional difference between the number of LN dissections and obesity (Fig. 3). One of the possible reasons is that the di culty of intraoperative identi cation of LNs is reduced. LNs are often surrounded by adipose tissue, which helps the surgeon identify the LNs and minimize the burden of identi cation. The second reason involves in ammatory factor stimulation, as obesity can cause adipose cells to secrete in ammatory factors, such as C-reactive protein, interleukin 6 (IL-6), IL-10, and tumor necrosis factor (TNF-α). These in ammatory factors may stimulate LN hyperplasia (33).

LNR
After discovering a positive relationship between obesity and the number of lateral neck LN dissections, we speculated whether more LN dissections would yield higher positivity rates, which could contribute to the radical cure of thyroid cancer in obese patients. However, by analyzing the LNR, we found that obesity did not increase the positive rate of lateral central LNs. On the other hand, obesity was found to increase the positive rate of central LNs. Hence, there are regional differences in the LNR and the number of LN dissections.
Obese patients have di culty in performing CNDs due to their short and thick necks. In this study, obesity was not associated with an increase in the number of CNDs, but it was associated with an increased LNR. Hence, surgeons should be cautious when performing CNDs for obese patients, as there may be more positive LNs in obese patients. If positive LNs are missed during surgery, then they will continue to cause damage in the body.
In the current study, we revealed that obesity promotes the metastasis of LNs and increases the average diameter of positive LNs. These two indicators are indicative of a poor prognosis. In addition, we considered the number of LN dissections and the LNR into the current analysis. We found a regional difference between obesity and these two indicators.

Limitations
The limitations of this study include its retrospective design, which is subject to incomplete data and potential selection bias. This paper did not analyze the relationship between obesity and LNs in various regions of the lateral neck. The number of positive nodes is often affected by the variability in nodal staging techniques, which may yield different numbers of excised nodes. Finally, this study did not analyze the disease-free survival rate or locoregional recurrence rate of patients, for which LN metastasis is a very important independent risk factor in patients with thyroid cancer after curative resection.

Conclusions
In summary, this study revealed that obesity is associated with poor prognoses in patients with PTC, which is related to the LNs. There is an inverse regional difference between obesity and the number of LN dissections and the LNR, suggesting that caution is needed when performing CNDs in obese patients.
This will help ensure that all positive LNs are dissected, which will increase the overall survival of patients with PTC.  Figure 1 Flow chart of study inclusion and exclusion criteria. The gure was adapted from a previously published paper.