Given that few studies attempted to determine whether the ultrasound appearance of the primary thyroid tumor could be used to predict cervical LNM, this study aimed to identify the ultrasound features of the primary thyroid tumor that could be associated with cervical LNM in PTC. The results suggest that age and tumor size were independently associated with CLNM in patients with PTC. Age, tumor size, and percentage of capsular extension at ultrasound were independently associated with LLNM.
This study revealed that among patients with suspected PTC, age < 45 years, tumor size > 10 mm, and capsular extension > 50% were independently associated with LNM. The overall incidence of cervical LNM in the present study was approximately 38.5% (95/247), consistent with previous reports . The incidence of skip metastasis (i.e., LLNM without CLNM) was 6.9% (17/247). Women accounted for more than 70% of the patients in each group, and the sex distribution was similar among the three groups. Although PTC incidence is higher among women, men require specialized thyroid checkups to enable the early detection of thyroid tumors . Younger age (< 45 years) has been associated with an increased risk of CLNM . In a previous study, the univariable analysis showed that LNM is more likely among men, patients younger than 45 years, patients with multifocal tumors, and patients with local infiltration . Consistent with this, the present study showed that patients were more likely to be younger than 45 years old in the CLNM (72.3%) and LLNM groups (64.6%) than in the NLNM group (36.8%).
Tumor shape and capsular extension at ultrasound were related to cervical LNM, but tumor margins, calcification, echogenicity, and consistency were not. Previous studies reported that a taller-than-wide shape is a useful predictor of thyroid malignancy [22,23,24]. The dense fibrosis in a PTC may lead to decreased compressibility and a taller-than-wide shape . In the present study, a taller-than-wide shape was significantly more common in the NLNM group (38.8%) than in the CLNM (23.4%) or LLNM group (18.7%).
Capsular extension, specifically the degree of capsular extension and capsular disruption, can predict extrathyroidal extension and invasive thyroid cancer . Skip metastases are more common in patients with PTC and primary tumor capsular invasion than patients without capsular invasion . In the present study, capsular extension > 50% was the most common in the LLNM group (35.4%). Irregular tumor margins are a sign of malignancy. In this study, the incidence of non-smooth margins was more than 90% in all three groups. This high incidence might be attributable to the use of the high-resolution US and strict observation standards. Microcalcification was frequently observed in the present study, but its incidence did not significantly differ among the three groups. Only a few studies included capsular extension as one of the US features [5, 14, 26, 27]. The present study suggests that capsular invasion is significantly associated with LLNM. In other words, when a capsular extension is found by US, the patient should be considered at a relatively high risk of LLNM, and the surgeons might consider performing lateral lymph node dissection.
Tumor size is associated with the extent of cancer cell proliferation, and a faster rate of cell proliferation correlates with a higher risk of CLNM . Accordingly, tumor size is a strong predictor of microscopic CLNM and LLNM in N0 PTC patients . In the present study, tumor size was largest in the LLNM group (P < 0.001). A tumor size > 10 mm was more frequent in the CLNM (70.2%) and LLNM groups (89.6%) than in the NLNM group (45.4%).
A recent study revealed that Delphian lymph node (DLN) metastasis in PTC patients was related to tumor location in the isthmus or upper third of the thyroid . Azizi et al.  found that thyroid nodules in the isthmus were more likely to be malignant than nodules elsewhere in the gland. In the present study, 55.1% (136/247) of the patients had tumors in the middle third of the thyroid. Only four patients had tumors located in the isthmus. Tumor location did not significantly differ among the three groups.
Most patients with PTC (18–87%) have multifocal tumors. Some studies found that multifocality is associated with a high risk of LNM among patients with thyroid microcarcinomas, but other studies reported that the risk of LNM does not differ between those with unifocal and multifocal PTC [21, 31]. In the present study, the incidence of multifocality did not differ among the three groups.
Hashimoto’s thyroiditis is the most common form of autoimmune thyroid disease, with an incidence rate of about 2% in the general population. Some investigators reported that Hashimoto’s thyroiditis is a risk factor for PTC, while others found no correlation between the two . A study revealed that the incidence of Hashimoto’s thyroiditis among PTC patients ranges from 9 to 58% . In the present study, the incidence of Hashimoto’s thyroiditis did not significantly differ among the three groups.
Partially cystic and iso- or hyperechoic nodules are generally benign, with a low malignancy risk. In the present study, most of the tumors were hypoechoic, and the incidence of hypoechogenicity did not differ among the three groups. PTC is almost always solid, and cystic changes in PTC are rare. Some studies showed that mixed echogenicity (cystic component > 50%), accompanied by a honeycomb appearance, is an indication of benignity .
There are several potential limitations to this study. First, this is a retrospective study, and further prospective study is necessary to confirm the results. Second, we did not compare the US features of lymph nodes with and without metastasis. Third, we did not examine the association between capsular extension at ultrasound and the actual capsular contact at the histological level. Fourth, the signal intensity-related parameters showed inter-individual differences and were affected by certain conditions. Fifth, the sub-types of PTC were not recorded. Sixth, the BRAF V600E mutation was not examined. Finally, only patients with PTC were included in this study and not patients with medullary or undifferentiated carcinomas. This study, therefore, does not represent all the pathological subtypes of thyroid carcinoma.