The stratification of patient risk depending on the size and ratio of metastatic lymph nodes in papillary thyroid carcinoma

In this study, we attempted to create a risk stratification for the 435 patients who underwent postoperative RAI ablation due to nodal metastasis with PTC, and we obtained significant results. The sTg values exhibited a tendency to increase with risk as defined by the stratification of the three risk groups based on the size of mLN and LNR. The RFS was significantly different between the three groups. Risk stratification based on the size of mLN and the LNR could be considered a useful method for the classification of postoperative patients with PTC. In other words, the patient risk groups I, II, and III in the current study may be referred to as low-, intermediate-, and high-risk groups.

The size of the primary tumor or the mLN, the number of mLNs, the presence of ETE, and the tumor stage have already been demonstrated to be factors that are associated in increased risk in previous studies [11‐13].

Bardet et al. argued that modifications to the staging systems for the risk of recurrence of PTC are needed. This argument is supported by the finding that the 4-year risk of persistent/recurrent disease significantly differs according to size of mLN (pN1mac 49% versus pN1mic 24%, p = 0.03; pN0 12%, p = 0.01; or pNx 6%, p < 0.001). Moreover, the pN1mac was an independent predictor of persistent disease, but the pN1mic was not. Therefore, these parameters were described as pN1mic and pN1mac at quite different stages. However, this study did not mention any relations to the LNR [16]. A meta-analysis was published in 2012 and referred to risk stratifications of differentiated thyroid cancers that are based on the specific LN characteristics, such as the size of mLN and the number of m-LNs. Lower risk N1 disease was associated with conditions such as N0 clinical status, micrometastasis, and fewer than five m-LNs; these conditions predicted a recurrence rate below 5%. In contrast, higher risk N1 disease was associated with the conditions of N1 clinical status, macrometastasis greater than 3 cm, and a number of m-LNs greater than five; these conditions predicted a recurrence risk above 20% [11]. Moreover, Ricarte-Filho et al. attempted to stratify PTC patients with the cervical LNM using the number of mLMs, the oncogenic BRAF status, and the presence of ETE [13]. Although different methods have been used for stratification, these attempts at risk stratification can all be considered to have the goal of predicting the recurrence rates or prognoses of patients with PTC [11‐13, 17]. In another attempt of the current study, the stratification based on the performance of MRND and the LNR did not exhibit meaningful results. Another three risk groups were made for 580 patients who underwent RAI ablation treatment following thyroid surgery including MRND, as follows: group I', MRND (-) with LNR ≤ 0.5; group II', MRND (-) with LNR > 0.5 or MRND (+) with LNR ≤ 0.5; and group III', MRND (+) with LNR > 0.5. The number of patients in the risk group III' was only 17 patients, they were too small patients compared to the other two groups. Also, the sTg value did not exhibit a gradually increasing pattern (group I': 0.98 ng/ml; group II': 3.12 ng/ml; group III': 2.83 ng/ml).

The prognosis and recurrence of PTC are related to the presence of LN metastasis or ETE, greater primary tumor size or a greater number of mLNs, and a greater LNR [2, 4, 17]. Interest in the size of mLN has increased in recent years, although little research has been conducted in this area. In the majority of studies, the sizes of mLN have been compared following the classification of the patients as having micrometastasis or macrometastasis. Consequently, macrometastasis has been confirmed as a predictor of recurrence and poorer prognosis [1, 16, 18]. As illustrated in Table 2, the presence of gETE, and the mLN ratio of >0.5 were significantly associated with macrometastasis. The other factors that were found to be associated with the size of mLN were not significantly different from the results of previous studies [5, 6, 18].

Regarding the LNR, the studies that have included only central LN metastases are quite numerous. The central LNR is an independent factor related to the recurrence of PTC, and the macrometastasis of a CLN is clinically meaningful as an independent variable during the follow-up after surgery [19, 20]. However, analyses that have included both of central and lateral LN metastases are scarce. Therefore, large and well-designed studies are required in the future.

Additionally, the LNR cutoff value is still under debated, and the definitions of this cutoff value vary considerably between institutions. Studies published in 2013 and in 2014 reported LNR cutoff values of 0.4 and 0.65, respectively [5, 19]. Our statistical analyses were performed after defining a cutoff value of 0.5 based on the ROC curve method. Consequently, the LNR and central LNR were among the factors that exhibited statistically significant differences according to the size of mLN and the patient risk group.

The serum Tg level has been reported to be among the useful predictors of postoperative recurrence or survival. During the follow-up period, a sudden increase in the serum Tg level could be indicative of disease recurrence [8, 21]. Moreover, postoperative stimulation Tg value was used in the decision to apply RAI ablation therapy following surgery. The cutoff value was not clear, but in general, a value of approximately 1 ng/ml is considered to indicate stable disease status [7, 22]. Among our results, the mean sTg value was observed to gradually increase with the three stratified risk groups, and 1.24 ng/ml was the mean sTg value of the risk group I. (Fig. 2). Therefore, we suggest that RAI ablation can be omitted for only in the group I, which means a low-risk group.

This study had several limitations. First, the follow-up period was not sufficiently long to obtain information about the mortality rates of the patients. Second, we were not able to analyze the recurrence rates according to distant metastases and locoregional recurrence due to the small number of recurrence events in the short follow-up period. Additional studies with longer follow-up times are needed in the future. Third, more than 160 patients were entirely excluded from the present study because BRAF mutation tests were not performed for several months in 2014. This issue may have resulted in selection bias.

Nevertheless, we used well-organized prospective data collection that resulted in a wealth of information, and the analyses were performed on the data from a relatively large number of patients. Our study is the first to create a patient risk stratification based on the size of mLN and LNR. The results of this study suggest that this stratification will be valuable for informing the decision to apply RAI ablation to postoperative patients.