PTC is the most commonly diagnosed thyroid malignancy (29). Increased awareness of thyroid nodular disease, wide availability of US and FNA, and improved accuracy of histopathological examination of surgical samples have been suggested to be reasons for the increased incidence of detection (30). Fortunately, PTC is usually treatable and has a good prognosis if it is diagnosed early, although it is also accompanied by a high incidence of CLNMs (31). There is no consensus about the need to routinely perform pCND in patients with cN0 PTC. However, the optimal management for those patients is achieved by performing the most appropriate surgery at the time of diagnosis to achieve the best prognosis and minimize the risk to those patients and the need for unnecessary secondary procedures (32). Therefore, identifying RFs of CLNMs could guide surgeons in considering which cN0 PTC patients require pCND.
There are several studies in the literature to identify RFs that may predict CLNMs. Xue et al. (33) indicated that RFs for CLNMs including age < 45 years, male gender, tumor size of ≥ 1 cm, and ETE predicted CLNMs. Age is considered among the most important prognostic factors for TC (34,35). Yuan et al. (36) showed that the rate of CLNMs was higher in patients < 45 years than that ≥45 years. Multivariate analysis showed age < 45 years was independent predictor of CLNMs in patients with cN0 PTMC. Kim et al. also observed a trend to an inverse relationship between the incidence of CLNMs and age in patients with cN0 PTC (37).In a recent meta-analysis conducted by Hafez et al. (38) showed age <45 years, male sex, multifocality, bilaterality, capsular invasion, lymphovascular invasion and ETE are the factors significantly associated with CLNMs. Also as reported by Pontius et al. the presence of lymphovascular invasion among patients with PTC is associated with significantly decreased survival (39). In our study, univariate analysis shows that male gender, young age, the presence of ETE, a primary tumor size ≥10 mm , and lymphatic invasion were RFs of CLNMs. In multivariate logistic regression analysis we found; <41 years of age [OR=2.59 (95% CI=1.23-5.45); p=0.013], male gender status [OR=2.26 (95% CI=1.37-3.71); p=0.001] and presence of lymphatic invasion [4.09 (95% CI=2.46-6.80); p<0.001] were found to be independent factors affecting pathologic LN involvement. These results indicate that careful preoperative assessment of LN status must be followed in young and male patients.
In our study, the positive nodes were found in almost 59% of patients of PTC. Statistic analysis showed that the larger the tumor was, the more likely to find CLNMs, with the incidence of 51.5, 65.9, 63, and 66.7% in tumor size of; < 1, 1–2, 2–4 cm, and ≥ 4 cm, respectively. These results consist with most reports. Wang et al. (40) also showed a positive correlation between tumor size and the incidence of CLNMs, with the incidence of 35.1, 53.9, 47.1, and 100.0% in tumor size of; < 1, 1–2, 2–4 cm, and ≥4 cm, respectively. At the same time they found that, tumor size was also an independent predictive factor for CLNMs in cN0 PTC patients. Roh et al. (41) using multivariate analyses reported that tumor size > 1 cm was an independent factor for ipsilateral CLNMs. Park et al. (42) reported that tumor size > 0.7 cm was an independent variable predictive of CLNMs in patients with cN0 PTMC. However, other studies reported that there was no significant association between the presence of CLNMs and tumor size (43,44). In this study, multivariate analysis showed that tumor size ≥ 1 cm was not an independent predictor of CLNMs in patients with cN0 PTC. However, the incidence of CLNMs was much higher in tumors ≥1 cm (65.2% versus 44.8%). The difference in size cutoffs in each of those studies might affect the association between the incidence of CLNMs and tumor size. Further randomized controlled multicenter study will be helpful to draw the right conclusion.
Both TNM staging system (AJCC) and risk stratification system (ATA) have important prognostic values for TC. The former is used to predict disease-specific mortality, and the latter is mainly used to estimate the LRR. Several studies proposed that pCND did not help LRR and suggested RAI therapy to control the disease (45,46). This is in contrast to the findings of Moo et al. (47), who found a trend toward lower recurrence in TT+pCND, which could be due to increased control over local metastases or increased doses of RAI ablation in patients with recognized LNMs. Barczynski et al. (26) reported that pCND for more accurate N staging, followed by personalized adjuvant RAI treatment, and significantly improved both 10-year DSS improving from 92.5% to 98.0% and locoregional control, without increasing the risk of permanent morbidity in patients with PTC. Randolph et al. (48) showed that prognosis is negatively impacted if ≥5 LNMs are identified during neck dissections. In our study, CLNMs were found in 58.8% (127 of 216 cases) of patients with cN0 PTC. Because of the CLNMs detected after pCND, 46.7% of patients over 55 years of age were upstaged. Most of our patients had other RFs that also upgraded them to the ATA intermediate-risk group, however, in 18 patients (14.2%), the CLNMs data was the only factor that led to their stratification to the intermediate risk category. From patients with CLNMs 58 (26.8%) were recommended to receive the RAI therapy after initial surgery only for CLNMs, regardless of other poor prognostic factors. These patients would have been inappropriately down staged and possibly undertreated if pCND had been omitted. Similar to our results, Zhang et al.(8) reported that the percentage of upstaged patients with cN0 PTMC was 16% and Nylen et al. (49) reported that 23% of patients upgraded to intermediate-risk group. They also noted that 4% of these patients were upgraded only because of LN data. Recent studies have suggested that approximately one-third of patients who underwent pCND may have been upstaged.(48,50) The present studies results agree with this observation, as 58.8% of our cN0 PTC patients who underwent pCND had CLNMs , and 36.1% of pCND received RAI therapy for N1a ≥0.2 cm. Bonnet et al. reported data suggesting that pCND resulted in increased use of RAI and, ultimately, in more favorable outcomes (51).
Another important reason for pCND is that there is no other reliable method to verify CLNMs, whether preoperative or intraoperative US or other methods are far from sufficient sensitivity to verify CLNMs (52). Stulak et al. (53) reported that in a total of 511 patients with PTC, 476 patients were cN0; preoperative US detected CLNMs in 10 patients (2.1%), but CLNMs were confirmed in 179 patients (32.5%). Preoperative US for the detection of CLNMs has high specificity (92%) and positive predictive value (81%–92%) but low sensitivity (51%–61%) and negative predictive value, especially for central LNs (63%–76%) (54,55). Therefore, before other reliable authentication methods arise, there is no better way to assess CLNMs other than pCND. The detection of CLNMs is associated with the administration of higher doses of RAI for postoperative ablation, decreased recurrence in patients undergoing pCND, and the need for reoperation.
Studies have reported higher rates of HPT and RLN palsy with reoperative surgery because of tumor recurrence and local invasion (56,57), and LNM is strongly associated with recurrence after TT (58). In the literature, however, the rate of transient HPT has been reported to be between 9.7 and 56.5% after TT+pCND. A permanent HPT has been reported in up to 19.4% of patients after pCND, compared with 0.6–8.1% after simple TT (59). In this study, results consistent with the literature were obtained. According to the results of a recent meta-analysis published by Yang et al. (60), the TT+pCND group had a significantly higher rate of transient HPT than that of the TT alone group. This finding was consistent with that of Moo et al. (47), who reported that the number of patients with transient HPTs in the TT+pCND group was greater than that in the TT alone group at the beginning of the study. Some researchers have reached similar conclusions (8, 61-64). This study revealed that the rates of transient HPT were higher when a pCND was performed, but the difference was not statistically significant (30 vs 18.8; p = .185). Importantly, the rate of permanent HPT was not different between the 2 groups. We implemented the policy of liberal autografting of parathyroid glands where necessary, which may explain the lower rate of permanent HPT.
The other main complication related to pCND is RLN injury. Roh et al. (63) demonstrated an increased risk of transient RLN palsy after TT+pCND, which could be explained by increased RLN injury caused by dissection of the LNs. However, Wang et al. (65) reported that there were no significant differences in the incidence of other complications, such as transient HPT, permanent HPT, transient RLN palsy and permanent RLN palsy, between the TT and TT+pCND groups. Our data revealed that the rates of transient and permanent RLN palsy were 6.9% and 1.4%, respectively. These results are in accordance with those reported in the literature, which reported rates of transient RLN palsy ranging from 1% to 13% and rates of permanent RLN injury ranging from 0% to 3.6%. (66-69). The addition of CND to TT has been reported to increase neither the risk of transient nor permanent RLN palsy in prophylactic or therapeutic procedures (61,70-73). In this study, TT+pCND did not increase the incidence of transient or permanent RLN palsy, but the incidence of transient RLN palsy was lower in the TT+pCND group than in the TT alone group. This may be explained by the fact that we avoid performing pCND in patients with RLN injury or signal loss. On the basis of our data, we can conclude that pCND can be safely performed with comparable permanent morbidity by experienced surgeons.
The incidence of TC is increasing worldwide. In a large cohort study, an overall 23.3-fold increased risk of TC was observed following the diagnosis of any benign thyroid gland disorder (compared with no diagnosis) (74). For this reason, especially in our country, which is endemic for thyroid diseases, the treatment approach for PTC gains much more importance. Another striking finding in our study was the high rate of tall cell variants, one of the aggressive subtypes of PTC (31.8%). In the literature, the incidence of tall cell variant is reported to be between 1.3% and 13% (75,76), and as is well known, tall cell variant has a poorer prognosis with significantly higher rates of LN and distant metastases and lower 5-year DSS (77,78). Furthermore, Machens et al. reported that microscopic lymphatic invasion in PTC was associated with LNMs and multifocal tumor growth (79). In our study, the rate of lymphatic invasion was 74% in the whole group and 96% in the group of patients with CLNMs.
The first limitation of this study was that it was a retrospective study from a single center, and there might have been selection bias. The second limitation of this study was that we could not evaluate the effects of pCND on the cancer-specific survival and recurrence rates of patients with cN0 PTC because the follow-up time was relatively short.