**3. Results**

We included in this study 399 patients with clinically uninvolved lymph nodes (cN0) and pathological diagnosis of DTC (Table 1). There were 101 (25.3%) males and 298 (74.7%) females with a mean age of 50.5 years. Autoimmune thyroiditis was present in 146 (36.6%) cases, and hyperthyroidism in 28 (7%). A familiar history of thyroid carcinoma was present in seven (1.8%) patients, none of these were in the context of Multiple Endocrine Neoplasia Syndrome. In 148 (37.1%) patients, FNAC demonstrated atypia of undetermined significance (including Tir3, Tir3a, and Tir3b), and in 178 (44.6%) FNAC indicated a suspicious malignancy (Tir4) or a malignant nodule (Tir5). In the remaining 73 (18.3%) patients, despite the fact that FNAC was not suspicious or indicative for malignancy, a thyroidectomy was planned based on family history of the patient, physical examination, or US features of the nodule.

**Table 1.** Univariate analysis of demographic, preoperative, operative data, and outcomes of patients with differentiated thyroid carcinoma and clinically uninvolved lymph nodes.


TT: total thyroidectomy; pCLND: prophylactic central compartment lymph node dissection; US: ultrasound; FNAC: fine-needle aspiration cytology; RLN: recurrent laryngeal nerve; RAI: radioactive iodine. Continuous variables are reported as the mean ± standard deviation of the mean.

The surgical procedure consisted of total thyroidectomy alone in 320 (80.2%) patients, and total thyroidectomy and prophylactic CLND in 79 (19.8%).

Patients were divided into two groups based on the surgical approach. Those who underwent prophylactic CLND were significantly younger (42.6 y.o. vs. 52.4 y.o.; *p* < 0.001) and more frequently affected with autoimmune thyroiditis (62% vs. 30.3%; *p* < 0.001). Furthermore, FNAC was significantly more frequently diagnostic for DTC in the pCLND group (69.6% vs. 38.4%; *p* < 0.001).

The mean operative time was significantly longer in the pCLND group (102.3 min vs. 92.5 min; *p* < 0.001) as well as postoperative stay (3.0 days vs. 2.8 days). The incidence of transient hypoparathyroidism (43% vs. 31.9%), permanent hypoparathyroidism (15.2% vs. 8.4%), transient RLN injury (3.8% vs. 2.2%), and permanent RLN injury (1.3% vs. 0.6%) was higher in the pCLND group, but these differences were not statistically significant. The mean follow-up time was 55.4 months.

Full histopathologic findings are reported in Table 2. Nodule size and thyroid weight were similar between the two groups. Conversely, the histotype was significantly different between the two groups (*p* < 0.001). In the pCLND group, the incidence of the tall cell variant of PTC (TCV-PTC) was nearly quadruple that in the other group (25.3% vs. 6.3%), whereas the incidence of follicular variant of PTC (FV-PTC) was less than half (16.5% vs. 37.5%) when compared to the TT group. The presence of aggressive features of the tumor including multicentricity, angioinvasivity, and extrathyroidal extension was similar between the two groups.


**Table 2.** Univariate analysis of pathological data of patients with differentiated thyroid carcinoma and clinically uninvolved lymph nodes.

TT: total thyroidectomy; pCLND: prophylactic central compartment lymph node dissection; PTC: papillary thyroid carcinoma; FV-PTC: follicular variant of PTC; TCV-PTC: tall cell variant of PTC; FTC: follicular carcinoma; HCC: hurtle cell carcinoma; LN: lymph node; ATA: American Thyroid Association. Continuous variables are reported as the mean ± standard deviation of the mean.

As defined in the Methods section, all patients in the pCLND group underwent lymphectomy of the level VI and level VII lymph nodes, whereas in the other group, an excision of some perithyroidal lymph nodes was performed in 99 (30.9%) patients.

Lymph node yield (8.9 vs. 2.1; *p* < 0.001) and lymph node metastases (25.3% vs. 4.7%; *p* < 0.001) were significantly higher in the pCLND group. In contrast, the lymph node ratio was lower in the pCLND group (0.3 vs. 0.6; *p* < 0.001).

Patients were classified according to the ATA stratification for risk of structural disease recurrence. Patients in the pCLND group were ranked more frequently in the intermediate class of risk (43% vs. 9.4%; *p* < 0.001) than those in the other group, whereas low and high risk classes were similar between the two groups.

Radioactive iodine (RAI) therapy was administrated significantly more often in the pCLND group (92.4% vs. 80.6%; *p* = 0.019).

Overall, 22 (5.5%) patients experienced recurrent disease; 16 of these were localized in the central compartment, whereas the others were in the lateral neck compartment. The crude incidence of disease recurrence was similar between the two groups: 5.6% in The TT group and 5.1% in the pCLND group (*p* = 0.936). Log-rank test on Kaplan–Meier curves, reported in Figure 2a, did not show any significant difference between the two groups (*p* = 0.0883; HR 0.9267, 95% CI 0.3203–2.6814).

**Figure 2.** Kaplan–Meier curves estimating disease-free survival according to surgical procedure, (**a**) including all the patients regardless of ATA class risk of recurrence: *p* = 0.883, HR 0.9267 (95% CI 0.3203–2.6814); (**b**) including only patients with moderate and high risk for disease recurrence according to ATA guidelines: *p* = 0.0493, HR 0.3299 (95% CI 0.1092–0.9967). pCLND: prophylactic central lymph node dissection; TT: total thyroidectomy; HR: hazard risk.

A subset analysis was performed considering the ATA stratification for risk of disease recurrence, as reported in Table 3. Considering the patients in the intermediate and high class of risk, the incidence of recurrent disease was significantly lower in the pCLND group (5.4% vs. 21.2%; *p* = 0.0392) than in the TT group. On the other hand, no significant differences were observed in the low class of risk. Log-rank test on Kaplan–Meier curves for patients in intermediate and high class of risk, reported in Figure 2b, demonstrated a significant difference between the two curves (*p* = 0.0439; HR 0.3299, 95% CI 0.1092–0.9967).


**Table 3.** Univariate analysis of recurrent disease in patients stratified for risk of structural recurrence according to the American Thyroid Association guidelines.

ATA: American Thyroid Association guidelines; TT: total thyroidectomy; pCLND: prophylactic central compartment lymph node dissection.

Overall, occult lymph node metastases in our series were found in 35 (8.8%) patients. As reported in Table 4, the incidence of disease recurrence was 20% in patients with lymph node metastases (pN+), 2.8% in patients with uninvolved lymph nodes in which an evaluation of N status was possible because at least one lymph node was excised (pN0), and 5% in patients in which N status was not assessed (pNx) (*p* < 0.001). The log-rank test on Kaplan–Meier curves representing patients with and without lymph node metastases (Figure 3) demonstrated a significant difference between the two groups (*p* < 0.001; HR 15.160, 95% CI 3.444–566.7289).


**Table 4.** Univariate analysis of recurrent disease in patients stratified according to N status.

pNx: pathological nodal status not assessed; pN0: lymph nodes uninvolved at pathological examination; pN+: lymph node metastases at pathological examination. Statistical significance: overall *p* < 0.001; pNx vs. pN0: *p* = 0.1132; pNx vs. pN+: *p* < 0.001; pN0 vs. pN+ *p* < 0.001.

**Figure 3.** Kaplan-Meier curves estimating overall disease-free survival according to N status: *p* < 0.001 (HR 15.160, 95% CI 3.4445–66.7289). pN0: No evidence of lymph node metastases at pathological examination. pN+: Lymph node metastases at pathological examination.
