Table 1 describes the characteristics of all our patients. Eighteen patients (25.0%) exhibited confirmed recurrence during the follow-up period. The sites of recurrence included: 6 (33.3%) patients on the central neck, 10 (55.6%) on the lateral neck, 7 (38.9%) with distant recurrence, and 5 (27.8%) with recurrence at multiple sites.
Table 1
Clinical characteristics of 72 patients with RLN invasion.
Characteristics
|
|
Overall
|
Recurrence
|
No recurrence
|
p-value
|
* p < 0.05
|
n.
|
|
72
|
18
|
54
|
|
|
f/u time month (median [IQR])
|
|
58.00 [30.25 ~ 83.00]
|
38.50 [31.00, 59.00]
|
64.50 [28.75, 93.00]
|
0.091
|
|
Sex (%)
|
F
|
56 (77.8)
|
11 (19.64)
|
45 (80.36)
|
0.098
|
|
|
M
|
16 (22.2)
|
7 (43.75)
|
9 (56.25)
|
|
|
Age (mean (SD))
|
|
57.59 (14.17)
|
57.52 (14.47)
|
57.61 (14.20)
|
0.982
|
|
Height (mean (SD))
|
|
159.48 (7.86)
|
161.75 (8.16)
|
158.71 (7.68)
|
0.176
|
|
Body weight (mean (SD))
|
|
61.63 (11.89)
|
65.43 (15.82)
|
60.34 (10.08)
|
0.213
|
|
BMI (mean (SD))
|
|
24.12 (3.38)
|
24.81 (4.07)
|
23.89 (3.12)
|
0.390
|
|
Operative method (%)
|
HT
|
1 (1.4)
|
0 (0.0)
|
1 (100)
|
0.836
|
|
|
TT /c CND
|
43 (59.7)
|
10 (23.26)
|
33 (76.74)
|
|
|
|
TT /c CND /c LND
|
28 (38.9)
|
8 (28.57)
|
20 (71.43)
|
|
|
RAI dose (%)
|
< 150
|
17 (26.2)
|
4 (23.53)
|
13 (76.47)
|
1.000
|
|
|
≥ 150
|
48 (73.8)
|
13 (27.08)
|
35 (72.92)
|
|
|
Tg level post ablation (%)
|
≤ 1
|
22 (34.4)
|
3 (13.64)
|
19 (86.36)
|
0.137
|
|
|
> 1
|
42 (65.6)
|
14 (33.33)
|
28 (66.67)
|
|
|
Pre op. VCP (%)
|
yes
|
37 (51.4)
|
10 (27.03)
|
27 (72.97)
|
0.788
|
|
|
no
|
35 (48.6)
|
8 (22.86)
|
27 (77.14)
|
|
|
RLN direction (%)
|
Rt.
|
32 (44.4)
|
8 (25.00)
|
24 (75.00)
|
1.000
|
|
|
Lt.
|
39 (54.1)
|
10 (25.64)
|
29 (74.36)
|
|
|
Tumor size (%)
|
≤ 20 mm
|
39 (54.2)
|
7 (17.95)
|
32 (82.05)
|
0.175
|
|
|
> 20 mm
|
33 (45.8)
|
11 (33.33)
|
22 (66.67)
|
|
|
Resection margin (%)
|
clear
|
39 (54.2)
|
5 (12.82)
|
34 (87.18)
|
0.014
|
*
|
|
involved
|
33 (45.8)
|
13 (39.39)
|
20 (60.61)
|
|
|
Intra op. RLN resection (%)
|
yes
|
47 (65.3)
|
15 (31.91)
|
32 (68.09)
|
0.088
|
|
|
no
|
25 (34.7)
|
3 (12.00)
|
22 (88.00)
|
|
|
RLN anastomosis (%)
|
yes
|
7 (9.7)
|
1 (14.29)
|
6 (85.71)
|
0.672
|
|
|
no
|
65 (90.3)
|
17 (26.15)
|
48 (73.85)
|
|
|
Multifocality (%)
|
yes
|
25 (34.7)
|
6 (24.00)
|
19 (76.00)
|
1.000
|
|
|
no
|
47 (65.3)
|
12 (25.53)
|
35 (74.47)
|
|
|
Trachea invasion (%)
|
yes
|
13 (18.1)
|
4 (30.77)
|
9 (69.23)
|
0.725
|
|
|
no
|
59 (81.9)
|
14 (23.73)
|
45 (76.27)
|
|
|
Esophagus invasion (%)
|
yes
|
15 (20.8)
|
5 (33.33)
|
10 (66.67)
|
0.504
|
|
|
no
|
57 (79.2)
|
13 (22.81)
|
44 (77.19)
|
|
|
Central meta (%)
|
yes
|
44 (61.1)
|
10 (22.73)
|
34 (77.27)
|
0.589
|
|
|
no
|
28 (38.9)
|
8 (28.57)
|
20 (71.43)
|
|
|
N. of central meta (mean (SD))
|
|
2.10 (2.66)
|
3.11 (4.17)
|
1.76 (1.85)
|
0.198
|
|
Lateral meta (%)
|
yes
|
28 (38.9)
|
8 (28.57)
|
20 (71.43)
|
0.589
|
|
|
no
|
44 (61.1)
|
10 (22.73)
|
34 (77.27)
|
|
|
†RLN : Recurrent laryngeal nerve, IQR : Inter quartile range, SD : Standard deviation, RAI : Radioactive iodine, VCP : Vocal cord palsy |
The median follow-up time was 58 months, females were predominant, and the median patient age was 57.59 years. We also collected height, weight, and body mass index as basic information.
One patient chose to undergo hemithyroidectomy despite having T4a PTC because they wished to preserve their remaining thyroid lobes. We considered the postoperative radioactive iodine (RAI) dose as a categorical parameter, wherein 48 patients (73.8%) underwent RAI therapy with over 150 mCi and 24 patients underwent RAI therapy with less than 150 mCi.
A previous study on locally invasive thyroid cancer found that post-ablation thyroglobulin (Tg) levels that were higher than 1 contributed to a significantly increased recurrence rate8. In this study, the Tg levels of post-ablation varied between 1, less than 1, or more than 1. However, the correct post-ablation Tg values could not be determined in eight patients (11.1%), as RAI therapy was performed at other hospitals. Among the remaining patients, the post-ablation Tg levels were less than 1 in 34.4% of the cases and exceeded 1 in 73.8% of the cases. There was no significant association between post-ablation Tg level and PTC recurrence (p > 0.05).
A total of 37 patients (51.4%) had VCP before surgery and 35 patients (48.6%) did not. The RLN invasion was similar on both sides of the body, wherein the right RLN was involved in 32 patients (44.4%) and the left side was involved in 39 patients (54.1%); one patient had bilateral invasions. The median tumour size 23.28 mm, and the category value was divided into 2 cm or less and 2 cm or more. Twenty-five patients (34.7%) underwent shaving after we intraoperatively confirmed RLN invasion, and the remaining patients underwent nerve resection. Of these, we immediately performed nerve anastomoses in 7 patients. In addition to RLN involvement, 13 (18.1%) and 15 patients (20.8%) showed tracheal and esophageal invasion, respectively, while 44 (61.1%) and 28 (38.9%) patients showed central and lateral neck metastasis, respectively. The presence of central and lateral lymph node (LN) metastasis was not associated with PTC recurrence. The number of central LN metastases was not significantly different between patients with and without PTC recurrence.
After the final pathological evaluation, 39 patients (54.2%) showed clear resection margins and 33 (45.8%) showed positive/involved resection margins. Of the 39 patients with clear margins, 5 (12.82%) showed a recurrence and the remaining 34 (87.18%) did not. Of the 33 patients with involved margins, 13 (39.39%) relapsed and 20 (60.61%) showed no recurrence––this indicated a relatively high recurrence rate in the involved-margin group (p = 0.014). In a simple analysis of these parameters, which did not account for survival time, only the positive resection margins were statistically significant.
Table 2 compares the preoperative VCP group and the intraoperative RLN-invasion group. There was no significant difference in the recurrence rates between the two groups (p > 0.05). There was also no difference between the two groups in terms of tumour size, resection margin, RLN anastomosis, ETE, multifocality, and tracheal or esophageal invasion (p > 0.05). Likewise, there was no difference between the two groups regarding the presence of central and lateral LN metastasis, and the number of lateral LN metastases. However, the mean number of central LN metastases was significantly higher in the preoperative VCP group (2.76) compared to the no-VCP group (1.40; p = 0.028).
Table 2
Comparison of the pre-operative VCP group with non pre-operative VCP group.
Variable
|
|
Overall
|
Pre-operative VCP
|
Non pre-operative VCP
|
p-value
|
* p < 0.05
|
n.
|
|
72
|
37
|
35
|
|
|
Recur (%)
|
yes
|
18 (25.0)
|
10 (27.0)
|
8 (22.9)
|
0.788
|
|
|
no
|
54 (75.0)
|
27 (73.0)
|
27 (77.1)
|
|
|
Tumor size (%)
|
≤ 20 mm
|
39 (54.2)
|
19 (51.4)
|
20 (57.1)
|
0.644
|
|
|
> 20 mm
|
33 (45.8)
|
18 (48.6)
|
15 (42.9)
|
|
|
Resection margin (%)
|
clear
|
39 (54.2)
|
17 (45.9)
|
22 (62.9)
|
0.165
|
|
|
involved
|
33 (45.8)
|
20 (54.1)
|
13 (37.1)
|
|
|
Intra op. RLN resection (%)
|
yes
|
47 (65.3)
|
31 (83.8)
|
16 (45.7)
|
0.001
|
*
|
|
no
|
25 (34.7)
|
6 (16.2)
|
19 (54.3)
|
|
|
RLN anastomosis (%)
|
yes
|
7 (9.7)
|
2 (5.4)
|
5 (14.3)
|
0.254
|
|
|
no
|
65 (90.3)
|
35 (94.6)
|
30 (85.7)
|
|
|
ETE (%)
|
yes
|
66 (91.7)
|
35 (94.6)
|
31 (88.6)
|
0.423
|
|
|
no
|
6 (8.3)
|
2 (5.4)
|
4 (11.4)
|
|
|
Multifocality (%)
|
yes
|
25 (34.7)
|
15 (40.5)
|
10 (28.6)
|
0.329
|
|
|
no
|
47 (65.3)
|
22 (59.5)
|
25 (71.4)
|
|
|
Trachea invasion (%)
|
yes
|
13 (18.1)
|
9 (24.3)
|
4 (11.4)
|
0.222
|
|
|
no
|
59 (81.9)
|
28 (75.7)
|
31 (88.6)
|
|
|
Esophagus invasion (%)
|
yes
|
15 (20.8)
|
9 (24.3)
|
6 (17.1)
|
0.566
|
|
|
no
|
57 (79.2)
|
28 (75.7)
|
29 (82.9)
|
|
|
Central meta (%)
|
yes
|
44 (61.1)
|
26 (70.3)
|
18 (51.4)
|
0.147
|
|
|
no
|
28 (38.9)
|
11 (29.7)
|
17 (48.6)
|
|
|
N. of central meta (mean (SD))
|
|
2.10 (2.66)
|
2.76 (3.04)
|
1.40 (2.00)
|
0.028
|
*
|
Lateral meta (%)
|
yes
|
28 (38.9)
|
18 (48.6)
|
10 (28.6)
|
0.095
|
|
|
no
|
44 (61.1)
|
19 (51.4)
|
25 (71.4)
|
|
|
N. of lateral meta (mean (SD))
|
|
2.03 (4.04)
|
2.59 (4.21)
|
1.43 (3.82)
|
0.222
|
|
† VCP : Vocal cord palsy, RLN : Recurrent laryngeal nerve, ETE : Extrathyroidal extension, SD : Standard deviation |
A total of 47 patients (65.3%) underwent RLN resection and 25 did not (34.7%). In the preoperative VCP group, a majority of patients (31; 83.8%) underwent RLN resection. However, in the intraoperatively detected RLN-invasion group, the RLN was sacrificed in 16 patients (45.7%) to preserve it as far as possible. The rate of resection without preserving the RLN was relatively high in the preoperative VCP group; this result was statistically significant (p = 0.001).
In the univariate analysis, we found two parameters––resection margin and the number of central LN metastases––to be significant risk factors for recurrence-free survival (RFS) (p < 0.05; Table 3). Sex, age, height, body weight, body mass index, RAI dose, post-ablation Tg level, preoperative VCP, RLN direction, tumour size, RLN resection, RLN anastomosis, ETE, multifocality, tracheal and esophageal invasion, and lateral LN metastasis were not significantly associated with recurrence as per our statistical analysis (p > 0.05). In the multivariate analysis that included five parameters with p-values of less than 0.2, we only found one independent variable––resection margin involvement––as a risk factor for RFS (hazard ratio = 3.331 [confidence interval 1.017–10.915]; p = 0.047).
Table 3
Factors predictive of recurrence by univariate and multivariate analysis.
Variable
|
Univariate analysis
|
Multiivariate analysis
|
|
HR [95% CI]
|
p-value
|
* p < 0.05
|
HR [95% CI]
|
p-value
|
* p < 0.05
|
Sex = M (%)
|
1.784 [0.683, 4.661]
|
0.237
|
|
|
|
|
Age (mean (SD))
|
0.999 [0.965, 1.035]
|
0.957
|
|
|
|
|
Height (mean (SD))
|
1.022 [0.965, 1.082]
|
0.465
|
|
|
|
|
Body weight (mean (SD))
|
1.024 [0.986, 1.064]
|
0.222
|
|
|
|
|
BMI (mean (SD))
|
1.070 [0.922, 1.241]
|
0.372
|
|
|
|
|
RAI dose ≥ 150 (%)
|
1.084 [0.353, 3.329]
|
0.888
|
|
|
|
|
Tg level post ablation > 1 (%)
|
2.759 [0.791, 9.627]
|
0.112
|
|
1.838 [0.494, 6.839]
|
0.364
|
|
Pre op. VCP = no (%)
|
0.888 [0.350, 2.253]
|
0.803
|
|
|
|
|
RLN direction = Lt. (%)
|
1.221 [0.481, 3.098]
|
0.674
|
|
|
|
|
Tumor size > 20 (%)
|
1.626 [0.626, 4.225]
|
0.318
|
|
|
|
|
Resection margin positive (%)
|
2.982 [1.056, 8.418]
|
0.039
|
*
|
3.331 [1.017, 10.915]
|
0.047
|
*
|
Intra op. RLN resection = no (%)
|
0.395 [0.114, 1.367]
|
0.143
|
|
0.428 [0.091, 2.001]
|
0.281
|
|
RLN anastomosis = no (%)
|
2.164 [0.288, 16.279]
|
0.453
|
|
|
|
|
ETE = no (%)
|
0.925 [0.122, 6.982]
|
0.939
|
|
|
|
|
Multifocality = no (%)
|
1.073 [0.402, 2.861]
|
0.888
|
|
|
|
|
Trachea invasion = no (%)
|
0.786 [0.258, 2.391]
|
0.671
|
|
|
|
|
Esophagus invasion = no (%)
|
0.596 [0.212, 1.676]
|
0.327
|
|
|
|
|
Central node meta = no (%)
|
1.050 [0.411, 2.679]
|
0.919
|
|
|
|
|
N. of central meta (mean (SD))
|
1.196 [1.049, 1.364]
|
0.007
|
*
|
1.120 [0.992, 1.264]
|
0.068
|
|
Lateral node meta = no (%)
|
0.704 [0.277, 1.787]
|
0.460
|
|
|
|
|
N. of lateral meta (mean (SD))
|
0.859 [0.701, 1.052]
|
0.142
|
|
0.801 [0.636, 1.008]
|
0.059
|
|
† HR : Hazard ratio, SD : Standard deviation, RAI : Radioactive iodine, Tg : Thyroglobulin, VCP : Vocal cord palsy, RLN : Recurrent laryngeal nerve, ETE : Extrathyroidal extension |
Figure 1 shows a Kaplan-Meier survival curve depicting the RFS of patients with involved resection margins. Of the 18 patients who showed recurrence during follow-up, 5 and 13 belonged to the clear- and involved-margin groups, respectively. As compared with the clear-margin group, the involved-margin group showed a significantly lower rate of RFS (82% vs. 43%; p = 0.03; Fig. 1).
In this study, 2 patients died because of recurrence accompanied by anaplastic change during follow-up. Since the mortality rate was so low, it was difficult to evaluate the factors associated with survival.