One hundred and twenty-four HPV-negative OPSCC patients, including 17 females and 107 males (Table 1), were enrolled. Patient age ranged from 28 to 88 years old, and the average age was 60.7 ± 12.0 years (Table 1). Most OPSCCs were located in the base of the tongue (94, 75.9%) and soft palate (22, 17.7%) (Table 1). The distribution of OPSCCs for pathological T stage was T1 (15, 12.1%), T2 (36, 29.0%), T3 (46, 37.1%) and T4 (27, 21.7%) (Table 1). Regarding the histologic grade of tumors, 62 (50%) OPSCCs were categorized as well differentiated; 54 (43.5%) and 8 (6.5%) OPSCCs were moderately and poorly differentiated, respectively (Table 2). Fourteen (11.3%) OPSCCs had LVI, and 20 (16.1%) OPSCCs had PNI (Table 2). Sixty-one (49.2%) patients received adjuvant radiotherapy, and 20 (16.1%) patients received adjuvant chemoradiotherapy (Table 2). The choice of SND did not correlate with age (p = 0.823), sex (p = 0.296), tumor subsite (p = 1.000), T stage (p = 0.991), N stage (p = 0.750), radiotherapy (p = 0.590) or chemoradiotherapy (p = 0.625) (Table 1).
Table 2
Variables
|
n (%)
|
DFS
|
OS
|
DSS
|
5-y
(%)
|
P
|
5-y (%)
|
P
|
5-y
(%)
|
P
|
All
|
|
124
|
58.5
|
|
61.1
|
|
68.4
|
|
Mean age, year
|
≤ 60
|
60.7 ± 12.0†
|
59.3
|
0.766‡
|
61.9
|
0.940‡
|
63.8
|
0.286‡
|
> 60
|
53.3
|
58.2
|
79.1
|
Gender
|
Female
|
17 (13.7)
|
61.8
|
0.830‡
|
61.8
|
0.921‡
|
67.7
|
0.926‡
|
Male
|
107 (86.3)
|
57.5
|
61.4
|
70.6
|
Subsite
|
Base of tongue
|
94 (75.8)
|
57.6
|
0.807‡
|
63.4
|
0.829‡
|
68.7
|
0.756‡
|
Soft palate
|
22 (17.7)
|
36.4
|
40.4
|
80.8
|
Tonsil
|
6 (4.8)
|
66.7
|
66.7
|
66.7
|
Pharyngeal walls
|
2 (1.6)
|
100
|
100
|
100
|
T stage
|
T1
|
15 (12.1)
|
73.3
|
0.421‡
|
86.7
|
0.024‡
|
86.7
|
0.018‡
|
T2
|
36 (29.0)
|
55.2
|
62.4
|
72.8
|
T3
|
46 (37.1)
|
52.8
|
50.1
|
55.7
|
T4
|
27 (21.8)
|
63.3
|
76.7
|
85.2
|
N stage
|
N0
|
76 (61.3)
|
61.9
|
0.003‡
|
62.8
|
0.712‡
|
70.8
|
0.749‡
|
N1
|
28 (22.3)
|
55.9
|
49.0
|
54.8
|
N2
|
16 (12.9)
|
60.0
|
71.4
|
71.4
|
N3
|
4 (3.2)
|
25.0
|
50.0
|
50.0
|
SND
|
I-III
|
62 (50.0)
|
55.0
|
0.914‡
|
58.9
|
0.778‡
|
74.0
|
0.290‡
|
I-IV
|
62 (50.0)
|
60.1
|
61.5
|
64.8
|
DOI, mm
|
≤ 5
|
21 (16.9)
|
63.0
|
0.546‡
|
62.1
|
0.864‡
|
62.1
|
0.666‡
|
5 < DOI ≤ 10
|
30 (24.2)
|
40.6
|
51.7
|
64.6
|
> 10
|
73 (58.9)
|
62.3
|
65.8
|
72.1
|
Histologic grade
|
Well
|
62 (50.0)
|
56.8
|
0.871‡
|
62.0
|
0.531‡
|
69.2
|
0.768‡
|
Moderate
|
54 (43.5)
|
50.0
|
68.2
|
70.2
|
Poor
|
8 (6.5)
|
75.0
|
75.0
|
75.0
|
LVI
|
Yes
|
14 (11.3)
|
50.0
|
0.786‡
|
65.0
|
0.600‡
|
75.0
|
0.673‡
|
No
|
110 (88.7)
|
59.5
|
45.2
|
68.0
|
PNI
|
Yes
|
20 (16.1)
|
70.0
|
0.646‡
|
80.0
|
0.524‡
|
80.0
|
0.783‡
|
No
|
104 (83.9)
|
56.2
|
59.0
|
69.5
|
Radiotherapy
|
Yes
|
61 (49.2)
|
63.9
|
0.421‡
|
70.8
|
0.235‡
|
75.5
|
0.415‡
|
No
|
63 (50.8)
|
53.8
|
57.6
|
66.4
|
Chemoradiotherapy
|
Yes
|
20 (16.1)
|
44.0
|
0.071‡
|
26.8
|
0.019‡
|
26.8
|
0.005‡
|
No
|
104 (83.8)
|
61.4
|
70.8
|
79.9
|
†Mean ± SD. |
‡A log-rank test. |
Abbreviations: DFS, disease-free survival; OS, overall survival; DSS, disease-specific survival; P, p value. SND, selective neck dissection; DOI, depth of invasion; LVI, lymphovascular invasion; PNI, perineural invasion. |
Distribution Of Lymph Node Metastasis
For the entire cohort, 112 patients underwent unilateral neck dissection, and 12 patients (4 patients in Group A and 8 patients in Group B) underwent bilateral neck dissection (all contralateral neck dissections were level I-III). Forty-eight patients (38.7%) had occult lymph node metastasis, among which none was found in the contralateral neck (0/12). N1 accounted for 28 cases (22.6%) and N2 and N3 for 16 (12.9%) and 4 (3.2%, all cases of N3 were N3b) (Table 1). In Group A, occult lymph node metastasis rates of the ipsilateral neck at levels I, II and III were 8.1% (n = 5), 25.8% (n = 16) and 8.1% (n = 5), respectively (Fig. 2A). In Group B, metastasis rates of the ipsilateral neck in level I-IV were 4.8% (n = 3), 22.6% (n = 14), 11.3% (n = 7) and 3.2% (n = 2), and 2 cases of IV metastasis originated from the base of the tongue and the lateral pharyngeal wall (Fig. 2B). Regarding the distributed numbers of metastatic lymph nodes of the ipsilateral neck, there were 6 (15.8%), 27 (71.1%) and 5 (15.8%) lymph nodes at levels I, II and III in Group A (Fig. 2C) and 4 (11.1%), 23 (63.9%), 7 (19.4%) and 2 (5.6%) lymph nodes in Group B (Fig. 2D).
Group comparison of regional control, surgical complications and survival analysis
The median follow-up time was 46 months (range, 22–71). At the last follow-up, only 1 patient (1/62, 0.8%) had ipsilateral neck recurrence, which occurred at level II (Group A); it was confirmed as lymph node metastasis by pathology. Group B had no recurrence in the ipsilateral neck (0/62, 0%). No recurrence in the contralateral neck was found, though 2 patients in Group B experienced contralateral metastasis at level II after ipsilateral neck dissection.
We recorded 4 kinds of surgical complications associated with neck dissection of level IV (Table 3). Two patients in the entire cohort (2/124, 1.6%) had complications of chylous leakage, all in Group B (2/124, 1.6%, p = 0.496) (Table 3). There was 1 patient in both Group A (1/64, 1.6%) and Group B (1/64, 1.6%, p = 1.000) who experienced phrenic nerve paralysis (Table 3). Compared with Group A (5/62, 8.1%), patients in Group B (7/62, 11.3%, p = 0.544) had a slightly higher incidence of hematoma (Table 3). Regarding infection, there were no differences between Group A (8/62, 12.9%) and Group B (8/62, 12.9%, p = 1.000) (Table 3).
Table 3: Morbidity of surgical complications after neck dissection in cohort
Variable
|
In total
|
SND with I-III
(n = 62)
|
SND with I-IV
(n = 62)
|
P
|
n (%)
|
Chylous leakage
|
2 (1.6)
|
0
|
2 (3.2)
|
0.496†
|
Phrenic nerve paralysis
|
2 (1.6)
|
1 (1.6)
|
1(1.6)
|
1.000†
|
Hematoma
|
12 (9.7)
|
5 (8.1)
|
7 (11.3)
|
0.544‡
|
Wound infection
|
16 (12.9)
|
8 (12.9)
|
8 (12.9)
|
1.000‡
|
†Fisher’s exact test.
‡Person’s chi-squared test.
Abbreviations: SND, selective neck dissection; P, p value.
For the entire cohort, the 5-year DFS, OS and DSS rates were 58.5%, 61.1% and 68.4%, respectively (Table 2). The 5-year DFS rate was 55.0% in Group A and 60.1% in Group B (p = 0.914) (Table 2 & Fig. 3A). The five-year OS rate of Group A was 58.9%, similar to that of Group B (61.5%, p = 0.778) (Table 2 & Fig. 3B). The five-year DSS rate was 74.0% in Group A and 64.8% in Group B (p = 0.290) (Table 2 & Fig. 3C).
Univariate Analysis
In univariate survival analysis, there was no significant difference between Group A and Group B (SND) with regard to 5-year DFS, OS and DSS (Table 2 & Fig. 3). N stage was the only significant factor associated with 5-year DFS (p = 0.003), and patients with N0-stage disease exhibited the highest 5-year DFS rate, at 61.9% (Table 2). Five-year OS correlated with T stage (p = 0.024) and adjuvant chemoradiotherapy (p = 0.019) (Table 2). Patients with T1 stage disease had the highest 5-year OS rate of 86.7%, whereas these patients did not benefit from adjuvant chemoradiotherapy (26.8%) (Table 2). The five-year DSS rate was related to T stage (p = 0.018) and chemoradiotherapy (p = 0.005) (Table 2). Similar to OS, T1 stage had the highest 5-year DSS rate, at 86.7%, and patients who underwent adjuvant chemoradiotherapy had a 5-year DSS rate of only 26.8% (Table 2). Adjuvant radiotherapy led to distinct improvements in DFS (63.9% vs. 53.8%, p = 0.421), OS (70.8% vs. 57.6%, p = 0.235) and DSS (75.5% vs. 66.4%, p = 0.415), none of which showed statistical significance (Table 2).
Multivariate Survival Analysis
SND and other covariates, which were screened from univariate analysis, were included in Cox proportional hazard models for multivariate analysis of 5-year DFS, OS and DSS. The results showed that SND was not an independent prognosticator for 5-year DFS (HR 1.02, p = 0.949), OS (HR 0.88, p = 0.731) or DSS (HR 1.42, p = 0.354) (Table 4).
N3 stage (HR 3.89, p = 0.030) and chemoradiotherapy (HR 2.07, p = 0.041) were both related to worse 5-year DFS (Table 4). Chemoradiotherapy, an independent prognosticator (HR 2.41, p = 0.731) for 5-year OS, correlated with a worse survival outcome. Conversely, T stage (p = 0.155) and radiotherapy (HR 0.46, p = 0.055) did not correlate with OS (Table 4). In multivariate analysis of 5-year DSS, no independent prognosticator was found (Table 4).
Table 4: Multivariate survival analysis
DFS
|
B
|
P
|
HR
|
95% CI HR
|
|
|
|
|
Inf
|
Sup
|
SND (I-IV)
|
0.019
|
0.949
|
1.019
|
0.564
|
1.841
|
N stage
|
|
0.185
|
|
|
|
N1
|
0.200
|
0.583
|
1.221
|
0.598
|
2.495
|
N2
|
0.302
|
0.510
|
1.353
|
0.551
|
3.332
|
N3
|
1.357
|
0.030
|
3.883
|
1.140
|
13.224
|
Chemoradiotherapy
|
0.727
|
0.041
|
2.068
|
1.029
|
4.155
|
OS
|
B
|
P
|
HR
|
95% CI HR
|
|
|
|
|
Inf
|
Sup
|
SND (I-IV)
|
-0.131
|
0.731
|
0.878
|
0.418
|
1.845
|
T stage
|
|
0.115
|
|
|
|
T2
|
0.061
|
0.940
|
1.063
|
0.214
|
5.283
|
T3
|
0.957
|
0.223
|
2.605
|
0.559
|
12.142
|
T4
|
0.174
|
0.843
|
1.190
|
0.214
|
6.622
|
Radiotherapy
|
-0.785
|
0.055
|
0.456
|
0.205
|
1.015
|
Chemoradiotherapy
|
0.880
|
0.042
|
2.411
|
1.031
|
5.636
|
DSS
|
B
|
P
|
HR
|
95% CI HR
|
|
|
|
|
Inf
|
Sup
|
SND (I-IV)
|
0.353
|
0.354
|
1.424
|
0.674
|
3.007
|
T stage
|
|
0.278
|
|
|
|
T1
|
0.172
|
0.833
|
1.187
|
0.242
|
5.831
|
T2
|
0.847
|
0.280
|
2.332
|
0.502
|
10.823
|
T3
|
0.075
|
0.930
|
1.078
|
0.200
|
5.819
|
Chemoradiotherapy
|
0.700
|
0.079
|
2.013
|
0.922
|
4.395
|
Cox proportional hazard model was used for multivariate survival analysis.
Abbreviations: DFS, disease-free survival; SND, selective neck dissection; OS, overall survival; DSS, disease-specific survival; B, β coefficient; P, p value; HR, hazard ratio; 95% CI, 95% Confidence Interval.