Patient characteristics
The final matched pair analysis included 102 patients (51 patients in the submental island flap group and 51 patients in the radial forearm free flap group) with a median follow-up time of 61 months (range, 5-147 months). Patient characteristics before and after matching are listed in Table 2. There were significant differences in the N stage, the number of patients with pathologically positive level I lymph nodes, type of neck dissection, and operative time between the two groups before matching. However, only operative time was significantly different after matching (197 minutes in the submental island flap group and 450 minutes in the radial forearm free flap group, p < 0.001).
Table 2
Patient baseline characteristics before and after matching
Variables
|
Before matching
n = 133
|
After matching
n = 102
|
SMFa
n = 52 (%)
|
RFFb
n = 81 (%)
|
p-value
|
SMFa
n = 51 (%)
|
RFFb
n = 51 (%)
|
p-value
|
Age (mean ± SD (range); years)
|
59.6 ± 11.1
(33–84)
|
56.9 ± 10.9
(32–80)
|
0.186
|
59.2 ± 10.9
(33–84)
|
58.6 ± 10.7
(32–80)
|
0.791
|
Sex
- Male
- Female
|
32 (61.5)
20 (38.5)
|
52 (64.2)
29 (35.8)
|
0.854
|
32 (62.7)
19 (37.3)
|
30 (58.8)
21 (41.2)
|
0.839
|
Smoke
- Yes
- No
|
34 (65.4)
18 (34.6)
|
52 (64.2)
29 (35.8)
|
1.000
|
33 (64.7)
18 (35.3)
|
34 (66.7)
17 (33.3)
|
1.000
|
Alcohol
- Yes
- No
|
33 (63.5)
19 (36.5)
|
53 (65.4)
28 (34.6)
|
0.854
|
34 (65.4)
18 (34.6)
|
34 (65.4)
18 (34.6)
|
1.000
|
DM
- Yes
- No
|
7 (13.5)
45 (86.5)
|
9 (11.1)
72 (88.9)
|
0.786
|
7 (13.7)
44 (86.3)
|
7 (13.7)
44 (86.3)
|
1.000
|
COPD
- Yes
- No
|
14 (26.9)
38 (73.1)
|
19 (23.5)
62 (76.5)
|
0.684
|
14 (27.5)
37 (72.5)
|
11 (21.6)
40 (78.4)
|
0.646
|
ECOG PSb
− 0 or 1
− 2
|
41 (78.8)
11 (21.2)
|
69 (85.2)
12 (14.8)
|
0.358
|
40 (78.4)
11 (21.6)
|
42 (82.4)
9 (17.6)
|
0.804
|
Primary site location
- Tongue
- Floor of the mouth
- Buccal mucosa
- Alveolar ridge
- Retromolar trigone
|
29 (55.8)
14 (26.9)
5 (9.6)
3 (5.8)
1 (1.9)
|
48 (59.3)
12 (14.8)
12 (14.8)
6 (7.4)
3 (3.7)
|
0.513
|
29 (56.9)
14 (27.5)
4 (7.8)
3 (5.9)
1 (2.0)
|
27 (52.9)
9 (17.6)
10 (19.6)
4 (7.8)
1 (2.0)
|
0.394
|
T stage
- T1
- T2
- T3
- T4
|
1 (1.9)
13 (25)
20 (38.5)
18 (34.6)
|
-
15 (18.5)
31 (38.3)
35 (42.3)
|
0.422
|
1 (2.0)
13 (25.5)
20 (39.2)
17 (33.3)
|
-
9 (17.6)
21 (41.2)
21 (41.2)
|
0.582
|
N stage
- N0
- N1
- N2
- N3
|
22 (42.3)
12 (23.1)
16 (30.8)
2 (3.8)
|
18 (22.2)
12 (14.8)
41 (50.6)
10 (12.3)
|
0.011f
|
21 (41.2)
12 (23.5)
16 (31.4)
2 (3.9)
|
15 (29.4)
8 (15.7)
22 (43.1)
6 (11.8)
|
0.206
|
Table 2
Patient baseline characteristics before and after matching (continue)
Variables
|
Before matching
n = 133
|
After matching
n = 102
|
SMFa
n = 52 (%)
|
RFFb
n = 81 (%)
|
p-value
|
SMFa
n = 51 (%)
|
RFFb
n = 51 (%)
|
p-value
|
Clinically + Level I lymph node
|
18 (34.6)
|
41 (50.6)
|
0.077
|
18 (35.3)
|
21 (41.2)
|
0.684
|
Pathologically + Level I lymph node
|
13 (25)
|
34 (41.9)
|
0.048f
|
15 (29.4)
|
20 (39.2)
|
0.404
|
Pathological ENE
|
4 (7.7)
|
11 (13.6)
|
0.295
|
3 (5.9)
|
6 (11.8)
|
0.487
|
Neck dissection
Side
- Unilateral
- Bilateral
Type
- Selective
- Comprehensive
|
18 (34.6)
34 (65.4)
22 (42.3)
30 (57.7)
|
23 (28.4)
58 (71.6)
18 (22.4)
63 (77.8)
|
0.564
0.020f
|
17 (33.3)
34 (66.7)
21 (41.2)
30 (58.8)
|
19 (37.3)
32 (62.7)
15 (29.4)
36 (70.6)
|
0.836
0.300
|
Flap size (median (IQR);cm2)c
|
30 (24-35.8)
|
30 (28–42)
|
0.177
|
30 (24–36)
|
32 (28–42)
|
0.193
|
Operation time (mean ± SD (range);min)
|
196 ± 52
(80–340)
|
454 ± 84
(290–690)
|
< 0.001f
|
197 ± 53
(80–340)
|
450 ± 78
(290–595)
|
< 0.001f
|
Pathological margin
- Clear
- Close or Positive
|
20 (38.5)
32 (61.5)
|
27 (33.3)
54 (66.7)
|
0.581
|
20 (39.2)
31 (60.8)
|
18 (35.3)
33 (64.7)
|
0.838
|
Postoperative treatment
- No
- RT or CRTe
|
9 (17.3)
43 (82.7)
|
10 (12.3)
71 (87.7)
|
0.454
|
9 (17.6)
42 (82.4)
|
7 (13.7)
44 (86.3)
|
0.786
|
a Submental island flap; b Radial forearm free flap; c The Kolmogorov-Smirnov normality test revealed that the data was not normally distributed (p-value < 0.05).; d Eastern Cooperative Oncology Group Performance Status; e Radiotherapy or Chemoradiotherapy; f Statistically significant |
Oncological Outcomes
The 5-year OS, DSS, and LRC rates between the patients reconstructed with a submental island flap and those reconstructed with a radial forearm free flap were 70.1% and 64.8% (p = 0.612; Fig. 4A), 77.3% and 83.7% (p = 0.857; Fig. 4B), and 76.1% and 73.3% (p = 0.664; Fig. 4C), respectively. The results showed that there were no significant differences found in the survival and locoregional control rates between the two groups.
The multivariate Cox proportional hazard regression model analysis was performed based on the type of flap reconstruction and clinicopathological variables for the oncological outcomes. The analysis revealed that ECOG PS 2 was a significant independent predictor of overall mortality (HR 6.74, 95% CI 1.89–23.95, p = 0.003). Furthermore, nodal stage N3 compared to N0 and pathologically positive level I lymph nodes were associated with an increased risk of locoregional recurrence (HR 1.85, 95% CI 1.14–5.14, p = 0.048, and HR 3.52, 95% CI 1.08–11.52, p = 0.038, respectively). However, no variable affecting disease-specific mortality was detected (Table 3). Notably, the type of flap reconstruction did not statistically influence the oncological outcomes. Concerning disease recurrence, no significant differences in local, regional, and distant organ recurrences were observed between the submental island flap group and the radial forearms free flap group (11.8% and 15.7%, p = 0.775, 13.7% and 9.8%, p = 0.769, and 7.8% and 15.7%, p = 0.357, respectively).
Table 3
Multivariate Cox proportional hazard regression model analysis based on reconstruction type and clinicopathological variables potentially affecting oncological outcomes.
Variables
|
HR (95% CI)
|
p-value
|
Overall mortality
ECOG PSa 2
|
6.74 (1.89–23.95)
|
0.003b
|
Locoregional recurrence
N0
N1
N2
N3
Pathologically positive level I lymph nodes
|
Reference
1.38 (0.08–1.69)
1.25 (0.06–0.99)
1.85 (1.14–5.14)
3.52 (1.08–11.52)
|
0.203
0.826
0.048b
0.038b
|
a Eastern Cooperative Oncology Group Performance Status; b Statistically significant |
Subgroup Analysis
We performed a subgroup analysis of patients reconstructed with a submental island flap in association with level I cervical lymph node status on the oncological outcomes. The results demonstrated that there were no statistical differences in the 5-year OS, DSS, and LRC rates in patients with clinically negative or positive level I lymph nodes (68.4% and 73.5%, p = 0.837, 74.8% and 81.7%, p = 0.382, and 70.7% and 87.4%, p = 0.132, respectively) (Fig. 5A-C). Furthermore, pathologically negative or positive level I lymph nodes also did not affect the outcomes (69.3% and 72%, p = 0.699, 79.7% and 71.1%, p = 0.827, and 78.4% and 70%, p = 0.636, respectively) (Fig. 6A-C).
Functional Outcomes
There were no statistically significant differences in speech and swallowing functions between the two groups (Table 4). No patients had bad speech function, and most patients in both groups had excellent to good speech results (88.2% in the submental island flap group and 96.1% in the radial forearm free flap group, p = 0.269). In addition, most of the patients in both groups were able to take at least a soft diet (78.4% in the submental island flap group and 90.2% in the radial forearm free flap group, p = 0.172). However, obligated tube feeding was necessary in 1 patient (2%) in the submental island flap group.
Table 4
Speech and swallowing outcomes of patients
Functional outcomes
|
SMFa
n = 51 (%)
|
RFFb
n = 51 (%)
|
p-value
|
Speech
- Excellent or Good
- Fair or Poor
|
45 (88.2)
6 (11.8)
|
49 (96.1)
3 (3.9)
|
0.269
|
Swallowing
- Full or Soft diet
- Liquid diet or Feeding tube
|
40 (78.4)
11 (21.6)
|
46 (90.2)
5 (9.8)
|
0.172
|
a Submental island flap; b Radial forearm free flap |
Surgical Complications
Donor site complications were detected more frequently in the radial forearm free flap group (15.7% and 2%, p = 0.031). Bleeding at the wound edge was recorded in 1 patient in the submental island flap group on the first postoperative day. Suturing at the bleeding site under local anesthesia was successfully performed. There were no other records of donor site complications in the submental island flap group, including no marginal mandibular branch paralysis. The donor site defects in all cases were primarily closed, with pleasing cosmesis and without restricted neck extension. However, in the radial forearm free flap group, there was partial loss of the skin graft in 2 patients (3.9%), and arm function, including grip strength, pinch strength, and wrist movements, was restricted in 5 patients (9.8%). In addition, 1 patient with bleeding at the raw surface under the skin graft was detected on the first postoperative day and treated with cauterization (Table 5).
Table 5
Surgical complications, duration of hospital stay, and hospital costs of patients
|
SMFa
n = 51 (%)
|
RFFb
n = 51 (%)
|
p-value
|
Donor site complications
- No
- Yes
- Bleeding
- Wound dehiscence
- Partial loss of skin graft
- Restricted neck/arm function
|
50 (98.0)
1 (2.0)
-
-
-
|
43 (84.3)
1 (1.9)
-
2 (3.9)
5 (9.8)
|
0.031d
|
Recipient site complications
- No
- Yes
- Orocutaneous fistula
- Wound dehiscence
- Minor infection
- Hematoma
|
43 (86.0)
-
4 (7.8)
2 (3.9)
2 (3.9)
|
34 (66.7)
3 (5.9)
7 (13.7)
7 (13.7)
2 (3.9)
|
0.034d
|
Flap Complications
- No
- Yes
- Partial flap loss
- Revision of the anastomosis
|
46 (90.2)
5 (9.6)
-
|
44 (86.3)
5 (9.6)
2 (3.9)
|
0.760
|
Hospital stay (median (IQR);days)c
|
19 (15–24)
|
22 (17–30)
|
0.018d
|
Hospital costs (median (IQR);USD)c
|
4,098
(3,206-5,104)
|
6,243
(5,389-6,966)
|
< 0.001d
|
a Submental island flap; b Radial forearm free flap; c The Kolmogorov-Smirnov normality test revealed that the data was not normally distributed (p-value < 0.05).; d Statistically significant |
Recipient site complications were also detected significantly more frequently in the radial forearm free flap group (33.3% and 14%, p = 0.034). Orocutaneous fistula was recorded in 3 patients (5.9%) in the radial forearm free flap group, no such complication was recorded in the submental island flap group. The fistulas were treated with necrotic tissue debridement followed by re-suturing of the wounds. Other complications, including wound dehiscence, minor local infection, and small hematoma, were detected in both groups and treated successfully with wound re-suturing, local wound care, and wound reopening with collection removal (Table 5).
With regard to flap complications, neither group had total flap loss, but partial flap loss (partial epithelial loss) occurred in 5 patients in the submental island flap group. However, re-epithelialization was complete within four weeks with conservative management. Five patients in the radial forearm free flap group had less than 10% loss of the flap volume, and these patients were treated with tissue debridement and wound re-suturing. Radial forearm free flap congestion was detected in 2 patients, and the venous anastomosis was revised within 6 hours of the first postoperative day (Table 5).
Duration Of Hospital Stay And Hospital Costs
The median duration of hospital stay was significantly shorter in the submental island flap group than in the radial forearm free flap group (19 days and 22 days, respectively; p = 0.018). Concerning median hospital costs during admission, submental island flap reconstruction was significantly less costly than those involving radial forearm free flap reconstruction (4,098 US Dollars and 6,243 US Dollars, respectively, p < 0.001) (Table 5).