The characteristics of included studies:
Based on the search strategy mentioned above, a total of 1110 articles were searched from Pubmed, Embase and Web of Science. Finally, 41 studies published between 1994 and 2019 with a total of 28, 973 patients, were included in our meta-analysis according to the inclusion and exclusion criteria(6-10, 20-55). The specific literature selection process was shown in Figure 1.
Table 1 summarizes the characteristics of all the studies included in this meta-analysis.All of the included studies were retrospective. Among these studies, 23 studies were from China(7, 9, 10, 30-49), 7 studies were from USA(6, 50-55), 7 studies were from Japan(8, 24-29), 1 study each from Italy(22), Australia(20), the Netherlands(21), and Turkey(23), participants in 39 studies were patients with SCC. Three of the 41 studies included patients who received chemoradiotherapy only (26, 48, 53) and the remaining 38 studies included patients who underwent surgical resection with or without chemoradiotherapy. Among 41 studies, 35 studies reported the relationship between tumor length of EC and OS, 10 studies reported disease-free survival (DFS), 2 reported progression-free survival (PFS)(48, 53), 1 reported cancer-specific survival (CSS)(55) and 1 reported disease-specific survival (DSS)(20). The cut-off value (cm) applied in all studies was between 1.5 and 6. Eighteen studies used a tumor length cut-off value≥5, and 23 studies used a tumor length cut-off value < 5. Two studies reported only relative risk (RR) data (41, 42), and we used RR to replace HR when pooled the data.
Table1: Main characteristics of all included studies.
Author
|
Publication year
|
Patient source
|
No. of patients
|
Age(years)(median and range)
|
Follow-up(months)(median and range)
|
Treatment
|
LN
metastasis (
% of total)
|
Histology
|
Tumor location
|
TNM stage
|
Ⅲ Ⅳ(% of total)
|
Cut-off value(cm)
|
Survival analusis
|
HR
reported
|
NOS
|
Upper
|
Middle
|
Lower
|
Tachibana, M.
|
1999
|
Japan
|
65
|
65
|
NR
|
Surg²
|
30.77
|
SCC
|
6
|
59
|
T1-2N±
|
NR
|
4
|
OS
|
MV
|
7
|
Shimada Hideaki
|
2004
|
Japan
|
374
|
65
|
NR
|
Surg²
|
65.78
|
SCC
|
98
|
276
|
Ⅰ-Ⅳ
|
77.0
|
5
|
OS
|
MV
|
6
|
Barbour, A. P.
|
2008
|
Australia
|
131
|
61(30-78)
|
61
|
Surg³
|
51.15
|
SCC,AC
|
4
|
28
|
85
|
0-Ⅲ
|
32.8
|
6
|
DSS
|
MV
|
8
|
Yendamuri , S.
|
2009
|
US
|
209
|
64(33-84)
|
NR
|
Surg²
|
45.45
|
SCC,AC
|
3
|
27
|
179
|
Ⅰ-Ⅲ
|
34.9
|
3
|
OS
|
UV/MV
|
7
|
Shitara Kohei
|
2010
|
Japan
|
363
|
63
|
67.2(25.2-94.8)
|
Surg¹
|
NR
|
SCC
|
NR
|
Ⅰ-Ⅳ
|
56.2
|
5
|
OS
|
UV/MV
|
7
|
Heijl, M.V
|
2010
|
The Netherlands
|
199
|
64(35-78)
|
60+
|
Surg²
|
72.36
|
SCC,AC
|
NR
|
NR
|
NR
|
4
|
OS/DFS
|
UV/MV
|
8
|
Gaur, P.
|
2011
|
US
|
164
|
65(26-84)
|
NR
|
Surg²
|
14.63
|
AC
|
0
|
8
|
156
|
NR
|
NR
|
2
|
OS
|
UV/MV
|
8
|
Lu C. L.
|
2011
|
China
|
127
|
59(39-77)
|
NR
|
Surg²
|
32.28
|
SCC
|
NR
|
Ⅰ-Ⅲ
|
30.7
|
2
|
OS/DFS
|
MV(RR)
|
7
|
Wang Bing-Yen
|
2011
|
China
|
582
|
65.4(30-88)
|
31.9
|
Surg²
|
54.64
|
SCC
|
77
|
340
|
165
|
T1-4N0-3
|
NR
|
3
|
OS
|
MV
|
7
|
Yamamoto Sachiko
|
2011
|
Japan
|
170
|
64
|
NR
|
Surg¹
|
NR
|
SCC
|
18
|
107
|
45
|
Ⅰ
|
NR
|
5
|
OS
|
MV
|
6
|
Song Zhengbo
|
2012
|
China
|
201
|
59(31-78)
|
52(30-136)
|
Surg
|
NR
|
SCC
|
12
|
102
|
87
|
T₁-₂
|
NR
|
3
|
OS
|
MV
|
7
|
Chen, J.
|
2012
|
China
|
945
|
NR
|
NR
|
Surg²
|
100
|
SCC
|
146
|
712
|
87
|
T1-4N+
|
NR
|
5
|
OS
|
UV/MV
|
7
|
Feng, J. F.
|
2013
|
China
|
132
|
73.6 ± 2.6
|
NR
|
Surg²
|
56.11
|
SCC
|
6
|
55
|
71
|
T1-4N0-3
|
NR
|
4
|
OS
|
UV/MV
|
6
|
Matsumoto , S.
|
2013
|
Japan
|
32
|
63
|
NR
|
Surg²
|
42.86
|
SCC
|
3
|
20
|
6
|
Ⅰ-Ⅲ
|
71.9
|
6
|
OS
|
MV
|
7
|
Zeybek Arife
|
2013
|
Turkey
|
116
|
60.0 (33-75)
|
39.7
|
Surg²
|
65.52
|
SCC,AC
|
10
|
36
|
70
|
Ⅱ-Ⅲ
|
64.7
|
3
|
DFS
|
MV
|
6
|
Chen Lujun
|
2014
|
China
|
103
|
58
|
NR
|
Surg²
|
47.57
|
SCC
|
NR
|
Ⅰ-Ⅳ
|
33.0
|
3.5
|
OS
|
UV/MV
|
7
|
Shridhar Ravi
|
2014
|
US
|
154
|
65
|
NR
|
Surg¹
|
84.42
|
AC
|
0
|
5
|
80
|
Ⅰ-Ⅳ
|
68.8
|
5
|
OS
|
MV
|
6
|
Freilich, J.
|
2015
|
US
|
232
|
64.4±11.3
|
25.9(2.5-124.2)
|
Surg¹
|
83.19
|
SCC,AC
|
8
|
17
|
118
|
Ⅰ-Ⅳ
|
72.8
|
5
|
OS
|
MV
|
7
|
Hulshoff, J. B.
|
2015
|
US
|
105
|
64(57-69)
|
29(15.–56.0)
|
Surg¹
|
73.33
|
SCC,AC
|
0
|
12
|
49
|
NR
|
NR
|
5
|
DFS
|
UV/MV
|
6
|
Ma Ming-Quan
|
2015
|
China
|
362
|
54.5
|
84(6-144)
|
Surg²
|
12.18
|
SCC
|
28
|
243
|
100
|
Ⅰ-Ⅲ
|
25.1
|
4
|
OS
|
MV(RR)
|
8
|
Miao Longsheng
|
2015
|
China
|
1342
|
59.5±7.9
|
30
|
Surg²
|
45.98
|
SCC,AC
|
NR
|
Ⅰ-Ⅳ
|
41.1
|
4
|
OS
|
UV/MV
|
7
|
Hirahara Noriyuki
|
2016
|
Japan
|
147
|
65.7
|
NR
|
Surg¹
|
59.86
|
SCC
|
8
|
65
|
52
|
Ⅰ-Ⅲ
|
37.4
|
3
|
CSS
|
UV/MV
|
7
|
Hwang Jwu-Yun
|
2016
|
China
|
294
|
55
|
20.4
|
Surg²
|
58.84
|
SCC
|
51
|
114
|
129
|
Ⅰ-Ⅳ
|
49.3
|
3.2
|
OS
|
UV/MV
|
7
|
Jia Wei
|
2016
|
China
|
83
|
NR
|
NR
|
Surg²
|
56.63
|
SCC
|
8
|
44
|
31
|
Ⅰ-Ⅲ
|
43.4
|
5
|
OS/DFS
|
MV
|
6
|
Ma Qilong
|
2016
|
China
|
725
|
58(32-80)
|
NR
|
Surg²
|
46.48
|
SCC
|
NR
|
Ⅰ-Ⅲ
|
38.2
|
5
|
OS
|
UV/MV
|
7
|
Sakanaka
|
2016
|
Japan
|
144
|
67(41-85)
|
48(13-88)
|
CRT
|
74.31
|
SCC
|
33
|
79
|
32
|
Ⅰ-Ⅳ
|
63.9
|
4
|
OS
|
MV
|
7
|
Valmasoni Michele
|
2016
|
Italy
|
357
|
62 ± 9.3
|
NR
|
Surg²
|
50.98
|
SCC
|
90
|
147
|
120
|
Ⅰ-Ⅲ
|
47.6
|
3
|
OS
|
MV
|
7
|
Valmasoni Michele
|
2016
|
Italy
|
305
|
63 ± 11.2
|
NR
|
Surg²
|
66.23
|
AC
|
0
|
6
|
299
|
Ⅰ-Ⅲ
|
58.4
|
3
|
OS
|
MV
|
7
|
Wu Jie
|
2016
|
China
|
1435
|
58.3
|
24(1-128)
|
Surg²
|
53.24
|
SCC
|
33
|
691
|
711
|
T1-4N0-3M0-1
|
NR
|
4
|
OS
|
UV/MV
|
7
|
Duan, J.
|
2016
|
China
|
328
|
61
|
44.9(3.4–107.6)
|
Surg²
|
42.4
|
SCC
|
24
|
202
|
102
|
Ⅰ-Ⅲ
|
57.6
|
4.2
|
OS/DFS
|
UV/MV
|
8
|
Gao, S. H.
|
2016
|
China
|
126
|
NR
|
NR
|
Surg¹
|
85
|
SCC
|
NR
|
Ⅰ-Ⅲ
|
54.8
|
4
|
OS/DFS
|
MV
|
6
|
Li, S. P.
|
2016
|
China
|
100
|
59.2±10.3
|
NR
|
Surg¹
|
55
|
SCC
|
12
|
41
|
37
|
Ⅰ-Ⅲ
|
54.0
|
4
|
OS/DFS
|
MV
|
6
|
Tian, R.
|
2016
|
China
|
442
|
60(20-88)
|
NR
|
Surg²
|
47.5
|
SCC
|
39
|
277
|
126
|
Ⅰ-Ⅲ
|
43.7
|
5
|
OS/DFS
|
UV/MV
|
7
|
Xi Mian
|
2017
|
US
|
496
|
67(20-92)
|
24.2(2.8-155.9)
|
CRT
|
70.97
|
SCC,AC
|
138
|
358
|
Ⅰ-Ⅲ
|
67.9
|
5
|
PFS
|
MV
|
7
|
Zeng, Y.
|
2017
|
US
|
1242
|
NR
|
NR
|
Surg
|
NR
|
SCC,AC
|
39
|
136
|
913
|
NR
|
NR
|
1.5
|
OS/CSS
|
MV
|
7
|
Li, J.
|
2017
|
China
|
294
|
58(38-70)
|
26
|
Surg²
|
67.3
|
SCC
|
6
|
139
|
149
|
T1-4N±
|
NR
|
5
|
OS
|
MV
|
7
|
Yang, Y. S.
|
2017
|
China
|
508
|
59
|
37.5(1-105)
|
Surg²
|
40.9
|
SCC
|
78
|
260
|
170
|
T1-4N±
|
NR
|
4
|
OS
|
UV/MV
|
7
|
Zhang, Xiangwei
|
2017
|
China
|
498
|
59(38-81)
|
47.2(6.–64.5)
|
Surg²
|
41
|
SCC
|
NR
|
T1-4N±
|
NR
|
3
|
OS
|
UV/MV
|
7
|
Bai G.
|
2018
|
China
|
80
|
54.7±12.6
|
28(9-62)
|
Surg²
|
52.5
|
SCC
|
24
|
36
|
20
|
T1-4N±
|
NR
|
5
|
OS/DFS
|
MV
|
6
|
Cheng Ya-Fu
|
2018
|
China
|
14394
|
NR
|
NR
|
Surg¹
|
73.5
|
SCC
|
2897
|
5028
|
3213
|
0-Ⅳ
|
71.5
|
5
|
OS
|
MV
|
6
|
Xu Hong-Yao
|
2018
|
China
|
751
|
65
|
56.6(25.2-112.5)
|
CRT
|
78.96
|
SCC+Others
|
157
|
420
|
128
|
T1-4N0-3
|
NR
|
5
|
PFS
|
UV/MV
|
7
|
Gu, L.
|
2019
|
China
|
116
|
NR
|
NR
|
Surg¹
|
85.34
|
SCC
|
NR
|
Ⅰ-Ⅳ
|
55.2
|
5
|
OS
|
UV/MV
|
6
|
NO=number, HR=hazard ratio, “M” means the HR come from multivariate analysis, “U” means the HR comes from univariate analysis, NOS=Newcastle –Ottawa Quality Assessment Scale, R=Retrospective , SCC=squamous cell carcinoma , AC=adenocarcinoma , OS=overall survival, PFS=progression -free survival, DFS=disease-free survival, DSS=disease-specific survival, CSS=cancer-specific survival, NR=not reported , CRT=chemoradiation therapy, Surg¹:±Neo CRT/±Surg/±Adj CRT; Surg²: Surg±Adj CRT; Surg³: Neo CRT+Surg
|
tumor length and OS in EC:
Among the 35 studies evaluating the relationship between tumor length of EC and OS,because of the significant heterogeneity between the included studies (I²= 66.8%; ph < 0.001), the pooled HR and 95% CI was calculated by a random-effect model. The pooled HR of 1.301(95%CI:1.210-1.399) suggests that long tumor is associated with poor OS (Fig. 2).
tumor length and DFS in EC:
Ten studies showed the HR and 95% CI of tumor length to DFS. The pooled data showed that long tumor was associated with poor DFS (HR= 1.378; 95% CI:1.179-1.609, p <.001). Due to obvious heterogeneity (I²= 76.7%, ph < 0.001), the random-effect model was still used to calculate the pooled HR and its 95% CI (Fig. 3).
tumor length and PFS in EC:
No significant heterogeneity was found in these studies of PFS evaluation, so a fixed-effect model was used to calculate the pooled HR and 95% CI (HR = 1.161; 95% CI: 1.008-1.337, p <.005). This data suggests that long tumor was associated with shorter PFS.
tumor length and CSS/DSS in EC:
Two studies reported preoperative tumor length and CSS / DSS data for EC. There was no significant heterogeneity (I²= 0.0%, ph = 0.551). The pooled HR and 95% CI (HR=1.856; 95% CI:1.173-2.937, p <.001), indicating that the long tumor was also related to the poor CSS / DSS.
Subgroup analyzes:
In order to analyze the impact of heterogeneity on the results of this study, we performed a subgroup analysis of the extracted data. Subgroup analyses by patient source (China and the others), histology (SCC, AC and mixed), treatment (Surg²and the others), median age (<60 and ≥ 60), cut-off value (<5 and ≥5), sample size (<200 and ≥ 200), percentage of patients with LN metastasis(<50% and ≥50%), percentage of patients with TNM stag Ⅲ, Ⅳ classification (<50% and ≥50%) and HR analysis method (MV and UV), to explore the potential sources of heterogeneity for pooled result of OS. Because the cut-off values of tumor length were different among the included studies, we performed subgroup analysis according to different cut-off values. Because most articles choose 5(cm)as the cut-off value of tumor length, we take this cut-off value as the boundary, for the cut-off value ≥ 5, the pooled HR was 1.259 (95%CI: 1.096-1.446 , I²=59.4%, ph = 0.002), for the cut-off value < 5, the pooled HR was 1.322 (95%CI: 1.210-1.443 , I²=67.9%, ph < 0.001).The results of subgroup analyses showed that the different classification methods have no obvious influence on HR (Table 2). In almost all subgroups, long tumor was significantly related to poor OS, which showed our pooled HR result on OS was stable and reliable. Considering the limited researchs on tumor length and DFS, PFS, DSS, and CSS, no other subgroup analysis is required.
Table 2: Subgroup analyses reflecting the association between tumor length and OS in EC patients.
|
Random-effects model
|
|
|
Fixed-effects model
|
|
|
|
Subgroup
|
No.of studies
|
HR (95% CI)
|
P
|
|
HR (95% CI)
|
P
|
I²(%)
|
Ph
|
Overall
|
35
|
1.301(1.210-1.399)
|
<.001
|
|
1.151(1.117-1.185)
|
<.001
|
66.8%
|
<.001
|
Patient source
|
|
|
|
|
|
|
|
|
China
|
23
|
1.290(1.191-1.397)
|
<.001
|
|
1.155(1.117-1.195)
|
<.001
|
64.6%
|
<.001
|
Japan
|
6
|
1.540(1.108-2.142)
|
.010
|
|
1.446(1.174-1.781)
|
<.001
|
47.0%
|
.093
|
US
|
5
|
1.457(0.851-2.495)
|
.170
|
|
1.288(1.046-1.586)
|
.017
|
84.1%
|
<.001
|
The Netherlands
|
1
|
1.080(1.008-1.157)
|
.029
|
|
1.080(1.008-1.157)
|
.029
|
—
|
—
|
Italy
|
1
|
1.258(0.890-1.778)
|
.193
|
|
1.283(1.001-1.644)
|
.049
|
—
|
—
|
Histology
|
|
|
|
|
|
|
|
|
SCC
|
28
|
1.332(1.227-1.445)
|
<.001
|
|
1.163(1.125-1.203)
|
<.001
|
64.0%
|
<.001
|
AC
|
3
|
1.616(0.950-2.748)
|
.077
|
|
1.469(1.103-1.957)
|
.009
|
67.9%
|
.044
|
mixed
|
5
|
1.128(0.895-1.421)
|
.307
|
|
1.095(1.029-1.166)
|
.004
|
78.5%
|
.001
|
Treatment
|
|
|
|
|
|
|
|
|
Surg²
|
24
|
1.311(1.207-1.423)
|
<.001
|
|
1.127(1.091-1.165)
|
<.001
|
67.3%
|
<.001
|
Others
|
11
|
1.270(1.066-1.514)
|
.007
|
|
1.250(1.170-1.336)
|
<.001
|
59.5%
|
.006
|
Sample size
|
|
|
|
|
|
|
|
|
≤200
|
16
|
1.458(1.245-1.707)
|
<.001
|
|
1.163(1.098-1.233)
|
<.001
|
56.3%
|
.003
|
>200
|
19
|
1.264(1.157-1.380)
|
<.001
|
|
1.146(1.108-1.186)
|
<.001
|
73.3%
|
<.001
|
Median age
|
|
|
|
|
|
|
|
|
<60
|
9
|
1.422(1.228-1.647)
|
<.001
|
|
1.124(1.071-1.179)
|
<.001
|
74.5%
|
<.001
|
≥60
|
14
|
1.281(1.132-1.450)
|
<.001
|
|
1.133(1.083-1.185)
|
<.001
|
67.5%
|
<.001
|
Cut-off value
|
|
|
|
|
|
|
|
|
<5cm
|
21
|
1.322(1.210-1.443)
|
<.001
|
|
1.123(1.086-1.162)
|
<.001
|
67.9%
|
<.001
|
≥5cm
|
14
|
1.259(1.096-1.446)
|
.001
|
|
1.241(1.169-1.317)
|
<.001
|
59.4%
|
.002
|
LN metastasis(% of total)
|
|
|
|
|
|
|
|
|
<50%
|
13
|
1.351(1.189-1.537)
|
<.001
|
|
1.106(1.061-1.153)
|
<.001
|
73.9%
|
<.001
|
≥50%
|
18
|
1.281(1.162-1.413)
|
<.001
|
|
1.194(1.144-1.246)
|
<.001
|
62.3%
|
<.001
|
TNM stage ⅢⅣ(% of total)
|
|
|
|
|
|
|
|
|
<50%
|
12
|
1.441(1.240-1.676)
|
<.001
|
|
1.377(1.245-1.524)
|
<.001
|
44.5%
|
.062
|
≥50%
|
14
|
1.243(1.088-1.421)
|
.001
|
|
1.190(1.134-1.249)
|
<.001
|
69.5%
|
<.001
|
HR estimate
|
|
|
|
|
|
|
|
|
UV
|
16
|
1.676(1.469-1.914)
|
<.001
|
|
1.218(1.180-1.256)
|
<.001
|
91.3%
|
<.001
|
MV
|
35
|
1.301(1.210-1.399)
|
<.001
|
|
1.151(1.117-1.185)
|
<.001
|
66.8%
|
<.001
|
HR = hazard ratio, CI = confidence interval, Ph = P value of Q test for heterogeneity test, SCC = squamous cell carcinoma, AC = adenocarcinoma; UVA = univariate analysis, MVA = multivariate analysis, LN= lymph node, Surg²: Surg±Adj CRT
|
Sensitivity analyses:
By omitting each individual study for sensitivity analysis, the results showed that when excluding any study, the pooled HRs did not change substantially, indicating our results were stable (Fig. 4,5).
Publication bias:
The publication bias was evaluated by Begg funnel plot and Egger linear regression test. Considering the limited number of studies involving the relationship between tumor length and DFS, PFS, DSS or CSS, we analyzed the publication bias of tumor length to OS (Pr> |z| =0.127 for the Begg test and p < 0.01 for the Egger test). (Figure 6) Therefore, we conducted a further trim-and-fill analysis (Figure 7).The adjusted result(HR=1.195, 95% CI: 1.111-1.284) was similar to our previous pooled result. This result showed that despite the publication bias, it did not substantially affect the results of our study.