Literature searching and study screening
In total, 1867 articles were obtained from the four databases. After eliminating 639 duplicate articles, 1228 studies were further screened. After screening the titles, abstracts, and full texts, 42 eligible studies[2, 7, 13-52] were finally included based on the selection criteria (Fig 1).
Quality evaluation and main characteristics of the eligible studies
The diagnostic meta-analysis analyzed 42 eligible studies[2, 7, 13-52] published between 2013 and 2020. Thirty-seven studies detected lncRNA expression in Asian population, while 5 studies detected lncRNA expression in Caucasian populations. Sample types included plasma, serum, and plasma/serum exosomes. All SC patients were pathologically confirmed, and the control groups consisted of healthy donor individuals and benign stomach disease patients. A total of 49 different lncRNAs were examined across all included studies; most of the lncRNAs were upregulated in SC (Table 2). The quality assessment is shown in Fig 2.
Table 2. Main characteristics of eligible studies for diagnosis.
First author, year
|
Race
|
Pathologictype(E/C)
|
Sample size(E/C)
|
Specimen
|
lncRNA
|
State
|
Sen
|
Spe
|
TP
|
FP
|
FN
|
TN
|
QUADAS-2
|
(Refs)
|
Liu W, 2019
|
Asian
|
GC/HD
|
89/73
|
Serum
|
FEZF1-AS1
|
Up
|
75.3%
|
65.8%
|
67
|
25
|
22
|
48
|
5
|
13
|
|
|
|
|
|
AFAP1-AS1
|
Up
|
76.4%
|
56.2%
|
68
|
32
|
21
|
41
|
|
|
Yoruke E, 2018
|
Caucasian
|
GC/non-GC
|
40/42
|
Plasma
|
H19
|
Up
|
87.2%
|
38.1%
|
35
|
26
|
5
|
16
|
6
|
14
|
Liu Y, 2019
|
Asian
|
GC/HD
|
100/100
|
Serum
|
MALAT1
|
Up
|
85.8%
|
74.5%
|
86
|
26
|
14
|
75
|
4
|
15
|
Hashad D, 2016
|
Caucasian
|
GC/HD
|
32/30
|
Plasma
|
H19
|
Up
|
68.8%
|
56.7%
|
22
|
13
|
10
|
17
|
5
|
16
|
Li Q, 2014
|
Asian
|
GC/HD
|
79/81
|
Plasma exosome
|
LINC00152
|
Down
|
48.1%
|
85.2%
|
38
|
12
|
41
|
69
|
6
|
17
|
Liu Z, 2014
|
Asian
|
GC/HD
|
83/80
|
Plasma
|
FER1L4
|
Up
|
67.2%
|
80.3%
|
56
|
16
|
27
|
64
|
7
|
18
|
Liu J, 2018
|
Asian
|
GC/HD
|
50/50
|
Plasma
|
CTC-501O10.1
|
Up
|
90.0%
|
51.0%
|
45
|
25
|
5
|
26
|
6
|
19
|
|
|
|
|
|
AC100830.4
|
Up
|
84.0%
|
58.0%
|
42
|
21
|
8
|
29
|
|
|
|
|
|
|
|
RP11-210K20.5
|
Up
|
89.0%
|
55.0%
|
45
|
23
|
6
|
28
|
|
|
Lu Q, 2017
|
Asian
|
EGC/HD
|
76/76
|
Serum
|
XIST
|
Up
|
84.6%
|
59.0%
|
64
|
31
|
12
|
45
|
5
|
20
|
|
|
|
|
|
BCYRN1
|
Up
|
67.9%
|
85.9%
|
52
|
11
|
24
|
65
|
|
|
|
|
|
|
|
RRP1B
|
Down
|
85.9%
|
56.4%
|
65
|
33
|
11
|
43
|
|
|
|
|
|
|
|
TDRG1
|
Down
|
73.1%
|
60.3%
|
56
|
30
|
20
|
46
|
|
|
Mohamed W, 2019
|
Caucasian
|
GC/HD
|
35/25
|
Serum
|
H19
|
Up
|
90.9%
|
100.0%
|
32
|
0
|
3
|
25
|
5
|
21
|
Piao H, 2020
|
Asian
|
GC/HD
|
281/80
|
Plasma exosome
|
CEBPA-AS1
|
Up
|
74.0%
|
88.0%
|
208
|
10
|
73
|
70
|
6
|
22
|
Zhou H, 2016
|
Asian
|
GC/HD
|
77/60
|
Plasma
|
ZFAS1
|
Up
|
76.6%
|
63.9%
|
59
|
22
|
18
|
38
|
5
|
23
|
Cai C, 2019
|
Asian
|
GC/HD
|
63/29
|
Serum exosome
|
PCSK2-2:1
|
Up
|
84.0%
|
86.5%
|
53
|
4
|
10
|
25
|
6
|
24
|
Zhou X, 2015
|
Asian
|
GC/HD
|
90/90
|
Plasma
|
H19
|
Up
|
82.9%
|
72.9%
|
75
|
24
|
15
|
66
|
7
|
25
|
Elsayed E, 2018
|
Caucasian
|
GC/HD
|
50/50
|
Plasma
|
HOTAIR
|
Up
|
86.0%
|
94.0%
|
43
|
3
|
7
|
47
|
4
|
26
|
Xian H, 2018
|
Asian
|
GC/HD
|
50/50
|
Plasma
|
HULC
|
Up
|
58.0%
|
80.0%
|
29
|
10
|
21
|
40
|
5
|
27
|
|
|
|
|
|
ZNFX1-AS1
|
Up
|
84.0%
|
68.0%
|
42
|
16
|
8
|
34
|
|
|
Feng W, 2019
|
Asian
|
GC/HD
|
107/87
|
Serum
|
B3GALT5‐AS1
|
Up
|
64.5%
|
87.4%
|
69
|
11
|
38
|
76
|
5
|
28
|
Fu M, 2017
|
Asian
|
GC/HD
|
72/72
|
Serum
|
LINC00978
|
Up
|
80.0%
|
70.0%
|
58
|
22
|
14
|
50
|
5
|
29
|
Gao J, 2015
|
Asian
|
GC/HD
|
20/20
|
Plasma
|
UCA1
|
Up
|
85.0%
|
96.3%
|
17
|
1
|
3
|
19
|
6
|
30
|
|
|
|
|
|
PVT1
|
Down
|
70.8%
|
91.3%
|
14
|
2
|
6
|
18
|
|
|
Ghaedi H, 2018
|
Asian
|
GC/HD
|
62/40
|
Plasma
|
H19
|
Up
|
74.2%
|
90.0%
|
46
|
4
|
16
|
36
|
6
|
31
|
|
|
|
|
|
MEG3
|
Down
|
77.4%
|
52.5%
|
48
|
19
|
14
|
21
|
|
|
Guo X, 2020
|
Asian
|
EGC/HD
|
217/219
|
Plasma exosome
|
GC1
|
Up
|
97.0%
|
83.0%
|
210
|
37
|
7
|
182
|
5
|
32
|
Arita T, 2013
|
Asian
|
GC/HD
|
43/34
|
Plasma
|
H19
|
Up
|
74.0%
|
58.0%
|
32
|
14
|
11
|
20
|
6
|
33
|
Ji B, 2019
|
Asian
|
GC/HD
|
168/74
|
Plasma
|
LINC00086
|
Down
|
72.6%
|
83.8%
|
122
|
12
|
46
|
62
|
7
|
34
|
Jiang H, 2019
|
Asian
|
GC/HD
|
317/100
|
Plasma
|
PCGEM1
|
Up
|
72.9%
|
88.9%
|
231
|
11
|
86
|
89
|
7
|
7
|
Lin L, 2018
|
Asian
|
GC/HD
|
51/60
|
Plasma exosome
|
UEGC1
|
Up
|
88.0%
|
82.0%
|
45
|
11
|
6
|
49
|
5
|
35
|
|
|
|
|
|
UEGC2
|
Up
|
89.0%
|
58.0%
|
45
|
25
|
6
|
35
|
|
|
Pan L, 2017
|
Asian
|
GC/HD
|
60/37
|
Serum exosome
|
ZFAS1
|
Up
|
71.7%
|
75.7%
|
43
|
9
|
17
|
28
|
5
|
36
|
Jin C, 2016
|
Asian
|
GC/HD
|
173/110
|
Serum
|
HULC
|
Up
|
82.0%
|
83.6%
|
142
|
18
|
31
|
92
|
6
|
37
|
Zhang X, 2018
|
Asian
|
GC/HD
|
57/29
|
Serum exosome
|
UFC1
|
Up
|
78.0%
|
80.0%
|
44
|
6
|
13
|
23
|
5
|
2
|
Zhao R, 2018
|
Asian
|
GC/HD
|
126/120
|
Serum exosome
|
HOTTIP
|
Up
|
69.8%
|
85.0%
|
88
|
18
|
38
|
102
|
4
|
38
|
Burock S, 2015
|
Caucasian
|
GC/non-GC
|
76/54
|
Plasma
|
MACC1
|
Up
|
68.0%
|
89.0%
|
52
|
6
|
24
|
48
|
5
|
39
|
Ke D, 2017
|
Asian
|
GC/HD
|
51/53
|
Plasma
|
INHBAAS1
|
Down
|
92.7%
|
74.5%
|
47
|
14
|
4
|
39
|
6
|
40
|
|
|
|
|
|
MIR4435-2HG
|
Down
|
90.2%
|
74.5%
|
46
|
14
|
5
|
39
|
|
|
|
|
|
|
|
CEBPA-AS1
|
Down
|
78.0%
|
76.6%
|
40
|
12
|
11
|
41
|
|
|
|
|
|
|
|
UCA1
|
Down
|
73.2%
|
82.3%
|
37
|
9
|
14
|
44
|
|
|
|
|
|
|
|
AK001058
|
Down
|
95.1%
|
72.3%
|
49
|
15
|
2
|
38
|
|
|
|
|
|
47/52
|
Plasma
|
INHBAAS1
|
Down
|
82.7%
|
59.6%
|
39
|
21
|
8
|
31
|
|
|
|
|
|
|
|
MIR4435-2HG
|
Down
|
65.4%
|
87.2%
|
31
|
7
|
16
|
45
|
|
|
|
|
|
|
|
CEBPA-AS1
|
Down
|
96.2%
|
57.4%
|
45
|
22
|
2
|
30
|
|
|
|
|
|
|
|
AK001058
|
Down
|
76.9%
|
92.3%
|
36
|
4
|
11
|
48
|
|
|
Liu Y, 2019
|
Asian
|
GC/HD
|
94/40
|
Serum
|
HOXA11-AS
|
Up
|
78.7%
|
97.8%
|
74
|
1
|
20
|
39
|
7
|
41
|
Shan L, 2019
|
Asian
|
GC/HD
|
117/100
|
Serum
|
UCA1
|
Up
|
93.2%
|
78.6%
|
109
|
21
|
8
|
79
|
6
|
42
|
Shao Y, 2016
|
Asian
|
GC/HD
|
83/90
|
Plasma
|
RMRP
|
Down
|
59.1%
|
67.8%
|
49
|
29
|
34
|
61
|
5
|
43
|
Yang Z, 2019
|
Asian
|
GC/HD
|
109/106
|
Plasma
|
FOXD2-AS1
|
Up
|
83.0%
|
50.0%
|
90
|
53
|
19
|
53
|
5
|
44
|
|
|
|
|
|
PANDAR
|
Up
|
85.0%
|
63.0%
|
93
|
39
|
16
|
67
|
|
|
|
|
|
|
|
SMARCC2
|
Up
|
90.0%
|
55.0%
|
98
|
48
|
11
|
58
|
|
|
Xu H, 2020
|
Asian
|
GC/HD
|
109/50
|
Serum
|
MIAT
|
Up
|
81.5%
|
87.5%
|
89
|
6
|
20
|
44
|
5
|
45
|
Xu Y, 2018
|
Asian
|
GC/HD
|
34/34
|
Plasma
|
DGCR5
|
Down
|
58.0%
|
87.0%
|
20
|
4
|
14
|
30
|
6
|
46
|
Xu Y, 2019
|
Asian
|
GC/HD
|
45/45
|
Plasma
|
LINC01225
|
Up
|
50.0%
|
90.0%
|
23
|
5
|
23
|
41
|
4
|
47
|
Yang T, 2016
|
Asian
|
GC/HD+GS
|
133/152
|
Serum
|
H19
|
Up
|
65.0%
|
53.0%
|
86
|
71
|
47
|
81
|
7
|
48
|
|
|
|
|
|
LINC00152
|
Up
|
40.0%
|
72.0%
|
53
|
43
|
80
|
109
|
|
|
Zheng P, 2020
|
Asian
|
GC/HD
|
60/60
|
Plasma
|
SLC2A12-10:1
|
Up
|
68.0%
|
75.0%
|
41
|
15
|
19
|
45
|
5
|
49
|
Zheng R, 2018
|
Asian
|
GC/HD
|
241/228
|
Plasma
|
FAM49B-AS
|
Up
|
58.0%
|
60.0%
|
140
|
91
|
101
|
137
|
6
|
50
|
|
|
|
|
|
GUSBP11
|
Up
|
46.0%
|
75.0%
|
111
|
57
|
130
|
171
|
|
|
|
|
|
|
|
CTDHUT
|
Up
|
73.0%
|
65.0%
|
176
|
80
|
65
|
148
|
|
|
Zhou Q, 2020
|
Asian
|
GC/GS+GA+GD
|
200/278
|
Serum
|
C5orf66-AS1
|
Down
|
77.5%
|
53.6%
|
155
|
129
|
45
|
149
|
6
|
51
|
Tan L, 2016
|
Asian
|
GC/HD
|
117/80
|
Plasma
|
GACAT2
|
Down
|
87.2%
|
28.2%
|
102
|
57
|
15
|
23
|
5
|
52
|
Note: E/C: experimental group/control group; GC: gastric cancer; EGC: early gastric cancer; HD: healthy donor individuals; GS: superficial gastritis; GA: atrophic gastritis; GD: gastric dysplasia; SEN: sensitivity; SPE: specificity; TP: true positive; FP: false positive; FN: false negative; TN: true negative; QUADAS-2: Quality Assessment of Diagnostic Accuracy Studies 2.
Diagnostic accuracy of lncRNA
A total of 42 eligible diagnostic studies were meta-analyzed. As illustrated in Fig 3, the pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were 0.78 (95% CI: 0.75-0.81), 0.75 (95% CI: 0.71-0.78), 3.09 (95% CI: 2.66-3.58), 0.29 (95% CI: 0.25-0.33), and 10.67 (95% CI: 8.34-13.65), respectively. As demonstrated in Fig 4A, the AUC value of the SROC was 0.83 (95% CI: 0.80-0.86). The SROC results were further evaluated through the HSROC model. As shown in Fig 4B, the β estimate was 0.11 (95% CI: -0.19-0.40) and the corresponding P value was 0.485. The lambda estimate was 2.38 (95% CI: 2.13-2.63).
Heterogeneity analysis
As illustrated in Fig 3, obvious heterogeneity was found in the pooled sensitivity (I2 = 88.93%, P < 0.01), specificity (I2 = 88.14%, P < 0.01), positive likelihood ratio (I2 = 88.49%, P < 0.01), negative likelihood ratio (I2 = 88.71%, P < 0.01), and diagnostic odds ratio (I2 = 100.00%, P < 0.01).
A nontypical shoulder arm appearance was observed in the ROC plane (Fig 5A). Twenty out of the 63 studies of the Galbraith star chart and 10 out of 42 studies of the bivariate box plot fell outside the 95% CI (Fig 5B and 5C). Fig 5D shows the meta-regression forest map. All studies were grouped according to race, pathological types of experimental groups, pathological types of control groups, sample size, specimen type, dysregulated state of lncRNAs, and lncRNA types. Table 3 shows the changes in sensitivity, specificity, and I2 values after meta-regression and subgroup analysis.
Table 3. Subgroup analysis of the diagnostic efficacy of lncRNA in stomach cancer.
Group
|
Subgroup
|
No. of studies
|
No. of patients
|
Sensitivity
|
Heterogeneity (I2; P value)
|
Specificity
|
Heterogeneity (I2; P value)
|
AUC
|
Meta-regression (P value)
|
Overall
|
|
42
|
7524
|
0.78 [0.75, 0.81]
|
88.93% ; <0.001
|
0.75 [0.71, 0.78]
|
88.14% ; <0.001
|
0.83 [0.80 - 0.86]
|
|
Race
|
Asian
|
60
|
7090
|
0.78 [0.75, 0.81]
|
89.38% ; <0.001
|
0.74 [0.70, 0.77]
|
87.51% ; <0.001
|
0.83 [0.79 - 0.86]
|
0.48
|
Caucasian
|
5
|
434
|
0.81 [0.70, 0.89]
|
75.16% ; <0.001
|
0.86 [0.52, 0.97]
|
94.22% ; <0.001
|
0.87 [0.84 - 0.90]
|
0.48
|
Pathologic types(E)
|
GC
|
60
|
6936
|
0.78 [0.74, 0.81]
|
88.19% ; <0.001
|
0.75 [0.71, 0.79]
|
88.35% ; <0.001
|
0.83 [0.80 - 0.86]
|
0.38
|
EGC
|
5
|
588
|
0.85 [0.72, 0.93]
|
91.39% ; <0.001
|
0.71 [0.58, 0.81]
|
90.27% ; <0.001
|
0.84 [0.80 - 0.87]
|
0.38
|
Pathologic types(C)
|
health
|
61
|
6549
|
0.79 [0.75, 0.82]
|
89.32% ; <0.001
|
0.75 [0.71, 0.79]
|
86.53% ; <0.001
|
0.84 [0.80 - 0.87]
|
0.15
|
non-GC
|
2
|
212
|
-
|
-
|
-
|
-
|
-
|
0.86
|
GS
|
2
|
763
|
-
|
-
|
-
|
-
|
-
|
0.10
|
Sample size
|
N≤100
|
21
|
1153
|
0.79 [0.74, 0.84]
|
74.67% ; <0.001
|
0.77 [0.69, 0.84]
|
84.82% ; <0.001
|
0.85 [0.82 - 0.88]
|
0.62
|
100<N≤200
|
28
|
2722
|
0.79 [0.74, 0.82]
|
79.14% ; <0.001
|
0.74 [0.68, 0.79]
|
85.75% ; <0.001
|
0.83 [0.80 - 0.86]
|
0.95
|
N>200
|
16
|
3649
|
0.77 [0.68, 0.84]
|
95.41% ; <0.001
|
0.73 [0.65, 0.79]
|
92.57% ; <0.001
|
0.81 [0.77 - 0.84]
|
0.58
|
Specimen
|
Plasma
|
23
|
3467
|
0.78 [0.74, 0.82]
|
87.25% ; <0.001
|
0.72 [0.67, 0.77]
|
86.87% ; <0.001
|
0.82 [0.79 - 0.85]
|
0.34
|
Serum
|
12
|
2468
|
0.78 [0.72, 0.83]
|
90.56% ; <0.001
|
0.75 [0.67, 0.82]
|
90.37% ; <0.001
|
0.84 [0.80 - 0.87]
|
0.99
|
Exosome
|
8
|
1589
|
0.81 [0.70, 0.89]
|
92.18% ; <0.001
|
0.81 [0.76, 0.86]
|
69.40% ; <0.001
|
0.87 [0.84 - 0.90]
|
0.17
|
Dysregulated state
|
Upregulated
|
46
|
6003
|
0.78 [0.74, 0.82]
|
90.19% ; <0.001
|
0.75 [0.71, 0.80]
|
87.40% ; <0.001
|
0.84 [0.80 - 0.87]
|
0.78
|
Downregulated
|
19
|
1521
|
0.79 [0.72, 0.84]
|
84.54% ; <0.001
|
0.72 [0.64, 0.79]
|
89.12% ; <0.001
|
0.82 [0.79 - 0.86]
|
0.78
|
lncRNA
|
H19
|
7
|
848
|
0.78 [0.70, 0.84]
|
75.98% ; <0.001
|
0.72 [0.49, 0.88]
|
89.62% ; <0.001
|
0.82 [0.78 - 0.85]
|
0.69
|
UCA1
|
3
|
361
|
0.87 [0.81, 0.91]
|
83.60% ; 0.02
|
0.82 [0.76, 0.88]
|
45.20% ; 0.161
|
0.92 [0.84 - 0.99]
|
0.13
|
CEBPA-AS1
|
3
|
564
|
0.77 [0.73, 0.81]
|
86.00% ; 0.001
|
0.76 [0.69, 0.82]
|
86.80% ; 0.001
|
0.88 [0.84 - 0.92]
|
0.63
|
Note: E/C: experimental group/control group; GC: gastric cancer; EGC: early gastric cancer; HD: healthy donor individuals; GS: superficial gastritis; AUC, area under the curve.
Sensitivity analysis and publication bias
First, sensitivity analysis was carried out to determine the stability of our results. The removal of individual studies exhibited no noticeable changes in pooled results (Additional file 1: Supplementary Figure 1, Fig. S1A). The P value of Deeks’ funnel plot asymmetry test was 0.12 (Additional file 1: Supplementary Figure 1, Fig. S1B).
Clinical values of lncRNAs for SC diagnosis
As shown in Fig 6, Fagan's nomogram revealed that if the patient had a positive lncRNA test result, the actual probability of suffering from SC was 76%, while the probability was 22% if a negative test result was obtained.