Patient enrollment and baseline characteristics
The flow diagram of the study is shown in Fig. 1. A total of 465 patients were screened for enrollment from Jan 2013 to Oct 2018. We excluded 126 patients who were not eligible for the study. Another 70 patients were excluded due to the absence of high-risk features, inadequate lymph node dissection, and adjuvant CCRT. Finally, 269 eligible patients were included in the study. The patient and tumor characteristics are shown in Table 1. Among the included patients, 110 received curative gastrectomy without adjuvant therapies (Group OBS); 64 had S-1 (Group S-1), and 95 had platinum-based doublet (Group P) adjuvant chemotherapies. The distributions of sex, H. pylori infection, HER2 status, baseline performance status, GEJ tumors, and serum cancer antigen 19 − 9 (CA19-9) levels were similar among the three groups. The median ages were similar in patients who received adjuvant chemotherapies, but patients who received surgery alone were significantly older (S-1 vs. P vs. OBS, 65.2 vs. 60.3 vs. 73.3 years). More patients had early stages of the disease, notably stage IB/II, T1-2 or N0 disease, in the Group OBS as compared with Group S-1 or P (OBS: stage IB/II, 58.2%; T1-2 58.2%; N0, 53.6%; all p < 0.001 when compared with S-1 or P). Similarly, Group S-1 had more stage IB/II and N0 patients as compared with Group P (S-1 vs. P: stage IB/II, 48.4% vs. 23.2%, p = 0.002; N0, 21.9% vs. 4.2%, p = 0.001). Lymphovascular invasion (LVI) was more evident in Group S-1 and P as compared with OBS, and the former two had no significant between-group differences (OBS vs. S-1 vs. P: 39.1% vs. 65.6% vs. 77.9%; p < 0.001 when OBS was compared with S1 or P; p = 0.088 in S1 vs. P). Lauren histology classification of the tumors were generally balanced, with the exception of more diffuse type tumors in Group P as compared with OBS or S1 (OBS vs. S-1 vs. P: 30.9% vs. 43.8% vs. 51.6%; p = 0.009 in OBS vs. P). The median serum CEA was higher in Group P as compared with S-1 or OBS (OBS vs. S-1 vs. P: 2.1 vs. 1.8 vs. 2.8 ng/mL; p = 0.007 via a Kruskal-Wallis test). A total of 8 patients in the study had received neoadjuvant therapies prior to surgery, where 3 had triplet; 1 had doublet; 2 had immune checkpoint inhibitors and chemotherapies, and 1 had CCRT.
Table 1
Group
|
OBS
|
S-1
|
P
|
|
|
Operation alone
(n = 110)
|
S-1
(n = 64)
|
Platinum-based doublets
(n = 95)
|
p
|
OBS vs. S-1
|
OBS vs. P
|
S-1 vs. P
|
Age, median (IQR)
|
73.3
|
(63.2–81.2)
|
65.2
|
(52.0-70.7)
|
60.3
|
(52.5–68.0)
|
< 0.001*
|
< 0.001*
|
0.066
|
Male, n (%)
|
63
|
(57.3)
|
32
|
(50.0)
|
58
|
(61.1)
|
0.353
|
0.264
|
0.224
|
AJCC stage, n (%)
|
0.321
|
< 0.001*
|
0.002*
|
IB/II
|
64
|
(58.2)
|
31
|
(48.4)
|
22
|
(23.2)
|
|
|
|
IIIA
|
14
|
|
13
|
|
14
|
|
|
|
|
IIIB/IIIC
|
32
|
|
20
|
|
59
|
|
|
|
|
pT stage, n (%)
|
|
|
|
T1/2
|
64
|
(58.2)
|
13
|
(20.3)
|
18
|
(18.9)
|
< 0.001*
|
< 0.001*
|
0.992
|
T3
|
34
|
|
32
|
|
38
|
|
|
|
|
T4
|
12
|
|
19
|
|
39
|
|
|
|
|
pN stage, n (%)
|
< 0.001*
|
< 0.001*
|
0.001*
|
N0
|
59
|
(53.6)
|
14
|
(21.9)
|
4
|
(4.2)
|
|
|
|
N1/2
|
42
|
|
36
|
|
40
|
|
|
|
|
N3
|
9
|
|
14
|
|
51
|
|
|
|
|
ECOG ≥ 2, n (%)
|
13
|
(11.8)
|
4
|
(6.3)
|
8
|
(8.4)
|
0.233
|
0.424
|
0.840
|
HER2-positive, n (%)
|
15
|
(13.6)
|
4
|
(6.3)
|
9
|
(9.5)
|
0.132
|
0.355
|
0.665
|
HER2-therapya, n
|
|
0
|
|
2
|
|
|
|
|
Tumor at GEJ, n (%)
|
7
|
(6.3)
|
3
|
(4.7)
|
9
|
(9.5)
|
0.647
|
0.408
|
0.415
|
Lauren histology, n (%)
|
0.195
|
0.009*
|
0.600
|
Intestinal
|
57
|
(51.8)
|
25
|
(39.1)
|
32
|
(33.7)
|
|
|
|
Diffuse
|
34
|
(30.9)
|
28
|
(43.8)
|
49
|
(51.6)
|
|
|
|
Mixed
|
16
|
|
10
|
|
14
|
|
|
|
|
Othersb
|
3
|
|
1
|
|
0
|
|
|
|
|
Borrmann classification, n (%)
|
|
|
|
|
|
|
0.190
|
< 0.001
|
0.058
|
Infiltrative
|
54
|
(49.1)
|
38
|
(59.4)
|
70
|
(73.7)
|
|
|
|
Non-infiltrative
|
56
|
|
26
|
|
25
|
|
|
|
|
LVI (+), n (%)
|
43
|
(39.1)
|
42
|
(65.6)
|
74
|
(77.9)
|
< 0.001*
|
< 0.001*
|
0.088
|
H. pylori infection, n (%)
|
21
|
(19.1)
|
13
|
(20.3)
|
28
|
(29.5)
|
0.845
|
0.082
|
0.267
|
Median CEA, ng/mL (IQR)
|
2.1
|
(1.3–3.4)
|
1.8
|
(1.3–2.6)
|
2.8
|
(1.5–4.8)
|
0.143
|
0.027*
|
0.002*
|
Median CA19-9, U/mL (IQR)
|
12.6
|
(8.4–26.0)
|
13.7
|
(6.4–24.1)
|
12.7
|
(6.8–32.5)
|
0.669
|
Neoadjuvant therapiesc, n (%)
|
0
|
|
3
|
(4.7)
|
5
|
(5.3)
|
-
|
-
|
0.836
|
Adjuvant chemotherapy, n
|
|
|
|
|
|
|
|
|
|
S-1
|
-
|
|
63
|
|
-
|
|
|
|
|
S-1/nivolumab
|
-
|
|
1
|
|
-
|
|
|
|
|
PFL
|
-
|
|
-
|
|
43
|
|
|
|
|
XELOX/FOLFOX
|
-
|
|
-
|
|
46
|
|
|
|
|
Othersd
|
-
|
|
-
|
|
6
|
|
-
|
-
|
-
|
a. Defined as any frontline or subsequent HER2-directed therapies in patients with HER2-positive disease. Group OBS, XELOX/trastuzumab (n = 1); Group P, XELOX/trastuzumab (n = 1) and lapatinib (n = 1). |
b. Group OBS, lymphoepithelioma (n = 3); Group S-1, adenocarcinoma with neuroendocrine differentiation (n = 1). |
c. Group S-1, DCF (n = 1) and FLOT (n = 2); Group P, pembrolizumab/cisplatin (n = 1), pembrolizumab/cisplatin/capecitabine (n = 1), ECF (n = 1), XELOX (n = 1) and 5-FU/cisplatin-based CCRT (n = 1). |
d. Pembrolizumab/cisplatin/capecitabine (n = 4), pembrolizumab/cisplatin (n = 1) and ECF (n = 1) |
IQR, interquartile range; ECOG, Eastern Cooperative Oncology Group; GEJ, gastroesophageal junction; LVI, lymphovascular invasion; CCRT, concurrent chemoradiotherapy; PFL, cisplatin/fluorouracil/leucovorin; XELOX, capecitabine/oxaliplatin; FOLFOX, oxaliplatin/fluorouracil/leucovorin; DCF, docetaxel/cisplatin/fluorouracil; FLOT, docetaxel/oxaliplatin/fluorouracil/leucovorin; ECF, epirubicin/cisplatin/ fluorouracil. |
PPS matching and post-match characteristics
Two independent matched results are demonstrated in Table 2. A total of 61 and 58 patients were matched in the analysis 1 (S-1 vs. P) and 2 (S-1 vs. OBS), respectively. Both post-match analyses revealed well-balanced characteristics in terms of age, sex, performance status, stage (disease, T, or N stage), and histology classification, with the exception of marginal disproportions in stage IB/II disease in analysis 1 and serum CEA in analysis 2. In Group S-1, all patients had S-1 monotherapy as adjuvant chemotherapy. In Group P, 29/61 patients had cisplatin/5-FU/leucovorin (PFL); 30/61 patients had oxaliplatin-based doublets, and 2/61 patients had pembrolizumab/cisplatin/capecitabine as the selected adjuvant chemotherapies.
Table 2
Post matching characteristics
|
Analysis 1 (1:1)
|
|
Analysis 2 (1:1)
|
|
Group
|
S-1
|
P-based
|
|
S-1
|
OBS
|
|
|
S-1
(n = 61)
|
Platinum-based (n = 61)
|
p
|
S-1
(n = 58)
|
Surgery alone (n = 58)
|
p
|
Age, median (IQR)
|
65.0
|
(55.9–75.3)
|
61.8
|
(55.5–69.4)
|
0.544
|
65.5
|
(57.8–76.0)
|
70.1
|
(60.6–76.2)
|
0.143
|
Male, n (%)
|
31
|
(51.7)
|
33
|
(54.1)
|
0.784
|
29
|
(50.0)
|
31
|
(53.4)
|
0.999
|
Disease stage, n (%)
|
|
0.470
|
|
0.370
|
IB/II
|
25
|
(40.1)
|
19
|
(31.1)
|
|
31
|
(53.4)
|
33
|
(56.9)
|
|
IIIA
|
13
|
|
13
|
|
|
9
|
|
13
|
|
|
IIIB/IIIC
|
23
|
|
29
|
|
|
18
|
|
12
|
|
|
pT stage, n (%)
|
|
0.825
|
|
|
|
|
0.999
|
T1/2
|
12
|
(19.7)
|
14
|
(23.0)
|
|
13
|
(22.4)
|
13
|
(22.4)
|
|
T3/4
|
49
|
|
47
|
|
|
45
|
|
45
|
|
|
pN stage, n (%)
|
|
0.499
|
|
0.980
|
N0
|
11
|
(18.0)
|
7
|
(11.5)
|
|
14
|
(24.1)
|
16
|
(27.6)
|
|
N1/2
|
36
|
|
36
|
|
|
32
|
|
31
|
|
|
N3
|
14
|
|
18
|
|
|
12
|
|
11
|
|
|
ECOG ≥ 2, n (%)
|
4
|
(6.6)
|
5
|
(8.2)
|
0.729
|
4
|
(6.9)
|
8
|
(13.8)
|
0.222
|
HER2-positive, n (%)
|
4
|
(6.6)
|
5
|
(8.2)
|
0.729
|
3
|
(5.2)
|
6
|
(10.3)
|
0.298
|
Tumor at GEJ, n (%)
|
2
|
(3.3)
|
3
|
(4.9)
|
0.995
|
3
|
(5.2)
|
3
|
(5.2)
|
0.999
|
Lauren histology type, n (%)
|
|
0.509
|
|
0.287
|
Intestinal
|
24
|
(39.3)
|
23
|
(37.7)
|
|
23
|
(39.7)
|
28
|
(48.3)
|
|
Diffuse
|
26
|
(42.6)
|
31
|
(50.8)
|
|
25
|
(43.1)
|
19
|
(32.8)
|
|
Mixed
|
10
|
|
7
|
|
|
9
|
|
10
|
|
|
Othersa
|
1
|
|
0
|
|
|
1
|
|
0
|
|
|
Borrmann classification, n (%)
|
|
|
|
|
0.253
|
|
|
|
|
0.699
|
Infiltrative
|
37
|
(60.7)
|
43
|
(70.5)
|
|
38
|
(65.5)
|
36
|
(62.1)
|
|
Non-infiltrative
|
24
|
|
18
|
|
|
20
|
|
22
|
|
|
LVI (+), n (%)
|
42
|
(68.9)
|
49
|
(80.3)
|
0.145
|
31
|
(53.4)
|
36
|
(62.1)
|
0.347
|
Median CEA ng/mL, (IQR)
|
1.9
|
(1.2–2.7)
|
1.9
|
(1.2–3.1)
|
0.535
|
1.9
|
(1.3–2.6)
|
2.1
|
(1.5–4.4)
|
0.072
|
H. pylori infection, n (%)
|
12
|
(19.7)
|
16
|
(26.2)
|
0.389
|
13
|
(22.4)
|
10
|
(17.2)
|
0.485
|
Adjuvant chemotherapy, n (%)
|
|
|
|
S-1
|
61
|
|
-
|
|
|
58
|
|
-
|
|
|
PFL
|
-
|
|
29
|
|
|
-
|
|
-
|
|
|
XELOX/FOLFOX
|
-
|
|
30
|
|
|
-
|
|
-
|
|
|
Othersb
|
-
|
|
2
|
|
|
-
|
|
-
|
|
|
a. Adenocarcinoma with neuroendocrine differentiation (n = 1) |
b. Pembrolizumab/cisplatin/capecitabine (n = 2) |
Survival: S-1 vs. platinum-based doublets
All patients were evaluated in a median follow-up time of 42.5 months. In analysis 1, patients who received adjuvant S-1 had a trend of shorter RFS and OS as compared with platinum-based doublets but it did not reach statistical significance (median RFS/OS: S-1 vs. P, 24.4/37.4 vs. 39.1/62.1 months, HR = 1.12/1.22, p = 0.667/0.514; Fig. 2A and 2B). Stratified by the N stage, the patients in Group P had significantly better RFS and OS in N2-3 disease as compared with Group S-1 (median RFS/OS: S-1 vs. P, 19.3/27.8 vs. 28.5/56.0 months, HR = 2.04/2.29, p = 0.010/0.007; Fig. 2C and 2D). However, the survival differences were not evident in terms of N0-1 disease (5-year RFS/OS rates: S-1 vs. P, 66.5%/67.1% vs. 60.0%/71.0%, HR = 0.65/0.57, p = 0.458/0.479; Fig. 2E and 2F). Stratified by staging, patients with stage IB to IIIA disease had comparable RFS and OS (5-year RFS/OS rates: S-1 vs. P, 62.3%/66.3% vs. 52.8%/71.3%, HR = 0.73/0.60, p = 0.431/0.371; Supplementary 1A and 1B). However, platinum-based doublets were associated with significantly prolonged survival rates in patients with stage IIIB/IIIC disease (median RFS/OS: S-1 vs. P, 12.8/27.0 vs. 28.5/55.6 months, HR = 2.87/2.89, p = 0.001/0.002; Supplementary 1C and 1D). We did not observe significant survival differences based on the S-1, oxaliplatin, or cisplatin-based chemotherapy in the matched population (median RFS: S-1 vs. oxaliplatin vs. cisplatin, 24.4 vs. 24.7 vs. 54.3 months; all p ≥ 0.05 between groups; Supplementary 1E) (median OS: S-1 vs. oxaliplatin vs. cisplatin, 37.4 vs. 55.67 vs. 64.2 months; all p ≥ 0.05 between groups; Supplementary 1F).
Survival: S-1 vs. OBS
In analysis 2, Group S-1 had a comparable survival outcome to that of OBS (median RFS/OS: S-1 vs. OBS, 23.6/37.4 vs. 24.0/64.6 months, HR = 1.17/1.23, p = 0.554/0.499; Fig. 3A and 3B). Conversely, patients with N0-1 disease in Group S-1 had a significantly better RFS and OS (median RFS/OS: S-1 vs. OBS, undefined/37.4 vs. undefined/61.0 months, HR = 0.41/0.27, p = 0.038/0.024; Fig. 3C and 3D), but the survival advantage disappeared in N2-3 disease as compared with OBS (median RFS/OS: S-1 vs. OBS, 16.1/27.8 vs. 16.2/28.9 months, HR = 1.10/1.03, p = 0.778/0.948; Fig. 3E and 3F). Stratified by staging, stage IB to IIIA patients in Group S-1 had a longer RFS and OS (5-yr RFS/OS rates: S-1 vs. OBS, 71.0%/65.4% vs. 50.1%/56.9%, HR = 0.41/1.03, p = 0.041/0.032; Supplementary 2A and 2B). However, dismal survival outcomes were observed in the stage IIIB and IIIC patients, irrespective of whether they received adjuvant S-1 or not (median RFS/OS: S-1 vs. OBS, 12.8/21.0 vs. 11.8/25.9 months, HR = 0.95/1.09, p = 0.871/0.841; Supplementary 2C and 2D).
Adverse events and dosage intensity
We present the adverse events and the grading in Table 3 for the total number of evaluable patients (S-1 monotherapy, n = 61; platinum-based doublets, n = 84). There were no deaths attributable to the adjuvant chemotherapies. Anemia was the most common adverse event in all patients and was significantly higher in the platinum-based doublets as compared with S-1 (grade III/IV anemia: S-1 vs. combination therapy, 4.9% vs. 9.5%, p = 0.028). Grade III/IV thrombocytopenia, renal insufficiency, and neutropenia were prevalent in the platinum-based doublets (thrombocytopenia, 17.9%; renal insufficiency, 4.8%; neutropenia, 3.6%). Scattered cases of severe diarrhea, anorexia, mucositis, and palmar plantar erythrodysesthesia were also observed in patients who received platinum-based doublets. In the S-1 group, except for a few cases with grade III/IV hepatitis (n = 2), thrombocytopenia (n = 1), and neutropenia (n = 1), there were no additional warning toxicities reported. One patient developed S-1-related grade III drug eruptions that were relieved after discontinuation and supportive care. The average dose intensity of S-1 was 40.9 ± 13.6 mg in patients with a body surface area (BSA) < 1.25 m2, 47.4 ± 13.4 mg with a BSA from 1.25 to 1.50 m2, and 58.5 ± 12.4 mg with a BSA > 1.50 m2 per day, which were lower than the suggested dose from the published studies, with an average relative intensity of 73.6% (Supplement 3). 10.9% of the patients required a dose reduction due to intolerable toxicity. The median duration of S-1 was 8.3 months, respectively.
Table 3
Adverse events, according to the treatment
|
S-1
(n = 61)
|
Platinum-based doublets
(n = 84)a
|
|
Events
|
Total, n
|
Grade I/II, n
|
Grade III/IV, n
|
Grade III/IV, (%)
|
Total, n
|
Grade I/II, n
|
Grade III/IV, n
|
Grade III/IV, (%)
|
pb
|
Any events
|
90
|
82
|
8
|
-
|
212
|
173
|
39
|
-
|
-
|
Anemia
|
22
|
19
|
3
|
4.9
|
47
|
39
|
8
|
9.5
|
0.028*
|
Thrombocytopenia
|
7
|
6
|
1
|
1.6
|
36
|
21
|
15
|
17.9
|
< 0.001*
|
Neutropenia
|
3
|
2
|
1
|
1.6
|
12
|
9
|
3
|
3.6
|
0.121
|
Renal insufficiency
|
5
|
5
|
0
|
0
|
17
|
13
|
4
|
4.8
|
0.046*
|
Hepatitis
|
13
|
11
|
2
|
3.3
|
15
|
13
|
2
|
2.3
|
0.759
|
Anorexia
|
22
|
22
|
0
|
0
|
37
|
33
|
4
|
4.8
|
0.427
|
Diarrhea
|
9
|
9
|
0
|
0
|
8
|
7
|
1
|
1.2
|
0.481
|
Vomiting
|
6
|
6
|
0
|
0
|
29
|
29
|
0
|
0
|
0.001*
|
Mucositis
|
1
|
1
|
0
|
0
|
4
|
3
|
1
|
1.2
|
0.398
|
PPE
|
0
|
0
|
0
|
0
|
6
|
5
|
1
|
1.2
|
-
|
Skin eruptions
|
2
|
1
|
1
|
1.6
|
1
|
0
|
1
|
1.2
|
0.573
|
a. XELOX (n = 41), PFL (n = 38) and FOLFOX (n = 5). |
b. Comparisons with incidence of all events. |
PPE, palmar plantar erythrodysesthesia |
Subgroup analysis
The subgroup analyses for RFS and OS are shown in Fig. 4 using forest plotting. In the total population, we did not observe significant differences among those treated with adjuvant S-1 or platinum-based doublets. In the subgroups, patients with stage III, T3-4, or N2-3 disease had a longer RFS on platinum-based doublets. Female gender, stage III or N2-3 disease were associated with favorable OS when treated with platinum-based doublets. To the contrary, H. pylori infection correlated with a survival benefit in patients who received S-1.
Factors associated with adjuvant S-1
We evaluated the factors associated with selecting adjuvant S-1 versus platinum-based doublets using a binary logistic regression, and the results are shown in Table 4. We observed that low serum CEA, N0-2 disease, increased patient age, low serum CA19-9, and early stages were associated factors in the univariate regression. Following multivariate adjustments, low serum CEA, N0-2 disease, and increased patient age were significant factors for adjuvant S-1 rather than platinum-based doublets.
Table 4
Associated factors for the selection of adjuvant S-1
|
|
Univariate regression
|
Multivariate regression
|
n
|
OR
|
95% CI
|
p
|
OR
|
95% CI
|
p
|
CEA
≥5 ng/mL
<5 ng/mL
|
40
119
|
1.00
5.39
|
2.11–13.79
|
< 0.001*
|
3.93
|
1.44–10.7
|
0.008*
|
N stage
N3
N1-2
N0
|
65
75
19
|
1.00
3.36
10.2
|
1.60–7.09
3.13–33.20
|
0.001*
< 0.001*
|
2.16
5.48
|
1.02–5.01
1.21–24.90
|
0.044*
0.028*
|
Age, increased by 1 year
|
159
|
1.03
|
0.99–1.05
|
0.058
|
1.04
|
1.01–1.07
|
0.041*
|
CA199
≥47 IU/mL
<47 IU/mL
|
23
136
|
1.00
3.75
|
1.21–11.61
|
0.022*
|
2.82
|
0.77–10.42
|
0.118
|
AJCC stage
III
I/II
|
107
52
|
1.00
3.31
|
1.66–6.60
|
0.001*
|
1.74
|
0.69–4.41
|
0.241
|
Gender
Male
Female
|
90
69
|
1.00
1.57
|
0.83–2.97
|
0.169
|
|
|
|
H. pylori infection
Presence
Absence
|
41
118
|
1.00
1.64
|
0.77–3.48
|
0.197
|
|
|
|
LVI
Presence
Absence
|
118
41
|
1.00
1.60
|
0.78–3.28
|
0.198
|
|
|
|
Lauren classification
Intestinal
Diffuse
Mixed
|
55
77
27
|
1.00
0.73
1.01
|
0.36–1.47
0.39–2.59
|
0.381
0.991
|
|
|
|
T stage
T1-2
T3
T4
|
31
70
58
|
1.17
0.68
|
0.50–2.74
0.27–1.66
|
0.725
0.391
|
|
|
|
Tumor differentiation
Well/Moderate
Poor
|
44
115
|
1.00
0.74
|
0.37–1.50
|
0.409
|
|
|
|
ECOG
<2
≥2
|
147
12
|
1.00
1.38
|
0.40–4.79
|
0.613
|
|
|
|
HER2 status
Positive
Negative
|
13
146
|
1.00
1.11
|
0.43–2.85
|
0.829
|
|
|
|
Borrmann classification
Non-infiltrative
Infiltrative
|
41
118
|
1.00
1.07
|
0.52–2.22
|
0.852
|
|
|
|
OR, odds ratio |