Cohort description and patient characteristics
A total of 11,967 patients were initially selected from the SEER database in this study. After the exclusion of patients with a lack of information concerning basic demographic data, clinicopathological features and survival time, 10,968 patients remained for further investigation. To accurately analyze surgical and chemotherapeutic effects in GC, patients who had received radiotherapy were excluded to keep the intervention methods comparable. Ultimately, 9,306 patients were selected for statistical analysis. The selection procedure is shown in Fig. 1.
This pool was divided into four groups according to therapeutic regimens: 1) patients who received neither surgery nor CT (n = 4,674); 2) patients who received surgery alone (n = 1,097); 3) patients who received CT alone (n = 2,969); and 4) patients who received both surgery and CT (n = 566). The percentages of the groups were 50.23%, 11.79%, 31.90% and 6.08%, respectively. The clinical features of this cohort are presented in Table 1. There were significant differences between the four intervention groups regarding sex, age, ethnicity, marital status, median income, insurance status, T stage, N stage, histologic type and grade (P < 0.05).
Table 1
Characteristics of old patients of gastric cancer at Ⅳ stage treated with four different intervention groups.
Clinicopathologic Characteristics | No surgery or CT | Surgery alone | CT alone | Surgery plus CT | P-value |
Gender | | | | | < 0.001 |
Male | 2767 (59.2%) | 593 (54.1%) | 2029 (68.3%) | 330 (58.3%) | |
Female | 1907 (40.8%) | 504 (45.9%) | 940 (31.7%) | 236 (41.7%) | |
Age | | | | | < 0.001 |
65–69 | 754 (16.1%) | 204 (18.6%) | 949 (32.0%) | 190 (33.6%) | |
70–74 | 862 (18.4%) | 238 (21.7%) | 776 (26.1%) | 164 (29.0%) | |
75–79 | 1053 (22.5%) | 256 (23.3%) | 687 (23.1%) | 121 (21.4%) | |
80–84 | 1040 (22.3%) | 241 (22.0%) | 378 (12.7%) | 65 (11.5%) | |
༞85 | 965 (20.6%) | 158 (14.4%) | 179 ( 6.0%) | 26 ( 4.6%) | |
Ethnicity | | | | | < 0.001 |
White | 3389 (72.5%) | 764 (69.6%) | 2258 (76.1%) | 408 (72.1%) | |
Black | 653 (14.0%) | 137 (12.5%) | 362 (12.2%) | 64 (11.3%) | |
Other1 | 632 (13.5%) | 196 (17.9%) | 349 (11.8%) | 94 (16.6%) | |
Region | | | | | 0.905 |
Urban | 765 (16.4%) | 174 (15.9%) | 454 (15.3%) | 85 (15.0%) | |
County | 3817 (81.7%) | 903 (82.3%) | 2454 (82.7%) | 469 (82.9%) | |
Rural | 92 ( 2.0%) | 20 ( 1.8%) | 61 ( 2.1%) | 12 ( 2.1%) | |
Marital status | | | | | < 0.001 |
Married | 2298 (49.2%) | 622 (56.7%) | 1985 (66.9%) | 373 (65.9%) | |
Single or divorced2 | 945 (20.2%) | 166 (15.1%) | 510 (17.2%) | 96 (17.0%) | |
Widowed | 1431 (30.6%) | 309 (28.2%) | 474 (16.0%) | 97 (17.1%) | |
Median income | | | | | 0.046 |
1(lowest) | 1193 (25.5%) | 271 (24.7%) | 723 (24.4%) | 138 (24.4%) | |
2 | 1266 (27.1%) | 297 (27.1%) | 714 (24.0%) | 140 (24.7%) | |
3 | 1144 (24.5%) | 267 (24.3%) | 761 (25.6%) | 150 (26.5%) | |
4(highest) | 1071 (22.9%) | 262 (23.9%) | 771 (26.0%) | 138 (24.4%) | |
Insurance | | | | | < 0.001 |
Insured3 | 2719 (58.2%) | 531 (48.4%) | 2066 (69.6%) | 345 (61.0%) | |
Medicaid | 533 (11.4%) | 129 (11.8%) | 246 ( 8.3%) | 52 ( 9.2%) | |
Uninsured | 54 ( 1.2%) | 7 ( 0.6%) | 25 ( 0.8%) | 11 ( 1.9%) | |
Unknown | 1368 (29.3%) | 430 (39.2%) | 632 (21.3%) | 158 (27.9%) | |
T stage | | | | | < 0.001 |
T0 | 22 ( 0.5%) | 4 ( 0.4%) | 14 ( 0.5%) | 4 ( 0.7%) | |
T1 | 930 (19.9%) | 70 ( 6.4%) | 612 (20.6%) | 52 ( 9.2%) | |
T24 | 436 ( 9.3%) | 333 (30.4%) | 427 (14.4%) | 165 (29.2%) | |
T3 | 132 ( 2.8%) | 296 (27.0%) | 122 ( 4.1%) | 143 (25.3%) | |
T4 | 803 (17.2%) | 287 (26.2%) | 435 (14.7%) | 130 (23.0%) | |
Tx | 2351 (50.3%) | 107 ( 9.8%) | 1359 (45.8%) | 72 (12.7%) | |
N stage | | | | | < 0.001 |
N0 | 1839 (39.3%) | 245 (22.3%) | 1069 (36.0%) | 134 (23.7%) | |
N1 | 1132 (24.2%) | 394 (35.9%) | 1037 (34.9%) | 223 (39.4%) | |
N2 | 56 ( 1.2%) | 225 (20.5%) | 80 ( 2.7%) | 106 (18.7%) | |
N3 | 20 ( 0.4%) | 114 (10.4%) | 37 ( 1.2%) | 52 ( 9.2%) | |
Nx | 1627 (34.8%) | 119 (10.8%) | 746 (25.1%) | 51 ( 9.0%) | |
Histologic type | | | | | < 0.001 |
Epithelial | 372 ( 8.0%) | 30 ( 2.7%) | 137 ( 4.6%) | 11 ( 1.9%) | |
Adenocarcinoma | 3434 (73.5%) | 785 (71.6%) | 2255 (76.0%) | 401 (70.8%) | |
Other5 | 868 (18.6%) | 282 (25.7%) | 577 (19.4%) | 154 (27.2%) | |
Grade | | | | | < 0.001 |
Well | 104 ( 2.2%) | 25 ( 2.3%) | 67 ( 2.3%) | 8 ( 1.4%) | |
Moderately | 864 (18.5%) | 199 (18.1%) | 610 (20.5%) | 123 (21.7%) | |
Poorly | 2340 (50.1%) | 722 (65.8%) | 1584 (53.4%) | 345 (61.0%) | |
Undifferentiated | 79 ( 1.7%) | 29 ( 2.6%) | 49 ( 1.7%) | 11 ( 1.9%) | |
Unknown | 1287 (27.5%) | 122 (11.1%) | 659 (22.2%) | 79 (14.0%) | |
CT chemotherapy |
1 American Indian/AK Native, Asian/Pacific Islander |
2 Single (never married), Divorced, Separated and Unmarried or Domestic Partner |
3 Insured, and Insured/No specifics |
4 T2a, T2b and T2NOS |
5 Squamous and complex epithelial gastric cancer |
Univariate and multivariate analysis of OS and CSS prognostic factors
To screen prognostic factors for OS and CSS, univariate analysis was first conducted. The results showed that age, race, marital status, insurance status, median income, N stage, histologic type and therapeutic methods were statistically associated with the OS of patients. Additionally, a race of American Indian/AK Native, Asian/Pacific Islander, the county region, being uninsured, T2 stage, poorly differentiated status and unknown grade were also correlated with OS (P < 0.05, Table 2). In addition, sex, age, black race, being married, unknown insurance status, N stage, histologic type, grade, and therapeutic method were significantly associated with CSS (Table 3). Furthermore, multivariate analysis was employed to screen out independent prognostic factors of OS and CSS. The modified model indicated that age, American Indian/AK Native, Asian/Pacific Islander race, being single (never married), being divorced, being separated and being unmarried or being with a domestic partner, the second lowest income bracket, therapeutic method, insurance status, T2 stage, and grade were independent prognostic factors for OS. The independent prognostic factors for CSS were similar to those for OS, including age, black race, the second lowest income bracket, unknown insurance status, N1 stage, grade and therapeutic method.
Table 2
Univariate and multivariate analysis of OS for old patients at Ⅳ stage with gastric cancer.
Clinicopathologic Characteristics | Total (n) | Univariate | P-value | Multivariate | P-value |
| HR | HR |
Gender | | | | | |
Male | 5719 (61.5%) | Reference | | Reference | |
Female | 3587 (38.5%) | 1.03 (0.98, 1.07) | 0.2170 | 0.91 (0.87, 0.96) | 0.0002 |
Age | | | | | |
65–69 | 2097 (22.5%) | Reference | | Reference | |
70–74 | 2040 (21.9%) | 1.13 (1.06, 1.20) | 0.0002 | 1.09 (1.02, 1.16) | 0.0086 |
75–79 | 2117 (22.7%) | 1.27 (1.20, 1.36) | < 0.0001 | 1.14 (1.07, 1.21) | < 0.0001 |
80–84 | 1724 (18.5%) | 1.44 (1.35, 1.54) | < 0.0001 | 1.14 (1.06, 1.22) | 0.0002 |
༞85 | 1328 (14.3%) | 1.79 (1.67, 1.93) | < 0.0001 | 1.27 (1.18, 1.38) | < 0.0001 |
Ethnicity | | | | | |
White | 6819 (73.3%) | Reference | | Reference | |
Black | 1216 (13.1%) | 1.06 (0.99, 1.12) | 0.0932 | 1.04 (0.97, 1.11) | 0.2450 |
Other1 | 1271 (13.7%) | 0.92 (0.87, 0.98) | 0.0128 | 0.91 (0.86, 0.98) | 0.0066 |
Region | | | | | |
Urban | 1478 (15.9%) | Reference | | Reference | |
County | 7643 (82.1%) | 0.93 (0.88, 0.99) | 0.0171 | 0.98 (0.92, 1.05) | 0.5474 |
Rural | 185 (2.0%) | 0.92 (0.78, 1.08) | 0.3223 | 0.97 (0.83, 1.14) | 0.7207 |
Marital status | | | | | |
Married | 5278 (56.7%) | Reference | | Reference | |
Single or divorced2 | 1717 (18.5%) | 1.14 (1.08, 1.21) | < 0.0001 | 1.07 (1.00, 1.13) | 0.0360 |
Widowed | 2311 (24.8%) | 1.27 (1.21, 1.34) | < 0.0001 | 1.03 (0.97, 1.09) | 0.3237 |
Median income | | | | | |
1(lowest) | 2325 (25.0%) | Reference | | Reference | |
2 | 2417 (26.0%) | 0.92 (0.87, 0.98) | 0.0065 | 0.88 (0.83, 0.95) | 0.0003 |
3 | 2322 (25.0%) | 0.96 (0.90, 1.01) | 0.1307 | 0.97 (0.91, 1.03) | 0.3511 |
4(highest) | 2242 (24.1%) | 0.91 (0.85, 0.96) | 0.0016 | 0.94 (0.88, 1.01) | 0.1145 |
Insurance | | | | | |
Insured3 | 5661 (60.8%) | Reference | | Reference | |
Medicaid | 960 (10.3%) | 1.03 (0.95, 1.10) | 0.4990 | 0.93 (0.86, 1.00) | 0.0415 |
Uninsured | 97 (1.0%) | 0.80 (0.63, 1.01) | 0.0655 | 0.78 (0.61, 0.99) | 0.0373 |
Unknown | 2588 (27.8%) | 1.14 (1.08, 1.19) | < 0.0001 | 1.04 (0.99, 1.09) | 0.1116 |
T stage | | | | | |
T0 | 44 (0.5%) | Reference | | Reference | |
T1 | 1664 (17.9%) | 0.85 (0.62, 1.17) | 0.3203 | 0.80 (0.58, 1.09) | 0.1620 |
T24 | 1361 (14.6%) | 0.66 (0.48, 0.90) | 0.0086 | 0.72 (0.52, 0.99) | 0.0427 |
T3 | 693 (7.4%) | 0.67 (0.48, 0.92) | 0.0125 | 0.77 (0.56, 1.07) | 0.1153 |
T4 | 1655 (17.8%) | 0.92 (0.67, 1.26) | 0.6119 | 0.89 (0.65, 1.22) | 0.4662 |
Tx | 3889 (41.8%) | 0.95 (0.70, 1.30) | 0.7652 | 0.84 (0.62, 1.16) | 0.2917 |
N stage | | | | | |
N0 | 3287 (35.3%) | Reference | | Reference | |
N1 | 2786 (29.9%) | 0.84 (0.80, 0.89) | < 0.0001 | 0.95 (0.90, 1.01) | 0.0988 |
N2 | 467 (5.0%) | 0.71 (0.64, 0.78) | < 0.0001 | 0.96 (0.86, 1.07) | 0.4568 |
N3 | 223 (2.4%) | 0.73 (0.64, 0.84) | < 0.0001 | 0.99 (0.85, 1.15) | 0.8989 |
Nx | 2543 (27.3%) | 1.17 (1.11, 1.24) | < 0.0001 | 1.04 (0.99, 1.11) | 0.1335 |
Histologic type | | | | | |
Epithelial | 550 (5.9%) | Reference | | Reference | |
Adenocarcinoma | 6875 (73.9%) | 0.75 (0.68, 0.82) | < 0.0001 | 0.95 (0.87, 1.05) | 0.3221 |
Other5 | 1881 (20.2%) | 0.78 (0.71, 0.86) | < 0.0001 | 1.01 (0.91, 1.12) | 0.8605 |
Grade | | | | | |
Well | 204 (2.2%) | Reference | | Reference | |
Moderately | 1796 (19.3%) | 1.08 (0.93, 1.26) | 0.3234 | 1.23 (1.05, 1.44) | 0.0093 |
Poorly | 4991 (53.6%) | 1.26 (1.09, 1.47) | 0.0023 | 1.51 (1.30, 1.76) | < 0.0001 |
Undifferentiated | 168 (1.8%) | 1.40 (1.13, 1.74) | 0.0022 | 1.65 (1.33, 2.05) | < 0.0001 |
Unknown | 2147 (23.1%) | 1.42 (1.22, 1.66) | < 0.0001 | 1.44 (1.23, 1.68) | < 0.0001 |
Therapeutic methods | | | | | |
No surgery or CT | 4674 (50.2%) | Reference | | Reference | |
Surgery alone | 1097 (11.8%) | 0.50 (0.47, 0.54) | < 0.0001 | 0.55 (0.50, 0.59) | < 0.0001 |
CT alone | 2969 (31.9%) | 0.36 (0.34, 0.38) | < 0.0001 | 0.37 (0.35, 0.39) | < 0.0001 |
Surgery plus CT | 566 (6.1%) | 0.25 (0.23, 0.28) | < 0.0001 | 0.28 (0.25, 0.31) | < 0.0001 |
CT chemotherapy; OS overall survival. |
1 American Indian/AK Native, Asian/Pacific Islander |
2 Single (never married), Divorced, Separated and Unmarried or Domestic Partner |
3 Insured, and Insured/No specifics |
4 T2a, T2b and T2NOS |
5 Squamous and complex epithelial gastric cancer |
Table 3
Univariate and multivariate analysis of CSS for old patients at Ⅳ stage with gastric cancer.
Clinicopathologic Characteristics | Total (n) | Univariate | P-value | Multivariate | P-value |
| HR | HR |
Gender | | | | | |
Male | 5719 (61.5%) | Reference | | Reference | |
Female | 3587 (38.5%) | 1.12 (1.07, 1.18) | < 0.0001 | 0.99 (0.94, 1.05) | 0.7952 |
Age | | | | | |
65–69 | 2097 (22.5%) | Reference | | Reference | |
70–74 | 2040 (21.9%) | 1.15 (1.07, 1.24) | 0.0001 | 1.11 (1.03, 1.19) | 0.0056 |
75–79 | 2117 (22.7%) | 1.30 (1.21, 1.40) | < 0.0001 | 1.16 (1.07, 1.25) | 0.0001 |
80–84 | 1724 (18.5%) | 1.48 (1.37, 1.59) | < 0.0001 | 1.17 (1.08, 1.26) | 0.0001 |
༞85 | 1328 (14.3%) | 1.87 (1.72, 2.03) | < 0.0001 | 1.32 (1.21, 1.44) | < 0.0001 |
Ethnicity | | | | | |
White | 6819 (73.3%) | Reference | | Reference | |
Black | 1216 (13.1%) | 1.13 (1.05, 1.21) | 0.0008 | 1.13 (1.05, 1.22) | 0.0013 |
Other1 | 1271 (13.7%) | 1.05 (0.98, 1.13) | 0.1491 | 1.03 (0.96, 1.10) | 0.4667 |
Region | | | | | |
Urban | 1478 (15.9%) | Reference | | Reference | |
County | 7643 (82.1%) | 0.96 (0.90, 1.03) | 0.2669 | 0.99 (0.92, 1.07) | 0.8013 |
Rural | 185 (2.0%) | 0.92 (0.77, 1.11) | 0.4015 | 0.98 (0.81, 1.18) | 0.8390 |
Marital status | | | | | |
Married | 5278 (56.7%) | Reference | | Reference | |
Single or divorced2 | 1717 (18.5%) | 1.11 (1.04, 1.18) | 0.0020 | 1.01 (0.94, 1.08) | 0.7566 |
Widowed | 2311 (24.8%) | 1.31 (1.24, 1.39) | < 0.0001 | 1.01 (0.95, 1.08) | 0.7529 |
Median income | | | | | |
1(lowest) | 2325 (25.0%) | Reference | | Reference | |
2 | 2417 (26.0%) | 0.95 (0.89, 1.02) | 0.1711 | 0.90 (0.84, 0.97) | 0.0089 |
3 | 2322 (25.0%) | 0.97 (0.90, 1.04) | 0.3368 | 0.98 (0.91, 1.06) | 0.5899 |
4(highest) | 2242 (24.1%) | 0.94 (0.88, 1.01) | 0.0705 | 0.96 (0.88, 1.04) | 0.3002 |
Insurance | | | | | |
Insured3 | 5661 (60.8%) | Reference | | Reference | |
Medicaid | 960 (10.3%) | 1.05 (0.97, 1.14) | 0.2052 | 0.93 (0.86, 1.01) | 0.0913 |
Uninsured | 97 (1.0%) | 0.87 (0.67, 1.13) | 0.3115 | 0.87 (0.67, 1.12) | 0.2807 |
Unknown | 2588 (27.8%) | 1.19 (1.13, 1.26) | < 0.0001 | 1.07 (1.02, 1.14) | 0.0108 |
T stage | | | | | |
T0 | 44 (0.5%) | Reference | | Reference | |
T1 | 1664 (17.9%) | 1.10 (0.73, 1.67) | 0.6378 | 1.05 (0.69, 1.59) | 0.8099 |
T24 | 1361 (14.6%) | 0.92 (0.61, 1.39) | 0.6856 | 1.03 (0.68, 1.56) | 0.8847 |
T3 | 693 (7.4%) | 0.94 (0.62, 1.43) | 0.7715 | 1.10 (0.72, 1.68) | 0.6490 |
T4 | 1655 (17.8%) | 1.32 (0.87, 1.99) | 0.1919 | 1.27 (0.84, 1.93) | 0.2512 |
Tx | 3889 (41.8%) | 1.27 (0.84, 1.91) | 0.2582 | 1.13 (0.75, 1.71) | 0.5540 |
N stage | | | | | |
N0 | 3287 (35.3%) | Reference | | Reference | |
N1 | 2786 (29.9%) | 0.82 (0.77, 0.87) | < 0.001 | 0.92 (0.86, 0.98) | 0.0069 |
N2 | 467 (5.0%) | 0.71 (0.63, 0.80) | < 0.001 | 0.91 (0.80, 1.04) | 0.1536 |
N3 | 223 (2.4%) | 0.76 (0.65, 0.89) | < 0.001 | 0.96 (0.82, 1.14) | 0.6534 |
Nx | 2543 (27.3%) | 1.18 (1.11, 1.25) | < 0.001 | 1.05 (0.99, 1.12) | 0.1255 |
Histologic type | | | | | |
Epithelial | 550 (5.9%) | Reference | | Reference | |
Adenocarcinoma | 6875 (73.9%) | 0.75 (0.67, 0.83) | < 0.0001 | 0.97 (0.87, 1.07) | 0.5200 |
Other5 | 1881 (20.2%) | 0.83 (0.74, 0.93) | 0.0010 | 1.06 (0.95, 1.19) | 0.2833 |
Grade | | | | | |
Well | 204 (2.2%) | Reference | | Reference | |
Moderately | 1796 (19.3%) | 1.09 (0.91, 1.31) | 0.3522 | 1.24 (1.03, 1.49) | 0.0218 |
Poorly | 4991 (53.6%) | 1.37 (1.15, 1.63) | 0.0005 | 1.60 (1.34, 1.91) | < 0.0001 |
Undifferentiated | 168 (1.8%) | 1.50 (1.17, 1.92) | 0.0013 | 1.75 (1.37, 2.25) | < 0.0001 |
Unknown | 2147 (23.1%) | 1.50 (1.25, 1.79) | < 0.0001 | 1.49 (1.24, 1.79) | < 0.0001 |
Therapeutic methods | | | | | |
No surgery or CT | 4674 (50.2%) | Reference | | Reference | |
Surgery alone | 1097 (11.8%) | 0.51 (0.47, 0.55) | < 0.0001 | 0.53 (0.49, 0.58) | < 0.0001 |
CT alone | 2969 (31.9%) | 0.34 (0.32, 0.36) | < 0.0001 | 0.36 (0.34, 0.39) | 0.0001 |
Surgery plus CT | 566 (6.1%) | 0.27 (0.24, 0.30) | < 0.0001 | 0.29 (0.26, 0.33) | < 0.0001 |
CT chemotherapy; CSS cancer-specific survival. |
1 American Indian/AK Native, Asian/Pacific Islander |
2 Single (never married), Divorced, Separated and Unmarried or Domestic Partner |
3 Insured, and Insured/No specifics |
4 T2a, T2b and T2NOS |
5 Squamous and complex epithelial gastric cancer |
Survival outcomes of different therapeutic methods
Kaplan-Meier survival analysis was employed to compare survival time and accumulative hazards of OS and CSS between the four intervention regimens (Fig. 2). The results showed that the patients who received surgery plus CT obtained the best survival benefit [(HR of OS = 0.28, 95% CI = 0.25–0.31, P < 0.0001) (HR of CSS = 0.29, 95% CI = 0.26–0.33, P < 0.0001)], while patients who received neither method obtained the worst survival benefit. The survival of the CT alone group [(HR of OS = 0.37, 95% CI = 0.35–0.39, P < 0.0001) (HR of CSS = 0.36, 95% CI = 0.34–0.39, P = 0.0001)] was better than that of the surgery alone group [(HR of OS = 0.55, 95% CI = 0.50–0.59, P < 0.0001) (HR of CSS = 0.53, 95% CI = 0.49–0.58, P < 0.0001)].
Stratified analysis for survival of elderly patients with stage IV disease
According to the univariate and multivariate analyses of OS and CSS, age and grade were important prognostic factors with gradient changes in terms of survival. Therefore, stratified analysis was conducted to investigate the effects of age and grade on OS (Figs. 3, 4) and CSS (Figs. 5, 6) by dividing the patients into four therapeutic groups to explore suitable subpopulations for surgery and CT. The hazards of OS and CSS after age stratification were similar to those identified in the unstratified results (Additional file 1, 2). The Cox proportional hazard model showed that the OS and CSS gaps between the CT alone group and surgery alone group were increasingly narrowed with increasing age. When patients were stratified according to the “Undifferentiated” grade parameter (Additional file 3, 4), the best OS and CSS values were shown in the surgery alone group [(HR of OS = 0.32, 95% CI = 0.16–0.63, P = 0.0009) (HR of CSS = 0.26, 95% CI = 0.12–0.56, P = 0.0006)]. Notably, the best CSS benefit with surgery plus CT treatment was shown in the well-differentiated group [(HR of CSS = 0.14, 95% CI = 0.04–0.56, P = 0.0049)], which decreased the CSS hazard ratio by at least 10% in comparison with the hazard ratios seen in the other stratified group.