Patient selection and PSM results
Between 2013 and 2019, a total of 180 patients with gastric carcinomas containing neuroendocrine cancer components underwent radical resection at Peking University Cancer Hospital, of which 55 were pure NEC and the remaining 125 were mixed tumors. Finally, 54 patients were included in our study (NEC: 15, < 30%G-HMiNEN: 15, G-HMiNEN: 20, > 70%G-HMiNEN: 4) (Fig. 1). Of included patients, a total of 22 patients received neoadjuvant therapy. We also randomly selected 600 patients with gastric adenocarcinoma who underwent radical surgery and followed up for more than 5 years. Among them, 381 patients received neoadjuvant therapy, and the remaining 228 patients were treated with surgery directly.
Prior to the OS comparison, PSM was performed to ensure that there were no significant statistical differences in patient gender, age, tumor location, tumor size, pathological staging, neoadjuvant chemotherapy, and adjuvant chemotherapy between the two groups. Details before and after the match are shown in Table 1.
Table 1
Patient characteristics before and after propensity matching in all patients with NEC components
Patient Characteristics | Unmatched comparison | | Matched comparison | |
Patients with NEC components (n = 54) | Gastric adenocarcinoma (n = 600) | P value | | Patients with NEC components (n = 45) | Gastric adenocarcinoma (n = 146) | P value | |
Age (year), mean ± SD | 64.5 ± 8.5 | 57.4 ± 10.9 | < 0.001 | | 63.0 ± 8.3 | 62.3 ± 9.0 | 0.624 | |
Gender (male/female) | 42/12 | 504/96 | 0.238 | | 37/8 | 122/24 | 0.833 | |
BMI, mean ± SD | 24.0 ± 3.8 | 23.5 ± 3.4 | 0.333 | | 23.6 ± 3.9 | 23.6 ± 3.5 | 0.937 | |
Neoadjuvant therapy | | | < 0.001 | | | | 0.249 | |
Yes | 19 (35.2) | 372 (62.0) | | | 19 (42.2) | 76 (52.1) | | |
No | 35 (64.8) | 228 (38.0) | | | 26 (57.8) | 70 (47.9) | | |
Adjuvant therapy | | | | | | | 0.404 | |
Yes | 47 (87.0) | 449 (74.8) | 0.045 | | 38 (84.4) | 115 (78.8) | | |
No | 7 (13.0) | 151 (25.2) | | | 7 (15.6) | 31 (21.2) | | |
Tumor location | | | < 0.001 | | | | 0.802 | |
Upper third | 36 (66.7) | 224 (37.3) | | | 32 (71.1) | 94 (64.4) | | |
Middle third | 13 (24.1) | 66 (11.0) | | | 8 (17.8) | 33 (22.6) | | |
Lower third | 5 (9.3) | 294 (49.0) | | | 5 (11.1) | 19 (13.0) | | |
Entire | 0 (0.0) | 16 (2.7) | | | 0 (0.0) | 0 (0.0) | | |
Tumor size | | | 0.102 | | | | 0.409 | |
< 5 cm | 32 (59.3) | 420 (70.0) | | | 29 (63.4) | 102 (69.9) | | |
≥5 cm | 22 (40.7) | 180 (30.0) | | | 16 (25.6) | 44 (30.1) | | |
Type of gastrectomy | | | < 0.001 | | | | 0.104 | |
Total gastrectomy | 44 (81.5) | 243 (40.5) | | | 36 (80.0) | 92 (63.0) | | |
Distal gastrectomy | 5 (9.3) | 292 (48.7) | | | 5 (11.1) | 28 (19.2) | | |
Proximal gastrectomy | 5 (9.3) | 65 (10.8) | | | 4 (8.9) | 26 (17.8) | | |
Surgical procedure | | | 0.002 | | | | < 0.001 | |
Open | 54 (100.0) | 510 (85.0) | | | 45 (100.0) | 116 (79.5) | | |
Laparoscopic | 0 (0.0) | 90 (15.0) | | | 0 (0.0) | 30 (20.5) | | |
T stage | | | < 0.001 | | | | < 0.001 | |
T1 | 2 (3.7) | 186 (28.0) | | | 1 (2.2) | 32 (21.9) | | |
T2 | 10 (18.5) | 54 (9.0) | | | 9 (20.0) | 8 (5.5) | | |
T3 | 24 (44.4) | 141 (23.5) | | | 20 (44.4) | 44 (30.1) | | |
T4 | 18 (33.3) | 237 (39.5) | | | 15 (33.3) | 62 (42.5) | | |
N stage | | | 0.017 | | | | 0.228 | |
N0 | 15 (27.8) | 268 (44.7) | | | 14 (41.1) | 54 (37.0) | | |
N1 | 12 (22.2) | 94 (15.7) | | | 10 (22.2) | 25 (17.1) | | |
N2 | 15 (27.8) | 88 (14.7) | | | 12 (26.7) | 23 (15.8) | | |
N3 | 12 (22.2) | 150 (25.0) | | | 9 (20.0) | 44 (30.1) | | |
TNM stage | | | 0.033 | | | | 0.629 | |
I | 8 (14.8) | 188 (31.3) | | | 8 (17.8) | 35 (24.0) | | |
II | 17 (31.5) | 135 (22.5) | | | 13 (28.9) | 35 (24.0) | | |
III | 29 (53.7) | 277 (46.2) | | | 24 (53.3) | 76 (52.1) | | |
NEC: neuroendocrine carcinoma; Patients with NEC components: NEC, < 30% high grade MiNEN, high grade MiNEN and > 70% high grade MiNEN |
Comparison of OS between all patients with NEC components and patients with gastric adenocarcinoma
PSM was performed on all patients with NEC components, including pure NEC patients and mixed tumor patients (Table 1). As a result, 45 patients with NEC components and 146 patients with gastric adenocarcinoma were matched. The proportion of patients receiving neoadjuvant therapy was 42.2% and 52.1% (Table 1, p = 0.249), respectively, with no statistical difference. Apparently, patients with NEC components had a poorer OS than that of gastric adenocarcinoma (Fig. 2a, p = 0.0424). To investigate whether neoadjuvant therapy had an effect on OS, patients in the surgical and neoadjuvant groups were analyzed separately (Supplement Table 1–2). The results showed that the OS of patients with NEC components was still poorer than that of gastric adenocarcinomas in the surgical group (Fig. 2b, p = 0.0347). While there was no significant difference between the patients receiving neoadjuvant therapy (Fig. 2c, p = 0.5396).
Table 2
Univariate and multivariate analyses of survival after PSM in mixed tumor surgical group
Patient Characteristics | Univariate analysis | | Multivariate analysis | |
HR | 95% CI | P value | | HR | 95% CI | P value | |
Age (year) | 0.961 | 0.925–0.998 | 0.039 | | 1.022 | 0.976–1.069 | 0.345 | |
Gender | | | 0.643 | | | | | |
female | 1 | 1 | | | | | | |
male | 0.807 | 0.333–1.956 | | | | | | |
BMI | 1.052 | 0.952–1.162 | 0.324 | | | | | |
Adjuvant therapy | | | 0.001 | | | | 0.139 | |
Yes | 1 | 1 | | | 1 | 1 | | |
No | 12.324 | 2.939–51.666 | | | 15.936 | 0.405-626.747 | | |
Tumor location | | | 0.263 | | | | | |
Upper third | 1 | 1 | | | | | | |
Middle third | 1.107 | 0.475–2.581 | | | | | | |
Lower third | 0.384 | 0.115–1.281 | | | | | | |
Tumor size | | | < 0.001 | | | | 0.005 | |
< 5 cm | 1 | 1 | | | 1 | 1 | | |
≥5 cm | 4.296 | 2.142–8.615 | | | 3.438 | 1.460–8.094 | | |
Tumor category | | | 0.024 | | | | 0.007 | |
Gastric adenocarcinoma | 1 | 1 | | | 1 | 1 | | |
<30%G-HMiNEN + G-HMiNEN + > 70%G-HMiNEN | 2.273 | 1.117–4.626 | | | 3.181 | 1.370–7.385 | | |
Type of gastrectomy | | | 0.120 | | | | | |
Total gastrectomy | 1 | 1 | | | | | | |
Distal gastrectomy | 0.220 | 0.052–0.929 | | | | | | |
Proximal gastrectomy | 0.911 | 0.373–2.223 | | | | | | |
Surgical procedure | | | 0.499 | | | | | |
Open | 1 | 1 | | | | | | |
Laparoscopic | 0.737 | 0.304–1.785 | | | | | | |
T stage | | | 0.009 | | | | 0.501 | |
T1 | 1 | 1 | | | 1 | 1 | | |
T2 | 4.048 | 0.676–24.230 | | | 0.913 | 0.125–6.660 | | |
T3 | 11.239 | 2.504–50.453 | | | 0.406 | 0.040–4.129 | | |
T4 | 8.865 | 2.034–38.638 | | | 0.296 | 0.033–2.636 | | |
N stage | | | 0.001 | | | | 0.196 | |
N0 | 1 | 1 | | | 1 | 1 | | |
N1 | 5.262 | 1.412–19.614 | | | 1.878 | 0.153–23.038 | | |
N2 | 4.372 | 1.315–14.541 | | | 1.121 | 0.110-11.452 | | |
N3 | 9.613 | 3.176–29.095 | | | 2.989 | 0.313–28.528 | | |
TNM stage | | | 0.002 | | | | 0.623 | |
I | 1 | 1 | | | 1 | 1 | | |
II | 10.251 | 2.125–49.443 | | | 0.731 | 0.209–2.554 | | |
III | 13.117 | 3.088–55.726 | | | - | - | | |
The data of TNM_III had a collinearity relation with the data of Adjuvant therapy, so it was deleted from the multivariate cox regression analysis |
Comparison of OS between patients with pure NEC and patients with gastric adenocarcinoma
In pure NEC group, PSM was also performed, and finally, 13 NEC cases and 46 gastric adenocarcinoma cases were successfully matched. The proportion of patients receiving neoadjuvant therapy in the two groups was 57.1% and 62.0% (Supplement Table 3, p = 0.742), respectively. OS in patients with pure NEC was also poorer than that of patients with gastric adenocarcinoma (Fig. 3a, p = 0.0150). We also evaluated the OS according to the neoadjuvant therapy status, the result was similar to that of all patients with NEC component. In the surgery group, the OS of patients with pure NEC was significantly poorer than that of patients with gastric adenocarcinoma (Fig. 3b, p = 0.0260), however, there was no significant difference in the neoadjuvant group (Fig. 3c, p = 0.9857). Detail patient characteristics before and after PSM in surgical group and neoadjuvant group were shown in Supplement Table 4–5.
Comparison of OS between patients with mixed tumors containing NEC components and patients with gastric adenocarcinoma
In the group of patients with mixed tumors containing NEC components, 32 cases were successfully matched with 101 gastric adenocarcinoma cases after PSM, with neoadjuvant therapy ratio of 34.4% and 41.6% (Supplement Table 6, p = 0.468), respectively. Significant differences in OS were also observed when compared with the OS of gastric adenocarcinoma (Fig. 4a, p = 0.0484). As with the pure NEC group, the OS of patients with mixed tumors containing NEC components was also poorer than that of gastric adenocarcinoma in the surgery group (Fig. 4b, p = 0.0198, Supplement Table 7), but there was still no significant difference in the neoadjuvant group (Fig. 4c, p = 0.9885, Supplement Table 8).
From the above results we can infer that in the surgical group, both pure NEC and mixed tumors patients had a poorer prognosis than that of gastric adenocarcinoma. To further explore the prognostic factors, Cox regression analysis was performed on matched patients in the surgical group of mixed tumors (Table 2). In univariate analysis, age, adjuvant therapy, tumor size ≥ 5 cm, tumor category (gastric adenocarcinoma or mixed tumor containing NEC components), T stage, N stage, and TNM stage are considered to be important factors affecting prognosis. Multivariate analysis identified tumor size ≥ 5 cm and tumor category as independent risk factors.
Comparison of OS between patients with mixed tumors containing different proportions of NEC components and patients with gastric adenocarcinoma in the surgery group
To investigate whether the level of NEC components had an effect on overall survival in the surgical group, < 30% G-HMiNEN, G-HMiNEN and > 70%G-HMiNEN were compared with gastric adenocarcinoma and found that even the group with the lowest content of NEC, < 30% G-HMiNEN group had a poorer OS than adenocarcinoma (Fig. 5a). And in the G-HMiNEN and > 70%G-HMiNEN groups with higher NEC content, the OS was also worse than that of gastric adenocarcinoma as speculated (Fig. 5b & 5c). Detailed clinical information after matching was shown in Supplement Table 9–11.