Received: 29 May 2021

| Revised: 19 June 2021

| Accepted: 27 June 2021

DOI: 10.1002/cam4.4147

RESEARCH ARTICLE

Worth it or not? Primary tumor resection for stage IV pancreatic

cancer patients: A SEER-­based analysis of 15,836 cases

Ningzhen Fu1,2,3,4 | Yu Jiang1,2,3,4 | Yuanchi Weng1,2,3,4 | Hao Chen1,2,3,4 |

Xiaxing Deng1,2,3,4 | Baiyong Shen1,2,3,4

1
Department of General Surgery,
Pancreatic Disease Center, Ruijin Abstract
Hospital affiliated to Shanghai Jiao Tong Background: Primary tumor resection (PTR) as a treatment option for patients with
University School of Medicine, Shanghai,
stage IV pancreatic cancer (PC) is controversial.
China
2
Shanghai Jiao Tong University School
Patients and methods: Stage IV PC patients, with treatment data from the National
of Medicine, Research Institute of Cancer Institute's Surveillance, Epidemiology, and End Results (SEER), were
Pancreatic Disease, Shanghai, China screened. The main outcomes were overall survival (OS) and cancer-­specific survival
3
State Key Laboratory of Oncogenes and
(CSS).
Related Genes, Shanghai, China
4 Results: We enrolled 15,836 stage IV PC patients in this study. Propensity score-­
Shanghai Jiao Tong University, Institute
of Translational Medicine, Shanghai, matched analyses revealed improved OS and CSS of patients receiving chemotherapy
China plus PTR versus chemotherapy (median survival time [MSTOS]: 13 vs. 9 months,
Correspondence
p = 0.024; MSTCSS: 14 vs. 10 months, p = 0.035), and chemoradiotherapy plus
Xiaxing Deng, Pancreatic Disease Center, PTR versus chemoradiotherapy (MSTOS: 14 vs. 7 months, p = 0.044; MSTCSS: 14
Shanghai Ruijin Hospital affiliated with vs. 7 months, p = 0.066). Multivariate adjusted analyses further confirmed these re-
Shanghai Jiaotong University School
of Medicine, No.197 Ruijin Er Road, sults. Stratified with different metastatic modalities, multivariate analyses suggested
Shanghai 200025, China. that PTR significantly improved the OS and CSS among patients with ≤1 metastatic
Email: kejiadxx@hotmail.com
organ, and that patients with brain metastasis might not benefit from chemotherapy
Baiyong Shen, Pancreatic Disease Center, treatment.
Shanghai Ruijin Hospital affiliated to
Conclusion: PTR improves the OS and CSS of stage IV PC patients on the basis of
Shanghai Jiaotong University School of
Medicine, Shanghai, China; Shanghai chemotherapy or chemoradiotherapy, provided that the metastases involve ≤1 organ.
Jiao tong University School of Medicine, Chemotherapy, however, should be carefully considered in patients with metastases
Research Institute of Pancreatic Disease,
involving the brain.
Shanghai, China; State Key Laboratory
of Oncogenes and Related Genes, No.197
Ruijin Er Road, Shanghai 200025, China. KEYWORDS
Email: shenby@shsmu.edu.cn cancer-­specific survival (CSS), metastatic disease, overall survival (OS), pancreatic cancer (PC),
primary tumor resection (PTR)
Funding information
This study was supported in part by the
National Natural Science Foundation of
China (Grant number: 81871906).

Co-­first author: Ningzhen Fu and Yu Jiang.

Co-­corresponding author: Xiaxing Deng and Baiyong Shen.

This is an open access article under the terms of the Creat​ive Commo​ns Attri​bution License, which permits use, distribution and reproduction in any medium, provided the original
work is properly cited.
© 2021 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.

5948 | 
wileyonlinelibrary.com/journal/cam4 Cancer Medicine. 2021;10:5948–5963.
FU et al.   
| 5949

1 | BACKG RO U N D PC, who had recorded treatment information. A total of

19,400 patients with stage IV PC were screened out with
Pancreatic cancer (PC) is well-­known for its high mortality clear metastatic site information. Excluding those with un-
rate and poor prognosis, causing 466,003 deaths in 2020.1 known race, unknown primary site surgery, unknown re-
Radical surgery is the only treatment option for this malignant gional/distant site surgery data, and no survival months,
disease. In 2020, 495,773 patients were diagnosed with PC,1 15,836 patients were enrolled in our study. The final cohort
and it was reported that approximately 60% of newly diag- was then divided into different comparison patterns for the
nosed PCs were metastatic.2 Even with the current advance- purpose of synchronous presentations, including those who
ments in chemotherapy and radiotherapy, these patients can received chemotherapy with or without PTR (N = 9515),
hardly survive for more than 1 year.3–­7 For these patients, nei- those undergoing chemoradiation with or without PTR
ther the National Comprehensive Cancer Network (NCCN)8 (N = 699), and those who received no treatment versus PTR
nor the American Society of Clinical Oncology (ASCO) only (N = 5403). (Figure S1) It was noteworthy that we
recommend resection of the primary tumor.9 Nevertheless, lacked detailed information, such as timing of treatments,
radical surgery to treat primary or metastatic sites has been whether the therapy paradigms were sequential or synchro-
accepted and conducted in an increasing number of meta- nized was unclear.
static tumors, including but not limited to neuroendocrine
neoplasm, breast cancer, and colorectal cancer.10–­13 Thus
far, primary tumor resection (PTR) for stage IV PC remains 2.2 | Methods
controversial.14–­22 A previous study by our group discovered
that surgery helped to prolong the overall survival (OS) of OS and CSS were calculated using the Kaplan–­Meier method,
patients with stage IV PC.14 Some also advocated PTR and univariate comparisons relied on the log-­rank test, and unad-
emphasized the importance of patient selection,15,20 while justed Cox models when necessary. Cox proportional hazard
others reported no survival benefits from PTR.19,22 Due to the regression models adjusted for other variables were applied
violation of established guidelines, approval for clinical tri- to calculate adjusted hazard ratios, 95% confidence intervals,
als to study this issue is difficult to acquire. Therefore, there and p values.
is no strong evidence to clarify the problem. Thus, we turn Propensity score matching (PSM) methods were used
to the Surveillance, Epidemiology, and End Results (SEER) to adjust differences among the aforementioned compari-
database, to evaluate the value of PTR among stage IV PC son patterns in our study. The caliper was set to 0.05, with a
patients. matching ratio of 1:1. The matched covariates are presented
The SEER is a clinical database that collects cancer inci- in the corresponding tables.
dence, prevalence, and survival data from US cancer regis- As timing of the enrollment relative to surgical resection
tries that cover approximately 34.6% of the US population.23 differed, sequential landmark analyses were implemented to
With the large-­volume multi-­center database, we screened evaluate survival in varied settings for patients surviving a
eligible stage IV PC patients with definite metastasis, fol- minimum of 0.5, 1, 2, and 3 years from diagnosis, to elimi-
low-­up, and treatment data. OS and cancer-­specific survival nate “time-­to-­treatment” bias.
(CSS) were utilized as the main outcomes, and analyzed with A forest plot was generated to evaluate the effect of PTR-­
multiple statistical methods to determine the value of PTR in combined therapy on OS by subgroups including age, race,
the treatment of stage IV PC patients. sex, grade, T/N stage, tumor location, metastatic site, and
number of involved organs, and further, to explore the poten-
tial candidates suitable for PTR.
2 | PATIE N TS A N D ME T HOD S Statistical analyses were performed using SPSS version
26, and R version 3.6.3. Normally distributed continuous
2.1 | Patients variables are displayed as average (standard deviation, SD),
while non-­normally distributed continuous variables are pre-
The eligibility criteria for patients in this study were patients sented as median (Q1–­Q3). The Kolmogorov–­Smirnov test
older than 18 years, who had stage IV PC with metastatic was used for normality testing of continuous variables. The
disease at diagnosis, from 1975 to 2016. All patients were Wilcoxon (Mann–­Whitney) and Kruskal–­Wallis tests were
pathologically diagnosed with primary malignant tumors used to evaluate continuous variables. Categorical variables
of the pancreas using ICD-­O-­3 codes of 8140/3 or 8500/3. are presented as percentages and analyzed using Pearson's
The primary site-­labeled column was C25.0–­C25.3, and test. A two-­sided p value of less than 0.05 was considered to
C25.7–­C25.9. In total, we identified 98.949 patients with be statistically significant.
TABLE 1 The baseline and demographic characteristics of total cohort and different comparison patterns

5950
All Comparison pattern 1 Comparison pattern 2 Comparison pattern 3

|
  
Chemotherapy Chemotherapy plus Chemoradiotherapy Chemoradiotherapy No treatment PTR only
(N = 15,836) (N = 9190) PTR (N = 325) p (N = 636) plus PTR (N = 63) p (N = 5,231) (N = 172) p
Age 67 (59–­75) 66 (58–­73) 65 (58–­71) 0.107 64 (57–­72) 62 (57–­68) 0.12 71 (62–­79) 70 (62–­78) 0.174
Female 7419 (46.8) 4211 (45.8) 142 (43.7) 0.449 284 (44.7) 34 (54) 0.157 2552 (48.8) 97 (56.4) 0.05
Race
White 12,549 (79.2) 7432 (80.9) 259 (79.7) 0.064 492 (77.4) 54 (85.7) 0.06 3998 (76.4) 143 (83.1) 0.121
Black 2097 (13.2) 1140 (12.4) 34 (10.5) 86 (13.5) 2 (3.2) 786 (15) 19 (11)
Other 1190 (7.5) 618 (6.7) 32 (9.8) 58 (9.1) 7 (11.1) 447 (8.5) 10 (5.8)
Grade
I 280 (1.8) 131 (1.4) 17 (5.2) <0.001 17 (2.7) 3 (4.8) <0.001 94 (1.8) 16 (9.3) <0.001
II 1612 (10.2) 870 (9.5) 124 (38.2) 66 (10.4) 27 (42.9) 439 (8.4) 72 (41.9)
III 1842 (11.6) 972 (10.6) 129 (39.7) 77 (12.1) 21 (33.3) 558 (10.7) 59 (34.3)
IV 65 (0.4) 34 (0.4) 1 (0.3) 6 (0.9) 0 (0) 19 (0.4) 3 (1.7)
Unknown 12,037 (76) 7183 (78.2) 54 (16.6) 470 (73.9) 12 (19) 4121 (78.8) 22 (12.8)
T stage
1 750 (4.7) 424 (4.6) 20 (6.2) <0.001 23 (3.6) 5 (7.9) 0.001 251 (4.8) 14 (8.1) <0.001
2 4951 (31.3) 2895 (31.5) 151 (46.5) 182 (28.6) 27 (42.9) 1575 (30.1) 61 (35.5)
3 4701 (29.7) 2868 (31.2) 109 (33.5) 156 (24.5) 22 (34.9) 1437 (27.5) 62 (36)
4 2423 (15.3) 1468 (16.0) 31 (9.5) 146 (23) 7 (11.1) 712 (13.6) 20 (11.2)
Unknown 3011 (19) 1535 (16.7) 14 (4.3) 129 (20.3) 2 (3.2) 1256 (24) 15 (8.7)
N stage
0 322 (2) 124 (1.3) 66 (20.3) <0.001 17 (2.7) 16 (25.4) <0.001 63 (1.2) 34 (19.8) <0.001
1 332 (2.1) 95 (1.0) 103 (31.7) 9 (1.4) 19 (30.2) 52 (1) 53 (30.8)
2 168 (1.1) 3 (0.0) 93 (28.6) 0 (0) 17 (27) 5 (0.1) 49 (28.5)
Unknown 15014 (94.8) 8968 (97.6) 63 (19.4) 610 (95.9) 11 (17.5) 5111 (97.7) 36 (20.9)
Tumor location
Head 5574 (35.2) 3109 (33.8) 183 (56.3) <0.001 248 (39) 38 (60.3) 0.001 1829 (35) 94 (54.7) <0.001
Bodytail 5913 (37.3) 3.656 (39.8) 110 (33.8) 220 (34.6) 19 (30.2) 1787 (34.2) 54 (31.4)
Overlapping 1470 (9.3) 910 (9.9) 17 (5.2) 52 (8.2) 5 (7.9) 465 (8.9) 8 (4.7)
Other 2879 (18.2) 1515 (16.5) 15 (4.6) 116 (18.2) 1 (1.6) 1150 (22) 16 (9.3)
(Continues)

FU et al.
 FU et al.
TABLE 1 (Continued)

All Comparison pattern 1 Comparison pattern 2 Comparison pattern 3

Chemotherapy Chemotherapy plus Chemoradiotherapy Chemoradiotherapy No treatment PTR only

(N = 15,836) (N = 9190) PTR (N = 325) p (N = 636) plus PTR (N = 63) p (N = 5,231) (N = 172) p
Msite bone 1133 (7.2) 509 (5.5) 6 (1.8) 0.004 209 (32.9) 3 (4.8) <0.001 293 (5.6) 6 (3.5) 0.233
Msite brain 102 (0.6) 29 (0.3) 0 (0) 0.311 25 (3.9) 0 (0) 0.109 28 (0.5) 0 (0) 0.336
Msite liver 12,151 (76.7) 7265 (79.1) 185 (56.9) <0.001 375 (59) 24 (38.1) 0.001 4072 (77.8) 90 (52.3) <0.001
Msite lung 3107 (19.6) 1816 (19.8) 31 (9.5) <0.001 158 (24.8) 5 (7.9) 0.002 1021 (19.5) 20 (11.6) 0.01
Msite number
0 2345 (14.8) 1226 (13.3) 114 (35.1) <0.001 125 (19.7) 32 (50.8) <0.001 762 (14..6) 65 (37.8) <0.001
1 10,840 (68.5) 6473 (70.4) 200 (61.5) 312 (49.1) 30 (47.6) 3620 (69.2) 98 (57)
2 2314 (14.6) 1332 (14.5) 11 (3.4) 144 (22.6) 1 (1.6) 755 (14.4) 9 (5.2)
3 323 (2) 154 (1.7) 0 (0) 53 (8.3) 0 (0) 92 (1.8) 0 (0)
4 14 (0.1) 5 (0.1) 0 (0) 2 (0.3) 0 (0) 2 (0) 0 (0)
Insurance status
Insured 15,201 (96) 8874 (96.6) 316 (97.2) 0.806 610 (95.9) 60 (95.2) 0.967 4971 (95) 165 (95.9) 0.79
Uninsured 421 (2.7) 215 (2.3) 6 (1.8) 17 (2.7) 2 (3.2) 163 (3.1) 5 (2.9)
Unknown 214 (1.4) 101 (1.1) 3 (0.9) 1.4 (9) 1 (1.6) 97 (1.9) 2 (1.2)
Married 9056 (57.2) 5725 (62.3) 226 (69.5) 0.008 373 (58.6) 38 (60.3) 0.797 2.472 (47.3) 105 (61) <0.001
Chemotherapy 10,214 (64.5) \ \ \ \ \ \
Radiotherapy 918 (5.8) \ \ \ \ \ \
PTR 562 (3.5) \ \ \ \ \ \
Distant/reginal 963 (6.1) 421 (4.6) 110 (33.8) <0.001 54 (8.5) 24 (38.1) <0.001 272 (5.2) 66 (38.4) <0.001
site resection
Abbreviations: Msite, metastatic site; PTR, primary tumor resection.

  
|
5951
5952
|    FU et al.

3 | R E S U LTS treatments, additional PTR significantly improved the OS

and CSS of patients with stage IV PC. (Figure 3; chemo-
Among the 15,836 stage IV PC patients, the median follow- therapy plus PTR vs. chemotherapy: OS: 13 vs. 9 months,
­up was 33 months (median survival time [MSTOS], 4 months; p = 0.024; CSS: 14 vs. 10 months, p = 0.035; chemora-
MSTCSS, 5 months). According to the different treatment diotherapy plus PTR vs. chemoradiotherapy: OS: 14 vs.
types, patients were categorized into three groups: chemo- 7 months, p = 0.044; CSS: 14 vs. 7 months, p = 0.066; PTR
therapy with or without PTR (N = 9,515), chemoradiation only vs. no treatment: OS: 4 vs. 3 months, p = 0.34; CSS: 4
with or without PTR (N = 699), and no treatment versus PTR vs. 4 months, p = 0.69).
only (N = 5,403). All patient characteristics are presented in To eliminate time-­to-­treatment bias, sequential landmark
Table 1. analyses with landmarks set as 0, 0.5, 1, and 2 years after diag-
Kaplan–­Meier analysis was performed targeting OS and nosis, were performed. (Figure 4, Figure S2) Chemotherapy
CSS categorized with different therapy modalities. (Figure 1) plus PTR, compared with chemotherapy, improved the OS
MST was calculated and listed in order from high to low: OS: and CSS in both univariate and multivariate survival anal-
chemoradiotherapy plus PTR, 15 months; chemotherapy plus yses. Chemoradiotherapy plus PTR was also advantageous
PTR, 13 months; radiotherapy plus PTR, 8 months; chemo- over chemoradiotherapy for CSS and OS. Nevertheless, no
radiotherapy, 7 months; chemotherapy, 6 months; PTR, significant differences were observed when comparing non-­
4 months; radiotherapy, 2 months; no treatment, 2 months; treatment and PTR-­only patients. (Table S1, Table S2).
CSS: chemoradiotherapy plus PTR, 15 months; chemother- Except for the primary site, the outcomes of patients
apy plus PTR, 14 months; chemoradiotherapy, 8 months; receiving distant/regional site resections were also evalu-
radiotherapy plus PTR, 8 months; chemotherapy, 7 months; ated (Table S3). The MSTs for patients with resections of
PTR, 5 months; radiotherapy, 3 months; no treatment, the primary site and distant/regional site (OS: 10 months;
2 months. The PTR intervention significantly improved OS CSS: 11 months), primary site only (OS: 10 months; CSS:
among all three comparison patterns. (Figure 1) Similarly, 11 months), the distant/regional site only (OS: 6 months; CSS:
beneficial outcomes of PTR were also observed when ana- 7 months), and no surgery (OS: 4 months; CSS: 5 months)
lyzing CSS. (Figure 1). were measured, which indicated that surgical interventions
After adjusting for confounding covariates, PTR also increased the OS and CSS in stage IV PC patients. (Figure 5)
significantly improved the OS and CSS in different com- The pairwise comparisons among the four groups above were
parison patterns. (Figure 2, Table 2) PSM was performed all significant (p < 0.001) except for that between primary
to reduce the confounding effects of covariates for three site plus distant/regional site and primary site only (OS:
pairs of comparison patterns in addition. The matched co- p = 0.819, CSS: p = 0.570). Additionally, in multivariate
horts are shown in Table 3. Compared with PTR-­absent analyses, PTR increased the OS and CSS more significantly

F I G U R E 1 (A) Comparison of OS among stage IV PC patients receiving different treatment modalities. (B) Comparison of CSS among stage
IV PC patients receiving different treatment modalities
FU et al.   
| 5953

than distant/regional resection did (Figure 5, multivariate CSS, p = 0.003). On the basis of primary site resection, dis-
Cox model: primary site vs. no-­surgery: OS, p < 0.001; CSS, tant/regional resection might not result in additional OS/CSS
p < 0.001; distant/regional site vs. no-­surgery: OS, p = 0.011; improvements. (Multivariate Cox Model: primary site plus

F I G U R E 2 (A) Comparison of OS between patients receiving chemotherapy and chemotherapy plus PTR with covariates adjusted. (B)
Comparison of CSS between patients receiving chemotherapy and chemotherapy plus PTR with covariates adjusted. (C) Comparison of OS
between patients receiving chemoradiotherapy and chemoradiotherapy plus PTR with covariates adjusted. (D) Comparison of CSS between patients
receiving chemoradiotherapy and chemoradiotherapy plus PTR with covariates adjusted. (E) Comparison of OS between patients receiving with
covariates adjusted. (A) Comparison of CSS between patients receiving no treatments and PTR only with covariates adjusted
TABLE 2 Univariate and multivariate survival analyses of different variables aiming at OS and CSS for patients with stage IV PC

5954
Univariate Multivariate

|
  
OS CSS OS CSS

HR (95% CI) p HR (95% CI) p aHR (95% CI) p aHR (95% CI) p
Age 1.016 (1.014–­1.017) <0.001 1.015 (1.013–­1.017) <0.001 1.010 (1.008–­1.012) <0.001 1.009 (1.008–­1.011) <0.001
Female 0.987 (0.971–­1.004) 0.125 0.992 (0.975–­1.010) 0.374
Race 0.008 0.226 0.084
White Ref Ref Ref
Black 1.079 (1.027–­1.133) 0.002 1.021 (0.982–­1.062) 0.297 1.058 (1.007–­1.112) 0.027
Other 1.032 (0.969–­1.099) 0.328 1.002 (0.957–­1.049) 0.936 1.015 (0.952–­1.081) 0.655
Grade <0.001 <0.001 <0.001 <0.001
I Ref Ref Ref Ref
II 1.130 (0.986–­1.294) 0.079 1.119 (0.972–­1.290) 0.119 1.283 (1.119–­1.471) <0.001 1.256 (1.089–­1.448) 0.002
III 1.420 (1.242–­1.625) <0.001 1.403 (1.219–­1.614) <0.001 1.612 (1.408–­1.846) <0.001 1.565 (1.359–­1.802) <0.001
IV 1.287 (0.959–­1.729) 0.093 1.219 (0.890–­1.669) 0.218 1.282 (0.954–­1.722) 0.100 1.191 (0.869–­1.632) 0.276
Unknown 1.430 (1.259–­1.624) <0.001 1.393 (1.220–­1.590) <0.001 1.447 (1.273–­1.645) <0.001 1.402 (1.227–­1.602) <0.001
T stage <0.001 <0.001 <0.001 <0.001
1 Ref Ref Ref Ref
2 1.112 (1.023–­1.209) 0.013 1.119 (1.025–­1.221) 0.012 1.139 (1.047–­1.238) 0.002 1.139 (1.044–­1.244) 0.003
3 1.226 (1.127–­1.333) <0.001 1.243 (1.138–­1.356) <0.001 1.243 (1.142–­1.353) <0.001 1.248 (1.142–­1.363) <0.001
4 1.145 (1.048–­1.250) 0.003 1.157 (1.055–­1.268) 0.002 1.196 (1.095–­1.307) <0.001 1.206 (1.099–­1.323) <0.001
Unknown 1.422 (1.304–­1.550) <0.001 1.254 (1.144–­1.374) <0.001 1.296 (1.185–­1.417) <0.001 1.182 (1.075–­1.299) 0.001
N stage <0.001 <0.001 <0.001 <0.001
0 Ref Ref <0.001 Ref Ref
1 1.081 (0.910–­1.286) 0.375 1.112 (0.928–­1.333) 0.248 1.159 (0.974–­1.379) 0.096 1.199 (0.999–­1.438) 0.051
2 1.120 (0.911–­1.378) 0.282 1.144 (0.921–­1.420) 0.224 1.427 (1.143–­1.781) 0.002 1.403 (1.114–­1.768) 0.004
Unknown 1.881 (1.659–­2.134) <0.001 1.878 (1.645–­2.144) <0.001 1.435 (1.251–­1.645) <0.001 1.469 (1.271–­1.697) <0.001
Tumor location <0.001 <0.001 <0.001 <0.001
Head 0.908 (0.873–­0.944) <0.001 0.897 (0.861–­0.934) <0.001 0.914 (0.878–­0.952) <0.001 0.91 (0.873–­0.949) <0.001
Bodytail Ref Ref Ref Ref
Overlapping 1.030 (0.970–­1.094) 0.339 1.047 (0.983–­1.114) 0.152 0.993 (0.943–­1.046) 0.792 1.016 (0.955–­1.082) 0.611
Other 1.154 (1.101–­1.209) <0.001 1.017 (0.966–­1.070) 0.521 1.000 (0.941–­1.062) 0.996 0.922 (0.872–­0.976) 0.005
Msite bone 1.251 (1.175–­1.333) <0.001 1.168 (1.091–­1.251) <0.001 1.157 (0.992–­1.35) 0.064 1.132 (0.952–­1.346) 0.159

FU et al.
(Continues)
 FU et al.
TABLE 2 (Continued)

Univariate Multivariate

OS CSS OS CSS

HR (95% CI) p HR (95% CI) p aHR (95% CI) p aHR (95% CI) p
Msite brain 1.469 (1.201–­1.795) <0.001 1.264 (1.006–­1.589) 0.045 1.309 (1.057–­1.62) 0.014 1.228 (0.964–­1.565) 0.096
Msite liver 1.241 (1.291–­1.343) <0.001 1.323 (1.269–­1.380) <0.001 1.313 (1.132–­1.523) <0.001 1.351 (1.143–­1.596) <0.001
Msite lung 1.114 (1.069–­1.162) <0.001 1.095 (1.048–­1.145) <0.001 1.007 (0.868–­1.168) 0.923 1.009 (0.854–­1.192) 0.915
Msite number <0.001 <0.001 <0.001 <0.001
0 Ref Ref Ref Ref
1 1.234 (1.177–­1.295) <0.001 1.249 (1.187–­1.314) <0.001 0.947 (0.814–­1.102) 0.480 0.928 (0.783–­1.1) 0.388
2 1.537 (1.446–­1.634) <0.001 1.551 (1.454–­1.655) <0.001 1.104 (0.827–­1.472) 0.502 1.091 (0.788–­1.509) 0.601
3 1.977 (1.754–­2.228) <0.001 1.768 (1.547–­2.020) <0.001 1.309 (0.848–­2.021) 0.224 1.178 (0.721–­1.924) 0.514
4 2.244 (1.326–­3.796) <0.001 1.993 (1.102–­3.606) 0.023
Insurance status 0.012 0.068 0.004 0.019
Insured Ref Ref Ref Ref
Uninsured 1.154 (1.043–­1.277) 0.006 1.131 (1.015–­1.260) 0.026 1.188 (1.072–­1.317) 0.001 1.17 (1.048–­1.305) 0.005
Unknown 1.088 (0.942–­1.257) 0.253 1.056 (0.904–­1.232) 0.493 1.040 (0.900–­1.202) 0.596 1.028 (0.881–­1.201) 0.723
Married 0.827 (0.800–­0.855) <0.001 0.846 (0.817–­0.876) <0.001 0.915 (0.884–­0.947) <0.001 0.922 (0.89–­0.956) <0.001
Chemotherapy 0.405 (0.391–­0.420) <0.001 0.421 (0.406–­0.437) <0.001 0.418 (0.403–­0.434) <0.001 0.428 (0.412–­0.445) <0.001
Radiotherapy 0.828 (0.771–­0.888) <0.001 0.790 (0.732–­0.852) <0.001 0.915 (0.849–­0.985) 0.018 0.888 (0.82–­0.961) 0.003
PTR 0.520 (0.473–­0.572) <0.001 0.543 (0.493–­0.599) <0.001 0.704 (0.613–­0.807) <0.001 0.732 (0.635–­0.845) <0.001
Distant/reginal site 0.762 (0.710–­0.818) <0.001 0.740 (0.687–­0.798) <0.001 0.900 (0.836–­0.968) 0.005 0.876 (0.81–­0.947) 0.001
resection
Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; CSS, cancer-­specific survival; HR, hazard ratio; Msite, metastatic site; OS, overall survival; p, p value; PTR, primary tumor resection.

  
|
5955
TABLE 3 Baseline and demographic characteristics of different comparison patterns after PSM

5956
Comparison pattern 1 Comparison pattern 2 Comparison pattern 3

|
  
Chemotherapy Chemotherapy plus Chemoradiotherapy Chemoradiotherapy No treatment PTR only
(N = 139) PTR (N = 139) p (N = 22) plus PTR (N = 22) p (N = 73) (N = 73) p
Age 64 (57–­71) 65 (58–­71) 0.6 63 (54–­72) 61 (56–­67) 0.481
Female 72 (51.8) 59 (42.4) 0.149 13 (59.1) 10 (45.5) 0.546 41 (56.2) 42 (57.5) 0.867
Race
White 107 (77.0) 108 (77.7) 0.805 20 (90.9) 20 (90.9) 1 56 (76.7) 56 (76.7) 0.585
Black 16 (11.5) 18 (12.9) 1 (4.5) 1 (4.5) 8 (11) 11 (15.1)
Other 16 (11.5) 13 (9.4) 1 (4.5) 1 (4.5) 9 (12.3) 9 (12.3)
Grade
I 7 (5.0) 8 (5.8) 0.656 2 (9.1) 1 (4.5) 0.727 6 (8.2) 7 (9.6) 0.222
II 40 (28.8) 43 (30.9) 4 (18.2) 7 (31.8) 22 (30.1) 27 (37)
III 35 (25.2) 41 (29.5) 5 (22.7) 4 (18.2) 12 (16.4) 18 (24.7)
IV 1 (0.7) 0 (0.0) 0 (0) 0 (0) 2 (2.7) 0 (0)
Unknown 56 (40.3) 47 (33.8) 11 (50.0) 10 (45.5) 31 (42.5) 21 (28.8)
T stage
1 7 (5.0) 10 (7.2) 0.608 2 (9.1) 1 (4.5) 0.962 7 (9.6) 4 (5.5) 0.825
2 53 (38.1) 61 (43.9) 10 (45.5) 9 (40.9) 27 (37) 24 (32.9)
3 42 (30.2) 40 (28.8) 4 (18.2) 5 (22.7) 17 (23.3) 19 (26)
4 25 (18.0) 17 (12.2) 4 (18.2) 5 (22.7) 11 (15.1) 12 (16.4)
Unknown 12 (8.6) 11 (7.9) 2 (9.1) 2 (9.1) 11 (15.1) 14 (19.2)
N stage
0 40 (28.8) 42 (30.2) 0.58 8 (36.4) 10 (45.5) 0.812 17 (23.3) 15 (20.5) 0.926
1 38 (27.3) 29 (20.9) 4 (18.2) 3 (13.6) 20 (27.4) 18 (24.7)
2 3 (2.2) 5 (3.6) 0 (0.0) 0 (0.0) 4 (5.5) 4 (5.5)
Unknown 58 (41.7) 63 (45.3) 10 (45.5) 9 (40.9) 32 (43.8) 36 (49.3)
Tumor location
Head 78 (56.1) 73 (52.5) 0.834 16 (72.7) 13 (59.1) 0.65 36 (49.3) 35 (47.9) 0.974
Bodytail 42 (30.2) 48 (34.5) 5 (22.7) 7 (31.8) 23 (31.5) 22 (30.1)
Overlapping 9 (6.5) 7 (5.0) 1 (4.5) 1 (4.5) 3 (4.1) 4 (5.5)
Other 10 (7.2) 11 (7.9) 0 (0.0) 1 (4.5) 11 (15.1) 12 (16.4)

(Continues)

FU et al.
 FU et al.
TABLE 3 (Continued)

Comparison pattern 1 Comparison pattern 2 Comparison pattern 3

Chemotherapy Chemotherapy plus Chemoradiotherapy Chemoradiotherapy No treatment PTR only

(N = 139) PTR (N = 139) p (N = 22) plus PTR (N = 22) p (N = 73) (N = 73) p
Msite bone 1 (0.7) 3 (2.2) 0.615 1 (4.5) 0 (0.0) 1 2 (2.7) 3 (4.1) 0.649
Msite brain 0 (0.0) 0 (0.0) \ 0 (0.0) 0 (0.0) \ 0 (0) 0 (0) \
Msite liver 76 (54.7) 90 (64.7) 0.112 12 (54.5) 11 (50.0) 1 34 (46.6) 40 (54.8) 0.321
Msite lung 15 (10.8) 13 (9.4) 0.842 2 (9.1) 1 (4.5) 1 10 (13.7) 9 (12.3) 0.806
Msite number
0 52 (37.4) 41 (29.5) 0.306 7 (31.8) 10 (45.5) 0.353 32 (43.8) 25 (34.2) 0.413
1 82 (59.0) 90 (64.7) 15 (68.2) 12 (54.5) 36 (49.3) 44 (60.3)
2 5 (3.6) 8 (5.8) 0 (0.0) 0 (0.0) 5 (6.8) 4 (5.5)
3 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
4 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Insurance status
Insured 135 (97.1) 135 (97.1) 0.766 20 (90.9) 21 (95.5) 1 70 (95.9) 69 (94.5) 0.843
Uninsured 2 (1.4) 3 (2.2) 0 (0) 0 (0) 1 (1.4) 2 (2.7)
Unknown 2 (1.4) 1 (0.7) 2 (9.1) 1 (4.5) 2 (2.7) 2 (2.7)
Married 87 (62.6) 92 (66.2) 0.616 15 (68.2) 13 (59.1) 0.754 43 (58.9) 40 (54.8) 0.616
Distant/reginal site 47 (33.8) 37 (26.6) 0.24 7 (31.8) 7 (31.8) 1 21 (28.8) 19 (26) 0.711
resection
PTR,.primary tumor resection. Msite, metastatic site.

  
|
5957
5958
|    FU et al.

distant/regional site vs. primary site only: OS, p = 0.486; 4 | DISCUSSION

CSS, p = 0.341).
A forest plot was constructed to verify whether the OS im- The majority of patients with PC (approximately 60%) were
provement seen with PTR existed in different subcategories. diagnosed with metastatic disease (stage IV).2 Current guide-
(Figure 6) It was revealed that insured patients with T2–­4, lines do not recommend surgery for these patients.8,9 And
N0–­1, and distant metastases involving ≤1 organ, could ben- whether these patients should be treated with PTR or not re-
efit from PTR regardless of age, sex, race, tumor location, or mains a controversy. Some advocate PTR as a treatment mo-
marital status. dality,14,15,20 while others disagreed.19,22 Gu et al. compared
For different organ involvement modalities, we com- patients that underwent PTR with those that underwent by-
pared the OS and CSS using log-­rank analyses. (Table S4) pass or exploratory laparotomy only, on the basis of a 3-­year
Stratified with different metastatic modalities, multivari- follow-­up.14 They reported a significantly prolonged OS rate
ate analyses were performed. (Table 4, Table S5) It was in patients that underwent PTR. However, the sample size
discovered that chemotherapy would no longer ameliorate was limited (34 for the PTR group), and adjuvant treatments
OS and CSS in patients with brain metastasis. The im- were not well incorporated into the analysis. The metastatic
pact of primary site surgeries was only significant when a patterns have not been well explored. The work of Wang et al.
single organ was involved. Distant/regional site surgeries was also limited by the lack of data on adjuvant therapy, and
were also found to be beneficial in patients with only liver metastatic patterns.15 Tachezy et al. studied this issue based
metastasis. on 69 cases and meta-­analysis, proving the survival benefits

F I G U R E 3 (A) Comparison of OS between patients receiving chemotherapy and chemotherapy plus PTR after PSM. (B) Comparison
of CSS between patients receiving chemotherapy and chemotherapy plus PTR after PSM. (C) Comparison of OS between patients receiving
chemoradiotherapy and chemoradiotherapy plus PTR after PSM. (D) Comparison of CSS between patients receiving chemoradiotherapy and
chemoradiotherapy plus PTR after PSM. (E) Comparison of OS between patients receiving after PSM (A) Comparison of CSS between patients
receiving no treatments and PTR only after PSM
FU et al.   
| 5959

F I G U R E 4 Sequential landmark Kaplan–­Meier analyses of OS at ≥0, ≥0.5, ≥1, and ≥2 years between patients receiving chemotherapy and
chemotherapy plus PTR, A–­D, patients receiving chemoradiotherapy and chemoradiotherapy plus PTR, E–­H, patients receiving no treatments and
PTR only, I–­L. The blue curves referred to the PTR-­absence treatments while the yellow curves referred to the PTR-­combined treatments

F I G U R E 5 (A) Comparison of OS among stage IV PC patients receiving different surgery treatment modalities. (B) Comparison of CSS
among stage IV PC patients receiving different surgery treatment modalities
5960
|    FU et al.

of PTR among hepatic oligometastatic PC patients, especially carefully selected, and that the positive impact of surgeries
for patients with tumors located at the head of the pancreas.20 should be further verified.16–­18,20,21 Hitherto, no consensus on
Although the neoadjuvant/adjuvant treatment data were this issue has been reached; however, majority still maintain
listed, they were not well utilized in the analyses. The main a positive attitude. Previous studies were limited by sample
limitations of the studies by Dünschede et al, and Gleisner size, lack of adjuvant treatment analyses, insufficient meta-
et al. were the sample sizes (23 and 17, respectively), and static pattern analyses, and uncorrected “time-­to-­treatment”
the nonmention of “time-­to-­treatment” bias.19,22 Some were bias. Thus, based on the SEER database with treatment and
speculative, and proposed that the PTR candidates should be metastatic pattern data, we sought to fix the limitations above

F I G U R E 6 Subgroup analysis of the

association between PTR and OS using
forest plot (chemotherapy plus PTR vs.
chemotherapy). CI, confidence interval;
T, tumor stage; N, nodal stage; Msite,
metastatic site, including brain, liver,
bone and lung; number, metastatic organ
summation
 FU et al.

TABLE 4 Multivariate Cox regression analyses of different treatments stratified with different metastatic organ modalities considered for overall survival

PTR Distant/reginal site resection Radiotherapy Chemotherapy

Number aHR 95% CI p aHR 95% CI p aHR 95% CI Sig aHR 95% CI p
B 229 0.277 0.133–­0.579 0.001 1.101 0.676–­1.794 0.698 1.27 0.954–­1.691 0.102 0.339 0.25–­0.461 <0.001
BC 7 16.102 0.003–­79,865.239 0.522 2.007 0.178–­22.642 0.573 0.003 0–­370.28 0.333
BCH 6 0.998 0.089–­11.259 0.999 15.119 0.066–­3459.531 0.327 0.005 0–­26.38 0.226
BCHP 14 5.146 0.371–­71.47 0.222 0.225 0.05–­1.008 0.051 0.204 0.043–­0.959 0.044
BCP 8 0.551 0.041–­7.458 0.654 1.876 0.26–­13.526 0.532 0 0–­4.237E+163 0.949
BH 444 0.554 0.076–­4.048 0.561 0.857 0.433–­1.696 0.658 0.909 0.718–­1.152 0.43 0.352 0.283–­0.439 <0.001
BHP 295 0.617 0.197–­1.931 0.407 0.928 0.695–­1.238 0.61 0.356 0.27–­0.47 <0.001
BP 130 0.566 0.175–­1.832 0.342 1.204 0.67–­2.163 0.535 1.108 0.754–­1.63 0.601 0.302 0.194–­0.47 <0.001
C 15 0 0–­9.851E13 0.989 0.84 0.238–­2.972 0.787 0.336 0.084–­1.338 0.122
CH 26 0.655 0.121–­3.541 0.623 1.182 0.461–­3.032 0.728 0.491 0.201–­1.197 0.118
CHP 14 1.45 0.293–­7.174 0.649 0.893 0.152–­5.257 0.9 0.171 0.019–­1.551 0.117
CP 12 2.064 0.156–­27.283 0.582 0.186 0.034–­1.03 0.054 0.494 0.08–­3.045 0.447
H 9657 0.545 0.476–­0.624 <0.001 0.803 0.722–­0.893 <0.001 0.869 0.771–­0.979 0.021 0.397 0.38–­0.416 <0.001
HP 1695 0.67 0.4–­1.122 0.128 0.941 0.673–­1.317 0.724 0.808 0.581–­1.126 0.208 0.41 0.368–­0.457 <0.001
Other 2345 0.586 0.497–­0.692 <0.001 0.95 0.839–­1.077 0.424 0.812 0.686–­0.96 0.015 0.384 0.25–­0.421 <0.001
P 939 0.663 0.458–­0.96 0.029 0.943 0.718–­1.239 0.674 1.038 0.751–­1.436 0.82 0.447 0.386–­0.519 <0.001
Organ involvement code: P-­lung, C-­brain, H-­liver, and B-­bone. The combination of the letters referred to multi-­organ involvements. Other referred to those IV stage patients with metastatic organs other than four organs
mentioned above.
Bold p value indicated those less than 0.001.
Abbreviations: aHR, adjusted hazard ratio; CI, confidence interval; p, p value; PTR, primary tumor resection.
  
|
5961
5962
|    FU et al.

with multivariate, PSM, stratification, and sequential land- it could be considered as a last complementary choice. The
mark analyses. survival prognoses of different metastatic organ modalities
In our study, additional PTR significantly increased the varied slightly (MST range ≤8 months). Therefore, the sub-
OS and CSS, compared with PTR-­absent treatments in mul- group of metastatic behavior was of no use. In summary, pa-
tivariable risk adjustment analyses, based on chemotherapy/ tients with multiorgan or liver involvement were observed to
chemoradiotherapy. PSM analyses were performed serving as have worse prognoses compared to those with brain, bone, or
the sensitivity analysis and the complementary verification. lung involvement.
“Time-­to-­treatment” bias arises in comparative research, This study had some limitations. First, its retrospective
when survival time is measured from enrollment (e.g., diag- nature limited the evidence value of our work. However,
nosis), and the receipt of treatment occurs during follow-­up. our discoveries provide more grounds for ethical commit-
Patients with poor performance status or significantly aggres- tees to approve clinical trials on the subject matter. Second,
sive disease, might die too early (before undergoing PTR), due to the limitations of the database, details about the
which creates an apparent survival disadvantage for patients treatments, such as the treatment time, chemotherapy reg-
who do not receive PTR. The time interval between the date imen, surgery details, and so on were not designed and re-
of initial diagnosis and the date of surgery, was unknown in corded. These findings should be taken into consideration
the SEER data. The longer the deferral, the greater the “time-­ in future prospective studies. Third, information on the total
to-­treatment” bias is in favor of PTR. We speculated that tumor burden (e.g., size and number of metastatic loci for
“time-­to-­treatment” bias could partially explain the treatment different involved organs), as well as the response to treat-
effect estimate in previous studies. In our study, we used the ment was lacking. Fourth, selection bias existed because
landmark analysis, one of the recommended corrective ap- of preference for surgery, for individuals whose general
proaches, to eliminate “time-­to-­treatment” bias, after which condition was more “acceptable.” Good performance status
PTR with synchronous chemotherapy/chemoradiotherapy and personal habits may influence the OS, even though the
was still associated with improved OS and CSS. NCCN guidelines bifurcated the strategies for stage IV PC
As suggested by Drs. Shi, Liu, and Shrikhande, the pop- patients based on performance status, while SEER lacked
ulation for PTR should be carefully selected.16–­18 Forest plot this information.
was then performed, which revealed the OS improvement Ultimately, “time-­to-­treatment” bias, the metastatic pat-
seen with the chemotherapy plus PTR modality compared tern analyses, and combined treatment analyses, which most
with chemotherapy only, among different subcategories. It previous studies neglected, were taken into consideration
was observed that insured patients with T2–­4, N0–­1, and dis- with a satisfactory sample size. In conclusion, we demon-
tant metastases involving ≤1 organ, could benefit from PTR strated improvements in OS and CSS in patients with stage
regardless of age, sex, race, tumor location, or marital sta- IV PC. If only one organ is involved, surgery should be per-
tus; this could serve as a reference for the screening of PTR formed on the basis of chemotherapy or chemoradiotherapy.
candidates. We supposed that there probably existed special The use of chemotherapy as a treatment option in patients
oncological behaviors in T1 metastatic patients. with brain metastasis, should be carefully considered due to
To treat metastatic loci, Dünschede et al. recommended the lack of significant improvement in OS and CSS in our
metachronous resection, while Tachezy et al. recommended studies.
synchronous resection for liver metastases. Dr. Liu, however,
believed that both metachronous and synchronous metastatic CONFLICTS OF INTEREST
resection, improved survival in lung metastases. However, The authors report no conflict of interest.
some researchers have recommended nonsurgical treatment
options.24–­26 In our study, PTR improved the OS and CSS ETHICS STATEMENT
more than distant/regional site resection. In multivariate Cox This study was exempt from institutional review board ap-
analyses, combined distant/regional site resection could not proval due to the nature of the study. Because all data were
contribute to OS/CSS increase on the basis of PTR. deidentified, patient consent was waived.
Due to the insufficiency of studies on metastatic patterns,
stratified with different metastatic organ involvement mo- DATA AVAILABILIT Y STATEMENT
dalities, we verified the impact of different treatments. PTR The data of our work are available and publicly accessible.
should be considered only if a single organ is involved, which The original data comes from the Surveillance, Epidemiology,
corresponds to the discoveries in the forest plot analysis. and End Results (SEER) database.
Distant/regional site resection can be applied when only the
liver is involved. If the brain is involved, the administration of ORCID
chemotherapy should be carefully considered. Radiotherapy, Ningzhen Fu https://orcid.org/0000-0001-7002-1745
however, was not beneficial for patients with stage IV PC; Baiyong Shen https://orcid.org/0000-0002-3994-248X
FU et al.   
| 5963

Epidemiology and End Results (SEER) analysis. Med Sci Monit.

R E F E R E NC E S 2019;25:8230-­8241.
1. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: 16. Shi S, Yu XJ. Time to think: selecting patients who may benefit
GLOBOCAN estimates of incidence and mortality worldwide for 36 from synchronous resection of primary pancreatic cancer and liver
cancers in 185 countries. CA A Cancer J Clin. 2021;71(3):209-­249. metastases. World J Gastroenterol. 2018;24(33):3677-­3680.
2. Vincent A, Herman J, Schulick R, Hruban RH, Goggins M. 17. Liu Q, Zhang R, Michalski CW, Liu B, Liao Q, Kleeff J. Surgery
Pancreatic cancer. Lancet. 2011;378(9791):607-­620. for synchronous and metachronous single-­organ metastasis of pan-
3. Kang J, Hwang I, Yoo C, et al. Nab-­paclitaxel plus gemcitabine creatic cancer: A SEER database analysis and systematic literature
versus FOLFIRINOX as the first-­line chemotherapy for patients review. Sci Rep. 2020;10(1):4444.
with metastatic pancreatic cancer: retrospective analysis. Investig 18. Shrikhande SV, Kleeff J, Reiser C, et al. Pancreatic resection
New Drugs. 2018;36(4):732-­741. for M1 pancreatic ductal adenocarcinoma. Ann Surg Oncol.
4. Conroy T, Desseigne F, Ychou M, et al. FOLFIRINOX ver- 2007;14(1):118-­127.
sus gemcitabine for metastatic pancreatic cancer. N Engl J Med. 19. Dünschede F, Will L, von Langsdorf C, et al. Treatment of meta-
2011;364(19):1817-­1825. chronous and simultaneous liver metastases of pancreatic cancer.
5. Peixoto RD, Ho M, Renouf DJ, et al. Eligibility of metastatic pan- Eur Surg Res. 2010;44(3–­4):209-­213.
creatic cancer patients for first-­line palliative intent nab-­paclitaxel 20. Tachezy M, Gebauer F, Janot M, et al. Synchronous resections of
plus gemcitabine versus FOLFIRINOX. Am J Clin Oncol. hepatic oligometastatic pancreatic cancer: disputing a principle in
2017;40(5):507-­511. a time of safe pancreatic operations in a retrospective multicenter
6. McBride A, Bonafede M, Cai Q, et al. Comparison of treat- analysis. Surgery. 2016;160(1):136-­144.
ment patterns and economic outcomes among metastatic 21. Bellon E, Gebauer F, Tachezy M, et al. Pancreatic cancer and liver
pancreatic cancer patients initiated on nab-­paclitaxel plus gem- metastases: state of the art. Update Surg. 2016;68(3):247-­251.
citabine versus FOLFIRINOX. Expert Rev Clin Pharmacol. 22. Gleisner AL, Assumpcao L, Cameron JL, et al. Is resection of
2017;10(10):1153-­1160. periampullary or pancreatic adenocarcinoma with synchronous
7. Azar I, Virk G, Esfandiarifard S, et al. Treatment and survival rates hepatic metastasis justified? Cancer. 2007;110(11):2484-­2492.
of stage IV pancreatic cancer at VA hospitals: a nation-­wide study. 23. Doll KM, Rademaker A, Sosa JA. Practical guide to surgical
J Gastrointest Oncol. 2019;10(4):703-­711. data sets: Surveillance, Epidemiology, and End Results (SEER)
8. Tempero MA. NCCN guidelines updates: pancreatic cancer. J Natl Database. JAMA Surg. 2018;153(6):588-­589.
Compr Canc Netw. 2019;17(6)(5.5):603-­605. 24. Timmer FEF, Geboers B, Nieuwenhuizen S, et al. Locoregional
9. Sohal DPS, Kennedy EB, Khorana A, et al. Metastatic pancreatic treatment of metastatic pancreatic cancer utilizing resection, ablation
cancer: ASCO clinical practice guideline update. J Clin Oncol. and embolization: a systematic review. Cancers (Basel). 2021;13(7).
2018;36(24):2545-­2556. 25. Yan X, Ning Z-­Y, Wang P, et al. Combined ablation-­chemotherapy
10. Jin K, Xu J, Chen J, et al. Surgical management for non-­ versus chemotherapy alone for pancreatic cancer with liver metastasis:
functional pancreatic neuroendocrine neoplasms with syn- a propensity score matching study. Expert Rev Gastroenterol Hepatol.
chronous liver metastasis: a consensus from the Chinese Study 2021;1-­10. https://doi.org/10.1080/17474​124.2021.1869937
Group for Neuroendocrine Tumors (CSNET). Int J Oncol. 26. Dong Y, Song Z, Luo Y, et al. A new proposal of utilizing intraop-
2016;49(5):1991-­2000. erative electron radiation therapy on the surface of liver to prevent
11. Margonis GA, Buettner S, Sasaki K, et al. The role of liver-­ postoperative liver metastasis of pancreatic cancer. Med Hypo.
directed surgery in patients with hepatic metastasis from pri- 2019;126:15-­19.
mary breast cancer: a multi-­institutional analysis. HPB (Oxford).
2016;18(8):700-­705.
12. Silberhumer GR, Paty PB, Denton B, et al. Long-­term oncologic
SUPPORTING INFORMATION
outcomes for simultaneous resection of synchronous metastatic
Additional supporting information may be found online in
liver and primary colorectal cancer. Surgery. 2016;160(1):67-­73.
13. Kopetz S, Chang GJ, Overman MJ, et al. Improved survival in the Supporting Information section.
metastatic colorectal cancer is associated with adoption of hepatic
resection and improved chemotherapy. J Clin Oncol. 2009;27(22):​
3677-­3683. How to cite this article: Fu N, Jiang Y, Weng Y, Chen
14. Gu J, Xu Z, Ma Y, et al. Surgical resection of metastatic pancreatic H, Deng X, Shen B. Worth it or not? Primary tumor
cancer: is it worth it? -­a 15-­year experience at a single Chinese resection for stage IV pancreatic cancer patients: A
center. J Gastrointest Oncol. 2020;11(2):319-­328. SEER-­based analysis of 15,836 cases. Cancer Med.
15. Wang L, Yang L, Chen L, et al. Do patients diagnosed with 2021;10:5948–­5963. https://doi.org/10.1002/cam4.4147
metastatic pancreatic cancer benefit from primary tumor sur-
gery? A propensity-­ adjusted, population-­ based Surveillance,