Treatment options in patients with pancreatic cancer: A 10-year multicenter epidemiological investigation in China

Abstract

BACKGROUND

The survival rate of pancreatic cancer is low, and there is a lack of effective treatment.

AIM

To explore the epidemiological characteristics of patients with pancreatic cancer in China and compare multiple chemotherapy regimens at different stages.

METHODS

This was a retrospective study conducted from 2005 to 2014, involving six cancer hospitals and eight general hospitals across seven geographical regions of China (East, South, North, Central, Southwest, Northwest, and Northeast). Stratified sampling was used based on the population distribution of each region. Efficacy assessments were conducted by Cox proportional hazards regression models. When assessing the effectiveness of various chemotherapy regimens, traditional drugs such as gemcitabine used as monotherapy served as the reference.

RESULTS

A total of 3256 patients were included. The median follow-up time was 407 days, and the median overall survival was 183 days. At diagnosis, 56% of patients were already in stage IV. Chemotherapy was administered to 39.73% of patients. In the adjuvant therapy phase, gemcitabine + fluorouracil was superior to gemcitabine monotherapy [hazard ratio (HR) = 0.35, 95% confidence interval (CI): 0.14-0.89]. In fluorouracil-based regimens, other combination regimens did not show effectiveness relative to monotherapy. For first-line treatment in patients with advanced disease, tegafur alone (HR = 0.20, 95%CI: 0.06-0.66), gemcitabine plus cisplatin (HR = 0.16, 95%CI: 0.04-0.70), and tegafur, gemcitabine plus platinum-based agents (HR = 0.32, 95%CI: 0.11-0.91) were associated with a lower risk of death compared to gemcitabine alone. In second-line treatment, there were no significant differences in efficacy among various drugs, but FOLFIRINOX (irinotecan + oxaliplatin + leucovorin + 5-fluorouracil) had an outstanding point estimate (HR = 0.10, 95%CI: 0.01-1.27).

CONCLUSION

In China, pancreatic cancer is often diagnosed at advanced stages, emphasizing the need for early diagnosis and treatment. Combined therapies in adjuvant and first-line settings may reduce the risk of death compared with monotherapy, and FOLFIRINOX might offer more significant benefits in second-line treatment.

Key Words: Pancreatic cancer; Epidemiology; Chemotherapy; Risk factors; Multicenter retrospective study

Core Tip: This study was based on 14 tertiary level A and oncology hospitals in seven major geographic regions in China. We included 3614 patients with pancreatic cancer over a 10-year period to compare the efficacy of different chemotherapy regimens and investigate the influence of risk factors and demographic characteristics on tumor prognosis.

INTRODUCTION

The incidence of pancreatic cancer has been steadily increasing, coupled with a low survival rate, making it one of the most malignant tumors. According to the latest cancer statistics from GLOBOCAN, approximately 496000 new cases and 466000 deaths globally were attributed to pancreatic cancer in 2020, ranking 12^th in cancer incidence and seventh in cancer-related deaths[1]. The annual increase in pancreatic cancer incidence ranges from 0.5% to 1.0%, and it is projected to become the second leading cause of cancer-related deaths by 2030[2]. The 2020 China Cancer Registration Annual Report revealed that pancreatic cancer ranked 13th in cancer incidence, with 30869 new cases, and seventh in cancer-related deaths, with 27907 deaths, among the registered cancer regions in 2017. Five-year overall survival rate for pancreatic cancer globally remains < 10%, and a summary from 17 cancer registration sites in China for 2012-2015 reported a dismal 5-year relative survival rate of 7.2; the lowest among all cancers[3,4]. Pancreatic cancer, manifesting discreetly, often leads to late-stage diagnosis for most patients, which means missing the optimal treatment window[5]. Therefore, effective treatment for pancreatic cancer is crucial to salvage survival. For patients with early-stage disease, surgical intervention is the primary treatment[6]. However, since a significant number of patients is diagnosed late, these patients require chemotherapy to shrink the tumor[7], either to make them eligible for surgery or for late-stage chemotherapy only[8]. Hence, the treatment outcomes for pancreatic cancer, particularly the efficacy of different chemotherapy regimens, significantly influence prognosis. Chemotherapy regimens for pancreatic cancer have been continuously updated in recent years. Currently, FOLFIRINOX (irinotecan + oxaliplatin + leucovorin + 5-fluorouracil) is the main regimens used for neoadjuvant therapy[9]. International guidelines recommend modified FOLFIRINOX (mFOLFIRINOX) and gemcitabine for adjuvant therapy[10], while S-1 is also commonly used in Japan and China[11]. For palliative treatment, FOLFIRINOX, gemcitabine plus nab-paclitaxel, or gemcitabine monotherapy are widely adopted internationally[12]. Currently, there is a paucity of prospective studies systematically evaluating the correlation between different treatment modalities and prognosis in pancreatic cancer. This study conducted a multicenter retrospective analysis of 3256 patients from 14 tertiary or specialized cancer hospitals in seven major regions nationwide, to analyze the current landscape of pancreatic cancer treatment in China.

MATERIALS AND METHODS

Study population

A retrospective collection of cases was conducted from 2005 to 2014, involving six cancer hospitals and eight general hospitals across seven geographical regions of China (East, South, North, Central, Southwest, Northwest and Northeast). Stratified sampling was used based on the population distribution of each region. Within each region, tertiary or specialized oncology hospitals were selected to ensure the acquisition of high-quality case information. Patient records, pathology reports, tissue samples, laboratory tests, imaging reports, and details of inpatient treatments (surgery, radiotherapy, chemotherapy, etc.) were collected through hospital record systems. Pancreatic-cancer-related deaths were determined based on admission records and telephone inquiries. The study adhered to the ethical principles outlined in the Declaration of Helsinki and received approval from the Ethics Committee of the Cancer Hospital of the Chinese Academy of Medical Sciences. Due to the absence of potential harm to the enrolled patients, a waiver of informed consent was obtained.

Inclusion/exclusion criteria

Inclusion criteria: (1) Patients with admission dates between January 1, 2005 and December 31, 2014; (2) Diagnosis of pancreatic cancer through pathology and cytology at the participating hospitals; (3) Records of case information; and (4) Various treatment modalities accepted, including surgery, chemotherapy, radiotherapy, and interventional therapy.

Exclusion criteria: (1) Patients with a history of participation in any anticancer clinical trials; (2) Confirmed cases of neuroendocrine tumors and benign pancreatic tumors, such as cystadenomas; (3) Confirmed cases of benign pancreatic conditions, including chronic pancreatitis, autoimmune pancreatitis, etc.; (4) Confirmed cases of tumors originating outside the pancreas in the periampullary region; (5) Suspected pancreatic cancer that could not be conclusively diagnosed; and (6) Repeated enrollment of patients.

Sampling method

Stratified sampling, convenient sampling, and random sampling were used to ensure the representativeness and randomness of sampling. Stratified sampling was conducted from the seven major regions of China based on population distribution to determine the cases that needed to be collected in each major region. We allocated the number of patients to each hospital and completed the collection of patients between 2005 and 2014. The case collection of each year was randomly selected from all the cases of that year. The tumor staging was determined based on the tumor-node-metastasis staging system.

Quality control

Case information was recorded by hospital physicians after training on the completion of investigation forms and documented in case information reports. Two data entry personnel entered the information into the Epidata database, compiled by the Epidemiology Department of the Cancer Hospital of the Chinese Academy of Medical Sciences. After data entry, Excel software was utilized for consistency and logical checks. Data errors were sent back to the regional centers as error lists, and data entry personnel made necessary improvements and corrections based on medical records until all errors were rectified. We traced the missing data back to the original hospital and had the hospital information department staff retrieve the patient's file for confirmation. Since paper-based cases were still predominant in China from 2005 to 2014, the paper-based materials of some patients were lost. For variables with a large amount of lost information, we will provide a full explanation in the Discussion.

Statistical analyses

Descriptive statistics, such as rates and proportions, were used for categorical data. Cox proportional hazards regression models were used for analyzing risk factors, with hazard ratios (HRs) and 95% confidence intervals (CIs) presented. Model validity was checked by proportional hazards assumption testing. Efficacy assessments were also conducted using Cox proportional hazards regression models, with single-agent therapy as the reference for evaluating specific drug monotherapy or combination therapy. The statistically significant risk factors in the aforementioned analysis were controlled. When assessing the effectiveness of various chemotherapy regimens, traditional drugs such as gemcitabine used as monotherapy served as the reference. P < 0.05 was considered statistically significant, and all data analyses were performed using R language version 4.3.2.

RESULTS

Basic demographic characteristics

Over the course of 10 years, information from 3614 patients was collected. After excluding 34 cases not admitted between 2005 and 2014, 324 cases of neuroendocrine tumors and nonpancreatic malignant tumors, 3256 patients with pancreatic cancer meeting the criteria were included. The median follow-up time was 407 days (range: 367-456 days), and the median survival period was 183 days (range: 155-228 days). The distribution of basic demographic characteristics is illustrated in Table 1. Patients aged < 55 years accounted for 30.90%, and those aged ≥ 75 years constituted 10.10%, with the remaining distributed between 55 and 75 years (58.8%). The majority were male (62.56%). Patients with body mass index (BMI) < 18.5 comprised 16.54%, those with BMI > 23.9 constituted 22.92%, and those with BMI between 18.5 and 23.9 accounted for 60.54%. Patients with a history of smoking or current smokers accounted for 29.82%, and those with a history of alcohol consumption or current drinkers constituted 24.75%. A history of diabetes was present in 15.50% of patients. Only 114 (4.10%) patients had history of pancreatitis, 93 3.70%) had a history of intestinal microbiota imbalance, and 39 (1.43%) had a family history of malignant tumors (98.57%). Only 15.7% of patients were diagnosed at stage I, 15.00% at stage II, 12.40% at stage III, and 56.90% at stage IV. Only 15.14% of patients had highly differentiated tumors, 40.78% had moderately differentiated tumors, and 44.09% had poorly differentiated tumors. Regarding molecular testing, only 15.10% of patients underwent Ki-67 testing, with a positivity rate of 96.03%. Only 14.19% of patients underwent cytokeratin (CK) or AE1/AE3 testing, with a positivity rate of 94.08%.

Table 1 Distribution of basic demographic characteristics.

Characteristic		Case	%
Age at first diagnosis, year
	< 55	549	30.90
	55-64	601	33.80
	65-74	444	25.0
	≥ 75	181	10.10
Sex
	Male	2037	62.56
	Female	1217	37.38
BMI
	< 18.5	197	16.54
	18.5-23.9	721	60.54
	> 23.9	273	22.92
History of pancreatitis
	No	2706	95.90
	Yes	114	4.10
History of intestinal flora disorders
	No	2428	96.30
	Yes	93	3.70
Family history of pancreatic malignancy
	No	2689	98.57
	Yes	39	1.43
Diabetes mellitus
	No	2601	84.50
	Yes	477	15.50
Smoking
	Never	2163	70.18
	Ever	364	11.81
	Now	555	18.01
Alcohol drinking
	Never	2314	75.25
	Ever	296	9.62
	Now	465	15.12
Staging of pancreatic cancer
	Ⅰ	373	15.70
	Ⅱ	358	15.00
	Ⅲ	295	12.40
	Ⅳ	1354	56.90
Differentiation degree
	Well	160	15.14
	Moderately	431	40.78
	Poorly	466	44.09
Surgery
	Yes	2134	66.80
	No	1063	33.20
Surgical margin
	Negative	673	80.00
	Positive	169	20.00
No. of lymph node metastasis
	0	941	73.60
	1-3	270	21.10
	> 3	68	5.30
Chemotherapy
	Yes	1076	39.73
	No	1632	60.26
Ki-67
	Not done	1130	84.90
	Positive	193	14.50
	Negative	8	0.60
CK or AE1/AE3
	Not done	1228	85.81
	Positive	191	13.35
	Negative	12	0.84

BMI: Body mass index; CK: Cytokeratin.

Treatment options

Surgery was performed in 2134 (66.80%) patients, with 80% achieving negative margins and 73.60% having no lymph node metastasis (Table 1). For individuals who opted for surgical intervention, curative surgery accounted for 49.67% and palliative surgery for 27.02% (Table 2). Among the implemented surgical approaches, traditional open surgery constituted 89.74%, while the remaining procedures had lower proportions. Chemotherapy was administered to only 1076 (38.20%) patients: 8.36% received neoadjuvant chemotherapy, 35.5% received adjuvant chemotherapy, and 56.13% received treatment for recurrence or metastasis. Seventy-six patients underwent targeted therapy, with 2.60% receiving it in the neoadjuvant stage, 51.30% in the adjuvant stage, and 46% in the stage of recurrence or metastasis. There were 362 patients who underwent radiotherapy: Synchronous radiotherapy accounted for 16.82%, palliative radiotherapy for 64.02%, postoperative adjuvant radiotherapy for 35.05%, and intraoperative radiotherapy for 38.79%.

Table 2 Treatments options for patients with pancreatic cancer in China from 2005 to 2014.

Treatment modalities		Cases	%
Surgery
Mode of operation
	Radical	1059	49.67
	Palliative	576	27.02
	Other	497	23.31
Mode of execution
	Traditional open surgery	2073	91.85
	Laparoscopic surgery	63	2.79
	Robotic surgery	3	0.13
	Other	118	5.23
Chemotherapy
	Neoadjuvant	90	8.36
	Adjuvant therapy	382	35.50
	Relapse/metastasis treatment	604	56.13
Targeted therapy information
	Neoadjuvant	2	2.60
	Adjuvant therapy	39	51.30
	Relapsing metastasis therapy	35	46.00
Radiation therapy information
	Concurrent	36	16.82
	Postoperative adjuvant	75	35.05
	Palliative	137	64.02
	Intraoperative	83	38.79
	Other	31	14.49

Drug statistics of chemotherapy include single drug and combination drug.

Factors influencing mortality risk

The Schoenfeld residual test was performed on the items included in the multivariate analysis, with P = 0.17, meeting the proportional hazards hypothesis. The older the patient, the higher the risk of pancreatic cancer-related mortality. Using < 55 years as the baseline, in univariate analysis, the HR for the 55-64 years age group was 1.37 (95%CI: 1.08-1.76; P = 0.010), for the 65-74 years age group, it was 1.58 (95%CI: 1.22-2.05; P < 0.001), and for the ≥ 75 years age group, it reached 3.19 (95%CI: 2.34-4.34; P < 0.001) (Table 3). These results remained robust after adjusting for multiple factors. There was no significant difference in pancreatic cancer mortality risk between sex, smoking status and alcohol consumption. Compared to individuals with normal BMI, those with BMI < 18.5 had an increased risk of death, with an HR of 1.60 (95%CI: 1.23-2.07; P < 0.001). After adjusting, the risk increase became more significant (adjusted HR = 2.13, 95%CI: 1.36-3.33; P = 0.002), while the population with BMI > 24 showed no difference from those with normal BMI. A history of pancreatitis increased the risk of pancreatic cancer mortality after adjusting for multiple factors (adjusted HR = 2.07, 95%CI: 1.02-4.19; P = 0.043). No increased risk of pancreatic cancer mortality was found with family history of pancreatic malignant tumors, family history of diabetes, and history of intestinal microbiota imbalance. Compared to stage I tumors, patients with stage III (HR = 2.32, 95%CI: 1.49-3.62; P < 0.001) and stage IV (HR = 2.12, 95%CI: 1.43-3.14; P = 0.001) tumors had an increased risk of death. After adjusting for multiple factors, the risk of death significantly increased for patients with stage III tumors (adjusted HR = 2.76, 95%CI: 1.02-7.44; P = 0.045). Patients with poorly differentiated tumors had an increased risk of death (HR = 2.00, 95%CI: 1.15-3.45; P = 0.014). Patients who underwent surgical treatment had a lower risk of death compared to those who did not (HR = 0.57, 95%CI: 0.47-0.69; P < 0.001, adjusted HR = 0.48, 95%CI: 0.36-0.64; P = 0.036). Patients with negative surgical margins had a lower risk of death (HR = 0.48, 95%CI: 0.33-0.69; P < 0.001, adjusted HR = 0.60, 95%CI: 0.24-0.99; P = 0.033). After adjusting for multiple factors, patients with > 3 metastatic lymph nodes had a significantly increased risk compared to those with no lymph node metastasis (adjusted HR = 4.32, 95%CI: 1.01-18.52; P = 0.007). Chemotherapy was associated with a reduced risk of death compared to those who did not undergo chemotherapy (HR = 0.59, 95%CI: 0.48-0.71; P < 0.001, adjusted HR = 0.58, 95%CI: 0.43-0.78; P < 0.001). Univariate analysis showed that positive result of Ki67 (HR = 3.43, 95%CI: 0.45-26.31; P = 0.235) and CK or AE1/AE3 (HR = 6.39, 95%CI: 0.8-51.0; P = 0.080) had no effect on the risk of death. Due to the small number of people undergoing laboratory tests, we did not include them in the Cox multivariate model.

Table 3 Analysis of risk factors affecting the prognosis of pancreatic cancer.

	Univariate Cox regression		Multivariate Cox regression
	HR (95%CI)	P value	HR (95%CI)	P value
Age, year
< 55	Reference		Reference
55-64	1.37 (1.08-1.76)	0.010	1.80 (1.29-2.50)	0.041
65-74	1.58 (1.22-2.05)	< 0.001	1.45 (1.03-2.04)	0.018
≥ 75	3.19 (2.34-4.34)	< 0.001	2.68 (1.67-4.32)	0.007
Sex
Male	Reference		Reference
Female	0.95 (0.78-1.16)	0.622	0.88 (0.65-1.18)	0.531
BMI
< 18.5	1.60 (1.23-2.07)	0.0003	2.13 (1.36-3.33)	0.002
(18.5-24)	Reference		Reference
(24-28)	0.75 (0.55-1.02)	0.062	0.97 (0.60-1.59)	0.394
≥ 28	0.75 (0.41-1.38)	0.355	0.28 (0.07-1.16)	0.697
History of pancreatitis
No	Reference		Reference
Yes	1.28 (0.76-2.14)	0.349	2.07 (1.02-4.19)	0.043
History of intestinal flora disorders
No	Reference		Reference
Yes	1.03 (0.63-1.67)	0.919	0.93 (0.54-1.61)	0.428
Family history of pancreatic malignancy
No	Reference		Reference
Yes	0.39 (0.10-1.57)	0.185	0.33 (0.14-1.35)	0.371
Smoking
Never	Reference		Reference
Ever	0.75 (0.54-1.05)	0.094	0.60 (0.35-1.04)	0.420
Now	0.97 (0.75-1.25)	0.818	0.88 (0.59-1.34)	0.782
Alcohol consumption
Never	Reference		Reference
Ever	1.00 (0.72-1.38)	0.993	1.10 (0.63-1.91)	0.346
Now	0.81 (0.60-1.09)	0.168	1.50 (0.91-2.46)	0.436
Diabetes mellitus
No	Reference		Reference
Yes	0.89 (0.68-1.18)	0.427	1.13 (0.77-1.66)	0.376
Staging of pancreatic cancer
Ⅰ	Reference		Reference
Ⅱ	1.48 (0.93-2.36)	0.099	1.72 (0.77-3.80)	0.183
Ⅲ	2.32 (1.49-3.62)	< 0.001	2.76 (1.02-7.44)	0.045
Ⅳ	2.12 (1.43-3.14)	0.001	2.03 (0.90-4.56)	0.085
Differentiation degree
Well	Reference		Reference
Moderately	0.91 (0.51-1.63)	0.757	0.98 (0.41-2.32)	0.964
Poorly	2.00 (1.15-3.45)	0.014	1.35 (0.58-3.13)	0.480
Surgery
No	Reference		Reference
Yes	0.57 (0.47-0.69)	< 0.001	0.48 (0.36-0.64)	0.036
Surgical margin
Positive	Reference		Reference
Negative	0.48 (0.33-0.69)	< 0.001	0.60 (0.24-0.99)	0.033
No. of lymph node metastasis
0	Reference		Reference
1-3	1.12 (0.84-1.49)	0.441	1.02 (0.56-1.87)	0.939
> 3	0.82 (0.46-1.46)	0.496	4.32 (1.01-18.52)	0.007
Chemotherapy
No	Reference		Reference
Yes	0.59 (0.48-0.71)	< 0.001	0.35 (0.21-0.60)	< 0.001
Ki-67
Negative	Reference		--
Positive	3.43 (0.45-26.31)	0.235	--
CK or AE1/AE3
Negative	Reference		--
Positive	6.39 (0.8-51.00)	0.080	--

Bold values indicated P < 0.05, which was statistically significant. Ki-67 and CK or AE1/AE3 were not included in multivariate Cox regression due to the small number of tested people. For the variables with P < 0.1 in the univariate analysis, we considered them to be factors that might affect the prognosis of pancreatic cancer and included them in the multivariate analysis. However, although other items had P > 0.01, they are still relevant risk factors suggested by previous studies, such as the history of pancreatitis and the history of intestinal flora disorders; therefore, these variables were included in the multivariate model. BMI: Body mass index; CI: Confidence interval; CK: Cytokeratin; HR: Hazard ratio.

Chemotherapeutic efficacy

Due to the limited number of patients undergoing neoadjuvant chemotherapy, the recorded treatment regimens were insufficient for data analysis. Therefore, efficacy comparisons were only conducted for adjuvant and late-stage chemotherapy (Table 4). In terms of adjuvant treatment, the efficacy of gemcitabine in combination with other drugs was also superior to monotherapy (HR = 0.40, 95%CI: 0.18-0.86; P = 0.019). Of all protocols, only gemcitabine + fluorouracil demonstrated superior efficacy compared to gemcitabine monotherapy (HR = 0.35, 95%CI: 0.14-0.89; P = 0.028). In fluorouracil-based regimens, other combination regimens have not shown effectiveness relative to monotherapy. In terms of first-line treatment for patients with advanced-stage disease, compared to gemcitabine monotherapy, the use of tegafur alone (HR = 0.02, 95%CI: 0.06-0.66; P = 0.008), gemcitabine + cisplatin (HR = 0.16, 95%CI: 0.04-0.70; P = 0.015), and gemcitabine + tegafur + cisplatin (HR = 0.32, 95%CI: 0.11-0.91; P = 0.033) were more effective in reducing the risk of patient mortality. In the second-line treatment for patients with advanced-stage disease, there were no significant differences in efficacy among other regimens compared to tegafur monotherapy.

Table 4 Comparison of efficacy of different chemotherapy regimens.

Adjuvant therapy (case number)		P value	HR (95%CI)	Relapse/metastasis treatment (case number)		P value	HR (95%CI)
1. Fluorouracil1	Monotherapy (15)		1.00	1. First-line	Gemcitabine monotherapy (40)		1.00
	Combination (73)	0.163	0.40 (0. 11-1.44)		Tegafur + gemcitabine + platinum (2)	0.343	0.38 (0.05-2.82)
2. Gemcitabine	Monotherapy (27)		1.00		Tegafur + monotherapy (9)	0.008	0.20 (0.06-0.66)
	Combination (80)	0.019	0.40 (0.18-0.86)		Gemcitabine + albumin-bound paclitaxel (20)	0.929	0.97 (0.54-1.76)
3. All regimens	Gemcitabine monotherapy (27)		1.00		Gemcitabine + cisplatin (18)	0.015	0.16 (0.04-0.70)
	Fluorouracil monotherapy (14)	0.720	0.79 (0.22-2.89)		Gemcitabine + oxaliplatin (53)	0.788	0.94 (0.58-1.52)
	Gemcitabine + fluorouracil (38)	0.028	0.35 (0.14-0.89)		Gemcitabine + 5-FU (15)	0.577	0.80 (0.37-1.74)
	Gemcitabine + platinum (39)	0.998	0.48 (0.20-1.19)		Gemcitabine + capecitabine (21)	0.912	0.97 (0.54-1.74)
	Gemcitabine + fluorouracil + platinum2 (6)	0.113	0.19 (0.02-1.51)		Tegafur + gemcitabine + platinum (12)	0.033	0.32 (0.11-0.91)
	Fluorouracil + platinum (21)	0.117	0.29 (0.08-1.06)		FOLFIRINOX (8)	0.854	0.93 (0.42-2.07)
4. Gemcitabine-based regimens	Monotherapy (27)		1.0	2. Second-line	Tegafur monotherapy (6)		1.00
	Gemcitabine + fluorouracil (38)	0.030	0.35(0.14-0.90)		Tegafur + oxaliplatin (9)	0.278	0.49 (0.14-1.76)
	Gemcitabine + platinum (39)	0.113	0.48 (0.20-1.19)		Gemcitabine + albumin-bound paclitaxel (2)	0.343	0.35 (0.04-3.08)
	Gemcitabine + fluorouracil + platinum (6)	0.124	0.20 (0.02-1.56)		Gemcitabine + oxaliplatin (3)	0.825	1.28 (0.14-11.52)
5. Fluorouracil-based regimens	Monotherapy (14)		1.00		FOLFIRINOX (5)	0.099	0.10 (0.01-1.27)
	Gemcitabine + fluorouracil (38)	0.219	0.43 (0.11-1.65)		Oxaliplatin + 5-FU + catarabine (6)	0.943	1.05 (0.31-3.56)
	Gemcitabine + fluorouracil + platinum (6)	0.212	0.24 (0.02-2.28)		Oxaliplatin + catarabine (6)	0.892	1.10 (0.29-4.16)
	Fluorouracil + platinum (21)	0.217	0.36 (0.07-1.82)

¹Fluorouracil includes tegafur, 5-FU, and capecitabine therapies.

²Platinum includes cis-platinum and oxaliplatin.

Bold values indicated P < 0.05, which was statistically significant. The confounding factors identified in Table 2 were controlled, including age, body mass index, and history of pancreatitis. The treatment options and patient stages were not controlled because the stages and choices of patients were differentiated through different chemotherapy stages. CI: Confidence interval; HR: Hazard ratio.

DISCUSSION

Data collected from 14 hospitals across seven major regions in China revealed that the median survival period for patients with pancreatic cancer was 183 days. The male population (62.56%) outnumbered the female population (37.38%), aligning with a concurrent study in Shanghai that reported a median survival period of 7.8 months[13]. A significant proportion (56%) of patients were diagnosed at stage IV, and well-differentiated tumors accounted for only 15.14%. This suggests that pancreatic cancer is often detected late in Chinese patients and is predominantly of intermediate to low differentiation, consistent with previous domestic studies on pancreatic cancer[14]. This late detection is associated with the inconspicuous clinical features of pancreatic cancer and its rapid progression[15]. A history of intestinal microbiota disorders and a family history of pancreatic cancer were low in our study population, so we did not find that these factors increase the risk of pancreatic cancer. A history of pancreatitis has emerged as a risk factor for pancreatic cancer, aligning with the understanding that chronic pancreatitis can induce tumor development through inflammation-related pathways[16]. Large-scale studies from the United States National Institutes of Health in the United States identified type 2 diabetes, obesity (BMI ≥ 30), and smoking as risk factors for pancreatic cancer[17]; however, our study did not find conclusive evidence supporting these associations, possibly due to insufficient data on diabetes subtypes. The population of smokers and drinkers in China may represent individuals with higher economic capabilities and social status, and in some studies, have even been identified as protective factors against cancer[18]. Our study discovered an increased risk of death among patients with low BMI (< 18.5); a finding not commonly reported in international research. Some studies have suggested that low body weight or underweight is unrelated to overall survival rates in patients with pancreatic cancer[19], while others consider low BMI as an adverse prognostic factor[20]. One possible explanation from our study was that 56% of patients were diagnosed at stage IV, and patients with late-stage pancreatic cancer often present with significant weight loss. Therefore, individuals with low BMI may have a certain association with advanced-stage tumors, contributing to an increased risk of death.

We found that 66.8% of patients underwent surgery, with 80% having negative margins and 73.6% showing no lymph node metastasis. However, the survival rate in our cohort remained low, which contradicts the expected prognosis given these favorable surgical outcomes. Among 1059 patients who received radical surgery, only 24% were stage IV. Of these, just 26.2% had negative margins, and only 23.9% with no lymph node metastasis were patients with stage IV tumor. Among patients with missing surgical data, stage IV cases accounted for 41.2% (missing margin data) and 45% (missing lymph node data). This suggests that many advanced cases were underrepresented in the surgical dataset, while patients with early-stage disease were more likely to have complete records-leading to overestimation of radical resection rates, negative margins, and nonmetastatic nodes. As a result, these estimates are biased and lack reference value. Nearly 50% of operations were radical and 27.02% were palliative. Of these, 91.85% were open procedures. Some studies suggest that compared to open surgery, laparoscopic procedures are associated with reduced blood loss, shorter hospital stays, and improved disease-free survival[21]. In the future, China should consider transitioning from open surgery to laparoscopic or robotic forms to enhance postoperative outcomes. Further analysis indicates that surgery reduces the risk by 42% compared to nonsurgical treatment, and negative surgical margins decrease the risk by 40% compared to positive margins. Patients with > 3 Lymph node metastases have a 4.32 times higher risk of death than those without lymph node metastasis. A meta-analysis of pancreatic cancer studies across several countries showed that achieving R0 resection ranged from 70% to 80%, and R0 was associated with a 12%-23% reduced risk compared to R1 resection[22]. Another meta-analysis found that lymph node metastasis was associated with poor prognosis[23]. That study, based on nationwide data in China, discovered that the rate of negative surgical margins for pancreatic cancer surgery was relatively high globally. Consistent with foreign studies, this research also identified negative surgical margins and absence of lymph node metastasis as important protective factors for pancreatic cancer prognosis.

The number of patients undergoing laboratory testing in China between 2005 and 2014 was low, with detection rates for Ki-67 and CK or AE1/AE3 at about 15%. However, both tests had a high positive rate, exceeding 90%. We conducted a trend analysis of the number of people undergoing laboratory tests from 2005 to 2014. The number of people undergoing Ki-67 test increased from 141 in 2005 to 315 in 2014, and people undergoing CK or AE1/AE3 increased from 131 to 316. Therefore, during the 10 years of the research being carried out, the number of people conducting experimental tests has also been gradually increasing. Many studies suggest that Ki-67 and CK or AE1/AE3 are important indicators for differentiating and staging pancreatic cancer, and their positive expression may suggest the invasiveness and metastatic ability of tumor cells[8,24,25]. This study indicates a high positive rate for Ki-67 and CK or AE1/AE3 in Chinese pancreatic cancer patients, potentially increasing the risk of death. It underscores the need to increase laboratory testing rates for patients with pancreatic cancer in China to provide better prognostic predictions.

Among all patients, 39.73% received chemotherapy, with 8.36% undergoing neoadjuvant treatment, 35.50% receiving adjuvant treatment, and 56.13% undergoing treatment for recurrence or metastasis. Due to the lack of detailed treatment information for most neoadjuvant therapies, the analysis of chemotherapy regimens and prognosis in our study focused only on adjuvant and recurrence/metastasis treatments. In China, the recommended adjuvant treatment regimens are based on gemcitabine or fluoropyrimidine drugs (5-FU, capecitabine, or tegafur)[26]. Our nationwide study found that combining drugs produced better results compared to monotherapy with gemcitabine. Gemcitabine plus fluorouracil was found to be superior to the reference regimen based on gemcitabine monotherapy. Similar results were reported in the ESPAC-4 study, which found a median overall survival of 28.0 months in the gemcitabine combined with capecitabine group and the median overall survival in the gemcitabine group was 25.5 (22.7-27.9) months. In fluorouracil-based regimens, other combination regimens have not shown effectiveness relative to monotherapy. In the context of recurrence and metastasis treatment, this study suggests that, in first-line treatment regimens, compared with gemcitabine alone, tegafur alone, gemcitabine + cisplatin combined therapy, tegafur + gemcitabine + platinum combined therapy can reduce the risk of death in patients. In China, common first-line treatment regimens include two-drug combinations with gemcitabine and three-drug combinations like FOLFIRINOX or mFOLFIRINOX[26]. FOLFIRINOX has been proven superior to gemcitabine monotherapy in large-scale randomized controlled trials[27], but it was not widely used in the Chinese population during the study period (2005-2014). Therefore, the study did not have a sufficient sample size to measure an effect of FOLFIRINOX. The study suggests that monotherapy with tegafur therapy can reduce the risk of patient mortality, consistent with a meta-analysis showing the superiority of the gemcitabine + tegafur regimen in treating advanced or metastatic pancreatic cancer, surpassing gemcitabine monotherapy[28]. Currently, there is a lack of direct comparisons between monotherapy with tegafur therapy and gemcitabine + oxaliplatin therapy, warranting further investigation. Gemcitabine + cisplatin indicated better clinical efficacy in our study, compared with gemcitabine monotherapy. A meta-analysis suggests that in locally advanced patients, two-drug combination regimens, such as gemcitabine + cisplatin, yield longer overall survival compared to gemcitabine monotherapy[29]. Tegafur + gemcitabine + platinum therapy in our study was superior to gemcitabine monotherapy. There are no direct clinical trials or meta-analyses to demonstrate superior efficacy of the combination of these three drugs, and future studies are needed. Regarding second-line treatment regimens, our study suggests that, compared to monotherapy with tegafur, other drugs do not show significant differences. The primary reason is the limited information on second-line treatment patients, with an insufficient sample size leading to wide CI and an inability to find significant differences. However, the point estimates indicate a trend towards a reduced risk of patient mortality with FOLFIRINOX compared to the reference regimen. Domestic guidelines suggest that for patients who have failed first-line treatment, nonoverlapping drugs can be chosen for second-line chemotherapy based on drugs previously used, patient comorbidities, and toxic side effects[26]. Our multicenter study using nationwide data also suggests that FOLFIRINOX provides significant benefits for second-line treatment of pancreatic cancer. The treatment regimens for pancreatic cancer have undergone significant changes in the past decade. For patients with metastatic or advanced pancreatic cancer, treatment has shifted from previous gemcitabine-based monotherapy to primarily FOLFIRINOX or gemcitabine plus nab-paclitaxel, with a growing emphasis on individualized first-line treatment strategies[27,30]. In the adjuvant setting, gemcitabine was once the standard, but mFOLFIRINOX has now become the preferred regimen, aiming to improve overall survival[31]. Neoadjuvant chemotherapy was rarely used in the past, but in recent years, neoadjuvant FOLFIRINOX has shown increasing advantages, particularly in improving surgical success rates[32]. While there were no standard targeted or immunotherapy options a decade ago, a small subset of patients today may benefit from PARP inhibitors or PD-1 checkpoint inhibitors[33,34]. FOLFIRINOX and its modified regimens have gradually gained international recognition since 2011. A clinical trial conducted in 2011 demonstrated that, compared with gemcitabine, FOLFIRINOX offers a survival advantage in patients with metastatic pancreatic cancer, albeit with increased toxicity[27]. The final results of two clinical trials, PRODIGE 4/ACCORD 11, established this regimen as a guideline-recommended treatment for advanced and metastatic pancreatic cancer[35]. Since 2018, the use of FOLFIRINOX in adjuvant therapy has also been validated by clinical trials, primarily through the PRODIGE 24/CCTG PA.6 studies[10]. Our research failed to generate evidence based on FOLFIRINOX. However, during the period from 2005 to 2014, other combination therapies showed advantages over the classic gemcitabine monotherapy, which might provide a new direction for future research. There were two main limitations to this study. Firstly, the research was based on real data from Chinese patients with pancreatic cancer from 2005 to 2014, and there is a certain time difference before now, so there is a certain lag. Secondly, at that time, neoadjuvant and targeted therapies were not commonly used, and there were insufficient data on neoadjuvant chemotherapy regimens. This limitation resulted in the absence of a comparison of neoadjuvant regimens in this study. Data sets were not uniform due to missing data. At the time of our study, Chinese hospitals were still in the era of paper-based patient records, and there was a problem of missing treatment records for some patients. Therefore, we do not have access to information on all subjects in the treatment selection and related data analysis. Furthermore, in items such as lymph node metastasis and surgical margins, our further analysis revealed that the patients with missing information were mainly those at the advanced stage, which also led to the distribution of some items being inconsistent with that of previous studies.

CONCLUSION

Based on information collected from 3614 patients in 14 hospitals across seven regions in China, this study reveals that the median survival period for pancreatic cancer was 183 d. More than half of the patients were diagnosed at stage IV, indicating late discovery of pancreatic cancer. Future efforts should focus on increasing early diagnosis and treatment. Patients with low BMI faced an increased risk of death. There were 66.8% of patients who underwent surgery, with 80% having negative margins, suggesting a high level of surgical proficiency in China. The Ki-67 and CK or AE1/AE3 positivity rates were high and potentially increased the risk of death. However, current laboratory testing rates are low. In terms of adjuvant therapy, combination therapy appeared to produce better outcomes compared to monotherapy with tegafur or gemcitabine. In first-line treatment, compared to gemcitabine alone, monotherapy with tegafur, gemcitabine plus cisplatin, and a combination of tegafur, gemcitabine, and platinum drugs reduced the risk of death. In second-line treatment, FOLFIRINOX may also provide more significant benefits to patients with pancreatic cancer.

ACKNOWLEDGEMENTS

We sincerely thank Xiaoting Ma for her valuable support in the chemotherapy-related aspects of pancreatic cancer. Her insights and guidance greatly contributed to the development and interpretation of this study.