Construction and Validation of Risk Factors and Prognostic Model for Liver Metastasis of Pancreatic Cancer

doi:10.21203/rs.3.rs-3244451/v1

Background

Pancreatic cancer (PC) is a common malignancy that often metastasizes to the liver. The presence of liver metastasis (LM) in PC significantly impacts treatment selection and prognosis, but factors affecting the occurrence and prognosis of pancreatic cancer with liver metastasis (PCLM) are not well described.

Methods

Patients diagnosed with PC between 2010 and 2015 were selected from the Surveillance Epidemiology and End Results (SEER) database. Independent risk factors for PCLM were identified using univariable and multivariable logistic regression. Independent prognostic factors affecting the overall survival (OS) of PCLM patients were analyzed by univariate and multivariate Cox regression, and two nomograms were constructed to predict the risk and prognosis of PCLM. Nomograms were evaluated by receiver operating characteristic (ROC) analysis, C-index, calibration plots, and decision curve analysis (DCA).

Results

Multivariate logistic regression showed that age, primary site, grade, histological subtype, N stage, radiotherapy, surgery, bone metastasis, and lung metastasis were independent risk factors for PCLM. Multivariable COX regression showed that age, grade, histological subtype, surgery, radiotherapy, chemotherapy and lung metastasis were independent prognostic factors for PCLM. Diagnostic and prognostic nomograms were constructed based on the ROC curves, C-index, calibration curves and DCA curves, and both nomograms showed good predictive performance and clinical utility.

Conclusion

The two nomograms constructed in this study exhibit good predictive performance in the risk and prognosis of PCLM and may thus serve as a guide for future clinical management of PCLM.

Pancreatic cancer

liver metastasis

predictors

nomogram

Surveillance Epidemiology and End Results (SEER) database

Pancreatic cancer (PC) is a common malignancy of the digestive system. According to the International Agency for Research on Cancer (IARC) GLOBOCAN cancer incidence and mortality estimates, there were approximately 500,000 new PC cases and 460,000 PC deaths in 2020, accounting for 2.6% of new cancer cases and 4.7% of all cancer deaths worldwide(1). PC is the seventh leading cause of cancer death in the world and sixth in China(1, 2). Unlike other cancers, PC incidence has been on the rise while its mortality rate remains high. Compared with other cancers, the 5-year survival rate of PC has been very low at only 8%(3). It was reported that the median survival is 6–10 months for locally advanced PC, and 3–6 months for patients with metastases(4).

The early symptoms of PC are insidious and difficult to detect, and many patients have already developed metastases by the time of diagnosis. The incidences of liver, lung, bone, and brain metastasis from PC are 45.1%, 11.4%, 3.8%, and 0.4%, respectively, suggesting that the liver is the most common organ of distant PC metastasis(5). Studies have shown that the organs of metastasis exhibit microscopic changes that favor metastatic growth, known as "pre-metastatic niches (PMNs) ". These PMNs are established by various signals originating from the primary PC, which allow cancer cells to grow once they have spread(6). The infiltration of immunosuppressive cells, such as tumor-associated macrophages, myeloid-derived suppressor cells, and regulatory T cells, impedes CD8 + T cell function and thereby promotes tumor growth, invasion, migration, and metastasis(7). Zhang et al.(8) found that PC with liver metastasis (PCLM) mobilizes and recruits a population of P2RX1-negative neutrophils that overexpress immunosuppressive molecules such as PD-L1 and Arg, and thus mediate the formation of an immunosuppressive tumor microenvironment.

Before reaching the target organ, tumor cells produce a wide range of tumor-derived secretory factors, such as exosomes, soluble growth factors and cytokines, which not only regulate the cells of the recipient organs but also recruit additional cells from a third site such as the bone marrow. Bone marrow-derived cells (BMDCs) are mobilized into the blood and seeded into the target organs before the tumor cells arrive, forming local PMNs in the target organs that are suitable for tumor growth(6). Exosomes are extracellular vesicles ranging from 40 to 150 nm in diameter and are secreted by a variety of cells within the tumor. A preclinical study(9) demonstrated that intravenous injection of exosomes isolated from the mouse PDAC cell line PAN02 induced the formation of hepatic PMNs, and chemokines contained in the tumor exosomes led to fibronectin production and increased BMDC retention in the liver, resulting in the release of transforming growth factor-β from hepatic macrophages and consequently enhanced metastatic colonization of PAN02 cells. The formation of PMNs is the initial step in the progression of metastasis. Distant metastasis to a specific target organ requires the co-adaptation of disseminated tumor cells from the primary tumor and resident cells of the target organ(10), and it is only when such co-adaptation is achieved can the metastatic tumor foci grow successfully.

The basic mode of PCLM is dissemination through the portal system, which is formed by the confluence of the superior mesenteric vein and splenic vein at the back of the pancreatic neck and accounts for 2/3 of the hepatic blood flow. When PC lesions encircle and invade the superior mesenteric vein and splenic vein, tumor cells are prone to enter the portal circulation and lead to LM. LM is found in 70% of patients with advanced PC (11), and most patients have multiple intrahepatic metastases. When the intrahepatic metastases are small, patients usually have no obvious symptoms and no significant abnormalities in liver function. However, as intrahepatic metastases gradually increase, patients often feel discomfort in the right upper abdomen, vague pain, and soreness, and may later develop pain, ascites and jaundice. PCLM tends to have a worse prognosis compared to other metastatic sites(12). Based on this, the current study retrospectively analyzed the clinical data of PC patients to identify the independent risk factors and prognostic factors for PCLM, and established nomograms for early PCLM prediction and guidance of clinical treatment selection.

2.1 Patient Selection

The data for this study were obtained from the SEER database using SEER∗Stat v8.4.0. SEER is a leading cancer surveillance database that collects cancer treatment follow-up data from approximately 30% of the U.S. population(13), and these data represent the epidemiologic status of that population. The inclusion criteria for this study were: (1) PC confirmed by pathology and liver metastases diagnosed by imaging; (2) pancreas as the only primary site; (3) complete follow-up data. The following cases were excluded: (1) cases diagnosed by death certificate or autopsy; (2) cases with missing or incomplete clinicopathological features. A final total of 11,529 patients diagnosed with PC, including 2,149 with PCLM, were included in this study from the SEER database. In addition, 178 patients who were diagnosed with PC (including 64 PCLM patients) at the Affiliated Hospital of Xuzhou Medical University between January 2012 and June 2021 were also included in the study using the same inclusion and exclusion criteria. The use of patient clinical data in the external validation group was approved by the institutional Ethics Committee.

2.2 Data Elements

The age, sex, race, primary site, grade, histological subtype, tumor size, TNM stage, treatment status (surgery, radiotherapy, chemotherapy), brain metastasis, bone metastasis, lung metastasis, liver metastasis, and survival time of patients were analyzed. Histological subtypes were classified according to the International Classification of Diseases in Oncology (ICD-O) into four categories: adenocarcinoma (histologic codes 8140, 8480), neuroendocrine carcinoma (histologic code 8246), infiltrating duct carcinoma (histologic code 8500) and others (histologic codes 8010, 8012, 8013, 8020, 8021, 8041, 8046, 8070, 8150, 8240, 8244,8249, 8481, 8490, 8560). Survival analysis was performed based on overall survival (OS), which is defined as the date from diagnosis to death (of any cause) or last follow-up visit.

2.3 Statistical Analysis

Data were analyzed using SPSS 26.0 and R software (version 4.2.0). Continuous data were compared using independent t-test or Mann-Whitney U test, and categorical data were compared using the chi-square test or Fisher's exact test. All variables were analyzed by univariate logistic regression, and those with P < 0.05 were included in the multivariate logistic regression to identify independent risk factors for PCLM. The 2149 PCLM patients were divided into the training group (n = 1612) and internal validation group (n = 537) at a 3:1 ratio. Variables in the training group were analyzed by univariate Cox proportional regression, and those with P < 0.05 were included in the multivariate Cox proportional regression to determine independent prognostic factors for PCLM. Two nomograms were then constructed based on these independent risk and prognostic factors. The accuracy of the nomograms was assessed by C-index and ROC, and their discriminability was verified with calibration plots. Clinical utility of the nomograms was evaluated by decision curve analysis (DCA). A P < 0.05 (two-sided) was considered statistically significant.

3.1 Patient Characteristics

We included 11,529 PC patients with complete baseline data from the SEER database, of which 2149 patients had PCLM. In the SEER database, 51.96% were males and 79.53% were Caucasians. The head of the pancreas was the most common primary site (59.49%), grade II (40.98%) was the most common degree of differentiation, and adenocarcinoma (54.87%) was the most common histological subtype. The most common tumor size was 2-4cm, accounting for 50.09% of all tumor sizes. In addition, the most common T and N stages were T3 (57.26%) and N1 (50.37%), respectively. In terms of treatment, 7109 cases (61.66%) underwent surgery, 2506 cases (21.74%) received radiotherapy, and 6854 cases (59.45%) received chemotherapy. Among the distant metastases of PC, 13 patients (0.11%) had brain metastases, 168 (1.46%) had bone metastases, and 500 (4.34%) had lung metastases. Aside from the SEER database, we also included 178 Chinese PC patients from our hospital, 64 of whom had PCLM (35.95%). Table 1 shows the clinicopathological characteristics and treatment options of PC patients in the SEER database and our Chinese cohort.

Table 1

Baseline clinical features and treatment regimens of pancreatic carcinoma patients in SEER database and Chinese cohort
	SEER database (n = 11529)			Chinese cohort (n = 178)
Characteristic	Without LM (n = 9380)	With LM (n = 2149)	P	Without LM (n = 114)	With LM (n = 64)	P
Age			0.001 ^a			0.013^b
Median/\(\stackrel{-}{x}\)±SD	66(58–74)	65(57–73)		63.86 ± 10.007	59.92 ± 10.081
Sex			< 0.001			0.813
Female	4606	933		43	23
Male	4774	1216		71	41
Race			0.003			NA
White	7463	1706		NA	NA
Black	974	265		NA	NA
Others	943	178		NA	NA
Primary site			< 0.001			0.002
Head of pancreas	5955	904		55	13
Body of pancreas	1035	341		6	9
Tail of pancreas	1271	510		8	13
Pancreatic duct	42	9		NA	NA
Other specified parts of pancreas	146	31		11	6
Overlapping lesion of pancreas	553	216		27	18
Pancreas, NOS	378	138		7	5
Grade			< 0.001			< 0.001
Highly differentiated: I	2124	235		5	2
Moderately differentiated: II	3990	735		56	12
Poorly differentiated: III	3108	1086		53	49
Undifferentiated; anaplastic: IV	158	93		0	1
Histological subtype			< 0.001			< 0.001
Adenocarcinoma	4857	1469		67	51
Infiltrating duct carcinoma	2654	162		45	9
Neuroendocrine carcinoma	664	224		0	3
Other	1205	294		2	1
Tumor size			< 0.001			0.180
≤ 2	1601	117		5	2
2 ~ 4	4910	865		55	22
>4	2869	1167		54	40
AJCC T stage			< 0.001			0.003
T1	930	71		4	2
T2	1520	727		49	18
T3	5761	841		49	23
T4	1169	510		12	21
AJCC N stage			< 0.001			0.044
N0	4508	1214		57	22
N1	4872	935		57	42
Surgery			< 0.001			< 0.001
No	2570	1850		4	44
Yes	6810	299		110	20
Radiotherapy			< 0.001			0.106
No	6999	2024		101	51
Yes	2381	125		13	13
Chemotherapy			0.919			0.139
No/Unknown	3801	874		20	6
Yes	5579	1275		94	58
Brain metastasis			< 0.001			1.000
No	9376	2140		113	63
Yes	4	9		1	1
Lung metastasis			< 0.001			0.090
No	9177	1852		109	56
Yes	203	297		5	8
Bone metastasis			< 0.001			0.048
No	9326	2035		112	58
Yes	54	114		2	6
^a Mann-Whitney U test
^b t test

3.2 Independent Risk Factors for PCLM

To identify factors that predispose a patient to develop PCLM, we performed univariate and multivariate logistic regression of all variables, and found that age, primary site, grade, histological subtype, N stage, surgery, radiotherapy, lung metastasis, and bone metastasis were independent predictors for PCLM (P < 0.05) (Table 2).

Table 2

Univariate and multivariate logistic regression of risk factors for liver metastasis in patients with pancreatic cancer.
Characteristic	Univariate analysis		Multivariate analysis
Characteristic	OR (95%CI)	P	OR (95%CI)	P
Age	0.994(0.990–0.998)	0.004	0.976(0.971–0.981)	< 0.001
Sex
Female	Reference		Reference
Male	1.257 (1.144–1.382)	< 0.001	1.123(0.999–1.263)	0.053
Race
White	Reference		Reference
Black	1.190(1.029–1.377)	0.019	1.003(0.836–1.203)	0.971
Others ^a	0.826(0.698–0.977)	0.026	0.787(0.640–0.967)	0.023
Primary site
Head of pancreas	Reference		Reference
Body of pancreas	2.170(1.885–2.499)	< 0.001	1.366(1.148–1.625)	< 0.001
Tail of pancreas	2.643(2.334–2.993)	< 0.001	2.406(2.038–2.841)	< 0.001
Pancreatic duct	1.412(0.685–2.910)	0.350	2.349(0.907–6.088)	0.079
Other specified parts of pancreas	1.399(0.943–2.074)	0.095	0.935(0.587–1.489)	0.776
Overlapping lesion of pancreas	2.573(2.166–3.056)	< 0.001	1.322(1.068–1.635)	0.010
Pancreas, NOS	2.405(1.955–2.958)	< 0.001	1.357(1.052–1.750)	0.019
Grade
Highly differentiated: I	Reference		Reference
Moderately differentiated: II	1.665(1.424–1.946)	< 0.001	2.231(1.833–2.716)	< 0.001
Poorly differentiated: III	3.158(2.714–3.674)	< 0.001	3.065(2.522–3.724)	< 0.001
Undifferentiated; anaplastic: IV	5.320(3.983–7.106)	< 0.001	3.416(2.367–4.931)	< 0.001
Histological subtype
Adenocarcinoma	Reference		Reference
Infiltrating duct carcinoma	0.197(0.167–0.233)	< 0.001	0.733(0.597–0.899)	0.003
Neuroendocrine carcinoma	1.083(0.923–1.271)	0.328	3.001(2.370–3.801)	< 0.001
Other	0.802(0.700-0.919)	0.002	1.207(1.000-1.458)	0.050
Tumor size
≤ 2	Reference		Reference
2 ~ 4	2.411(1.971–2.948)	< 0.001	1.345(0.948–1.909)	0.097
>4	5.566(4.559–6.796)	< 0.001	1.855(1.309–2.629)	0.001
AJCC T stage
T1	Reference		Reference
T2	6.265(4.845–8.101)	< 0.001	1.862(1.188–2.916)	0.007
T3	1.912(1.486–2.460)	< 0.001	1.216(0.786–1.881)	0.379
T4	5.715(4.394–7.432)	< 0.001	1.009(0.643–1.583)	0.969
AJCC N stage
N0	Reference		Reference
N1	0.713(0.648–0.783)	< 0.001	1.293(1.140–1.466)	< 0.001
Surgery
No	Reference		Reference
Yes	0.061(0.054–0.069)	< 0.001	0.066(0.056–0.078)	< 0.001
Radiotherapy
No	Reference		Reference
Yes	0.182(0.151–0.219)	< 0.001	0.192(0.155–0.238)	< 0.001
Chemotherapy
No/Unknown	Reference
Yes	0.994(0.903–1.093)	0.900
Brain metastasis
No	Reference		Reference
Yes	9.858(3.033–32.041)	< 0.001	1.628(0.432–6.136)	0.471
Lung metastasis
No	Reference		Reference
Yes	7.250(6.023–8.726)	< 0.001	1.907(1.536–2.368)	< 0.001
Bone metastasis
No	Reference		Reference
Yes	9.675(6.974–13.421)	< 0.001	3.696(2.476–5.517)	< 0.001

3.3 Establishment and Validation of Diagnostic Nomogram

Next, we constructed a predictive nomogram for PCLM based on the independent predictors identified by multivariate logistic regression (Fig. 1), and created a user-friendly free online version of the model (https://pclm.shinyapps.io/DynNomapp/). The AUC of the nomogram was 0.877 (0.869–0.884), indicating good predictive accuracy (Fig. 2A). Comparison with the calibration curve also indicated good agreement between the model-predicted probability and actual probability (Fig. 2B). In addition, the DCA curve demonstrated that the nomogram is of clinical utility (Fig. 2C). Collectively, our results indicate that the nomogram has better predictive accuracy and clinical utility than TNM staging.

The Chinese PC cohort from our hospital was used for external validation of the model. The AUC of ROC was 0.846 (0.777–0.915), confirming that the nomogram has good predictive accuracy (Fig. 3A). Furthermore, the calibration curve also revealed good agreement between the predicted and actual probabilities of the model (Fig. 3B). The externally validated DCA curve shows that the model has considerable clinical utility (Fig. 3C).

3.4 Independent Prognostic Factors for PCLM

The 2149 PCLM patients were randomly divided into the training and validation cohorts at a 3:1 ratio, and the clinical characteristics and treatment regimens of the two cohorts are summarized in Table 3. All variables were not significantly different between the training and validation cohorts. Univariate and multivariate COX regression showed that age, grade, histological subtype, surgery, radiotherapy, chemotherapy and lung metastasis were independent prognostic factors for PCLM in the training cohort (Table 4).

Table 3

Demographic and clinicopathological characteristics of patients with PCLM.
Characteristic	Training cohort (N = 1612)		Internal validation cohort (N = 537)		Z/χ²	P
Characteristic	n	%	n	%	Z/χ²	P
Age					-0.482	0.630
Median	65(57–74)		65(57–73)
Sex					0.218	0.641
Female	705	43.7	228	42.5
Male	907	56.3	309	57.5
Race					1.986	0.371
White	1291	80.1	415	77.3
Black	193	12.0	72	13.4
Others ^a	128	7.9	50	9.3
Primary site					3.780	0.710
Head of pancreas	686	42.6	218	40.6
Body of pancreas	260	16.1	81	15.1
Tail of pancreas	372	23.1	138	25.7
Pancreatic duct	6	0.4	3	0.6
Other specified parts of pancreas	23	1.4	8	1.5
Overlapping lesion of pancreas	157	9.7	59	11.0
Pancreas, NOS	108	6.7	30	5.6
Grade					4.444	0.217
Highly differentiated: I	184	11.4	51	9.5
Moderately differentiated: II	564	35.0	171	31.8
Poorly differentiated: III	795	49.3	291	54.2
Undifferentiated; anaplastic: IV	69	4.3	24	4.5
Histological subtype					4.629	0.201
Adenocarcinoma	1087	67.4	382	71.1
Infiltrating duct carcinoma	132	8.2	30	5.6
Neuroendocrine carcinoma	170	10.5	54	10.1
Other	223	13.8	71	13.2
Tumor size					2.819	0.244
≤ 2	95	5.9	22	4.1
2 ~ 4	651	40.4	214	39.9
>4	866	53.7	301	56.1
AJCC T stage					1.254	0.740
T1	57	3.5	14	2.6
T2	543	33.7	184	34.3
T3	633	39.3	208	38.7
T4	379	23.5	131	24.4
AJCC N stage					1.039	0.308
N0	900	55.8	314	58.5
N1	712	44.2	223	41.5
Surgery					0.954	0.329
No	1395	86.5	455	84.7
Yes	217	13.5	82	15.3
Radiotherapy					0.072	0.788
No	1520	94.3	504	93.9
Yes	92	5.7	33	6.1
Chemotherapy					0.519	0.471
No/Unknown	648	40.2	226	42.1
Yes	964	59.8	311	57.9
Brain metastasis					1.825	0.240
No	1607	99.7	533	99.3
Yes	5	0.3	4	0.7
Lung metastasis					0.154	0.695
No	1386	86.0	466	86.8
Yes	226	14.0	71	13.2
Bone metastasis					0.449	0.503
No	1530	94.9	505	94.0
Yes	82	5.1	32	6.0

Table 4

Univariate and multivariate Cox proportional hazards regression of PCLM patients.
Characteristic	Univariate analysis		Multivariate analysis
Characteristic	HR (95%CI)	P	HR (95%CI)	P
Age	1.028 (1.024–1.033)	0.000	1.016(1.011–1.021)	< 0.001
Sex
Female	Reference
Male	1.085 (0.979–1.202)	0.120
Race
White	Reference
Black	1.109(0.949–1.296)	0.193
Others ^a	0.917(0.757–1.111)	0.377
Primary site
Head of pancreas	Reference		Reference
Body of pancreas	1.169(1.010–1.353)	0.036	1.008(0.869–1.169)	0.918
Tail of pancreas	0.773(0.676–0.884)	< 0.001	0.907(0.791–1.040)	0.163
Pancreatic duct	0.998(0.447–2.230)	0.996	0.909(0.405–2.040)	0.817
Other specified parts of pancreas	1.014(0.669–1.537)	0.948	0.906(0.596–1.377)	0.643
Overlapping lesion of pancreas	1.025(0.858–1.226)	0.783	0.933(0.779–1.118)	0.454
Pancreas, NOS	0.803(0.645–1.001)	0.051	0.828 (0.662–1.035)	0.097
Grade
Highly differentiated: I	Reference		Reference
Moderately differentiated: II	2.561(2.096–3.130)	< 0.001	1.954(1.579–2.417)	< 0.001
Poorly differentiated: III	4.384(3.592–5.350)	< 0.001	3.328(2.685–4.125)	< 0.001
Undifferentiated; anaplastic: IV	3.336 (2.459–4.527)	< 0.001	3.225(2.357–4.414)	< 0.001
Histological subtype
Adenocarcinoma	Reference		Reference
Infiltrating duct carcinoma	0.761(0.634–0.913)	0.003	1.163(0.959–1.409)	0.125
Neuroendocrine carcinoma	0.253(0.207–0.309)	< 0.001	0.354(0.286–0.439)	< 0.001
Other	0.432(0.365–0.510)	< 0.001	0.597(0.501–0.711)	< 0.001
Tumor size
≤ 2	Reference
2 ~ 4	1.197(0.953–1.504)	0.122
>4	1.209(0.965–1.513)	0.099
AJCC T stage
T1	Reference		Reference
T2	1.534(1.137–2.069)	0.005	1.311(0.968–1.776)	0.080
T3	1.134(0.842–1.527)	0.408	1.084 (0.802–1.465)	0.599
T4	1.518(1.119–2.058)	0.007	1.233(0.904–1.683)	0.186
AJCC N stage
N0	Reference
N1	0.942(0.850–1.043)	0.251
Surgery
No	Reference		Reference
Yes	0.297(0.250–0.354)	< 0.001	0.376(0.308–0.458)	< 0.001
Radiotherapy
No	Reference		Reference
Yes	0.753(0.604–0.939)	0.012	0.761(0.610–0.951)	0.016
Chemotherapy
No/Unknown	Reference		Reference
Yes	0.706(0.635–0.785)	< 0.001	0.454(0.404–0.509)	< 0.001
Brain metastasis
No	Reference
Yes	1.625(0.675–3.912)	0.279
Lung metastasis
No	Reference		Reference
Yes	1.670(1.446–1.929)	< 0.001	1.261(1.088–1.463)	0.002
Bone metastasis
No	Reference
Yes	1.077(0.858–1.351)	0.523

3.5 Establishment and Validation of Prognostic Nomogram

We constructed a prognostic nomogram based on the above 7 prognostic factors (Fig. 4), and created a user-friendly free online version of the model (https://pclm.shinyapps.io/PCLM_Cox_Nomo/). ROC analysis of the nomogram showed that the AUCs of 6-, 12-and 18-month OS were 0.813, 0.842, and 0.884 in the training cohort (Fig. 5A-C), and 0.828, 0.848 and 0.863 in the internal validation cohort, respectively (Fig. 5D-F), which were significantly higher than that of TNM staging. The C-index was 0.755 (95% CI: 0.741–0.769) for the training cohort and 0.760 (95% CI: 0.736–0.784) for the internal validation cohort, indicating good predictive accuracy. The calibration curves indicated strong agreement between actual observations and predictions (Fig. 6). Moreover, DCA revealed that the nonogram has superior clinical utility than the TNM system (Fig. 7). Likewise, ROC analysis of the prognostic nomogram for the external validation cohort showed that the AUCs of 6-, 12-and 18-month OS were 0.692, 0.732, and 0.720, respectively, demonstrating considerable predictive accuracy (Fig. 8A-C).

PC remains a major threat to human life and health worldwide. Most PC patients (about 60%) have metastases at the time of diagnosis(14), with the liver being the most common site of metastasis. Timely diagnosis of liver metastasis in PC patients is important for prognosis. However, traditional imaging tests, such as ultrasound, magnetic resonance imaging (MRI), or computed tomography (CT), have low diagnostic sensitivity and specificity for liver metastasis, and there are potential biases and limitations in different imaging techniques. For example, ultrasound detection of small liver metastasis is often operator-dependent and may result in false-negative results due to the presence of intestinal gas or obesity. CT may have limitations in detecting small liver metastasis, especially in patients with fatty liver or cirrhosis, where there may be diffuse changes in the liver parenchyma that make metastasis difficult to distinguish. Although MRI may improve the detection of small liver metastasis, this technique is more time-consuming and expensive than CT and is not readily available in all clinical settings. Therefore, it is important to develop a model that accurately predicts the probability of LM in PC patients. Despite the great progress in understanding the molecular mechanisms of PCLM(15, 16), these findings are often not applicable to the objective and accurate assessment of prognosis, and hence a simple and easy-to-use model that can accurately predict the prognosis of PCLM patients is urgently needed. In this study, we developed a diagnostic nomogram and a prognostic nomogram for predicting the risk and survival of PCLM based on the patient cohort from SEER database, respectively, and externally validated these models using our own patient cohort.

Similar to previous findings, we found that age, primary site, degree of differentiation, histological subtype, N stage, surgery, radiotherapy, lung metastasis, and bone metastasis are risk factors for PCLM(17, 18). Furthermore, the prognostic factors for PCLM, including age, degree of differentiation, histological subtype, surgery, radiotherapy, chemotherapy, and lung metastases, were largely consistent with those for liver metastases of gastric cancer(19) and rectal cancer(20), bone metastases of pancreatic cancer(21), and lung metastases of PC(22). It was reported that the median survival of patients with untreated distant metastases is no more than 6 months(23) and the median survival of PCLM patients in this study was 5 months for both the SEER cohort and our cohort. Therefore, clinicians should keep a close eye on these risk factors and prognostic factors in PCLM patients.

We found that younger patients are more likely to develop LM with greater malignancy or poorer histological features compared with older patients(24, 25),which is also consistent with findings in other types of tumors(28, 29). Chang et al.(27) found that early-onset colorectal adenocarcinoma was more likely to have positive circumferential margins, venous invasion, and nerve invasion, which may be attributed to the greater susceptibility of tumor cells to DNA damage in younger patients (26). Therefore, younger patients are more likely to develop LM. However, elderly PCLM patients have a lower survival rate, which is similar to patients with distant metastases of gastrointestinal tract tumors(30, 31). This decreased survival may be due to the high incidence of comorbidities and poor tolerance to radiotherapy and chemotherapy in elderly patients. We found that primary tumors in the body and tail of the pancreas are more likely to result in liver metastasis than tumors in the head of the pancreas, possibly due to the presence of more organs and blood vessels and nerves near the body and tail of the pancreas, such as the spleen, stomach, transverse colonic mesentery, abdominal aorta and superior mesenteric vein, which facilitate easier access to the liver through circulation. When tumors in the tail of the pancreas invade the surrounding tissues and organs, the clinical symptoms are often not obvious and may become apparent later. However, when the symptoms are actually obvious, the tumor may have already invaded the surrounding lymphatic, vascular and nerve tissues and developed distant metastasis(32).

Our nomogram shows that the risk of LM increases when the tumor is less differentiated. The rates of LM for highly differentiated, moderately differentiated, poorly differentiated and undifferentiated PC were 10.0%, 15.5%, 25.9% and 37.1%, respectively. Tumor cells of poorly differentiated cancer have been shown to be highly invasive and motile, and can often invade the basement membrane of blood vessels or lymphatic vessels, reflecting the inherent characteristics of tumor biological behavior(33). The degree of differentiation is also an independent risk factor for PCLM patients. Our data showed that median survival significantly varied among patients with highly differentiated (30.00 months), moderately differentiated (6.00 months), poorly differentiated (3.00 months), and undifferentiated (4.00 months) PC (P < 0.05). For histological subtypes, neuroendocrine carcinoma had the highest probability of LM, adenocarcinoma was the most common histologic subtype, and adenocarcinoma and infiltrating duct carcinoma had the worse prognosis. Median survival was significantly different among PCLM patients with adenocarcinoma (3.00 months), infiltrating duct carcinoma (6.00 months), and neuroendocrine carcinoma (28.50 months) (P < 0.05), consistent with previous findings(34). Lymph node metastasis is common prior to distant metastasis(35, 36), suggesting that distant metastasis should be closely monitored in PC patients with lymph node metastasis.

Our study showed that 67.9% of patients with bone metastases had LM and 59.4% of patients with lung metastases had LM, suggesting that multiple distant metastases may occur at the same time, which is consistent with other studies(37, 38).When distant metastases from other sites are included in the prediction model, bone metastasis and lung metastasis are risk factors for PCLM. Though, only lung metastasis affected the prognosis of PCLM patients as the median survival of PCLM patients decreased from 5.00 months without lung metastasis to 2.00 months with lung metastasis. Therefore, clinicians should pay close attention to the development of lung metastasis in PCLM patients. Among the treatment modalities, both surgical treatment and radiotherapy resulted in a lower incidence of LM, suggesting that both modalities are highly efficacious in preventing LM in PC patients (39).

Systemic chemotherapy is currently the preferred modality for PC patients with metastases. Although the FOLFIRINOX and NG regimens have prolonged the survival of these patients, the median survival time is always less than 12 months (40, 41). Conroy T et al. (41) showed that the FOLFIRINOX regimen prolonged the median survival of patients by more than 4 months (from 6.8 to 11.1 months) but had a poorer safety profile compared with gemcitabine. Meanwhile, nab-paclitaxel combined with gemcitabine has been reported to improve median OS compared with gemcitabine alone (8.5 months vs. 6.7 months)(42). The median survival was 7.00 months for PC patients treated with chemotherapy and 1.00 month for those not treated with chemotherapy. Furthermore, it was found that the objective remission rate of first-line chemotherapy is only 50%(43), which indicates that chemotherapy is somewhat effective in improving advanced PC. Pancreatic ductal adenocarcinoma (PDAC) is primarily driven by genetic alterations, particularly K-Ras mutations, and the tumor microenvironment also plays a critical role in facilitating treatment resistance and immune evasion of malignant pancreatic cells(44). The involvement of the immune system in PC, including the application of immune checkpoint inhibitors (ICIs), is an area of active research. Several studies(45, 46) have shown that ICIs may be more effective when combined with chemotherapy or radiotherapy and targeted treatment in advanced PC. In addition, targeting the associated signaling pathways and immune checkpoint molecules, alone or in combination with conventional therapies, has yielded promising results in the treatment of PC(DOI: 10.3390/biomedicines9040373). Taken together, these findings suggest that a more comprehensive treatment approach should be taken for PCLM, which takes into account the molecular mechanisms of PDAC development and progression, as well as the involvement of the immune system in cancer progression and immune evasion. It is only by understanding the molecular and cellular mechanisms driving PDAC and liver metastasis can clinicians be better equipped to provide personalized treatments for patients with the aggressive form of this cancer.

In recent years, advances in drug therapy have fostered the emergence of new treatment approaches, such as the possibility of post-treatment surgery for smaller and fewer liver metastases (47–49). In 2019, Japanese scholars reported a case of PCLM in which both pancreatic lesions and liver metastases were no longer detectable by imaging after 6 months of FOLFIRNOX chemotherapy. The patient subsequently underwent pancreaticoduodenectomy with postoperative pathology revealing a 2-mm reduction of the pancreatic lesions from the initial 15 mm (50). Hackert et al.(51) reported the outcome of 128 PC patients with limited oligometastases (including 85 with liver metastases and 43 with para-aortic lymph node metastases) who underwent resection of the primary tumor and metastases. The authors found that the incidences of surgical complications and 30-d mortality rates for simultaneous metastasectomy were 45% and 2.9%, respectively, and the median OS after metastasectomy was 12.3 months. The 5-year postoperative survival rates were 8.1% in the LM group and 10.1% in the para-aortic lymph node resection group, both of which were better than the historical data of patients without surgery within the same period. In the present study, median survival was significantly different between PCLM patients with surgery (25.00 months) and without surgery (4.00 months) (P < 0.05). However, given the presence of heterogeneity and selection and publication bias in these studies, as well as the subjectivity in surgical indications, the risks and benefits of the proposed surgical approach should be carefully assessed and surgery should only be performed based on each patient’s conditions. Radiotherapy is a palliative treatment that helps to relieve patients' suffering and improve their quality of life. The median survival of PCLM patients with and without radiotherapy in this study was 6.00 months and 4.00 months, respectively. Altogether, these data indicate that a multidisciplinary approach is needed for the treatment of PCLM to enhance patient prognosis.

In this study, we identified the risk and prognostic factors for PCLM and established a considerably complete assessment system. The diagnostic nomogram has an AUC of 0.877 for the internal validation cohort, an AUC of 0.846 for the external validation cohort, which ascertained that the diagnostic nomogram could accurately identify PC patients at high risk of LM. In both the training and internal validation cohorts, the AUC and C-index of the prognostic nomogram was above 0.8 and 0.75, respectively. The calibration curves of the two cohorts were in good agreement, and the DCA curves showed that the model is of good clinical utility and can better predict the prognosis of PCLM patients than the TNM system, indicating that both of our nomograms are valuable tools for assessing the risk and outcome of PCLM.

Several limitations should be considered for this study. First, this was a retrospective study of the SEER database and may introduce selection bias and confounding factors into the results. The factors included in model construction are influenced by the data records, which may lack some key clinical data and hematological indicators. Second, 75% of patients in the SEER database were Caucasians. Although univariate and multivariate analyses revealed that race was not a significant influencing factor for PCLM, it is not certain that this prediction model is applicable to all races. Last, the external validation results for the prognostic nomogram were unsatisfactory. Since the external validation cohort came from a single center and the sample size was small (64 cases), further validation of the nomogram with multi-center large cohort data is warranted. Despite the limitations of this retrospective study, the nomograms constructed in this study are valid and instructive tools that serve as a simple, quick and noninvasive approach for clinical workup.

In summary, we identified the risk and prognostic factors for PCLM and established a diagnostic and a prognostic nomogram. These two nomograms are useful for clinical selection of personalized treatment plans and provide relatively accurate individualized assessment of patient prognosis.

DATA AVAILABILITY STATEMENT

Publicly available datasets were analyzed in this study. This data can be found here: https://seer.cancer.gov/.

ETHICS STATEMENT

The studies involving human participants were reviewed and approved by the ethics committees of Afliated Hospital of Xuzhou Medical University (permission number. XYFY2022-KL328-01).

AUTHOR CONTRIBUTIONS

ZZ, YS and SF contributed to conception and design of the study. ZZ organized the database. ZB, KW performed the statistical analysis. ZZ, HF wrote the first draft of the manuscript. SF, YZ, LW, and RL wrote sections of the manuscript. All authors revised and reviewed the manuscript, and approved the submitted version.

FUNDING

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin (2021) 71(3):209-249. doi:10.3322/caac.21660
Chen W, Zheng R, Baade PD, Zhang S, Zeng H, Bray F, et al. Cancer statistics in China, 2015. CA Cancer J Clin (2016) 66(2):115-32. doi:10.3322/caac.21338
Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin (2018) 68(1):7-30. doi:10.3322/caac.21442
Allendorf JD, Lauerman M, Bill A, DiGiorgi M, Goetz N, Vakiani E, et al. Neoadjuvant chemotherapy and radiation for patients with locally unresectable pancreatic adenocarcinoma: Feasibility, efficacy, and survival. J Gastrointest Surg (2008) 12(1):91-100. doi:10.1007/s11605-007-0296-7
Ren Y, Wang S, Wu B, Wang Z. Clinicopathological Features, Prognostic Factors and Survival in Patients With Pancreatic Cancer Bone Metastasis. Front Oncol (2022) 12:759403. doi:10.3389/fonc.2022.759403
Gumberger P, Bjornsson B, Sandström P, Bojmar L, Zambirinis CP. The Liver Pre-Metastatic Niche in Pancreatic Cancer: A Potential Opportunity for Intervention. Cancers (Basel) (2022) 14(12):3028. doi:10.3390/cancers14123028
Murakami T, Hiroshima Y, Matsuyama R, Homma Y, Hoffman RM, Endo I. Role of the tumor microenvironment in pancreatic cancer. Ann Gastroenterol Surg (2019) 3(2):130-137. doi:10.1002/ags3.12225
Wang X, Hu LP, Qin WT, Yang Q, Chen DY, Li Q, et al. Identification of a subset of immunosuppressive P2RX1-negative neutrophils in pancreatic cancer liver metastasis. Nat Commun (2021) 12(1):174. doi:10.1038/s41467-020-20447-y
Costa-Silva B, Aiello NM, Ocean AJ, Singh S, Zhang H, Thakur BK, et al. Pancreatic cancer exosomes initiate pre-metastatic niche formation in the liver. Nat Cell Biol (2015) 17(6):816-26. doi:10.1038/ncb3169
Curtis M, Kenny HA, Ashcroft B, Mukherjee A, Johnson A, Zhang Y, et al. Fibroblasts Mobilize Tumor Cell Glycogen to Promote Proliferation and Metastasis. Cell Metab (2019) 29(1):141-155. doi:10.1016/j.cmet.2018.08.007
Klein AP. Pancreatic cancer: a growing burden. Lancet Gastroenterol Hepatol (2019) 4(12):895-896. doi:10.1016/s2468-1253(19)30323-1
Oweira H, Petrausch U, Helbling D, Schmidt J, Mannhart M, Mehrabi A, et al. Prognostic value of site-specific metastases in pancreatic adenocarcinoma: A Surveillance Epidemiology and End Results database analysis. World J Gastroenterol (2017) 23(10):1872-1880. doi:10.3748/wjg.v23.i10.1872
Qiu L, Song P, Chen P, Wang H, Li F, Shu M, et al. Clinical Characteristics and Prognosis of Patients With Pulmonary Mucoepidermoid Carcinoma: A SEER-Based Analysis. Front Oncol (2021) 11:601185. doi:10.3389/fonc.2021.601185
Vincent A, Herman J, Schulick R, Hruban RH, Goggins M. Pancreatic cancer. Lancet (2011) 378(9791):607-20. doi:10.1016/s0140-6736(10)62307-0
Nielsen SR, Quaranta V, Linford A, Emeagi P, Rainer C, Santos A, et al. Macrophage-secreted granulin supports pancreatic cancer metastasis by inducing liver fibrosis. Nat Cell Biol (2016) 18(5):549-60. doi:10.1038/ncb3340
Young ED, Manley SJ, Beadnell TC, Shearin AE, Sasaki K, Zimmerman R, et al. Suppression of pancreatic cancer liver metastasis by secretion-deficient ITIH5. Br J Cancer (2021) 124(1):166-175. doi:10.1038/s41416-020-01093-z
Li Q, Bai L, Xing J, Liu X, Liu D, Hu X. Risk Assessment of Liver Metastasis in Pancreatic Cancer Patients Using Multiple Models Based on Machine Learning: A Large Population-Based Study. Dis Markers (2022) 2022:1586074. doi:10.1155/2022/1586074
He C, Zhong L, Zhang Y, Cai Z, Lin X. Development and validation of a nomogram to predict liver metastasis in patients with pancreatic ductal adenocarcinoma: a large cohort study. Cancer Manag Res (2019) 11:3981-3991. doi:10.2147/cmar.S200684
Dong Z, Zhang Y, Geng H, Ni B, Xia X, Zhu C, et al. Development and validation of two nomograms for predicting overall survival and cancer-specific survival in gastric cancer patients with liver metastases: A retrospective cohort study from SEER database. Transl Oncol (2022) 24:101480. doi:10.1016/j.tranon.2022.101480
Liu C, Hu C, Huang J, Xiang K, Li Z, Qu J, et al. A Prognostic Nomogram of Colon Cancer With Liver Metastasis: A Study of the US SEER Database and a Chinese Cohort. Front Oncol (2021) 11:591009. doi:10.3389/fonc.2021.591009
Zhang W, Ji L, Wang X, Zhu S, Luo J, Zhang Y, et al. Nomogram Predicts Risk and Prognostic Factors for Bone Metastasis of Pancreatic Cancer: A Population-Based Analysis. Front Endocrinol (Lausanne) (2021) 12:752176. doi:10.3389/fendo.2021.752176
Zhang W, Ji L, Zhong X, Zhu S, Zhang Y, Ge M, et al. Two Novel Nomograms Predicting the Risk and Prognosis of Pancreatic Cancer Patients With Lung Metastases: A Population-Based Study. Front Public Health (2022) 10:884349. doi:10.3389/fpubh.2022.884349
Deeb A, Haque SU, Olowokure O. Pulmonary metastases in pancreatic cancer, is there a survival influence? J Gastrointest Oncol (2015) 6(3):E48-51. doi:10.3978/j.issn.2078-6891.2014.114
Wang J, Liu J, Pan H, Jiang C, Liu S, Zhu Z, et al. Young age increases the risk of lymph node positivity in papillary thyroid cancer patients: a SEER data-based study. Cancer Manag Res (2018) 10:3867-3873. doi:10.2147/cmar.S167774
Mete O, Asa SL. Pathological definition and clinical significance of vascular invasion in thyroid carcinomas of follicular epithelial derivation. Mod Pathol (2011) 24(12):1545-1552. doi:10.1038/modpathol.2011.119
Romei C, Ciampi R, Elisei R. A comprehensive overview of the role of the RET proto-oncogene in thyroid carcinoma. Nat Rev Endocrinol (2016) 12(4):192-202. doi:10.1038/nrendo.2016.11
Chang DT, Pai RK, Rybicki LA, Dimaio MA, Limaye M, Jayachandran P, et al. Clinicopathologic and molecular features of sporadic early-onset colorectal adenocarcinoma: an adenocarcinoma with frequent signet ring cell differentiation, rectal and sigmoid involvement, and adverse morphologic features. Mod Pathol (2012) 25(8):1128-39. doi:10.1038/modpathol.2012.61
Lin Z, Yan S, Zhang J, Pan Q. A Nomogram for Distinction and Potential Prediction of Liver Metastasis in Breast Cancer Patients. J Cancer (2018) 9(12):2098-2106. doi:10.7150/jca.24445
Shiroyama T, Suzuki H, Tamiya M, Tamiya A, Tanaka A, Okamoto N, et al. Clinical Characteristics of Liver Metastasis in Nivolumab-treated Patients with Non-small Cell Lung Cancer. Anticancer Res (2018) 38(8):4723-4729. doi:10.21873/anticanres.12779
Tang M, Wang H, Cao Y, Zeng Z, Shan X, Wang L. Nomogram for predicting occurrence and prognosis of liver metastasis in colorectal cancer: a population-based study. Int J Colorectal Dis (2021) 36(2):271-282. doi:10.1007/s00384-020-03722-8
Lin Z, Wang R, Zhou Y, Wang Q, Yang C-Y, Hao B-C, et al. Prediction of distant metastasis and survival prediction of gastric cancer patients with metastasis to the liver, lung, bone, and brain: research based on the SEER database. Ann Transl Med (2022) 10(1):16. doi:10.21037/atm-21-6295
Paye F, Micelli Lupinacci R, Bachellier P, Boher JM, Delpero JR. Distal pancreatectomy for pancreatic carcinoma in the era of multimodal treatment. Br J Surg (2015) 102(3):229-36. doi:10.1002/bjs.9708
Li HB, Zhou J, Zhao FQ. A Prognostic Nomogram for Disease-Specific Survival in Patients with Pancreatic Ductal Adenocarcinoma of the Head of the Pancreas Following Pancreaticoduodenectomy. Med Sci Monit (2018) 24:6313-6321. doi:10.12659/msm.909649.
Pokrzywa CJ, Abbott DE, Matkowskyj KA, Ronnekleiv-Kelly SM, Winslow ER, Weber SM, et al. Natural History and Treatment Trends in Pancreatic Cancer Subtypes. J Gastrointest Surg (2019) 23(4):768-778. doi:10.1007/s11605-019-04113-3
Wu SG, Zhang WW, Sun JY, Li FY, Lin Q, He ZY. Patterns of Distant Metastasis Between Histological Types in Esophageal Cancer. Front Oncol (2018) 8:302. doi:10.3389/fonc.2018.00302
Gaitanidis A, Alevizakos M, Tsaroucha A, Tsalikidis C, Pitiakoudis M. Predictive Nomograms for Synchronous Distant Metastasis in Rectal Cancer. J Gastrointest Surg (2018) 22(7):1268-1276. doi:10.1007/s11605-018-3767-0
Oweira H, Petrausch U, Helbling D, Schmidt J, Mannhart M, Mehrabi A, et al. Prognostic value of site-specific metastases in pancreatic adenocarcinoma: A Surveillance Epidemiology and End Results database analysis. World J Gastroenterol (2017) 23(10):1872-1880. doi:10.3748/wjg.v23.i10.1872
Shiroyama T, Suzuki H, Tamiya M, Tamiya A, Tanaka A, Okamoto N, et al. Clinical Characteristics of Liver Metastasis in Nivolumab-treated Patients with Non-small Cell Lung Cancer. Anticancer Res (2018) 38(8):4723-4729. doi:10.21873/anticanres.12779
Labori KJ, Katz MH, Tzeng CW, Bjørnbeth BA, Cvancarova M, Edwin B, et al. Impact of early disease progression and surgical complications on adjuvant chemotherapy completion rates and survival in patients undergoing the surgery first approach for resectable pancreatic ductal adenocarcinoma - A population-based cohort study. Acta Oncol (2016) 55(3):265-77. doi:10.3109/0284186x.2015.1068445
Von Hoff DD, Ervin T, Arena FP, Chiorean EG, Infante J, Moore M, et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med (2013) 369(18):1691-703. doi:10.1056/NEJMoa1304369
Conroy T, Desseigne F, Ychou M, Bouché O, Guimbaud R, Bécouarn Y, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med (2011) 364(19):1817-25. doi:10.1056/NEJMoa1011923
Tabernero J, Kunzmann V, Scheithauer W, Reni M, Shiansong Li J, Ferrara S, et al. nab-Paclitaxel plus gemcitabine for metastatic pancreatic cancer: a subgroup analysis of the Western European cohort of the MPACT trial. Onco Targets Ther (2017) 10:591-596. doi:10.2147/ott.S124097
Ramanathan RK, McDonough SL, Philip PA, Hingorani SR, Lacy J, Kortmansky JS, et al. Phase IB/II Randomized Study of FOLFIRINOX Plus Pegylated Recombinant Human Hyaluronidase Versus FOLFIRINOX Alone in Patients With Metastatic Pancreatic Adenocarcinoma: SWOG S1313. J Clin Oncol (2019) 37(13):1062-1069. doi:10.1200/jco.18.01295
Javadrashid D, Baghbanzadeh A, Derakhshani A, Leone P, Silvestris N, Racanelli V, et al. Pancreatic Cancer Signaling Pathways, Genetic Alterations, and Tumor Microenvironment: The Barriers Affecting the Method of Treatment. Biomedicines (2021) 9(4):373. doi: 10.3390/biomedicines9040373
Weiss GJ, Blaydorn L, Beck J, Bornemann-Kolatzki K, Urnovitz H, Schütz E, et al. Phase Ib/II study of gemcitabine, nab-paclitaxel, and pembrolizumab in metastatic pancreatic adenocarcinoma. Invest New Drugs (2018) 36(1):96-102. doi:10.1007/s10637-017-0525-1
Liu L, Huang X, Shi F, Song J, Guo C, Yang J, et al. Combination therapy for pancreatic cancer: anti-PD-(L)1-based strategy. J Exp Clin Cancer Res (2022) 41(1):56. doi:10.1186/s13046-022-02273-w
Pausch TM, Liu X, Cui J, Wei J, Miao Y, Heger U, et al. Survival Benefit of Resection Surgery for Pancreatic Ductal Adenocarcinoma with Liver Metastases: A Propensity Score-Matched SEER Database Analysis. Cancers (2022) 14(1):57. doi:10.3390/cancers14010057
McKenzie S, Mailey B, Artinyan A, Kim J, Ellenhorn JDI. The incidence and outcomes of pancreatectomy in patients with metastatic pancreatic adenocarcinoma. JOP (2010) 11(4):341-7
Shi S, Yu X-J. Time to think: Selecting patients who may benefit from synchronous resection of primary pancreatic cancer and liver metastases. World J Gastroenterol (2018) 24(33):3677-3680. doi:10.3748/wjg.v24.i33.3677
Shimmura H, Kuramochi H, Jibiki N, Katagiri S, Nishino T, Araida T. Dramatic response of FOLFIRINOX regimen in a collision pancreatic adenocarcinoma patient with a germline BRCA2 mutation: a case report. Jpn J Clin Oncol (2019) 49(11):1049-1054. doi:10.1093/jjco/hyz141. Cited in: Pubmed
Hackert T, Niesen W, Hinz U, Tjaden C, Strobel O, Ulrich A, et al. Radical surgery of oligometastatic pancreatic cancer. Eur J Surg Oncol (2017) 43(2):358-363. doi:10.1016/j.ejso.2016.10.023

No competing interests reported.

Construction and Validation of Risk Factors and Prognostic Model for Liver Metastasis of Pancreatic Cancer

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1. INTRODUCTION

2. MATERIALS AND METHODS

2.1 Patient Selection

2.2 Data Elements

2.3 Statistical Analysis

3. RESULTS

3.1 Patient Characteristics

3.2 Independent Risk Factors for PCLM

3.3 Establishment and Validation of Diagnostic Nomogram

3.4 Independent Prognostic Factors for PCLM

3.5 Establishment and Validation of Prognostic Nomogram

4. DISCUSSION

5. CONCLUSION

Declarations

References

Additional Declarations

Status:

Version 1