search for


Prognostication for recurrence patterns after curative resection for pancreatic ductal adenocarcinoma
Ann Hepatobiliary Pancreat Surg 2024 May;28(2):248-61
Published online May 31, 2024;
Copyright © 2024 The Korean Association of Hepato-Biliary-Pancreatic Surgery.

Andrew Ang1,2, Athena Michaelides1,2, Claude Chelala2, Dayem Ullah2, Hemant M. Kocher1,2; Barts Pancreas Tissue Bank

1The Royal London Hospital, Barts Health NHS Trust, London, UK,
2Barts Cancer Institute, Queen Mary University of London, John Vane Science Centre, Charterhouse Square, London, UK
Correspondence to: Hemant M. Kocher, MS, MD, FRCS
Barts Cancer Institute, Queen Mary University of London, John Vane Science Centre, Charterhouse Square, London EC1M 6BQ, UK
Tel: +44-0-2078823573, Fax: +44-0-2078823884, E-mail:
Received November 15, 2023; Revised February 2, 2024; Accepted February 10, 2024.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Backgrounds/Aims: This study aimed to investigate patterns and factors affecting recurrence after curative resection for pancreatic ductal adenocarcinoma (PDAC).
Methods: Consecutive patients who underwent curative resection for PDAC (2011-21) and consented to data and tissue collection (Barts Pancreas Tissue Bank) were followed up until May 2023. Clinico-pathological variables were analysed using Cox proportional hazards model.
Results: Of 91 people (42 males [46%]; median age, 71 years [range, 43–86 years]) with a median follow-up of 51 months (95% confidence intervals [CIs], 40–61 months), the recurrence rate was 72.5% (n = 66; 12 loco-regional alone, 11 liver alone, 5 lung alone, 3 peritoneal alone, 29 simultaneous loco-regional and distant metastases, and 6 multi-focal distant metastases at first recurrence diagnosis). The median time to recurrence was 8.5 months (95% CI, 6.6–10.5 months). Median survival after recurrence was 5.8 months (95% CI, 4.2–7.3 months). Stratification by recurrence location revealed significant differences in time to recurrence between loco-regional only recurrence (median, 13.6 months; 95% CI, 11.7–15.5 months) and simultaneous loco-regional with distant recurrence (median, 7.5 months; 95% CI, 4.6–10.4 months; p = 0.02, pairwise log-rank test). Significant predictors for recurrence were systemic inflammation index (SII) ≥ 500 (hazard ratio [HR], 4.5; 95% CI, 1.4–14.3), lymph node ratio ≥ 0.33 (HR, 2.8; 95% CI, 1.4–5.8), and adjuvant chemotherapy (HR, 0.4; 95% CI, 0.2–0.7).
Conclusions: Timing to loco-regional only recurrence was significantly longer than simultaneous loco-regional with distant recurrence. Significant predictors for recurrence were SII, lymph node ration, and adjuvant chemotherapy.
Keywords : Overall survival; Disease free survival; Multivariate analysis; Lymph node ratio; Systemic inflammation index

Pancreatic cancer is the fourth most common cause of cancer related deaths in the UK [1]. Unlike other types of cancer, it has not seen an overall decline in deaths over the last two decades in the USA or Europe [1]. Pancreatic cancer has a poor mortality to incidence ratio of 98% and a poor 5-year overall survival (OS) rate of 6% [2,3]. Only surgery offers a potential cure. However, most cases are already metastatic at the time of diagnosis. Only 15%–20% of cases have a potentially resectable disease [4]. A proportion of these patients will then undergo curative resection and adjuvant therapy, which has been shown to increase 5-year OS rates to 28%–43% in the best outcome scenario [5,6]. Despite advances in modern medicine, a recent systematic review has shown that even after surgery, 47%–92% of patients will have pancreatic ductal adenocarcinoma (PDAC) recurrence [7].

Recurrence occurs mainly within the first two years after resection [5,8]. However, evidence suggests that recurrence at different sites behaves differently. Studies have reported that liver recurrence occurs earlier than recurrence at other sites while lung recurrence occurs later [9,10]. This also impacts survival after diagnosis of recurrence where lung recurrence has a better prognosis than recurrence at other sites [11]. Despite multiple studies looking into PDAC recurrence, there has been no consensus on predictive factors of recurrence. To date, there are few studies reporting PDAC recurrence patterns and predictive factors in the UK. There are no national guidelines defining surveillance after resection for PDAC. Defining recurrence patterns and predictors could provide insight into how surveillance could be conducted. Thus, the aim of this study was to investigate patterns and factors affecting first site recurrence after curative resection of PDAC.


Patient selection and characteristics

All consented patients from a prospectively maintained database Barts Pancreas Tissue Bank (Hampshire B Research Ethics Committee supported in January 2014, ref: 13/SC/0592; renewed 2019, ref: 18/SC/0630) were screened for eligibility (curative resection of PDAC) for this study. Patients with distant metastases diagnosed at the time of surgery were excluded. Patients with neoadjuvant chemotherapy were included. Only eight patients had venous resection. None had an arterial resection. The TNM 7th or 8th edition staging system was used. Resection margin status (R) was defined as R0 when the resection margin was ≥ 1 mm away from the tumour. It was defined as R1 when the resection margin was < 1 mm away from the tumour. Patients were excluded if they had an R2 margin status. The cut-off surgical resection date for inclusion into this study was November 1st, 2021. Follow-up was censored at May 1st, 2023. Of a total of 799 patients screened, 91 patients were included in this study (Fig. 1).

Fig 1. Flow chart of patient enrolment in this study. BPTB, Barts Pancreas Tissue Bank; PDAC, pancreatic ductal adenocarcinoma.

Data collection and follow-up

Data missing from Barts Pancreas Tissue Bank database were gathered from Electronic Health Records of Barts Health NHS Trust. Preoperative blood test results were collected within two months prior to surgery—at the time of referral, preoperative assessment or on the day of surgery, whichever was the latest. Postoperative CA19-9 levels were taken within three months of the date of recurrence. Symptom information was collected preoperatively. If a patient underwent a PET scan within two months prior to surgery, SUVmax data were obtained. Posterior resection margin was defined as posterior surface (retroperitoneum). Anterior margin was defined as anterior surface of the pancreas. Loco-regional recurrence was defined as involvement of the pancreatic bed, remnant pancreas, superior mesenteric vessels, regional lymph nodes, and anastomotic site.

Standard follow-up regimen after curative surgical resection for PDAC was a computed tomography (CT) scan, CA19-9 blood test, and a clinic appointment at 6-monthly intervals for a duration of five years. Patients who were referred to oncology postoperatively had regular follow-up appointments and CT scans as directed by the oncologist, occasionally from the hospital that referred them. CT scans were the main modality of choice for detecting recurrence. Ratification occurred at multi-disciplinary team (MDT) meeting without histological confirmation of recurrence. Only the first site(s) of recurrence were recorded and categorised into loco-regional alone, liver alone, lung alone, peritoneal alone, simultaneous loco-regional with distant metastases, and multiple distant metastases. These patients were censored on the last known date of follow-up during data analysis. There was one patient with PDAC within intraductal papillary mucinous neoplasm who developed a second primary pancreatic cancer in the tail of the pancreas five years following a pylorus-preserving pancreaticoduodenectomy. This patient was censored at the date of diagnosis of the second primary during data analysis.

Statistical analysis

Survival and recurrence analysis were undertaken by first converting continuous variables into categorical variables. This was achieved using the X-tile software (Rimm Lab, Yale School of Medicine) to identify optimal cut-off values after Monte Carlo correction [12]. Cut-off values were then rounded off to the nearest round number if applicable. All other statistical analyses were undertaken using SPSS version 28 (IBM Corp.). Univariate analysis was undertaken using the Cox proportional hazards regression model. Kaplan–Meier curves were generated for each variable. A cut-off p-value of 0.05 was used to determine statistically significant covariates. Covariates with significant numbers of missing values (> 10%) were not included in multivariate analysis. Age was also excluded from multivariate analysis as it was not a biological factor of PDAC. Remaining covariates were then subject to a test for multi-collinearity and excluded if predictor variables were highly correlated with a variance inflation factor ≥ 5 and/or a tolerance value < 0.2 [13]. Multivariate survival analysis was then performed for remaining covariates with a cut-off p-value of 0.05 for significance.


Patient demographic

A total of 91 patients were identified in the Barts Pancreas Tissue Bank database who underwent curative surgical resection for PDAC between March 2011 and November 2021 (Fig. 1). The study population consisted of 42 males (46%) with a median age of 71 years (range, 43–86 years) at the time of surgery. The median follow-up time was 51 months (95% confidence interval [CI], 40–61 months). There were 61 patients (67%) with PDAC in the head of the pancreas, 16 (18%) with PDAC in the tail, 4 (4%) with PDAC in the body, 3 (3%) with PDAC in the uncinated process, and 7 (8%) with multifocal PDAC. Nine patients (10%) received neo-adjuvant chemotherapy, of which only one had gemcitabine + capecitabine while the rest had folfirinox. Fifty-six patients (62%) received adjuvant chemotherapy, of which 17 (19%) had folfirinox and 38 (42%) had gemcitabine +/– capecitabine while 1 (1%) had insufficient data.

Recurrence pattern analysis

The recurrence rate in this study was 72.5% (n = 66; loco- regional alone [n = 12], liver alone [n = 11], lung alone [n = 5], peritoneal alone [n = 3], simultaneous loco-regional and distant metastasis [n = 29], and multiple distant metastatic recurrence [n = 6]) (Table 1, 2). No patient had solitary distant metastasis. One-year and 2-year OS rates were 63% and 38%, respectively, in the recurrence group and 96% and 85%, respectively, in the non-recurrence group (Fig. 2). The 5-year OS rate for the non-recurrence group was 47% (deaths due to other causes). The median time to recurrence was 8.5 months (95% CI, 6.6–10.5 months) and median survival after recurrence was 5.3 months (95% CI, 4.0–6.5 months). Stratification by recurrence location revealed significant differences in disease-free survival between loco-regional only recurrence (median, 13.6 months; 95% CI, 11.7–15.5 months) and simultaneous loco-regional with distant recurrence (median, 7.5 months; 95% CI, 4.6–10.4 months; p = 0.02, pairwise log-rank test) (Fig. 3, 4, Table 3).

Table 1 . Patient characteristics, variables, survival analysis, and recurrence analysis

VariablenOverall survivalDisease-free survival
Univariate analysisMultivariate analysisUnivariate analysisMultivariate analysis
p-valueHR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)
Age (yr, at surgery)0.004*0.112
< 787511
≥ 78162.5 (1.3–4.6)1.6 (0.9–3.0)
Female490.9 (0.5–1.4)1 (0.6–1.6)
Afro-Caribbean71.1 (0.4–3.1)1.4 (0.6–3.5)
Asian61 (0.4–2.5)1.2 (0.5–3.0)
BMI (kg/m2)0.4090.162
< 221811
≥ 22720.8 (0.4–1.4)0.7 (0.4–1.2)
PPPD450.6 (0.3–1.1)0.7 (0.4–1.3)
DP220.4 (0.2–0.9)0.5 (0.3–1.1)
Others50.4 (0.1–1.3)0.3 (0.1–1.3)
Vessel resection0.4530.573
Yes81.4 (0.6–3.6)1.3 (0.5–3.0)
T stage0.40.056
1100.8 (0.2–4.4)0.4 (0.1–2.7)
2401.6 (0.4–6.9)2.1 (0.5–8.9)
3381.8 (0.4–7.6)2 (0.5–8.3)
N stage< 0.001*0.002*< 0.001*0.08
1463.6 (1.8–7.3)0.5 (0.2–1.5)3.3 (1.7–6.4)3.1 (1.1–8.6)
2197.7 (3.3–17.9)2.1 (0.6–7.5)4.9 (2.2–10.6)2.3 (0.7–7.1)
Tumour differentiationa)0.1190.371
Poor411.5 (0.9–2.5)1.3 (0.8–2.0)
Tumour resection status0.009*0.8680.004*0.308
R1542.1 (1.2–3.6)1 (0.5–1.8)2.2 (1.3–3.7)1.4 (0.7–2.9)
Number of R1 margins0.02*0.004*
0 (R0)3511
1161.6 (0.8–3.6)1.5 (0.7–3.2)
2232.3 (1.2–4.4)2.6 (1.4–4.9)
≥ 3132.7 (1.3–5.7)3.2 (1.5–6.7)
R1 locationb)
Anterior160.1251.7 (0.9–3.2)0.061.8 (1.0–3.4)
Posterior230.9551 (0.5–1.9)0.4840.8 (0.4–1.5)
SMV250.3261.4 (0.7–2.5)0.1391.6 (0.9–2.9)
SMA60.4731.4 (0.6–3.6)0.0922.3 (0.9–6.1)
Uncinate280.041*2 (1.0–3.9)0.015*2.3 (1.2–4.3)
Invasionc)< 0.001*0.361< 0.001*0.188
Both663.4 (1.6–7.1)1.5 (0.6–3.7)3.3 (1.6–6.6)2 (0.7–5.3)
Max tumour size (mm)0.033*0.7980.006*0.223
< 22151111
≥ 22712.3 (1.1–4.8)1.1 (0.4–3.2)3 (1.4–6.7)2 (0.7–5.6)
Tumour location0.0920.064
Head, uncinate6311
Tail160.8 (0.4–1.4)0.9 (0.5–1.7)
Body, neck100.3 (0.1–1.0)0.3 (0.1–0.9)
Lymph node ratio< 0.001*0.026*
< 0.123411
≥ 0.12555.4 (2.9–10.2)3.7 (1.2–11.4)
Lymph node ratio< 0.001*0.005*
< 0.336211
≥ 0.33274.7 (2.6–8.3)2.8 (1.4–5.8)
ASA score0.7510.764
1, 24411
3, 4380.9 (0.6–1.5)0.9 (0.6–1.5)
Smoking status0.1010.042*0.291
Ex-smoker271.8 (1.0–3.2)1.9 (1.1–3.3)1.7 (0.9–3.3)
Current smoker181.1 (0.5–2.2)1 (0.5–2.0)1.4 (0.6–3.0)
Yes340.8 (0.5–1.3)0.7 (0.4–1.2)
Neo-adjuvant chemotherapy0.6190.173
Yes90.8 (0.3–1.9)0.5 (0.2–1.3)
Adjuvant chemotherapy0.002*0.01*0.004*0.002*
Yes560.5 (0.3–0.8)0.4 (0.2–0.8)0.5 (0.3–0.8)0.4 (0.2–0.7)
Adjuvant radiotherapy0.1850.44
Yes40.4 (0.1–1.6)0.6 (0.2–2.0)
Recurrence< 0.001*< 0.001*
Yes6611.7 (4.1–33.3)36.8 (6.0–224.6)
Type of recurrenced)
Loco-regional410.9481 (0.6–1.7)
Liver320.043*1.7 (1.0–3.0)
Lung200.3491.3 (0.8–2.3)
Peritoneal140.1711.5 (0.8–2.8)
Simultaneous loco-regional + distant metastases290.02*1.9 (1.1–3.1)
Multiple distant metastases140.004*2.5 (1.3–4.8)
Blood tests
Preoperative CA19-9 (U/mL)0.003*
< 500521
≥ 500212.5 (1.4–4.6)
Preoperative CA19-9 (U/mL)0.001*
< 250431
≥ 250302.5 (1.4–4.4)
Pre-recurrence CA19-9 (U/mL)< 0.001*
< 1,500481
≥ 1,500143.8 (1.8–7.7)
Pre-recurrence CA19-9 (U/mL)0.017*
< 100241
≥ 100382.1 (1.1–3.7)
CRP (mg/L)0.440.587
< 3711
≥ 3331.5 (0.6–3.8)1.3 (0.5–3.4)
Albumin (g/L)0.3280.702
< 35711
≥ 35810.7 (0.3–1.5)0.8 (0.3–2.1)
Platelets (× 109/L)0.7230.135
< 3005111
≥ 300371.1 (0.7–1.8)1.5 (0.9–2.4)
Neutrophil (× 109/L)0.003*0.3331.4 (0.7–2.7)0.013*0.804
< 546111
≥ 5422.3 (1.3–3.8)1.9 (1.1–3.1)1.1 (0.6–2.0)
Lymphocyte (× 109/L)0.1120.288
< 1711
≥ 1810.5 (0.2–1.2)0.6 (0.2–1.5)
Monocyte (× 109/L)0.2170.219
< 0.75311
≥ 0.7351.4 (0.8–2.3)1.4 (0.8–2.2)
Preoperative bilirubin (µmol/L)0.08
≤ 12371
> 12481.6 (1.0–2.7)
Preoperative bilirubin (µmol/L)0.013*0.642
< 103411
≥ 10512 (1.1–3.4)0.8 (0.4–1.8)
Blood test-based scoring systems
Glasgow prognostic score0.9040.92
≥ 1131 (0.5–2.3)1 (0.5–2.4)
Systemic inflammation index0.01*0.007*
< 3501111
≥ 350773.4 (1.3–8.6)11.4 (2.0–66.1)
Systemic inflammation index0.011*0.01*
< 5002111
≥ 500672.3 (1.2–4.3)4.5 (1.4–14.3)
Neutrophil-lymphocyte ratio0.004*0.1890.014*
< 220111
≥ 2682.7 (1.4–5.2)0.4 (0.1–1.6)2.3 (1.2–4.4)
Platelet-lymphocyte ratio0.0840.038*0.778
< 15041111
≥ 150471.6 (0.9–2.6)1.7 (1.0–2.8)0.9 (0.4–2.1)
Prognostic nutrition index0.091
< 50311
≥ 50570.6 (0.4–1.1)
Prognostic nutrition index0.284
< 55631
≥ 55250.7 (0.4–1.3)
Lymphocyte-monocyte ratio0.01*0.086
< 21911
≥ 2690.4 (0.2–0.8)0.5 (0.2–1.1)
Lymphocyte-monocyte ratio0.026*0.638
< 4.57611
≥ 4.5120.4 (0.1–0.9)1.4 (0.4–4.8)
PET SUVMax0.846
< 7121
≥ 781.1 (0.3–3.8)
PET SUVMax0.176
< 681
≥ 6122.3 (0.7–7.5)
Pain0.012*< 0.001*0.099
Yes381.9 (1.1–3.1)3.6 (1.8–7.2)1.5 (0.9–2.4)
Yes472.1 (1.2–3.5)0.9 (0.4–1.7)1.9 (1.2–3.1)0.5 (0.2–1.2)
Weight loss0.9170.42
Yes381 (0.6–1.6)0.8 (0.5–1.3)
Yes201.1 (0.6–2.0)1.2 (0.7–2.1)
Yes120.8 (0.4–1.8)1 (0.5–2.1)
Yes210.9 (0.5–1.7)0.8 (0.5–1.5)
Yes131.6 (0.8–3.0)1.7 (0.9–3.2)
Other symptoms0.950.526
Yes231 (0.5–1.8)0.8 (0.5–1.5)
Post recurrence therapy0.403
Yes321.2 (0.7–2.1)

BMI, body mass index; PD, pancreaticoduodenectomy; PPPD, pylorus-preserving pancreaticoduodenectomy; DP, distal pancreatectomy; SMV, superior mesentric vein; SMA, superior mesentric artery; ASA, American Society of Anesthesiologists; CRP, C-reactive protein; HR, hazard ratio; CI, confidence interval.

a)There were only 2 patients with well-differentiated tumour gradings. These patients were combined with patients with moderately differentiated tumour gradings for data analysis. b)Analysis did not include R0 patients as all margins were significant on inclusion of R0 patients. Each variable was analysed against all other margins. c)Lymphovascular and perineural invasion were combined for statistical analysis due to small numbers. d)Each site of recurrence was compared against any other site of recurrence.

*p < 0.05.

Table 2 . Recurrence statistics

Recurrence locationLocation specific onlyTotal recurrence at these sites
Loco-regional12 (18)41
Liver11 (17)32
Lung5 (8)20
Peritoneal3 (5)14
Simultaneous loco-regional + distant metastasis29 (44)29
Multiple distant metastases6 (9)14

Values are presented as number (%) or number only.

Table 3 . Kaplan–Meier analysis of disease-free survival stratified by recurrence location

Recurrence location and p-valueLoco-regional onlyLiver onlyLung onlyPeritoneal onlySimultaneous loco-regional + distant metastasisMultiple distant metastases
Loco-regional only
Liver only0.09
Lung only0.060.38
Peritoneal only0.40.940.43
Simultaneous loco-regional + distant metastasis0.02*0.480.880.64
Multiple distant metastases0.10.50.990.80.97

*p < 0.05.

Fig 2. Kaplan–Meier OS curve for recurrence vs. no recurrence. OS, overall survival.

Fig 3. Scatter plot of disease-free survival stratified by location specific recurrence.

Fig 4. Kaplan–Meier survival curve for DFS stratified by recurrence location. DFS, disease-free survival.

Univariate analysis revealed that N stage, positive tumour resection status, number of R1 margins, uncinate positive margin, simultaneous lymphovascular and perineural invasion, maximum tumour size ≥ 22 mm, lymph node ratio (LNR) ≥ 0.33, smoking status, adjuvant chemotherapy, preoperative CA19-9 level ≥ 250 U/mL, postoperative CA19-9 level ≥ 100 U/mL, neutrophil level ≥ 5 × 109/L, preoperative bilirubin > 10 µmol/L, systemic inflammation index (SII) ≥ 500, neutrophil-lymphocyte ratio (NLR) ≥ 2, platelet-lymphocyte ratio ≥ 150, lymphocyte-monocyte ratio (LMR) ≥ 4.5, and jaundice were significant predictors for recurrence.

In multivariate analysis, preoperative CA19-9 level ≥ 250 U/mL and postoperative CA19-9 level ≥ 100 U/mL were excluded due to missing data. Number of R1 margins and uncinate positive margins were also excluded as information was not available for the total study population. NLR ≥ 2 was excluded due to multi-collinearity. A total of 82 cases were included in the multivariate analysis. Significant predictors for recurrence were SII ≥ 500, LNR ≥ 0.33, and adjuvant chemotherapy. Kaplan–Meier survival curves were generated for each significant predictor (Fig. 57).

Fig 5. Kaplan–Meier DFS curve for SII < 500 vs. ≥ 500. SII, systemic inflammation index; DFS, disease-free survival.

Fig 6. Kaplan–Meier DFS curve for LNR < 0.33 vs. ≥ 0.33. LNR, lymph node ratio; DFS, disease-free survival.

Fig 7. Kaplan–Meier DFS curve for adjuvant chemotherapy vs. no adjuvant chemotherapy. DFS, disease-free survival.

Resection margin analysis

Resection margin status was analysed further. It was found be to significantly predictive of loco-regional recurrence (p = 0.034; hazard ratio [HR], 2.1; 95% CI, 1.1–4.1), but not predictive of other types of recurrence (liver, p = 0.255, HR: 1.5, 95% CI: 0.7–3.2; lung, p = 0.064, HR: 2.7, 95% CI: 1.0–7.5; peritoneal, p = 0.245, HR: 2, 95% CI: 0.6–6.7; simultaneous loco-regional and distant sites, p = 0.173, HR: 1.7, 95% CI: 0.8–3.7; multiple distant sites, p = 0.305, HR: 1.9, 95% CI: 0.6–6.1). However, 57% of patients (n = 20) with R0 resections still developed recurrence.

Survival analysis

At the time of this study, 63 patients (69%) had died. The median OS was 22 months (95% CI, 13.6–30.4 months). One-year and 2-year OS rates were 71% and 50%, respectively.

In univariate analysis, age ≥ 78, N stage, positive tumour resection status, number of R1 margins, uncinate positive margin, simultaneous lymphovascular and perineural invasion, maximum tumour size ≥ 22 mm, LNR ≥ 0.12, adjuvant chemotherapy, PDAC recurrence status, any liver recurrence, simultaneous locoregional + distant metastases recurrence, multiple distant metastases recurrence, preoperative CA19-9 level ≥ 500 U/mL, postoperative CA19-9 level ≥ 1,500 U/mL, neutrophil level ≥ 5 × 109/L, SII ≥ 350, NLR ≥ 2, LMR ≥ 2, pain, and jaundice were significant prognostic factors for OS.

In multivariate analysis, preoperative CA19-9 level ≥ 500 U/mL and postoperative CA19-9 level ≥ 1,500 U/mL were excluded due to missing data. Uncinate positive margin, any liver recurrence, simultaneous loco-regional with distant recurrence, and multiple distant site recurrence covariates were also excluded as these covariates were not present in the total study population. The number of R1 margins as a covariate was then excluded due to multi-collinearity. A total of 82 cases were included in the multivariate analysis. Significant predictors for prognosis in multivariate analysis were N stage, LNR ≥ 0.12, adjuvant chemotherapy, recurrence, SII, and pain.

Recurrence location and overall survival

OS stratified by recurrence location revealed significant differences between loco-regional only recurrence and simultaneous loco-regional with distant recurrence (p = 0.01, pairwise log-rank test) as well as locoregional only recurrence and multiple distant site recurrence (p = 0.02, pairwise log-rank test) (Fig. 8, Table 4). The OS was 32.5 months (95% CI, 12.8–52.2 months) for loco-regional alone, 20.4 months (95% CI, 3.8–37 months) for liver alone, 13.2 months (95% CI, 5.7–20.6 months) for lung alone, 27.4 months (95% CI, 15.3–39.5 months) for peritoneal alone, 11.3 months (95% CI, 8.0–14.5 months) for simultaneous loco-regional and distant metastatic recurrence, and 13.7 months (95% CI, 3.7–23.7 months) for multiple distant metastases.

Table 4 . Kaplan–Meier analysis of overall survival stratified by recurrence location

Recurrence location and p-valueLoco-regional onlyLiver onlyLung onlyPeritoneal onlySimultaneous loco-regional + distant metastasisMultiple distant metastases
Loco-regional only
Liver only0.07
Lung only0.230.68
Peritoneal only0.90.250.34
Simultaneous loco-regional + distant metastasis0.01*0.380.560.2
Multiple distant metastases0.02*0.410.410.170.95

*p < 0.05.

Fig 8. Kaplan–Meier survival curve for OS stratified by recurrence location. OS, overall survival.

Analysis comparing any liver recurrence against any other recurrence location demonstrated a significantly worse OS (p = 0.041; HR, 1.7; 95% CI, 1.0–2.9). Similarly, simultaneous loco-regional recurrence had worse OS (p = 0.02; HR, 1.9; 95% CI, 1.1–3.1) as well as multiple distant site recurrence (p = 0.004; HR, 2.5; 95% CI, 1.3–4.8). No significantly worse OS was found for loco-regional (p = 0.948; HR, 1; 95% CI, 0.6–1.7), lung (p = 0.349; HR, 1.3; 95% CI, 0.8–2.3), or peritoneal (p = 0.171; HR, 1.5; 95% CI, 0.8–2.8) recurrence.


Patterns of recurrence

Despite a complex major surgery for resection of PDAC with an intention to cure, recurrence is common. This study found a significant difference between first site loco-regional only recurrence and simultaneous loco-regional and distant recurrence. Other studies have reported a relatively shorter disease-free interval for liver recurrence than lung recurrence [5,14,15]. A Japanese study analysing 524 patients who underwent PDAC resections found that the median time to recurrence was five months for liver as compared to 18 months for lung [9]. Histological differences such as venous invasion were noticed between liver and lung recurrence [9,16]. Kubo et al. [16] also found that even in patients receiving neo-adjuvant therapy, the presence of microscopic invasion could increase the risk of recurrence, particularly hepatic relapse. It has been postulated that the close proximity of PDAC to the portal venous system might lead to earlier recurrence in the liver [17]. Molecular alterations in PDAC such as SMAD4 gene mutation as well as those affecting DNA repair mechanisms could potentially lead to different sites of recurrence [18,19]. These variables could not be assessed in this study. They could explain the difference in recurrence patterns from other studies.

The recurrence rate within the first year after surgery was 48% (n = 44), decreasing to 21% (n = 19) in the second year and 3% (n = 3) in the third year, consistent with other studies demonstrating a high rate of recurrence within the first year (40%–61%) [9,20,21] possibly due to undiagnosed micro-metastases at the time of surgery [22]. These sub-clinical micro-metastases could then be classified as postoperative recurrence within a short postoperative timeframe [8,23]. In order to diagnose micro-metastases, circulating tumour cells or circulating tumour DNA may help [24,25]. Adjuvant and/or neo-adjuvant chemotherapy may reduce recurrence rates and alter recurrence patterns [6,26]. However, the natural history of PDAC without systemic therapy points to an aggressive disease [27,28].

Predictors of recurrence

This study investigated more than 40 covariates, significantly more than most other studies. Only LNR, adjuvant chemotherapy, and SII were found to be significant predictors of recurrence in multivariate analysis. While most studies looked at LNR and its association with prognosis, several studies looked at its association with recurrence and found significant results [10,29,30]. A large study with 692 patients found that LNR was associated with recurrence and only distant recurrence on further analysis [10]. However, this was disputed by another recent study which found that LNR was significantly associated with overall recurrence, but not with local or distant recurrence specifically [29]. In the present study, LNR was significantly predictive of recurrence in all locations apart from peritoneal recurrence possibly due to the low number. Implications of LNR as a predictor of recurrence could help guide prognosis selection for adjuvant treatment. Further mechanistic studies to determine how LNR predicts recurrence could guide therapy or increase surveillance. This study used the optimum cut-off value to stratify the covariate into a categorical variable, which was opposed to other studies where either the median value was used as the cut-off value or how the value was obtained was not mentioned. More work is needed in this area to determine a specific cut-off value for further clinical application.

Adjuvant chemotherapy is known to reduce recurrence after surgical resection for PDAC as evidenced by several recent randomised controlled trials (RCTs) [5,6]. However, even in a secondary analysis of ESPAC4, one of the largest RCTs for adjuvant treatment, there was no statistically significant difference in recurrence or survival statistics between recurrence sites [5]. However, not all patients would be fit enough to undergo adjuvant chemotherapy after a major resection for PDAC. Neo-adjuvant therapy has been reported to reduce the risk of early local recurrence and recommended to be used together with adjuvant therapy [31]. A systematic review has confirmed that local recurrence, but not distant metastasis, is reduced when neo-adjuvant therapy is used [32]. In the present study, neo-adjuvant therapy was not significantly associated with overall, local, or distant recurrence, although numbers were small.

SII can be easily obtained preoperatively. It has been reported to be predictive of OS and disease-free survival in different types of cancer [33,34]. There is at least one study in advanced pancreatic cancer, but several others with resection of other cancers [35-37]. A large prospective study with 321 patients has reported that SII is predictive of OS, but not recurrence [38]. However, another study with 590 patients has found that SII is an independent predictor of recurrence [39]. The present study corroborates findings of the latter study. However, there are limitations to its use. Neo-adjuvant therapy was not accounted for when studying SII. Furthermore, it has also been reported that its accuracy is reduced when bilirubin levels are high as it can cause immunomodulatory effects [39].

Resection margin status (R1 vs. R0) in pancreatic cancer has been studied extensively. Several studies have reported that R1 margin is correlated with recurrence [7,10,40,41]. However, others have found no significant difference in recurrence with R1 margins vs. R0 margins [42,43]. Some studies have analysed this further and found that the resection margin status is only significantly associated with local recurrence [7,10,44], consistent with findings of the present study, which found that R1 status was significantly associated with loco-regional recurrence. However, heterogeneity exists when reporting resection margin status. A large South Korean study with 558 patients has reported that R1 margin can predict early recurrence [41]. However, it only studied patients with distal pancreatectomies. Although a large trial with 912 patients reported no difference in DFS, all patients had adjuvant therapy in that trial [43]. A systematic review also noted issues with studies reporting total recurrence as opposed to isolated recurrence [7], while another study noted existing differences in definitions of R1 status [45]. In the present study, 57% of patients with R0 resection still developed recurrence. In several studies, resection margin status was reported to be not significantly associated with OS [43,45]. Standardised resection margin reporting will help reduce reporting variations of R1 margin after PDAC resection.


This study was limited by its small population size which reduced the power of the study. There was also heterogeneity in surveillance for patients in this study as some were followed up in the tertiary HPB center while some were followed up by their oncologists. Follow-up regimens varied significantly across oncologist groups. Furthermore, some patients also diagnosed with recurrence after developing symptoms. Some factors made it difficult to determine the exact time point when recurrence occurred. Lastly, in some instances, postoperative CT scans were difficult to interpret as postoperative inflammation appeared similar to recurrence. Thus, MDT decision was used as consensus reporting.

In conclusion, timing to loco-regional only recurrence was significantly longer than simultaneous loco-regional with distant recurrence. Significant predictors for recurrence were LNR, adjuvant chemotherapy, and SII. More research with large national cohorts is needed to investigate patterns and predictive factors affecting recurrence after curative resection for PDAC.


Barts Pancreas Tissue Bank is funded by Pancreatic Cancer Research Fund. The Operations Group is chaired by Prof Hemant Kocher/Prof Claude Chelala and Tissue Access Committee by Professor Richard Grose. Vickna Balarajah, Abhirup Banerjee, Mafalda CostaNeves, Konstantinos Stasinos, Mohamed Zardab, Andreas Konstantinou helped with clinical verification of data collection. Past members helping in running of Barts Pancreas Tissue Bank include, Archana Ambily, Thomas Dowe, Ajith Vajrala, and Ope Banwo. We are grateful to consultant surgeons at Barts: Satyajit Bhattacharya, Robert Hutchins, Ajit Abraham, Vincent Yip, Deepak Hariharan, Omar Mowanah for access to their patients. We are grateful to Dr Joanne Chin-Aleong, Consultant Pathologist and Dr David Propper/Dr Sarah Slater, Consultant Oncologist for their valuable advice.




No potential conflict of interest relevant to this article was reported.


Conceptualization: AA, HMK. Data curation: AA, AM, HMK. Methodology: All authors. Visualization: AA, DU, HMK. Writing - original draft: AA, HMK. Writing - review & editing: All authors.

  1. Carioli G, Malvezzi M, Bertuccio P, Boffetta P, Levi F, La Vecchia C, et al. European cancer mortality predictions for the year 2021 with focus on pancreatic and female lung cancer. Ann Oncol 2021;32:478-487.
    Pubmed CrossRef
  2. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, et al. GLOBOCAN 2012: Estimated Cancer Incidence and Mortality Worldwide in 2012 v1.0: IARC cancerbase No. 11. International Agency for Research on Cancer, 2012.
  3. Eliison LF, Wilkins K. An update on cancer survival. Heal Rep 2010;21:55-60.
  4. Ducreux M, Cuhna AS, Caramella C, Hollebecque A, Burtin P, Goéré D, et al. Cancer of the pancreas: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2015;26 Suppl 5:v56-68.
    Pubmed CrossRef
  5. Jones RP, Psarelli EE, Jackson R, Ghaneh P, Halloran CM, Palmer DH, et al. Patterns of recurrence after resection of pancreatic ductal adenocarcinoma: a secondary analysis of the ESPAC-4 randomized adjuvant chemotherapy trial. JAMA Surg 2019;154:1038-1048.
    Pubmed KoreaMed CrossRef
  6. Conroy T, Castan F, Lopez A, Turpin A, Ben Abdelghani M, Wei AC, et al. Five-year outcomes of FOLFIRINOX vs gemcitabine as adjuvant therapy for pancreatic cancer: a randomized clinical trial. JAMA Oncol 2022;8:1571-1578.
    Pubmed KoreaMed CrossRef
  7. Kalisvaart M, Broadhurst D, Marcon F, Pande R, Schlegel A, Sutcliffe R, et al. Recurrence patterns of pancreatic cancer after pancreatoduodenectomy: systematic review and a single-centre retrospective study. HPB (Oxford) 2020;22:1240-1249.
    Pubmed CrossRef
  8. Tanaka M, Mihaljevic AL, Probst P, Heckler M, Klaiber U, Heger U, et al. Meta-analysis of recurrence pattern after resection for pancreatic cancer. Br J Surg 2019;106:1590-1601.
    Pubmed CrossRef
  9. Hayashi K, Ono Y, Takamatsu M, Oba A, Ito H, Sato T, et al. Prediction of recurrence pattern of pancreatic cancer post-pancreatic surgery using histology-based supervised machine learning algorithms: a single-center retrospective study. Ann Surg Oncol 2022;29:4624-4634.
    Pubmed CrossRef
  10. Groot VP, Rezaee N, Wu W, Cameron JL, Fishman EK, Hruban RH, et al. Patterns, timing, and predictors of recurrence following pancreatectomy for pancreatic ductal adenocarcinoma. Ann Surg 2018;267:936-945.
    Pubmed CrossRef
  11. Watanabe Y, Nishihara K, Niina Y, Kudo Y, Kurata K, Okayama T, et al. Patients with lung recurrence after curative resection for pancreatic ductal adenocarcinoma have a better prognosis than those with recurrence at other sites. JOP J Pancreas 2017;18:54-61.
  12. Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res 2004;10:7252-7259.
    Pubmed CrossRef
  13. Kim JH. Multicollinearity and misleading statistical results. Korean J Anesthesiol 2019;72:558-569.
    Pubmed KoreaMed CrossRef
  14. Groot VP, Gemenetzis G, Blair AB, Ding D, Javed AA, Burkhart RA, et al. Implications of the pattern of disease recurrence on survival following pancreatectomy for pancreatic ductal adenocarcinoma. Ann Surg Oncol 2018;25:2475-2483.
    Pubmed KoreaMed CrossRef
  15. Kolbeinsson H, Hoppe A, Bayat A, Kogelschatz B, Mbanugo C, Chung M, et al. Recurrence patterns and postrecurrence survival after curative intent resection for pancreatic ductal adenocarcinoma. Surgery 2021;169:649-654.
    Pubmed CrossRef
  16. Kubo H, Ohgi K, Sugiura T, Ashida R, Yamada M, Otsuka S, et al. The association between neoadjuvant therapy and pathological outcomes in pancreatic cancer patients after resection: prognostic significance of microscopic venous invasion. Ann Surg Oncol 2022;29:4992-5002.
    Pubmed CrossRef
  17. Zheng B, Ohuchida K, Yan Z, Okumura T, Ohtsuka T, Nakamura M, et al. Primary recurrence in the lung is related to favorable prognosis in patients with pancreatic cancer and postoperative recurrence. World J Surg 2017;41:2858-2866.
    Pubmed CrossRef
  18. Ferguson MD, Dong L, Wan J, Deneve JL, Dickson PV, Behrman SW, et al. Molecular alterations associated with DNA repair in pancreatic adenocarcinoma are associated with sites of recurrence. J Gastrointest Cancer 2019;50:285-291.
    Pubmed CrossRef
  19. Shin SH, Kim HJ, Hwang DW, Lee JH, Song KB, Jun E, et al. The DPC4/SMAD4 genetic status determines recurrence patterns and treatment outcomes in resected pancreatic ductal adenocarcinoma: a prospective cohort study. Oncotarget 2017;8:17945-17959.
    Pubmed KoreaMed CrossRef
  20. Ng KYY, Chow EWX, Jiang B, Lim C, Goh BKP, Lee SY, et al. Resected pancreatic adenocarcinoma: an Asian institution's experience. Cancer Rep 2021;4:e1393.
    Pubmed KoreaMed CrossRef
  21. Maggino L, Malleo G, Crippa S, Belfiori G, Nobile S, Gasparini G, et al. CA19.9 response and tumor size predict recurrence following post-neoadjuvant pancreatectomy in initially resectable and borderline resectable pancreatic ductal adenocarcinoma. Ann Surg Oncol 2023;30:207-219.
    Pubmed KoreaMed CrossRef
  22. Haeno H, Gonen M, Davis MB, Herman JM, Iacobuzio-Donahue CA, Michor F. Computational modeling of pancreatic cancer reveals kinetics of metastasis suggesting optimum treatment strategies. Cell 2012;148:362-375.
    Pubmed KoreaMed CrossRef
  23. Barnes CA, Aldakkak M, Christians KK, Clarke CN, Dua K, George B, et al. Radiographic patterns of first disease recurrence after neoadjuvant therapy and surgery for patients with resectable and borderline resectable pancreatic cancer. Surgery 2020;168:440-447.
    Pubmed CrossRef
  24. Park Y, Jun HR, Choi HW, Hwang DW, Lee JH, Song KB, et al. Circulating tumour cells as an indicator of early and systemic recurrence after surgical resection in pancreatic ductal adenocarcinoma. Sci Rep 2021;11:1644.
    Pubmed KoreaMed CrossRef
  25. Hugenschmidt H, Labori KJ, Borgen E, Brunborg C, Schirmer CB, Seeberg LT, et al. Preoperative CTC-detection by cellsearch® is associated with early distant metastasis and impaired survival in resected pancreatic cancer. Cancers 2021;13:485.
    Pubmed KoreaMed CrossRef
  26. Uesaka K, Boku N, Fukutomi A, Okamura Y, Konishi M, Matsumoto I, et al. Adjuvant chemotherapy of S-1 versus gemcitabine for resected pancreatic cancer: a phase 3, open-label, randomised, non-inferiority trial (JASPAC 01). Lancet 2016;388:248-257.
    Pubmed CrossRef
  27. Hernandez JM, Morton CA, Al-Saadi S, Villadolid D, Cooper J, Bowers C, et al. The natural history of resected pancreatic cancer without adjuvant chemotherapy. Am Surg 2010;76:480-485.
    Pubmed CrossRef
  28. Hishinuma S, Ogata Y, Tomikawa M, Ozawa I, Hirabayashi K, Igarashi S. Patterns of recurrence after curative resection of pancreatic cancer, based on autopsy findings. J Gastrointest Surg 2006;10:511-518.
    Pubmed CrossRef
  29. Honselmann KC, Pergolini I, Castillo CF, Deshpande V, Ting D, Taylor MS, et al. Timing but not patterns of recurrence are different between node-negative and node-positive resected pancreatic cancer. Ann Surg 2020;272:357-365.
    Pubmed KoreaMed CrossRef
  30. Walston S, Salloum J, Grieco C, Wuthrick E, Diaz DA, Barney C, et al. Identifying clinical factors which predict for early failure patterns following resection for pancreatic adenocarcinoma in patients who received adjuvant chemotherapy without chemoradiation. Am J Clin Oncol 2018;41:1185-1192.
    Pubmed KoreaMed CrossRef
  31. Suto H, Okano K, Oshima M, Ando Y, Takahashi S, Shibata T, et al. The predictors and patterns of the early recurrence of pancreatic ductal adenocarcinoma after pancreatectomy: the influence of pre- and post- operative adjuvant therapy. BMC Surg 2019;19:186.
    Pubmed KoreaMed CrossRef
  32. Schorn S, Demir IE, Samm N, Scheufele F, Calavrezos L, Sargut M, et al. Meta-analysis of the impact of neoadjuvant therapy on patterns of recurrence in pancreatic ductal adenocarcinoma. BJS Open 2018;2:52-61.
    Pubmed KoreaMed CrossRef
  33. Zhou Q, Su S, You W, Wang T, Ren T, Zhu L. Systemic inflammation response index as a prognostic marker in cancer patients: a systematic review and meta-analysis of 38 cohorts. Dose Response 2021;19:15593258211064744.
    Pubmed KoreaMed CrossRef
  34. Cao W, Shao Y, Zou S, Wang N, Wang J. Prognostic significance of systemic immune-inflammation index in patients with bladder cancer: a systematic review and meta-analysis. Medicine (Baltimore) 2022;101:e30380.
    Pubmed KoreaMed CrossRef
  35. Liu XC, Jiang YP, Sun XG, Zhao JJ, Zhang LY, Jing X. Prognostic significance of the systemic immune-inflammation index in patients with cholangiocarcinoma: a meta-analysis. Front Oncol 2022;12:938549.
    Pubmed KoreaMed CrossRef
  36. Zhang K, Hua YQ, Wang D, Chen LY, Wu CJ, Chen Z, et al. Systemic immune-inflammation index predicts prognosis of patients with advanced pancreatic cancer. J Transl Med 2019;17:30.
    Pubmed KoreaMed CrossRef
  37. Liu X, Sun X, Liu J, Kong P, Chen S, Zhan Y, et al. Preoperative C-reactive protein/albumin ratio predicts prognosis of patients after curative resection for gastric cancer. Transl Oncol 2015;8:339-345.
    Pubmed KoreaMed CrossRef
  38. Jomrich G, Gruber ES, Winkler D, Hollenstein M, Gnant M, Sahora K, et al. Systemic immune-inflammation index (SII) predicts poor survival in pancreatic cancer patients undergoing resection. J Gastrointest Surg 2020;24:610-618.
    Pubmed KoreaMed CrossRef
  39. Aziz MH, Sideras K, Aziz NA, Mauff K, Haen R, Roos D, et al. The systemic-immune-inflammation index independently predicts survival and recurrence in resectable pancreatic cancer and its prognostic value depends on bilirubin levels: a retrospective multicenter cohort study. Ann Surg 2019;270:139-146.
    Pubmed CrossRef
  40. Kim HJ, Lee WJ, Kang CM, Hwang HK, Bang SM, Song SY, et al. Risk factors associated with loco-regional failure after surgical resection in patients with resectable pancreatic cancer. PLoS One 2016;11:e0157196.
    Pubmed KoreaMed CrossRef
  41. Kwon J, Lee SR, Park SY, Lee JH, Song KB, Hwang DW, et al. Effect of resection margin status on recurrence pattern and survival in distal pancreatectomy for left-sided pancreatic ductal adenocarcinoma. J Hepatobiliary Pancreat Sci 2023;30:633-643.
    Pubmed CrossRef
  42. Raut CP, Tseng JF, Sun CC, Wang H, Wolff RA, Crane CH, et al. Impact of resection status on pattern of failure and survival after pancreaticoduodenectomy for pancreatic adenocarcinoma. Ann Surg 2007;246:52-60.
    Pubmed KoreaMed CrossRef
  43. Kurreck A, Weckwerth J, Modest DP, Striefler JK, Bahra M, Bischoff S, et al. Impact of completeness of adjuvant gemcitabine, relapse pattern, and subsequent therapy on outcome of patients with resected pancreatic ductal adenocarcinoma - A pooled analysis of CONKO-001, CONKO-005, and CONKO-006 trials. Eur J Cancer 2021;150:250-259.
    Pubmed CrossRef
  44. Tummers WS, Groen JV, Sibinga Mulder BG, Farina-Sarasqueta A, Morreau J, Putter H, et al. Impact of resection margin status on recurrence and survival in pancreatic cancer surgery. Br J Surg 2019;106:1055-1065.
    Pubmed KoreaMed CrossRef
  45. Yamamoto T, Uchida Y, Terajima H. Clinical impact of margin status on survival and recurrence pattern after curative-intent surgery for pancreatic cancer. Asian J Surg 2019;42:93-99.
    Pubmed CrossRef


May 2024, 28 (2)
Full Text(PDF) Free
PubMed Central

Social Network Service


Cited By Articles
  • CrossRef (0)

Author ORCID Information