Consensus statement on the use of systemic therapies in pancreatic ductal adenocarcinoma in Asia

Abstract

Systemic therapies, primarily cytotoxic chemotherapies, are the mainstay treatment option for pancreatic ductal adenocarcinoma (PDAC) alongside surgery. Several guidelines exist for managing PDAC within Asia, but recommendations for systemic therapies can differ significantly. Consequently, geographical areas across Asia not covered by international or regional guidelines often have disparate care for PDAC patients. To address this, we utilised the Delphi method to establish a collective opinion on the optimal use of systemic therapies for PDAC patients in Asia. Fourteen comprehensive consensus recommendations are reported. For resectable/localised and borderline resectable disease, recommendations were developed on specific chemotherapeutic regimens for adjuvant and neoadjuvant settings and best practice monitoring. Recommendations on downstaging/conversion therapy and therapy options were developed for locally advanced disease. In metastatic disease, recommendations for first- and second-line therapy options were developed, covering overall treatment strategy and best supportive care. In addition, consensus recommendations for molecular and genomic testing in PDAC were developed. To our knowledge, we report the first consensus statement on the optimal use of systemic therapies in PDAC for a broad, cross-border Asian population. These expert recommendations may serve as a starting point to improve the standardisation of treatment practices and care for PDAC patients within Asia.

Highlights

• •National guidelines for managing PDAC within Asia are sparse.
• •Different systemic therapy use across Asia affects care quality for PDAC patients.
• •Expert recommendations for PDAC systemic therapy in Asia have been developed.
• •Asia-specific consensus recommendations aim to help uniform PDAC treatment.

Introduction

Pancreatic cancer, particularly pancreatic ductal adenocarcinoma (PDAC) making up 90% of cases, is known for its aggressive nature and dismal prognosis, with global incidence rates close to mirroring its high mortality rates.^1,2 Despite a static 5-year survival rate of ∼4%, incidence in Asia is expected to surge by 81% by 2040.1, 2, 3 Diagnosis challenges often lead to late detection, making systemic therapies crucial, especially for advanced stages. For early-stage disease, surgery aims for cure, supported by systemic (neo)adjuvant therapies to enhance outcomes.^3,4

International guidelines from the American Society of Clinical Oncology (ASCO) 2016 and updated 2020 guidelines,^5,6 the National Comprehensive Cancer Network (NCCN) 2024 clinical practice guidelines⁴ and the European Society for Medical Oncology (ESMO) 2015 practice guidelines³ and the recent 2023 update are available.⁷ In addition, several national guidelines that focus on the use of systemic therapy for pancreatic cancer within Asia are available, including in China (2022),⁸ Hong Kong (2024),⁹ Japan (2019, 2022),10, 11, 12 Korea (2021)¹³ and Singapore (2015).¹⁴ National guidelines covering surgical approaches separately in Korea (2022)¹⁵ are also available. However, there are instances where these guidelines differ in their recommendations, especially for borderline resectable or locally advanced pancreatic cancer, where evidence from randomised clinical trials is limited due to patient exclusion criteria.¹⁶ Recommendations for adjuvant chemoradiotherapy in resected patients differ significantly between the European and US guidelines.^3,4 NCCN 2024 guidelines recommend adjuvant radiotherapy for patients with a high risk for local recurrence.⁴ The ESMO 2015 guidelines do not recommend adjuvant chemoradiation after surgery outside clinical trials.³ No sufficient evidence for adjuvant radiotherapy for resectable pancreatic cancer outside a clinical trial was found in China.⁸ In contrast, the 2021 Korean guidelines recommend conditional adjuvant radiotherapy based on moderate evidence in this setting.¹³ In Singapore, adjuvant radiotherapy concurrent with fluoropyrimidine is only recommended routinely for PDAC patients when the resection margin is positive.¹⁴ S-1 is an oral fluoropyrimidine derivative recommended for treating PDAC in various settings in Asian populations.^8,17 S-1 monotherapy is recommended as a first-line treatment option for metastatic pancreatic cancer in the 2022 Japan guidelines, albeit a weak recommendation, based on the results of a phase III study (GEST study) conducted in Japan and Taiwan.^10,18,19 However, S-1 monotherapy in this metastatic setting is not currently recommended in the 2015 Singaporean guidelines or for Caucasian populations in the latest ESMO or NCCN guidelines.^3,4,14,20

These differences in strategies, compounded by socioeconomic disparities and health care access challenges in rural Asian areas, highlight the issues of uneven care quality and outcomes in PDAC management.^21,22 This manuscript aims to provide a foundational starting point for Asia-specific recommendations on systemic therapy for pancreatic cancer, tailored to the region’s unique clinical, economic and demographic contexts. Using expert consensus, we aim to address variations in care and propose standardisation opportunities to improve patient outcomes. This effort does not intend to replicate or encompass the comprehensive scope of formal international guidelines, but rather complements existing frameworks by focusing on the specific needs and challenges faced in Asia.

Methods

A multidisciplinary steering committee of 10 pancreatic cancer experts (2 surgeons, 1 gastroenterologist, 7 oncologists) from across Asia (China, Hong Kong, Japan, Korea and Singapore) met in May 2023 to discuss providing expert consensus recommendations on the use of systemic therapies in PDAC. The authors for this publication comprise the steering committee members. The steering committee reviewed the literature on existing guidelines and drafted statements covering different disease stages and considerations, refined in a face-to-face meeting. These were reviewed by a broader 19-expert panel from Asian countries, including Taiwanese representatives, using a modified Delphi method for consensus.²³ Panel members voted on each statement (or component thereof) independently and anonymously using a 5-point Likert scale of acceptability (1 = strongly agree; 2 = agree; 3 = neutral; 4 = disagree; 5 = strongly disagree). Statements reaching over 75% agreement were accepted; those falling short were revised for a second voting round, ensuring the recommendations reflected broad expert agreement. Generally, the statements were structured according to stage of disease: resectable/localised, borderline resectable, locally advanced and metastatic, with an additional set for other considerations. A list of the members of the Delphi panel is provided in Supplementary Table S1, available at https://doi.org/10.1016/j.esmogo.2025.100158.

Evidence level, grades and recommendation strengths were assessed using the Centre for Evidence-Based Medicine 2011 Levels of Evidence (CEBM)²⁴ and graded by the expert panel during the consensus process, as detailed in Supplementary Tables S2 and S3, available at https://doi.org/10.1016/j.esmogo.2025.100158.

No new patients participated in this study. All data/statements are sourced from existing publications or the opinions of the authors.

Results

Literature evidence and statements

A summary of clinical and real-world evidence used to develop the recommendation statements is shown in Table 1. A total of 41 studies, including 23 phase II/III randomised studies of which 2 were double-blinded, 2 multicentre prospective studies and 2 single-centre prospective studies, 1 multicentre and 6 single-centre retrospective studies, 2 observational studies, 4 meta-analyses of various studies and 1 case series study, are included as evidence to support the recommendations on the optimal use of systemic therapies in PDAC.

Table 1.

Key clinical efficacy evidence for systematic therapies in PDAC

Study name (ID)	Design	Population (n)	Drug(s)	Primary endpoint	Key efficacy outcomes	Refs	Level of evidence
Resectable/localised PDAC
APACT (NCT01964430)	Multicentre, randomised, open-label phase III trial	866 treatment-naïve patients with resected PDAC (597 completed treatment)	Adjuvant nab-paclitaxel + GEM (GnP) versus GEM	Disease-free survival	The primary endpoint was not met. Median DFS was similar in the GnP group versus GEM alone (19.4 versus 18.8 months; HR 0.88, 95% CI 0.729-1.063, P = 0.18); however, a favourable OS was seen with GnP. At the 5-year follow-up, median OS was 41.8 versus 37.7 months for GnP versus GEM, respectively (HR 0.80, 95% CI 0.678-0.947, P = 0.0091).	Tempero et al. 202325	I
Kang 2022	Single-centre, retrospective study	292 Asian patients with resected PDAC	Adjuvant GEM versus GEMCAP	Efficacy and safety	With a median follow-up of 39.4 months, median OS was 36.8 versus 46.1 months for GEM versus GEMCAP, respectively (HR 0.7, 95% CI 0.5-1.0, P = 0.07). Median recurrence-free survival was 14.3 versus 17.0 months, respectively (P = 0.5). AEs were more common in the GEMCAP group, including hand–foot skin reactions (15.3% versus 0.6%), neutropenia (78.6% versus 67.7%) and thrombocytopenia (30.5% versus 20.5%).	Kang et al. 202226	III
Tang 2022	Single-centre, retrospective study	241 Asian patients with resected PDAC	Adjuvant GEM + S-1 (GS) versus S-1	Efficacy and safety	Median DFS (15.1 versus 15.9 months, P = 0.524) and OS (34.8 versus 27.1 months, P = 0.338) did not differ significantly between the GS and S-1 groups. Chemotherapy completion rates were higher with S-1 (75.7% versus 52.4%, P = 0.006), while grade 3-4 neutropenia was more frequent with GS (49.5% versus 18.2%, P = 0.015).	Tang et al. 202227	III
PRODIGE-24/ACCORD-24/CCTG-PA-6 (NCT01526135)	Multicentre, randomised, open-label phase III trial	493 patients with resected PDAC	Adjuvant mFFX versus GEM	Disease-free survival	At a median follow-up of 33.6 months, median DFS was 21.6 versus 12.8 months for mFFX versus GEM (HR 0.58, 95% CI 0.46-0.73, P < 0.001), with 3-year DFS rates of 39.7% versus 21.4%.	Conroy et al. 201828	I
ESPAC-4 (ISRCTN96397434)	Multicentre, randomised, open-label phase III trial	730 patients with resected PDAC	Adjuvant GEM mono versus GEMCAP	Overall survival	Median OS was significantly higher in the GEMCAP group versus the GEM mono group at 28.0 versus 25.5 months (HR 0.82, 95% CI 0.68-0.98, P = 0.032).	Neoptolemos et al. 201729	I
JASPAC 01 (UMIN000000655)	Multicentre, randomised, open-label, noninferiority phase III trial	385 patients with resected PDAC	Adjuvant GEM versus S-1	Overall survival	S-1 was superior to GEM, with HR for mortality 0.57 (95% CI 0.44-0.72, P < 0.0001), with 5-year OS rates of 24.4% versus 44.1% for GEM versus S-1.	Uesaka et al. 201630	II
GERCOR/FFCD-9203 (EORTC-40013-22012)	Multicentre, randomised, open-label phase II trial	90 patients with resected PC	Adjuvant GEM versus GEM + RT	Feasibility and tolerability of treatment	GEM-based RT was feasible and well tolerated. Median DFS was 12 (95% CI 10.1-19.3) months in the GEM + RT group and 11 (95% CI 8.3-16.0) months in the GEM group. Median OS was 24 months in both arms. First local recurrence was less frequent in the GEM + RT group (11% versus 24%).	Van Laethem et al. 201031	II
CONKO-001 (ISRCTN34802808)	Multicentre, randomised, open-label, parallel-group phase III trial	368 patients with resected PC randomised (354 ITT population)	Adjuvant GEM versus observation	Disease-free survival	Median DFS was significantly higher in the GEM group versus the observation group at 13.4 versus 6.7 months for treatment versus observation (HR 0.55, 95% CI 0.44-0.69, P < 0.001).	Oettle et al. 2013, 200732,33	II
ESPAC-1 (ISRCTN21587955)	Multicentre, 2 × 2 factorial design, randomised trial	541 patients with resected PDAC	Adjuvant 5-FU + RT versus 5-FU versus 5-FU + RT and 5-FU versus observation	Two-year survival rate	Overall, there was no benefit for adjuvant chemoRT but a significant survival benefit for chemotherapy. Median survival was 15.5 months with chemoRT versus 16.1 months in patients without (HR 1.18, 95% CI 0.90-1.55, P = 0.24). For chemotherapy versus no chemotherapy, median survival was 19.7 versus 14.0 months (HR 0.66, 95% CI 0.52-0.83, P = 0.0005).	Neoptolemos et al. 2001, 200434,35	II
Borderline resectable PDAC
Cloyd 2020	Meta-analysis of 6 prospective RCTs	850 patients with resectable or borderline resectable PDAC	Various neoadjuvant chemotherapy versus upfront surgery	OS	Neoadjuvant improved OS compared with surgery first (HR 0.73, 95% CI 0.61-0.86). This effect was consistent across anatomical classifications (potentially retractable: HR 0.73, 95% CI 0.59-0.91; borderline retractable: HR 0.51, 95% CI 0.28-0.93) and neoadjuvant types (chemoradiation: HR 0.77, 95% CI 0.61-0.98; chemotherapy alone: HR 0.68, 95% CI 0.54-0.87).	Cloyd et al. 202036	II
Prep-02/JSAP05 (UMIN000009634)	Multicentre, two-arm, open-label, randomised phase II/III trial	362 patients with resectable PDAC	Neoadjuvant GEM and S-1 (GS) followed by surgery versus upfront surgery	Overall survival	The neoadjuvant chemotherapy group (GS) had significantly longer survival than the upfront surgery group. Median OS was 36.7 versus 26.6 months, respectively (HR 0.72, P = 0.015), 2.5 years after final enrolment.	Satoi et al. 201937	II
Verstejine 2018	Meta-analysis of 38 studies (3 RCTs, 9 phase I/II, 26 prospective/retrospective cohort studies)	3484 patients with resectable or borderline resectable PC; 1738 patients had neoadjuvant treatment	Various—but at least chemotherapy as neoadjuvant (mainly GEM) versus upfront surgery	Overall survival	Median OS ITT was 18.8  months for neoadjuvant versus 14.8  months for upfront surgery, with a larger difference seen in resected tumours (26.1 versus 15.0 months, respectively).	Versteijne et al. 201816	II
Christians 2014	Single-centre, observational study	18 patients with borderline resectable PC	Neoadjuvant FFX followed by GEM or CAP chemoRT	Surgical resectability	Initial experience with FFX followed by chemoRT as neoadjuvant suggests favourable resection rates. Neoadjuvant therapy was completed by 15 of 18 patients (83%), with 12 (67%) proceeding to pancreatectomy. All 12 resected patients had negative (R0) margins, and only 2 (17%) were node positive. Portal vein resection/reconstruction was carried out in 10 (83%). No in-hospital or 30-day mortalities, clinical pancreatic leaks or reoperations were reported. Survival: Of the 12 patients who completed all therapy, 7 (58.3%) are alive, including 5 without evidence of disease (median time from diagnosis 22 months, range 18-35 months). The six patients who did not complete planned therapy are deceased (6.9-17.5 months from diagnosis).	Christians et al. 201438	III
Neoadjuvant GS	Multicentre, single-arm phase II trial	35 patients with resectable borderline PDAC	Neoadjuvant GEM and S-1 (GS)	Two-year survival rate	Two-year survival rate was 45.7%. Patients who underwent resection without metastases after neoadjuvant chemotherapy with GS (n = 27) had an increased median OS (34.7 months) compared with those who did not undergo resection (P = 0.0017).	Motoi et al. 201339	II
Locally advanced PDAC
Yoo 2020	Single-centre, retrospective study	199 patients with borderline resectable or locally advanced unresectable PC	Neoadjuvant FFX	Clinical outcomes	For borderline resectable, median PFS was 11.1 (95% CI 8.8-13.5) months, and the median OS was 18.4 (95% CI 16.1-20.8) months. For locally advanced, median PFS was 10.1 (95% CI 8.4-11.8) months, and median OS was 17.1 (95% CI 13.2-20.9) months. Conversion surgery (R0/R1) was achieved in 36.0% of borderline resectable and 29.0% of locally advanced patients.	Yoo et al. 202040	III
Perri 2020	Case series	485 treatment-naïve patients with localised PC	First-line FFX versus GnP	Survival/radiographic/serologic measures of response	FFX was associated with higher rates of RECIST partial response, observed in 19% compared with 6% with GnP, and pancreatectomy was carried out in 27% versus 16%, respectively. OS was similar between regimens (HR 1.48, 95% CI 0.97-2.26, P = 0.07), as were local tumour downstaging rates.	Perri et al. 202041	IV
Napolitano 2019	Single-centre, prospective study	74 patients evaluated with locally advanced PC; 56 patients treated	Neoadjuvant FFX (n = 35) versus GnP (n = 21)	Clinical outcomes	For locally advanced, 14 patients in the FFX group (40%) and 6 in the GnP group (28.6%) underwent surgery. Median DFS was 77.10 weeks in the FFX group and 58.65 weeks in the GnP group (P = 0.625). For unresected patients, median PFS was 49.4 weeks with FFX and 30.9 weeks with GnP (HR 0.345, 95% CI 0.138-0.862, P = 0.0029). Median OS in the unresected population was 72.10 weeks for FFX and 53.30 weeks for GnP (P = 0.06).	Napolitano et al. 201942	III
Suker 2016	Patient-level meta-analysis of 13 studies (1 prospective non-randomised phase II study, 1 prospective cohort study and 11 retrospective cohort studies)	689 patients; 355 had locally advanced PC	Neoadjuvant FFX	Overall survival	For locally advanced treated with FFX, median OS ranged from 10.0 to 32.7 months across studies with a patient-level median OS of 24.2 (95% CI 21.6-26.8) months. Median PFS was 15.0 (95% CI 13.8-16.2) months.	Suker et al. 201643	II
GEST (NCT00498225)	Multicentre (Japan and Taiwan) randomised phase III trial	834 treatment-naïve patients with locally advanced or metastatic PC	GEM versus S-1 versus GEM + S-1	Overall survival	Median OS was 8.8, 9.7 and 10.1 months in the GEM, S-1 and GEM + S-1 groups, respectively. The noninferiority of S-1 to GEM was confirmed (HR 0.96, 97.5% CI 0.78-1.18, P < 0.001 for noninferiority), whereas the superiority of GEM + S-1 over GEM was not demonstrated (HR 0.88, 97.5% CI 0.71-1.08, P = 0.15).	Ueno et al. 201318 and Okusaka et al. 201719	I
Kim 2013 (NCT00658840)	Single-centre, Korean, retrospective phase II study	98 patients with locally advanced PC	CAP versus 5-FU plus radiotherapy	Primary tumour and overall response, toxicity, tumour response and OS	Grade ≥3 toxicities were significantly lower in the capecitabine group than in the 5-FU group (P = 0.045). Primary tumour (30.7% versus 28.2%, P = 0.658) and overall (13.7% versus 15.2%, P = 0.273) response rates and median OS time (12.5 versus 11.6 months, P = 0.655) were similar in the two groups.	Kim et al. 201344	III
Xie 2010	Meta-analysis of 18 prospective RCTs	4237 patients with unresectable locally advanced or metastatic PC	GEM versus GEM-based doublets (GEM/CAP, GEM/CIS, GEM/5-FU, GEM/IRI and GEM/OXA)	Overall survival	Survival benefit observed when GEM was either combined with CAP or OXA. Risk ratios (RRs) for 6-month OS were 0.85 (P = 0.04), 0.99 (P = 0.88), 0.95 (P = 0.46), 1.03 (P = 0.77) and 0.80 (P = 0.001), respectively, for GEM/CAP, GEM/CIS, GEM/5-FU, GEM/IRI and GEM/OXA. A preliminary subgroup analysis in patients with a poor PS (ECOG 2) indicated a worse survival benefit from GEM-based cytotoxic doublets chemotherapy. For poor PS patients, RRs were 1.17 (95% CI 1.01-1.36, P = 0.04) for 6-month OS and 1.09 (95% CI 1.01-1.19, P = 0.04) for 1-year OS.	Xie et al. 201045	I
Boeck 2007	Single-centre, prospective study	39 GEM-pretreated patients with advanced PC	CAP	Objective tumour response rate, carbohydrate antigen 19-9 (CA19-9) tumour marker response, time to progression, OS and toxicity	After a median follow-up of 6.6 months, no complete or partial responses were observed, but 15 patients (39%) had stable disease. A CA19-9 reduction of >20% after two cycles of CAP was documented in six patients (15%). Median time to progression was 2.3 (range 0.5-45.1) months and median OS (since start of CAP treatment) was 7.6 (range 0.7-45.1) months. Predominant grade 2/3 toxicities were hand–foot syndrome 28% (13% grade 3), anaemia 23%, leg oedema 15%, diarrhoea 13%, nausea/vomiting 10% and leukocytopenia 10%.	Boeck et al. 200746	III
Burris 1997	Phase III RCT	126 patients (18 patients with locally advanced PC)	GEM versus 5-FU	Clinical benefit response	GEM was more effective than 5-FU in alleviating some disease symptoms. 23.8% versus 4.8% of GEM- versus 5-FU-treated patients (P = 0.0022) experienced clinical response benefit. The median survival for GEM mono was 5.65 months compared with 4.41 months for 5-FU (P = 0.0025). Survival rate at 12 months was 18% for GEM patients and 2% for 5-FU patients.	Burris et al. 199747	I
Metastatic PDAC
NAPOLI 3 (NCT04083235)	International, open-label, randomised phase III trial	770 patients with metastatic PDAC	First-line NALIRIFOX versus GnP	Overall survival	Median OS was 11.1 (95% CI 10.0-12.1) months versus 9.2 (95% CI 8.3-10.6) months with NALIRIFOX versus GnP, respectively (HR 0.83, 95% CI 0.70-0.99, P = 0.036).	Wainberg et al. 202348	II
JCOG1611-GENERATE	Multicentre, open-label, randomised phase II/III trial	527 patients with ECOG 0-1 and metastatic PC enrolled; 426 patients analysed	GnP, mFFX or S-IROX	Overall survival	Median OS was 17.1 months in the GnP arm, 14.0 months in the mFFX arm (HR 1.31, 95% CI 0.97-1.77) and 13.6 months in the S-IROX arm (HR 1.35, 95% CI 1.00-1.82).	Ohba et al. 202349	II
KCSG (Park 2021)	Multicentre, retrospective study	378 patients with metastatic PC who had progressed following GEM-based therapy	Second-line nal-IRI + 5-FU/LV versus FFX	Clinical efficacy and safety	Similar efficacy was observed in both groups. Median PFS was 3.7 months for nal-IRI versus 4.6 months for FFX (P = 0.44). Median OS was 7.7 months for nal-IRI versus 9.7 months for FFX (P = 0.13).	Park et al. 202150	III
Riedl 2021	Observational cohort study	455 patients with advanced PDAC	First-line GnP versus FFX	Clinical effectiveness	FFX and GnP have similar effectiveness in the palliative first-line advanced PDAC setting. Median, 1- and 2-year OS were 10.1 (95% CI 9.3-11.4) months, 42% (95% CI 37% to 47%) and 18% (95% CI 11% to 19%) in the GnP group, compared with 11.2 months (95% CI not provided), 45% (95% CI not provided) and 12% (95% CI not provided) in the FFX group, respectively (P = 0.783).	Riedl et al. 202151	III
Prager 2021 (NCT02555813)	Multicentre, prospective, non-interventional study	299 metastatic PC patients	First-line GnP	Tolerability and effectiveness in younger (≤70 years) versus elderly (>70 years) patients	OS is similar for younger and elderly patients. Objective response rate was 36% in younger versus 48% in elderly patients, median PFS was 5.6 versus 5.5 months, respectively (HR 1.03, 95% CI 0.8070-1.314, P = 0.81), and OS was 10.6 versus 10.2 months, respectively (HR 0.89, 95% CI 0.6859-1.163, P = 0.4).	Prager et al. 202152	III
Ueno 2020	Multicentre, prospective, open-label, randomised phase II study	79 Japanese patients with metastatic PC following GEM	Second-line nal-IRI plus 5-FU/LV versus 5-FU/LV	PFS	Investigator-assessed median PFS increase with nal-IRI + 5-FU/LV was clinically meaningful and statistically significant versus 5-FU/LV, 2.7 (95% CI 1.5-5.0) months versus 1.5 (95% CI 1.4-1.6) months with HR 0.60. Independently assessed median PFS showed similar trends at 1.7 (95% CI 1.5-3.6) months versus 1.6 (95% CI 1.4-1.6) months, respectively (HR 0.79).	Ueno et al. 202053
Matsumoto 2020	Single-centre, retrospective study	23 Japanese patients with metastatic PC following GEM + GnP failure	Second-line FFX (n = 12) or mFFX (n = 11)	Clinical efficacy and safety	FFX has potential activity in metastatic PDAC patients who have failed GEM + GnP treatment. Median PFS was 5.3 (95% CI 2.5-8.9) months, and the median OS was 12.1 (95% CI 4.0-14.2) months.	Matsumoto et al. 202054	III
Sawada 2020	Single-centre, retrospective study	104 patients with metastatic PC who failed GnP	Second-line mFFX	Efficacy and tolerability of mFFX	mFFX was a tolerable and effective second-line therapy after GnP. Median OS was 7.0 (95% CI 6.2-9.8) months and PFS was 3.9 (95% CI 2.8-5.0) months. The objective response rate was 10.6% and the disease control rate 56.7%.	Sawada et al. 202055	III
Ozaka 2018	Multicentre, prospective Japanese study	69 Japanese chemotherapy-naïve metastatic PC patients	First-line mFFX	OS and incidence of grade ≥3 neutropenia	Median OS was 11.2 (95% CI 9.0-NR) months. Incidence of grade ≥3 neutropenia was 47.8%.	Ozaka et al. 201856	III
NAPOLI 1 (NCT01494506)	International, open-label, randomised phase III trial	417 patients with metastatic PDAC who had received GEM-based therapy	Second-line nal-IRI mono versus 5-FU/LV versus nal-IRI + 5-FU/LV	Overall survival	Median OS was 6.1 (95% CI 4.8-8.9) versus 4.2 (95% CI 3.3-5.3) months in the nal-IRI + 5-FU/LV versus the 5-FU/LV group, respectively (HR 0.67, 95% CI 0.49-0.92, P = 0.012).	Wang-Gillam et al. 201657	II
	Extended follow-up analysis (survival of ≥1 year)				OS for nal-IRI + 5-FU/LV versus 5-FU/LV was maintained, 6.2 (95% CI 4.8-8.4) versus 4.2 (95% CI 3.3-5.3) months, respectively (HR 0.75, 95% CI 0.57-0.99, P = 0.039).	Wang-Gillam et al. 201958	II
GEST (NCT00498225)	Multicentre (Japan and Taiwan) randomised phase III trial	834 treatment-naïve patients with locally advanced or metastatic PC	GEM versus S-1 versus GEM + S-1	Overall survival	Median OS was 8.8, 9.7 and 10.1 months in the GEM, S-1 and GEM + S-1 groups, respectively. The noninferiority of S-1 to GEM was demonstrated (HR 0.96, 97.5% CI 0.78-1.18, P < 0.001 for noninferiority), whereas the superiority of GEM + S-1 was not (HR 0.88, 97.5% CI 0.71-1.08, P = 0.15).	Ueno et al. 201318	II
MPACT (ABI-007) (NCT00844649)	International, multicentre, open-label, randomised phase III trial	861 patients with resectable advanced PC	Neoadjuvant GnP versus GEM	Overall survival	Median OS was significantly longer in the GnP group versus the GEM group, 8.5 versus 6.7 months (HR for death 0.72, 95% CI 0.617-0.835, P < 0.000015).59 An exploratory sub-analysis of 457 patients treated until PD showed a median OS of 9.8 versus 7.5 months (HR 0.69, 95% CI 0.56-0.84, P < 0.001) for the GnP versus GEM group, respectively.60	Von Hoff et al. 201359 Vogel et al. 201660	II
PRODIGE4/ACCORD11 (NCT00112658)	Multicentre, randomised phase III trial	342 patients with metastatic PC with ECOG PS 0-1	First-line FFX versus GEM	Overall survival	FFX was associated with a survival advantage as compared with GEM. Median OS was 11.1 (95% CI 9.0-13.1) months versus 6.8 (95% CI 5.5-7.6) months (HR for death 0.57, 95% CI 0.45-0.73, P < 0.001).	Conroy et al. 201161	I
Burris 1997	Phase III RCT	126 patients (93 patients with metastatic PC)	GEM versus 5-FU	Clinical benefit response	GEM was more effective than 5-FU in alleviating some disease symptoms. 23.8% versus 4.8% of GEM- versus 5-FU-treated patients (P = 0.0022) experienced clinical response benefit. The median survival for GEM mono was 5.65 months compared with 4.41 months for 5-FU (P = 0.0025). Survival rate at 12 months was 18% for GEM patients and 2% for 5-FU patients.	Burris et al. 199747	I
Molecular and/or genetic testing in PDAC
O’Reilly 2020	Multicentre, phase II RCT	50 patients with untreated gBRCA/PALB2+ PDAC with measurable stage III-IV disease and ECOG PS of 0-1	CIS, GEM and veliparib versus CIS and GEM	Overall response rate	The response rate was 74.1% for CIS plus GEM plus veliparib and 65.2% for CIS plus GEM (P = 0.55); both groups exceeded the prespecified activity threshold. Concurrent veliparib did not improve the response rate.	O’Reilly et al. 202062	II
KEYNOTE-158 (NCT02628067)	Non-randomised, open-label phase II trial	233 patients with non-colorectal cancer and high MSI/MMR-deficient cancer, including 22 patients with PC	Pembrolizumab	Objective response rate	MSI/MMR-deficient cancers, including PC, are sensitive to immune checkpoint blockade with anti-programmed death-1 therapy. Objective response rate for pancreatic cancer was 18.2% (95% CI 5.2% to 40.3%).	Marabelle et al. 202063	II
POLO (NCT02184195)	International, double-blind, placebo-controlled phase III RCT	154 randomised metastatic PC patients with germline BRCA1/2 mutation	Olaparib versus placebo	PFS	Median PFS was significantly longer in the olaparib group than in the placebo group: 7.4 versus 3.8 months. HR for PD or death was 0.53 (95% CI 0.35-0.82, P = 0.004).	Golan et al. 201964	I

5-FU, 5-fluorouracil; 5-FU/LV, fluorouracil and folinic acid; AEs, adverse events; CAP, capecitabine; CI, confidence interval; CIS, cisplatin; DFS, disease-free survival; EORTC, European Organisation for Research and Treatment of Cancer; ESPAC1, European Study Group for Pancreatic Cancer 1; FFCD, Federation Francophone de Cancérologie Digestive; FFX, FOLFIRINOX; FOLFIRINOX, fluorouracil–leucovorin–irinotecan–oxaliplatin; GEM, gemcitabine; GEMCAP, GEM plus CAP; GERCOR, Groupe Coopérateur Multidisciplinaire en Oncologie; GnP, GEM plus nab-paclitaxel; HR, hazard ratio; ID, study identification; ISRCTN, International Standard Randomised Controlled Trial Number; IRI, irinotecan; ITT, intent to treat; JSAP, Japanese Study Group of Adjuvant Therapy for Pancreatic Cancer; mFFX, modified FFX; MMR, mismatch repair; mono, monotherapy; MSI, microsatellite instability; nab, albumin-bound; nal-IRI, nanoliposomal irinotecan; NCT, ClinicalTrials.gov identifier; NR, not reached; OS, overall survival; OXA, oxaliplatin; PC, pancreatic cancer; PD, disease progression; PDAC, pancreatic ductal adenocarcinoma; PFS, progression-free survival; PREP, Study Group of Preoperative Therapy for Pancreatic Cancer; PRODIGE 24-ACCORD 24 and CCTG PG 6, Actions Concertées dans les Cancers Colorectaux et Digestifs Canadian Cancer Trials Group Pancreatic Adenocarcinoma; RCT, randomised controlled trial; RR, risk ratio; RT, radiotherapy; S-IROX, S-1, irinotecan, and oxaliplatin; UMIN, Japanese University Hospital Medical Information Network.

In the Delphi process’s first round, 6 out of 18 statements lacked consensus. After revision, a second round involving 11 statements led to 14 consensus recommendations (R1-R14), categorised by disease status: resectable/localised, borderline resectable, locally advanced, metastatic and other considerations, all reaching over 75% agreement.

Recommendations by disease status

Resectable/localised

Three recommendations (R1-R3, Table 2) on the optimal use of systemic therapies in resectable/localised PDAC reached 80%, 95% and 100% consensus, respectively. Seven randomised clinical trials and two retrospective real-world studies are included as evidence for three recommendations in this setting (Table 1). In the adjuvant setting, evidence from one phase II³¹ and one phase III^34,35 randomised clinical trials evaluating adjuvant chemoradiotherapy versus chemotherapy alone is conflicted. Although these pivotal trials were published more than a decade ago, randomised evidence is still lacking surrounding the benefit of adjuvant chemoradiotherapy, and existing guideline recommendations for chemoradiotherapy differ significantly. Given this, the expert panel concurs that chemotherapy should remain the optimal treatment approach, aiming to reduce recurrence following surgery. In close consultation with the patient, fit [Eastern Cooperative Oncology Group (ECOG) performance status (PS) 0-1] patients can be aggressively treated with the modified FOLFIRINOX (mFFX) regimen (preferred option; R1). This recommendation is based on the results of the PRODIGE-24/ACCORD-24/CCTG-PA-6 study, which demonstrated significantly greater overall survival (OS) for FFX versus gemcitabine (GEM) alone (Table 1).²⁸ Alternatively, depending on the treating physician’s judgement, balancing efficacy with tolerability, combination therapy [GEM with capecitabine (CAP)²⁹ or GEM with albumin-bound (nab)-paclitaxel (GnP)]²⁵ or monotherapy (S-1^27,30 or GEM^32,33) can be considered (R2), which have been evaluated in phase III randomised trials ESPAC-4, APACT, JASPAC 01 and CONKO-001, respectively (Table 1). In the ESPAC-4 study, adjuvant GEM with CAP significantly prolonged OS when compared with GEM alone [28.0 versus 25.5 months; hazard ratio (HR) 0.82, 95% confidence interval (CI) 0.68-0.98, P = 0.032; Table 1].²⁹ Although the primary endpoint of disease-free survival in the APACT trial was not met, the OS (secondary endpoint) at the time of the primary analysis trended in favour of adjuvant GnP versus GEM alone.²⁵ S-1 monotherapy is a GEM-sparing option that was shown to be significantly more effective in an Asian population, extending OS versus GEM alone (46.5 versus 25.5 months; HR 0.57, 95% CI 0.44-0.72, P < 0.0001 for superiority; Table 1).³⁰ For patients with poor PS, GEM combined with CAP^26,29 and monotherapy with GEM^32,33 or S-1^27,30 can be considered based on the clinical data from ESPAC-4 and CONKO-001, respectively, and real-world evidence for GEM plus CAP, as highlighted in Table 1 (R1).

Table 2.

Consensus recommendations on the optimal use of systemic therapies in resectable/localised PDAC

Recommendations		% Consensus	Level of evidence
Adjuvant treatment with systemic therapy is recommended following R0/R1 surgical resection. Adjuvant therapy should be administered for a maximum of 6 months:
R1	For patients with ECOG PS 0-1, (m)FOLFIRINOX is the preferred treatment. Alternatives include GEM with CAP, S-1 monotherapy, GEM monotherapy or GnP depending on the treating physician’s clinical judgement.	80	IA
R2	For patients with ECOG PS ≥2, GEM with CAP and GEM or S-1 monotherapy can be considered depending on the treating physician’s clinical judgement, balancing efficacy and tolerability. Note: For elderly or frail patients, less intensive options such as GEM or S-1 monotherapy may be preferred.	95	IA
R3	To detect recurrence and monitor overall well-being after adjuvant therapy, patients should be monitored every 3 months for up to 2 years and every 6 months thereafter for up to 5 years post-therapy using CT or other imaging techniques.	100	VA

(m), modified dose; CAP, capecitabine; CT, computed tomography; ECOG, Eastern Cooperative Oncology Group; GEM, gemcitabine; GnP, GEM plus nab-paclitaxel; PDAC, pancreatic ductal adenocarcinoma; PS, performance status; R0, resection for cure or complete remission; R1, microscopic residual tumour.

To our knowledge, no evidence-based guidelines for patient monitoring following adjuvant therapy exist. Based on our clinical experience, we recommend monitoring patients for signs of recurrence using computed tomography or other appropriate imaging techniques every 3 months for up to 2 years and then every 6 months for at least 5 years (R3). Treating physicians may consider other diagnostic tests useful to augment imaging results and monitor overall health. These follow-up visits can also be used to assess general well-being and quality of life (QoL).

Borderline resectable PDAC

Three consensus recommendations (R4-R6, Table 3) on the optimal use of systemic therapies in borderline resectable PDAC reached 85%, 100% and 85% consensus, respectively. There is limited evidence for systemic therapies in borderline resectable PDAC patients from randomised controlled trials (RCTs). Two open-label phase II/III randomised trials, one single-arm phase II trial, two meta-analyses of predominantly prospective/retrospective studies and one observational single-centre study support the expert consensus in this setting (Table 1). In patients with borderline resectable PDAC, neoadjuvant systemic therapy is recommended to increase the possibility of R0 resection and treat any micrometastases (R4). When fully considering the patient’s condition, a few agents used in the metastatic setting can be considered for neoadjuvant therapy, namely FFX³⁸ and GEM-based (with either nab-paclitaxel or S-1³⁷) regimens (R4a). This latter recommendation is based on the neoadjuvant treatment strategy used in the Prep-02/JSAP05 phase II/III trial^37,65 and the phase II trial reported by Motoi et al. (2013).³⁹ However, as shown in Table 1, a paucity of evidence, specifically high-quality randomised trials, is available in this setting. In the meta-analysis by Cloyd et al. (2020), the likelihood of R0 resection was increased with neoadjuvant therapy versus upfront surgery [risk ratio (RR) 1.51, 95% CI 1.18-1.93, I² = 0%].³⁶ The meta-analysis also showed an OS benefit of neoadjuvant treatment compared with surgery alone (pooled HR 0.73, 95% CI 0.61-0.86, I² = 0%), with a median OS of 25.4 months (95% CI 22.4-28.7 months) versus 19.4 months (95% CI 17.2-21.8 months, P < 0.001), respectively.³⁶ This finding was independent of disease stage (potentially resectable HR 0.73, 95% CI 0.59-0.91; borderline resectable HR 0.51, 95% CI 0.28-0.93) and treatment type (chemoradiotherapy HR 0.77, 95% CI 0.61-0.98; chemotherapy alone HR 0.68, 95% CI 0.54-0.87).³⁶ Similarly, in the meta-analysis by Versteijne et al. (2018), R0 resection rate was significantly greater with neoadjuvant treatment versus surgery alone (86.8% versus 66.9%, 95% CI 84.6% to 88.7% and 64.2% to 69.6%, respectively, P < 0.001).¹⁶

Table 3.

Consensus recommendations on the optimal use of systemic therapies in borderline resectable PDAC

Recommendations		% Consensus	Level of evidence
R4	Neoadjuvant treatment with systemic therapy is recommended for patients with borderline resectable disease.	85	IIB
a.	Recommended neoadjuvant therapies are those used for metastatic PDAC and associated with higher response in these patients; specifically, GnP and FOLFIRINOX should be considered. Other GEM-based combination regimens, such as GEM + S-1, can also be considered.
b.	GEM + S-1 should continue for a maximum of two cycles (21-day cycle is repeated for two cycles) or until disease progression, intractable adverse events, patient refusal, etc.
c.	Due to the heterogeneity in the definitions of disease extent and the patient inclusion and response evaluation criteria used in clinical trials, there is a need for procedural standardisation in resection criteria, radiological and pathological reports, the chemotherapeutic regimen and the radiation protocol.
R5	A multidisciplinary team should reassess patients for surgical intervention following neoadjuvant treatment with re-imaging, restaging and assessment.	100	VA
R6	For patients having undergone resection following neoadjuvant therapy, adjuvant (post-operative) treatment depends on various factors, including the stage, the patient’s overall health and the specific treatments used during neoadjuvant therapy: •Perioperative chemotherapy for an intended duration of 6 months may be considered, with the regimen individualised. •Adjuvant chemotherapy with an intended duration of 6 months may be considered, with the regimen individualised.	85	VA

FOLFIRINOX, fluorouracil–leucovorin–irinotecan–oxaliplatin; GEM, gemcitabine; GnP, GEM plus nab-paclitaxel; PDAC, pancreatic ductal adenocarcinoma; S-1, oral 5-fluorouracil prodrug.

If GEM with S-1 is used in the neoadjuvant setting, treatment should proceed for a maximum of two cycles (21-day cycle is repeated for two cycles; R4b).³⁷ It should be stopped in the event of disease progression, intractable adverse events (AEs) or patient refusal to continue with treatment (R4b). The steering committee acknowledges the heterogeneity in the definitions of disease extent and the patient inclusion and response evaluation criteria used in clinical trials, so recommendation R4c was included based on the 2021 Korean clinical practice guidelines¹³ to highlight the need for procedural standardisation.

Based on our clinical experience, we recommend (R5) that borderline resectable patients should be reassessed by a multidisciplinary team for surgical intervention following neoadjuvant treatment with re-imaging and restaging. In addition, we recommend (R6) individualised adjuvant (peri- and post-operative) chemotherapy for ∼6 months for patients having undergone resection following neoadjuvant chemotherapy.

Locally advanced PDAC

Two consensus recommendations (R7 and R8, Table 4) on the optimal use of systemic therapies in localised advanced PDAC both reached 95% consensus. Based on expert opinion, for conversion/downstaging therapy, the treatment strategy should be based on each institution’s multidisciplinary tumour board discussion (R7).^66,67 Notably, there is limited randomised clinical evidence for systemic therapies in this setting. Two phase III randomised trials, four single-centre studies (two prospective, two retrospective), two meta-analyses of predominantly prospective/retrospective studies and one case series support the expert consensus in localised advanced PDAC (Table 1).

Table 4.

Consensus recommendations on the optimal use of systemic therapies in locally advanced PDAC

Recommendation		% Consensus	Level of evidence
R7	For conversion/downstaging therapy, the treatment strategy should be based on each institution’s multidisciplinary tumour board discussion:	95	VA
a.	Resection of the primary lesion after multidisciplinary treatment is one of the treatment options for patients with locally advanced pancreatic cancer in whom the disease was judged as being unresectable at the first examination because good survival or recurrence-free survival may be obtained in patients in whom the treatment is successful, and the cancer becomes resectable.
R8	Treatment of locally advanced disease with systemic therapy is recommended. Recommendations depend on the patient’s performance status:	95
a.	For patients with ECOG PS 0-1, FOLFIRINOX and GnP are preferred.		IIIB
b.	For patients with ECOG PS 2, single-agent chemotherapy with GEM, S-1 or CAP is preferred.		IB
c.	For some selected patients with ECOG PS 2, doublet chemotherapy may be considered.		IIB

CAP, capecitabine; ECOG, Eastern Cooperative Oncology Group; FOLFIRINOX, fluorouracil–leucovorin–irinotecan–oxaliplatin; GEM, gemcitabine; GnP, GEM plus nab-paclitaxel; PDAC, pancreatic ductal adenocarcinoma; PS, performance status; S-1, oral 5-fluorouracil prodrug.

Systemic therapy is the main intervention for patients with locally advanced PDAC not suitable for conversion/downstaging. The aim of systemic therapy in this setting is to prolong life through a reduction in tumour burden. When selecting systemic therapy, the ECOG PS of each patient should be considered (R8a-c). In patients with ECOG PS 0-1, combination therapy with either FFX or GnP is recommended (R8a). This recommendation is largely based on data from a multicentre phase III trial in the adjuvant/neoadjuvant setting for resectable PDAC, which showed improved OS and progression-free survival (PFS) for these regimens.²⁸ Data specific to locally advanced PDAC include the meta-analysis by Suker et al. (2016), which demonstrated a patient-level median OS of 24.2 months (95% CI 21.6-26.8 months) for FFX-treated patients (Table 1).⁴³ The 1-year OS rate for these patients was 80.0% (95% CI 74.7% to 84.4%), and the 2-year OS was 50.2% (95% CI 42.9% to 57.5%) with a median PFS of 15.0 months (95% Cl 13.8-16.2 months).⁴³ One single-centre prospective study (Napolitano et al., 2019)⁴² and one single-centre retrospective study (Yoo et al., 2020)⁴⁰ evaluated adjuvant FFX versus GnP and FFX, respectively, as first-line treatments in unresected locally advanced PDAC. The latest NCCN 2024 guidelines state that while there is high-level evidence supporting the use of liposomal irinotecan/5-fluorouracil (5-FU)/leucovorin (LV)/oxaliplatin (NALIRIFOX) over GnP in the locally advanced disease setting,⁶⁸ this regimen does not appear to have an advantage over FFX and is considerably more expensive compared with FFX.⁴

There is no strong evidence about optimal treatment for patients with non-metastatic locally advanced PDAC. Based on a pivotal randomised trial reported over a decade ago,⁴⁷ patients with an ECOG PS 2 can be treated using GEM monotherapy (R8b). In treatment-naïve patients with metastatic PDAC, GEM showed a modest modified OS benefit versus 5-FU (5.65 versus 4.41 months, P = 0.0025) (Table 1) and a 12-month survival rate of 18% and 2%, respectively. Clinical benefit response was experienced by 23.8% of GEM-treated patients and 4.8% of 5-FU-treated patients (P = 0.0022; Table 1).⁴⁷ In Japanese and Taiwanese patients with locally advanced or metastatic pancreatic cancer, monotherapy with S-1 demonstrated noninferiority to GEM in OS in a phase III study (Table 1).¹⁸ S-1 monotherapy is recommended as one of the standard first-line treatment options for locally advanced pancreatic cancer in the 2022 Japan guidelines,¹⁰ albeit a weak recommendation, based on the results of a phase III study (GEST study) conducted in Japan and Taiwan.^18,19 CAP may also be considered in this setting (R8b).^44,46 In a phase II, single-centre, retrospective Korean study of concurrent chemoradiotherapy with CAP in unresectable locally advanced pancreatic cancer patients, CAP plus radiotherapy was shown to have similar efficacy and low rates of toxicities compared with 5-FU plus radiotherapy.⁴⁴ A prospective single-centre study showed that CAP monotherapy may be a possible treatment option for GEM-pretreated patients with advanced pancreatic cancer.⁴⁶ A meta-analysis of 18 prospective RCTs compared GEM with GEM-based doublets chemotherapy in 4237 advanced PDAC patients with a good PS versus poor PS (ECOG PS 2).⁴⁵ Overall, results showed that first-line GEM-based doublets chemotherapy had a substantial OS advantage compared with GEM monotherapy in advanced pancreatic cancer.⁴⁵ However, in the poor PS subgroup, patients with a poor PS appeared to have a worse survival benefit from GEM-based cytotoxic doublets chemotherapy (RR 1.17, 95% CI 1.01-1.36, P = 0.04 for 6-month OS, and RR 1.09, 95% CI 1.01-1.19, P = 0.04 for 1-year OS, respectively).⁴⁵ Since GEM-based cytotoxic doublets chemotherapy increased the risk of death in advanced PDAC patients with a poor PS in this study, they should be reserved for selected patients only (R8c).⁴⁵

Metastatic PDAC

Four consensus recommendations (R9-R12, Table 5) on the optimal use of systemic therapies in metastatic PDAC reached 90%, 100%, 95% and 90% consensus, respectively. Eight phase II/III randomised trials, of which five were open-label, two multicentre prospective studies, three retrospective studies and one observational cohort study support the expert consensus in metastatic PDAC (Table 1).

Table 5.

Consensus recommendations on the optimal use of systemic therapies in metastatic PDAC

Recommendations		% Consensus	Level of evidence
R9	Treatment of metastatic disease with first-line systemic therapy is recommended. Recommendations depend on the patient’s performance status:	90	IIIA
a.	First-line combination therapy is recommended for patients with ECOG PS 0-1 and amenability (patient and physician) to treatment; GnP or NALIRIFOX; (m)FOLFIRINOX may be considered as an alternative unless toxicity is unsuitable, in which case gemcitabine-based therapies may be prioritised (mFFX can be considered as a second-line treatment in such scenarios).
b.	For patients with ECOG PS ≥2, monotherapy with GEM (preferred), GnP (ECOG PS 2) or S-1 (rarely) can be considered.
c.	Best supportive care should be considered for patients with ECOG PS ≥2.
R10	In patients who have progressed on first-line therapy, treatment of metastatic disease with second-line systemic therapy is recommended:	100	IIIB
a.	Patients should be exposed to both GEM-based and 5-FU-based regimens across their treatment scheme.
b.	Recommendations depend on the patient’s performance status and the prevailing level of evidence for second-line therapy.
c.	Best supportive care interventions include monitoring bilirubin levels, early detection and management of gastric outlet obstruction, decompression of bile duct obstruction, pain control, nutritional support, management of depression and pancreatic enzyme replacement therapy.
R11	Treatment strategy should be considered before commencing therapy for metastatic disease. When agreeing on a treatment strategy with the patient, factors including efficacy, adverse event profile, age and amenability of the patient to aggressive treatment should be considered, along with any prior (neo)adjuvant therapy. The following should be considered:	95	IIIB
a.	mFOLFIRINOX is generally associated with a less favourable adverse event profile than GnP. Therefore, it should be reserved for patients with ECOG PS 0-1 and amenability to aggressive treatment.
b.	The patient’s age should also be considered, as GEM has been shown to be effective in patients older than 70 years.
c.	The use of GnP or other GEM-based regimen in the first line allows for the use of a second-line option that has been assessed in an RCT (nal-IRI + 5-FU/LV).
R12	Active intervention best supportive care should be pursued. Some patients may be converted to a better ECOG PS:	90	VB
a.	Best supportive care interventions include monitoring bilirubin levels, early detection and management of gastric outlet obstruction, decompression of bile duct obstruction, pain control, nutritional support, management of depression and pancreatic enzyme replacement therapy.

5-FU, 5-fluorouracil; ECOG, Eastern Cooperative Oncology Group; FOLFIRINOX, fluorouracil–leucovorin–irinotecan–oxaliplatin; GEM, gemcitabine; GnP, GEM plus nab-paclitaxel; mFOLFIRINOX, modified FOLFIRINOX; PDAC, pancreatic ductal adenocarcinoma; PS, performance status; S-1, oral 5-fluorouracil prodrug.

Systemic therapy is the mainstay of treatment for metastatic PDAC, with the aim of prolonging life. This treatment also frequently provides significant palliative benefits, particularly in terms of reducing disease-related symptoms such as pain, fatigue and gastrointestinal discomfort, thereby improving the patient’s QoL. When deciding on a therapeutic approach, the patient’s PS, amenability to aggressive treatment and potential for treatment sequencing should be considered (R9-R12). Based on the evidence shown in Table 1 as well as expert experience, we recommend either mFFX, GnP or NALIRIFOX as first-line options for fit patients (ECOG PS 0-1) with metastatic disease (R9a). In the PRODIGE 4/ACCORD 11 RCT, FFX demonstrated improved OS compared with GEM alone (HR 0.57, 95% CI 0.45-0.73, P < 0.001; Table 1).⁶¹ This result was mirrored in the MPACT study wherein combination therapy with GnP demonstrated an OS benefit versus GEM alone (HR 0.72, 95% CI 0.62-0.83, P < 0.001; Table 1).^59,60 A multicentre, prospective study of 299 metastatic pancreatic cancer patients treated with GnP showed similar tolerability and effectiveness in younger (≤70 years) versus older patients (Table 1).⁵² Conditions in clinical practice can differ from those in clinical trials as some patients with PDAC have contraindications to administration of the optimal therapy. The first prospective study of mFFX in Asian chemotherapy-naïve patients with metastatic pancreatic cancer showed that mFFX improved safety profile and maintained efficacy in this setting.⁵⁶ Based on the results of this phase II trial, FFX was changed to mFFX in Japanese clinical practice. Moreover, in another retrospective Japanese cohort of 23 patients, mFFX showed similar efficacy to FFX in patients with metastatic PDAC after GnP failure.⁵⁴ Similarly, a retrospective study reported by Sawada et al. (2020) showed that mFFX was a tolerable second-line treatment for metastatic pancreatic cancer patients after GnP failure, with a median OS of 7.0 months (95% CI 6.2-9.8 months) and PFS of 3.9 months (95% CI 2.8-5.0 months).⁵⁵ To clarify, mFFX is recommended as a first-line treatment option for fit patients (ECOG PS 0-1) with metastatic disease, given its demonstrated efficacy and improved safety profile compared with FFX. However, in cases where its toxicity profile limits its suitability, alternative regimens, such as GEM-based therapies, may be prioritised as first-line options, with mFFX considered as a second-line treatment in such scenarios. Recently, based on a planned interim analysis, first-line GnP was reported to be superior to mFFX and S-IROX (S-1, irinotecan and oxaliplatin) in a Japanese phase III trial (JCOG1611-GENERATE) evaluating an initial 527 metastatic or recurrent pancreatic cancer patients with good PS.⁴⁹ Therefore, we recommend mFFX rather than FFX as a first-line option for fit Asian patients with metastatic disease. In the absence of a direct comparison, we cannot recommend a preferred regimen in the first-line metastatic setting. However, an observational cohort study by Riedl et al. (2021) in 455 advanced PDAC patients demonstrated FFX and GnP to have similar effectiveness in the palliative first-line setting.⁵¹ Furthermore, recent data from the NAPOLI-3 study showed improved OS with the NALIRIFOX regimen versus GnP, with a median OS of 11.1 months (95% CI 10.0-12.1 months) and 9.2 months (95% CI 8.3-10.6 months), respectively (HR 0.83, 95% CI 0.70-0.99, P = 0.036).⁴⁸ PFS was also significantly longer with NALIRIFOX (7.4 months, 95% CI 6.0-7.7 months) versus GnP (5.6 months, 95% CI 5.3-5.8 months) (HR 0.69, 95% CI 0.58-0.83, P < 0.0001).⁴⁸ We would, therefore, advise that NALIRIFOX is considered over GnP in the treatment schema. For patients with metastatic disease and an ECOG PS ≥2, GEM is recommended due to an improved median survival for GEM monotherapy versus 5-FU (5.65 versus 4.41 months, respectively) (R9b).⁴⁷ For patients with advanced PDAC and an ECOG PS of 2, GnP can be considered based on improved survival versus GEM in a phase III study reported by Von Hoff et al. (2013).⁵⁹ In a randomised phase III study, S-1 demonstrated noninferiority to GEM in OS with good tolerability and is therefore recommended as a possible option for locally advanced and metastatic PDACs in chemotherapy-naïve patients (R9b).¹⁸

Notably, in existing guidelines from Asia and Western countries, there is no clear consensus on the recommended second-line treatment after failure of GnP therapy for metastatic pancreatic cancer. We recommend that patients who have progressed on first line are considered for second-line therapy (R10). In this setting, we recommend physicians consider previous treatments in all (neo)adjuvant and first-line settings. Patients previously exposed to a GEM-based regimen should receive a 5-FU-based regimen; otherwise, patients previously exposed to 5-FU should receive a GEM-based regimen (R10a). When selecting a second-line therapy, the patient’s PS should be considered with any additional evidence available at the time of treatment (R10b). This recommendation is based on evidence of 5-FU-based regimens in the second-line metastatic setting from the phase III NAPOLI 1 study,^57,58 as well as from a prospective, open-label, randomised, multicentre phase II study in Japanese patients who progressed or recurred following prior GEM-based therapy,⁵³ as well as retrospective data from the multicentre Korean Cancer Study Group (KCSG).⁵⁰ The global, phase III NAPOLI 1 study demonstrated that the combination of nanoliposomal irinotecan (nal-IRI) with 5-FU and folinic acid significantly extended survival rates for patients with metastatic PDAC who had previously been treated with GEM-based therapy.⁵⁷ NAPOLI 1 found that patients treated with both nal-IRI and 5-FU and folinic acid had a median OS rate of 6.1 months, compared with 4.2 months for those treated with only 5-FU and folinic acid.⁵⁷ The results suggest that the combination of nal-IRI plus 5-FU and folinic acid may offer a new treatment option for these patients due to its ability to prolong survival with manageable side-effects.⁵⁷ Notably, the survival advantage of nal-IRI plus 5-FU/LV versus 5-FU/LV was maintained over 12 months in the extension study of NAPOLI 1.⁵⁸ Similarly to participants in the NAPOLI 1 study, clinically meaningful and statistically significant gains in investigator-assessed PFS and objective response rate were observed with the combination regimen of nal-IRI plus 5-FU/LV versus 5-FU/LV in Japanese patients, with no new or unexpected safety signals.⁵³

With an expanding therapeutic landscape, the treatment sequencing strategy for metastatic disease should be considered upfront, with factors for consideration including efficacy, AE profile, age and patient amenability (R11). While real-world data suggest similar effectiveness,⁵¹ the steering committee is of the opinion that the AE profile of FFX is generally less favourable than that of GnP, particularly with respect to haematological AEs (R11a). This regimen should be reserved for patients with ECOG PS 0-1, which is in line with the entry criteria for the PRODIGE 4/ACCORD 11 study (R11a).⁶¹ Additionally, while the PRODIGE 4/ACCORD 11 study enrolled only patients <76 years of age, the MPACT study enrolled patients up to 86 years of age in the GEM with nab-paclitaxel arm. The results of MPACT (discussed earlier) are supported by real-world evidence suggesting comparable effectiveness between patients ≤70 years and >70 years of age (R11b).⁵² Lastly, using a GEM-based regimen as first-line treatment allows for the use of the RCT-assessed combination of nal-IRI with 5-FU/LV as second line, as it is indicated in GEM-refractory patients (R11c).

We recommend active intervention with best supportive care, listing several interventions that may result in conversion to a better PS. In such cases, more aggressive treatment or additional lines of therapy may be warranted (R12). These interventions may also improve patient comfort and QoL.

Other considerations

One recommendation (R13, Table 6) is provided for clinical molecular testing and one for genetic testing in PDAC (R14, Table 6), which reached 85% and 90% consensus, respectively. One double-blind, placebo-controlled, phase III randomised trial, one multicentre phase II randomised trial and one open-label phase II study support the expert consensus recommendations for molecular and genetic testing (Table 1).

Table 6.

Consensus recommendations regarding other considerations in metastatic PDAC

Recommendations		% Consensus	Level of evidence
R13	Clinical molecular testing should be considered for patients with locally advanced or metastatic disease. PDAC-specific targets include: •MSI/MMR •Germline BRCA •Tumour-agnostic genes: BRAF, RET, NTRK, TMB-H	85	IC
R14	Clinical genomic testing should be considered for patients with locally advanced or metastatic disease:	90
a.	MSI/MMR testing by NGS should be considered for patients with locally advanced or metastatic disease.
b.	Note: Treatment with immune checkpoint inhibitors is appropriate for patients with high MSI or abnormalities in MMR genes.

BRAF, v-Raf murine sarcoma viral oncogene homolog B; BRCA, breast cancer gene; MSI/MMR, microsatellite instability/mismatch repair; NGS, next-generation sequencing; NTRK, neurotrophic tyrosine receptor kinase; PDAC, pancreatic ductal adenocarcinoma; RET, rearranged during transfection; TMB-H, tumour mutational burden—high.

Molecular testing should be considered in locally advanced or metastatic PDAC, as it may provide additional information relevant to treatment selection (R13). For instance, in patients with BRCA1/2 mutations, the combination of GEM and cisplatin was shown as an effective regimen in an open-label, randomised phase II study and is a recommended option in the NCCN 2024 guidelines.^4,62 In the phase III Pancreas Cancer Olaparib Ongoing (POLO) trial, maintenance therapy with olaparib in patients with a germline BRCA mutation and metastatic PDAC that had not progressed during first-line platinum-based chemotherapy led to significant PFS benefit but no difference in terms of OS.⁶⁴ Genomic testing using next-generation sequencing should also be considered in these settings (R14a). Patients with microsatellite instability-high or mismatch repair-deficient tumours may benefit from treatment with immune checkpoint inhibitors, such as pembrolizumab (R14b).^4,63 Low prevalence of actionable mutations limits the wide use of genetic testing. The availability of targeted therapies for specific mutations may vary by region, and their effectiveness can depend on the individual patient’s overall health and the extent of the disease.

Our study has limitations, including a non-systematically selected Delphi panel possibly causing bias, and a narrative review based on selective research that lacks systematic review rigour. The 75% consensus threshold used is arbitrary, posing another limitation. Despite these issues, the modified Delphi method effectively identifies best practices for PDAC treatment in Asia, offering valuable insights but not replacing clinical judgement. This consensus statement aims to enhance, rather than replace, international/local guidelines.

Conclusions

In summary, this is the first expert consensus statement to provide comprehensive and rigorous recommendations on using systematic therapies for PDAC in Asia. Implementing this consensus has important implications as it will provide harmonised treatment strategies in patients with PDAC, thus improving patient care and enabling better comparison of new trial findings and standardising routine clinical practice.