Pancreatic cancer, the majority of which is pancreatic ductal adenocarcinoma (PDAC), accounted for almost as many new deaths (~466,000) as new cases (~496,000) in 2020 due to its poor prognosis and it is the 7^th leading cause of cancer mortality in both sexes (1). Advanced PDACs (APACs) with involvement of major arteries (e.g., celiac trunk and superior mesenteric artery) or with metastasis to distant sites, which we previously found to comprise ~64-81% of all PDACs in Europe and the US (2), are often deemed unresectable. Chemotherapy remains the major treatment modality for APAC, and can significantly prolong survival and improve quality of life (3). The prognosis for APACs is extremely poor, the 3-year overall survival (OS) being 2-5% for total patients <60 years and 1-2% for those ≥60 years in Europe and the US, where only slight improvements in survival were observed over time (4).

Growing evidence supports that maintenance therapy offers clinical benefits in pancreatic, colorectal, and lung cancers (5–7). Although the POLO study demonstrated that progression-free survival (PFS) was longer with olaparib maintenance than with placebo (7.4 vs. 3.8 months) among patients with metastatic pancreatic cancer and a germline BRCA mutation, PARP inhibitors were less commonly used than chemotherapy agents in the maintenance setting due to the low proportion of germline BRCA mutation (8). A phase II study showed that induction gemcitabine and S−1 (GS) followed by chemoradiotherapy and S−1 maintenance for locally-advanced disease was feasible, well-tolerated, and highly active, with median survival of 22.9 months and 3-year survival rate of 30.4% (9). Maintenance with capecitabine for metastatic disease treated with first-line FOLFIRINOX was also effective, with median OS (mOS) of 17 months and first PFS of 5 months (10).

Among various chemotherapy regimens for APAC, nano-albumin-bound (nab)-paclitaxel plus S-1 (NPS) is effective under careful toxicology surveillance. As first-line chemotherapy for metastatic disease, it significantly prolonged median time to progression (7.1 vs. 3.6 months) and OS (10.2 vs. 6.0 months) compared to GS in a retrospective study (11). Our previous phase II trial (clinicaltrials.gov registration number: NCT02124317) supported that NPS as first-line treatment with S-1 as optional subsequent maintenance therapy had encouraging objective response rate (ORR; 50.0%) and survival (mOS, 9.4 months; median PFS [mPFS], 5.6 months) and manageable toxicities for APAC, suggesting this strategy as an effective alternative treatment regimen (12).

In this study, we described the patient and tumor characteristics, tumor responses, and safety profiles in APAC patients treated with induction NPS therapy and subsequent S-1 maintenance. We further described the survival and explored the prognostic factors among patients receiving S-1 maintenance both overall and in various stratifications. Our findings suggested that for APAC patients the NPS therapy may not need to be administered until disease progression and may be replaced by S-1 maintenance after objective response or stabilization of disease, which is effective with low toxicity.

PDAC is an aggressive and devastating disease. The majority of PDAC patients have unresectable, locally advanced, or metastatic disease at diagnosis and the survival is extremely poor (15, 16). The main reason for the refractoriness of local primary PDAC is that pancreatic cancer cells are surrounded by an intense desmoplastic reaction which may create a barrier to the drugs penetrating into the tumor and the stroma has been addressed as a potential therapeutic target. The role of the stroma proportion might explain the resistance of majorly local primary PDAC to chemotherapy (17) but not that of metastatic disease. Nevertheless, the control of disease at the primary site might help to relieve the metastatic burden. Nab-paclitaxel is an innovative molecule which depletes tumor stroma through interaction between albumin and secreted acidic protein which is rich in cysteine. In 2013, the Food and Drug Administration (FDA) approved nab-paclitaxel plus gemcitabine as the standard first-line chemotherapy for metastatic pancreatic cancer based on the MPACT trial (18). However, several common and severe adverse events, such as neutropenia, thrombocytopenia, and peripheral neurotoxicity, affected the compliance and clinical application of this regimen. S-1, an oral fluoropyrimidine derivative, has been shown as a very promising treatment agent with good tolerability for PDAC (19). Addition of nab-paclitaxel to S-1 as first-line treatment has shown encouraging activity and efficacy in APAC with promising ORR and survival in our previous NASPAC study (12), which was considered as one of the standards of care in APAC therapy (20–22).

For APAC where cure is scarce, the main aim of management should be prolongation of survival and improvement of quality of life, through inhibition of cancer progression without severely compromising physical status. Although the strategy of continuing intensive chemotherapy until disease progression or for 6 months has been shown to improve the survival of metastatic PDAC based on two famous phase III trials (18, 23), recently increasing evidence seems to favor the maintenance strategy for APAC (8, 24–26). Following our previous phase II trial on NPS for APAC where part of the patients further received S-1 maintenance and showed good prognosis (12), another single-arm phase II trial also showed that NPS regimen as first-line therapy followed by S-1 maintenance for APAC presented encouraging ORR and survival and manageable toxicities (27). Herein we further described the characteristics, tumor responses, survival, and safety profiles of patients receiving subsequent S-1 maintenance therapy, and explored prognostic factors in cases further having S-1 maintenance.

We observed that the ORR and the mOS in the total 182 patients with APAC receiving NPS (ORR, 53.8%; mOS, 11.2 months) were similar with those reported in our previous phase II NASPAC trial (12) (ORR, 50.0%; mOS, 9.4 months), and in another study on NPS followed by S-1 maintenance for APAC (ORR, 53.1%; mOS, 11.2 months) (27), suggesting the good verifiability of the efficacy of the NPS regimen. In this study, 123 (67.6%) patients without progression within the first 4 months during NPS treatment, of whom 74 patients received S-1 maintenance according to the physicians’ discretion based on the patients’ preference and ECOG performance status as well as on evaluations of efficacies and benefits. However, it is worthwhile to suggest that it might be better to change the strategy of APAC treatment from continuous intensive chemotherapy to the single-agent maintenance mode at a proper time-point.

Favorable outcomes remain challenging in PDAC, in part due to the lack of validated markers for patient and treatment selection and thus optimal clinical decision-making. Increasingly, however, therapeutic advancement for PDAC is accompanied by evaluations of response markers (28). Tumor biological behavior was considered as an important prognostic factor. In borderline resectable PDAC, patients with initially elevated CA19-9 levels, who do not have a decline to a sustained low level in response to chemotherapy, are at risk for disease progression and poor survival (29), which highlight the importance of CA19-9 response in achieving satisfactory ORR and favorable survival. In addition, tumor burden was a widely recognized prognostic factor in many kinds of cancers including pancreatic cancer. We could speculate that the maintenance strategy might not work well if the tumor burden was too large and that best supportive care instead of continuous chemotherapy or sequential treatment with two different chemotherapy regimens might be a better option.

Our previous phase II trial showed that in APAC patients receiving NPS chemotherapy, remarkable survival benefits were observed in female patients (mPFS: 7.7 months, mOS: 18.2 months) with CA19-9 response (mPFS: 6.8 months, mOS: 11.8 months), objective response (mPFS: 6.9 months, mOS: 12.2 months), or baseline ECOG score of 0 (mPFS: 7.5 months, mOS: 16.1 months) (12). Herein, we further found that only stable disease best responded to chemotherapy, with an ECOG score of 1, and metastatic disease significantly and independently predicted worse survival in patients receiving S-1 maintenance after NPS chemotherapy.

In patients receiving S-1 maintenance after first-line NPS therapy, the most common all-grade and Grade 3 or above hematological adverse events were neutropenia (82.4% and 40.5%, respectively), and the most common all-grade and Grade 3 or above non-hematological adverse events were peripheral neurotoxicity (66.2% and 6.8%, respectively). S-1 maintenance was significantly associated with a 17.5% unit increase in all-grade hand-foot syndrome incidence and 18.1% increase in Grade 3 or above neutropenia incidence. The incidence of Grade 3 or above neutropenia was higher in our study than in a previous phase II study with similar research focus (27.6%) (27), which could be due to the differences in study protocol and patient and tumor characteristics. Also considering the higher ORR and survival in our study, there may be room for modification in the chemotherapy regimen with further optimized and well-balanced oncologic benefits and adverse events. The prognostic role of treatment-related peripheral neuropathy was limited during nab-paclitaxel plus gemcitabine treatment for patients with metastatic pancreatic cancer (30), and the replace of gemcitabine with S-1 is expected to be associated with safer profiles. While generally manageable, various adverse events during treatment call for meticulous and specific monitoring.

There are several important points to further discuss. Pancreas body/tail cancers comprised the majority in our study, where only locally advanced or metastatic unresectable pancreatic cancers were included. This site distribution might not be in compliance with the distribution in overall pancreatic cancers, as pancreatic body/tail cancers might more often involve major vasculatures rendering them unresectable and might be more insidious in disease course without causing obvious symptoms or signs, like jaundice, at the early stage of disease, making them more often diagnosed at an advanced stage. In contrast, pancreatic head/neck cancers might be more often detected and surgically managed at an earlier state as they are relatively more distant from major vasculatures and more often surgically manageable, having more often symptoms and/or signs (especially obstructive jaundice) at an earlier stage (2). The relatively small case number could be another reason.

S-1 has been majorly used in Asian countries based on some major clinical trials conducted in Asia (31, 32). The S-1 agent has been shown to potentially improve the efficacy of oral fluorinated pyrimidines by combining the antineoplastic agent with two biologic modulators aimed at reducing bowel toxicity and increasing the antitumor efficacy of the fluoro-pyrimidine. Patient and cancer characteristics might differ across different ethnicities, with possibly different efficacies and/or safety profiles of S-1 treatment. The outcomes based on Asian patients might not be directly generalizable or applicable to Western patients before further relevant evidence in the Western population is revealed, as with the case in gastric cancer (33). In Western populations tolerable doses of S-1 may be lower as a result of higher folate levels and genetic differences in drug metabolism between Western and Asian populations (33).

In our real-world study, some patients with stable disease within a certain time period during NPS treatment received subsequent S-1 monotherapy as maintenance. For patients with progressive pancreatic cancers not responsive to induction combination NPS chemotherapy including S-1, S-1 monotherapy, as maintenance, might not help to prevent disease from further progression. For patients with stable disease rather than progressive disease during induction chemotherapy, the regimen components in S-1 might help to inhibit cancer progression to some extent. Notably, the combination NPS regimen, especially the nab-paclitaxel component, might cause severe adverse events including peripheral neurotoxicity and neutropenia, which significantly deteriorates patients’ quality of life; and the toxicities might accumulate with longer duration of use. It is uncertain whether the possibly limited survival benefits brought about by the addition of nab-paclitaxel to S-1 might outweigh or be compromised by the toxicities caused, especially in the longer term. S-1, an oral fluoropyrimidine derivative, has been shown to have good efficacy and tolerability with milder adverse events (31, 32).

Our study had some limitations. Metastatic and locally advanced pancreatic cancers were both included in this study, which might present different prognostic profiles. However, S-1 maintenance followed NPS showed good compliance and efficacy in both metastatic and locally advanced diseases. Compared to other chemotherapeutic drugs, S-1 is more easily administrable with better health economic indices. Before further randomized evidence is obtained, our analysis, based on the previous phase II trial and with careful and rigorous statistics and interpretations, provides important evidence supporting S-1 maintenance following NPS chemotherapy for APAC, thus laying key foundations for further endeavors and offering new hope for this disease of despair.

In conclusion, S-1 maintenance after tumor response or stable disease induced by first-line NPS treatment was effective and well-tolerated for some patients with APAC, which offers a promising alternative treatment strategy with encouraging survival for APAC. Further relevant investigations with larger sample sizes and with randomized design are warranted to validate our findings.

The datasets presented in this article are not readily available because the data were used under license for the current study. Requests to access the datasets should be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the institutional review boards at the Chinese PLA General Hospital, Beijing, China. The patients/participants provided their written informed consent to participate in this study.

GD and YS conceived and designed the study. YS, HY, YL, SG, LH, QH, and ZW collected the original data. JY and JL contributed to blinded pathological review. YS, HY, YL, LH, QH, and GD analyzed and interpreted the data. YS drafted the paper. All authors participated in revising the manuscript critically. All authors read and approved the final manuscript.

This work was supported by Suzhou collaborative medicine foundation BG-BJ-3752. The funding body has no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.

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

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.