Nab-paclitaxel plus S-1 induction chemotherapy for patients with locally advanced pancreatic cancer: a multicenter, open-label phase 2 study Authors: Zhe Cao, Jiangdong Qiu, Wenhao Luo, Juan Li, Hao Chen, Yueze Liu, Shengmian Li, Fei Li, Yinmo Yang, Yuejuan Cheng, Taiping Zhang, Huadan Xue, Chunmei Bai, Yupei Zhao Published: 2026 ABSTRACT Locally advanced pancreatic cancer (LAPC) is associated with a poor prognosis and limited therapeutic options. This study aimed to evaluate the efficacy and safety of nab-paclitaxel plus S-1 (SnP) as a first-line treatment in patients with previously untreated LAPC. In this multicenter, single-arm phase 2 trial conducted across four centers in China, eligible patients received up to eight cycles of nab-paclitaxel combined with S-1, followed by investigator-selected continuation strategies in the absence of disease progression or intolerable toxicity. The primary endpoint was the 6-month progression-free survival (PFS). Between April 2019 and March 2023, 60 patients were enrolled, of whom 48.3% completed induction therapy and 18.3% subsequently underwent surgical resection, with 10 achieving R0/R1 status. The 6-month PFS rate was 71.0% (95% CI: 57.6%–80.9%), while the median PFS and overall survival (OS) were 11.1 months and 20.2 months, respectively. The 12-month OS rate reached 83.3%. The objective response rate was 26.7%, and the disease control rate was 90.0%. Grade ≥3 treatment-related adverse events occurred in 53.3% of patients, most commonly neutropenia and leukopenia, with no treatment-related deaths observed. These findings suggest that SnP demonstrates promising antitumor activity with manageable toxicity in LAPC and may facilitate conversion to surgical resectability in a subset of patients. This clinical study was registered at ClinicalTrials.gov (NCT03885219). Introduction Pancreatic cancer is a highly malignant tumor of the digestive system. The morbidity and mortality are increasing, and the 5-year survival rate is only 13%.1,2 Radical resection is the only way to cure pancreatic cancer, but only 10%-20% of newly diagnosed patients are eligible for surgical resection.3 Although there is no distant metastasis, locally advanced pancreatic cancer (LAPC) always invades adjacent critical vasculature, leading to unresectable diseases.4 In recent years, chemotherapy has been applied to the treatment of LAPC, including the FOLFIRINOX regimen, gemcitabine, nab-paclitaxel, and S-1.4,5 Research shows that using more intensified chemotherapy regimens may improve the surgical resection rates in some patients with LAPC. A systematic literature review of studies on FOLFIRINOX (leucovorin, fluorouracil, irinotecan, and oxaliplatin) as a first-line treatment for LAPC revealed an average resection rate of 28%, ranging from 0% to 43%, across 12 studies.6 While several systemic therapies have demonstrated some degree of benefit in pancreatic ductal adenocarcinoma overall, a major evidence gap remains for LAPC. Current treatment approaches are largely extrapolated from data in metastatic PDAC rather than based on LAPC-specific evidence. Because of this lack of robust randomized phase III studies or large, high-quality single-arm investigations, no universally accepted standard first-line regimen for LAPC has been established. Consequently, the NCCN recommendations for LAPC remain Category 2 A rather than Level 1 evidence. These gaps highlight the need to further explore chemotherapy strategies specifically tailored to LAPC. The international multicenter randomized controlled trial (MPACT study) found that compared with gemcitabine monotherapy, the median overall survival (OS) and progression-free survival (PFS) of gemcitabine combined with nab-paclitaxel (GnP) were significantly prolonged, and the 1-year survival rate increased from 22% to 35%.7 The FDA approved GnP for first-line treatment of metastatic pancreatic cancer. The results of the international multicenter phase II trial (LAPACT study) showed that GnP was used as the first-line treatment for LAPC, with a median PFS of 10.9 months and a median OS of 18.8 months.8 NCCN guidelines recommend this regimen for the first-line treatment of metastatic and LAPC. In Asian populations, tolerability to FOLFIRINOX may, to some extent, limit its use in routine practice, particularly because of treatment-related toxicities and the need for dose modifications in some patients. S-1 is a fluorouracil derivative oral anticancer agent. According to the results of a phase III clinical study, S-1 was not inferior to gemcitabine in the first-line treatment of LAPC and metastatic disease, and it is safe and well tolerated, especially in the Asian population.9 Preclinical studies have confirmed that nab-paclitaxel plus S-1 (SnP) has synergistic antitumor effects.10 A phase II clinical study included 32 locally advanced/metastatic pancreatic cancer patients treated with SnP, and the PFS and OS were 8.5 months and 13.7 months, respectively.11 SnP as a second-line treatment for patients who had previously failed gemcitabine treatment was effective and well tolerated. In summary, preclinical and early clinical studies have shown that SnP has high efficacy and good tolerance. The objective of this study was to investigate the efficacy and safety of SnP as induction therapy for LAPC in a prospective phase II clinical trial. Results Patient characteristics In total, 72 patients were screened for enrollment between April 25, 2019, and March 29, 2023. Sixty patients were finally enrolled in this study. The patient flow chart is shown in Fig. 1. Twenty-nine patients (48.3%) completed the induction, including 10 patients who received radical resection and 31 patients (51.7%) who discontinued induction. The median number of completed cycles was 6, and all patients received 346 cycles of treatment in total. Twelve patients discontinued treatment for adverse events or intolerance, and another 13 patients discontinued treatment because of progressive diseases. Other reasons for discontinued treatment included alteration in treatment (4 [6.7%]) and patients’ decision to withdraw from the study (2 [3.3%]). The characteristics of these 60 patients are summarized in Table 1.Fig. 1The flowchart of the trial profile. Seventy-two patients were assessed, and 60 were included for induction. Finally, 29 patients completed induction treatmentTable 1Baseline characteristicsCharacteristicsIntention-to-treat population (n = 60)Patients who completed induction treatment and received treatment per investigator’s choice (n = 29)Continued nab-paclitaxel/S-1 combination (n = 4)Chemoradiation (n = 8)Surgery (n = 11)Other systemic treatment (n = 5)Radiation (n = 1)Sex, Male37 (61.7)1 (25.0)5 (62.5)8 (72.7)3 (60.0)0 (0.0) Female23 (38.3)3 (75.0)3 (37.5)3 (27.3)2 (40.0)1 (100.0)ECOG PS 035 (58.3)2 (50.0)6 (75.0)8 (72.7)2 (40.0)0 (0.0) 125 (41.7)2 (50.0)2 (25.0)3 (27.3)3 (60.0)1 (100.0)Age (years), Median (range)61.0 (29.0-73.0)63.5 (60.0-67.0)62.5 (53.0-67.0)63.0 (48.0-66.0)61.0 (46.0-68.0)37.0 (37.0-37.0) <6541 (68.3)2 (50.0)4 (50.0)6 (54.5)4 (80.0)1 (100.0) 65-7519 (31.7)2 (50.0)4 (50.0)5 (45.5)1 (20.0)0 (0.0)Albumin (g/L), Median (Range)41.4 (32.3-48.0)41.2 (38.7,47.1)40.0 (35.0,44.0)41.9 (37.0,47.0)43.0 (41.0,47.0)-Ratio of neutrophil/lymphocyte ≤552 (86.7)3 (75.0)7 (87.5)8 (72.7)5 (100.0)0 (0.0) å> 55 (8.3)1 (25.0)0 (0.0)2 (18.2)0 (0.0)0 (0.0) Unknown3 (5.0)0 (0.0)1 (12.5)1 (9.1)0 (0.0)1 (100.0)CEAa (ng/ml), Median (Range)3.4 (1.0-37.4)2.4 (1.9-7.9)3.7 (1.8-16.7)3.4 (1.2-37.4)2.7 (1.4-4.1)- Normal39 (65.0)2 (50.0)6 (75.0)7 (63.6)5 (100.0)1 (100.0) Abnormal19 (31.7)1 (25.0)2 (25.0)4 (36.4)0 (0.0)0 (0.0) Unknown2 (3.3)1 (25.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)CA125a (U/ml), Median (Range)16.3 (4.9-149.7)10.7 (8.8-16.3)23.2 (14.3-82.6)15.3 (5.1-32.6)12.3 (9.8-53.6)- Normal44 (73.3)3 (75.0)6 (75.0)9 (81.8)3 (60.0)1 (100.0) Abnormal9 (15.0)1 (25.0)2 (25.0)2 (18.2)2 (40.0)0 (0.0) Unknown7 (11.7)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)CA19-9a (U/ml), Median (Range)257.2 (1.7-3509.0)346.4 (54.7-2174.0)538.3 (110.7-1795.0)257.9 (1.7-1000.0)115.8 (33.4-792.0)1220.0 (1220.0-1220.0) Normal10 (16.7)0 (0.0)0 (0.0)4 (36.4)1 (20.0)0 (0.0) Abnormal49 (81.7)3 (75.0)8 (100.0)7 (63.6)4 (80.0)1 (100.0) Unknown1 (1.7)1 (25.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)Tumor site Pancreatic head34 (56.7)2 (50.0)4 (50.0)6 (54.5)1 (20.0)1 (100.0) Pancreatic body and tail26 (43.3)2 (50.0)4 (50.0)5 (45.5)4 (80.0)0 (0.0)Sum of the longest diameters of tumors (cm), Median (Range)6.7 (2.3-12.7)5.4 (4.5, -11.0)7.0 (3.2-12.7)7.0 (3.5-11.4)6.2 (3.5-10.4)7.7 (7.7-7.7)Lymph node status Negative18 (30.0)2 (50.0)1 (12.5)7 (63.6)1 (20.0)0 (0.0) Positive40 (66.7)2 (50.0)7 (87.5)4 (36.4)4 (80.0)1 (100.0) Unknown2 (3.3)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)Vessels involved Celiac axis15 (25.0)1 (25.0)2 (25.0)2 (18.2)3 (60.0)0 (0.0) Superior mesenteric artery17 (28.3)1 (25.0)3 (37.5)3 (27.3)2 (40.0)1 (100.0) Common hepatic artery14 (23.3)0 (0.0)2 (25.0)3 (27.3)2 (40.0)1 (100.0) Proper hepatic artery2 (3.3)1 (25.0)1 (12.5)0 (0.0)0 (0.0)0 (0.0) Portal vein30 (50.0)2 (50.0)6 (75.0)5 (45.5)3 (60.0)1 (100.0) Superior mesenteric vein36 (60.0)1 (25.0)6 (75.0)7 (63.6)2 (40.0)1 (100.0)Data are median (IQR) or n (%), unless otherwise specified. ECOG, Eastern Cooperative Oncology GroupaNormal range of tumor markers: CA19-9: 0-34 U/ml; CEA: 0-5 ng/ml; CA125: 0-35 U/ml The flowchart of the trial profile. Seventy-two patients were assessed, and 60 were included for induction. Finally, 29 patients completed induction treatment Table 1Baseline characteristicsCharacteristicsIntention-to-treat population (n = 60)Patients who completed induction treatment and received treatment per investigator’s choice (n = 29)Continued nab-paclitaxel/S-1 combination (n = 4)Chemoradiation (n = 8)Surgery (n = 11)Other systemic treatment (n = 5)Radiation (n = 1)Sex, Male37 (61.7)1 (25.0)5 (62.5)8 (72.7)3 (60.0)0 (0.0) Female23 (38.3)3 (75.0)3 (37.5)3 (27.3)2 (40.0)1 (100.0)ECOG PS 035 (58.3)2 (50.0)6 (75.0)8 (72.7)2 (40.0)0 (0.0) 125 (41.7)2 (50.0)2 (25.0)3 (27.3)3 (60.0)1 (100.0)Age (years), Median (range)61.0 (29.0-73.0)63.5 (60.0-67.0)62.5 (53.0-67.0)63.0 (48.0-66.0)61.0 (46.0-68.0)37.0 (37.0-37.0) <6541 (68.3)2 (50.0)4 (50.0)6 (54.5)4 (80.0)1 (100.0) 65-7519 (31.7)2 (50.0)4 (50.0)5 (45.5)1 (20.0)0 (0.0)Albumin (g/L), Median (Range)41.4 (32.3-48.0)41.2 (38.7,47.1)40.0 (35.0,44.0)41.9 (37.0,47.0)43.0 (41.0,47.0)-Ratio of neutrophil/lymphocyte ≤552 (86.7)3 (75.0)7 (87.5)8 (72.7)5 (100.0)0 (0.0) å> 55 (8.3)1 (25.0)0 (0.0)2 (18.2)0 (0.0)0 (0.0) Unknown3 (5.0)0 (0.0)1 (12.5)1 (9.1)0 (0.0)1 (100.0)CEAa (ng/ml), Median (Range)3.4 (1.0-37.4)2.4 (1.9-7.9)3.7 (1.8-16.7)3.4 (1.2-37.4)2.7 (1.4-4.1)- Normal39 (65.0)2 (50.0)6 (75.0)7 (63.6)5 (100.0)1 (100.0) Abnormal19 (31.7)1 (25.0)2 (25.0)4 (36.4)0 (0.0)0 (0.0) Unknown2 (3.3)1 (25.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)CA125a (U/ml), Median (Range)16.3 (4.9-149.7)10.7 (8.8-16.3)23.2 (14.3-82.6)15.3 (5.1-32.6)12.3 (9.8-53.6)- Normal44 (73.3)3 (75.0)6 (75.0)9 (81.8)3 (60.0)1 (100.0) Abnormal9 (15.0)1 (25.0)2 (25.0)2 (18.2)2 (40.0)0 (0.0) Unknown7 (11.7)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)CA19-9a (U/ml), Median (Range)257.2 (1.7-3509.0)346.4 (54.7-2174.0)538.3 (110.7-1795.0)257.9 (1.7-1000.0)115.8 (33.4-792.0)1220.0 (1220.0-1220.0) Normal10 (16.7)0 (0.0)0 (0.0)4 (36.4)1 (20.0)0 (0.0) Abnormal49 (81.7)3 (75.0)8 (100.0)7 (63.6)4 (80.0)1 (100.0) Unknown1 (1.7)1 (25.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)Tumor site Pancreatic head34 (56.7)2 (50.0)4 (50.0)6 (54.5)1 (20.0)1 (100.0) Pancreatic body and tail26 (43.3)2 (50.0)4 (50.0)5 (45.5)4 (80.0)0 (0.0)Sum of the longest diameters of tumors (cm), Median (Range)6.7 (2.3-12.7)5.4 (4.5, -11.0)7.0 (3.2-12.7)7.0 (3.5-11.4)6.2 (3.5-10.4)7.7 (7.7-7.7)Lymph node status Negative18 (30.0)2 (50.0)1 (12.5)7 (63.6)1 (20.0)0 (0.0) Positive40 (66.7)2 (50.0)7 (87.5)4 (36.4)4 (80.0)1 (100.0) Unknown2 (3.3)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)Vessels involved Celiac axis15 (25.0)1 (25.0)2 (25.0)2 (18.2)3 (60.0)0 (0.0) Superior mesenteric artery17 (28.3)1 (25.0)3 (37.5)3 (27.3)2 (40.0)1 (100.0) Common hepatic artery14 (23.3)0 (0.0)2 (25.0)3 (27.3)2 (40.0)1 (100.0) Proper hepatic artery2 (3.3)1 (25.0)1 (12.5)0 (0.0)0 (0.0)0 (0.0) Portal vein30 (50.0)2 (50.0)6 (75.0)5 (45.5)3 (60.0)1 (100.0) Superior mesenteric vein36 (60.0)1 (25.0)6 (75.0)7 (63.6)2 (40.0)1 (100.0)Data are median (IQR) or n (%), unless otherwise specified. ECOG, Eastern Cooperative Oncology GroupaNormal range of tumor markers: CA19-9: 0-34 U/ml; CEA: 0-5 ng/ml; CA125: 0-35 U/ml Baseline characteristics Data are median (IQR) or n (%), unless otherwise specified. ECOG, Eastern Cooperative Oncology Group aNormal range of tumor markers: CA19-9: 0-34 U/ml; CEA: 0-5 ng/ml; CA125: 0-35 U/ml After completion of the induction, 29 patients received subsequent therapy according to the investigators’ choice: 4 continued SnP, 5 received other systemic treatment, 8 received chemoradiation, 1 received radiation and 11 underwent surgery. Treatment response Of the 60 patients, 59 had a baseline assessment and at least one had a postbaseline assessment. The tumor response after administration of SnP is summarized in Table 2. No patient had a complete response (CR), as determined by RECIST v1.1. PR was achieved in 16 (26.7%) patients, and 38 (63.3%) patients had SD. PD occurred in 5 (8.3%) patients. The overall ORR was 26.7% (95% CI: 16.1%-39.7%), and the disease control rate (DCR) was 90.0% (95% CI: 79.5%-96.2%). The best percentage change in tumor from baseline is displayed in Fig. 2. A total of 56 patients were tested for CA19-9 before and after treatment. Forty-six had abnormal baseline CA19-9 levels (>34 U/mL) and at least one posttreatment assessment. Thirty-two (69.6%) patients had a decrease in CA19-9 over 50%, and 27 (58.7%) patients had a decrease in CA19-9 over 70%. The CA19-9 levels of 17 (37.0%) patients decreased to normal levels. The changes in CA19-9 during treatment are displayed in Fig. 3.Fig. 2The best percentage change in the primary tumor from baseline. *New lesion except for the primary tumor occurred during induction. Lesion diameters were assessed by the study investigators. During induction, 16 of 60 patients (26.7%) achieved a reduction of at least 30% in the sum of diameters of the target lesions from baseline (gray dotted line)Fig. 3The best percentage change in CA19-9 during treatment. Of the 56 patients with at least one CA19-9 measurement, 46 had abnormal baseline CA19-9 levels (>34 U/mL) and at least one posttreatment assessment. Among these 46 patients, the rates of CA19-9 reduction ≥50%, ≥70%, and normalization were 69.6% (32/46), 58.7% (27/46), and 37.0% (17/46), respectivelyTable 2Best response during the induction phase in the intention-to-treat populationBest responsePatients (n = 60)Complete response (CR)0 (0.0)Partial response (PR)16 (26.7)All stable disease (SD)38 (63.3) SD ≥ 16 weeks31 (51.7) SD ≥ 24 weeks25 (41.7)Progressive disease (PD)5 (8.3)Not all evaluated1 (1.7)Objective response rate (ORR)16 (26.7) 95%CI16.1%-39.7%Disease Control Rate (DCR)54 (90.0) 95%CI79.5%-96.2%Tumor response was assessed by investigators in accordance with RECIST v1.1 guidelines, with evaluations performed after any nonprotocol anticancer treatments or surgical interventions were omitted. The results are presented as counts and percentages [n (%)] The best percentage change in the primary tumor from baseline. *New lesion except for the primary tumor occurred during induction. Lesion diameters were assessed by the study investigators. During induction, 16 of 60 patients (26.7%) achieved a reduction of at least 30% in the sum of diameters of the target lesions from baseline (gray dotted line) The best percentage change in CA19-9 during treatment. Of the 56 patients with at least one CA19-9 measurement, 46 had abnormal baseline CA19-9 levels (>34 U/mL) and at least one posttreatment assessment. Among these 46 patients, the rates of CA19-9 reduction ≥50%, ≥70%, and normalization were 69.6% (32/46), 58.7% (27/46), and 37.0% (17/46), respectively Table 2Best response during the induction phase in the intention-to-treat populationBest responsePatients (n = 60)Complete response (CR)0 (0.0)Partial response (PR)16 (26.7)All stable disease (SD)38 (63.3) SD ≥ 16 weeks31 (51.7) SD ≥ 24 weeks25 (41.7)Progressive disease (PD)5 (8.3)Not all evaluated1 (1.7)Objective response rate (ORR)16 (26.7) 95%CI16.1%-39.7%Disease Control Rate (DCR)54 (90.0) 95%CI79.5%-96.2%Tumor response was assessed by investigators in accordance with RECIST v1.1 guidelines, with evaluations performed after any nonprotocol anticancer treatments or surgical interventions were omitted. The results are presented as counts and percentages [n (%)] Best response during the induction phase in the intention-to-treat population Tumor response was assessed by investigators in accordance with RECIST v1.1 guidelines, with evaluations performed after any nonprotocol anticancer treatments or surgical interventions were omitted. The results are presented as counts and percentages [n (%)] Surgical outcomes Eleven (18.3%) patients underwent surgery. Among them, 10 patients (90.9%) underwent radical resection after induction therapy, and all of them (90.9%) had R0/R1 resection. Tumor resection failed in 1 patient after exploration during the operation, and a bypass operation was performed. Among these patients, 3 had distal pancreatectomy with celiac axis resection (DP-CAR), and 6 had SMV/PV resection. Two patients had grade 3 or higher postoperative complications. The postoperative pathology showed that 10 patients had a pathological response. After the operation, 6 patients received adjuvant therapy, including SnP, S-1, mFOLFIRINOX and immunotherapy. One patient did not receive adjuvant therapy, and 1 patient died within 90 days after the operation. Detailed information about the surgery is listed in Table 3.Table 3Outcomes of surgical operation after induction therapyPatients (n = 11)Surgical procedure PD4 (36.4) DP3 (27.3) DP-CAR3 (27.3) Bypass1 (9.1)PV/SMV resection Yes6 (54.5) No5 (45.5)Artery resection Yes3 (27.3) No8 (72.7)Postoperative complication Grade Ⅰ1 (9.1) Grade Ⅱ7 (63.6) Grade Ⅲ1 (9.1) Grade V1 (9.1) Unknown1 (9.1)Pathological differentiation Moderately4 (36.4) Poorly3 (27.3) pCR1 (9.1) Unknown3 (27.3)T stage T14 (36.4) T24 (36.4) T31 (9.1) T42 (18.2)N stage N07 (63.6) N14 (36.4)Pathological response (Evans) Grade IIa4 (36.4 Grade IIb4 (36.4) Grade III2 (18.2) Unknown1 (9.1)R0/R1 resection Yes10 (90.9) No1 (9.1)Adjuvant therapy Other systemic treatment5 (45.5) Continuing nab-paclitaxel/S-11 (9.1) No adjuvant therapy1 (9.1) Unknown3 (27.3) Postoperative mortality (90 days)1 (9.1)OS (95% CI) (months)17.8 (9.7,26.0)Data are n (%)PD Pancreaticoduodenectomy, DP Distal Pancreatectomy, DP-CAR Distal pancreatectomy with En bloc Celiac Axis Resection, PV Portal Vein, SMV Superior Mesenteric Vein, pCR Pathological Complete Response, OS Overall survival Table 3Outcomes of surgical operation after induction therapyPatients (n = 11)Surgical procedure PD4 (36.4) DP3 (27.3) DP-CAR3 (27.3) Bypass1 (9.1)PV/SMV resection Yes6 (54.5) No5 (45.5)Artery resection Yes3 (27.3) No8 (72.7)Postoperative complication Grade Ⅰ1 (9.1) Grade Ⅱ7 (63.6) Grade Ⅲ1 (9.1) Grade V1 (9.1) Unknown1 (9.1)Pathological differentiation Moderately4 (36.4) Poorly3 (27.3) pCR1 (9.1) Unknown3 (27.3)T stage T14 (36.4) T24 (36.4) T31 (9.1) T42 (18.2)N stage N07 (63.6) N14 (36.4)Pathological response (Evans) Grade IIa4 (36.4 Grade IIb4 (36.4) Grade III2 (18.2) Unknown1 (9.1)R0/R1 resection Yes10 (90.9) No1 (9.1)Adjuvant therapy Other systemic treatment5 (45.5) Continuing nab-paclitaxel/S-11 (9.1) No adjuvant therapy1 (9.1) Unknown3 (27.3) Postoperative mortality (90 days)1 (9.1)OS (95% CI) (months)17.8 (9.7,26.0)Data are n (%)PD Pancreaticoduodenectomy, DP Distal Pancreatectomy, DP-CAR Distal pancreatectomy with En bloc Celiac Axis Resection, PV Portal Vein, SMV Superior Mesenteric Vein, pCR Pathological Complete Response, OS Overall survival Outcomes of surgical operation after induction therapy Data are n (%) PD Pancreaticoduodenectomy, DP Distal Pancreatectomy, DP-CAR Distal pancreatectomy with En bloc Celiac Axis Resection, PV Portal Vein, SMV Superior Mesenteric Vein, pCR Pathological Complete Response, OS Overall survival Survival outcomes At a median follow-up of 15.6 months (range: 1.31-72.1), PFS events were documented in 51 out of the 60 patients (including 44 cases of disease progression and 7 deaths). The median PFS was 11.1 months (95% CI: 8.1-14.2). The 6-month PFS rate was 71.0% (95% CI: 57.6%-80.9%), the 12-month PFS rate was 45.0% (95% CI: 32.0%-57.2%), and the 24-month PFS rate was 14.8% (95% CI: 6.8%-25.8%). Thirty-three of the 60 patients encountered survival events. The median OS was 20.2 months (95% CI: 11.2-29.2). The 12-month OS rate was 83.3% (95% CI: 70.3%-90.9%), the 24-month OS rate was 45.7% (95% CI: 31.3%-59.0%), and the 36-month OS rate was 25.1% (95% CI: 10.7%-42.6%). The survival outcomes are shown in Fig. 4.Fig. 4Progression-free survival (A) and overall survival (B) in the intention-to-treat population Progression-free survival (A) and overall survival (B) in the intention-to-treat population Safety of SnP In the treated population of 60 patients, 58 patients (96.7%) had all grades of TRAEs. Among these patients, 32 patients (53.3%) had TRAEs of grade 3 or higher. The most common grade 3 or higher TRAEs were neutropenia (23 [38.3%] of 60 patients), followed by leukopenia (13 [21.7%]) and fatigue (4 [6.7%]). The most common grade 1 and 2 TRAEs were anemia (28 [46.7%]), followed by peripheral neuropathy (25 [41.7%]). Detailed information about the main TRAEs is listed in Table 4. Delayed chemotherapy (≥14 days) occurred in 13 patients (21.7%), and 5 of them (8.3%) had adverse events. In the total of 346 finished treatment cycles, delayed chemotherapy occurred in 16 cycles (4.6%). In addition, dose reduction occurred in 13 patients (21.7%), and all of them contributed to adverse events. In the total of 346 finished treatment cycles, dose reduction occurred in 14 cycles (4.0%) because of adverse events. No treatment-related deaths due to toxicity were reported.Table 4Main treatment-related adverse events (n = 60)G1-2G3-4All gradesAny TRAEs26 (43.3)32 (53.3)58 (96.7)Hematological toxicity Neutropenia15 (25.0)23 (38.3)38 (63.3) Leukopenia22 (36.7)13 (21.7)35 (58.3) Anemia28(46.7)1 (1.7)29 (48.3) Thrombocytopenia16 (26.7)1 (1.7)17 (28.3) Lymphopenia2 (3.3)0 (0.0)2 (3.3) Febrile neutropenia0 (0.0)1 (1.7)1 (1.7)Nonhaematological toxicity (≥10% incidence) Peripheral neuropathy25 (41.7)2 (3.3)27 (45.0) Fatigue21 (35.0)4 (6.7)25 (41.7) Nausea19 (31.7)0 (0.0)19 (31.7) Constipation15 (25.0)0 (0.0)15 (25.0) Diarrhea10 (16.7)2 (3.3)12 (20.0) Anorexia12 (20.0)0 (0.0)12 (20.0) Fever11 (18.3)0 (0.0)11 (18.3) Oral mucositis7 (11.7)3 (5.0)10 (16.7) Rash10 (16.7)0 (0.0)10 (16.7) Hand-foot syndrome8 (13.3)1 (1.7)9 (15.0) Alopecia9 (15.0)0 (0.0)9 (15.0) Skin hyperpigmentation8 (13.3)0 (0.0)8 (13.3) Blood bilirubin increased7 (11.7)0 (0.0)7 (11.7) AST increased6 (10.0)1 (1.7)7 (11.7)Data are n (%). AST, aspartate aminotransferase Table 4Main treatment-related adverse events (n = 60)G1-2G3-4All gradesAny TRAEs26 (43.3)32 (53.3)58 (96.7)Hematological toxicity Neutropenia15 (25.0)23 (38.3)38 (63.3) Leukopenia22 (36.7)13 (21.7)35 (58.3) Anemia28(46.7)1 (1.7)29 (48.3) Thrombocytopenia16 (26.7)1 (1.7)17 (28.3) Lymphopenia2 (3.3)0 (0.0)2 (3.3) Febrile neutropenia0 (0.0)1 (1.7)1 (1.7)Nonhaematological toxicity (≥10% incidence) Peripheral neuropathy25 (41.7)2 (3.3)27 (45.0) Fatigue21 (35.0)4 (6.7)25 (41.7) Nausea19 (31.7)0 (0.0)19 (31.7) Constipation15 (25.0)0 (0.0)15 (25.0) Diarrhea10 (16.7)2 (3.3)12 (20.0) Anorexia12 (20.0)0 (0.0)12 (20.0) Fever11 (18.3)0 (0.0)11 (18.3) Oral mucositis7 (11.7)3 (5.0)10 (16.7) Rash10 (16.7)0 (0.0)10 (16.7) Hand-foot syndrome8 (13.3)1 (1.7)9 (15.0) Alopecia9 (15.0)0 (0.0)9 (15.0) Skin hyperpigmentation8 (13.3)0 (0.0)8 (13.3) Blood bilirubin increased7 (11.7)0 (0.0)7 (11.7) AST increased6 (10.0)1 (1.7)7 (11.7)Data are n (%). AST, aspartate aminotransferase Main treatment-related adverse events (n = 60) Data are n (%). AST, aspartate aminotransferase Discussion In this phase II study, SnP showed a promising efficacy signal with acceptable tolerability as induction therapy for LAPC. The regimen was deliverable across multiple centers and allowed some patients to proceed to subsequent local treatment and, in selected cases, surgical management. These results warrant confirmation in future randomized studies to clarify the comparative benefit and optimal sequencing strategy. Systemic chemotherapy is administered to most patients with LAPC to slow tumor progression and relieve disease-related symptoms. The MPACT trial7 showed positive results for GnP; the NAPOLI-3 trial12 showed positive outcomes with NALIRIFOX in metastatic disease; however, few regimens have undergone a dedicated prospective analysis in patients with LAPC. To the best of our knowledge, this is the first trial to evaluate SnP in the induction therapy of LAPC. In the present study, acceptable tolerability was one of the advantages of SnP, especially in the Asian population. Compared with the LAPACT study, where grade 3 or higher AEs occurred in 80% of patients, the rate of grade 3 or higher AEs in SnP was 53.3%.8 The most common grade 3 or higher adverse event was neutropenia, with a rate of 38.3%, which was comparable with GnP.7,8 However, the rates of other adverse events, such as leukopenia and fatigue, were lower in SnP than in GnP.8 In addition, the rate of dose reduction was lower in SnP than in GnP, FOLFIRINOX and NALIRIFOX.12,13 Regarding treatment efficacy, SnP showed potential superiority. The 6-month PFS rate was selected as the primary endpoint of this study to reflect disease control in a short period, and a 71.0% 6-month PFS rate was achieved. Compared with other studies involving LAPC, SnP achieved a higher DCR of 90.0%. In the LAPACT trial (gemcitabine + nab-paclitaxel), the DCR was 77.6%.8 In the present study, 16 patients had a complete or partial response, and the ORR was 26.7%, which was comparable to that of nab-paclitaxel plus gemcitabine. However, we should be mindful of the clinical relevance and limitations of cross-trial comparisons and interpret them with caution. For example, JCOG140714 is the only published randomized trial in LAPC that directly compared mFOLFIRINOX with GnP. The results showed that GnP achieved a higher disease control rate (DCR, 87.3% vs. 96.5%). In our study, the median DCR with SnP was 90.0%, which is comparable to that reported in JCOG1407. In the future, further randomized controlled studies are warranted. In addition, SnP had advantages in survival over other regimens based on gemcitabine or nab-paclitaxel. In the present study, the median PFS was 11.1 months (95% CI: 8.1-14.2), and the median OS was 20.2 months (95% CI: 11.2-29.2). In contrast, in the SCALOP trial, the median OS was 12.7 months (95% CI: 11.0-14.5),15 and in the LAP07 trial, the median OS was 12.8 months (95% CI: 11.8-14.1),16 whereas in the LAPACT trial, it was 18.8 months (90% CI: 15.0-24.0).8 Another phase III randomized trial involving neoadjuvant gemcitabine showed that the PFS was only 7.7 months (95% CI: 6.2-9.2), and the OS was only 15.4 months (95% CI: 11.7-18.6).17 However, the survival benefit of SnP in LAPC may not be as superior as FOLFIRINOX, which had a median OS of 24.2 months (ranging from 10.0 to 32.7 months) according to a systemic review, and further prospective study is needed.6 Interpretation of mPFS and mOS should be undertaken with caution because only 48.4% of patients (29/60) completed induction therapy, and a substantial proportion subsequently received heterogeneous postinduction treatments. At present, the optimal postinduction strategy for LAPC remains controversial. Postinduction management is typically individualized based on an integrated assessment of treatment tolerability, performance status, and efficacy evaluation (e.g., radiographic response, disease control, and reassessment of resectability). Albumin-bound paclitaxel plus S-1 has not been established as a standard regimen for LAPC. The present trial was originally designed as an exploratory study, prioritizing participant safety while assessing feasibility and preliminary signals of efficacy. Accordingly, although postinduction treatment options were recommended, the final selection was left to the investigators’ clinical judgment. While this approach more closely reflects real-world practice, it may reduce the reproducibility of our findings across other settings and therefore represents a limitation of the study. Similar to prior induction studies in LAPC (e.g., LAPACT),8 the postinduction strategy was not prespecified as a single uniform approach. In LAPACT, 58% of patients (62/107) completed induction therapy; 47 patients (44%) then received investigator-selected subsequent treatments, including continuation of nab-paclitaxel plus gemcitabine (12/107, 11%), chemoradiotherapy (18/107, 17%), and surgery (17/107, 16%; including 7 R0 resections and 9 R1 resections), whereas 15 patients (14%) completed induction therapy but did not receive the recommended therapy. These data highlight a pragmatic reality: even in prospective LAPC trials, postinduction treatment choices vary. Therefore, the mPFS and mOS observed in our study reflect outcomes under a sequential treatment strategy rather than the effect of induction chemotherapy with SnP alone. Recent studies suggest that the CA19-9 response could be an independent prognostic factor.18,19 The LAPACT study8 showed that among the 92 patients, 68 (74%) achieved a reduction of at least 50% in CA19-9 levels. In this study, serum CA19-9 levels declined during induction therapy in most patients: a reduction of ≥50% was observed in 32 (69.6%) of 46 patients, and a reduction of ≥70% was seen in 27 patients (58.7%). The CA 19-9 levels of 17 (37.0%) patients decreased to normal levels. At present, the optimal postinduction sequential strategy for LAPC remains controversial. The phase III randomized trial conducted in LAPC (CONKO-007)20 showed that delivering chemoradiotherapy (CRT) after induction therapy can increase the R0 resection rate among patients who proceed to surgery and may translate into improved outcomes. In contrast, the LAP07 trial used a gemcitabine-based induction regimen, whereas the ESPAC-5 study evaluated CRT with concurrent capecitabine without induction chemotherapy. Collectively, CONKO-007 suggests that induction chemotherapy may help identify LAPC patients who are more likely to benefit from subsequent CRT and that FOLFIRINOX plus radiotherapy may be more effective than gemcitabine-based approaches. These findings highlight the need for more effective induction regimens in LAPC to address early micrometastatic spread and achieve meaningful tumor downsizing. In our study, the 6-month progression-free survival rate with SnP was 71.0%, with favorable tolerability. This provides supportive evidence for considering SnP as an induction option before CRT, particularly for patients who are unable to tolerate mFOLFIRINOX. Nevertheless, this hypothesis requires confirmation in future phase III randomized trials. In addition, the disease control rate during the induction phase was promising, indicating that SnP had antitumor activity. Considering that the pathological assessment criteria may vary in different centers, we calculated the R0/R1 resection rate instead of the R0 resection rate. Surgical resection is uncommon in patients with LAPC. Although all individuals in this study were classified as unresectable at baseline, 10 of 60 achieved R0 or R1 resection following the induction phase, suggesting that treatment with SnP may convert a subset of initially unresectable tumors to resectable disease. There are some inherent limitations to the present study. First, the design of this study is single-arm, not randomized, which brought challenges in evaluating the survival and conversion to resection by SnP. In addition, the treatment options after induction therapy were decided by investigators, which may be affected by the investigators’ personal judgment, especially the criteria for surgical resection. Thus, the investigators’ choice may influence the survival outcomes of patients. Furthermore, the imaging before and after induction were not evaluated by central review, and disease status was determined by investigators according to the protocol criteria from NCCN guidelines. In our study, the sample size was limited, and all participants were enrolled from only four centers in China. This relatively narrow enrollment may reduce the statistical robustness of our estimates and limit how confidently these findings can be extended to other geographic regions and non-Asian populations, where patient characteristics, clinical practice patterns, and supportive care may differ. Moreover, the median follow-up of 15.6 months is relatively short and may limit the maturity of the survival outcomes. The present trial was the first to prospectively evaluate the efficacy and safety of SnP in patients with LAPC. The findings from the present study demonstrated that SnP for LAPC showed superior efficacy with mild toxicity, which may prolong patient survival. These results suggested that SnP could be an alternative induction therapy option for LAPC, especially in patients who cannot tolerate mFOLFIRINOX and NALIRIFOX. Materials and methods Study design This study was a multicenter, open-label, single-arm, phase II study of patients with LAPC conducted at 4 hospitals in China. Patients with LAPC received eight cycles of induction chemotherapy with SnP. The primary endpoint was the 6-month PFS rate. The secondary endpoints were PFS, OS, objective response rate (ORR), disease control rate (DCR), R0/R1 resection rate and safety. Post hoc analyses examined the changes from baseline in serum carbohydrate antigen 19-9 (CA19-9) concentrations. PFS is defined as the interval from the first dose of study treatment to the earliest occurrence of disease progression or death from any cause. Participants who neither progress nor die will be censored at the most recent date on which they were confirmed to be progression-free. If a participant initiates new anticancer therapy or receives palliative radiotherapy or surgery at a lesion site before documented progression or death, censoring will occur at the last tumor assessment demonstrating progression-free status prior to that intervention. Participants with two or more consecutive missing response assessments before a visit where progression or death is recorded will be censored at the last assessment date at which they were documented as progression-free. This study was conducted in compliance with the ethical principles of the Declaration of Helsinki, and was approved by the Institutional Review Board (IRB) at each participating center, including: Institutional Review Board of Chinese Academy of Medical Sciences & Peking Union Medical College Hospital (Approval No. HS-1751). Institutional Review Board of Xuanwu Hospital, Capital Medical University (Approval No. [2019]035). Institutional Review Board of Peking University First Hospital (Approval No.[2019]171). Institutional Review Board of the Fourth Hospital of Hebei Medical University (Approval No. 2020014). All patients provided written informed consent before participating in the study. This clinical study was registered at ClinicalTrials.gov (NCT03885219). Participants Patients eligible for this study had pancreatic cancer classified as LAPC based on resectability status according to the National Comprehensive Cancer Network (NCCN) guidelines (Version 1, 2019 criteria). LAPC was defined as solid tumor contact >180° of the superior mesenteric artery (SMA) or celiac axis (CA) and unreconstructable superior mesenteric vein/portal vein (SMV/PV) or common hepatic artery. The inclusion criteria were as follows: (1) patients with good compliance and signing written informed consent; (2) age ≥18 years; (3) histological or cytological diagnosis of pancreatic adenocarcinoma; (4) imaging assessment of LAPC according to NCCN guidelines; (5) Eastern Cooperative Oncology Group (ECOG) performance status of 0-1; (6) measurable lesions according to Response Evaluation Criteria in Solid Tumors (RECIST v1.1); (7) estimated survival ≥3 months; (8) adequate hematological, hepatic and renal functions (including neutrophils ≥1.5×109/L, platelets ≥100×109/L, hemoglobin ≥ 90 g/L; serum creatinine ≤1.5 upper limit of normal value (ULN); AST and ALT ≤ 2.5 ULN; total bilirubin ≤ 1.5 ULN); and (9) no contraindications to the use of nab-paclitaxel or S-1. The exclusion criteria were as follows: (1) current or previous peripheral neuropathy of grade 2 or higher; (2) severe systemic infection or diseases; (3) allergy to or intolerance to chemotherapy drugs or their excipients; (4) antitumor treatment currently or within the previous 5 years; (5) within pregnancy or lactation; (6) digestive or metabolic dysfunction affecting absorption of S-1; (6) participation in other trial studies within 30 days; and (7) patients deemed by the investigators to be ineligible for the trial. Procedures We referenced the dosing regimens used in previous successful phase II clinical studies in locally advanced/metastatic pancreatic cancer11,21. Induction therapy with nab-paclitaxel and S-1 was administered on a 21-day cycle. Nab-paclitaxel was administered intravenously at 120 mg/m2 over 30 min on days 1 and 8. A daily dose of S-1 was defined based on the patient’s body surface area as follows: <1.25 m2, 80 mg/d; 1.25-1.5 m2, 100 mg/d; and >1.5 m2, 120 mg/d. S-1 was administered orally twice a day on days 1-14. Patients received treatment until disease progression or unacceptable toxicity. Every 2 cycles of chemotherapy regimen, patients were re-evaluated for response to treatment. The investigators were recommended the following subsequent antitumor therapies: continuation of the original regimen, concurrent chemoradiotherapy, and surgery; however, these were not mandatory. Patients showing a marked response to treatment were allowed to undergo surgery before finishing all eight cycles of SnP, and such cases were regarded as having completed the induction phase. During the induction phase, patients underwent contrast-enhanced CT or MRI scans at baseline, followed by scans every 42 days (+7 or -3 days). Evaluations will be conducted at a follow-up assessment conducted 28 days after the final treatment. During treatment selected by the investigator, imaging assessments were conducted every 56 days. A follow-up evaluation was carried out 28 days after the final dose of chemotherapy or chemoradiotherapy or 28 days after surgery. Patients were planned to undergo serial contrast-enhanced CT or MRI examinations, continuing until disease progression, consent withdrawal, loss to follow-up, or death, depending on which occurred first. In addition, survival status and any subsequent anticancer therapies were monitored every 90 days through telephone contact or review of medical records, and follow-up continued until death, withdrawal of consent, or loss to follow-up. All assessments were performed at the study center. Tumor response was determined by the investigator in accordance with RECIST version 1.1, and resection outcomes (R0, R1, or R2) were classified based on institutional criteria. Medical monitors and drug safety physicians continuously reviewed safety data throughout the study. Patients who discontinued treatment underwent a safety follow-up visit 28 days after stopping treatment. For patients who underwent chemoradiotherapy or surgery, safety assessments were performed both before the procedure and 28 days after completion of the treatment. Adverse events were graded according to version 5.0 of the Common Terminology Criteria for Adverse Events (CTCAE) by the National Cancer Institute. Statistical analysis The sample size was calculated for the estimated PFS at 6 months after treatment initiation for patients with LAPC. The null hypothesis was that the 6-month PFS rate for patients with LAPC was 30%. In the present study, the planned sample size was 52 patients, which was calculated as per an expected 6-month PFS rate of 48%, a two-sided alpha error of 0.05 and a power of 0.9. We hypothesized that SnP would not be inferior to single-agent S-1 or gemcitabine and would be comparable to GnP or GS. In the GEST trial,9 the 6-month PFS rates with S-1 and gemcitabine monotherapy were 26.9% and 29.8%, respectively; therefore, 30% was chosen as the null (ineffective) hypothesis. In MPACT,7 the 6-month PFS rate with GnP was 44%, and in GEST, the 6-month PFS rate with GS was 47.9%; accordingly, 48% was set as the alternative hypothesis (H1). A total of 58 patients were scheduled for enrollment, with a projected dropout rate of 10%. OS and PFS were evaluated using the Kaplan–Meier approach. Both OS and PFS were calculated from the initiation of treatment until the occurrence of the relevant event or the most recent follow-up. Efficacy assessments were conducted in the intention-to-treat population. The intention-to-treat population comprised all participants enrolled in the study, whereas the per-protocol population included those who met the eligibility criteria, received at least one dose of nab-paclitaxel or S-1, and had no major protocol deviations. The treated group included everyone who received at least one dose of nab-paclitaxel or S-1, which was used for all safety analyses. Supplementary information Statistical Analysis Plan (SAP) Study Protocol CONSORT 2010 Checklist Statistical Analysis Plan (SAP) Study Protocol CONSORT 2010 Checklist