Safety and efficacy of neoadjuvant chemoradiotherapy with moderately hypofractionated intensity‐modulated radiotherapy for resectable pancreatic cancer: A prospective, open‐label, phase II study

Abstract Background Resectable pancreatic cancer (RPC) is potentially resectable on admission, and the impact of neoadjuvant therapy on these tumors is controversial. Moreover, the safety and efficacy of neoadjuvant chemoradiotherapy with moderately hypofractionated intensity‐modulated radiation therapy (NACIMRT) for RPC have not been studied. Here, we conducted a phase II study to evaluate the safety and efficacy of hypofractionated NACIMRT for RPC. Methods A total of 54 RPC patients were enrolled and treated according to the study protocol. We used moderately hypofractionated (45 Gy in 15 fractions) IMRT with gemcitabine to shorten the duration of radiotherapy and reduce gastrointestinal toxicity. The primary endpoint was overall survival (OS), and we subsequently analyzed the microscopically margin‐negative resection (R0) rate, disease‐free survival (DFS), and histologic effects and safety of NACIMRT. Results Median OS for the cohort was 40.0 months. Forty‐two patients (77.8%) underwent pancreatectomy after NACIMRT. Median DFS was 20.3 months. The R0 resection rate was 95.2% (40/42) per protocol and 85.2% (46/54) for the cohort. There were no intervention‐related deaths during the study period. Local treatment response, as assessed by the CAP classification, showed no residual tumor in 4.8% of patients. Overall, 23.9% of patients experienced CTCAE grade 3 or 4 during NACIMRT. Adjuvant therapy was initiated in 88% of patients undergoing resection. Postoperative complications grade ≥3b on the Clavien–Dindo scale occurred in 4.8% of patients. CA19‐9 level at enrollment was an independent prognostic factor for OS and DFS. Conclusions This is the first prospective study of hypofractionated IMRT as neoadjuvant therapy for RPC. Hypofractionated NACIMRT for RPC could be safely introduced with a high induction rate of adjuvant chemotherapy, with an overall survival of 40.0 months.


| BACKGROUND
Pancreatic cancer has one of the poorest prognoses and highest mortality rates of any cancer. 1In Japan, pancreatic cancer is the fourth leading cause of cancer death, and its incidence is increasing.Although surgical resection results in a high cure rate, pancreatic cancer often recurs, and the 5-year recurrence-free survival rate in Japan is 15%. 2 As defined by the National Comprehensive Cancer Network, pancreatic cancer is classified as resectable pancreatic cancer (RPC), borderline resectable pancreatic cancer (BRPC), and unresectable pancreatic cancer. 3RPC is considered if it meets the following criteria: (1) no arterial tumor contact with the celiac axis, superior mesenteric artery (SMA), or common hepatic artery (CHA); and (2) no tumor contact with the superior mesenteric vein (SMV) or portal vein or no more than 180° contact without venous contour irregularity.Neoadjuvant therapy for RPC is controversial because the lesions are resectable on admission.A recent randomized controlled study showed the benefits of neoadjuvant chemotherapy with gemcitabine plus S-1 4 ; however, another analysis showed a survival rate comparable to the non-neoadjuvant approach in RPC in the PREOPANC trial. 5In addition, several studies are retrospective in nature and do not have information on how many RPC patients received neoadjuvant therapy, 6 often resulting in an analysis of post-resection survival rates rather than an intention-to-treat analysis.
Intensity-modulated radiation therapy (IMRT) allows higher doses to be focused on an area with minimal exposure of normal tissue, thus maximizing the effect on the target area with minimal toxicity to surrounding organs.IMRT has been used for prostate and nasopharyngeal cancers, but precise irradiation is difficult because intraabdominal tissues move with respiration.In addition, there is little evidence to support the efficacy of IMRT for pancreatic cancer.Recently, moderately hypofractionated irradiation has been employed for breast 7 and prostate cancers. 8The benefits of neoadjuvant hypofractionated irradiation include a shortened radiotherapy period, allowing patients to undergo surgery before adverse events (AEs) occur.We have reported that neoadjuvant with moderately hypofractionated IMRT results in favorable outcomes with minimal toxicity in BRPC patients 9 ; however, the efficacy and safety of this technique in RPC patients have not been studied.Therefore, we prospectively conducted a phase II study to evaluate the safety and efficacy of neoadjuvant chemoradiotherapy with gemcitabine in combination with hypofractionated IMRT (NACIMRT) for RPC.

| Study design and patients
This study was conducted as a prospective phase II study of neoadjuvant chemotherapy with IMRT plus gemcitabine (jRCT1051190064) for RPC.The primary endpoint was overall survival (OS), and the secondary endpoints were the microscopically margin-negative resection (R0) rate, diseasefree survival (DFS), and histologic effects and safety of NACIMRT.All patients diagnosed with RPC according to the NCCN 2016 guidelines 3 at our hospital between June 2016 and August 2021 who provided informed consent were enrolled in this study.RPC status was assessed by multidetectorrow computed tomography (MDCT) using a multiphase contrast-enhanced technique and evaluated by a multidisciplinary team consisting of physicians from the departments of surgery, gastroenterology, radiation oncology, clinical oncology, and diagnostic imaging.We retrospectively reviewed pre-NACIMRT CT images of eligible patients and classified them according to the NCCN 2023 guidelines. 3All patients fell into the RPC category of the NCCN 2023 guidelines.
Inclusion criteria for this study were RPC patients with no arterial tumor contact with the CA, SMA, or CHA, and no tumor contact with the SMV or portal vein or contact of not more than 180° without venous contour irregularity on contrast-enhanced MDCT.Other inclusion criteria were as follows: histologically confirmed pancreatic ductal adenocarcinoma; age between ≥20 and <80 years; ECOG performance status of 0 or 1; no distant metastases in the thorax, abdomen, or pelvis on dynamic contrast-enhanced MDCT, positron emission tomography with 2-deoxy-2-[ 18 F]fluoro-D-glucose (FDG-PET), and magnetic resonance imaging Conclusions: This is the first prospective study of hypofractionated IMRT as neoadjuvant therapy for RPC.Hypofractionated NACIMRT for RPC could be safely introduced with a high induction rate of adjuvant chemotherapy, with an overall survival of 40.0 months.

K E Y W O R D S
chemoradiotherapy, IMRT, intensity-modulated radiotherapy, neoadjuvant therapy, pancreatic cancer, surgery (MRI) with contrast medium of gadolinium-ethoxybenzyl diethylenetriamine pentaacetic acid (EOB-MRI); no pretreatment for current pancreatic cancer; and no hematologic or major organ dysfunction.
Exclusion criteria were as follows: interstitial pneumonitis; history of irradiation to the upper abdomen; severe comorbidities (heart failure, renal failure, liver failure, bleeding peptic ulcer, intestinal paralysis, intestinal obstruction, and uncontrolled diabetes); moderate or severe ascites or pleural effusion; history of active cancer (concurrent multiple cancers or heterogeneous multiple cancers with a disease-free interval of <3 years); or expectant or nursing women.The Kyoto University Ethics Committee approved this study (YC1171), and each patient gave informed consent prior to participation.

| Neoadjuvant chemoradiotherapy
The treatment flow of NACIMRT is shown in Figure 1.Following histologic confirmation of adenocarcinoma, patients were initially treated with gemcitabine at a dose of 1000 mg/m 2 three times (Days 1, 8, and 22) before chemoradiotherapy.At the start of radiotherapy, gemcitabine was administered at a dose of 1000 mg/m 2 (Days 1, 8, and 22) concurrently with IMRT.If grade 4 or worse neutropenia or thrombocytopenia occurred, chemotherapy was discontinued for 1 week.For IMRT planning, the gross tumor volume (GTV) included pancreatic tumors and lymph nodes >1 cm in diameter.The clinical target volume (CTV) included the retropancreatic regions between the CA and SMA in addition to the GTV plus a 5-mm margin, according to our institutional contouring guidelines.The planning target volume (PTV) was defined as the CTV with a 5-mm margin in all directions.GTV-PRV and PTV-PRV were volumes calculated by subtracting the stomach plus 10-mm and the duodenum plus 5-mm margins from the GTV and PTV, respectively.Organs at risk of injury were the stomach, duodenum, small intestine, colon, liver, kidneys, and spinal cord, which were delineated on expiratory-phase CT.The prescribed dose was 45 Gy as D50% of GTV-PRV and 42 Gy as D95% of PTV-PRV (Dxx: The dose covering xx% of the target structure) in 15 fractions; PTV D98% ≥36 Gy was preferred.
Intensity-modulated radiation therapy was performed with a treatment plan optimized for each patient.The breath-hold method was used to minimize tumor motion, and cone-beam CT was used before each treatment to determine daily setup errors.Radiation treatment was delivered using volumetric modulated arc therapy techniques.

| Resection and adjuvant chemotherapy
Patients were evaluated for feasibility of resection within 4 weeks of neoadjuvant therapy using MDCT, EOB-MRI, and FDG-PET.In the absence of unresectable factors such as distant metastases, resection was attempted between 4 and 8 weeks after completion of neoadjuvant radiotherapy.Pancreaticoduodenectomy, distal or total pancreatectomy, and resection of all involved tissues were performed.Operative findings, surgical complications, and histopathology were recorded.S-1 at a dose of 80 mg/m 2 /day was administered on Days 1-28 of a 42-day cycle for 6 months as adjuvant chemotherapy, starting 4-8 weeks after resection.

| Assessment
Resection margins were considered positive (R1) if malignant cells were observed on the surface of the resected specimen (0-mm margin rule), in the plexus around the SMA or CHA, duodenum, bile duct, or retroperitoneal tissue.If a vein was concomitantly resected, the vein margin was also examined.
Follow-up data up to December 2022 were analyzed.Patients were evaluated by contrast-enhanced CT every 3 months for the first 2 years and every 6 months thereafter.A new low-density mass in the peripancreatic and mesenteric root area was considered a locoregional recurrence.DFS was calculated as the time from the date of resection to recurrence and included off-protocol patients who underwent resection (n = 7).OS was calculated as the time from enrollment to death.Radiological tumor response was evaluated by comparing tumor size on MDCT images before and after NACIMRT according to the Response Evaluation Criteria in Solid Tumors (RECIST v1.0), and pathological responses were evaluated by CAP grade.Toxicity events were recorded using the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. 10Weekly complete blood counts and liver function tests were performed from the start of radiotherapy until 2 weeks after the end of chemoradiotherapy.Pretreatment serum carbohydrate antigen 19-9 (CA19-9) levels were assessed after biliary drainage.No eligible patients were lost to follow-up during the study period.

| Statistics
The primary endpoint of OS and the secondary endpoints of DFS, response rate, pathologic response, R0 resection rate, surgical morbidity, and acute and late toxicity were evaluated 6 months after the completion of enrollment.The final dataset was carefully reviewed for clerical errors by three physicians (T.M., K.N., and T.A.).Data for continuous variables were expressed as median and range.Fisher's exact test and the Mann-Whitney U test were used for categorical and continuous data, respectively.Kaplan-Meier curves were created to estimate OS, and comparisons between groups were estimated using log-rank tests.Multivariate Cox proportional hazards regression analysis was used to identify prognostic factors independently associated with survival.p < 0.05 was considered significant.In the regression analysis model for OS, multivariate analysis was performed using diameter, CEA, and CA19-9, which have been reported as factors associated with OS.For DFS, multivariate analysis was performed using the forward selection method, and variables with p values less than 0.1 in univariate analysis were entered in addition to previously reported factors.All statistical analyses were performed on an intention-to-treat basis using JMP version 16.0 software (SAS Institute).

| Patient characteristics
Fifty-four patients were enrolled in this study.Baseline characteristics of the cohort are summarized in Table 1A.

| Safety and clinical outcomes of neoadjuvant chemoradiotherapy with moderately hypofractionated intensity-modulated radiation therapy
The study diagram is shown in Figure 2. The median time from staging MDCT to the start of neoadjuvant induction chemotherapy with gemcitabine was 14 days (range, 2-21).Of the 54 patients, 46 completed IMRT with gemcitabine (85.2%, 46/54).Seven patients discontinued due to severe bone marrow suppression and one patient due to cholangitis; seven patients underwent upfront surgery, and one patient was switched to gemcitabine with S-1 due to liver metastases after neoadjuvant therapy.There were two patients with grade 3 leukopenia and six patients with grade 3 neutropenia; one patient with grade 3 leukopenia recovered after omitting concurrent gemcitabine, and the other seven patients did not recover within 2 weeks and underwent upfront resection.There were also one patient with grade 3 fatigue who received continuous infusion, one patient with grade 3 cholangitis who refused to continue NACIMRT and underwent chemotherapy, and one patient with grade 3 pancreatitis who developed during NACIMRT but recovered after the replacement of the stent in the pancreatic duct.The median relative dose intensity of gemcitabine was 83.5%.The incidence of grade 3/4 toxicity for NACIMRT was 16.7% (9/54; Table 1B).There were no cases of grade 3/4 gastroenterological toxicity.

| Surgical outcomes and pathological effects
The median interval from completion of IMRT to surgery was 26 days (range, 20-48 days).Of the 45 patients who underwent surgery, one had a positive washing cytology, and two had positive distant lymph node metastasis in the para-aortic region on laparotomy, resulting in 42 patients (77.8%) undergoing pancreatectomy.In the 25 patients with pancreatic head cancer, 12 underwent biliary drainage with metallic stent, with no differences in operative time (521.7 min.vs. 517.8min, p = 0.926), blood loss (577 mL vs. 691 mL, p = 0.565), or postoperative hospital stay (32.3 days vs. 23.4days, p = 0.149) between patients with metal stent and those without.Clavien-Dindo 3a or higher was observed in one case in the metal stent group and in two cases in the non-stent group, with no significant difference (p = 0.085).R0 resection was achieved in 40/42 patients (95.2%;Table 2A).The overall R0 resection rate was 85.2% (46/54 patients including seven who did not complete NACIMRT and underwent pancreatectomy).According to the CAP classification, no residual tumor, <10% residual tumor, and >50% residual tumor were observed in 2 (4.7%), 7 (16.7%),and 23 (54.8%) patients, respectively.Pathologic lymph node metastasis was observed in 10 patients (23.8%).
Median albumin and cholinesterase levels were within normal ranges 1 month after surgery in the per-protocol cohort.Of the 42 patients with R0/1 resection after completion of NACIMRT, 37 (88.1%)started postoperative adjuvant therapy with S-1 within 6 weeks of resection (Table 2A); 35 (83.3%) completed 6 months of S-1 treatment; and 26 (61.9%) received 100% of the S-1 relative dose intensity.Five patients could not start adjuvant chemotherapy, two of whom had clinically relevant pancreatic fistulas; two refused to take S1; and one had delayed gastric emptying.

| Survival
Kaplan-Meier plots for survival are shown in Figure 3.The median follow-up for censored patients was 31.5 months.Median OS and DFS were 40.0 months (95% confidence interval [CI]: 36.5-n.a.; Figure 3A) and 20.3 months (95% CI: 13.1-30.8;Figure 3B), respectively.OS and DFS of patients with normal CA19-9 levels were significantly longer than those with elevated CA19-9 levels (OS: 65.3 vs. 38.0months, p = 0.030; DFS: not reached vs. 14 months, p = 0.001; Figure 3C,D).To elucidate the independent factors related to prognosis, we analyzed factors affecting OS and identified normal serum CA19-9 levels before NACIMRT to be independently associated with OS.For factors associated with DFS, univariate analysis showed that a normal initial CA19-9 level, tumor diameter >20 mm, pathological N0 status, and R0 status were associated with DFS; however, only a normal initial CA19-9 level and pathological N0 status were independent factors for DFS (Table 3).

| DISCUSSION
This study prospectively evaluated the safety and efficacy of neoadjuvant gemcitabine with moderately hypofractionated IMRT in 54 RPC patients who had undergone rigorous evaluation for resectability and distal metastases using EOB-MRI, dynamic CT, and FDG-PET.In our cohort, 46 patients completed IMRT with concurrent gemcitabine with no grade 4 or 5 AEs.In terms of efficacy, the median survival time (MST) of OS was 40.0 months, and R0 was achieved in 95.2% (40/42 patients) of the perprotocol cohort and 85.2% (46/54 patients) of intentionto-treat analysis (ITT) including patients who did not complete NACIMRT.These results are better than those of our previous phase II study of gemcitabine plus IMRT as neoadjuvant therapy for BRPC-A, in which 93.1% of the per-protocol cohort achieved R0 resection and 55.3% of patients in ITT. 9 In contrast to postoperative adjuvant therapy with various regimens for RPC, only a few prospective clinical studies have been conducted on neoadjuvant therapy for RPC due to the possibility of tumor resection on admission  2) 0 (0) 0 (0) 0 (0) 0 (0) Wound infection 0 (0) 1 ( 2) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Intra-abdominal abscess 0 (0) 1 ( 2) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Abdominal bleeding 0 (0) 0 (0) 0 (0) 1 ( 2) 0 (0) 0 (0) 0 (0) Cholangitis 0 (0) 3 ( 7) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Chylous ascites 2 (5) 2 ( 5) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Diarrhea 0 (0) 2 ( 5) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Portal vein embolism 0 (0) 1 ( 2 and concerns about disease progression after neoadjuvant therapy.Nevertheless, several neoadjuvant therapies have been conducted to date to improve survival in RPC patients.In these clinical studies, chemotherapy or chemoradiotherapy has been administrated.In the SWOG S1505 trial, in which FOLFIRINOX or gemcitabine plus nabpaclitaxel was used as neoadjuvant therapy, 76% of initial patients underwent resection, and 85% of resected specimens were negative for R0 margins.Seventy-eight percent of patients started and 63% completed adjuvant therapy. 11n another study, 182 RPC patients received gemcitabine plus S1 as neoadjuvant therapy, 86% of patients underwent resection, and a median OS of in ITT was 36.7 months. 4n the HOPS-R01 trial using S-1 as neoadjuvant therapy, 63% (31/49 patients) underwent pancreatectomy with a median OS of 35.5 months and ITT-based progression-free survival of 17.4 months. 12Regarding neoadjuvant chemoradiotherapy for RPC, Eguchi et al. 13 applied 50.4 Gy of three-dimensional (3D) conformal radiotherapy with gemcitabine plus S1 as neoadjuvant therapy to 62 RPC patients, 87% of patients were resected, and median OS was 55.3 months in ITT.In the PREOPANC study, a randomized phase III study of neoadjuvant gemcitabine plus radiotherapy for RPC and BRPC, hypofractionated 3D conformal radiotherapy (36 Gy in 15 fractions) with gemcitabine demonstrated a 66% resection rate with a median OS of 14.6 months in RPC patients. 5is study was conducted with the same purpose as the other studies of improving survival, with some differences in the approach to neoadjuvant therapy for RPC, and our approach has not been previously studied.We used IMRT to deliver therapeutic doses to affected tissue with minimal irradiation to surrounding normal organs compared to conventional radiotherapy.IMRT may have contributed to the low incidence of AEs in this study, resulting in a high completion rate of neoadjuvant chemoradiotherapy (85.2%, 46/54 patients).We used moderately hypofractionated schedules, although SBRT (5 fractions) or conventionally fractionated radiotherapy (25-30 fractions) has been used in neoadjuvant radiotherapy for pancreatic cancer. 14Hypofractionated irradiation may have contributed to the high completion rate in this study because it shortened the duration of irradiation, thereby reducing the incidence of AEs by chemotherapy and allowing patients to proceed to surgery more quickly.The incidence of gastroenterological AEs was low in this study, and the types of AEs appeared to be slightly different from other approaches.In the HOPS-R01 trial using neoadjuvant S-1, 12 the proportion of patients with grade 3 gastroenterological toxicity (25%) was higher than that in this study (0%).In this study, 16% of patients experienced grade 1 to 2 abdominal pain, which could have been caused by IMRT.In the PREP-02 phase III trial using neoadjuvant gemcitabine and S-1, 4 median OS was 39.2 months and no abdominal pain was observed, but grade 3 neutropenia developed in 35% of patients (11% in this study).
Besides the purpose of reducing toxicity, we used IMRT to increase the irradiation dose to the primary tumor.The biological effective dose at α/β of 10 (BED10) of 45 Gy in 15 fractions was 58.5 Gy, while the BED10 of the conventional standard treatment dose (45-50.4Gy in 1.8-Gy fractions) was 53.1-59.5 Gy.Given the shorter treatment duration but higher doses, the treatment intensity of hypofractionated therapy should be higher than that of standard treatment.We calculated the α/β of pancreatic cancer to be 10, which has not been established in the literature.If the α/β of pancreatic cancer were <10, the benefits of hypofractionated therapy would be even greater.Furthermore, the use of IMRT increased treatment intensity compared to the PREOPANC trial, which used conventional 3D conformal radiotherapy with 36 Gy in 15 fractions (BED10: 44.6 Gy).Moreover, our current strategy of increasing treatment intensity while minimizing AEs not only increases the R0 resection rate, but also increases the likelihood of maintaining local control even in inoperable cases, as shown by the favorable local control rate (1-year locoregional progression-free survival: 73.1%) in our previous study of hypofractionated IMRT for locally advanced pancreatic cancer. 15djuvant therapy with S-1 is important to improve the long-term survival of RPC patients.The JASPAC-01 study with upfront surgery showed better survival (OS, 46.5 months in MST) with S-1 compared to gemcitabine. 16n this study, postoperative adjuvant S-1 therapy was administered to 88% of patients who underwent resection; 83.3% completed 6 months of adjuvant S-1 therapy with the median relative dose intensity of 100%.Our current S-1 induction/completion rates were higher, compared to those of other GS-based neoadjuvant chemoradiation studies for RPC 13 (70.3% and 55.5%, respectively).
In this study, elevated CA19-9 levels (>37 U/mL) at enrollment were independently associated with OS and DFS.Recently, the term bio-borderline RPC has been proposed to describe tumors with elevated CA19-9 levels (>37 U/mL) at clinical stages I-II (TNM) that are anatomically resectable but have a poor prognosis. 17,18The present data and this concept suggest that initial CA19-9 is also important in RPC patients receiving neoadjuvant chemoradiotherapy.Furthermore, it suggests that the differences in initial CA19-9 levels between cohorts may influence the outcome of clinical trials.Some studies argue that normalization of CA19-9 levels after neoadjuvant therapy has been considered important for long-term survival after resection. 19In our cohort, initial high CA19-9 levels were more significantly associated with survivals rather than normalized CA19-9.Because high CA19-9 levels at enrollment were associated with those after neoadjuvant therapy (data not shown), our current neoadjuvant chemoradiotherapy may not be intense enough to normalize CA19-9 levels regardless of the initial level.This is consistent with our previous study for BRPC with arterial abutment 9 with the same neoadjuvant NACIMRT treatment, in which a serum CA19-9 level ≥400 IU/mL was an independent prognostic factor but normalized CA19-9 was not.
This study has several limitations.First, this was a single-center study; standardizing IMRT treatment in a multicenter study is difficult because GTV and CTV must be designed as standard procedures according to tumor location, size, and vascular anatomy.Second, the current treatment protocol uses gemcitabine monotherapy as concurrent therapy, which is outdated for the newer doublet or triplet combination therapy.Given the recurrence rate of distant metastases (21/23, Table 2A), more potent chemotherapy, such as gemcitabine/nab-paclitaxel, in combination with less toxic IMRT may be more effective.Third, because this is not a randomized clinical trial, it is difficult to compare results with other therapies.Future welldesigned randomized controlled trials should be prepared.

| CONCLUSIONS
This is the first report of neoadjuvant with moderately hypofractionated IMRT for RPC, resulting in a median OS of 40.0 months and a median DFS of 20.3 months.RPC patients who had received neoadjuvant hypofractionated IMRT had high induction and completion rates of postoperative adjuvant chemotherapy, which may have an impact on post-resection survival.

F I G U R E 2
CONSORT diagram of the study flow.A total of 46 patients (85.2%) completed neoadjuvant chemoradiotherapy with moderately hypofractionated intensity-modulated radiation therapy (NACIMRT) with gemcitabine, and 77.8% (42/54) of the patients underwent pancreatectomy after NACIMRT.T A B L E 1 B Common Terminology Criteria for AdverseEvents grade for neoadjuvant chemoradiotherapy with moderately hypofractionated intensity-modulated radiation therapy comorbidities (N = 54).

F
I G U R E 3 Intention-to-treat analyses of (A) overall survival and (B) disease-free survival in 54 resectable pancreatic cancer patients.(C) Overall survival and (D) disease-free survival according to CA19-9 levels at enrollment.Patients with CA19-9 level > 37 U/mL had significantly worse overall survival (p = 0.030) and disease-free survival (p = 0.007) using the log-rank test.
Surgical outcome and pathological features.