Neoadjuvant Chemotherapy With Chemoradiotherapy for Patients With Borderline Resectable or Locally Advanced Pancreatic Ductal Adenocarcinoma—Retrospective Review From a Tertiary Care Hospital

ABSTRACT

Aim

Patients diagnosed with borderline resectable pancreatic cancer (BRPC) or locally advanced pancreatic cancer (LAPC) have historically worse survival rates compared to those with resectable pancreatic cancer. The study aimed to assess the feasibility and efficacy of neoadjuvant chemotherapy and chemoradiotherapy in BRPC/LAPC. Additionally, we evaluated the R0 resection rates for patients who progressed to surgery.

Methods

This retrospective study included patients diagnosed with BRPC/LAPC between January 2019 and December 2023 at The Canberra Hospital (TCH), a tertiary care setting. A total of 115 patients were screened, of whom 37 were eligible for inclusion. Demographic data, CA19‐9 levels, treatment regimens, surgical outcomes, resection rates, disease‐free survival (DFS), and overall survival were analysed.

Results

A total of 20 (54%) patients (15 FOLFIRINOX and 5 gemcitabine/nab‐paclitaxel) completed their planned chemotherapy, and 17 (46%) of these patients had chemoradiotherapy, majority receiving 45–50 Gy of conventional radiation with capecitabine. Tumor marker Ca19‐9 normalized after chemotherapy in seven patients (19%). In total, 23 patients (62.2%) progressed to surgery. The median DFS for all patients was 12.7 months (95% CI 5.5–15.9), and the median OS was 21 months (95% CI 13.7–44.9).

Conclusion

This study suggests that neoadjuvant treatment is feasible for BRPC/LAPC, allowing patients to undergo surgery and achieve R0 resection. However, further randomized controlled trials with larger cohorts are needed to validate these findings and refine treatment protocols.

A single‐centre retrospectively analyzed 37 BRPC/LAPC patients receiving neoadjuvant chemotherapy ± chemoradiotherapy. Surgical resection was feasible in 62% of patients, with 38% of patients achieving R0 resections. The median overall survival was 21 months, and disease‐free survival was 12.7 months. The neoadjuvant protocol facilitates surgical resection in BRPC/LAPC.

1. Introduction

Pancreatic ductal adenocarcinoma (PDAC) is a devastating cancer with the eighth‐highest incidence rate (17.1 per 100,000 cases) and the fourth‐highest mortality rate (13.9 per 100,000 cases) within the Australian population [1]. Despite these alarming statistics, the primary curative treatment for PDAC is surgical resection. Among newly diagnosed PDAC patients, 20% present with resectable disease, 30% present with BRPC/LAPC, and 50% present with metastatic pancreatic cancer [2]. Even with curative intent surgery, prognosis remains poor: The median survival for resectable and non‐resectable localized PDAC is limited to 32 and 12 months, respectively [3]. The 5‐year survival for those eligible for surgery is only 20%–21%, whereas for those with metastatic disease is just 3% [4].

According to NCCN guidelines, PDAC resectability is defined based on the extent of encasement or abutment of nearby vessels [5]. BRPC has solid tumor contact with the superior mesenteric vein (SMV) of > 180° or contact of ≤ 180° if complete resection and vein reconstruction can be completed. With the superior mesenteric artery (SMA), there is ≤ 180° solid tumor contact. LAPC has occlusion of the SMV and would have incomplete vessel reconstruction. With the SMA, there is > 180° solid tumor contact.

The European Society of Medical Oncology (ESMO) guidelines for BRPC/LAPC recommend that patients consider a sequence of neoadjuvant therapy prior to surgery, followed by adjuvant therapy after curative intent resection. Disease resectability depends on various factors, including response to neoadjuvant therapy and the extent of disease following initial treatment [6].

One approach used in the neoadjuvant setting is the FOLFIRINOX (fluorouracil, oxaliplatin, leucovorin, and irinotecan) regimen. Findings of the ESPAC5 trial, a multicenter study comparing upfront surgery to neoadjuvant therapies (FOLFIRINOX, gemcitabine plus capecitabine, or chemoradiotherapy with capecitabine), demonstrated that 1‐year DFS was more significant in patients who received neoadjuvant therapies. The 1‐year disease‐free survival (DFS) was 33% (95% CI 19–58) for patients who underwent upfront surgery compared to 59% (46–74) for patients in the combined neoadjuvant therapies group (hazard ratio 0.53 [95% CI 0.28–0.98], p = 0.016). In this trial, there was no difference in resection rates for the cohort who had upfront surgery compared to the neoadjuvant therapy cohort [7].

The role of radiotherapy in managing patients with BRPC/LAPC is still evolving, with previous studies showing mixed results in improving outcomes [8, 9]. The A0121501 Phase II randomized controlled trial (RCT), which compared neoadjuvant mFOLFIRINOX to mFOLFIRINOX plus hypofractionated radiotherapy in 126 patients with BRPC, found an 18‐month overall survival (OS) rate of 68% for the mFOLFIRINOX group and 47% for the combination group across all patients based on intention to treat. However, patients who progressed to surgery after receiving either of the two neoadjuvant treatments had higher 18‐month OS rates: 88% for mFOLFIRINOX and 79% for the mFOLFIRINOX plus radiotherapy group [8]. Similarly, an earlier Phase II study involving 48 patients with BRPC/LAPC who received eight cycles of neoadjuvant FOLFIRINOX followed by individualized short‐ or long‐course chemoradiotherapy reported that 32 patients progressed to surgical resection, with 31 achieving an R0 resection rate [8]. These results laid the foundation for the current treatment protocol for BRPC/LAPC patients at The Canberra Hospital (TCH) [9].

The primary aim of this study was to investigate the feasibility and completion rates of planned chemotherapy/chemoradiotherapy, the ability of the neoadjuvant approach to improve resectability and R0 resection rates, and the analysis of adverse events during chemotherapy. Secondary outcomes were the biochemical response after chemotherapy, DFS, and OS.

2. Methods

We conducted a retrospective review of all patients diagnosed with PDAC between January 1, 2019, and December 31, 2023, at TCH, and patient data were extracted from the electronic medical record (EMR) and the PURPLE pancreatic registry [10]. A total of 115 patients' data were extracted and screened for further analysis. The inclusion criteria were patients diagnosed with BRPC/LAPC between 2019 and 2023 who received care in TCH and were deemed suitable for a neoadjuvant chemotherapy approach based on multidisciplinary assessment. Patients with diagnosis of pancreatic cancer before 2019, presenting with upfront resectable disease; diagnosed with metastatic disease, BRPC, or LAPC with multiple comorbidities; and poor performance status deemed unsuitable for neoadjuvant approach—were excluded.

This data included demographic information (age, gender, Eastern Cooperative Oncology Group (ECOG) performance status, and Charlson Comorbidity Index [CCI]), tumor marker CA19‐9 levels, preoperative laparoscopy rates, neoadjuvant treatments used, completion rates of planned treatments, granulocyte colony‐stimulating factor (G‐CSF) use, rate of surgical management, and R0 resection rate. This project was approved by the ACT Health Human Research Ethics Committee (2019.STE.0215).

Diagnosis with either BRPC or LAPC was based on computed tomography (CT) imaging as per NCCN guidelines [5]. This consisted of a staging CT contrast scan of the abdomen and chest. An optional positron emission tomography (PET) or magnetic resonance imaging (MRI) was done at the physician's discretion or recommendations at the multidisciplinary meeting (MDM). Baseline laboratory results, including tumor marker CA19‐9 levels, were also collected. Histological or cytological confirmation of the diagnosis was obtained through endoscopic ultrasound (EUS)‐guided biopsy. Staging laparoscopy was performed selectively to exclude intra‐abdominal metastases. All patients were discussed at the gastrointestinal cancer MDM to establish the staging, resectability, and consensus management plan. Patients presenting with obstructive jaundice underwent an endoscopic retrograde cholangiopancreatography (ERCP) with stent placement. OS was defined as the time from diagnosis to death or the last clinic review for still‐alive patients. DFS was defined as the duration of time from diagnosis without evidence of disease progression, local recurrence, or distant metastasis.

Continuous data were expressed as means and standard deviations (medians and interquartile [IQR] ranges). Differences in continuous variables were assessed using the Student's t‐test or Mann–Whitney U test. Pearson's chi‐square test (or Fisher's exact test) was used to compare proportions. Survival was determined by the Kaplan–Meier method, and differences in survival were analyzed using the log‐rank Wilcoxon–Breslow–Gehan test. A significance level of 0.05 was used for all analyses. All data management and statistical analyses were executed in SAS software.

3. Results

Between January 1, 2019, and December 31, 2023, 37 patients met the inclusion criteria for diagnosis of BRPC/LAPC (Figure 1). There was a higher percentage of females than males (43% male and 57% female). The mean age at diagnosis was 65 (39–84) years old. The most common location for a pancreas tumor was the head of the pancreas in 32 patients (86%), followed by the body in 3 patients (8%), and the tail in 2 patients (5%). All patients in this cohort had an ECOG performance status of 0 or 1. The mean Charlson's Comorbidity Index (CCI) was 3 (Table 1). A total of 31 (83%) patients presented with an elevated CA19‐9 level (>37 kU/L) at diagnosis.

FIGURE 1.

Flowchart for participation criteria in this review.

TABLE 1.

Demographic data.

Total number of patients (n = 37)
BRPC	32
LAPC	5
Male to female ratio	16:21 (43%:57%)
Age (mean, years)	65
Age range (years)	39‐84
Charlson Comorbidity Index (CCI) status mean	3.22
0	1
1	3
2	10
3	9
4	6
5	5
6	2
7	0
8	1
Location of tumor
Head	32 (86%)
Body	3 (8%)
Tail	2 (5%)
Tumor marker CA 19.9
Elevated CA19‐9 before/at diagnosis (> 37 kU/L) (n, %)	31 (83%)
Normal CA19‐9 before/at diagnosis (> 37 kU/L) (n, %)	5 (14%)
Normal CA19‐9 following neoadjuvant treatment (> 37 kU/L) (n, %)	7 (19%)
Staging laparoscopy
Yes	28 (76%)
No	9 (24%)

All 37 patients started neoadjuvant treatment, with FOLFIRINOX being the most used regimen (32 patients, 86%), followed by gemcitabine/nab‐paclitaxel (5 patients, 14%) (Table 2).

TABLE 2.

Neoadjuvant therapy and surgery data.

Chemotherapy agents
FOLFIRINOX	32 (86%)
Gemcitabine and nab‐paclitaxel	5 (14%)
Completed all planned cycles of neoadjuvant treatment (n = 20; 15 FOLFIRINOX and 5 gemcitabine and nab‐paclitaxel)
Progressed to surgery	17 (85%)
No surgery	3 (13%)
G‐CSF received during chemotherapy
Yes	31 (84%)
No	4 (11%)
Not collected	2 (5%)
Didn't complete planned neoadjuvant chemotherapy (n = 16)
Progression of disease	9 (56%)
Eligible for surgery	6 (38%)
Died	1 (6%)
Number received chemoradiotherapy 17 (46%) (all patients except 1 proceeded to surgery)
Concurrent chemotherapy agents
Capecitabine	14 (82%)
5‐Fluorouracil	3 (18%)
Surgery (n = 23, 64%)
Unable to proceed with surgery	3 (13%)
Pancreatoduodenectomy	17 (74%)
Total pancreatectomy	2 (9%)
Distal pancreatectomy	1 (4%)
Resection status
R0	14 (70%)
R1	6 (30%)

A total of 20 patients (54%) completed all planned cycles of neoadjuvant chemotherapy. One patient, who developed toxicity after two cycles of FOLFIRINOX, completed four cycles of gemcitabine/nab‐paclitaxel instead (this patient was included in the gemcitabine/nab‐paclitaxel group).

Sixteen patients (43%) did not complete the planned chemotherapy. Among these, nine patients (56%) had disease progression, six (38%) proceeded to surgery, and one (6%) died. Of these 16 patients, 5 required hospital admission due to various chemotherapy‐related toxicities.

A total of 17 (46%) patients started chemoradiotherapy. The majority of patients received long‐course chemoradiation with 45–50.4 Gy in 1.8–2 Gy fractions over 5–5.5 weeks. One patient was planned for chemoradiotherapy but had disease progression to the lungs before the start of chemoradiation. Capecitabine (n = 14, 82%) was the most common chemotherapy agent, followed by 5‐fluorouracil (n = 3, 18%) (Table 2).

A total of 23 patients (64%) progressed to surgery. However, three (13%) of these surgeries were discontinued. Intraoperatively, one patient was found to have liver metastases, while the other had peritoneal metastasis. The details of the third patient's inoperability are unknown. All three (13%) patients completed all cycles of FOLFIRINOX and chemoradiotherapy.

Of the six patients who did not complete the planned number of neoadjuvant chemotherapy cycles but proceeded to surgery, two had completed seven cycles of chemotherapy but discontinued treatment after being hospitalized for toxicities. These two patients subsequently underwent chemoradiation therapy before surgery. The remaining four patients, who experienced only minor chemotherapy‐related toxicities, did not receive chemoradiation and proceeded directly to surgery. The mean and median number of cycles of neoadjuvant chemotherapy completed before discontinuation for these 16 patients was 4.5 and 5 cycles, respectively.

In this analysis, five patients (14%) developed neutropenia following neoadjuvant chemotherapy, which led to a dose reduction in one of these patients. Due to the addition of G‐CSF, none of the patients suffered from any neutropenic sepsis events while on therapy.

The other toxicity in our cohort of patients was that seven patients (19%) experienced peripheral neuropathy, and seven patients (19%) developed diarrhea. Biliary sepsis and mucositis were each observed in two patients (5%). A total of 10 patients (27%) required hospital admissions during treatment. Four patients (40%) were admitted due to diarrhea‐related complications. One patient died from hemorrhagic shock due to an upper gastrointestinal bleed.

Following neoadjuvant treatment, 7 patients (19%) had normalized CA19‐9 levels, while a further 13 patients (35%) showed a reduction in CA19‐9 levels, although they did not reach the normal range. A total of five patients (14%) had normal CA19‐9 levels at diagnosis. Staging laparoscopy was conducted in 28 (76%) patients as part of the pancreatic protocol.

There were five (13.5%) patients with LAPC in our series, four (80%) received FOLFIRINOX, and one (20%) received gemcitabine and nab‐paclitaxel. Three of these patients had staging laparoscopies with no evidence of peritoneal disease. One patient in the FOLFIRINOX group had normalization of Ca19‐9 and underwent R0 resection and was alive at the final follow‐up; one patient on gemcitabine and nab‐paclitaxel had R1 resection and was lost to follow‐up. Both patients received a neoadjuvant chemoradiation protocol. The remaining three patients died due to progressive disease.

Of the 20 patients who completed surgical resections, 14 (70%) had R0 resections, and 6 (30%) had R1 resections. Nine (70%) of the 14 patients with R0 resections received chemoradiation, and 4 (67%) of the 6 patients with R1 resections received chemoradiation.

The DFS and OS for the 37 patients were 12.77 months (95% CI 5.52–15.90) and 20.96 months (95% CI 13.67–44.94), respectively.

4. Discussion

This retrospective dataset evaluated the feasibility of neoadjuvant chemotherapy with or without combined chemoradiotherapy for BRPC/LAPC, intending to achieve successful surgical resection with an R0 margin. Of the 37 patients who began neoadjuvant chemotherapy, 20 patients (54%) were successfully able to proceed with surgery.

Toxicity remains a significant concern in neoadjuvant treatment, with patients experiencing numerous Grade 3 or 4 adverse events. In two separate trials assessing the safety and efficacy of FOLFIRINOX and gemcitabine‐based neoadjuvant treatments, Grade 3 or 4 neutropenia was a common adverse event [11, 12]. In the current analysis, none of the patients suffered from neutropenic sepsis, as 84% of patients received primary G‐CSF prophylaxis; however, five (14%) with neutropenia required dose reductions.

A patient was switched from neoadjuvant FOLFIRINOX to gemcitabine/nab‐paclitaxel due to severe adverse events of diarrhea and Grade III mucositis after his second cycle. While all grades of diarrhea and vomiting were not systematically documented due to incomplete chart records, these symptoms were observed in several patients and likely impacted their quality of life (QOL). Both neoadjuvant FOLFIRINOX and gemcitabine/nab‐paclitaxel have shown an increased rate of potentially curative surgery having R0 resections and improved OS and DFS. However, the adverse events and toxicities have been shown to affect patient QOL [7, 11, 12].

The NCCN guidelines recommend that patients diagnosed with BRPC have staging laparoscopy and imaging studies before starting neoadjuvant treatment [5, 13]. In a single‐center study, staging laparoscopy resulted in upstaging of 9 (15%) upfront resectable patients; this included 5 (9%) BRPC patients and 10 (25%) LAPC patients [13]. Several other studies have shown benefits to staging laparoscopy in identifying unsuspected metastatic disease or upstaging resectable PC [14, 15]. A separate single‐center review also found that tumors in the pancreas' body and tail were twice as likely to harbor unsuspected metastases compared to those in the head (53% vs. 28%) [16]. In our cohort of patients, 28 (76%) patients had staging laparoscopy, while 9 (24%) did not. Although no patients in this study were upstaged at the time of diagnosis, laparoscopy remains an additional tool to determine disease stage accurately.

The current protocol for neoadjuvant chemotherapy of BRPC/LAPC is eight cycles of FOLFIRINOX [8, 17]. Various clinical studies have debated the efficacy of neoadjuvant treatment and its effectiveness in reducing the cancer burden for surgical resection. A meta‐analysis of 13 studies using FOLFIRINOX as the neoadjuvant treatment for LAPC reported an overall R0 resection rate of 74%. However, five studies were excluded from this analysis due to a lack of R0 resection data [18].

In another randomized Phase II clinical study assessing the effectiveness of neoadjuvant treatment in BRPC, 84% of patients (43/51) progressed to surgery, and 67% of patients (33/51) achieved an R0 resection. This study included two treatment protocols: FOLFIRINOX and gemcitabine/nab‐paclitaxel. Among patients who received FOLFIRINOX, 19 (73%) achieved an R0 resection, compared to 14 (56%) who received gemcitabine/nab‐paclitaxel (p = 0.202) [17]. This study was not designed to compare the two protocols but rather to show the effectiveness of neoadjuvant chemotherapy in reaching surgical resections and R0 resections [19].

In this cohort, eight of the nine patients who completed all neoadjuvant treatment and had R0 resections were on FOLFIRINOX. Only one patient in the R0 resection group was on gemcitabine/nab‐paclitaxel. Five patients in our cohort received gemcitabine/nab‐paclitaxel due to toxicity concerns or pre‐existing conditions that favored this regimen. Four of these five patients were able to undergo potentially curative surgery. The small cohort limits a direct comparison between the two treatment protocols. However, based on current literature and our findings, gemcitabine/nab‐paclitaxel can be considered a viable alternative when needed [11, 12, 19–20].

The results of neoadjuvant chemoradiotherapy in current trials have been mixed with newer studies debating its efficacy in PC [8, 21, 22]. A Phase II clinical trial reported a 97% R0 resection rate in a small patient population receiving neoadjuvant chemotherapy and chemoradiotherapy [9]. The PREOPANC study, which compared neoadjuvant chemoradiotherapy to upfront surgery in resectable pancreatic cancer and BRPC, found that the neoadjuvant treatment group had a significantly higher R0 resection rate of 72% compared to 43% in the upfront surgery group (p <. 001) [23]. The PREOPANC trial is criticized for using suboptimal single‐agent gemcitabine as a neoadjuvant protocol compared to FOLFIRINOX and gemcitabine/nab‐paclitaxel. The LAP‐07 and CONKO‐007 studies were unable to reach their primary endpoints with the addition of radiotherapy [24, 25, 26]. This was also seen indirectly in the Phase II Alliance A021501 study, where there was a lower percentage of patients completing surgical resections in the chemoradiotherapy group compared to the chemotherapy‐only group [8]. The results of the Phase II Alliance A021501 study led to a change in the neoadjuvant protocol at our center. Two patients in our analysis were treated with eight cycles of chemotherapy but without chemoradiotherapy and still progressed to R0 resections.

Based on both the available literature and our findings, neoadjuvant chemoradiotherapy may improve R0 resection rates, but its impact on OS remains unclear. Outside clinical trials, the role of neoadjuvant chemoradiation or stereotactic body radiation therapy (SBRT) is still unclear in the management of BRPC and LAPC, and there are multiple prospective Phases II and III trials evaluating the role of these modalities in these subgroups of patients [27].

Although our study was a retrospective series with a small cohort, it still makes a valuable contribution to the current literature. It suggests the feasibility of the neoadjuvant approach FOLFIRINOX and gemcitabine/nab‐paclitaxel chemotherapy with or without chemoradiation in patients with BRPC/LAPC in routine clinical practice. The strengths of our cohort are comprehensive staging with imaging and laparoscopy, and all patients' management plans were discussed at MDM. The small number of patients who received chemoradiation therapy were able to complete the planned treatment with no significant complications. This study also adds to the knowledge that the biological behavior of pancreatic cancer can be tested by using the neoadjuvant approach with a reasonable number of patients achieving R0 resection with improvement in DFS and OS. It also suggests that eight cycles of FOLFIRINOX or four cycles of nab‐paclitaxel and gemcitabine are feasible as a neoadjuvant approach.

5. Limitations of study

The present study has several limitations. As a retrospective series, the data collected may not always reflect the clinical situation. For instance, although patients were assigned an ECOG performance status of 0 or 1 at diagnosis, this may not have been explicitly stated. Since this was a non‐randomized cohort, it lacked a control group, which meant no direct comparison for OS and DFS. The small sample size was another limitation; a more extensive multicenter real‐world study could have provided further insight into current protocols while awaiting further randomized control trials. As this was a retrospective analysis, certain aspects of data, like adverse events and toxicities, may have been missed. A total of six (16%) patients were lost to follow‐up, which may limit the exact estimation of OS and DFS. We are also unable to provide the precise difference in the outcome of BRPC/LAPC due to the limited number of LAPC patients in this study. We are also unable to provide the details of adjuvant approaches adopted for these patients, and this remains an area of active investigation, especially after 4 months of neoadjuvant approach.

6. Conclusion

This retrospective series suggests that neoadjuvant FOLFIRINOX or gemcitabine/nab‐paclitaxel is a feasible and potentially effective treatment for BRPC/LAPC. It offers patients the opportunity to undergo surgery, achieve R0 resection, and improve DFS and OS with a manageable toxicity profile. Outside clinical trials, the role of neoadjuvant chemoradiation therapy remains uncertain in terms of improvement in long‐term survival. However, further randomized controlled trials with larger cohorts are needed to validate these findings and refine treatment protocols.

Author Contributions

Avtaj Nerwal: validation, software, investigation, formal analysis, writing – original draft, writing – review and editing, visualization. Desmond Yip: conceptualization, resources, writing – review and editing, visualization, supervision. Sivakumar Gananadha: visualization, writing – review and editing, supervision. Amy Shorthouse: visualization, supervision. Belinda Lee: resources, software, writing – review and editing, visualization. Ankit Jain: conceptualization, software, validation, formal analysis, resources, writing – review and editing, visualization, supervision.

Ethics Statement

The study was approved by the ACT Health Human Research Ethics Committee Canberra.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.