Pancreatic cancer is the tenth most commonly diagnosed malignancy and fourth leading cause of cancer related mortality in the United States (1). Surgical resection has been considered the only curative treatment modality, however treatment options include surgical resection alone, surgical resection followed by adjuvant therapy, and neoadjuvant therapy prior to surgical resection with or without adjuvant therapy. The neoadjuvant approach has gained increasing popularity in the last decade despite controversy over its risks and benefits. Neoadjuvant therapy offers the theoretical advantages of down-staging borderline resectable or locally advanced tumors, enabling more patients to be candidates for surgical resection, increasing the rate of margin negative resection, treating occult micrometastatic disease, optimizing selection of surgical candidates, and increasing overall survival (2–4). Conversely, neoadjuvant therapy in pancreatic adenocarcinoma has variable response rates and delaying surgical resection may risk missing the opportunity for the only potentially curative modality (4, 5)

Standard definitions for resectability designation were created in an effort to standardize patients into groups based on likelihood and potential for margin negative resection. The AHPBA/SSAT/SSO guideline uses objective CT imaging criteria to designate pancreatic adenomas as resectable, borderline resectable, locally advanced, and metastatic (6). In addition to defining the role of neoadjuvant therapy in resectable and borderline resectable pancreatic cancer, more data is needed to compare different neoadjuvant regimens, including the use of systemic chemotherapy (SCT), chemoradiation therapy (CRT), or both. Several studies have demonstrated decreased rates of local recurrence and improved survival outcomes using various regimens in the preoperative setting, however more evidence is needed to elucidate the optimal treatment protocol (7, 8).

Greer et al. demonstrated that neoadjuvant gemcitabine-based CRT decreased rates of local recurrence in pancreatic adenocarcinoma (9). The PRODIGE 4/ACCORD 11 and MPACT trials demonstrated improved survival outcomes in patients with metastatic disease using neoadjuvant SCT with FOLFIRINOX and gemcitabine/abraxane over gemcitabine alone. It is still unknown whether similar results are observed in localized, non-metastatic pancreatic adenocarcinomas (10, 11). A recent study by Jang et al. found that gemcitabine based neoadjuvant CRT offered survival benefits over up front surgery in patients with borderline resectable pancreatic adenocarcinoma. This clinical trial not only demonstrated the efficacy of neoadjuvant therapy in borderline resectable disease, but also the efficacy of gemcitabine-based CRT in the absence of SCT (12).

Thus, while there is data supporting both SCT with and without CRT, as well as CRT alone, more evidence is needed to determine what the optimal neoadjuvant protocol entails in the setting of resectable and borderline resectable cancer. In this study, we compiled 13 years of data to compare the outcomes of neoadjuvant CRT with and without SCT in resectable and borderline resectable pancreatic adenocarcinoma. In addition to the previous most commonly used regimen of gemcitabine-based CRT alone and the currently popular systemic regimens of FOLFIRINOX and gemcitabine/abraxane, this retrospective study includes our institution's experience with two other neoadjuvant protocols: radio-sensitizing gemcitabine/cetuximab and systemic gemcitabine/docetaxel.

The Dartmouth-Hitchcock interdisciplinary gastrointestinal tumor board saw 1,020 patients with pancreatic adenocarcinoma between 2004 and 2017. Of those cases, 742 had imaging available for retrospective AHPBA/SSAT/SSO resectability designation. Two hundred eleven patients had resectable pancreatic cancer, and 126 of those patients initiated neoadjuvant therapy. Of those 126 patients with resectable pancreatic adenocarcinoma who initiated neoadjuvant therapy, 108 patients completed treatment (86%). Seventy-five of those patients underwent surgical exploration, with 100% completing surgical resection. Of those 75 patients, 39 received preoperative CRT only and 36 completed SCT and CRT in the preoperative setting. Of the 39 patients who underwent preoperative CRT, 11 were treated with gemcitabine alone and 28 were treated with gemcitabine plus cetuximab. Of the 36 patients who received SCT and CRT, 29 received gemcitabine plus docetaxel, six received gemcitabine/abraxane, and one received FOLFIRINOX.

The borderline resectable pancreatic adenocarcinoma group consisted of 270 patients. One hundred ninety-four of those patients initiated neoadjuvant therapy. One hundred and forty-three patients out of the 194 patients who initiated neoadjuvant therapy completed treatment (74%). Hundred patients underwent surgical exploration, and 74 of those patients completed surgical resection (74%). Of those 74 patients who completed NAT and surgical resection, 34 patients received CRT alone and 40 completed SCT with CRT. Of the 34 patients who underwent preoperative CRT, six were treated with gemcitabine alone, and 28 were treated with gemcitabine plus cetuximab. Of the 40 patients who underwent SCT and CRT, 29 received gemcitabine plus docetaxel, nine received gemcitabine/abraxane, and two received FOLFIRINOX. Figure 1 maps the inclusion and progression of patients in this study. Figure 1A shows initial patient selection through Tumor Board and AHPBA/SSAT/SSO designation. Figure 1B outlines the progression of patients with resectable pancreatic adenocarcinoma through neoadjuvant therapy and surgical resection, and Figure 1C charts the progression of patients with borderline resectable pancreatic adenocarcinoma through neoadjuvant therapy and surgical resection. The follow up period for this study was 3 years, starting from date of diagnosis.

There was no difference in age, gender, endobiliary stent presence, tumor location, tumor diameter, T classification, N classification, or CA 19-9 between the two treatment groups in resectable or borderline resectable disease. These baseline characteristics are displayed in Table 2 for resectable pancreatic adenocarcinoma and Table 3 for borderline resectable cancer. There was no significant difference in adjuvant therapy treatment following surgical resection in the two treatment groups for patients with resectable disease: 28% of patients who received CRT without SCT also received adjuvant therapy compared to 33% in patients who received neoadjuvant SCT with CRT (p = 0.65). There was no significant difference in patients who received adjuvant therapy following resection in borderline resectable cancer, with 22 and 34% in the CRT alone and SCT with CRT treatment groups, respectively (p = 0.26).

In the resectable cancers, 74% of patients who received only CRT had R0 margins compared to 78% in the SCT with CRT group (p = 0.73). For both treatment groups, 92% of patents lived past 1 year (p = 0.92). There was no significant difference in overall 3 years survival: 46% of patients who received CRT alone survived at least 3 years compared to 44% in the SCT with CRT treatment group (p = 0.86). The recurrence free survival rates did not differ significantly between the treatment groups. At 1 year, 74% of patients achieved recurrence free survival in the CRT alone group compared to 75% in patients treated with SCT and CRT (p = 0.95). 29% of patients had recurrence free survival at 3 years in the CRT alone group compared to 25% in patients who received SCT with CRT (p = 0.71). The resection and survival outcomes for resectable pancreatic adenocarcinoma are displayed in Table 4.

In borderline resectable cancers, patients who received CRT alone had 68% R0 margins compared to 57% in patients who received SCT and CRT (p = 0.37). At 1 year, 85% of patients survived in the CRT alone group compared to 93% in patients treated with SCT and CRT (p = 0.32). There was no significant difference in overall 3 years survival: 41% of patients who received only neoadjuvant CRT survived at least 3 years compared to 40% in the SCT with CRT treatment group (p = 0.92). 56% of patients had recurrence free survival at 1 year in the CRT alone group compared to 73% in patients who received SCT with CRT (p = 0.14). The 3 years recurrence free survival was 26 and 30% for the CRT alone group and the SCT plus CRT group, respectively (p = 0.76). The resection and survival outcomes for borderline resectable pancreatic adenocarcinoma are displayed in Table 5.

The Kaplan-Meier curve for resectable pancreatic adenocarcinoma showed no difference in the survival function up to 3 years between the two treatment groups, with a mean survival of 28.19 months for SCT with CRT and 28.93 for CRT alone (p = 0.87) (Figure 2). For borderline resectable adenocarcinoma, the Kaplan-Meier curve also showed no difference in the survival function between the two treatment groups, with a mean survival of 26.83 months for SCT with CRT and 25.94 for CRT alone (p = 0.58) (Figure 3). There was no significant difference in recurrence patterns (local, distant, or both) between the treatment groups in either resectable or borderline resectable pancreatic adenocarcinoma (Tables 6, 7, respectively).

Additional survival analysis was done using intention to treat methodology and included all patients who initiated neoadjuvant therapy, regardless of neoadjuvant therapy completion or surgical resection. In the intention to treat survival analysis for resectable pancreatic adenocarcinoma, there was no significant difference in 1 year overall survival: 38% of patients who received CRT alone compared to 37% in the SCT with CRT treatment group (p = 0.66). There was also no difference in 3 years overall survival with 17% in CRT patients and 14% in SCT with CRT patients (p = 0.62). These outcomes are displayed in Table 8. The intention to treat Kaplan-Meier curve for resectable pancreatic adenocarcinoma showed no difference in the survival function up to three years between the two treatment groups, with a mean survival of 22.32 months for SCT with CRT and 23.81 for CRT alone (p = 0.89) (Figure 4).

The intention to treat survival analysis for borderline resectable pancreatic adenocarcinoma demonstrated no significant difference in overall all survival rate. At 1 year, 34% of patients survived in the CRT alone group compared to 36% in patients treated with SCT and CRT (p = 0.27). At 3 years post diagnosis, there was a 9% survival rate in the CRT alone group compared to 12% in patients treated with SCT and CRT (p = 0.53). The intention to treat survival outcomes for borderline resectable pancreatic adenocarcinoma are displayed in Table 9. Additionally, the intention to treat Kaplan-Meier curve for borderline resectable disease showed no difference in the survival function between the two treatment groups, with a mean survival of 19.99 months for SCT with CRT and 17.08 for CRT alone (p = 0.19) (Figure 5).

While the efficacy of neoadjuvant therapy in localized pancreatic adenocarcinoma is still highly debated, neoadjuvant therapy is increasingly being utilized at institutions across the world. With modern systemic chemotherapy, oncologic outcomes in clinical trials show improved overall survival benefits not only in metastatic disease, but in localized cancer as well (7–9, 13, 14). In addition to defining the role of neoadjuvant therapy, more insight is needed on what the optimal treatment regimen is and whether that includes SCT, CRT, or both. The existing data comparing CRT alone and SCT plus CRT in locally advanced disease is contradictory: in comparing SCT with and without CRT, Hugeut et al. found survival benefits in patients who received SCT with CRT, while Hammal et al. saw no improvement in survival in patients who received SCT and CRT (15, 16). While multiple trials have compared different regimens and combinations of SCT, CRT, or both in resectable and borderline resectable pancreatic adenocarcinoma, there is still more work to elucidate the optimal treatment regimen (17–21).

This study reports 13 years of experience using neoadjuvant therapy for pancreatic adenocarcinoma with regimens including SCT with CRT, and CRT alone. The majority of patients with resectable and borderline pancreatic carcinoma who initiated neoadjuvant therapy completed their regimens and made it to surgical resection. We compared margin negative resection rates as well as recurrence patterns and saw no difference between the two treatment groups in either resectable or borderline resectable disease. We also did not find any significant advantage in overall or recurrence free survival between patients treated with SCT plus CRT vs. CRT alone. Our intention to treat survival analysis was consistent with our findings from examining only patients who achieved surgical resection.

Our findings suggest that for resectable and borderline resectable pancreatic adenocarcinoma, SCT with CRT does not offer any survival or recurrence benefits compared to CRT alone. This brings into question whether neoadjuvant SCT is necessary if neoadjuvant CRT alone is just as effective. SCT with CRT not only increases the delay in surgical resection compared to CRT alone, but it also subjects patients to increased side effects and drug toxicity (4). An alternative treatment approach to consider is using neoadjuvant CRT followed by surgical resection and systemic adjuvant therapy. This approach may provide patients with the benefits of neoadjuvant therapy, while decreasing further delay in potentially curative resection and offering the benefits of systemic therapy. A clinical trial is needed to provide insight on whether SCT in the neoadjuvant or adjuvant setting provides more benefits in patients receiving neoadjuvant CRT with resectable and borderline resectable pancreatic adenocarcinoma.

This study has several limitations. First, it is limited by its sample size and non-randomized observational design. The type of neoadjuvant therapy a patient received was made on an individual basis by the treating physicians. There was no over-arching guide, and it is possible that patients with more adverse clinical features were preferentially given one type of therapy over another, as this was not controlled for in our analysis. Additionally, grouping the different regimens into SCT with CRT vs. CRT alone does not allow us to compare each individual regimen directly. The heterogeneity in the SCT with CRT group includes the currently popular systemic regimens of FOLFIRINOX and gemcitabine/abraxane as well as old regimens no longer commonly used. This may account for the lack of benefit seen over CRT alone, as recent studies focusing on these newer regimens have shown promising results in borderline and locally advanced disease, and trials in resectable disease are beginning to shed more light (22–25). Although intention to treat survival analysis was done, some of our other data analysis excluded patients who were not surgically resected. This number was similar in both treatment groups for both resectable and borderline resectable cancer however, and was done to allow analysis of variables such as resection margins and recurrence patterns. Lastly, the follow up time of this study limits the detection of a difference in the overall and recurrence free survival beyond 3 years.

In conclusion, both SCT with CRT and CRT alone are effective options for neoadjuvant therapy in resectable and borderline resectable pancreatic adenocarcinoma. Our findings bring into question whether SCT is necessary to increase survival benefits if neoadjuvant CRT alone is just as effective. This argues for the need for better systemic therapy to justify its use, or whether neoadjuvant CRT alone should be used with adjuvant SCT following resection. While there is need for a prospective randomized clinical trial, our study is an early attempt at elucidating the effects of neoadjuvant SCT with CRT compared to CRT alone in resectable and borderline resectable pancreatic adenocarcinoma.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

The studies involving human participants were reviewed and approved by Dartmouth Committee for the Protection of Human Subjects. The patients/participants provided their written informed consent to participate in this study.

KT, DF, TG, and KS contributed to the study concept, design, acquisition of data, analysis, interpretation, drafting, and editing of the manuscript, and critical revision. All authors contributed to the article and approved the submitted version.

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