Abstract
Background: Neoadjuvant chemotherapy (NAC) is established in the treatment of ductal pancreatic adenocarcinoma for downsizing borderline-resectable pancreatic cancer (BRPC) and may affect nodal positivity and rates of R0 resection. This study aimed to identify the impact of NAC on postoperative histopathological parameters with a prognostic relevance. Patients and Methods: A one-to-three matched-pair analysis, including an overall total of 132 patients (25% treated with NAC and subsequent resection and 75% undergoing upfront surgery) was performed. Influence of NAC on nodal positivity, lymphatic, vascular and perineural invasion, as well as resection stage and grading, was examined. Furthermore, perioperative complications, in-hospital stay, re-admission rates, mortality, as well as preoperative body mass index and American Association of Anesthesiologist classification scores, were evaluated. Results: Patients treated with NAC significantly less frequently had lymphatic tissue invasion (lymph node invasion: 51.5% vs. 72.7%; p=0.032, and lymphatic vessel invasion 9.4% vs. 55.3%; p=0.0004), whereas vascular and perineural invasion, as well as grading and resection state were not significantly different. Carbohydrate antigen 19-9 regression in correlation with nodal positivity also did not differ, and both groups showed comparable perioperative complication rates. Occurrence and severity of postoperative pancreatic fistula (18.2% vs. 24.3%; p=0.034) were significantly lower in patients who had undergone NAC. Conclusion: NAC significantly affects postoperative histopathological tumour stage in BRPC and appears to be a safe treatment option without increased perioperative complications, re-admission, in-hospital stay, or mortality. Further studies are mandatory to underline the suitability of NAC for ductal pancreatic adenocarcinoma subgroups in order to guide clinicians in their daily decision-making comprehensively.
- Borderline resectable pancreatic cancer
- BRPC
- neoadjuvant chemotherapy
- NAC
- histopathological tumour stage
Pancreatic ductal adenocarcinoma (PDAC) remains a malignant tumour entity with a grim prognosis, and surgical treatment is still considered the only option for cure, with currently increasing importance for those with borderline resectable pancreatic cancer (BRPC) (1). Following resection of PDAC, there is an indication for adjuvant chemotherapy. The regimen chosen differs depending on the general condition of the patients and their accompanying diseases. However, due to postoperative complications and delayed recovery, 20-30% of patients do not receive the indicated adjuvant chemotherapy (2).
The use of neoadjuvant chemotherapy (NAC) in the treatment of PDAC is mainly focused on BRPC. Resectability is defined as the absence of distant metastasis as well as the distance to the superior mesenteric artery, common hepatic artery and coeliac axis as well as an involvement of less than 180° of the circumference of the portal vein and the superior mesenteric vein (3). BRPC is defined as potentially resectable, with a high risk of infiltrated resection margins due to encasement of vascular structures, and thereby represents a target for downsizing by neoadjuvant treatment (NAT). Positive resection margins often imply a prognostic deterioration with high perioperative morbidity in pancreatic surgery. There are no unified criteria for the definition of BRPC. The revised National Comprehensive Cancer Network (NCCN) 2012 classification, supported by the International Study Group for Pancreatic Surgery (4), the MD Anderson Cancer Center criteria and the American Hepatopancreaticobiliary Association/Society of Surgery of the Alimentary Tract/Society of Surgical Oncology classification are most commonly applied (5). The NCCN 2012 classification determines BRPC as having no abutment or encasement with the coeliac axis, abutment or short segment encasement with the common hepatic artery, abutment with the superior mesenteric artery and abutment with impingement or narrowing to superior mesenteric vein and portal vein.
In patients with BRPC as well as local advanced pancreatic cancer (LAPC), nodal positivity as well as positive resection margins were associated with a decrease of overall survival (OS) (6).
NAC diminishes the rate of nodal positivity compared to upfront surgery (7). In order to identify additional local histopathological factors with prognostic relevance influenced by NAC, in this study, patients who underwent upfront surgery were matched by their postoperative T-stage in a one-to-three matched-pair analysis to patients who underwent NAC in our surgical Department. Involvement of local lymph nodes, lymphatic, vascular and perineural invasion, grading and resection state were analysed. A recent meta-analysis showed a positive influence of NAC on nodal positivity, perineural invasion, R0 resection rate and lymphatic vessel invasion (8). To the best of our knowledge, no matched-pair analysis aiming to identify the impact of NAC on local histopathological parameters with prognostic relevance has ever been performed. As local nodal positivity is an important prognostic factor influenced by NAC, correlation between initial pre-chemotherapeutic serum carbohydrate antigen 19-9 (CA19-9) level and regression after NAC on the one hand and nodal positivity on the other, were examined in order to evaluate CA19-9 regression as a predictive factor for nodal positivity in patients undergoing NAC and secondary resection for BRPC.
NAC and subsequent resection appeared to be a safe treatment option considering complication rates such as pancreatic fistula and other perioperative complications (9), whereas the complication burden was higher for patients who underwent NAC (10). The utility of NAC should not only be evaluated by its influence on the postoperative tumour stage. We therefore additionally assessed perioperative parameters and overall complication rates.
Patients and Methods
Data extraction and matching criteria. A retrospective single-centre analysis was conducted at a tertiary referral centre for pancreatic surgery. Following permission from the local Institutional Review Board (Charité Ethics Committee Ref. EA2/035/14), clinicopathological data of patients with BRPC who underwent pancreatic resection following NAC from 2016 to 2018 were consecutively collected in a database and further reviewed. A one-to-three matched-pair analysis was performed including data on postoperative T-stage, sex, age and resection performed [pylorus-preserving pancreatic head resection (PPPD), distal pancreatectomy (DP), total pancreatectomy (TP)] with patients who underwent upfront surgery from our database. Figure 1 depicts the flow diagram for matching. Postoperative local histopathological tumour stage, assessed by an attending pathologist based on the current TNM classification criteria, was extracted and analysed. Additionally, serum CA19-9 levels measured before and after applying NAC were assessed and correlated with postoperative nodal positivity.
The following data were collected for each patient: American Association of Anesthesiologists score (ASA) and body mass index (BMI); postoperative pancreatic fistula (POPF), postoperative pancreatic hemorrhage, surgical site infections, delayed gastric emptying and insufficiency of biliodigestive anastomosis; in-hospital stay, intensive care unit stay; required intervention or reoperation; re-admission within 90 days; other complications including cardiovascular and pulmonary complications (pulmonary embolism, pneumonia, respiratory insufficiency, myocardial infarction) as well as postoperative delirium and complications of the urinary tract (such as cystitis and urinary retention). Furthermore, complications were subdivided based on the Clavien/-Dindo classification (11) and mortality rates after 30 days were evaluated.
Inclusion and exclusion criteria. All patients undergoing resection after NAC for BRPC were included, whereas patients treated for other entities such as neuroendocrine tumours or metastasis were excluded. Patients with complete pathological responses (postoperative T0 stage) after NAT were excluded due to impossible matching (Figure 1).
Neoadjuvant treatment and preoperative assessment. A multimodal therapeutic approach was individually formulated for every oncological patient by a multidisciplinary Tumour Board, consisting of attending radiologists, medical oncologists, specialized radiation therapists and surgeons. Patients with ongoing chemotherapy were either observed by our outpatient clinic or an oncologicaI practitioner. Initial and pre-operative imaging included computed tomography (CT) with contrast agents or magnetic resonance imaging (MRI) assessing tumour extent, distant metastasis and adherence to vascular structures. Standard preoperative clinical diagnostics additionally included a physical examination and routine laboratory testing, as well as measuring of serum CA19-9 level.
Postoperative evaluation. Postoperative care was standardized and patients were monitored at a specialized surgical intensive care unit for at least 1 day. All resected specimens were histologically examined by an attending pathologist to identify tumour entity, to determine local tumour stage, and to evaluate the resection margin. R0 resection was defined as a complete resection with surgical margins microscopically negative for tumour cells. Each case was individually discussed by our interdisciplinary Tumour Board for indication of adjuvant therapy.
Patient characteristics.
Statistics. Bivariant two-tailed Pearson's chi-square test was performed using SPSS version 25.0 (IBM, Armonk, NY, USA) to compare incidences of specific histopathological criteria in patients with and without treatment prior to resection. Two-tailed Pearson's chi-square test and Friedman test were used to compare preoperative condition factors such as ASA and BMI and postoperative complications depending on ordinal or nominal scaled variables. Significance tests were two-sided, and a value of p<0.05 was considered to be statistically significant.
Results
Patient characteristics. Overall, 132 patients were included, a quarter of them undergoing NAT, whereas the other patients underwent upfront surgery without NAC. 50.8% of all patients were female and 49.2% male with a mean age of 62 (35-94) years. Most patients (51.5%) of both groups underwent PPPD and had stage T2 or T3 postoperatively (Table I). Chosen regimens for neoadjuvant treatment were: FOLFIRINOX (folic acid, 5-fluorouracil, irinotecan, oxaliplatin) in 76% of patients (n=25), gemcitabine in combination with nab-paclitaxel in 18% (n=6) and gemcitabine monotherapy for 6% (n=2). In the majority of the patients (n=30), at least four cycles of chemotherapy were applied. Three patients received less than four cycles of chemotherapy either due to a good response and early surgery, or chemotherapy-related complications with maintained resectability.
Correlation of NAT and postoperative histopathological criteria. Patients undergoing NAC significantly less frequently had lymphatic tissue invasion (lymph node invasion 51.5% vs. 72.7%; p=0.032; lymphatic vessel invasion (9.4% vs. 55.3%; p=0.0004), whereas vascular and perineural invasion, as well as grading and resection state, were not significantly affected. The most significant impact on nodal positivity was seen in patients with a postoperative T4 stage, while in the group that underwent upfront surgery without NAC, patients with a postoperative T2 stage had a higher N0 rate (Table II).
Flow diagram for matching in the one-to-three matched-pair analysis. PDAC: Pancreatic ductal adenocarcinoma.
Correlation of post-resectional nodal positivity and serum CA19-9 regression. Nodal positivity was not significantly related to the pre-therapeutic serum levels of CA19-9. Nor was it related to the absolute or relative regression of CA19-9 level by comparing pretherapeutic and pre-resectional levels (Table III).
Preoperative condition and postoperative morbidity. Patients who underwent NAC and those who underwent upfront surgery showed comparable preoperative conditions. NAC did not lead to a significantly lower BMI or increased pre-resectional morbidity due to complicative chemotherapeutic regimens (Table IV).
Postoperative morbidity was also not significantly influenced by NAT, with a in-hospital stay and mortality rates being comparable. Major complications (Clavien–Dindo ≥3a) were also comparable in frequency (39.3% vs. 45.5%, p=0.544). The rate and also the severity of POPF was even significantly lower in patients undergoing NAC (18.2% vs. 24.3%, p=0.034) (Table IV).
Postresectional histopathological tumour stage in patients based on a subdivision by neoadjuvant chemotherapy or upfront surgery.
Carbohydrate antigen 19-9 (CA19-9) at point of diagnosis and 1 day before surgery and regression of CA19-9.
Discussion
As PDAC still has a poor prognosis, additive adjuvant chemotherapy can generate a prolonged OS and reduce recurrence. European Study group for Pancreatic Cancer (ESPAC) trials prospectively compared consecutive different adjuvant regimens. The ESPAC-4 trial suggested a combination of gemcitabine and capecitabine for adjuvant treatment of PDAC (12). A recent prospective trial compared FOLFIRINOX to gemcitabine, with a favorable OS of 21.6 months with FOLFIRINOX to 12.8 months with gemcitabine monotherapy (13). One of the main problems of adjuvant chemotherapy after such invasive surgeries remains the high rate of patients not receiving the indicated therapy due to surgical complications, such as POPF or surgical site infections, prolonged in-hospital stay or limited general condition. Additionally, adjuvant chemotherapy does not immediately influence the post-resectional histopathological tumour stage, such as the rate of R0 resection and nodal positivity. It thereby remains an option for resectable pancreatic cancer.
Various neoadjuvant regimens with and without additional radiotherapy, analogous to regimens implemented for adjuvant therapy, have been evaluated. The first approach was made using 5-fluorouracil and mitomycin C in combination with radiation in 1993. Thereby a resection rate of 38% with a 5-year survival rate of 58% was achieved (14). Subsequently, gemcitabine-based regimens were introduced. The first prospective trial in 2018, comparing patients who underwent neoadjuvant gemcitabine-based chemoradiation with those who underwent upfront surgery for BRPC, showed a significantly longer median survival (21 vs. 12 months) and R0 resection rate (15). A recent meta-analysis showed a favorable OS of 22.2 months in patients undergoing NAC with FOLFIRINOX and an R0 rate of 83.9% including 283 patients with BRPC from 24 studies (16).
Preoperative assessment and perioperative complications.
Currently, NAC is mainly considered for a PDAC subtype on the margin, BRPC. Resecting BRPC has a high risk of positive resection margins, a factor with prognostic relevance that can significantly be influenced by NAC (17). Our study did not show a significant increase of R0 rate, which may be explained by the exclusion of patients with postoperative T0 and subsequent R0 states. A meta-analysis assessed 11 studies with 9,388 patients overall comparing upfront surgery with 5-fluorouracil-based or gemcitabine-based chemoradiation for patients with resectable PDAC, and showed no significant advantage in OS, but a higher incidence of R0 resection (18).
Even in patients with LAPC, NAC with FOLFIRINOX resulted in a resectability rate of 60% with a favorable OS (19). R0 resection rates after NAC for patients (resectable, BRPC, and LAPC) are described in up to 85% (17).
Predictive parameters for prolonged survival after NAC and consequent resection were recently identified, including regimens with more than six cycles of chemotherapy, CA19-9 regression and significant pathological response (20).
Our study revealed that not only the lymph node class, but also the lymphatic vessel class is significantly affected by NAC, reducing the risk for lymphatic invasion and metastasis. NAC had the greatest effect on nodal positivity in patients with a postoperative T4 stage (N0: 60% vs. 6.7%). Vascular invasion was not significantly affected nor was perineural invasion, leaving these routes open for systemic invasion and metastasis. A positive effect on perineural invasion was described in a meta-analysis without pre-matched T-stages for patients who underwent NAT (8). NAT did not significantly influence the differentiation of tumour cells (grading) in our cohort. A recent subgroup analysis in a meta-analysis showed a slight decrease in those with G3 tumors. However, patients in these studies were not matched by postoperative T-stage. Matching by postoperative T-stage ensures comparability of correlated histopathological parameters. In our group, patients with a postoperative T2 stage undergoing upfront surgery had a higher N0 rate than those who underwent NAC. Intrinsic lower nodal positivity rate in patients with lower T-stage may serve as a possible explanation for this. However, further assessment is mandatory.
Characterising a response to NAT in pancreatic cancer with means of radiographic imaging is challenging. Response evaluation criteria in solid tumours (RECIST) are commonly used, defining a complete response (CR) as the complete cure, partial response (PR) as a decrease of 30% of the longest tumour diameter, stable disease (SD) as no significant growth or regression, and the progressive disease (PD) as an increase of at least 20% of the longest tumour diameter (21). CR and PR are scarcely seen in PDAC as tumour necrosis can hardly be distinguished radiographically from vital tumour tissue. The absolute serum CA19-9 level after NAT and the proportion of regression supplement the radiographical criteria for indication of surgical exploration during or after NAC. In our study, initial serum CA19-9 levels and CA19-9 regression after NAC did not significantly predict nodal positivity or negativity, but was still useful for therapy decision as well as pre- and postresectional monitoring and follow-up.
BMI is considered to be of prognostic relevance for perioperative and long-term morbidity and mortality. As chemotherapy is related to peri-therapeutic nausea and weight loss, we compared the preoperative BMI of both groups and found no significant difference. ASA scores indicating morbidity and general condition were aIso comparable for both groups.
NAT may also be considered to influence local tissue quality by causing tumour necrosis and scarring. In our study, there was no significant difference in mortality rates after 30 days and common morbidities in pancreatic surgery such as insufficiency of anastomosis, surgical site infections, postoperative pancreatic haemorrhage and delayed gastric emptying. Furthermore, reoperation and intervention rates were equal for both groups. In our cohort, NAC was correlated with a significantly lower POPF rate and severity, causing fewer type C cases of POPF, as described before (10). NAC and subsequent pancreatic resection thus appears to be a safe and reliable treatment option (22).
We consider PDAC to be a systemic disease with a reduced overall prognosis. Our study was able to identify local histopathological parameters with prognostic relevance that remain unaffected by additional chemotherapeutic treatment.
The present study was limited by common biases that are mainly due to the retrospective character of this analysis, although we used a carefully matched control group to counter this condition.
Conclusion
Our study showed that NAC and subsequent pancreatic surgery is a safe and reliable treatment option leading to a favourable postoperative local histopathological tumour stage in BRPC. Applicability of these findings to other subgroups, such as resectable PDAC, needs further assessment in prospective trials.
Footnotes
Authors' Contributions
Study design was drafted by LT, TM and MB. LT and TM identified patients, collected all data, performed a one-to-three matched pair analysis and statistics. Data interpretation was performed by all Authors. The final article was drafted by LT, TM and FK. All Authors read and approved the final article.
Availability of Data and Materials
All data analyzed are included in this article. The datasets used during the current study are available from the corresponding Author on reasonable request.
Conflicts of Interest
All authors declare that they have no competing interests. All Authors declare they have no commercial associations that might represent conflicts of interest with the contents of this article and have nothing to disclose.
- Received August 29, 2019.
- Revision received September 14, 2019.
- Accepted September 17, 2019.
- Copyright© 2019, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
