Predictive factors for survival in borderline resectable and locally advanced pancreatic cancer: are these really two different entities?

All consecutive patients who underwent a surgical exploration for BR or LA at our hospital between January 2011 and June 2021 were identified, based on the review of clinical files and surgical records. At diagnosis all patients underwent a triphasic contrast-enhanced computed tomography (CT), occasionally with the addition of magnetic resonance imaging (MRI). Diagnostic work-up included a chest CT and baseline CA 19.9. The tumors were classified as resectable, BR or LA according to the NCCN guidelines [3]. In summary, a tumor was considered BR when there was contact with the superior mesenteric artery (SMA) to an extent inferior to 180° or with the common hepatic artery without extension to the coeliac axis. When tumor contact with the superior mesenteric vein (SMV) or portal vein was wider than 180° or venous thrombosis was present, the tumor remained classified as BR, if segmental resection and venous reconstruction remained feasible. A tumor was classified as LA when there was solid tumor contact with SMA wider than 180°, extension to the coeliac axis or unreconstructible involvement of the SMV or portal vein. Evaluation of iconographic response after NACHT was performed according to the RECIST guidelines [11]. All patients were discussed at a multidisciplinary oncology meeting, with extensive review of all imaging by radiologists and surgeons. The NCCN classification was determined and the indication for upfront surgery or NACHT was agreed upon. For patients who had multiple cycles of NACHT with interval evaluations, only the last assessment before surgery was considered for this study. A biopsy confirming the diagnosis of PDAC was required before start of NACHT. The data collected included demographics, details on the surgical procedure and postoperative period, morbidity during index admission according to Clavien-Dindo classification (minor complications classified as grade I and II, major complications corresponding to grade III or IV), 90-day mortality and oncologic follow-up [12].

The primary objective of the study was to identify factors associated with survival since surgery. Several factors that could potentially influence survival outcomes in patients with PDAC were investigated including surgical resection, lymph node involvement, use of NACHT, chemotherapy regimen, RECIST and NCCN classification. Secondary objectives were the assessment of OS, surgical resectability and postoperative outcomes such as post operative morbidity and length of hospital stay.

NACHT with FOLFIRINOX or other scheme were administrated according to standard of care and response was evaluated every 2 months with triphasic CT abdomen, CT chest, CA 19-9 and occasionally MRI abdomen. The value of CA19-9 was considered normal if inferior to 37 kU/L. In the absence of iconographic progression, and decreased or stable CA19.9 after NACHT in patients with good performance status, surgical exploration was performed to evaluate resectability. Surgical exploration commenced with exclusion of peritoneal and liver metastases, followed by evaluation of local resectability. When fibrous tissue was present in contact with a venous vessel, a vascular resection was preferred. When fibrous tissue was in contact with de SMA or common hepatic artery, periadventitial dissection was performed at the blood vessel with arterial divestment and the hard tissue was sampled for frozen section analyses [13]. Dissection only continued if frozen section analysis showed no malignancy. The finding of arterial tumoral invasion at a later point of the dissection, after committing to resection and achieving a point of no return, was considered a R2 resection. During the surgical exploration, tumors requiring an arterial resection and reconstruction to obtain a complete tumoral excision were considered unresectable. Arterial resection was only considered in case of left pancreatectomy with coeliac axis resection [14]. Venous invasion was not a limitation for resection if a reconstruction was deemed possible. Pancreatic resections were performed according to the tumoral location with pancreaticoduodenectomy, left pancreatectomy or total pancreatectomy with associated lymphadenectomy. The department of hepatopancreatobiliary surgery of Ghent University Hospital is considered a high-volume referral center for pancreatic surgery. Pancreatic specific postoperative complications were analyzed according to the definitions of the international study group of pancreatic surgery and only the clinically relevant pancreatic fistula type B and C were considered [15‐17]. Follow-up occurred according to the guidance of the oncologist in charge following the existent guidelines at the referral center, after postoperative surgical consultation. The first CT was performed within 3-6 weeks after the surgery, before starting adjuvant chemotherapy. When neoadjuvant chemotherapy was administered, the adjuvant therapy was given until completion of 12 cycles. In case of R1 or R2 resection, radiotherapy was advised after completion of adjuvant chemotherapy. Standardized macroscopic histopathological assessment of pancreatic resection specimens was performed according to the Leeds protocol in all specimens with evaluation of transection margins and dissection surfaces [18, 19]. The pancreatic and biliary transection margins were analyzed with frozen section. The circumferential resection margins were carefully analyzed. A margin was considered positive if the tumor was at or within 1 mm (≤1 mm) of the margin (R1). Tumor response after NACHT was scored in resected pancreatic specimens according to the grading system of College of American Pathologists (0: No viable cancer cells; 1: Single cells or rare small groups of cancer cells; 2: Residual cancer with evident tumor regression, but more than single cells or rare small groups of cancer cells; 3: Extensive residual cancer with no evident tumor regression) [20]. Lymph node ratio (LNR) was calculated in the patients who underwent a surgical resection by dividing the number of invaded lymph nodes by the number of retrieved lymph nodes and categorized as low LNR (0 – 0.2) or high LNR (>0.2) [21].

Continuous variables were described with median and interquartile range (IQR) and they were compared between groups using student’s t-test in case of a normal distribution, otherwise Mann-Whitney U test was applied. Categorical data were described with counts and percentages. Chi-square test or Fisher’s exact test was used to compare categorical variables. Survival from the date of surgery was estimated using the Kaplan-Meier method. The log-rank test was used to compare survival curves between groups. Both simple and multiple Cox proportional hazards models were fitted to identify preoperative factors associated with survival. Surgical resection and NACHT were respectively analyzed in conjugation with lymph node involvement and RECIST instead of being considered separately as dichotomic variables because of their dependence relation. Lymph node status is only known in patients who had a formal resection and RECIST is only determined for patients who received NACHT. This allowed a better fitting of the proportional hazards models. Hazard ratios (HRs) from simple models and conditional hazard ratios (cHRs) from the multiple model were reported with 95% Wald confidence interval (CI) and Wald p-value. All statistical analyses were performed with IBM® SPSS® statistics 28.0 (2020, Armonk, New York, United States) and open-source R-software [22]. A two-sided P-value of < 0.05 was considered statistically significant.