Pancreatic cancer surgery

Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1] ; Associate Editor(s)-in-Chief: Fahad Hasan, M.D.[2] Sudarshana Datta, MD [3]

Complete (R0) surgical resection is the only potentially curative treatment for pancreatic ductal adenocarcinoma (PDAC). It is appropriate only for localized disease and is one part of multimodality therapy that nearly always includes systemic chemotherapy.^[1][2] A multidisciplinary team assigns resectability from pancreatic-protocol imaging. The four categories are resectable, borderline resectable, locally advanced, and metastatic, and they are defined mainly by tumor contact with the celiac axis, superior mesenteric artery, common hepatic artery, superior mesenteric vein, and portal vein.^[3][4]

For resectable disease, the standard approach is upfront resection followed by adjuvant chemotherapy. For borderline resectable disease, neoadjuvant therapy is given before resection.^[3][5] Venous resection with reconstruction is accepted when it achieves R0. Arterial resection is reserved for selected responders at expert centers.^[3][6] Outcomes depend strongly on hospital volume.^[7] Metastatic and unconvertible locally advanced disease are managed with endoscopic or surgical palliation, not resection.^[3]

• Determine resectability at a multidisciplinary tumor board.
  • Use dedicated pancreatic-protocol multiphase CT angiography or MRI, ideally within 4 weeks before surgery.
  • Image before biliary stent placement when feasible.^[3]
• Use a structured radiology reporting template.
  • "Abutment" means ≤180° circumferential tumor–vessel contact.
  • "Encasement" means >180° contact or vessel contour deformity.^[4]
• Tumor size alone is not a criterion for unresectability. Contiguous organ involvement that can be resected en bloc does not by itself preclude resection.^[3]
• The 2017 International Association of Pancreatology consensus adds two further dimensions to anatomy:^[8]
  • Biological factors, such as markedly elevated CA 19-9
  • Conditional (physiological) factors, such as performance status
• Locally advanced disease is not necessarily permanently unresectable. A subset of patients becomes resectable after neoadjuvant therapy.^[3]

CA, celiac axis; SMA, superior mesenteric artery; CHA, common hepatic artery; SMV, superior mesenteric vein; PV, portal vein; IVC, inferior vena cava. Adapted from NCCN v4.2026.^[3]

• Neoadjuvant regimen
  • In the phase 3 PREOPANC-2 trial, overall survival did not differ between neoadjuvant FOLFIRINOX and neoadjuvant gemcitabine-based chemoradiotherapy (median 21.9 vs 21.3 months; HR 0.88, 95% CI 0.69–1.13). Both regimens are acceptable.^[12]
  • In the phase 3 PACT-21 CASSANDRA trial, preoperative PAXG (cisplatin, nab-paclitaxel, capecitabine, gemcitabine) improved event-free survival over mFOLFIRINOX in patients aged 18–75 years (median 16.0 vs 10.2 months; HR 0.63, 95% CI 0.47–0.84).^[13]
  • Regimen details belong in the medical therapy microchapter.
• Restaging after neoadjuvant therapy
  • Base the decision to operate on overall response:^[3]
    • Radiologic response (RECIST)
    • Biochemical response (CA 19-9)
    • Metabolic response (FDG-PET)
  • Do not base the decision on persistent vascular contact alone. CT cannot distinguish treatment-related fibrosis from viable tumor.
  • The more complex the planned operation, the greater the response needed to justify it.
  • The final determination is often made intraoperatively.
• CA 19-9 limitations
  • Interpret only after biliary decompression; obstruction and cholangitis falsely raise CA 19-9.
  • CA 19-9 is uninformative in Lewis antigen–negative patients.^[3]
• Adjuvant delivery
  • Completing systemic therapy is a key survival determinant.
  • Giving chemotherapy first ensures it is delivered to patients whose postoperative recovery would otherwise delay or prevent adjuvant therapy.^[3]

• Margins
  • A tumor-free margin of ≥1 mm is increasingly used as the clinically relevant R0 definition, although definitions vary internationally.^[15]
• Lymphadenectomy
  • Extended lymphadenectomy does not improve survival and increases morbidity compared with standard lymphadenectomy.^[16]
• Recurrence
  • Recurrence occurs in about 70–80% of patients after curative-intent resection, most often at distant sites.
  • Median recurrence-free survival is about 12 months.^[15]

• Venous (SMV/PV) resection
  • An oncologically appropriate extension of pancreatoduodenectomy when it allows R0 resection.^[3]
  • Reconstruction options:
    • Primary venorrhaphy or patch
    • Segmental resection with end-to-end anastomosis
    • Interposition graft: internal jugular, saphenous, or femoral vein, or allograft
  • It is associated with greater operative complexity, more blood loss, and a higher risk of positive margins.^[3]
  • Contemporary series report acceptable short-term mortality, with survival similar to standard resection when R0 is achieved.^[17][18]
• Arterial (CHA, CA, SMA) resection
  • Carries substantially higher morbidity and mortality than venous resection.
  • Restrict to highly selected patients with objective response to neoadjuvant therapy, at centers experienced in arterial reconstruction.
  • Counsel patients explicitly that the oncologic benefit is unproven.^[3]
  • In a 10-year single-center series of 715 patients, vascular resection was not independently associated with oncologic outcome after adjustment for perioperative therapy.^[19]
• Arterial divestment
  • Periadventitial, artery-sparing dissection is an emerging alternative to arterial resection in selected patients.^[18]
• 2019 French recommendations^[6]
  • Give neoadjuvant therapy before any planned vascular resection.
  • Avoid pancreatoduodenectomy with SMA resection when the periarterial frozen section is positive.

• Distal pancreatectomy
  • Minimally invasive (laparoscopic or robotic) distal pancreatectomy is an accepted approach.
  • In the international DIPLOMA randomized trial, it was noninferior to open surgery for radical (R0) resection of resectable PDAC.^[20]
  • Advantages include less blood loss and a shorter length of stay.^[21]
• Pancreatoduodenectomy
  • The LEOPARD-2 trial was stopped early because of excess complication-related mortality in the laparoscopic arm.^[22]
  • Meta-analyses show longer operative time, less blood loss, and shorter stay, with similar morbidity, mortality, R0 rates, and nodal yield.^[23]
  • DIPLOMA-2 randomized 288 patients at 14 high-volume European centers.^[24] Minimally invasive pancreatoduodenectomy was noninferior for 90-day complications and gave faster functional recovery.^[25]
  • Randomized oncologic data for robotic pancreatoduodenectomy in pancreatic cancer are pending (DIPLOMA-2x2).^[24]
  • Perform minimally invasive pancreatoduodenectomy only in experienced, high-volume programs.

• Volume–outcome evidence
  • Meta-analysis: higher hospital volume is associated with lower postoperative mortality (OR 0.35) and fewer major complications.
  • High surgeon volume is also associated with lower mortality (OR 0.29).^[7]
• Volume thresholds
  • NCCN recommends resection at institutions performing at least 15–20 pancreatic resections per year.^[3]
  • The best outcomes are reported for surgeons performing more than 20 pancreatoduodenectomies per year.^[2]
• Hospital versus surgeon volume
  • For short-term outcomes, hospital volume appears to outweigh individual surgeon volume, likely reflecting multidisciplinary rescue of complications.^[26]
• Timing versus volume
  • Delayed resection at a high-volume center was associated with lower mortality than earlier resection at a low-volume center (HR 0.81).^[27]
• Limits of volume as a quality marker
  • Outcomes vary about 2-fold even among surgeons who meet volume standards.^[28]
  • Target-trial emulation suggests confounding of the volume–outcome association.^[29]

At high-volume centers, operative mortality is below 4–5%, but overall morbidity remains 40–50%.^[2]

• Do not routinely drain the biliary tree before upfront resection.
  • In a randomized trial, routine preoperative drainage increased serious complications without benefit compared with early surgery.^[30]
• Reserve drainage for the following situations, using a self-expanding metal stent (SEMS):^[3]
  • Cholangitis
  • Severe symptomatic jaundice, such as intense pruritus
  • Anticipated delay to surgery, including neoadjuvant therapy

• Neoadjuvant therapy for anatomically resectable disease
  • This remains unresolved; randomized phase 2 trials have not shown a survival benefit.^[9][10]
  • Phase 3 trials are ongoing.
• Optimal neoadjuvant regimen and role of radiotherapy
  • Both are undefined.^[12][13]
  • Consolidative chemoradiation improves local control and margin negativity, but its overall survival benefit is inconsistent.^[3]
• Arterial resection and divestment
  • Oncologic benefit is unproven.
  • Classification of celiac axis involvement in body/tail tumors (distal pancreatectomy with celiac axis resection) differs between institutions and guideline panel members.^[3][31]
• Volume as a quality metric
  • The volume–outcome association is strong but partly confounded.
  • Regionalization involves trade-offs in access and travel.^[29][32]
• Robotic pancreatoduodenectomy
  • It lacks randomized oncologic outcome data.^[24]

• Deciding resectability
  • Calling tumors unresectable because of size or isolated reconstructible venous contact, rather than referring to an expert center.^[3]
  • Judging resectability after neoadjuvant therapy by persistent vascular contact on CT alone.^[3]
• Treatment sequencing
  • Operating upfront on borderline resectable disease instead of giving neoadjuvant therapy first.^[5]
  • Pursuing arterial resection without neoadjuvant response or adequate counseling.^[6]
• Preoperative management
  • Routinely stenting the bile duct before upfront resection.^[30]
  • Interpreting CA 19-9 during jaundice or cholangitis, or in Lewis antigen–negative patients.^[3]
• Operative technique
  • Preserving the spleen during distal pancreatectomy for adenocarcinoma.^[3]
  • Omitting frozen-section assessment of the pancreatic neck and bile duct margins.^[3]
• Choice of center
  • Attempting resection at a low-volume center, which carries higher rates of positive margins and failure to rescue complications.^[7]