Pancreatic cancer

Overview

• ~60,000 cases/yr in the US, M:F 1:1, ↑ incidence in African Americans. 5-yr relative survival ~44% localized, ~17% regional, ~3% distant (SEER); ~80 to 85% present with advanced/unresectable disease.
• Risk factors: smoking, obesity, age, DM, chronic pancreatitis, pancreatic precursor lesions: PanIN (microscopic, noninvasive, small ducts), IPMN, and mucinous cystic neoplasms (three distinct entities).
• New-onset diabetes can predate the diagnosis by up to 2 years; depression can also predate it.
• Most common mutations: KRAS (90%), TP53 (Li-Fraumeni), SMAD4 (formerly DPC4), BRCA (~5%, hereditary breast-ovarian syndrome), ATM (ataxia-telangiectasia), PALB2 (breast ca), PRSS1 (hereditary pancreatitis, chronic inflammation → ↑ risk), STK11 (Peutz-Jeghers), p16/CDKN2A (FAMMM, familial atypical multiple mole and melanoma), MLH1/MSH2/MSH6/PMS2 (Lynch/HNPCC). Familial pancreatic ca (unknown genetics): diagnosis in younger patients (40s), multifocal dysplasia.
• Genetic testing for all pancreatic ca regardless of family history or stage, as ~5 to 10% carry a germline mutation.
• Screening with MRCP and EUS for high-risk patients:
  • First-degree relatives of pancreatic ca patients (familial kindreds).
  • Carriers of p16/CDKN2A (regardless of family history), or BRCA2 with an affected first-degree relative.
  • Peutz-Jeghers syndrome.
  • Lynch syndrome with an affected first-degree relative with pancreatic ca.
• Diagnosis: triple-phase CT (late arterial and portal venous phases). Late arterial phase shows a hypodense hypovascular mass and best assesses arterial anatomy; portovenous phase is best for liver metastases and venous anatomy. Also MRI, EUS with biopsy, and ERCP ± stent for biliary obstruction (~70% obstructed at presentation).
• CA 19-9: made by pancreatic ca cells (also stomach, colon, bile duct), so it can be elevated in biliary obstruction, cholangitis, and pancreatitis. It derives from the Lewis a blood antigen, and ~10% of Caucasians lack this antigen and cannot synthesize CA 19-9. A rising CA 19-9 precedes radiographic progression, but always rule out other causes with imaging and/or MRCP/ERCP.

Symptom management

• Refer to palliative care at any stage.
• Pain: celiac plexus neurolysis is superior to opioids alone.
• Metoclopramide for malignant gastroparesis (delayed gastric emptying without anatomic obstruction, from tumor infiltration of autonomic nerves).
• Gastric stent for gastric outlet obstruction; consider gastrojejunostomy and/or venting PEG for decompression.
• Pancreatic enzymes for malabsorption from exocrine insufficiency (50 to 80%): 40 to 50K IU lipase per meal, 20 to 25K per snack, with meals (not before), add PPI/H2 blocker, do not recommend a low-fat diet.
• DM management for endocrine insufficiency (~50%).
• Biliary stents: metallic (preferred, patent >6 mo) if life expectancy >6 mo; plastic has ↑ reobstruction risk (~3 mo); when preoperative drainage is needed, a short (often fully covered) metal stent is preferred, especially before neoadjuvant chemo. Consider percutaneous biliary drainage with subsequent internalization. Drain the obstruction and let bilirubin normalize before FOLFIRINOX (trial required bilirubin ≤1.5x ULN; irinotecan dosing not established for bilirubin >2 mg/dL).
• VTE: all pancreatic and stomach ca patients have a Khorana score ≥2, so prophylactic anticoagulation should be considered. Enoxaparin better than warfarin; consider DOACs in patients without luminal tumors.

Resectability

• Vessel definitions: arteries = celiac axis (CA), superior mesenteric artery (SMA), common hepatic artery (CHA); veins = superior mesenteric vein (SMV), portal vein (PV). Abutment = <180° contact; encasement = >180° contact. Teardrop sign: PV/SMV deformed to a teardrop by tumor encasement or fibrosis. Contour irregularity and thrombosis suggest vascular invasion (irregularity especially significant in arteries, which have thicker walls than veins).

Treatment principles

• Clinical TNM staging has prognostic value but plays little role in treatment decisions.
• Resectability is defined by tumor relationship with blood vessels.
• Systemic therapy is used in all stages.
• Squamous/adenosquamous ca are treated the same as adenocarcinoma.
• Staging laparoscopy in high-risk patients: ↑ CA 19-9, large primary, LN+, ↑ weight loss, ↑ pain.
• Consider neoadjuvant chemo in resectable patients with these high-risk features.

Adjuvant therapy

• Gemcitabine + capecitabine vs gemcitabine (ESPAC-4): ↑ 5-yr OS 29% vs 16%, but benefit mainly in R0 (R0 mOS 39.5 vs 27.9 mo; R1 23.7 vs 23.0 mo). Does not improve recurrence.
• FOLFIRINOX vs gemcitabine (PRODIGE-24): ↑ DFS 21.6 vs 12.8 mo, ↑ mOS 54 vs 35 mo, 5-yr OS 43.2% vs 31.4%.
• Gemcitabine vs observation (CONKO-1): ↑ 5-yr OS 20.7% vs 10.4%, 10-yr OS 12.2% vs 7.7%.
• Gemcitabine + nab-paclitaxel is NOT used adjuvantly (can be used neoadjuvant, locally advanced, or metastatic): APACT missed its primary DFS endpoint (19.4 vs 18.8 mo, HR 0.88, p=.18); final OS was nominally improved (41.8 vs 37.7 mo, HR 0.80, p=.0091; 5-yr OS ~38% vs 31%), a secondary finding that does not reverse the negative primary.
  • Nab-paclitaxel uses albumin as a vehicle. SPARC (Secreted Protein Acidic and Rich in Cysteine) is highly expressed in pancreatic tumor stroma; the albumin component was proposed to bind SPARC in the tumor stroma, but SPARC-mediated delivery is unproven (SPARC did not predict benefit in MPACT).
• ChemoRT is infrequent (ESPAC-1 showed no benefit vs chemo alone), but consider it after adjuvant chemo in selected patients (e.g., R1 resection; RTOG 0848 does not support it for LN+). (RTOG 0848, JCO 2026: adding CRT after adjuvant gemcitabine did not improve OS, HR 0.96; benefit suggested only in node-negative patients.)

Neoadjuvant therapy (resectable and borderline resectable)

• Perioperative mFOLFIRINOX vs gem/nab-paclitaxel (SWOG S1505, phase 2): similar outcomes (mOS 23.2 vs 23.6 mo); not a formal equivalence trial.
• mFOLFIRINOX or Gem/Cis for BRCA1/2 or PALB2 mutations.
• ESPAC-5F (borderline resectable): neoadjuvant CRT with cape vs mFOLFIRINOX vs Gem/Cape vs surgery alone; neoadjuvant ↑ 1-yr OS 77 to 84% (FOLFIRINOX 84%) vs 39% immediate surgery; resection and R0 rates similar (R0 23% vs 14%, p=.49).
• PREOPANC: neoadjuvant gemcitabine × 3 total (36 Gy/15 fx concurrent with cycle 2), surgery, adjuvant gemcitabine × 4 vs surgery then adjuvant gemcitabine × 6: ↑ R0 71% vs 40%, ↑ OS significant at long-term follow-up (HR 0.73, P = .025; 5-yr OS 20.5% vs 6.5%; Versteijne JCO 2022).
• ALLIANCE A021806 (neoadjuvant then adjuvant FOLFIRINOX): pending. PREOPANC-2 (neoadjuvant FOLFIRINOX × 8 vs gem-RT-gem sequence; Lancet Oncol 2025): no OS difference (mOS 21.9 vs 21.3 mo, HR 0.88), either approach acceptable.

Locally advanced unresectable

• Give definitive chemo, or chemo plus CRT, and reassess resectability. Chemo options: mFOLFIRINOX and gemcitabine/nab-paclitaxel.
  • 6 mo of combination chemo then CRT, reassess, and resect if resectable.
  • Gem/nab-paclitaxel then mFOLFIRINOX, reassess, and resect if resectable.
• LAP07: gemcitabine ± erlotinib, then RT vs continued chemo in stable disease; capecitabine-based chemoRT reduced local progression (32% vs 46%, p=.03) and lengthened the treatment-free interval but gave no PFS or OS benefit.
• TTFields (Optune Pax) + gem/nab-paclitaxel. PANOVA-3 (Babiker, JCO 2025, PMID 40448572) phase 3 in unresectable locally advanced PDAC (n=571): mOS 16.2 vs 14.2 mo (HR 0.82, p=0.039); PFS not improved (pain-free survival and distant PFS improved). FDA approved Feb 12, 2026 for LAPC with gem/nab-paclitaxel. Non-invasive alternating-electric-field adjunct; long wear requirement (≥18 hr/day) limits adoption. First phase 3 to show an OS benefit in LAPC (LAP07 was negative for OS).

Metastatic pancreatic cancer

• FOLFIRINOX for good PS, no biliary obstruction, no infection.
  • FOLFIRINOX vs gemcitabine (PRODIGE 4/ACCORD 11, 2011): ↑ OS 11.1 vs 6.8 mo (HR 0.57), PFS 6.4 vs 3.3 mo, ORR 31.6% vs 9.4%; first time mOS reached double digits. ↑ grade 3/4 neutropenia 46% vs 21% (febrile neutropenia 5.4%).
  • mFOLFIRINOX (now standard): no 5-FU bolus, irinotecan reduced from 180 to 150 mg/m².
• Gem/nab-paclitaxel vs gemcitabine (MPACT): ↑ OS 8.5 vs 6.7 mo (HR 0.72), ORR 23% vs 7%; ↑ neutropenia 38%, neuropathy 17%.
• NALIRIFOX (5-FU/LV + liposomal irinotecan + oxaliplatin) (NAPOLI-3, Wainberg Lancet 2023, PMID 37708895): 1L metastatic PDAC; mOS 11.1 vs 9.2 mo gem/nab-paclitaxel (HR 0.83), mPFS 7.4 vs 5.6 mo. FDA Feb 13, 2024. A 1L option alongside mFOLFIRINOX and gem/nab-paclitaxel.
• Gemcitabine/cisplatin for gBRCA1/2 or PALB2 mutations: mOS 16.4 mo, 3-yr OS 17.8%.
• Single-agent gemcitabine for ECOG 2 or comorbidities.
• Maintenance / chemo holiday after 4 to 6 months in responders or stable disease:
  • After mFOLFIRINOX: capecitabine, 5-FU/leucovorin, FOLFOX, or FOLFIRI.
  • After gem/nab-paclitaxel: single-agent gemcitabine or gem/nab-paclitaxel.
  • Olaparib maintenance (POLO): for germline BRCA1/2 with stable disease after 4 months of platinum chemo; improves PFS only (no OS benefit); vs placebo ORR 23.1% vs 11.5%. The control arm (placebo after stopping chemo at 16 weeks) was not a standard of care; in practice, after 4 to 6 months of disease control, de-escalate to maintenance or consider a treatment break; continuous combination chemotherapy is not mandatory.
  • Rucaparib maintenance for germline or somatic BRCA1/2 or PALB2.
• Second-line options:
  • Progressed on mFOLFIRINOX: gem/nab-paclitaxel, or Gem/Cis for BRCA1/2 and PALB2 (5-FU/liposomal irinotecan is indicated after gemcitabine-based therapy, not here).
  • Progressed on gem/nab-paclitaxel: 5-FU/liposomal irinotecan (category 1), mFOLFIRINOX, or 5-FU/Ox/FOLFIRI.
    • 5-FU/liposomal irinotecan vs 5-FU (NAPOLI-1): ↑ OS 6.1 vs 4.2 mo (HR 0.67).
    • 5-FU/oxaliplatin (weekly infusional 5-FU and oxaliplatin days 8 and 22 then a 3-week break, not FOLFOX) (CONKO-003): ↑ mOS 5.9 vs 3.3 mo (HR 0.66).

Precision/targeted therapy (NGS on all patients)

• KRAS context: mutated in ~90% (most common driver). KRAS G12C is uncommon in pancreatic ca; adagrasib or sotorasib for KRAS G12C. Broader KRAS-directed agents are in development.
• Daraxonrasib (Rasonque), oral RAS(ON) multi-selective inhibitor: RASolute 302 (metastatic PDAC after 1 prior line, n = 500) vs physician's-choice chemo, mOS 13.2 vs 6.7 mo (HR 0.40), mPFS 7.2 vs 3.6 mo (HR 0.49). FDA Aug 26, 2026 (regular approval) for metastatic PDAC after ≥1 prior systemic therapy or not a candidate for multiagent therapy; 300 mg PO daily; label does not require RAS testing.
• Pembrolizumab or dostarlimab for dMMR/MSI-H (very rare, <1%): pembrolizumab ORR 62.5% (Hu et al); KEYNOTE-158 pancreatic cohort ORR 18%, mOS ~4 mo.
• Entrectinib, larotrectinib, or repotrectinib (Augtyro) for NTRK fusion (tumor-agnostic; repotrectinib FDA Jun 13, 2024 for NTRK+ solid tumors, TRK TKI-naive or pretreated).
• Dabrafenib + trametinib for BRAF V600E.
• Zenocutuzumab (Bizengri), HER2/HER3 bispecific, for NRG1 fusion-positive pancreatic ca (rare, <1%). FDA accelerated approval Dec 4, 2024 for advanced/metastatic NRG1+ pancreatic ca after prior systemic therapy (eNRGy trial, Schram NEJM 2025; pancreatic ORR 42%).

IPMN

• Main-duct IPMN has higher cancer risk than side-branch IPMN. Diagnose with MRI/MRCP or pancreatic-protocol CT.
• EUS-FNA if IPMN has size ≥30 mm, thickened/enhancing cyst walls, enhancing mural nodule <5 mm, associated pancreatitis, dilated main pancreatic duct 5 to 9 mm, or abrupt caliber change with distal atrophy.
• Surgery by imaging: main pancreatic duct ≥10 mm, an enhancing solid component, or obstructive jaundice (a high-risk stigma); other cyst-attributable symptoms such as pancreatitis are worrisome features warranting EUS, not automatic surgery.
• Surgery by FNA: high-grade dysplasia, invasive carcinoma, or features concerning for malignancy.
• IPMN <10 mm: surveillance imaging.

Undifferentiated carcinoma with osteoclast-like giant cells (UC-OGCs)

• Very rare tumor with two elements: spindle/ovoid mononuclear cells and osteoclast-like giant cells. IHC: malignant mononuclear cells express vimentin, variable/focal keratin, and aberrant p53; the osteoclast-like giant cells and reactive histiocytes are CD68-positive and keratin-negative (LCA marks the reactive cells, not the malignant cells). Rapid growth, presents as a mass with hemorrhage and necrosis; prognosis is heterogeneous (resected pure UC-OGC without a ductal adenocarcinoma component can do better than PDAC). When systemic therapy is needed, extrapolate from PDAC (FOLFIRINOX or gem/nab-paclitaxel); Gem/Cis is for germline BRCA1/2 or PALB2 disease, not UC-OGC histology itself.