The pump: what it does, what it costs
A man with colorectal cancer had his liver metastases resected last month. He now has a metal disc the size of a hockey puck under the skin of his abdomen, and comes to clinic for a "pump fill". What is in it, and what do you check before it is filled?
Concepts first
- HAI
- Hepatic artery infusion: chemotherapy given into the artery of the liver, here through a pump implanted at surgery.
- The pump
- A sealed reservoir under the skin. Its catheter sits in the gastroduodenal artery (a branch of the common hepatic artery) with the tip at the junction, so the drug flows into the hepatic artery. Gas-driven, fixed rate; no battery, no programming.
- Floxuridine
- Also called FUDR: a fluoropyrimidine, the family of fluorouracil (5-FU), with a short half-life that the liver extracts on its first pass. That is why it, not 5-FU, goes in the pump.
- Perfusion scan
- Radioactive albumin injected through the pump before the first dose, to prove the drug reaches the whole liver and not the stomach or duodenum.
- Biliary sclerosis
- Scarring and narrowing of the bile ducts caused by floxuridine. The bile ducts depend only on the artery, so they get the full dose.
Why the artery.Tumor concentration up to 400 times that of systemic infusion, per the device maker; near-complete hepatic extraction keeps the rest of the body almost untouched (Lidsky 2023).
One cycle, 4 weeks
The Memorial Sloan Kettering (MSK) recipe, as written in their trials.Floxuridine 0.12 mg/kg/day and dexamethasone 1 mg/day, with heparin and saline, given for 14 days. Because the pump runs at a fixed rate, the amount loaded is scaled by the pump volume divided by its measured flow rate. On day 15 the pump is emptied and refilled with heparin and saline. Systemic chemotherapy goes in on days 1 and 15. Treatment starts at least 2 weeks after surgery (D'Angelica 2015). Each program sets its own starting dose; the liver tests then adjust it.
Why dexamethasone is in the pump
A randomized answer from 1992.In 50 patients, floxuridine with hepatic arterial dexamethasone versus floxuridine alone: a bilirubin rise of 200% or more in 9% versus 30% (P = 0.07), and response in 71% versus 40% (P = 0.03). Small trial, but the steroid stayed in the pump ever since (Kemeny 1992).
What it costs
Bile ducts
6% at 5 yearsBiliary sclerosis needing a stent or drain, 2239 patients at MSK. Rises with each cycle and with cumulative dose.
Stomach and duodenum
Ulcer, gastritisIf drug leaks into a branch artery. The perfusion scan before the first dose is meant to prevent it.
Pump and artery
17% severePocket infection, pump or catheter dysfunction, dislodgement, arterial dissection or thrombosis: grade 3 (severe) in 17% of the PUMP trial.
The bile ducts pay.Stent-requiring sclerosis: 5.5% after adjuvant pump therapy (Ito 2012); 6% at 5 years in 2239 patients (Choubey 2026). Bevacizumab with pump floxuridine raised biliary toxicity enough to be dropped from an MSK trial (D'Angelica 2015).
At every floxuridine fill (day 1)
Living with a pump
Gas drives the pump, so heat and altitude speed it up.No heating pads, saunas or long hot tubs over the pump; tell the team before trips above 1,000 feet or flights in unpressurized planes; no scuba or contact sports. It sets off metal detectors, not body scanners: carry the device card (Mayo Clinic patient guide).
Memory hooks
Tap a card to flip it.
Check yourself
Question 1Why floxuridine in the pump and not 5-FU?
Floxuridine is almost completely extracted by the liver on its first pass and has a short half-life: high dose to the tumor, little to the rest of the body.
Question 2Fill day: bilirubin 3.1, was 0.8 two weeks ago. What goes in the pump?
Heparin and saline, no floxuridine. Image the bile ducts; a stricture needs a stent. Look for tumor progression too.
Question 3A patient on pump therapy has new epigastric pain and a duodenal ulcer. Why?
Drug reaching the stomach or duodenum through a branch artery. Hold floxuridine (heparin and saline), endoscopy, repeat the perfusion scan, proton pump inhibitor.
Sources
Retrieved from PubMed, the device maker and patient education pages, read on Nov 3, 2026.
- D'Angelica MI, Correa-Gallego C, Paty PB, et al. Phase II trial of hepatic artery infusional and systemic chemotherapy for patients with unresectable hepatic metastases from colorectal cancer: conversion to resection and long-term outcomes. Ann Surg 2015;261:353-360. doi:10.1097/SLA.0000000000000614
- Kemeny N, Seiter K, Niedzwiecki D, et al. A randomized trial of intrahepatic infusion of fluorodeoxyuridine with dexamethasone versus fluorodeoxyuridine alone in the treatment of metastatic colorectal cancer. Cancer 1992;69:327-334. doi:10.1002/1097-0142(19920115)69:2<327::aid-cncr2820690209>3.0.co;2-u
- Ito K, Ito H, Kemeny NE, et al. Biliary sclerosis after hepatic arterial infusion pump chemotherapy for patients with colorectal cancer liver metastasis: incidence, clinical features, and risk factors. Ann Surg Oncol 2012;19:1609-1617. doi:10.1245/s10434-011-2102-8
- Choubey AP, Seier K, Schleimer L, et al. Hepatic artery infusion pump chemotherapy associated biliary sclerosis after treatment for colorectal cancer liver metastasis. Ann Surg Oncol 2026;33:5990-6000. doi:10.1245/s10434-026-19505-4
- Kemeny NE, Niedzwiecki D, Hollis DR, et al. Hepatic arterial infusion versus systemic therapy for hepatic metastases from colorectal cancer: a randomized trial of efficacy, quality of life, and molecular markers (CALGB 9481). J Clin Oncol 2006;24:1395-1403. doi:10.1200/JCO.2005.03.8166
- Cercek A, Boerner T, Tan BR, et al. Assessment of hepatic arterial infusion of floxuridine in combination with systemic gemcitabine and oxaliplatin in patients with unresectable intrahepatic cholangiocarcinoma: a phase 2 clinical trial. JAMA Oncol 2020;6:60-67. doi:10.1001/jamaoncol.2019.3718
- Lidsky ME, Cercek A, Maithel SK, Kemeny N, D'Angelica MI. Hepatic arterial infusion pump chemotherapy: is it ready for prime time in the United States? ASCO Post, April 13, 2023.
- Boston Scientific. Intera 3000 hepatic artery infusion pump, product page. bostonscientific.com, accessed Nov 3, 2026.
- Mayo Clinic Staff. Hepatic artery infusion pump (HAIP) chemotherapy. Tests and procedures, mayoclinic.org, May 20, 2025.
Who benefits: three settings, one rule
Tumor board: a 58-year-old with 14 colorectal liver metastases in both lobes, nothing outside the liver, after 4 months of FOLFOX (fluorouracil, leucovorin, oxaliplatin), mismatch-repair proficient. The surgeon says unresectable. Someone asks: "What about a pump?"
Concepts first
- Colorectal liver metastases
- Spread of colorectal cancer to the liver; in many patients the liver is the only site, for a long time.
- Intrahepatic cholangiocarcinoma
- Bile duct cancer arising inside the liver; when unresectable, it usually kills through the liver.
- Adjuvant
- Given after all visible tumor was removed, to prevent recurrence.
- Conversion
- Shrinking unresectable disease until surgery can remove it all.
- Progression-free survival
- Time alive without growth or recurrence anywhere. Hepatic progression-free survival counts only the liver; the pump moves this first. Overall survival is the harder test.
- HR
- Hazard ratio: under 1 favors the pump.
- Clinical risk score
- 0 to 5 points for features of the primary and the liver disease (node-positive primary, liver metastases within 12 months of the primary, more than one metastasis, largest over 5 cm, carcinoembryonic antigen (CEA) over 200 ng/mL). Low risk is 0 to 2.
The evidence, setting by setting
randomized, positiverandomized, negativenot randomized
After resection of colorectal liver metastases
NEJM 1999, 156 patients
86% vs 72%Alive at 2 years, pump + systemic fluorouracil (5-FU) vs systemic alone. Free of liver recurrence at 2 years: 90% vs 60%.
Memorial Sloan Kettering 2017, 2368 patients
67 vs 44 monthsMedian survival with vs without pump; all pump patients also had systemic chemotherapy. Propensity-adjusted HR 0.67.
PUMP 2026, 241 patients
15.0 vs 15.5 monthsProgression-free survival, resection + pump vs resection alone, no systemic chemotherapy in either arm; clinical risk score 0 to 2. HR 0.92.
Unresectable colorectal liver metastases
CALGB 9481, 135 patients
24.4 vs 20 monthsMedian survival, pump alone vs systemic 5-FU. Response 47% vs 24%.
MSK phase 2, 49 and 64 patients
47% and 52%Converted to resection with pump + modern systemic chemotherapy; most pretreated, median 13 to 14 tumors. 5-year survival 36% (64-patient trial).
Unresectable intrahepatic cholangiocarcinoma
MSK phase 2, 38 patients
25.0 monthsMedian survival with pump + gemcitabine and oxaliplatin. Partial response 58%.
PUMP-2, 50 patients
22.3 monthsMedian survival with pump, plus gemcitabine and cisplatin in 74%. Alive at 3 years: 31.5%.
Pooled, 4 trials, 142 patients
26 monthsMedian survival; alive at 5 years 15%. Disease control 96%.
Read the dates.Both positive randomized trials compared the pump with 5-FU-based chemotherapy, before oxaliplatin, irinotecan and biologics. In CALGB 9481 the pump alone controlled the liver but disease grew outside it sooner (7.7 versus 14.8 months), which is why US programs pair the pump with systemic chemotherapy. Modern randomized trials are running: PUMP-IT in the Netherlands (306 chemotherapy-naive patients with synchronous, unresectable liver-only colorectal metastases, primary endpoint overall survival) and EA2222 in the US.
The negative trial, in detail
PUMP, ASCO 2026.Twelve Dutch centers randomized 241 patients with resectable liver metastases and a clinical risk score of 0 to 2 to resection plus 6 cycles of pump floxuridine or resection alone, with no systemic chemotherapy in either arm. In 220 analyzed, median progression-free survival 15.0 versus 15.5 months (HR 0.92, 0.65 to 1.31); hepatic progression-free survival 37.6 versus 21.5 months (HR 0.82, not significant); overall survival immature and numerically worse with the pump (HR 1.34, 0.77 to 2.32). Ninety-day mortality 3.6% versus 0.9%, none attributed to treatment; grade 3 pump complications 17%. The MSK data had suggested low-risk patients gained most.
Is this a pump candidate?
Selection is the treatment.A 2026 real-world series of 40 patients (30 with regional nodes or disease outside the liver, all pretreated) reported a median overall survival of 9.4 months, far from the trial numbers (Galuia, ASCO GI 2026).
Memory hooks
Tap a card to flip it.
Check yourself
Question 1The tumor board patient: 14 bilobar colorectal metastases, liver only, after FOLFOX. Pump?
Yes, a candidate for referral: liver-only unresectable disease, the setting where pump plus systemic chemotherapy converted about half of patients to resection in MSK phase 2 trials.
Question 2Same patient, but PET shows two lung nodules and a peritoneal implant.
No. The pump treats the liver; disease outside it will decide survival. Systemic therapy.
Question 3Unresectable intrahepatic cholangiocarcinoma, 9 cm, liver only, resectable hilar node. Pump?
Yes, a candidate: phase 2 trials allowed resectable regional nodes and reported median survival around 2 years with pump plus systemic chemotherapy.
Question 4Resected single 3 cm metastasis, node-negative primary. A colleague cites the 1999 NEJM trial to place a pump. Your answer?
That trial was positive at 2 years against 5-FU alone, but the 2026 PUMP trial in exactly this low-risk group showed no progression-free benefit from pump floxuridine. Discuss at tumor board; not routine.
Sources
Retrieved from PubMed and meeting abstracts, read on Nov 3, 2026.
- Kemeny N, Huang Y, Cohen AM, et al. Hepatic arterial infusion of chemotherapy after resection of hepatic metastases from colorectal cancer. N Engl J Med 1999;341:2039-2048. doi:10.1056/NEJM199912303412702
- Groot Koerkamp B, Sadot E, Kemeny NE, et al. Perioperative hepatic arterial infusion pump chemotherapy is associated with longer survival after resection of colorectal liver metastases: a propensity score analysis. J Clin Oncol 2017;35:1938-1944. doi:10.1200/JCO.2016.71.8346
- Outmani L, Buisman FE, Filipe WF, Bolhuis K, et al. Adjuvant hepatic arterial infusion pump chemotherapy with floxuridine for patients with resectable colorectal liver metastases and a low clinical risk score: a randomized controlled trial, the PUMP trial. J Clin Oncol 2026;44(17 suppl):LBA3506. doi:10.1200/JCO.2026.44.17_suppl.LBA3506; numbers as reported by OncoDaily.
- Kemeny NE, Niedzwiecki D, Hollis DR, et al. Hepatic arterial infusion versus systemic therapy for hepatic metastases from colorectal cancer (CALGB 9481). J Clin Oncol 2006;24:1395-1403. doi:10.1200/JCO.2005.03.8166
- D'Angelica MI, Correa-Gallego C, Paty PB, et al. Phase II trial of hepatic artery infusional and systemic chemotherapy for patients with unresectable hepatic metastases from colorectal cancer. Ann Surg 2015;261:353-360. doi:10.1097/SLA.0000000000000614
- Pak LM, Kemeny NE, Capanu M, et al. Prospective phase II trial of combination hepatic artery infusion and systemic chemotherapy for unresectable colorectal liver metastases: long term results and curative potential. J Surg Oncol 2018;117:634-643. doi:10.1002/jso.24898
- Cercek A, Boerner T, Tan BR, et al. Assessment of hepatic arterial infusion of floxuridine in combination with systemic gemcitabine and oxaliplatin in patients with unresectable intrahepatic cholangiocarcinoma: a phase 2 clinical trial. JAMA Oncol 2020;6:60-67. doi:10.1001/jamaoncol.2019.3718
- Franssen S, Rousian M, Filipe WF, et al. Hepatic arterial infusion pump chemotherapy in patients with unresectable intrahepatic cholangiocarcinoma: PUMP-2 trial. J Clin Oncol 2026;44:455-465. doi:10.1200/JCO-25-00923
- Rousian M, Alessandris R, Schleimer L, et al. Hepatic artery infusion pump chemotherapy for unresectable intrahepatic cholangiocarcinoma: pooled individual patient-level analysis of four clinical trials. J Hepatol 2026. doi:10.1016/j.jhep.2026.06.022
- Ter Kuile, Kok, Westerink, Swijnenburg, et al. Hepatic arterial infusion pump chemotherapy combined with systemic therapy versus systemic therapy alone as induction treatment for initially unresectable colorectal liver metastases: study protocol of the PUMP-IT RCT. BMC Cancer 2026. doi:10.1186/s12885-026-16130-y
- Galuia M, Tejani MA, Bennett S, et al. Real-world outcomes of hepatic arterial infusion pump therapy. J Clin Oncol 2026;44(2 suppl):69. doi:10.1200/JCO.2026.44.2_suppl.69
- Lidsky ME, Cercek A, Maithel SK, Kemeny N, D'Angelica MI. Hepatic arterial infusion pump chemotherapy: is it ready for prime time in the United States? ASCO Post, April 13, 2023 (EA2222 design).
- ECOG-ACRIN. EA2222 physician fact sheet: eligibility and exclusions.