Outcome of patients with perihilar cholangiocarcinoma and previous biliary instrumentation: an observational study.

BACKGROUND: To assess the outcome of previously untreated patients with perihilar cholangiocarcinoma who present to a cancer referral center with or without pre-existing trans-papillary biliary drainage. METHODS: Consecutive patients with a diagnosis of perihilar cholangiocarcinoma presenting between January 1, 2013, and December 31, 2017, were identified from a prospective surgical database and by a query of the institutional database. Of 237 patients identified, 106 met inclusion criteria and were reviewed. Clinical information was obtained from the Electronic Medical Record and imaging studies were reviewed in the Picture Archiving and Communication System. RESULTS: 73 of 106 patients (69%) presenting with a new diagnosis of perihilar cholangiocarcinoma underwent trans-papillary biliary drainage (65 endoscopic and 8 percutaneous) prior to presentation at our institution. 8 of the 73 patients with trans-papillary biliary drainage (11%) presented with and 5 developed cholangitis; all 13 (18%) required subsequent intervention; none of the patients without trans-papillary biliary drainage presented with or required drainage for cholangitis (p = 0.008). Requiring drainage for cholangitis was more likely to delay treatment (p = 0.012) and portended a poorer median overall survival (13.6 months, 95%CI [4.08, not reached)] vs. 20.6 months, 95%CI [18.34, 37.51] p = 0.043). CONCLUSION: Trans-papillary biliary drainage for perihilar cholangiocarcinoma carries a risk of cholangitis and should be avoided when possible. Clinical and imaging findings of perihilar cholangiocarcinoma should prompt evaluation at a cancer referral center before any intervention. This would mitigate development of cholangitis necessitating additional drainage procedures, delaying treatment and potentially compromising survival.

Treatment of Chylous Ascites with Peritoneovenous Shunt (Denver Shunt) following Retroperitoneal Lymph Node Dissection in Patients with Urological Malignancies: Update of Efficacy and Predictors of Complications.

PURPOSE: We investigated the efficacy and analyzed the complication risk factors of peritoneovenous shunt in treating refractory chylous ascites following retroperitoneal lymph node dissection in patients with urological malignancies. MATERIALS AND METHODS: From April 2001 to March 2019 all patients with refractory chylous ascites after retroperitoneal lymph node dissection treated with peritoneovenous shunt were reviewed. Demographic characteristics, technical success, efficacy, patency period and complications were studied. Univariate and multivariate logistic regression analysis was performed to identify predictors of complications. RESULTS: Twenty patients were included in this study. Testicular cancer was the most common malignancy (85%). The mean number of days from surgery to detection of chylous ascites was 21 days (SD 15, range 4 to 65). Ascites permanently resolved after peritoneovenous shunt in 18 patients (90%), leading to shunt removal in 17 patients (85%) between 46 and 481 days (mean 162, SD 141). The mean serum albumin level increased 24% after shunt placement (mean 3.0±0.6 gm/dl before, 3.9±0.8 gm/dl after, p <0.05). The most common complication was occlusion (30%). Relative risk of complications increased significantly when shunt placement was more than 70 days after surgery and in patients with more than 5 paracenteses before peritoneovenous shunt placement (AR 0.71% vs 0.25%, RR 2.9, p <0.048 and AR 0.6% vs 0.125%, RR 4.8, p <0.04, respectively). CONCLUSIONS: Peritoneovenous shunt permanently treated chylous ascites in 90% of patients after retroperitoneal lymph node dissection. Peritoneovenous shunt was removed in 85% of patients. Shunt placement is an effective and safe treatment option for refractory chylous ascites. These patients might benefit from earlier intervention, after 4 to 6 weeks of conservative management as opposed to 2 to 3 months.

Safety and Efficacy of Hepatic Artery Embolization in Treating Solitary Fibrous Tumor Metastatic to the Liver.

The aim of this study was to evaluate safety and survival following hepatic artery embolization (HAE) for metastatic solitary fibrous tumor (SFT) in the liver. All patients with SFT metastatic to liver treated with HAE were retrospectively analyzed. Tumor response was evaluated using mRECIST. Objective response, overall survival (OS), and progression-free survival (PFS) were evaluated using Kaplan-Meier and multivariate Cox proportional hazard ratio. Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0. Twelve patients (6 males and 6 females, mean age: 42.5 ± 13 years; 24-65) were treated with 33 embolizations. Anatomical sites of origin for SFT were the head and neck (n = 6; 50%), pelvis (n = 2), pleura (n = 2), retroperitoneal (n = 1), and thigh (n = 1). The median follow-up from first HAE was 4.5 years (3-7.9). 84% of the patients showed objective response [42% complete response (CR) plus 42% partial response (PR)] to HAE by mRECIST (95% CI, 60-99%). Patients with CR to HAE had significantly higher OS compared to others (p < 0.02). The postembolization median OS was 4 years (95% CI, 2.3-5.2), and mean PFS, for intra- or extrahepatic progression of disease, was 6 months (95%, CI, 3.2-7.1). One patient developed pneumonia/sepsis and died 27 days postembolization, possibly not directly related to embolization. No grade III or IV adverse events were identified in the remaining patients. In conclusion, HAE for metastatic liver SFT is a relatively safe treatment option with high response rate and should be considered as a treatment option for metastatic liver SFT. In our cohort of patients with metastatic SFT to the liver, we observed a median OS of 4 years following HAE. Further studies are needed to confirm the efficacy of HAE.

Incidence and Consequence of Nontarget Embolization Following Bland Hepatic Arterial Embolization.

PURPOSE: Estimate the incidence of nontarget embolization (NTE) as identified on immediate post-hepatic artery embolization CT. MATERIALS AND METHODS: Two hundred hepatic embolizations performed with particles alone (bland embolization) in 147 patients between August 16, 2013 and August 26, 2014 with immediate post-procedure CT were retrospectively reviewed. Arterial anatomy, vessels treated, imaging findings of NTE, patient demographics, length of hospital stay following embolization, and procedure-related complications were recorded. The data were analyzed using two-sided t-tests and chi-squared tests. RESULTS: Evidence of NTE was seen on post-procedure CT in 64 of 200 cases (64/200, 32%). Six organs were affected, with 69 discrete sites in 64 patients. The majority (49/69, 71.0%) involved the gallbladder. The mean length of hospital stay (LOS) for patients with and without NTE was 2.9 ± 1.5 nights (range 1-7) and 2.9 ± 2.3 nights (range 0-21), respectively (P = 0.81). NTE was more common following embolization of replaced or accessory hepatic vessels. There were three complications in the NTE group (3/64, 4.7%) following the embolization procedure, one of which was cholecystitis directly related to NTE. The other two were one incidence each of contrast-induced nephropathy and pneumonia. In the group without NTE, seven complications occurred (7/136, 5.1%, P = 0.889), including one death resulting from hepatic failure, two gastrointestinal bleeds, two hepatic abscesses, flash pulmonary edema, and pancreatitis. CONCLUSION: Unanticipated NTE is not uncommon after bland hepatic artery embolization, particularly after treating accessory or replaced vessels, but does not increase complications or LOS. LEVEL OF EVIDENCE: Level 2b, Retrospective Cohort.

Autologous Blood Patch Injection versus Hydrogel Plug in CT-guided Lung Biopsy: A Prospective Randomized Trial.

Purpose To compare the effect of autologous blood patch injection (ABPI) with that of a hydrogel plug on the rate of pneumothorax at CT-guided percutaneous lung biopsy. Materials and Methods In this prospective randomized controlled trial ( https://ClinicalTrials.gov , NCT02224924), a noninferiority design was used for ABPI, with a 10% noninferiority margin when compared with the hydrogel plug, with the primary outcome of pneumothorax rate within 2 hours of biopsy. A type I error rate of 0.05 and 90% power were specified with a target study population of 552 participants (276 in each arm). From October 2014 to February 2017, all potential study participants referred for CT-guided lung biopsy (n = 2052) were assessed for enrollment. Results The data safety monitoring board recommended the trial be closed to accrual after an interim analysis met prespecified criteria for early stopping based on noninferiority. The final study group consisted of 453 participants who were randomly assigned to the ABPI (n = 226) or hydrogel plug (n = 227) arms. Of these, 407 underwent lung biopsy. Pneumothorax rates within 2 hours of biopsy were 21% (42 of 199) and 29% (60 of 208); chest tube rates were 9% (18 of 199) and 13% (27 of 208); and delayed pneumothorax rates within 2 weeks after biopsy were 1.4% (three of 199) and 1.5% (three of 208) in the ABPI and hydrogel plug arms, respectively. Conclusion Autologous blood patch injection is noninferior to a hydrogel plug regarding the rate of pneumothorax after CT-guided percutaneous lung biopsy. © RSNA, 2018 Online supplemental material is available for this article.

DAXX Mutation Status of Embolization-Treated Neuroendocrine Tumors Predicts Shorter Time to Hepatic Progression.

PURPOSE: To identify common gene mutations in patients with neuroendocrine liver metastases (NLM) undergoing transarterial embolization (TAE) and establish relationship between these mutations and response to TAE. MATERIALS AND METHODS: Patients (n = 51; mean age 61 y; 29 men, 22 women) with NLMs who underwent TAE and had available mutation analysis were identified. Mutation status and clinical variables were recorded and evaluated in relation to hepatic progression-free survival (HPFS) (Cox proportional hazards) and time to hepatic progression (TTHP) (competing risk proportional hazards). Subgroup analysis of patients with pancreatic NLM was performed using Fisher exact test to identify correlation between mutation and event (hepatic progression or death) by 6 months. Changes in mutation status over time and across specimens in a subset of patients were recorded. RESULTS: Technical success of TAE was 100%. Common mutations identified were MEN1 (16/51; 31%) and DAXX (13/51; 25%). Median overall survival was 48.7 months. DAXX mutation status (hazard ratio = 6.21; 95% confidence interval [CI], 2.67-14.48; P < .001) and tumor grade (hazard ratio = 3.05; 95% CI, 1.80-5.17; P < .001) were associated with shorter HPFS and TTHP on univariate and multivariate analysis. Median HPFS was 3.6 months (95% CI, 1.7-5.3) for patients with DAXX mutation compared with 8.9 months (95% CI, 6.6-11.4) for patients with DAXX wild-type status. In patients with pancreatic NLMs, DAXX mutation status was associated with hepatic progression or death by 6 months (P = .024). DAXX mutation status was concordant between primary and metastatic sites. CONCLUSIONS: DAXX mutation is common in patients with pancreatic NLMs. DAXX mutation status is associated with shorter HPFS and TTHP after TAE.

Does Enhancement or Perfusion on Preprocedure CT Predict Outcomes After Embolization of Hepatocellular Carcinoma?

RATIONALE AND OBJECTIVE: The objective of this study was to evaluate whether quantitative enhancement or perfusion measurements on preprocedure triphasic computed tomography (CT) can be used to predict response or overall survival after embolization of hepatocellular carcinoma. MATERIALS AND METHODS: The institutional review board approved this retrospective review of 63 patients with hepatocellular carcinoma treated with particle embolization between March 2009 and December 2014. Quantitative enhancement and perfusion measurements were performed on the target tumor and the background liver on the triphasic CT performed before treatment. Microvascular invasion (MVI) and degree of differentiation were determined from a core biopsy specimen. Quantitative enhancement and perfusion values were then correlated with pathology (two-tailed t test), response to embolization on modified Response Evaluation Criteria In Solid Tumors (two-tailed t test), and overall survival after embolization (Cox proportional hazards model). RESULTS: Arterial enhancement did not predict immediate response or overall survival after embolization. The degree of differentiation or presence of MVI also did not predict immediate response or overall survival after embolization. However, high hepatic artery coefficient or low portal vein coefficient, both in the tumor (P = .011 and P = .004) and in the background liver (P = .015 and P = .009), were associated with worse survival. Hepatic artery coefficient, both in the tumor (P = .025) and in the background liver (P = .013), were independent predictors of survival in a multivariate model including the Child-Pugh score and the BCLC stage. CONCLUSIONS: Tumor and liver perfusion parameters estimated from preprocedure triphasic CT were predictive of survival after embolization. Arterial-phase enhancement and histology (degree of differentiation or MVI) did not predict immediate response or overall survival after particle embolization.

Comparison of biliary brush biopsy and fine needle biopsy in the diagnosis of biliary strictures.

PURPOSE: The purpose of this study is to evaluate the accuracy of percutaneous fine needle biopsy (FNB) and brush biopsy (BB) at a cancer center. MATERIAL AND METHODS: Retrospective analysis of all bile duct biopsies performed in Interventional Radiology between January 2000 and January 2015 was performed. FNB was performed under real-time cholangiographic guidance using a notched needle directed at the bile duct stricture. BB was performed by advancing a brush across the stricture and moving it back and forth to scrape the stricture. Biopsy results were categorized as true positive (TP), true negative (TN), false positive (FP) and false negative (FN) based on pathology reports and confirmed by surgical specimens or clinical follow-up of at least six months. Fisher's exact test was used to compare the rate of TP in FNB and BB. RESULTS: One-hundred and nineteen patients underwent FNB or BB. Fifteen were censored because of lack of follow-up. The remaining 104 patients underwent a total of 117 bile duct biopsies during the study period: 34 FNB and 83 BB. There were no complications in either group. In the FNB group 22/34 (64%) biopsies were TP, 4/34(12%) were TN and there were 8(24%) FN biopsies. In the BB group, 20/83 (24%) were TP, 38/83 (46%) TN and 25/83 (30%) FN biopsies. There were no FP biopsies in either group. The sensitivity of detecting malignancy by FNB was significantly higher than that by BB (73% vs 44%, p < .0005). There were no complications associated with FNB or BB. CONCLUSIONS: FNB of bile duct strictures is safe and has a higher sensitivity for detecting malignancy than BB.

Symptomatic Fluid Drainage: Peritoneovenous Shunt Placement.

Ascites causes significant discomfort and has negative impact on patient's quality of life. Medical therapies including dietary restriction and diuretics are successful in only 40 to 44% of patients with malignant ascites and repeated paracentesis only provides temporary symptomatic relief. Therefore, a more permanent solution is necessary. Indwelling catheters or peritoneovenous shunt placement can provide more permanent symptomatic relief and improve patients' quality of life. Unlike indwelling catheters, peritoneovenous shunts do not limit patients' life style and therefore should be offered as first option in patients who are good candidates. Denver shunt (CareFusion-BD Worldwide) is the current available peritoneovenous shunt. In this article, the indications, contraindications, technical aspects of shunt placement, and techniques to prevent postprocedure complications will be discussed.

Effect of abdominopelvic abscess drain size on drainage time and probability of occlusion.

BACKGROUND: The purpose of this study is to determine whether larger abdominopelvic abscess drains reduce the time required for abscess resolution or the probability of tube occlusion. METHODS: 144 consecutive patients who underwent abscess drainage at a single institution were reviewed retrospectively. RESULTS: Larger initial drain size did not reduce drainage time, drain occlusion, or drain exchanges (P > .05). Subgroup analysis did not find any type of collection that benefitted from larger drains. A multivariate model predicting drainage time showed that large collections (>200 mL) required 16 days longer drainage time than small collections (<50 mL). Collections with a fistula to bowel required 17 days longer drainage time than collections without a fistula. Initial drain size and the viscosity of the fluid in the collection had no significant effect on drainage time in the multivariate model. CONCLUSIONS: 8 F drains are adequate for initial drainage of most serous and serosanguineous collections. 10 F drains are adequate for initial drainage of most purulent or bloody collections.

Peripheral Blood Regulatory T-Cell and Type 1 Helper T-Cell Population Decrease after Hepatic Artery Embolization.

PURPOSE: To evaluate changes in T-cell populations in peripheral blood after bland hepatic artery embolization (HAE). MATERIALS AND METHODS: Bland HAE was performed in 12 patients to treat primary (n = 5) or metastatic (n = 7) liver tumors, using microspheres and polyvinyl alcohol (n = 8) or microspheres alone (n = 4). Patient peripheral blood samples were collected within 1 month before HAE, within 1 week after HAE (early period after HAE), and 2-8 weeks after HAE (follow-up period). Peripheral blood populations of cytotoxic T lymphocytes, CD4(+) T cells, type 1 helper T cells (Th1) and type 2 helper T cells (Th2), and regulatory T cells (Treg) were evaluated using flow cytometry. Changes in T-cell populations before and after bland HAE were compared using paired t tests. RESULTS: Peripheral blood CD4(+) T-cell populations decreased significantly in the early period after HAE (44.0% ± 2.2 to 34.4% ± 3.6, P < .01) and in the follow-up period (44.0% ± 2.2 to 36.3% ± 3.0, P < .01). Among the individual CD4(+) T-cell subtypes, Treg (2.5% ± 0.3 to 1.7% ± 0.2, P < .02) and Th1 (8.1% ± 1.8 to 5.6% ± 1.6, P < .02) decreased significantly in the early period after HAE only. The presence of extrahepatic disease was associated with decreasing Treg (P < .04). CONCLUSIONS: After HAE, the peripheral blood T-cell environment is changed with decreases in Treg and Th1.

Therapeutic Application of Percutaneous Peritoneovenous (Denver) Shunt in Treating Chylous Ascites in Cancer Patients.

PURPOSE: To evaluate the safety and efficacy of percutaneous peritoneovenous shunt (PPVS) placement in treating intractable chylous ascites (CA) in patients with cancer. MATERIALS AND METHODS: Data from 28 patients with refractory CA treated with PPVS from April 2001 to June 2015 were reviewed. Demographic characteristics, technical success, efficacy, laboratory values, and complications were recorded. Univariate and multivariate logistic regression analysis was performed. RESULTS: Technical success was 100%, and ascites resolved or symptoms were relieved in 92.3% (26 of 28) of patients. In 13 (46%) patients with urologic malignancies, whose ascites had resulted from retroperitoneal lymph node dissection, the ascites resolved, resulting in shunt removal within 128 days ± 84. The shunt provided palliation of symptoms in 13 of the remaining 15 patients (87%) for a mean duration of 198 days ± 214. Serum albumin levels increased significantly (21.4%) after PPVS placement from a mean of 2.98 g/dL ± 0.64 before the procedure to 3.62 g/dL ± 0.83 (P < .001). The complication rate was 37%, including shunt malfunction/occlusion (22%), venous thrombosis (7%), and subclinical disseminated intravascular coagulopathy (DIC) (7%). Smaller venous limb size (11.5 F) and the presence of peritoneal tumor were associated with a higher rate of shunt malfunction (P < .05). No patient developed overt DIC. CONCLUSIONS: PPVS can safely and effectively treat CA in patients with cancer, resulting in significant improvement in serum albumin in addition to palliation of symptoms.

Randomized Trial of Hepatic Artery Embolization for Hepatocellular Carcinoma Using Doxorubicin-Eluting Microspheres Compared With Embolization With Microspheres Alone.

PURPOSE: Transarterial chemoembolization is accepted therapy for hepatocellular carcinoma (HCC). No randomized trial has demonstrated superiority of chemoembolization compared with embolization, and the role of chemotherapy remains unclear. This randomized trial compares the outcome of embolization using microspheres alone with chemoembolization using doxorubicin-eluting microspheres. MATERIALS AND METHODS: At a single tertiary referral center, patients with HCC were randomly assigned to embolization with microspheres alone (Bead Block [BB]) or loaded with doxorubicin 150 mg (LC Bead [LCB]). Random assignment was stratified by number of embolizations to complete treatment, and assignments were generated by permuted blocks in the institutional database. The primary end point was response according to RECIST 1.0 (Response Evaluation Criteria in Solid Tumors) using multiphase computed tomography 2 to 3 weeks post-treatment and then at quarterly intervals, with the reviewer blinded to treatment allocation. Secondary objectives included safety and tolerability, time to progression, progression-free survival, and overall survival. This trial is currently closed to accrual. RESULTS: Between December 2007 and April 2012, 101 patients were randomly assigned: 51 to BB and 50 to LCB. Demographics were comparable: median age, 67 years; 77% male; and 22% Barcelona Clinic Liver Cancer stage A and 78% stage B or C. Adverse events occurred with similar frequency in both groups: BB, 19 of 51 patients (38%); LCB, 20 of 50 patients (40%; P = .48), with no difference in RECIST response: BB, 5.9% versus LCB, 6.0% (difference, -0.1%; 95% CI, -9% to 9%). Median PFS was 6.2 versus 2.8 months (hazard ratio, 1.36; 95% CI, 0.91 to 2.05; P = .11), and overall survival, 19.6 versus 20.8 months (hazard ratio, 1.11; 95% CI, 0.71 to 1.76; P = .64) for BB and LCB, respectively. CONCLUSION: There was no apparent difference between the treatment arms. These results challenge the use of doxorubicin-eluting beads for chemoembolization of HCC.

Symptomatic Perihepatic Fluid Collections After Hepatic Resection in the Modern Era.

BACKGROUND: Improvements in liver surgery have led to decreased mortality rates. Symptomatic perihepatic collections (SPHCs) requiring percutaneous drainage remain a significant source of morbidity. STUDY DESIGN: A single institution's prospectively maintained hepatic resection database was reviewed to identify patients who underwent hepatectomy between January 2004 and February 2012. RESULTS: Data from 2173 hepatectomies performed in 2040 patients were reviewed. Overall, 200 (9%) patients developed an SPHC, the majority non-bilious (75.5%) and infected (54%). Major hepatic resections, larger than median blood loss (≥360 ml), use of surgical drains, and simultaneous performance of a colorectal procedure were associated with an SPHC on multivariate analysis. Non-bilious, non-infected (NBNI) collections were associated with lower white blood cell (WBC) counts, absence of a bilio-enteric anastomosis, use of hepatic arterial infusion pump (HAIP), and presence of metastatic disease, and resolved more frequently with a single interventional radiology (IR) procedure (85 vs. 46.5%, p < 0.001) more quickly (15 vs. 30 days, p = 0.001). CONCLUSIONS: SPHCs developed in 9% of patients in a modern series of hepatic resections, and in one third were non-bilious and non-infected. In the era of modern interventional radiology, the need for re-operation for SPHC is exceedingly rare. A significant proportion of minimally symptomatic SPHC patients may not require drainage, and strategies to avoid unnecessary drainage are warranted.

Corrigendum to "Asymptomatic Liver Abscesses Mimicking Metastases in Patients after Whipple Surgery: Infectious Complications following Percutaneous Biopsy-A Report of Two Cases".

[This corrects the article DOI: 10.1155/2012/817314.].

Safety and efficacy of percutaneous cecostomy/colostomy for treatment of large bowel obstruction in adults with cancer.

PURPOSE: To assess the safety and efficacy of image-guided percutaneous cecostomy/colostomy (PC) in the management of colonic obstruction in patients with cancer. MATERIALS AND METHODS: Twenty-seven consecutive patients underwent image-guided PC to relieve large bowel obstruction at a single institution between 2000 and 2012. Colonic obstruction was the common indication. Patient demographics, diagnosis, procedural details, and outcomes including maximum colonic distension (MCD; ie, greatest transverse measurement of the colon on radiograph or scout computed tomography image) were recorded and retrospectively analyzed. RESULTS: Following PC, no patient experienced colonic perforation; pain was relieved in 24 of 27 patients (89%). Catheters with tip position in luminal gas rather than mixed stool/gas or stool were associated with greater decrease in MCD (-40%, -12%, and -16%, respectively), with the difference reaching statistical significance (P = .002 and P = .013, respectively). Catheter size was not associated with change in MCD (P = .978). Catheters were successfully removed from six of nine patients (67%) with functional obstructions and two of 18 patients (11%) with mechanical obstructions. One patient underwent endoscopic stent placement after catheter removal. Three patients required diverting colostomy after PC, and their catheters were removed at the time of surgery. One major complication (3.7%; subcutaneous emphysema, pneumomediastinum, and sepsis) occurred 8 days after PC and was successfully treated with cecostomy exchange, soft-tissue drainage, and intravenous antibiotic therapy. CONCLUSIONS: Image-guided PC is safe and effective for management of functional and mechanical bowel obstruction in patients with cancer. For optimal efficacy, catheters should terminate within luminal gas.

Successful control of liver metastases from pancreatic solid-pseudopapillary neoplasm (SPN) using hepatic arterial embolization.

No systemic agents that are known to be effective for the treatment of solid-pseudopapillary neoplasm (SPN) are available. We report the prolonged and sustained control of metastatic pancreatic SPN to the liver using hepatic arterial embolization (HAE), where a total of 13 HAE sessions were performed over a 6-year period.

Hepatic artery embolization for liver metastasis of gastrointestinal stromal tumor following imatinib and sunitinib therapy.

PURPOSE: The purpose of the study is to determine the efficacy of hepatic artery embolization (HAE) as a therapy for gastrointestinal stromal tumor (GIST) in patients who are refractory to imatinib and sunitinib. METHODS: After institutional review board approval, a retrospective review revealed 11 patients with GIST metastatic to the liver who underwent 15 HAEs between February 2002 and May 2013. These patients were stratified into two groups according to the previous treatment: (a) those treated with HAE as second-line treatment after failing first-line imatinib (n = 3) and (b) those treated with HAE as third-line therapy after failing first-line imatinib and second-line sunitinib (n = 8). Initial therapeutic response, overall survival (OS), progression-free survival (PFS), and safety were evaluated. RESULTS: Initial therapeutic response rates at 3 months after HAE were 27.3 % (95 % confidence interval (CI), 6.0-61.0 %) by Response Evaluation Criteria in Solid Tumor (RECIST) version 1.0 and 45.5 % (95 % CI, 16.7-76.6 %) by modified RECIST (mRECIST). The median OS and PFS after HAE were 14.9 and 3.9 months in group A and 23.8 and 3.4 months in group B, respectively. No procedure-related mortality or major complication was observed. CONCLUSIONS: HAE is an effective and well-tolerated therapeutic option for GIST liver metastases. Although larger studies are necessary, HAE should be considered as an alternative or adjuvant to third-line or even second-line systemic treatment.