1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Radioembolization (RE)-induced liver disease (REILD) has been defined as jaundice and ascites appearing 1 to 2 months after RE in the absence of tumor progression or bile duct occlusion. Our aims were to study the incidence of REILD in a large cohort of patients and the impact of a series of changes introduced in the processes of treatment design, activity calculation, and the routine use of ursodeoxycholic acid and low-dose steroids (modified protocol). Between 2003 and 2011, 260 patients with liver tumors treated by RE were studied (standard protocol: 75, modified protocol: 185). REILD appeared only in patients with cirrhosis or in noncirrhosis patients exposed to systemic chemotherapy prior to RE. Globally, the incidence of REILD was reduced in the modified protocol group from 22.7% to 5.4% and the incidence of severe REILD from 13.3% to 2.2% (P < 0.0001). Treatment efficacy was not jeopardized since 3-month disease control rates were virtually identical in both groups (66.7% and 67.2%, P = 0.93). Exposure to chemotherapy in the 2-month period following RE and being treated by the standard protocol were independent predictors of REILD among noncirrhosis patients. In cirrhosis, the presence of a small liver (total volume <1.5 L), an abnormal bilirubin (>1.2 mg/dL), and treatment in a selective fashion were independently associated with REILD. Conclusion: REILD is an uncommon but relevant complication that appears when liver tissue primed by cirrhosis or prior and subsequent chemotherapy is exposed to the radiation delivered by radioactive microspheres. We designed a comprehensive treatment protocol that reduces the frequency and the severity of REILD. (HEPATOLOGY 2013)

Radioembolization (RE) is a form of brachytherapy in which intraarterially injected microspheres loaded with Yttrium-90 [90Y] (a pure beta emitter with a 2.6-day half-life and an average 2.5-mm tissue penetration) serve as sources of internal radiation. In one treatment session, tumoricidal doses of radiation can be delivered to multiple liver tumors while sparing the nontumoral liver from harmful doses of radiation. This aim is accomplished by millions of sources of radiation released into the hepatic arteries, which are delivered preferentially to the tumors because of their higher arterial blood supply. However, the nontumoral liver absorbs a certain dose of radiation, which to date cannot be accurately predicted (for a more detailed discussion about dosimetry of RE see Ref.1).

The most frequent liver toxicity observed after external beam radiation is radiation-induced liver disease (RILD). Patients with RILD, also called anicteric ascites, develop ascites accompanied by raised alkaline phosphatase levels.2 Factors that impact the occurrence of RILD include the dose of radiation delivered, the volume of liver tissue involved, cirrhosis, prior liver function, and concurrent use of other antineoplastic therapies.3 A different and more disquieting syndrome may be observed when whole-body irradiation together with ablative doses of chemotherapy are used before bone marrow transplantation. In this so-called combined modality-induced liver disease (CMILD), patients develop jaundice, ascites, and fluid retention and their blood tests mainly show increased bilirubin levels.2 In 2006, our group described a clinical syndrome that complicated the course of patients with liver tumors treated by RE and consisted of jaundice and ascites appearing 1 to 2 months after treatment in the absence of tumor progression or bile duct occlusion. We called this radioembolization-induced liver disease (REILD) to differentiate it from RILD and an association with a younger age and higher treatment intensity was suggested.4 The aim of this study was to analyze the incidence of REILD in a large cohort of patients with and without cirrhosis and the impact of a series of changes to the treatment design, activity calculation, and concomitant medication.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Population Under Study.

Consecutive patients treated by RE between September 2003 and September 2011 at our institution were analyzed. Patients were considered to have cirrhosis determined by either histological or clinical-radiological criteria, including the presence of signs of cirrhosis on liver imaging (nodular liver contour, atrophy of the right lobe with hypertrophy of the left and caudate lobe) or by the presence of laboratory or imaging signs of portal hypertension (platelet count <100/pL, maximal spleen diameter >12 cm, or the presence of portal-collateral vessels). Only those patients who had a minimum follow-up of 2 months were included in the analysis.

Candidates for Radioembolization.

Patients were considered for RE provided that they had an unequivocal diagnosis of unresectable primary or metastatic liver cancer. The indication for RE was established by the HPB Oncology Multidisciplinary Team on the basis of the individual oncological history and available therapies. However, this has evolved over time due to the experience gained in RE, the advent of new systemic treatments, and changes in the treatment paradigm of each primary tumor. As a general rule, all patients with chemosensitive tumors had been exposed to at least one and usually several lines of chemotherapy. Histological confirmation was not obtained for those patients with a history of biopsy-proven extrahepatic primary liver tumors in which imaging procedures were consistent with liver metastases, or for those who met noninvasive criteria for hepatocellular carcinoma.5 Moreover, patients were only considered for RE if they had an Eastern Cooperative Oncology Group (ECOG) performance status not higher than 2, as well as preserved liver (serum total bilirubin <3 mg/dL), hematological (platelet count >40/pL), and renal function (serum creatinine <2 mg/dL), no contraindication to angiography, and were able to provide informed consent. Authorization for this retrospective study was obtained from the local Ethics Committee.

Treatment Procedure and Activity Calculation.

RE was performed as described6 using 90Y-labeled resin microspheres (Sir-Spheres, Sirtex Medical Europe, Bonn, Germany). Activity was initially calculated by either the body surface area (BSA) method (in which activity is empirically calculated based on liver and tumor volumes) or the Partition Model (which assumes a nonuniform distribution that can be estimated from the injection of 99mTc-macroaggregated albumin [MAA]). After our initial description of REILD,4 a number of major changes in the criteria for treatment design and the process of activity calculation in our institution were gradually introduced, together with the routine use of concomitant medication. The threshold for contraindicating RE due to liver dysfunction was reduced from 3 to 2 mg/dL of total bilirubin. Selective injections were used to spare as much liver tissue as possible from irradiation and a number of changes were introduced in the way activity is calculated. They are summarized in Fig. 1 and fully described in a Supporting section available online. The BSA formula was used to calculate the activity for those patients in whom essentially the entire liver volume was exposed to some radiation (whole-liver treatment), and the Partition Model was used when at least two segments of the liver were spared from exposure to radiation (selective treatment). For whole-liver treatment, a 10%-20% reduction in the calculated activity was applied on a general basis, and this reduction could be particularly intense for those patients with cirrhosis or heavily exposed to prior chemotherapy who had a small tumor burden or a reduced liver volume. For selective treatment, the Partition Model was used in two different ways according to the treatment aim and the importance of preserving the integrity of the nontumoral liver parenchyma. When large volumes were targeted, particularly in cirrhosis patients with hepatocellular carcinoma, the model was used to estimate a very safe activity for the nontumoral compartment. By contrast, when small volumes of liver were targeted or patients had an excellent liver function, the model was used to calculate a most effective activity, irrespective of the dose delivered to the surrounding liver tissue. This last procedure has recently been described as radiation segmentectomy,7 although ablation of the target area was neither the aim nor was it always caused. Besides these modifications in the treatment design and activity calculation, ursodeoxycholic acid was also given for 2 months starting the day of RE at a dose of 300 mg twice a day orally, in addition to methyl-prednisolone 8 mg once a day for 1 month and 4 mg once a day for a second month. This decision was based on the results of two randomized controlled trials that evaluated the efficacy of ursodeoxycholic acid prophylaxis for CMILD.8, 9 Hereafter, by modified protocol we will refer to the use of this modified system for activity calculation alongside ursodeoxycholic acid and steroid therapy as opposed to the standard protocol used before September 2006.

thumbnail image

Figure 1. Work-up of activity calculation in our modified protocol. NTC, nontumor compartment; TC, tumor compartment.

Download figure to PowerPoint

Follow-up and Variables Under Study.

We excluded from analysis those patients who died or were lost to follow-up within the first 2 months (most of them were patients who came from distant centers in Spain or abroad). Although follow-up was variable according to the therapeutic strategy individually designed for each patient, imaging procedures were available for all between 2 and 4 months post-RE. More than 100 pre- and posttreatment variables were prospectively collected for each patient including general characteristics at baseline and at the time of first diagnosis of cancer (from alcohol consumption to presence of steatosis on imaging or liver function tests), relevant features of prior cancer therapies (liver resection or liver-directed therapies, exposure to different chemotherapy agents), liver status at the time of RE, characteristics of the procedures (target area or treatment intensity), and chemotherapy or other relevant therapies (ursodeoxycholic acid or steroids) received after RE. REILD was defined as the appearance of a serum total bilirubin of 3 mg/dL or higher and ascites (clinically or by imaging) 1 to 2 months after RE in the absence of tumor progression or bile duct obstruction. According to the NCI Common Toxicity Criteria for Adverse Events (CTCAE) v. 3.0,10 REILD was considered to be grade 3 or higher when in the first 3 months after RE the maximal bilirubin level was higher than 3.6 mg/dL or a paracentesis was deemed necessary for the evacuation of ascites. To estimate the influence of the above-mentioned changes in treatment outcome, the status of tumor disease within the liver according to RECIST criteria11 at 3 months was also recorded. Local disease control was established when no new liver lesions had appeared and the sum of maximal diameters of target lesions had not increased over 20%.

Statistical Analysis.

All statistical analysis was performed by using SPSS for Windows, v. 13.0.1 (Chicago, IL). Data are presented as mean (standard deviation) for continuous variables with a normal distribution and as median (interquartile range) for continuous variables with a nonnormal distribution. For comparison between groups, a univariate analysis was first performed using Student's t test or Mann-Whitney's U test for continuous variables and the chi-square test or Fischer exact test for categorical variables. Those variables that were both clinically relevant and statistically significant in the univariate analysis at the level of P = 0.1 were used to construct multivariate models using logistic regression analysis. The Hosmer-Lemeshow test was used to confirm the goodness of fit of the models. To test the reliability and consistency of the modified method for activity calculation, internal consistency and test-retest reliability were assayed. Two experienced investigators (M.I. and B.S.) were provided with all the essential factors (prior history of resection, chemotherapy, and cirrhosis, treatment intention, volumes, BSA, T/N ratio, etc.) and were independently asked to prescribe the activity for each and every patient in a consecutive manner. Internal consistency was expressed as a Cronbach's α value (a coefficient of reliability that also measures the internal consistency of any method that produces a numeric result) and test-retest reliability as an intraclass correlation coefficient (commonly used in the assessment of reproducibility of quantitative measurements made by different observers measuring the same quantity since it operates on data structured as groups, rather than paired observations). All tests were two-tailed and the level of statistical significance was established for P < 0.05.


  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information


Between September 2003 and September 2011, 299 patients with primary or secondary liver tumors, one-third of them with cirrhosis, were treated by RE. Clinical, laboratory, or imaging follow-up for the first 2 months after RE were not available for 39 patients. Thirteen of these 39 patients died in the first 2 months, all of them for reasons other than suspected REILD and without jaundice or ascites. The baseline characteristics for the remaining 260 patients are presented in Tables 1 and 2 according to the presence of cirrhosis.

Table 1. Baseline Characteristics of 88 Cirrhosis Patients Treated by Radioembolization
CharacteristicsGlobal SeriesStandard ProtocolModified ProtocolP
  1. Variables are expressed as median (IQR). TA(C)E: transarterial bland embolization or chemoembolization.

Number of patients882464 
Age (years)67 (57-72)62 (53-71)67 (61-73)0.045
Male Sex73 (83%)20 (83.3%)53 (82.8%)0.614
Primary tumor
 Hepatocellular carcinoma86 (97.7%)23 (95.8%)63 (98.4%)0.473
 Cholangiocarcinoma2 (2.3%)1 (4.2%)1 (1.6%)0.897
Prior Therapies    
 Resection11 (12.5%)2 (8.3%)9 (14.1%)0.374
 TA(C)E28 (31.8%)14 (58.3%)14 (21.9%)0.002
 Prior chemotherapy3 (3.4%)1 (4.2%)2 (3.1%)0.620
Data at the time of radioembolization
 Total bilirubin (mg/dL)1.15 (0.77-1.51)1.22 (0.76-1.45)1.13 (0.77-1.68)0.864
 ALT (IU/mL)36 (21-56)45 (30.5-78.5)32 (20-53)0.133
 GGTP (IU/mL)96 (58-171)86 (61-176.5)101 (52-171)0.819
 Leukocyte count (/pL)4.9 (3.7-6.6)4.5 (3.2-6.7)4.9 (3.8-6.5)0.656
 Platelet count (/pL)126 (94-188)125 (88-214)126 (95-181)0.563
 Prothrombin activity (%)84 (71-95)78 (69-85)87 (73-97)0.021
 Tumor volume (mL)115 (54-344)189 (39-1100)101 (54-258)0.089
 Total liver volume (L)1.78 (1.36-2.20)1.86 (1.25-2.68)1.72 (1.42-2.15)0.397
 Tumor involvement (%)8.6 (3.7-27.4)9.2 (3.7-48.8)8 (3.5-17.6)0.153
Medications in the 2 months following RE
 Ursodeoxycholic acid62 (70.5%)062 (96.9%) 
 Methyl-prednisolone63 (71.6%)063 (98.4%) 
 Chemotherapy post-RE2 (2.3%)02 (3.1%)0.527
Table 2. Baseline Characteristics of 172 Noncirrhosis Patients Treated by Radioembolization
CharacteristicsGlobal SeriesStandard ProtocolModified ProtocolP
  1. Variables are expressed as median (IQR). TA(C)E: transarterial bland embolization or chemoembolization.

Number of patients17251121 
Age (years)60 (50-69)56 (46-61)63 (52-70)0.002
Male Sex98 (57%)38 (74.5%)60 (49.6%)0.003
Primary tumor
 Hepatocellular carcinoma27 (15.7%)12 (23.5%)15 (12.4%)0.057
 Colorectal cancer67 (39%)16 (31.4%)51 (42.1%)0.231
 Neuroendocrine tumors25 (14.5%)7 (13.7%)18 (14.9%)0.845
 Others47 (27.3%)13 (25.5%)34 (28.1%)0.726
Prior therapies
 Resection23 (13.4%)10 (19.6%)13 (10.7%)0.142
 TA(C)E13 (7.6%)7 (13.7%)6 (5%)0.060
 Prior chemotherapy116 (67.4%)36 (70.6%)80 (66.1%)0.598
Data at the time of radioembolization
 Total bilirubin (mg/dL)0.61 (0.42-0.94)0.65 (0.54-0.98)0.58 (0.4-0.89)0.028
 ALT (IU/mL)20 (14-29)20 (14-29)20 (13-30)0.810
 GGTP (IU/mL)82 (39-183)112 (50-210)78 (34-180)0.085
 Leukocyte count (/pL)6.1 (4.9-8.2)6.6 (4.4-9.1)6 (5-7.7)0.700
 Platelet count (/pL)227 (163-282)222 (174-284)228 (160-282)0.704
 Prothrombin activity (%)96 (86-101)96 (86-100)96 (85-106)0.181
 Tumor volume (mL)182 (61-499)239 (66-655)161 (50-493)0.256
 Total liver volume (L)1.70 (1.36-2.15)1.85 (1.58-2.36)1.63 (1.32-2.10)0.033
 Tumor involvement (%)12.8 (3.8-27.5)15.4 (3.8-33.6)10.6 (3.6-26.2)0.507
Medications in the 2 months following RE
 Ursodeoxycholic acid93 (54.1%)093 (76.9%) 
 Methyl-prednisolone98 (57%)098 (81%) 
 Chemotherapy post-RE56 (32.6%)13 (25.5%)43 (35.5%)0.217

Almost all cirrhosis patients had hepatocellular carcinomas and very rarely received chemotherapy either before or after RE. The two subgroups of cirrhosis patients treated by standard and modified protocols (Table 1) were well balanced in terms of liver function and volume, tumor burden, and prior surgery. However, patients in the modified protocol group were significantly older and had received transarterial chemoembolization (TACE) less frequently. Noncirrhosis patients had a broader range of primary tumors, mainly metastatic gastrointestinal tumors including colorectal carcinomas. Most of them had received prior chemotherapy but one-third were chemonaive, usually those with metastatic neuroendocrine tumors or hepatocellular carcinoma developing in a noncirrhosis liver, and one-third of them received chemotherapy in the 2 months following RE. The two subgroups of noncirrhosis patients treated by standard and modified protocols were well balanced in terms of frequency of primary tumors, tumor burden, and post-RE chemotherapy, or in the values of most liver function tests. However, patients in the modified protocol group were significantly older and more frequently females, and had smaller livers and lower bilirubin levels.

Treatment Characteristics.

The reliability and consistency of the modified algorithm for activity calculation, measured by asking two different prescribing physicians to recalculate the activity for each patient in a blind fashion, was very high, as shown by a Cronbach's α value of 0.963 and an intraclass correlation coefficient of 0.963 (0.945-0.975). It is known that occasionally the prescribed activity cannot be injected due to blood flow stasis. This happened in 9% of our patients, and the median fraction of the prescribed dose that was actually injected in these patients was 66% (range: 42%-92%). As shown in Table 3, selective treatments were performed in 80.7% of cirrhosis patients and 36.6% of noncirrhosis patients. Globally, cirrhosis and noncirrhosis patients treated with the modified protocol had lower target liver volumes and received less activity both absolute and relative to target volume. When stratified according to treatment fashion, the prescribed activity was reduced for all subgroups of patients by using the modified protocol. Activity related to target volume (that somehow reflects treatment intensity) was also reduced in all subgroups when the modified protocol was used but differences were statistically significant only for noncirrhosis patients receiving whole-liver treatment.

Table 3. Injected Activity Absolute and Relative to Targeted Liver Volume According to the Presence of Cirrhosis and Treatment Design
Noncirrhosis PatientsCirrhosis Patients
All treatments (n=172)All treatments (n=88)
 Standard (n=51)Modified (n=121)P Standard (n=24)Modified (n=64)P
Activity (GBq)1.74 (1.3-2.1)1.2 (0.9-1.6)0.001Activity (GBq)1.99 (1.22-2.23)1 (0.6-1.5)0.001
Target volume (L)1.65 (1.34-2.09)1.45 (1.12-1.80)0.022Target volume (L)1.47 (0.87-2.22)0.90 (0.48-1.42)0.005
Activity/target volume (GBq/L)1.01 (0.74-1.22)0.83 (0.7-1.02)0.017Activity/target volume (GBq/L)1.04 (0.86-1.76)1.20 (0.85-1.78)0.005
Whole-liver treatments (n=109, 63.4%)Whole-liver treatments (n=17, 19.3%)
 Standard (n=37)Modified (n=72)P Standard (n=11)Modified (n=6)P
Activity (GBq)1.74 (1.36-2.06)1.32 (1.0-1.6)0.001Activity (GBq)1.94 (1.53-2.20)1.30 (1.19-1.75)0.180
Target volume (L)1.75 (1.41-2.35)1.59 (1.32-2.02)0.082Target volume (L)1.55 (1.29-2.80)1.84 (1.39-2.36)0.961
Activity/target volume (GBq/L)1 (0.76-1.15)0.77 (0.64-0.90)0.003Activity/target volume (GBq/L)0.98 (0.69-1.14)0.76 (0.59-1.04)0.301
Selective treatment (n=63, 36.6%)Selective treatment (n=71, 80.7%)
 Standard (n=14)Modified (n=49)P Standard (n=13)Modified (n=58)P
Activity (GBq)1.46 (1.23-2.25)1.1 (0.8-1.5)0.009Activity (GBq)2.05 (0.97-2.75)1 (0.60-1.52)0.004
Target volume (L)1.26 (0.83-1.82)1.17 (0.83-1.54)0.612Target volume (L)1.13 (0.49-2.03)0.85 (0.40-1.31)0.281
Activity/target volume (GBq/L)1.34 (0.71-2.17)0.98 (0.79-1.16)0.233Activity/target volume (GBq/L)1.62 (0.98-3.39)1.27 (0.91-1.92)0.228


As summarized in Table 4, REILD occurred in patients with and without cirrhosis. The patient profile was nevertheless different. In noncirrhosis patients, with a single exception, REILD occurred among patients previously exposed to chemotherapy and receiving whole-liver treatments, and the use of the modified protocol resulted in a marked reduction in the incidence and severity of REILD. In cirrhosis, REILD occurred among patients receiving whole-liver or selective treatments, and the use of the modified protocol helped reducing the incidence and severity in both subgroups. Globally, the modified protocol resulted in a reduction in the incidence of REILD from 22.7% to 5.4% (P < 0.001). The severity of REILD was also improved as illustrated by a decrease in CTCAE grade 3 or higher events from 13.3% to 2.2% (P = 0.001). In most patients, REILD was a stable or transient condition. However, death REILD was probably the cause of death in four patients treated by the standard protocol and one patient treated by the modified protocol (5.3% versus 0.5%, respectively; P = 0.025). Importantly, the reduction in the activity prescribed by the modified protocol to patients treated in a whole-liver fashion did not seem to jeopardize the final outcome, as indicated by overlapping 3-month disease control rates.

As shown in Table 5, noncirrhosis patients more likely to develop REILD were those below 60 years, submitted to liver resection, previously treated by chemotherapy (specifically capecitabine), with higher serum levels of gamma-glutamyl transpeptidase (GGT), treated in a whole-liver fashion, treated with chemotherapy shortly after RE, not receiving steroids or ursodeoxycholic acid, or treated with the standard protocol. In the multivariate model that was constructed based on these findings, being treated by the standard protocol and being exposed to chemotherapy after RE were the only two factors independently associated with the development of REILD.

Table 4. Distribution of REILD Cases According to Main Basal Prognostic Indicators
   Standard Protocol (n=75)Modified Protocol (n=185)
  • *

    P < 0.001 for comparison with standard protocol group.

  • **

    P < 0.05 for comparison with standard protocol group. CTC: grade by NCI Common Toxicity Criteria v. 3.0.

CirrhosisTreatmentPrior ChemoAll CasesCTC≥3All CasesCTC≥3
No Cirrhosis 10/51 (19.6%)74/121 (3.3%)*2**
Whole-liverNo Chemo1/6 (16.7%)10/200
Chemo9/31 (29.0%)64/52 (7.7%)**2**
SelectiveNo Chemo0/9 0/21 
Chemo0/5 0/28 
Global disease control28 (54.9%)58 (47.9%)
Local disease control32 (62.7%)72 (59.5%)
Cirrhosis 7/24 (29.1%)36/64 (9.3%)**2
Whole-liverNo Chemo3/9 (33.0%)10/50
Chemo2/2 (100%)11/1 (100%)0
SelectiveNo Chemo2/13 (15.4%)15/57 (8.8%)2
Chemo 0/1 
Global disease control18 (75%)46 (71.9%)
Local disease control18 (75%)51 (79.7%)

Cirrhosis patients more likely to develop REILD were those with small livers (total volume <1.5 L), an abnormal bilirubin at baseline (>1.2 mg/dL), hypersplenism (platelets <100/pL), treated in a selective fashion, not receiving steroids or ursodeoxycholic acid, or treated by the standard protocol. In the corresponding multivariate model, only the presence of a small liver, an abnormal bilirubin, and treatment in a selective fashion were independently associated to the development of REILD. The reason for this apparent contradiction may be that treatment was more intense when delivered in a selective rather than in a whole-liver fashion (average activity relative to target volume: 1.79 GBq/L versus 0.92 GBq/L, P = 0.002).

When the whole series is analyzed, REILD did not appear when RE was performed in an otherwise healthy liver but in those patients who had a well-known liver condition such as cirrhosis or had been exposed to any form of potential liver insult, most usually chemotherapy. Although an absolute threshold of treatment intensity (activity per target liver volume) associated with the appearance or greatest severity of REILD was not identified, most cases occurred over the level of 0.8 GBq/L, as illustrated in Fig. 2. For patients treated in a whole-liver fashion, receiving an activity at least 10% lower than the one prescribed by the BSA formula was associated with a significantly lower likelihood of developing REILD (odds ratio 0.23, 95% confidence interval [CI] 0.08-0.63; P = 0.004). In fact, the difference in estimated normal liver absorbed radiation between patients with and without REILD did not reach statistical significance (38.9 [24.9-46.5] Gy versus 30.0 [22.0-38.0] Gy, P = 0.063). None of the patients with REILD who were exposed to post-RE chemotherapy (or biological agents) were chemonaïve at the time of RE. For those patients treated with chemotherapy before RE, when each chemotherapeutic agent used pre- or post-RE was analyzed individually, no association with REILD was found, as shown in Table 6 available online.

thumbnail image

Figure 2. Treatment intensity among patients treated in a whole-liver fashion according to the occurrence of REILD. (A,B) The individual values of administered activity relative to target liver volume for noncirrhosis and cirrhosis patients, respectively. (C,D) An expanded view of the values around 0.8 GBq/L for better identification of this cutoff value.

Download figure to PowerPoint

Table 5. Analysis of Risk Factors for REILD According to the Presence of Cirrhosis
   Univariate AnalysisMultivariate Analysis
Parameter nREILDPOR (95%CI)POR (95%CI)
  • Post-RE: received 0-2 months after radioembolization. Selective means not whole-liver treatment (lobar or sublobar).

  • *

    Six patients had missing values.

Patients without cirrhosis
Age < 60 yearsYes91120.0226.0 (1.30-27.6)  
Prior surgeryYes2340.0922.92 (0.83-10.26)  
Prior chemotherapyYes116130.0656.94 (0.88-54.4)  
Prior capecitabineYes88120.0176.39 (1.38-29.5)  
Whole-liver treatmentYes109140.0029.41 ()  
GGT > median (86 IU/mL)*Yes82120.0137.02 (1.52-32.4)  
Chemotherapy post-REYes56100.0036.08 (1.81-20.3)0.0096.46 (1.60-25.91)
Ursodeoxycolic acidYes15590.0060.29 (0.12-0.69)  
Methyl-prednisoloneYes161930.26 (0.11-0.62)  
Modified protocolYes12140.0010.14 (0.04-0.47)0.0360.85 (0.01-0.85)
Patients with cirrhosis
Total liver volume < 1.5 LYes27100.0029.21 (2.26-37.5)0.0307.58 (1.21-47.4)
Bilirubin > 1.2 mg/dLYes3990.0334.50 (1.12-17.9)0.0447.20 (1.05-49.3)
Platelets < 100/pLYes2470.0174.69 (1.32-16.6)  
Selective treatmentYes7170.0134.98 (1.40-17.6)0.0440.13 (0.01-0.94)
Ursodeoxycolic acidYes6260.0510.29 (0.08-0.97)  
Methyl-prednisoloneYes6360.0430.29 (0.08-0.91)  
Modified protocolYes6460.0260.25 (0.07-0.84)  


  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

In the largest series so far reported, we analyzed the incidence and severity of REILD and the potentially beneficial effects that a number of improvements in the process of treatment design, activity calculation, and concomitant therapy may have had in the safety of RE among patients with both primary and secondary liver tumors.

Our results indicate that REILD is more likely to occur in the setting of a liver primed by the existence of cirrhosis or by the exposure to anticancer chemotherapy before RE, particularly when it is followed by more chemotherapy shortly after RE. These data are particularly relevant due to the remarkable similarities between REILD and CMILD,1, 12 another radiation-induced syndrome described in patients exposed to total body irradiation and systemic high-dose chemotherapy as a conditioning procedure for bone marrow transplantation. Contrary to RILD, jaundice is the hallmark of CMILD and REILD, and both syndromes may occur in a situation of low-grade liver irradiation. In fact, CMILD has also been described after chemoirradiation of liver tumors. In a large cohort of patients with hepatocellular carcinoma receiving external beam radiation concurrently with chemotherapy delivered either by TACE (low systemic exposure) or intraarterial infusion (high systemic exposure), the incidence of RILD was similar in both groups (5.6% versus 3.7%) but CMILD occurred only in the latter (0% versus 8.8%).13

Regarding the potential mechanisms underlying REILD, a number of studies have suggested that a hypercoagulable state is a key factor in the pathogenesis of CMILD14 and that recombinant tissue plasminogen activator (tPA) may be effective in the treatment of established severe CMILD.15 In this setting, and taking into account that liver cirrhosis is increasingly recognized as a procoagulant situation,16 one may speculate that an uncontrolled local liver response to irradiation may be favored by subclinical underlying liver damage or a prothrombotic milieu. Although an impaired regenerative ability caused by antiproliferative agents cannot be ruled out as an alternative explanation, it seems rather unlikely since chemotherapy has been shown not to alter liver regeneration in an animal model of 70% hepatectomy17 or in patients simultaneously submitted to portal vein embolization.18 The observation in our previous study that REILD was less frequent among elderly patients has been confirmed in this larger series, and the higher number of elderly patients treated in recent years may have accounted for the reduction in the incidence of REILD. Yet we have no biological explanation for this statistical association between age and REILD. Further basic and clinical research is clearly warranted to shed light on the pathophysiology of REILD.

Our results argue for the sparing of as much liver volume as possible in RE by means of superselective catheterization. Only patients with cirrhosis developed REILD after lobar or sublobar RE and liver dysfunction in these cases was usually transient, most probably because the hypertrophy of the spared liver volume7 helps improve the global liver function. In those situations in which the entire liver volume needs to be targeted, our data strongly support a conservative approach to activity calculation. The reduction in the incidence and severity of REILD achieved by the implementation of our modified protocol may be due to the lower activities used in patients with potentially damaged livers, to the use of ursodeoxicolic acid and prolonged low-dose steroids, or to a combination of these factors. Only further prospective comparative studies will help elucidate this point. However, the use of both agents is safe even in cirrhosis and can be considered in clinical practice together with our algorithm for activity calculation. The absence of a clear dose-response relationship between the activity prescribed and the appearance of REILD argues against the use of purely dosimetric approaches to activity calculation, in which activity depends only on the estimated dose of radiation delivered to the tumor and nontumoral tissues and not on the volume, status, and function of the targeted liver.

The concurrent use of RE in combination with either folinic acid-5FU-oxaliplatin (FOLFOX) or irinotecan as first- and second-line therapy, respectively, has produced little liver toxicity in patients with liver metastasis from colorectal cancer.19, 20 Yet the higher likelihood of developing REILD among patients preexposed to chemotherapy who are also treated with chemotherapy after RE is in line with the observed toxicity when RE was combined with infusional fluorouracil (5FU) as rescue therapy in heavily pretreated patients.21 Our results suggest that when RE is used as a salvage approach, a 2-month chemotherapy-free interval following RE is highly recommended. Among HCC patients, REILD has not been specifically evaluated but similar or higher rates of CTCAE grade 3 bilirubin toxicities have been described.22-24

A limitation of our study is the lack of information about the histological basis of REILD among our patients. The concept of REILD was originally described in patients without liver cirrhosis as a form of sinusoidal obstruction syndrome that produces jaundice and ascites in the absence of tumor progression or bile duct obstruction.4 Among patients with cirrhosis, jaundice and ascites may also occur due to the progression of cirrhosis. However, it is very unlikely for patients with a low total bilirubin level to decompensate rapidly in the absence of an additional liver insult. Although sinusoidal obstruction has not been demonstrated in these patients, even in the few liver samples that we had the opportunity to analyze, we feel that the term REILD may also apply to this form of acute-on-chronic liver failure occurring in cirrhosis after RE. Whether or not they have a common histological background, REILD in cirrhosis and noncirrhosis appear in the same time period after RE and have been similarly affected by the changes we introduced in the modified protocol.

Activity calculation is different for resin and glass 90Y microspheres. In the latter, it depends only on the target liver volume (inclusive of tumor and nontumor tissue). For resin microspheres, users should select between BSA and the Partition Model. In our protocol, this decision is based on the volume of liver that will be targeted and the aim of the treatment. We believe that the results of this study may help users of resin microspheres by means of a better standardization of the treatment protocol. The pharmacological prophylaxis may also be pertinent for users of glass microspheres.

In conclusion, REILD is an uncommon, but clinically relevant, complication of RE that tends to occur when a large volume of liver tissue usually harboring an unfavorable milieu as a result of cirrhosis, prior or subsequent chemotherapy is exposed to radiation delivered radioactive microspheres. It may also appear when partial liver volumes are treated in patients with cirrhosis with a reduced liver functional reserve. We have designed a comprehensive treatment plan that includes an algorithm for individualized activity calculations that reduces both the frequency and the severity of REILD.


  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Mrs. Carmen Fuertes for excellent work in patient care and data management.


  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information
  • 1
    Sangro B, Iñarrairaegui M, Bilbao JI. Radioembolization for hepatocellular carcinoma. J Hepatol 2012; 56: 464473.
  • 2
    Lawrence TS, Robertson JM, Anscher MS, Jirtle RL, Ensminger WD, Fajardo LF. Hepatic toxicity resulting from cancer treatment. Int J Radiat Oncol Biol Phys 1995; 31: 12371248.
  • 3
    Pan CC, Kavanagh BD, Dawson LA, Li XA, Das SK, Miften M, et al. Radiation-associated liver injury. Int J Radiat Oncol Biol Phys 2010; 76: S94100.
  • 4
    Sangro B, Gil-Alzugaray B, Rodriguez J, Sola I, Martinez-Cuesta A, Viudez A, et al. Liver disease induced by radioembolization of liver tumors: description and possible risk factors. Cancer 2008; 112: 15381546.
  • 5
    Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. HEPATOLOGY 2011; 53: 10201022.
  • 6
    Sangro B, Bilbao JI, Boan J, Martinez-Cuesta A, Benito A, Rodriguez J, et al. Radioembolization using 90Y-resin microspheres for patients with advanced hepatocellular carcinoma. Int J Radiat Oncol Biol Phys 2006; 66: 792800.
  • 7
    Riaz A, Gates VL, Atassi B, Lewandowski RJ, Mulcahy MF, Ryu RK, et al. Radiation segmentectomy: a novel approach to increase safety and efficacy of radioembolization. Int J Radiat Oncol Biol Phys 2011; 79: 163171.
  • 8
    Essell JH, Schroeder MT, Harman GS, Halvorson R, Lew V, Callander N, et al. Ursodiol prophylaxis against hepatic complications of allogeneic bone marrow transplantation. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 1998; 128: 975981.
  • 9
    Ohashi K, Tanabe J, Watanabe R, Tanaka T, Sakamaki H, Maruta A, et al. The Japanese multicenter open randomized trial of ursodeoxycholic acid prophylaxis for hepatic veno-occlusive disease after stem cell transplantation. Am J Hematol 2000; 64: 3238.
  • 10
    Trotti A, Colevas AD, Setser A, Rusch V, Jaques D, Budach V, et al. CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 2003; 13: 176181.
  • 11
    Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, Rubinstein L, et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: 205216.
  • 12
    Bayraktar UD, Seren S, Bayraktar Y. Hepatic venous outflow obstruction: three similar syndromes. World J Gastroenterol 2007; 13: 19121927.
  • 13
    Shim SJ, Seong J, Lee IJ, Han KH, Chon CY, Ahn SH. Radiation-induced hepatic toxicity after radiotherapy combined with chemotherapy for hepatocellular carcinoma. Hepatol Res 2007; 37: 906913.
  • 14
    Scrobohaci ML, Drouet L, Monem-Mansi A, Devergie A, Baudin B, D'Agay MF, et al. Liver veno-occlusive disease after bone marrow transplantation changes in coagulation parameters and endothelial markers. Thromb Res 1991; 63: 509519.
  • 15
    Bearman SI, Shuhart MC, Hinds MS, McDonald GB. Recombinant human tissue plasminogen activator for the treatment of established severe venocclusive disease of the liver after bone marrow transplantation. Blood 1992; 80: 24582462.
  • 16
    Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl J Med 2011; 365: 147156.
  • 17
    Rickenbacher A, DeOliveira ML, Tian Y, Jang JH, Riener MO, Graf R, et al. Arguments against toxic effects of chemotherapy on liver injury and regeneration in an experimental model of partial hepatectomy. Liver Int 2011; 31: 313321.
  • 18
    Covey AM, Brown KT, Jarnagin WR, Brody LA, Schwartz L, Tuorto S, et al. Combined portal vein embolization and neoadjuvant chemotherapy as a treatment strategy for resectable hepatic colorectal metastases. Ann Surg 2008; 247: 451455.
  • 19
    Sharma RA, Van Hazel GA, Morgan B, Berry DP, Blanshard K, Price D, et al. Radioembolization of liver metastases from colorectal cancer using yttrium-90 microspheres with concomitant systemic oxaliplatin, fluorouracil, and leucovorin chemotherapy. J Clin Oncol 2007; 25: 10991106.
  • 20
    van Hazel GA, Pavlakis N, Goldstein D, Olver IN, Tapner MJ, Price D, et al. Treatment of fluorouracil-refractory patients with liver metastases from colorectal cancer by using yttrium-90 resin microspheres plus concomitant systemic irinotecan chemotherapy. J Clin Oncol 2009; 27: 40894095.
  • 21
    Hendlisz A, Van den Eynde M, Peeters M, Maleux G, Lambert B, Vannoote J, et al. Phase III trial comparing protracted intravenous fluorouracil infusion alone or with yttrium-90 resin microspheres radioembolization for liver-limited metastatic colorectal cancer refractory to standard chemotherapy. J Clin Oncol 2010; 28: 36873694.
  • 22
    Salem R, Lewandowski RJ, Mulcahy MF, Riaz A, Ryu RK, Ibrahim S, et al. Radioembolization for hepatocellular carcinoma using Yttrium-90 microspheres: a comprehensive report of long-term outcomes. Gastroenterology 2010; 138: 5264.
  • 23
    Hilgard P, Hamami M, Fouly AE, Scherag A, Muller S, Ertle J, et al. Radioembolization with yttrium-90 glass microspheres in hepatocellular carcinoma: European experience on safety and long-term survival. HEPATOLOGY 2010; 52: 17411749.
  • 24
    Sangro B, Carpanese L, Cianni R, Golfieri R, Gasparini D, Ezziddin S, et al. Survival after yttrium-90 resin microsphere radioembolization of hepatocellular carcinoma across Barcelona clinic liver cancer stages: a European evaluation. HEPATOLOGY 2011; 54: 868878.

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

HEP_26191_sm_SuppTab1.doc25KSupporting Information Table 1.
HEP_26191_sm_SuppInfo.pdf25KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.