management of high hepatopulmonary shunting in patients ... · hepatopulmonary shunting is measured...

CLINICAL STUDY

Management of High Hepatopulmonary Shunting inPatients Undergoing Hepatic Radioembolization

Thomas J. Ward, MD, Anobel Tamrazi, MD, PhD,Marnix G.E.H. Lam, MD, PhD, John D. Louie, MD, Peter N. Kao, MD, PhD,Rajesh P. Shah, MD, Michael A. Kadoch, MD, and Daniel Y. Sze, MD, PhD

ABSTRACT

Purpose: To review the safety of hepatic radioembolization (RE) in patients with high (Z 10%) hepatopulmonary shuntfraction (HPSF) using various prophylactic techniques.

Materials and Methods: A review was conducted of 409 patients who underwent technetium 99m–labeled macroaggregatedalbumin scintigraphy before planned RE. Estimated pulmonary absorbed radiation doses based on scintigraphy and hepaticadministered activity were calculated. Outcomes from dose reductions and adjunctive catheter-based prophylactic techniquesused to reduce lung exposure were assessed.

Results: There were 80 patients with HPSF Z 10% who received RE treatment (41 resin microspheres for metastases, 39 glassmicrospheres for hepatocellular carcinoma). Resin microspheres were used in 17 patients according to consensus guideline–recommended dose reduction; 38 patients received no dose reduction because the expected lung dose was o 30 Gy. Prophylactictechniques were used in 25 patients (with expected lung dose r 74 Gy), including hepatic vein balloon occlusion, varicealembolization, or bland arterial embolization before, during, or after RE delivery. Repeated scintigraphy after prophylactictechniques to reduce HPSF in seven patients demonstrated a median change of �40% (range, þ32 to �69%). Delayedpneumonitis developed in two patients, possibly related to radiation recall after chemoembolization. Response was lower inpatients treated with resin spheres with dose reduction, with an objective response rate of 13% and disease control rate of 47%compared with 56% and 94%, respectively, without dose reduction (P ¼ .023, P ¼ .006).

Conclusions: Dose reduction recommendations for HPSF may compromise efficacy. Excessive shunting can be reduced byprophylactic catheter-based techniques, which may improve the safety of performing RE in patients with high HPSF.

ABBREVIATIONS

HCC = hepatocellular carcinoma, HPSF = hepatopulmonary shunt fraction, HVTT = hepatic vein tumor thrombus, MIRD = medical

internal radiation dose, PVA = polyvinyl alcohol, PVTT = portal vein tumor thrombus, RE = radioembolization, RP = radiation

pneumonitis, 99mTc-MAA = technetium 99m–labeled macroaggregated albumin, 90Y = yttrium-90

Complications and toxicities after yttrium-90 (90Y) radio-embolization (RE) can result from nontarget depositionof microspheres, such as RE-induced liver disease fromdeposition in functional liver, ulceration from depositionin the stomach or bowel, and radiation pneumonitis

From the Divisions of Interventional Radiology (T.J.W., J.D.L. R.P.S., D.Y.S.),Pulmonology and Critical Care (P.N.K.), and Thoracic Imaging (M.A.K.),Stanford University Medical Center, H-3646, 300 Pasteur Drive, Stanford,CA 94305-5642; Division of Vascular and Interventional Radiology (A.T.),Johns Hopkins University, Baltimore, Maryland; and Division of NuclearMedicine (M.G.E.H.L.), University Medical Center Utrecht, Utrecht, TheNetherlands. Received May 4, 2014; final revision received August 24, 2015;accepted August 25, 2015. Address correspondence to T.J.W.; E-mail:[email protected]

D.Y.S. is on the advisory boards of Surefire Medical, Inc (Westminster, Colorado),Koli Medical, Inc (Fremont, California), Northwind Medical, Inc (San Jose,California), Treus Medical, Inc (Redwood City, California), RadiAction Medical, Ltd(Tel Aviv, Israel), and EmbolX, Inc (Los Altos, California), and is a consultant for

Downloaded from ClinicalKey.com atFor personal use only. No other uses without permission. C

(RP) from deposition in the lungs via hepatopulmonaryshunting (1). Pathologic arteriovenous communications(arterioportal and arteriohepatic venous) are common inliver tumors (2,3). Microspheres injected into the hepaticartery can pass through arteriovenous shunts 4 30 μm

Amgen, Inc (Thousand Oaks, California), BTG, Inc (West Conshohocken, Pennsyl-vania), Sirtex Medical, Ltd (North Sydney, Australia), W. L. Gore & Associates, Inc(Flagstaff, Arizona), Covidien, Inc (Mansfield, Massachusetts), Guerbet (Villepinte,France), Cook, Inc (Bloomington, Indiana), and Codman & Shurtleff (Raynham,Massachusetts). M.G.E.H.L. is a speaker for Sirtex Medical, Ltd (North Sydney,Australia), and a consultant for Bayer HealthCare AG (Leverkusen, Germany). Noneof the other authors have identified a conflict of interest.

From the SIR 2013 Annual Meeting.

& SIR, 2015

J Vasc Interv Radiol 2015; 26:1751–1760

http://dx.doi.org/10.1016/j.jvir.2015.08.027

Dignity Health July 15, 2016.opyright ©2016. Elsevier Inc. All rights reserved.

mailto:[email protected]

dx.doi.org/10.1016/j.jvir.2015.08.027

dx.doi.org/10.1016/j.jvir.2015.08.027

dx.doi.org/10.1016/j.jvir.2015.08.027

Table 1. Demographics of 80 Patients with HPSF 4 10%

Treated by 90Y RE

n or Mean SD or %

Age 64 � 12

Male-to-female ratio 54:26 68%/32%

Diagnosis

HCC 36 45%

Metastatic colorectal carcinoma 17 21%

Metastatic neuroendocrine tumor 9 11%

Cholangiocarcinoma 5 6%

Metastatic renal cell carcinoma 4 5%

Other 9 11%

Macrovascular invasion

HVTT 4 5%

PVTT 20 25%

Previous treatment before RE

Prior transarterial chemoembolization 15 19%

Prior ablation 3 4%90Y microsphere treatment

Glass 39 49%

Ward et al ’ JVIR1752 ’ Management of High HPSF in Patients Undergoing Hepatic RE

in luminal diameter, traverse the heart, and lodge inpulmonary arterioles. In addition, in patients with portalhypertension and varices, microspheres can pass througharterioportal shunts, exit the liver via hepatofugal portalflow, traverse varices into a systemic vein, and thenlodge in the lungs. These hepatopulmonary shunts canlead to clinically significant RP (4).Hepatopulmonary shunting is measured by whole-

body planar scintigraphy after injection of technetium99m–labeled macroaggregated albumin (99mTc-MAA)into a hepatic artery to simulate future 90Y microspheredistribution. Early studies on whole-liver RE treatmentassociated the risk of RP with the hepatopulmonaryshunt fraction (HPSF) (5) and led to consensusguidelines dictating that patients with high HPSFshould undergo RE with reduction of administeredactivity or should not undergo RE at all (6,7). Thepurpose of our study was to review the safety andefficacy of performing RE on patients with high HPSF(Z 10%) with the use of prophylactic dose reduction orcatheter-based shunt mitigation techniques or both.

Resin 41 51%

HCC ¼ hepatocellular carcinoma; HPSF ¼ hepatopulmonary

shunt fraction; HVTT ¼ hepatic vein tumor thrombus; PVTT ¼portal vein tumor thrombus; RE ¼ radioembolization; 90Y ¼yttrium-90.

MATERIALS AND METHODS

PatientsInstitutional review board approval was obtained for thisretrospective study. All data were handled in compliancewith the Health Insurance Portability and AccountabilityAct. Between 2004 and 2014, 409 patients who under-went 99mTc-MAA scintigraphy (Jubilant DraxImage,Kirkland, Quebec, Canada) before planned REfor primary or metastatic hepatic neoplasms at a singlecenter were reviewed. Combined with prescribed andadministered 90Y activities, HPSF was used to calculateexpected absorbed dose to the lungs for each patient.Tumor cell type, presence of portal vein tumor thrombus(PVTT) or hepatic vein tumor thrombus (HVTT), andhistory of previous liver-directed therapies (ie, resection,transarterial chemoembolization, and percutaneousor laparoscopic ablation) were recorded. Of patients,15 underwent 99mTc-MAA scintigraphy but did notreceive RE treatment; 394 patients received RE treat-ment. Retrospective analysis was performed of 80 high-risk patients treated with RE with HPSF Z 10%(median, 14.1%; range, 10%�54%) (Table 1).The different methods to address elevated HPSF in

these 80 patients (68% male; mean age, 64 y � 12; celltype, hepatocellular carcinoma [HCC] 45%, other 55%;microspheres used, glass 49%, resin 51%) were reviewed.Patients were divided into three groups (Fig 1). The firstgroup consisted of 17 patients treated earlier whoreceived resin microspheres with administered activityreduced by 20%�40%, in adherence to publishedguidelines (SIR-Spheres yttrium-90 microspheres [pack-age insert] Lane Cove, Australia: Sirtex Medical, Ltd,2004.). The second group consisted of 38 patients in


whom the expected single-administration lung dose waso 30 Gy and the cumulative lung dose was o 50 Gy,and no dose reductions or prophylactic interventionswere performed. Expected lung doses were calculatedfrom administered activity, HPSF, and medical internalradiation dose (MIRD) formulas based on estimatedmass of lung tissue including blood being 1 kg (8). Thethird group consisted of 25 patients who underwentprophylactic catheter-based techniques to decreaseHPSF. Five patients treated with prophylactic interven-tions also had dose reduction.

Catheter-Based Shunt Reduction

TechniquesProphylactic catheter-based techniques used in attemptsto mitigate shunting included temporary hepaticvein balloon occlusion (9), bland arterial embolizationpreemptively or immediately after RE administra-tion, and portosystemic variceal embolization. Oneor more techniques with or without dose reductionwere used depending on clinical practice at the time,expected lung absorbed dose, and angiographic findings.Temporary hepatic vein balloon occlusion was used ineight patients. Dominant venous drainage was identifiedon imaging performed before the procedure andconfirmed by early hepatic vein enhancement on cone-beam C-arm computed tomography (CT). To matchthe diameter of the draining veins, compliant balloonsup to 14 mm diameter (Python; Applied MedicalResources Corp, Rancho Santa Margarita, California)


Figure 1. Flow chart of 409 patients in the study cohort. BAE ¼ bland arterial embolization; HV ¼ hepatic vein.

Volume 26 ’ Number 12 ’ December ’ 2015 1753

or valvuloplasty balloons up to 30 mm diameter(TYSHAK; B. Braun Interventional Systems, Inc, Beth-lehem, Pennsylvania) were used. Balloons were intro-duced via the right internal jugular vein or femoral vein,inflated with dilute contrast material to low pressure (o2 atm) immediately before 90Y administration, and keptinflated for 15 minutes after completion of administra-tion (Fig 2a). No anticoagulation was administered.Preemptive bland arterial embolization was performed

on four patients with prominent arteriovenous shuntingon the day of 90Y administration, using spherical(Embospheres [300–500 μm, n ¼ 2; 500–700 μm, n ¼ 1];Merit Medical Systems, South Jordan, Utah) or particu-late embolic agents (polyvinyl alcohol [PVA; 1,000–1,400μm]; Cook, Inc, Bloomington, Indiana). Scintigraphywas repeated before administration of 90Y microspheres.In another patient, preemptive transarterial chemoembo-lization was performed, followed by repeated scintigraphyand RE treatment 5 weeks later.In the four patients who received preemptive bland

arterial embolization, one patient who received trans-arterial chemoembolization, and an additional 12 patients,bland arterial embolization was performed immediatelyafter administration of 90Y (Fig 2b). The celiac axis wasselected using an introducer guide (7-F RDC; CordisCorp, Warren, New Jersey), and two parallel micro-catheters were placed (Progreat Omega; Terumo Medi-cal Corporation, Somerset, New Jersey; and RenegadeHI-FLO; Boston Scientific, Marlborough, Massachusetts)


with the 90Y microcatheter tip slightly distal to the blandarterial embolization microcatheter tip. For glass micro-spheres, only one 20-mL flush of the V-vial was per-formed, instead of the standard three flushes. Likewisefor resin, only one 20-mL water or 5% dextrose flush andone air flush were performed. Immediately after visiblepassage of microspheres through the transparent deliverysystem tubing, concentrated bland arterial embolizationmaterial suspended in contrast medium was rapidlyinjected via the other microcatheter until stasis wasachieved. Different embolic materials tested included 1-mm cubed gelatin foam slurry (SURGIFOAM [n ¼ 1];Ethicon, Inc, Cincinnati, Ohio), particle embolic agents(PVA [90–180 μm; n ¼ 2), spherical embolic agents(Embospheres [100–300 μm, n ¼ 3; 500–700 μm, n ¼ 7;500–700 μm, n ¼ 1), and N-butyl cyanoacrylate gluewith ethiodized oil (Histoacryl; Aesculap, Inc, CenterValley, Pennsylvania, or Lipiodol; Guerbet, Villepinte,France; n ¼ 1) (Fig 2c). Delivered activity was calcu-lated by subtracting residual activity from the initialtotal activity. Two patients had bland arterial emboli-zation and hepatic vein balloon occlusion.Preemptive variceal embolization was performed in

two patients in whom arteriography demonstrated arte-rioportal shunting, hepatofugal portal flow, and drain-age through portosystemic varices. Embolization withcoils (Nester; Cook, Inc) of the right adrenal vein in onepatient and the coronary vein in the other patient wasperformed (Fig 2d, e). 99mTc-MAA scintigraphy was


Figure 2. (a) Hepatic arteriography in a 55-year-old woman with metastatic pancreatic neuroendocrine carcinoma showed early filling of

the right hepatic vein from arteriovenous shunting (arrowheads), with flow diminished by inflation of an occlusion balloon in the

downstream right hepatic vein (arrow). (b) A 57-year-old man with HCC and PVTT with arterioportal shunting had parallel 2.8-F

microcatheters (arrows) positioned in the left hepatic artery via a 7-F introducer guide advanced into the common hepatic artery.90Y microspheres were delivered as a bolus from the distal microcatheter, followed immediately by rapid injection of bland embolic

material from the proximal microcatheter to achieve prompt stasis. (c) In a 72-year-old woman with HCC, PVTT, and HVTT fed by segment

2 and 3 hepatic arteries, parallel microcatheters were placed, and glass microspheres were followed with immediate embolization to stasis

with N-butyl cyanoacrylate glue opacified with Lipiodol, filling arteries to segments 2 and 3 (arrows). The microcatheters had been swiftly

removed to avoid glue complications. (d) Venous phase image of common hepatic arteriography performed in a 58-year-old man with

HCC and PVTT showed avid arterioportal shunting, with hepatofugal drainage through an atypical varix (arrows) originating from the left

portal vein, traversing the porta hepatis, and feeding gastroesophageal varices. (e) Before RE treatment, percutaneous access into the left

portal vein was obtained, the varix was selected, and venography confirmed portosystemic drainage into gastroesophageal varices. The

varices were sclerosed with 2% ethanolamine oleate, and then coil embolization was performed.


repeated immediately, followed by RE treatment; onepatient was treated with immediate bland arterial embo-lization because of persistent arterioportal shunting intonontargeted liver.Follow-up included clinical evaluation with physical

examination at 4 weeks and 12 weeks and cross-sectionalimaging at 10–12 weeks. To study the effect of dosereduction on efficacy in patients treated with resin micro-spheres, objective radiographic response, defined as eithera complete response or a partial response, and diseasecontrol rate, defined as complete response þ partialresponse þ stable disease according to Response Evalua-tion Criteria in Solid Tumors (version 1.1), were assessed(10). One author (T.J.W.), blinded to intervention or dosereduction, reviewed all images. When comparing responserates, the five patients treated with catheter interventions


and dose reductions were included in the dose reductiongroup. Respiratory complications including RP were theprimary toxicity outcome of interest, graded accordingto the Common Terminology Criteria for Adverse Events(version 4.03) (11). RP was diagnosed based on radio-graphic and clinical criteria, including peribronchovas-cular consolidation and a restrictive ventilatory pattern onpulmonary function tests without evidence of active infec-tion. Surviving patients with residual or recurrent diseasetypically proceeded to other treatment options, such assystemic chemotherapy or transarterial chemoembolization.

Statistical AnalysisDescriptive data are presented as mean and SD (normal)or median and range (nonnormal). Categorical data are


Intervention

duringRE

RP

BO

No

BO

No

E(gelatinfoam

slurry)

No

E(700–900μm

SE)

No

E(300–500μm

SE)

No

E(300–500μm

SE)

No

REperform

ed

No

n;HVTT¼

hepaticvein

tumor

ics;99mTc-M

AA

¼technetium


presented as percentages and counts. Pearson correlationand stepwise multiple linear regression analyses wereperformed to evaluate risk factors (pathology, HVTT,PVTT, liver-directed therapy) for elevated HPSF. Forcomparison between groups, Fisher exact test (catego-rical variables) was used. A P value of o .05 wasconsidered statistically significant. Statistical analyseswere performed with IBM SPSS Statistics for Windowsversion 20.0 (IBM Corporation, Armonk, New York).

Table

2.Changein

HPSFofSevenPatients

WhoUnderw

entRepeat99mTc-M

AA

ScintigraphyafterShuntReductionIntervention

CellType

HVTT

PVTT

InitialHPSF(%

)Intervention

PostHPSF(%

)

Absolute

HPSF

Reduction(%

)

RelativeHPSF

Reduction(%

)

HCC

No

No

44

RetroperitonealVCEandBAE

(500–700μm

SE)

21

23

51

HV

HCC

No

Yes

25

Coronary

vein

VCE

17

829

HV

HCC

Yes

No

25

BAE(1,000–1,400μm

PVA)

26

�1�4

BA

HCC

No

Yes

40

BAE(700–900μm

SE)

24

16

40

BA

HCC

Yes

Yes

50

BAE(300–500μm

SE)

27

23

46

BA

HCC

Yes

No

54

Transarterialchemoembolization

17

37

69

BA

HCC

Yes

No

23

Transarterialchemoembolization�

230

�7�3

2No

BAE¼

blandarterialembolization;HCC

¼hepatocellularcarcinoma;HPSF¼

hepatopulm

onary

shuntfraction;HVBO

¼hepaticvein

balloonocclusio

thrombus;PVA

¼polyvinylalcohol;PVTT¼

portalvein

tumorthrombus;RE¼

radioembolization;RP¼

radiationpneumonitis;SE¼

sphericalembol

99m–labeledmacroaggregatedalbumin;VCE¼

varicealcoilembolization.

RESULTS

Of 409 patients who underwent HPSF measurement byscintigraphy, 89 had HPSF 4 10%, and 80 of thesepatients received RE treatment (Fig 1). Chemoembo-lization alone was performed in two patients, and noliver-directed treatment was performed in seven becauseof clinical deterioration (n ¼ 6) or variceal hemorrhage(n ¼ 1). One patient had HPSF of 23% with expectedlung dose of 151 Gy, but preemptive transarterial chemo-embolization performed twice, once with doxorubicin-eluting 100–300 μm LC Beads (BTG International,Farnham, United Kingdom), once with doxorubicin-eluting 50–100 μm QuadraSpheres (Merit Medical Sys-tems), increased HPSF to 30%, so RE was not performed(Table 2).

Risk FactorsThe mean and median HPSF in all 409 patients were8.3% (SD � 6.9%) and 6.4% (range, 0.5%–54%), res-pectively. HPSF 4 10% was demonstrated in 32% (n ¼43 of 134) of patients with HCC and 17% (n ¼ 46 of 275)of patients with other cell types (Fig 3). Multiple linearregression analysis performed to evaluate risk factors forelevated HPSF demonstrated that HCC, HVTT, andPVTT correlated with increased HPSF (correlationcoefficients 0.20, 0.27, 0.30; all P o .001). Previoustransarterial chemoembolization and previous ablationwere not correlated with HPSF (P ¼ .27 and P ¼ .47).When stepwise multiple regression analysis was per-formed to control for covariables, only HVTT andPVTT (and not HCC) demonstrated significant impactson HPSF (both P o .001), with this model explaining14% of the variability in HPSF (P o .001). The β coeffi-cients associated with HVTT and PVTT were 10.5% and5.9%, so that a patient with HVTT and PVTT wouldhave 16.4% absolute higher HPSF than a patient with-out tumor thrombus in this model.Of patients with HPSF 4 10%, 89% (n ¼ 80 of 89)

received RE treatment. Excluding patients who were nottreated because of clinical deterioration, dose reductionand adjunctive catheter techniques allowed for the treat-ment of 97.5% (n ¼ 80 of 82) of patients with HPSF 410%. Of the treated patients, the median HPSF was14.1% (range, 10%–54%); 54 (68%) were male, and36 (45%) were treated for HCC. The efficiency percentage

Downloaded from ClinicalKey.com at Dignity Health July 15, 2016.For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

Figure 3. Distribution of patients with high HPSF according to risk factors. High HPSF is defined as Z 10%. mCRC ¼ metastatic

colorectal cancer.


of microsphere delivery after a single flush of the V-vialwas 94.4% for glass microspheres and 89.3% for resinmicrospheres compared with 97.0% for three or moreflushes in patients who underwent normal delivery.

Repeated ScintigraphyIn seven patients who underwent repeat scintigraphyafter HPSF mitigation, a median absolute reduction of16% (range, 7% increase to 37% decrease) and a relativereduction of 40% (range, 32% increase to 69% decrease)was achieved (Table 2). In the two patients whose HPSFwas not reduced, one underwent bland arterial emboliza-tion with very large PVA particles (1,000–1,400 μm), andthe other underwent two drug-eluting microsphere trans-arterial chemoembolization procedures. The other fivepatients all had an absolute reduction of at least 7.2%(range, 7.2%–37%) and relative reduction of at least 29%(range, 29%–69%) (Fig 4a, b). In these patients withsuccessful mitigation, embolic particle size was 300–900μm. The patients who underwent variceal embolizationshowed relative reductions of 51% (in conjunction withbland arterial embolization) and 29% and did not mani-fest signs or symptoms of worsened portal hypertension.

Pulmonary ToxicityFollow-up CT scans of the lungs were available in 55 of80 treated patients (median follow-up period, 154 d;range, 13–2,050 d). Clinical follow-up data were avail-able in all patients (median follow-up, 137 d; range, 6–2,134 d). Within the first 4 months of follow-up, nopatients developed respiratory complications suspiciousfor RP. Of the 17 patients treated with dose reductionsonly, 12 were treated with 20% reduction, and 5 weretreated with 40% reduction. None of the patients showedevidence of pulmonary toxicity. Similarly, of the 38patients who received no dose reduction or shuntmitigation because of low expected lung dose, noneshowed evidence of pulmonary toxicity.Two patients (2.6%) with high HPSF treated with

shunt mitigating techniques developed delayed respira-tory complications and radiographic findings suspicious


for RP. Both patients were asymptomatic after RE butdeveloped delayed complications after subsequent trans-arterial chemoembolization. In the 80 patients withHPSF 4 10%, 16 patients underwent 23 transarterialchemoembolization procedures after RE (median, 1;range, 1–6; conventional transarterial chemoembolization,n ¼ 11; drug-eluting microsphere transarterial chemo-embolization, n ¼ 12). RP was not observed in any ofthe 64 patients who did not undergo transarterial chemo-embolization after RE (12.5% vs 0%, P ¼ .04).The first patient was a 58-year-old woman with meta-

stases from intracranial hemangiopericytoma (whole-liver volume ¼ 3,192 mL, estimated tumor replacement¼ 50%) (Fig 5a). Using the body surface area method,whole-liver activity prescription was 2.28 GBq. HPSFwas 27.7%, and estimated lung exposure without shuntmitigation was 32 Gy. Hepatic vein balloon occlusionand immediate bland arterial embolization using 4 vialsof 300–500 μm Embospheres were performed. Thisasymptomatic patient underwent transarterial chemo-embolization 3 months later to treat residual diseasesupplied by parasitized phrenic arteries, receiving 150mg doxorubicin on 100–300 μm and 300–500 μm LCBeads. She developed dyspnea on exertion 1 month later,which was 4 months after RE. CT revealed coarseperibronchial consolidation and ground-glass opacity,suggestive of RP (Fig 5b, c). Pulmonary function testsshowed restrictive physiology with forced vital capacity54% of predicted and diffusion capacity of the lung forcarbon monoxide 62% of predicted. She resumed recrea-tional physical activities on corticosteroid therapy (pred-nisone 40 mg/d tapering to 12.5 mg/d), and pulmonaryfunction tests improved to forced vital capacity of 66%and diffusion capacity of the lung for carbon monoxideof 80%. She tolerated four additional transarterialchemoembolization treatments and remains on activesurveillance 46 months after RE treatment, with scarringand mild bronchiectasis on follow-up imaging.The second patient was a 72-year-old woman with

focal HCC in segment 7 and infiltrative HCC with lefthepatic vein and inferior vena cava tumor thrombus insegments 2 and 3. The HPSF was 22.6%. Using 200-Gy


Figure 4. (a) Anterior view of planar scintigraphy after administration of 99mTc-MAA into the proper hepatic artery in a 60-year-old man

with HCC and PVTT showed pronounced uptake in the lungs, reflecting a high HPSF of 43.7%. (b) The patient underwent coil

embolization of the right adrenal vein, which was the systemic outflow from arterioportal and portosystemic shunting, and bland

arterial embolization of the right hepatic artery using 700–900 μm spherical embolic agents 14 days later. The HPSF was reduced to

21.4%, and focal uptake was evident in the intrahepatic tumors.

Figure 5. (a) Right anterior oblique surface rendering of a contrast-enhanced cone-beam C-arm CT scan of a 58-year-old woman with

metastatic hemangiopericytoma before RE demonstrated large, hypervascular tumors with early opacification of the right and left

hepatic veins (arrows). Dilute contrast material (150 mg iodine/mL) was injected into the common hepatic artery at 2 mL/s for 12

seconds, and imaging was performed in the last 8 seconds after a 4-second delay. 99mTc-MAA scintigraphy yielded HPSF of 27.7%,

which combined with an activity prescription of 2.28 GBq and estimated lung mass of 1 kg would have resulted in a pulmonary

absorbed dose of 32 Gy. This patient underwent whole-liver RE with temporary occlusion balloons in the right and left hepatic veins and

bland arterial embolization using 4 vials of 300–500 μm spherical embolic agents immediately after RE administration. (b) Coronal

reconstruction of thoracic CT image obtained 10 weeks after RE showed no abnormalities in the asymptomatic patient. (c) Coronal

reconstruction of thoracic CT image obtained 4 weeks after transarterial chemoembolization and 4 months after RE demonstrated

bilateral coarse peribronchial consolidation and ground-glass opacity that spared the periphery. This imaging pattern, combined with

pulmonary function tests indicating a restrictive pattern with decreased diffusion, was diagnostic of RP, possibly triggered by a

radiation recall mechanism after exposure to doxorubicin.


target segmentectomy dose, the prescription was 3.48GBq total (2.26 GBq to segments 2 and 3 and 1.28 GBqto segment 7), and estimated lung exposure withoutshunt mitigation was 39 Gy. Only segments 2 and 3 withtumor thrombus were treated with prophylactic techni-ques. Dose administration was followed by immediateN-butyl cyanoacrylate glue embolization via a parallelmicrocatheter. Follow-up showed a 92% decline inserum α-fetoprotein (221 ng/mL to 18 ng/mL), completeresponse in segment 7, and nearly complete response in


segments 2 and 3 by modified Response EvaluationCriteria in Solid Tumors criteria (12). A chronic coughcaused by Mycobacterium avium complex was stable,and the lungs were clear on CT. The patient underwenttransarterial chemoembolization 4 months later to treatresidual segment 4 disease, receiving 75 mg doxorubicinon 100–300 μm LC Beads. The patient developeddyspnea and tachypnea 2 months later, 6 months afterRE, requiring home supplemental oxygen. CT showedmetastatic nodules and peribronchial consolidation not



typical of Mycobacterium. She declined further treat-ment and died in hospice.

Imaging Response and Dose ReductionResponse evaluation at 3-month follow-up evaluationwas possible in 31 of 41 patients treated with resinmicrospheres; 16 (including eight treated with adjunctivetechniques) were treated without dose reduction, and 15(including five treated with adjunctive techniques) weretreated with dose reduction. Objective radiographicresponse was observed in 13% of patients (n ¼ 2 of 15)treated with dose reduction compared with 56% (n ¼ 9of 16) treated without dose reduction (P ¼ .023). Diseasecontrol rate was 47% (n ¼ 7 of 15) in patients treatedwith dose reduction compared with 94% (n ¼ 15 of 16)without dose reduction (P ¼ .006) (Table 3).

DISCUSSION

Excessive shunting of radioactive microspheres to thelungs from hepatic RE treatment can result in RP and isan established contraindication for this therapy (4,5). RPresults from pulmonary radiation damage and repair,which typically manifests 1–2 months after exposure.Patients present with nonspecific symptoms of fever,nonproductive cough, and dyspnea. Initial injury resultsin a fibrinous exudate and organizing pneumonia, even-tually leading to pulmonary fibrosis and potentiallydebilitating chronic disease. Imaging shows peribroncho-vascular interstitial disease (13), and pulmonary functiontests typically show a restrictive pattern with decreaseddiffusion. Treatment is primarily with systemic orinhaled corticosteroids, or both, which may be requiredindefinitely to maintain pulmonary function (5).RP after RE is a rare occurrence. The largest series of

five cases among 80 patients treated using activities of3.4–8.4 GBq suggested increased risk of RP in patientswith HPSF 4 13% (5). Preemptive bland arterial embo-lization in three patients using “inert particles” of 150–300 μm PVA performed an unspecified amount of time

Table 3. Objective Response and Disease Control Rates after

RE for Patients Treated with Resin Microspheres

Dose

Reduction

(n ¼ 15)

No Dose

Reduction

(n ¼ 16) P Value

CR 0 0

PR 2 9

SD 5 6

Objective response

(CR þ PR)

13% 56% .023

Disease control

(CR þ PR þ SD)

47% 94% .006

CR ¼ complete response; PR ¼ partial response; RE ¼ radio-

embolization; SD ¼ stable disease.


before the treatment decreased the measured HPSF by65%–78% but inexplicably did not protect against pneu-monitis. Consensus guidelines subsequently were developedfor the use of resin microspheres recommending decrea-sing the administered activity by 20% in patients with10%–15% HPSF, decreasing the administered activity by40% in patients with 15%–20% HPSF, and avoiding REtreatment in patients with HPSF 4 20% (SIR-Spheresyttrium-90 microspheres [package insert] Lane Cove,Australia: Sirtex Medical, Ltd, 2004.). Guidelines alsorecommend restricting radiation-absorbed dose to thelungs to o 30 Gy in a single exposure or 50 Gy over alifetime (14).These recommendations are not universally accepted.

In a series of 58 patients with a cumulative lung dose4 30 Gy, no patients were found to have radiographicor clinical evidence of RP (15). This cohort included19 patients with cumulative lung dose 4 50 Gy. Guide-lines for whole-liver treatment are not valid for treat-ment of smaller territories; treatment of a fraction of theliver results in a smaller total number of microspheresshunted if shunting is uniform throughout the liver.A large percentage of a small dose often does not exceedthe 30-Gy threshold. Conversely, focal shunting may beunderestimated and averaged by whole-liver 99mTc-MAA scintigraphy. Because response correlates in partto tumor absorbed dose (16), decreasing administeredactivity could lead to subtherapeutic dosing and poorresponse. Although still subject to error, calculation ofabsorbed dose to the lungs should be a more logical basisfor adjustment of administered activity than the currentoversimplified guidelines.Even when administered correctly with the micro-

catheter positioned at the planned location of RE,99mTc-MAA distribution may not exactly model 90Yactivity distribution (17). 99mTc-MAA particle sizes areheterogeneous, and approximately 10% are o 10 μm indiameter, resulting in an overestimation of HPSF (18).Scatter from high 99mTc-MAA concentration in the livermay also artifactually increase calculated deposition inthe right lung base (19). Streaming and clumping areother possible sources of mismatch between 99mTc-MAAand 90Y microsphere distribution. Other techniques ofcalculating HPSF, such as dynamic contrast enhance-ment studies of hepatic vein opacification (20) or usingimageable 30-μm microspheres (18), may improve accu-racy in the future.High HPSF occurs more frequently in patients with

HCC but is uncommon with other cell types, leadingsome investigators to suggest foregoing scintigraphy inpatients with cell types other than HCC (21). Our cohortdemonstrated that a substantial number of metastatictumors also result in high HPSF. After controlling formacrovascular invasion, HCC was not significantlyassociated with increased HPSF.Previously reported techniques to reduce hepatopulmo-

nary shunting include systemic sorafenib (22) or transarterial



chemoembolization (23,24). Sorafenib treatment dura-tion ranged from 72 to 297 days before repeat 99mTc-MAA imaging and subsequent 90Y treatment and wasable to reduce HPSF by 62%–87%. Transarterial chemo-embolization in patients with infiltrative HCC withportal venous invasion resulted in reduction of HPSFby 25%–57%. Although sorafenib and transarterialchemoembolization may also have antitumoral efficacy,their use to reduce shunting may delay RE treatment,and many candidates for RE are patients who havealready failed these treatments.Shunt mitigation techniques that may be employed on

the day of treatment include temporary balloon occlu-sion of hepatic veins, which can reduce HPSF by anestimated 80%–90% (9). Similarly, embolization ofvarices in patients with arterioportal and portosystemicshunting blocks routes carrying shunted material to thelungs. A similar technique of temporary portal vein bal-loon occlusion has been used for transarterial chemo-embolization in patients with arterioportal shunting (25).Bland arterial embolization long before RE was

shown to be ineffective and resulted in several cases offatal RP, likely from recurrence of shunting (26). Blandarterial embolization is probably more effective per-formed on the day of treatment (26), but its effective-ness in reducing shunting is difficult to measure and tooptimize. Too much bland arterial embolization beforeRE could redirect the 90Y microspheres to nontargetbackground liver or cause premature stasis and insuffi-cient dose. Too little or too late bland arterial emboliza-tion could allow 90Y microspheres to traverse shuntsbefore achieving stasis. Repeat scintigraphy after blandparticle embolization in a few patients established arange of expected effect of up to 50% reduction, withbetter shunt reduction using midsized particles. To avoidwashout of microspheres before stasis is achieved, wefound that use of two parallel microcatheters and glassmicrospheres to achieve a tight bolus allowed us toachieve almost instantaneous stasis after delivery of90Y microspheres.Both of our patients who developed delayed RP presen-

ted with signs or symptoms only after subsequent trans-arterial chemoembolization. It is unknown whether theirpulmonary toxicity was simply delayed in onset or wastriggered by transarterial chemoembolization. RP secon-dary to external-beam radiotherapy typically manifestswithin 8 weeks after exposure (13), and previous accountsof RE-related RP reported onsets 1–4 months after REtreatment (4,5). Circulating chemotherapeutic agent aftertransarterial chemoembolization possibly played a rolein our patients. “Radiation recall” is a well-describedphenomenon where systemic chemotherapy, especiallytaxanes and anthracyclines such as doxorubicin, givenafter radiation can reactivate and accentuate the effect ofradiation on the skin or lungs (27). Although the exactmechanism of action is unclear, interference with repairof irradiated lung parenchyma, changes in microvascular


permeability, and damage of stem cells have been pro-posed. Some data suggest that sequential radiotherapyand chemotherapy may be more toxic to the lungsthan concomitant therapy (28). However, corticosteroidtherapy is typically effective and may allow additionalchemotherapy treatment, as with one of our patients.Patients treated with resin microspheres according to

dose reduction guidelines had worse outcomes, withlower objective response and disease control rates. Theseresults corroborate prior reports that found response andsurvival to correlate with hepatic and tumor absorbeddoses (16,29). It is unknown if the use of adjunctivetechniques, particularly bland arterial embolization, con-tributed to better response rates in patients treated with-out dose reduction. Patients with high HPSF are atparticular risk for subtherapeutic dosing because highHPSF and dose reduction reduce the actual absorbedtumor dose. For example, if a patient with a 20% HPSFis treated with the recommended 40% dose reduction,the absorbed dose to the liver is only 48% of the origi-nally intended dose: (100% � dose reduction) � (100%� HPSF). Ideally, real-time monitoring of tumor andliver dose would help define RE endpoints, and intra-procedural positron emission tomography may providesuch feedback (30).Limitations of this study include its retrospective

nature and multiple treatment groups with small samplesizes and changing treatment algorithms over time. Thevariety of pathologic diagnoses (14 different cell types)and the various adjunctive techniques employed pre-cluded a more detailed subgroup analysis because oflimited sample sizes. Estimated lung absorbed dosesbased on MIRD calculations used only estimated lungmass, which was not accurately measurable. Only clinicaltoxicity was analyzed because actual pulmonary deposi-tion of 90Y was not measurable. Diagnosis of RP lackedcertainty because chronology was atypical and findingscould have overlapped with other neoplastic, infectious,or toxic conditions including transarterial chemoemboli-zation toxicity (31,32). Despite the suspicion for radiationrecall, the actual etiology of toxicity is uncertain.In conclusion, current recommendations for dose acti-

vity reduction for high HPSF are oversimplified, maycompromise therapeutic efficacy, and are most oftenavoidable without substantially increased risk of radia-tion pneumonitis. Instead of reducing the treatmentdose, catheter-based techniques can be used to decreasethe transit of microspheres into the lungs. Rather thanusing only the shunt fraction percentage, MIRD calcu-lation of anticipated dose to the lungs, especially whentotal administered activity is low, should be a bettermetric for risk assessment. None of our patients withexpected lung dose o 30 Gy experienced complications,including patients who underwent no prophylactic mea-sures. For patients with higher anticipated lung dose,catheter-based techniques during RE treatment maysuccessfully mitigate shunting.



REFERENCES

1. Riaz A, Lewandowski RJ, Kulik LM, et al. Complications followingradioembolization with yttrium-90 microspheres: a comprehensive liter-ature review. J Vasc Interv Radiol 2009; 20:1121–1130.

2. Chan WS, Poon WL, Cho DH, Chiu SS, Luk SH. Transcatheterembolisation of intrahepatic arteriovenous shunts in patients withhepatocellular carcinoma. Hong Kong Med J 2010; 16:48–55.

3. Sugano S, Miyoshi K, Suzuki T, Kawafune T, Kubota M. Intrahepaticarteriovenous shunting due to hepatocellular carcinoma and cirrhosis, andits change by transcatheter arterial embolization. Am J Gastroenterol1994; 89:184–188.

4. Wright CL, Werner JD, Tran JM, et al. Radiation pneumonitis followingyttrium-90 radioembolization: case report and literature review. J VascInterv Radiol 2012; 23:669–674.

5. Leung TW, Lau WY, Ho SK, et al. Radiation pneumonitis after selectiveinternal radiation treatment with intraarterial 90yttrium-microspheres forinoperable hepatic tumors. Int J Radiat Oncol Biol Phys 1995; 33:919–924.

6. Kennedy A, Nag S, Salem R, et al. Recommendations for radioembo-lization of hepatic malignancies using yttrium-90 microsphere brachyther-apy: a consensus panel report from the radioembolization brachytherapyoncology consortium. Int J Radiat Oncol Biol Phys 2007; 68:13–23.

7. Lau WY, Kennedy AS, Kim YH, et al. Patient selection and activityplanning guide for selective internal radiotherapy with yttrium-90 resinmicrospheres. Int J Radiat Oncol Biol Phys 2012; 82:401–407.

8. Snyder WS, Ford MR, Warner GG, Watson SB. MIRD pamphlet no.11: "S" absorbed dose per unit cumulated activity for selected radio-nuclides and organs. Reston, VA: Society of Nuclear Medicine; 1975,Available at: http://interactive.snm.org/docs/MIRD%20Pamphlet%2011.1-69.pdf. Accessed October 1, 2015.

9. Bester L, Salem R. Reduction of arteriohepatovenous shunting bytemporary balloon occlusion in patients undergoing radioembolization.J Vasc Interv Radiol 2007; 18:1310–1314.

10. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluationcriteria in solid tumours: revised RECIST guideline (version 1.1). EurJ Cancer 2009; 45:228–247.

11. US Department of Health and Human Services; National Institutes ofHealth; National Cancer Institute. Common Terminology Criteria forAdverse Events (CTCAE) Version 4.0. Available at: http://evs.nci.nih.gov/ftp1/CTCAE/CTCAE_4.03_2010-06-14_QuickReference_5x7.pdf. AccessedFebruary 1, 2015.

12. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment forhepatocellular carcinoma. Semin Liver Dis 2010; 30:52–60.

13. Graves PR, Siddiqui F, Anscher MS, Movsas B. Radiation pulmonarytoxicity: from mechanisms to management. Semin Radiat Oncol 2010;20:201–207.

14. Ho S, Lau WY, Leung TW, Chan M, Johnson PJ, Li AK. Clinicalevaluation of the partition model for estimating radiation doses fromyttrium-90 microspheres in the treatment of hepatic cancer. Eur J NuclMed 1997; 24:293–298.

15. Salem R, Parikh P, Atassi B, et al. Incidence of radiation pneumonitisafter hepatic intra-arterial radiotherapy with yttrium-90 microspheresassuming uniform lung distribution. Am J Clin Oncol 2008; 31:431–438.

16. Lam MG, Banerjee A, Goris ML, et al. Fusion dual-tracer SPECT-basedhepatic dosimetry predicts outcome after radioembolization for a wide


range of tumour cell types. Eur J Nucl Med Mol Imaging 2015; 42:1192–1201.

17. Wondergem M, Smits ML, Elschot M, et al. 99mTc-macroaggregatedalbumin poorly predicts the intrahepatic distribution of 90Y resin micro-spheres in hepatic radioembolization. J Nucl Med 2013; 54:1294–1301.

18. Elschot M, Nijsen JF, Lam MG, et al. (99m)Tc-MAA overestimates theabsorbed dose to the lungs in radioembolization: a quantitative evaluationin patients treated with 166Ho-microspheres. Eur J Nucl Med MolImaging 2014; 41:1965–1975.

19. Yu N, Srinivas SM, Difilippo FP, et al. Lung dose calculation withSPECT/CT for 90yttrium radioembolization of liver cancer. Int J RadiatOncol Biol Phys 2013; 85:834–839.

20. Pinjaroen N, Kaufman J, Kolbeck KJ, Farsad K, Watson L. Clinicalcorrelation between hepatic vein opacification and hepatopulmonaryshunt fraction for 90Y evaluation. J Vasc Interv Radiol 2013; 24:S164.

21. Gaba RC, Zivin SP, Dikopf MS, et al. Characteristics of primary andsecondary hepatic malignancies associated with hepatopulmonary shunt-ing. Radiology 2014; 271:602–612.

22. Theysohn JM, Schlaak JF, Muller S, et al. Selective internal radiationtherapy of hepatocellular carcinoma: potential hepatopulmonary shuntreduction after sorafenib administration. J Vasc Interv Radiol 2012; 23:949–952.

23. Rose SC, Hoh CK. Hepatopulmonary shunt reduction using chemo-embolization to permit yttrium-90 radioembolization. J Vasc Interv Radiol2009; 20:849–851.

24. Gaba RC, Vanmiddlesworth KA. Chemoembolic hepatopulmonaryshunt reduction to allow safe yttrium-90 radioembolization lobectomyof hepatocellular carcinoma. Cardiovasc Intervent Radiol 2012; 35:1505–1511.

25. Murata S, Tajima H, Nakazawa K, Onozawa S, Kumita S, NomuraK. Initial experience of transcatheter arterial chemoembolization duringportal vein occlusion for unresectable hepatocellular carcinoma withmarked arterioportal shunts. Eur Radiol 2009; 19:2016–2023.

26. Gramenzi A, Golfieri R, Mosconi C, et al. Yttrium-90 radioembolization vssorafenib for intermediate-locally advanced hepatocellular carcinoma: acohort study with propensity score analysis. Liver Int 2015; 35:1036–1047.

27. Ding X, Ji W, Li J, Zhang X, Wang L. Radiation recall pneumonitisinduced by chemotherapy after thoracic radiotherapy for lung cancer.Radiat Oncol 2011; 6:24.

28. Vogelius IR, Bentzen SM. A literature-based meta-analysis of clinicalrisk factors for development of radiation induced pneumonitis. ActaOncol 2012; 51:975–983.

29. Garin E, Rolland Y, Edeline J, et al. Personalized dosimetry withintensification using 90Y-loaded glass microsphere radioembolizationinduces prolonged overall survival in hepatocellular carcinoma patientswith portal vein thrombosis. J Nucl Med 2015; 56:339–346.

30. Bourgeois AC, Chang TT, Bradley YC, Acuff SN, Pasciak AS. Intra-procedural yttrium-90 positron emission tomography/CT for treatmentoptimization of yttrium-90 radioembolization. J Vasc Interv Radiol 2014;25:271–275.

31. Brown KT. Fatal pulmonary complications after arterial embolizationwith 40-120- micro m tris-acryl gelatin microspheres. J Vasc Interv Radiol2004; 15:197–200.

32. Kwok PC, Lam TW, Lam CL, Lai AK, Lo HY, Chan SC. Rare pulmonarycomplications after transarterial chemoembolisation for hepatocellularcarcinoma: two case reports. Hong Kong Med J 2003; 9:457–460.


http://refhub.elsevier.com/S1051-0443(15)00849-0/sbref1
























dx.doi.org/http://interactive.snm.org/docs/MIRD%20Pamphlet%2011.1-69.pdf

dx.doi.org/http://interactive.snm.org/docs/MIRD%20Pamphlet%2011.1-69.pdf







http://evs.nci.nih.gov/ftp1/CTCAE/CTCAE_4.03_2010-06-14_QuickReference_5x7.pdf

http://evs.nci.nih.gov/ftp1/CTCAE/CTCAE_4.03_2010-06-14_QuickReference_5x7.pdf








































































management of high hepatopulmonary shunting in patients ... · hepatopulmonary shunting is measured...

Documents