Radioembolization with 90Yttrium Microspheres: A State-of ... · Radioembolization with 90Yttrium...

Radioembolization with 90Yttrium Microspheres:A State-of-the-Art Brachytherapy Treatment forPrimary and Secondary Liver MalignanciesPart 1: Technical and Methodologic Considerations

Riad Salem, MD, MBA, and Kenneth G. Thurston, MA

Microsphere and particle technology represent the next-generation agents that have formed the basis of interventionaloncology, an evolving subspecialty of interventional radiology. One of these platforms, yttrium-90 microspheres, israpidly being adopted in the medical community as an adjunctive therapeutic tool in the management of primary andsecondary liver malignancies. Given the complexity of the treatment algorithm of patients who may be candidates forthis therapy and the need for clinical guidance, a comprehensive review of the methodologic and technical consid-erations was undertaken. This experience is based on more than 900 90Y infusions performed over a 5-year period.

J Vasc Interv Radiol 2006; 17:1251–1278

Abbreviations: ECOG � Eastern Cooperative Oncology Group, FDA � Food and Drug Administration, GDA � gastroduodenal artery, HCC � hepatocellularcarcinoma, LSF � lung shunt fraction, MAA � macroaggregated albumin, PET � positron emission tomography, PS � performance status, SPECT � singlephoton emission computed tomography, TACE � transcatheter arterial chemoembolization

YTTRIUM-90 microspheres are 20- to40-�m particles that emit �-radiation.Because the microspheres are deliv-ered via the hepatic arterial route, theprocess can be considered as internalrather than external radiation. Thetreatment algorithm is analogous tothat followed with transarterial che-moembolization (TACE). Clinical his-tory, physical examination, laboratoryvalues, and performance status (PS)are evaluated. Patients’ conditions areinitially evaluated and their disease isstaged with cross-sectional imagingtechniques: computed tomography

(CT), magnetic resonance (MR) imag-ing, and/or positron emission tomog-raphy (PET). When a patient is consid-ered a possible candidate for therapy,evaluation with mesenteric angiogra-phy followed by treatment on a lobarbasis is undertaken. Patients are fol-lowed up clinically to assess toxicitiesand response before treatment of theother lobe is undertaken.

TheraSphere (glass microsphere;MDS Nordion, Kanata, ON, Canada)was approved in 1999 by the U. S.Food and Drug Administration (FDA)under a humanitarian device exemp-tion for the treatment of unresectablehepatocellular carcinoma (HCC) in pa-tients who can have appropriately po-sitioned hepatic arterial catheters (1).Medical professionals are directed topublished FDA guidance documentson humanitarian device exemptionsfor uses in diseases other than HCC(2). SIR-Spheres (resin microsphere;Sirtex Medical, Lane Cove, Australia)were granted full premarketing ap-proval in 2002 by the FDA for thetreatment of colorectal metastases inconjunction with intrahepatic floxuri-

dine (3). Given the dearth of publishedliterature on the technical and meth-odologic considerations required forproper 90Y implementation and usage,this comprehensive overview was un-dertaken. For the purposes of this ar-ticle, unless otherwise stated, 90Y mi-crospheres, radioembolization, andselective internal radiation therapywill refer to the use of TheraSphereand SIR-Spheres.

The use of 90Y for the primary andsecondary treatment of liver malig-nancies is not investigational or exper-imental (2). Given the FDA approvalfor both devices, their use in HCC andcolorectal cancer represents their ap-proved indication. For disease statesother than the strict indication, the useof 90Y represents the practice of med-icine. This article is the first of a seriesof three that will be published on thetopic of radioembolization. The firstpart will focus on the technical andmethodologic considerations. The sec-ond will discuss special topics as theyrelate to 90Y microspheres. The thirdwill provide a comprehensive litera-ture review on the topic of radioembo-

From the Department of Radiology (R.S.), Robert H.Lurie Comprehensive Cancer, 676 North St Clair,Suite 800, Chicago, Illinois 60611; and MDS Nordion(K.G.T.), Kanata, Canada. Received January 9, 2006;accepted June 7. Address correspondence to R.S.;E-mail: [email protected]

R.S. is a consultant for MDS Nordion and has lec-tured for Sirtex Medical. K.G.T. is Director of Clin-ical Affairs for MDS Nordion. No research supportwas provided for this manuscript, and neither man-ufacturer had any input in this document.

© SIR, 2006

DOI: 10.1097/01.RVI.0000233785.75257.9A

1251

lization and discuss future directionsfor this technology. It should be notedthat some of the discussions presentedin parts 1 and 2 represent the opinionsof the authors, whereas part 3 repre-sents a strict review of the literature.This experience is based on more than900 90Y infusions performed over a5-year period by a multidisciplinaryteam including investigators frommedical oncology, hepatology, sur-gery, transplantation, and interven-tional radiology as the authorized us-ers.

TECHNICAL ANDMETHODOLOGICCONSIDERATIONS

Overview

Radioembolization is defined as theinjection of micron-sized embolic par-ticles loaded with a radioisotope byuse of percutaneous transarterial tech-niques. There are two distinct aspectsto the procedure. The first is the injec-tion of embolic particles (ie, “emboli-zation”) as the vehicle; the second isthe delivery and administration viathis embolic vehicle of radiation (“ra-dio-”). Fluoroscopic guidance, angio-graphic endpoints of embolization andstasis, and the need to modify this onthe basis of angiographic findingsmake this treatment a true emboliza-tion procedure. In addition, dosimetryplanning, the administration and de-livery of radiation, modification ofdose on the basis of tumor and hepaticvolume, and the required knowledgeof radiation effects on tissue make thisa brachytherapy procedure as well.Although the term “selective internalradiation therapy” has also been usedto describe this therapy, “radioembo-lization” more accurately describes theactual mode of action of 90Y micro-spheres according to the rationale de-scribed. Hence, for technologies thatrequire embolic particles to carry ra-dioisotopes to the targeted tumors, wepropose that the term “radioemboliza-tion” be formally recognized.

HCC represents one of the mostcommon forms of cancer, with morethan 1 million new cases estimated an-nually worldwide. In the UnitedStates, the incidence of HCC hassteadily increased during the past twodecades, an estimated 18,900 newcases having been diagnosed in 2004

(4). Traditionally, these patients havehad few treatment options (5). Thesafety and therapeutic benefit of 90Ymicrospheres for HCC is well docu-mented in the literature (6–8).

The evaluation of unresectableHCC is significantly different fromthat of metastatic disease. Curative op-tions include liver transplantation andresection (9). Unfortunately, only10%–15% of patients are candidatesfor curative therapy (9). Ideal candi-dates for treatment with 90Y micro-spheres include patients with a perfor-mance status (PS) of 0–2, intact liverfunction, and a patent portal vein. Un-like patients with metastatic disease tothe liver, pathologic confirmation ofHCC is not always necessary and maybe established in patients with classichistory (ie, alcohol or viral hepatitis),imaging findings (ie, hypervasculartumors, cirrhosis) and a serum �-feto-protein level greater than 400 ng/mL(10). The benefits of radioembolizationwith 90Y in patients with HCC hasbeen previously described (7,11–19).

Patients with metastatic cancer tothe liver often have complex medicalhistories. In cases of colon cancer, ifthe disease is detected in the earlystages, resection of the primary tumorwithout lymph node involvement mayobtain long-term cure. In some cases,patients with stage IV disease withliver metastases may be treated withsurgical resection alone, also provid-ing a chance for long-term cure. Inpatients with surgically unresectableliver disease with or without extrahe-patic disease, systemic chemotherapyhas become the standard for first- andsecond-line treatment (20,21). Combi-nation therapy with angiogenesis in-hibitors and surgical resection hasnow become an integral part of first-and second-line therapies (22). Pa-tients with liver-dominant disease inwhich standard first- and second-linetherapies have failed may be consid-ered for treatment with 90Y.

The liver is the most frequent site ofmetastases, primarily as a result of thespread of cancer cells through the por-tal circulation. In fact, approximately60% of patients with colorectal carci-noma eventually have liver disease asthe predominant site (23). Similarly toHCC, surgical resection of metastatichepatic disease is the treatment ofchoice. However, surgical resection isfeasible in fewer than 20% of patients

(23). The benefits of radioembolizationwith 90Y in these patients has beenreported in many studies (24–29).

PATIENT SCREENING ANDSELECTION

Clinical Presentation and ImagingCorrelates in HCC

The selection process for patientsundergoing 90Y treatment is multifac-torial. Simply put, ideal patientsshould have liver-only or liver-domi-nant disease, minimal comorbidities,and normal liver function test results.Patients with HCC may have a clinicalhistory of viral (hepatitis B or C virus)or alcoholic cirrhosis. In rare instances,patients may present with cirrhosis ofuncertain cause, a condition often re-ferred to as nonalcoholic steatohepati-tis (30). Depending on the severity ofthe disease, patients can manifestother sequelae of cirrhosis such as en-cephalopathy, ascites, and portal hy-pertension with or without portal veinthrombosis. Patients with HCC mayhave varied surgical and therapeutichistories, including previous resection,radiofrequency ablation, or TACE.Clinical considerations in these pa-tients include the degree of hepaticcompromise and imaging findings.

Hepatoma findings on imaging arequite variable (31). If ultrasound (US)is the initial diagnostic modality, ad-ditional cross-sectional imagingshould be performed. Other than op-erator dependence, altered hepaticechotexture can result in a high false-negative rate, especially for smaller le-sions. Also, in some cases, infiltrativetumors can be misdiagnosed as pelio-sis hepatis (32). For patients with cir-rhosis, any hepatic mass should beconsidered a hepatoma until provenotherwise, warranting further investi-gation. Triple-phase CT is highly sen-sitive in the detection of hepatoma(33). Because many of these tumors arehypervascular, scanning in the earlyphases allows for detection. Later-phase imaging is necessary to detectother less vascular lesions and multi-focality, as well as to identify portalvein patency (33). Alternatively, MRimaging can also be used to identifyand characterize HCC lesions, withspecific attention to diffusion-weighted imaging sequences and oxy-genation (34,35). The diagnosis of

1252 • Radioembolization with 90Y Microspheres for Liver Malignancies August 2006 JVIR

HCC can be confirmed by invasivemeans (ie, fine needle aspiration orcore biopsy) or noninvasive means (ie,imaging criteria, �-fetoprotein level�400 ng/mL) (10). If HCC can be di-agnosed according to imaging andbiochemical criteria, biopsy should beavoided, given the risk of tract seedingafter percutaneous needle proceduressuch as radiofrequency ablation (36).CT and MR imaging are useful in theposttreatment identification of necro-sis and cell death, allowing assessmentof tumor response to 90Y therapy(10,37).

On the basis of a retrospectivelyperformed review of historical data,Okuda et al (38) developed a stagingsystem for patients with HCC. Fromthis, historical controls were devel-oped, and median survival for un-treated HCC was estimated to be 3–6months. The variability results in partfrom the late stage of disease withwhich many of these patients present.With proper screening of patients athigh risk, it is possible to detect hepa-toma at the point at which adequatehepatic reserve is present. The idealpatient for 90Y therapy has less than50% tumor burden, no ascites, normalbilirubin level, and albumin levelgreater than 3 g/dL, translating intoOkuda stage 1. Although the lobartreatment approach increases thesafety profile of 90Y therapy, patientswith increased liver function test re-sults (�1.2 g/dL) should be treatedcautiously (discussed later). In thesepatients, if the tumor can be isolatedangiographically, segmental/subseg-mental infusion (just as with TACE)should be undertaken to provide ther-apeutic benefit (such as downstagingfor transplantation, radiofrequency

ablation, or resection) while minimiz-ing the risk of hepatic decompensation(39). Recently, other staging systemshave supplanted the Okuda system,such as the Cancer of the Italian LiverProgram, Child-Pugh, or BarcelonaClinic Liver Cancer systems (40). Inour center, we stage disease with theOkuda, Child-Pugh, Cancer of the Ital-ian Liver Program, and BarcelonaClinic Liver Cancer scales.

In summary, HCC should be stagedappropriately to provide valuableprognostic information. The patient’scondition should be quickly assessedfor possible resection or transplanta-tion. If not, it is possible that the pa-tient might be a candidate for down-staging of HCC (41). Because themajority of HCC patients have unre-sectable disease, palliative optionssuch as TACE, drug-eluting micro-spheres, and 90Y treatment may beconsidered.

Clinical Presentation and ImagingCorrelates in Metastatic Disease

Metastatic disease to the liver is themost common form of hepatic malig-nancy (42). In particular, given the in-creasing incidence of colorectal cancer,percutaneous management will con-tinue to play an increasingly impor-tant role in this disease (43). Patientswith metastatic cancer usually havebetter PS than those with hepatomaand have successfully completed twoor three different courses of chemo-therapy. First- and second-line thera-pies usually include 5-fluorouracil/leucovorin, irinotecan, and oxaliplatin,with or without bevacizumab. The in-herently rich blood supply of somemetastatic tumors such as renal or

neuroendocrine tumors makes themideal for intraarterial therapy.

Factors to consider in patients withmetastatic disease to the liver includea history of chemotherapy, surgical re-section, infusion pump placementwith surgically altered vascularity,and imaging findings of liver-domi-nant disease. Of particular interest isthe increasing pool of patients under-going liver-directed therapy after fail-ure of growth-factor inhibitors such asbevacizumab and cetuximab. How-ever, in all patients, one of the mostimportant factors in determining eligi-bility for 90Y treatment is Eastern Co-operative Oncology Group (ECOG)PS. Patients presenting with clearlycompromised functional status (ECOG2–4) (Table 1) are at high risk for rapidonset of liver failure and treatment-re-lated morbidity. Although each patientdeserves individual consideration, care-ful thought should precede treatment ofthese clinically compromised patients.

Findings on cross-sectional imagingin patients with liver metastases arerelatively consistent. Unlike HCC, inwhich a diagnosis can be made ac-cording to imaging criteria, these donot exist for metastatic lesions. Hence,when a mass is identified in a patientwith metastatic disease, pathologicconfirmation is usually necessary. PETshould play an integral role in the dis-ease staging and clinical follow-up ofpatients receiving 90Y therapy. Giventhe recognized limitations of tumorsize in the follow-up of patients receiv-ing liver-directed therapy, PET appearsto provide more salient information re-garding functional improvement aftertreatment (28,42,44,45). The use of PETin this setting is analogous to its clinicaluse in patients with lymphoma (46).

Irrespective of the cause of the livertumors, evaluation of a patient forpossible treatment with 90Y should bedriven by the patient’s ECOG PSrather than the clinical disease stage.In the current authors’ experience, PShas been the most reliable indicator ofa patient’s ability to tolerate 90Y treat-ment (16,42). From a practical stand-point, patients with an ECOG PS of0–2 (equivalent to Karnofsky score of�60%) should be considered for ther-apy (Table 1). For patients who do notfulfill this criterion, the decision totreat should be based on individualpatient evaluation and clinical judg-ment.

Table 1ECOG Performance Status and Karnofsky Score

ECOGScale Characteristics

EquivalentKarnofsky Score

(%)

0 Asymptomatic and fully active 1001 Symptomatic; fully ambulatory; restricted in physically

strenuous activity80–90

2 Symptomatic; ambulatory; capable of self-care; �50% ofwaking hours are spent out of bed

60–70

3 Symptomatic; limited self-care; spends �50% of timein bed

40–50

4 Completely disabled; no self-care; bedridden 20–30

Salem and Thurston • 1253Volume 17 Number 8

In summary, patients with meta-static disease to the liver should alsohave their disease staged appropri-ately. They should have liver-only orliver-dominant disease. Some maypresent with painful bulky tumorsthat require palliation. Unless contra-indicated, patients should have com-pleted standard-of-care chemother-apy. In such cases, palliative optionssuch as TACE, bland embolization,and 90Y therapy may be considered(42,47).

Liver Function Parameters andTumor Markers in HCC andMetastatic Disease

As described in the oncologic liter-ature, the best indicators of overallliver function include prothrombintime and levels of albumin and totalbilirubin (48). This is particularly trueof patients with HCC. Just as patientswith normal findings in these threeparameters have better outcomes ofsystemic chemotherapy, 90Y patientswill also have better long-term out-comes if these three biochemical fac-tors are within normal range (48). Ide-ally, before treatment with 90Y,patients should have recent determi-nation of laboratory values, includingliver function and complete bloodcount with differential. Patients withHCC should also have recent mea-surements of prothrombin time andInternational Normalized Ratio.

Other parameters that should beevaluated before treatment of livercancers include tumor markers, mostcommonly �-fetoprotein for hepatoma(�400 ng/mL is usually diagnostic),carcinoembryonic antigen for colorec-tal malignancies, cancer antigen 19–9for tumors of pancreaticobiliary ori-gin, and serotonin/5-hydroxyindoleacetic acid/chromogranin A for someneuroendocrine tumors. Although notall tumors produce these markers,they may be helpful during follow-up,when response to treatment is beingassessed and quantified biochemi-cally. If these tumor markers are mea-sured immediately after treatment, afalse increase caused by tumor lysismay be observed. Although nonspe-cific, other markers such as lactate de-hydrogenase and C-reactive proteinmay be helpful in posttreatment mon-itoring. C-reactive protein has also

been shown to have prognostic valuein patients with HCC (49).

VASCULAR ANATOMY, TARGETVASCULAR BED, PULMONARYSHUNT, ANDGASTROINTESTINAL FLOW

Given the propensity for arterialvariants and hepatic tumors to exhibitarteriovenous shunting, all patientsbeing evaluated for 90Y must undergopretreatment mesenteric angiography(17,39,50). In summary, meticulous,detailed, and power-injection imagingof the hepatic and visceral vasculatureshould be performed in all patientsbefore treatment. Vessels that must beinterrogated include the celiac artery,common and proper hepatic arteries,gastroduodenal artery (GDA), andright and left hepatic arteries. Knowl-edge of the arterial anatomy allows forthe administration of 90Y, with onetreatment to each target vascular bed(usually the hepatic lobe or segment)at 30- to 60-day intervals. However,when variants are present, treatmentplans must be tailored to ensure safeand accurate delivery of 90Y micro-spheres. For instance, in some cases, amiddle hepatic artery may be present,usually arising from the right hepaticartery. Despite accurate catheter place-ment, this anatomy may preclude de-livery of 90Y to the segment suppliedby the middle hepatic artery, given theflow dynamics. Therefore, three treat-ments (one each to the right, middle,and left hepatic arteries) may be nec-essary to cover the entire liver. An-other example might include a patientwith an accessory right hepatic artery,necessitating a third treatment (left he-patic [segments 2–4], right hepatic[segments 5/8], and accessory righthepatic [segments 6/7]) (50).

Pretreatment angiography allowsfor accurate calculation of the targetvolumes. Although discussed in latersections, target volume deserves spe-cial attention. Irrespective of the 90Yagent used, it is imperative that do-simetry calculations be based on thevolume of the target vascular bed sup-plied by the artery to be catheterized.For instance, the simplest scenariomight include a patient with standardanatomy, with single right and left he-patic arteries. In this case, if the lesionsto be treated are located within theright lobe, the target volume includes

the entire right lobe, with the middlehepatic vein (or gallbladder) as theseparator between the right and leftlobes (see Treatment Process section).In the case of a patient with right andaccessory right hepatic arteries, tech-niques in target volume calculationswill have to be altered. Usually, theaccessory right hepatic artery will sup-ply the posterior segment of the rightlobe (segments 6/7). The right hepaticartery branching off the proper he-patic artery will supply the anteriorsegment of the right lobe (segments5/8), and the left hepatic artery willsupply the left lobe (segments 2–4).Knowledge of this vascular anatomypermits accurate lobar volume calcu-lations for prescribing the correct ac-tivity to deliver the optimal 90Y dose.To summarize, the volume used fordose calculation is determined by thevolume of the liver segment(s) beingsupplied by the artery to be infused. Inaddition, the determination of targetvolume is best performed by the inter-ventional radiologist who performedthe initial angiographic assessment ofthe patient. It is the interventional ra-diologist who can optimally correlatethe CT findings (ie, location of tumors,burden) and angiographic findings (ie,standard vs variant anatomy). Hence,it is highly recommended that calcula-tions of volumes be determined by theinterventional radiologist who hasperformed the diagnostic mesentericangiogram on the patient to be treated.

Pulmonary Shunting andTechnetium-99m MacroaggregatedAlbumin Scanning

In contrast to metastatic tumors tothe liver, one of the angiographic char-acteristics of HCC, other than portalvein thrombosis, is direct arterio-venous shunting bypassing the capil-lary bed (51). As a result, one of theconcerns with administration of 20- to40-�m 90Y microspheres is directshunting to the lungs, possibly result-ing in radiation pneumonitis (52).When the catheter is placed in theproper hepatic artery, 4- to 5-mCi99mTc macroaggregated albumin(MAA) is administered intraarterially.Because the size of 99mTc-MAA parti-cles closely mimics that of 90Y micro-spheres, it is assumed that the distri-bution of the microspheres will be


identical with that of 99mTc-MAA, andthis distribution is used in the plan-ning process. Lung shunting is as-sessed with planar and/or single pho-ton emission CT (SPECT) �-cameras.In our experience, we have used pla-nar and fixed dual-head image acqui-sition techniques to assess shuntingimmediately after injection. Single-projection anterior planar techniqueshave the advantage of speed and easeof use. SPECT �-camera techniques intwo or three projections have the the-oretical advantage of higher resolutionand sensitivity. However, for SPECTto be performed, patients must betransferred from the angiographysuite to the nuclear medicine depart-ment. The transfer time may allow forthe normal degradation of 99mTc-MAAwith time and may falsely increase thedegree of shunting to the lungs. In ourexperience, we have found no clini-cally significant difference in the de-gree of lung shunting as measured byplanar or SPECT techniques (53). Ei-ther technique with the use of a por-table �-camera or fixed dual-headSPECT imaging in the nuclear medi-cine department is acceptable for theaccurate assessment of shunting. Lungshunting fraction is described in thepackage inserts for glass and resin mi-crospheres (1,3) and is determined bythe following equation:

Lung shunt fraction �

total lung counts/(total lung �

liver counts)

However, there are several technicalnuances to the assessment of shunting.If a patient has a solitary HCC and ifonly one treatment is planned, injec-tion of the 99mTc-MAA into the arterythat is planned for injection is indi-cated (in lobar or segmental infusion).However, if the diagnosis is multifocalbilobar HCC, 99mTc-MAA injectionand lung shunting should be assessedbefore each treatment at the lobarlevel. This is because HCC tumors lo-cated in different lobes may shunt tovarying degrees. Without this infor-mation, the total cumulative pulmo-nary dose may be inadvertently ex-ceeded.

In patients with metastatic disease,significant shunting is rare unless thetumor burden is very high. Hence,lung shunting can be assessed once

with catheter placement and 99mTc-MAA injection within the proper he-patic artery at the time of planningvisceral arteriography. Overall, in ourinstitution, we have found the inci-dence of significant lung shuntingwith HCC or metastatic disease to bewell below 10%. As a result, we havenow adopted injection of the 99mTc-MAA within the proper hepatic arteryfor all disease states (ie, HCC, metas-tases) unless gross arteriovenous/por-tal shunting is noted angiographically.If angiographic shunting is observed,lobar 99mTc-MAA imaging is per-formed.

Timing of imaging after injection of99mTc-MAA is also important. Thereare three potential causes of the over-estimation of lung shunt. First, giventhat 99mTc-MAA represents a radioiso-tope/protein structure, there is a time-dependent breakdown into smallerparticles of the proteinaceous 99mTc-MAA, with subsequent migration ofthese smaller fragments via the nor-mal capillary bed to the lungs. Thismay be a problem if too much timeelapses between the injection of 99mTc-MAA in the angiography suite and thetime of imaging in the nuclear medi-cine department. As a result, it is im-portant that patients undergo imagingas quickly as possible after injection of99mTc-MAA. Second, although thenormal manufactured size of 99mTc-MAA particles is 30–90 �m, statisti-cally, a small percentage of these par-ticles will fall outside of this range.Particles that are initially manufac-tured smaller than 8–10 �m will shuntthrough the normal capillary system,which will result in an increase in per-ceived lung shunt. According to mostmanufacturers, fewer than 10% of theMAA particles are smaller than 10 �m(54). Third, free technetium may resultin an inaccurate estimation of lungshunt. The interpretation of the lungshunt fraction and gastrointestinal up-take must take into consideration thepresence of 99mTc-MAA. Uptake in thethyroid and salivary glands and kid-neys, as well as diffuse gastric mucosaluptake, should not be consideredshunting. Uptake in the gastric mu-cosa, the small bowel, or pancreas inthe absence of salivary and thyroiduptake should be interpreted withcaution because it may represent truegastrointestinal shunting.

In patients with variant anatomy in

which whole liver shunting assess-ment is planned, fractionated injectionof 99mTc-MAA is recommended. Forexample, a patient with a replaced lefthepatic artery should receive 1–2 mCiof 99mTc-MAA in the replaced left he-patic artery, with the other 2–3 mCiinjected in the right hepatic artery. De-pending on the variant anatomy that isidentified, the 4- to 5-mCi vial of99mTc-MAA should be fractionatedduring the planning angiography pro-cedure in such a manner that the en-tire liver is imaged in one setting.

Finally, it should be stated that theinjection of 99mTc-MAA should beused to absolutely determine lungshunting fraction only. An authorizeduser should not rely solely on theSPECT images of the upper abdomentract to absolutely exclude gastrointes-tinal shunting. Rather, it should beconsidered an adjunctive imaging mo-dality of the gastrointestinal tract. Ex-clusion of gastrointestinal flow shouldbe accomplished by use of the com-bined information obtained from me-ticulous hepatic angiography, three-dimensional CT angiography, andSPECT imaging. Fusion of CT and99mTc-MAA SPECT images may alsobe helpful in the identification of ex-trahepatic flow of 99mTc-MAA.

The lungs can tolerate 30 Gy pertreatment session and 50 Gy cumula-tively (55). Therefore, when treatmentplanning is undertaken and lungshunting is identified, the total cumu-lative pulmonary dose must be calcu-lated. Patients may be treated if theircumulative pulmonary dose will notexceed 50 Gy for all planned infusionsof 90Y. However, in patients with com-promised pulmonary function, such asthat resulting from chronic obstructivepulmonary disease or previous lungresection, caution must be taken as to-tal cumulative doses are approached.In these instances, the decision to treatmust be individualized and based onoverall clinical status. In some 90Ytreatment practices, conservative val-ues of 15 Gy per treatment and 30 Gycumulatively are used as dose limita-tions.

Gastrointestinal Flow viaExtrahepatic Vessels

The final consideration in the an-giographic evaluation of HCC andmetastatic disease is the presence of


flow to the gastrointestinal tract. Dur-ing each step of visceral angiography,assessment for potential gastric orsmall-bowel flow must be made. Theceliac arteriogram is used to detect thepresence of arterial variants such asreplaced left hepatic artery or doublehepatic arteries, as well as parasitiza-tion of blood flow to the liver tumors(50,56,57). Depending on the proxim-ity of the left gastric branch to theother hepatic branches (such as with agastrohepatic trunk), coil embolizationof this vessel may be warranted, be-cause this will minimize the risk ofreflux during 90Y infusion. The GDAshould be identified and prophylacti-cally embolized, particularly if themore embolic SIR-Spheres are beingconsidered. The right gastric arteryshould also be identified during an-giography. Although this vessel is of-ten seen to arise from the left hepaticartery, it can also arise from the com-mon, proper, or right hepatic artery aswell as the GDA (50,57,58). Dependingon the anatomic location of this vesseland the relative ease with which distalcatheterization may be achieved forinfusion, prophylactic coil emboliza-tion should be undertaken to mini-mize inadvertent deposition of 90Yinto the gastric bed. Other vessels thatshould be searched for include the cys-tic, supraduodenal, retroduodenal, fal-ciform, accessory left gastric, and rightand left inferior phrenic arteries(50,57).

Given all the complexities of angio-graphic assessment and the absoluterequirement for the prevention of gas-trointestinal flow of microspheres, thethreshold for prophylactic emboliza-tion of the GDA, right gastric artery,and other extrahepatic arteries shouldbe low, irrespective of the proximity ofthe intended infusion site (50). Pro-phylactic embolization of these vesselsis of no clinical consequence, becausethis represents the analogue to surgi-cally placed pumps for floxuridine in-fusion in patients with colorectal can-cer. Surgeons placing such pumpsroutinely ligate the GDA and rightgastric artery and interrupt any vascu-lar communication between the liverand stomach, small bowel, or dia-phragm. Reflux of microspheres intogastroduodenal or gastric circulationwill almost always result in grave clin-ical consequences, including severe ul-ceration, gastrointestinal bleeding, or

pancreatitis. Given the serious clinicalrisks of microsphere reflux into non-target organs when prophylactic em-bolization is not performed versus thelack of clinical consequences of pro-phylactic embolization of the GDA,right gastric artery, and/or other ex-trahepatic vessels, the latter approachis strongly favored (50,57,58).

The use of balloon occlusion tech-niques for liver-directed therapy hasbeen described previously (59). Thisapproach to 90Y infusion should bediscouraged for several reasons. First,balloon occlusion techniques wereused historically, given catheter sizeand fluoroscopic imaging limitations.Imaging techniques have improvedsignificantly since then, and microcath-eters and microcoils are now available.Second, the use of relatively large bal-loon occlusion systems may cause ves-sel injury, spasm, or dissection. Also,administration of 90Y is a flow-depen-dent therapy. If the vessels flowing totumor are patent, tumor hypervascular-ity will permit the microspheres to be“absorbed.” When balloon occlusion isused, hypervascular tumor flow dy-namics are eliminated, and micro-spheres are “pushed” within the targettissue and tumor. This may result ininadequate distribution. Finally, balloonocclusion physiology may facilitate the“layering” of microspheres, given thatthey are pushed into the target bed.

ALTERING VASCULARANATOMY TO OPTIMIZE 90YDELIVERY

The topic of alteration in vascularanatomy to optimize 90Y delivery isquite complex. This complexity rein-forces the importance of highlytrained and dedicated interventionalradiologists as an integral part of theteam to safely perform 90Y therapy. Asdescribed earlier, the proper identifi-cation of vessels involved in the he-patic and gastric distribution is crucialto successful treatment. Interventionalradiologists with proper trainingshould be able to perform the basictasks of catheter placement and coilembolization when necessary in 95%of potential candidates. However,there are instances when patientscould have vascular anatomy that maybe altered percutaneously to achievemore favorable anatomy. The simplestof these examples includes coil embo-

lization of the right gastric artery,thereby minimizing the possibility ofgastrointestinal toxicity. However,more complex alterations may be re-quired. This topic has been extensivelyaddressed by other investigators(50,57,58,60).

The GDA and right gastric arteriesrepresent the two most common ves-sels that will require angiographic at-tention and evaluation. AlthoughTheraSphere and SIR-Spheres repre-sent significantly different embolicloads, the prophylactic embolizationof the GDA and right gastric arteries isadvocated. As described earlier, theclinical sequelae of this are nonexist-ent, whereas the potential clinicalsafety benefits are enormous. With re-spect to the small-sized right gastricartery, despite difficulty in its identifi-cation, infusion of microspheres intothis vessel almost invariably leads togastrointestinal ulceration. At times,significant difficulty in catheterizationof the right gastric artery may be en-countered. In such circumstances, thenormal right/left gastric anastomoticarcade should permit successful em-bolization of the right gastric artery ina retrograde fashion via the left gastricartery (50,57).

The topic of redistribution of he-patic flow deserves special mention. Insome cases, such as small hepatic ves-sels, redistribution of blood flow maybe considered. This is accomplishedby the prophylactic embolization of amain vessel (eg, left hepatic arteryarising off the left gastric artery) or theembolization of a small middle hepaticartery to convert the vascular bed intoa standard or whole lobe to simplifythe 90Y infusion. Although hepaticvessels such as the left hepatic arteryare routinely ligated in preparation forintraarterial pump placement andfloxuridine administration, this hasnot been definitively shown to be ap-plicable to 90Y with use of percutane-ous techniques (61). Although floxuri-dine infused into a pump placed in theGDA with a ligated left hepatic arterywill perfuse the entire liver, it remainsto be determined whether this phe-nomenon is replicated with the infu-sion of 20- to 60-�m 90Y particles. It isnot clear whether 90Y microsphereswill travel through the hepatic sinu-soids as effectively as floxuridine.Nevertheless, investigators infusing90Y through surgical pumps have con-


sistently observed positive tumor re-sponses (25,26). Further investigationon flow distribution and the effect ontumor response as it relates to 90Y in-fusion is necessary.

SELECTION OF TREATMENTAPPROACH: SIMPLIFIEDTREATMENT MATRIX

Despite a growing body of litera-ture clarifying the pathophysiology ofHCC and metastatic disease to theliver, there remains an element of un-certainty in determining which pa-tients should receive treatment with90Y microspheres. In addition, after thepatient has been selected, the safestand most effective technique may notbe obvious to the operator who is notwell versed in 90Y microsphere use.

HCC

We have adopted a simplified ap-proach to the treatment of patientswith unresectable HCC who are notcandidates for transplantation or re-section. This approach involves twovariables: tumor presentation (ie, uni-focal or multifocal/bilobar disease)and total bilirubin level. For patientswith unifocal tumors and normal bili-rubin levels, treatment may be con-ducted via a lobar or angiographicallyisolatable vessel(s) (ie, segmental infu-sion). For patients with unifocal tumorand increased bilirubin levels, treat-ment with 90Y should proceed only if asegmental feeding vessel may be iso-lated, resulting in a high dose deliv-ered to the tumor and no radiation tothe normal parenchyma. The intent inthese patients would be definitivetherapy or downstaging of disease topermit transplantation. For patientswith multifocal tumor and normal bil-irubin levels, staged and lobar infu-sions should be undertaken, analo-

gous to the classical TACE paradigm.For patients with multifocal diseaseand increased bilirubin levels, therisks of 90Y treatment in this settingmay be unacceptably high (as dis-cussed later).

Table 2 displays a simplified 2-by-2decision matrix that has been createdto simplify and guide the user on theuse of 90Y in unresectable HCC. Thissimple table describes the most com-monly encountered clinical scenariosin patients with HCC. The manage-ment of these four scenarios is depen-dent on the following assumptions: (i)90Y microspheres, if administered on alobar basis, will preferentially flow tothe tumor but will invariably result insome irradiation of normal hepatic pa-renchyma; (ii) segmental/subsegmen-tal infusions will result in microspheredelivery to the target vascular bedonly, with no irradiation of the re-maining noninfused parenchyma; (iii)multifocal disease is defined by thepresence of multiple tumors within ahepatic lobe that derive their bloodsupply from the artery feeding thelobe; (iv) increased bilirubin levels arecaused by liver failure itself and not anobstructive process; and (v) in the ab-sence of biliary obstruction and/ormetabolic conditions causing an in-creased total bilirubin level (eg, Gil-bert syndrome), the definition of in-creased bilirubin levels should betailored appropriately. At the presentauthors’ institution, increased total bil-irubin level is defined as greater than1.3 mg/dL. Given these assumptions,it is possible to appropriately tailor thetreatment plan for each clinical sce-nario.

In patients with normal bilirubinlevels and multifocal/bilobar disease(scenario A) (Table 2), a TACE-likeparadigm should be followed, becausethis will allow for optimal and com-plete tumor coverage. In patients with

normal bilirubin and localized or uni-focal disease (scenario B), a lobar orsegmental infusion is appropriate.However, in this scenario, the autho-rized user must realize that perform-ing a lobar infusion in a case in whicha segmental infusion could technicallybe performed will result in the unnec-essary irradiation of normal paren-chyma. Given the propensity for newtumors to develop in cirrhotic liver,we will perform segmental infusionswhen possible, sparing the normal pa-renchyma and thereby allowing for fu-ture 90Y infusion to nonirradiated pa-renchyma. In cases of localized diseasewith increased bilirubin levels (sce-nario D), patients should be assessedfor possible treatment with use of asegmental approach only. The in-creased bilirubin level implies intrinsichepatocyte dysfunction or failure,with a tumor that has developedwithin the diseased liver. In such acase, if angiographic factors permitand a segmental feeding vessel is iden-tified on angiography, a patient maybe a candidate for palliative therapywith 90Y without injury to normal pa-renchyma (41). In some cases, patientsmay have liver failure with a liver tu-mor that exceeds the size criteria forliver transplantation. These patientsshould, at minimum, undergo diag-nostic mesenteric angiography in asearch for a segmental vessel to thetumor. If such a vessel is found, thesepatients can undergo segmental/sub-segmental infusion without risk of non-target radiation or alteration in Child-Pugh score (39). Treatment with 90Y insuch cases may result in sufficientdownstaging of disease to allow trans-plantation (62).

The management of the last cate-gory may represent end-stage liverdisease and is clearly the most difficult(scenario C) (Table 2). Patients withmultifocal disease and increased bili-rubin level are clearly at high risk forhepatic decompensation after anytreatment. In such cases, and in therecognition that an infusion of 90Y mi-crospheres will irradiate some normalparenchyma and may further worsenthe liver failure, treatment shouldprobably be withheld (as discussedlater). In cases in which treatment issought, hepatic artery chemoemboli-zation or bland particle embolizationin a lobar or preferably segmental

Table 2Treatment Decision Matrix for Liver Cancers

Yes No

TOTALBILIRUBIN

Normal Lobar infusions(scenario A)

Lobar or segmented infusions(scenario B)

Elevated Uncertain (scenario C) Segmented infusions only(scenario D)


manner might be considered if the bil-irubin level is 3.0 mg/dL or lower (63).

Metastatic Disease to the Liver

Table 2 may also be used to de-scribe the most commonly encounteredclinical scenarios in patients with met-astatic disease to the liver and can alsobe used to guide users on the use of90Y in such patients. The managementof the four scenarios in the 2-by-2 ma-trix, just as with HCC, is dependent onthe following assumptions, with somemodifications: (i) 90Y microspheres, ifadministered on a lobar basis, willpreferentially flow to the tumor butwill also result in some irradiation ofnormal hepatic parenchyma; (ii) seg-mental/subsegmental infusions willresult in microsphere delivery to thetarget vascular bed only, with no irra-diation of the remaining noninfusedparenchyma; (iii) multifocal disease isdefined by the presence of multipletumors within a hepatic lobe that de-rive their blood supply from the lobarartery; (iv) as opposed to HCC, pa-tients with metastatic disease to theliver with increased bilirubin levelsare more likely to have an obstructiveprocess such as portal lymphadenop-athy; (v) increased bilirubin level, inthe absence of biliary obstruction, isvery uncommon de novo in these pa-tients and is more likely to be the re-sult of chemotherapy toxicity (eg,from capecitabine) and/or infiltrative,end-stage metastatic disease withinthe liver that may or may not be ap-parent on imaging studies; and (vi) inthe absence of biliary obstructionand/or metabolic conditions causingan increased total bilirubin level (eg,Gilbert syndrome), the definition ofincreased bilirubin level should be tai-lored appropriately. At this authors’institution, increased total bilirubin isdefined by a level greater than 1.3 mg/dL. Given these assumptions, it is pos-sible to appropriately tailor the treat-ment plan for each clinical scenario.

In patients with normal bilirubinlevels and multifocal/bilobar disease(scenario A), lobar infusions should beperformed, because this will allow foroptimal and complete tumor coverage.In patients with normal bilirubin leveland localized or unifocal disease (sce-nario B), a lobar or segmental infusionis appropriate. However, just as withHCC, the treating physician must re-

alize that performing a lobar infusionin a case in which the blood supplycan be isolated and performed on asegmental basis will lead to irradiationof normal, non–tumor-containing pa-renchyma. Because new tumors maydevelop in patients with metastases astheir disease progresses, the ability torepeatedly treat portions of liver thathave never been irradiated may proveto be beneficial for the patient. In ourinstitution, given the propensity fornew tumors to develop in patientswith metastatic disease, we will per-form segmental infusions when possi-ble, sparing the normal parenchyma,thereby allowing for future infusion tononirradiated parenchyma. However,scenario B is uncommon in patientswith metastatic disease, because theyusually present with bilobar multifo-cal metastases.

The scenario of localized diseasewith elevated bilirubin (scenario D)(Table 2) is also quite uncommon inpatients with metastatic disease to theliver. Nevertheless, these patients maybe assessed for possible treatmentwith a segmental approach only. Inthis scenario, the increased bilirubinlevel implies compromised liver func-tion. In such a case, if other factorspermit (eg, PS), a patient might be acandidate for 90Y treatment with seg-mental techniques without injury tothe normal parenchyma. In the presentauthors’ experience, follow-up imag-ing in this clinical scenario usuallydemonstrates the development of liverlesions in previously unsuspected por-tions of the liver (ie, micrometastases).If this develops and the bilirubin levelcontinues to be increased (scenario C),treatment with 90Y is not advocated,given the inevitable radiation of someof the remaining functional paren-chyma and deterioration of liver func-tion. In such cases, embolic therapywith or without chemotherapy (ie,TACE/bland embolization) may beconsidered (as discussed in the secondpart of this three-part series).

THE TREATMENT PROCESS

The Interdisciplinary Team: ReferralPatterns and “Drivers”

The development and establish-ment of an interdisciplinary team iscrucial to the success of a radioem-bolization with 90Y brachytherapy

program. The team should be wellrepresented with members from inter-ventional radiology; medical, radia-tion, and surgical oncology; transplantsurgery; nuclear medicine; hepatol-ogy; medical physics; and radiationsafety. Patients may be referred fortherapy from various sources. In pa-tients with hepatoma, the initial diag-nosis may be made by the hepatolo-gist. Depending on the tumor size andlocation, patients may then be referredto surgical or transplant oncology de-partments for possible resection. If thepatient is not a surgical candidate, heor she may be seen by medical oncolo-gists for possible treatment under aclinical trial, systemic therapy, embol-ic-type therapy (eg, TACE), or 90Ytherapy. In general, patients with HCCare usually referred by a team com-posed of hepatology, transplant sur-gery, and HCC-focused medical on-cology personnel. For patients withmetastatic disease to the liver, the re-ferrals may come directly from themedical oncology or surgical oncologydepartments because the surgical op-tions may have been exhausted. Alter-natively, given the limited ability todeliver external-beam therapy to theliver, the radiation oncologist may re-fer the patient for 90Y treatment in anattempt to maximize the therapeuticeffect of intraarterial brachytherapy. Itis essential that interventional radiol-ogy play an integral role in all of theseteams.

The drivers of the 90Y process maycome from any of the aforementionedmedical subspecialties. Medical andsurgical oncologists, interventional ra-diologists, or radiation oncologistsmay oversee the process. Irrespectiveof the driver, the complexity of thetreatment planning process necessi-tates adequate representation from alldisciplines. Interventional radiologistsare experts in radiation physics andcontribute the technical and imagingexpertise, all of which is crucial forsuccessful therapy. Radiation oncolo-gists have expertise in brachytherapy,dosimetry, and radiation biology andtreatment, and a sound understandingof radiation effects on tissue. Surgicaland medical oncologists provide thetreatment backbone for all malignan-cies. Nuclear medicine specialists havewell-established expertise in handlingradiopharmaceuticals, as well as nu-clear and functional imaging. Radia-


tion safety specialists oversee theproper and safe use of 90Y under theregulations dictated by the NuclearRegulatory Commission or local stateboards.

Clinical and Diagnostic Evaluation

The first step in the evaluation ofpatients for therapy includes collect-ing a history and conducting a physi-cal examination. Patients should beable to tolerate treatment, as best as-sessed by Okuda, ECOG, and Karnof-sky score evaluation. Total bilirubinlevel and prothrombin time are impor-tant predictors of which patients willtolerate treatment. Relevant informa-tion to be elicited includes a history ofrenal or hepatic failure, as well as pul-monary compromise such as chronicobstructive pulmonary disease. Tu-mor markers such as �-fetoproteinshould be measured and may be usedin the assessment of treatment re-sponse. If patients are receiving che-motherapy, it is important to discon-tinue the treatments 2–3 weeks beforethe beginning of treatment with 90Y toclearly identify the agent responsiblefor any subsequent therapeutic re-sponse. More important, it is essentialto identify those patients receivingagents known to be radiation sensitiz-

ers, such as 5-fluorouracil, capecitab-ine, and gemcitabine. Radiation hepa-titis, a potentially fatal complication, isa theoretical concern for patients re-ceiving 90Y microspheres, particularlyif it is used concurrently with radia-tion sensitizers.

From a diagnostic imaging stand-point, careful review of recent CT orMR imaging (with 2�4 weeks) is war-ranted. Controversy exists about thediagnostic pathway of a patient withcirrhosis and a liver mass. In manyinstitutions, including the authors’, abiopsy for the diagnosis of hepatomais usually performed. Although by-passing this step might prevent bleed-ing, tract seeding, and false-negativeresults, these potential complicationshave not been observed clinically (36).For metastatic disease to the liver, bythe time the patient is referred for 90Ytherapy to the liver, a diagnostic bi-opsy of the liver has been performed.When the cross-sectional imaging mo-dality has been reviewed and the pa-tient has been deemed a candidate,liver volume calculations are obtained.

Calculation of Liver Volumes: TheLobar Approach

Triple-phase CT provides the fast-est and most reproducible imaging of

the liver for volume (TheraSphere)and tumor burden (SIR-Spheres) cal-culation. Because the treatment ap-proach for 90Y is most commonly lo-bar, proper imaging and volumecalculation is essential for dosimetrypurposes. The ability to understandhepatic anatomy relies on the soundunderstanding of the Couinaud he-patic segments (64). Anatomically, themiddle hepatic vein separates theright and left lobes. When regions ofinterest are drawn and lobar volumesare calculated, it is the middle hepaticvein that should be used as the ana-tomic delineator between the right andleft lobes. If the middle hepatic veincannot be seen, the gallbladder fossaand its axis relative to the liver may beused. This technique assumes stan-dard arterial anatomy with singleright and left hepatic arteries. If vari-ants are observed angiographically(eg, an accessory right hepatic artery),accurate angiographic correlationsmust be performed when the regionsof interest for lobar or segmental vol-umes are drawn. This will ensure thataccurate volumes are obtained andthree or more treatments are adminis-tered.

The most common angiographicfindings and variants, with their asso-ciated target segments (and hence re-quired volumes), are listed in Table 3(50,58). It is incumbent on the inter-ventional radiologist to have a soundknowledge of these anatomic variantsand their effects on dosimetry and mi-crosphere distribution. The followingcase illustrates typical anatomy andthe middle hepatic vein as the land-mark used for volume calculation. InFigure 1, Ethiodol (Savage Laborato-ries, Melville, NY) was injected intothe left hepatic artery as part of aTACE procedure. The venous systemis not visualized, because no intrave-nous contrast medium has been ad-ministered. The area that is high-lighted by the Ethiodol is the left lobeof the liver: segments 2, 3, 4A, and 4B,and part of segment 1. Regions of in-terest around the right and left lobes inthe same patient who received intra-venous contrast medium are depictedin Figures 2 and 3. This demonstratesconclusively the location of the middlehepatic vein and its role as the separa-tor between the right and left lobe byCT. Figures 4 and 5 represent coronalmaximum-intensity projection views

Table 3Couinaud Segments Based on Angiographic Findings

Angiographic Findings

CorrespondingTarget Segments

Right Left

Standard right and left hepatic arteries 1, 5, 6, 7, 8 2, 3, 4Replaced right hepatic with flow to medial segment

left lobe1, 4, 5, 6, 7, 8 2, 3

Replaced right hepatic artery without flow tomiddle lobe, left hepatic artery with flow tomedial and lateral segments left lobe

1, 5, 6, 7, 8 2, 3, 4

Replaced left hepatic artery without flow to mediallobe

2, 3

Replaced left hepatic artery with flow to mediallobe

2, 3, 4

Accessory right hepatic artery 6, 7Right hepatic artery in the presence of an accessory

right hepatic5, 8

Middle hepatic artery (irrespective of origin) 4

Note.—Vascular anatomy subject to variation. There is an assumption that caudatelobe (segment 1) derives blood supply from right hepatic artery. Among the caudatelobe (segment 1), right anterior lobe (segments 5/8), right posterior lobe (segments6/7), left medial lobe (segment 4), left lateral lobe (segments 2 and 3), depending ontumor location, these variants will guide the volume calculation process.


of the three-dimensional reconstruc-tion of the right and left lobes, witheach respective calculated volume.

The dosimetry for SIR-Spheres isbased on (i) body surface area andpercentage tumor burden or (ii) per-centage tumor burden alone. The lat-ter dosimetry model is based on tumorburden with respect to the entire liver.As an example, Figures 6 and 7 illus-trate the right (937.3 mL) and left(972.6 mL) lobe volumes for a patient

to be treated with SIR-Spheres. Figure8 demonstrates three-dimensional re-construction of tumor volume (311.2mL). If tumor burden model is used,SIR-Spheres dosimetry is then basedon a whole liver tumor burden of16.3% (311.2 / [937.3 � 972.6]).

In this example, the TheraSpheredose calculation would be based onthe lobar volume to be infused (937.3mL) and would be independent of tu-mor burden. For TheraSphere dosim-

etry, the following statement is alwaystrue and should be remembered whendosimetry calculations are performed:the volume that is entered in the for-mula is the volume of liver tissue thatis perfused by the vessel that will beinfused. In other words, it is the vol-ume of the liver segments being per-fused by the vessel of interest. For ex-ample, if the right lobe is to be infusedwith standard hepatic arterial anat-omy, the volume to be entered is that

Figures 1, 2. 1. Noncontrast CT of the liver in a patient with standard anatomy after TACE to the left lobe illustrates the middle hepaticvein as the anatomic separator between the right and left lobes. 2. Regions of interest drawn around the right lobe to calculate volume(segments 5, 6, 7, 8, and a portion of segment 1). The middle hepatic vein is used as the delineator between the right and left lobes.

Figures 3, 4. 3. Regions of interest drawn around the left lobe to calculate the volume (segments 2, 3, 4, and a portion of segment 1).In patients with standard hepatic arterial anatomy, the middle hepatic vein separates the right and left lobes. 4. Three-dimensional CTreconstruction following regions of interest around the right lobe determine the right lobe volume (1,398.4 mL). This volume is now usedto calculate the activity of TheraSphere required.


of segments 1, 5, 6, 7, and 8. If themiddle hepatic artery is the vessel tobe infused, the volume to be entered isthat of segment 4. If the accessory righthepatic artery is to be infused, seg-ments 6 and 7 represent the volume tobe entered. These concepts are furtherdescribed later in the article.

Dose Calculation, Ordering, andInfusion

TheraSphere.—As described in theproduct insert, TheraSphere consistsof insoluble glass microspheres inwhich 90Y is an integral constituentof the glass (65). The mean sphere di-ameter ranges from 20 to 30 �m.Each milligram contains 22,000–73,000microspheres. TheraSphere is suppliedin 0.05 mL of sterile, pyrogen-free wa-ter contained in a 0.3-mL v-bottom vialsecured within a 12-mm clear acrylicvial shield. TheraSpheres are dis-pensed weekly by the manufacturer(MDS Nordion) on Wednesdays, arecalibrated for 12:00 noon (EasternStandard Time) of the following Sun-day, and are available in the follow-

ing six activity sizes: 3 GBq (81 mCi),5 GBq (135 mCi), 7 GBq (189 mCi),10 GBq (270 mCi), 15 GBq (405 mCi),and 20 GBq (540 mCi) (1). The corre-sponding number of microspheresper vial is 1.2 million, 2 million, 2.8million, 4 million, 6 million, and 8million, respectively. The activity permicrosphere is approximately 2,500Bq (66).

The recommended activity ofTheraSphere that should be deliveredto a lobe of the liver-containing tu-mor is between 80 Gy and 150 Gy.This wide range exists to give thetreating physician clinical flexibility.Patients with significant cirrhosisshould be treated more conserva-tively (80–100 Gy), whereas patientswithout cirrhosis may be treatedmore aggressively (100–150 Gy). Inour institution, the most commonlyused dose range is 100–120 Gy, arange that balances safety and effi-cacy. If there is any uncertainty as tothe tolerability of treatment, the ap-proach of dose fractionation shouldbe considered, such as two doses of50–75 Gy rather than a single dose of

100–150 Gy. Dose fractionation is awell-accepted principle in radiationoncology and should be consideredas one of the treatment options forpatients undergoing treatment with90Y. The one caveat to repeat infusionis the principle of augmentation ofhepatopulmonary shunting after treat-ment with 90Y (67). In our institution,we have observed that relativeshunting may increase after treat-ment with TheraSphere or SIR-Spheres. Therefore, if treatment ac-cording to a fractionation model is tobe followed, repeat assessment oflung shunting should be undertakenbefore the second treatment to en-sure that the dose to the lung will re-main below the threshold of 50 Gycumulative absorbed dose (52). As-suming that TheraSphere 90Y micro-spheres distribute in a uniform mannerthroughout the liver and 90Y under-goes complete decay in situ, radioac-tivity required to deliver the desireddose to the liver can be calculated ac-cording to the following formula (17):

TheraSphere: A (GBq) � [D (Gy) �

M (kg)]/50, or

Activity required (GBq) �

[desired dose (Gy)] �

[target liver mass (kg)]/50

Given that a fraction of the micro-spheres will flow into the pulmonarycirculation without lodging in the ar-terioles, when lung shunt fraction(LSF) and vial residual is taken intoaccount, the actual dose delivered tothe target volume after the vial is in-fused becomes (17):

D (Gy) � [A (GBq) �

50 � (1 � LSF � R)]/M (kg), or

Dose (Gy) � 50 [measured pre-infusion

activity (GBq) � (1 � LSF) � (1 � R)]/

target liver mass (kg)

where A is activity infused to the tar-get liver, D is the absorbed delivereddose to the target liver mass, and M istarget liver mass. Liver volume (incm3) is estimated with CT and thenconverted to mass with a conversionfactor of 1.03 mg/mL.

Figure 5. Three-dimensional CT reconstruction following regions of interest around theleft lobe determine the left lobe volume (484.0 mL). This volume is now used to calculatethe activity of TheraSphere required.


For example, an authorized userwishes to treat a patient with Thera-Sphere with the following characteris-tics: standard hepatic arterial anat-omy, target vascular volume (ie, rightlobe, segments 1, 5, 6, 7, and 8) of 1,000mL (1.030 kg), desired dose of 120 Gy,LSF of 5%, and vial residual of 2%. Thetumor(s) is in the right lobe.

Activity required for treatment �

(120 � 1.030)/50 � 2.47 GBq

The patient receives an infusion of 2.47GBq to the target volume. Given 5%LSF and 2% residual in the vial, theactual delivered dose to the targetliver (right lobe) would be:

D (Gy) � [2.47 � 50 �

(1 � 0.05) � (1 � 0.02)]/1.030 � 112 Gy

The lung dose calculation (see sectionson Calculation of Lung Dose and Cal-culation of Residual Activity and Pro-portional Activity Delivered) is asfollows:

D (lung) �50 * infused activity

corrected for residual in vial * LSF

D (Gy) � 50 � (2.47 [1 � 0.02] �

0.05) � 6.1 Gy

Note that the dosimetry for Thera-Sphere is independent of tumor bur-den and is dependent on the mass ofthe targeted/infused liver tissue (inthis case, 1.030 kg). In this scenario, thedose received by the left lobe was 0 Gyand the dose received by the right lobe(segments 1, 5, 6, 7, and 8) was 112 Gy.Finally, because the LSF is a knownvariable, it is possible to use the afore-mentioned formulas to compensate forthis and to prescribe a higher overalldose that will correspond to 120 Gy tothe targeted liver tissue. In this case,prescribing 129 Gy, assuming thesame 5% LSF and 2% residual, wouldresult in the desired dose to the rightlobe of 120 Gy. The lung dose wouldbe 6.5 Gy.

In a second example, an authorizeduser wishes to treat a patient withTheraSphere with the following char-acteristics: variant anatomy (accessoryright hepatic artery), target vascularvolume (ie, posterior segment of the

right lobe, segments 6 and 7) of 650mL (0.670 kg), desired dose of 136 Gy,LSF of 8%, vial residual of 1%. Thetumor(s) is in the posterior segment ofthe right lobe.

Activity required for treatment �

(136 � 0.670)/50 � 1.82 GBq

The patient receives an infusion of 1.82GBq to the target volume. Given 8%LSF and 1% residual in the vial, theactual delivered dose to the targetliver (posterior segment right lobe,segments 6 and 7) is:

D (Gy) � [1.82 � 50 �

(1 � 0.08) � (1 � 0.01)]/0.670 � 124 Gy

The lung dose calculation (see sectionson Calculation of Lung Dose and Cal-culation of Residual Activity and Pro-portional Activity Delivered) is asfollows:

D (Gy) � 50 � (1.82 � [1 � 0.01]

� 0.08) � 7.2 Gy

Note again that the dosimetry forTheraSphere is independent of tumorburden and is dependent on the massof the targeted/infused liver tissue (inthis case 0.670 kg). In this scenario, thedose received by the anterior segmentof the right lobe and the left lobe was0 Gy, and the dose received by theposterior segment of the right lobe(segments 6 and 7) was 124 Gy. In thiscase, prescribing 149 Gy and assumingthe same 8% LSF and 1% residualwould result in the desired dose to theposterior segment of the right lobe of136 Gy. The lung dose would be 7.9Gy.

TheraSphere administration/physics.—Because TheraSphere is not shipped asa specific patient unit dose, the appro-priate vial must be ordered and the ac-tivity decayed to the required treat-ment activity. The usable shelf life of aTheraSphere dose is 7 days from thedose calibration date. Doses are sched-uled to arrive on the day before thescheduled treatment date. If an institu-tion plans to perform more than fouradministrations per week, a licensepossession limit of 240 GBq is recom-mended. This possession limit allowsfor the storage of the highest dose or-dered and the residual activity from

Figure 6. Right lobe volume in a patient being treated with SIR-Spheres.


administrations. A clinical treatmentdose range of 80–150 Gy also allowsfor a 12-hour flexible schedule formost patients. Most patients can betreated earlier in the week with thelower-activity vials (3, 5, or 7 GBq) orlater in the week with higher-activityvials (7, 10, 15, or 20 GBq).

Before administration, the vial ac-tivity can be verified with two differ-ent techniques. The dose can be mea-sured in a standard dose calibrator.The dose calibrator must be calibratedfor the TheraSphere vial configuration,which includes the acrylic shield. As-saying the first dose and adjusting thereading until it is equal to the manu-facture’s decayed activity is the stan-dard method of obtaining the calibra-tion correction factor. The othermethod confirms the dose vial activityby measuring the dose rate with a por-table ionization chamber (ie, a “cutiepie”) that is placed at a calibrated dis-tance in such a manner that 1 mrem/h

is equal to 1 GBq. The center of theportable ionization chamber is posi-tioned approximately 30 cm from thecenter of the acrylic-shielded vial.Again, the reading is only accurate atbest to within 10%. If the facility hasmultiple dose vials at hand, it is cru-cial to use both methods just beforeinfusion to avoid a potential misad-ministration or medical event causedby confusion with dose vials. Becausehuman error is a possibility, an indi-vidual other than the one using theionization chamber for dose verifica-tion should measure the dose in thedose calibrator.

The TheraSphere administration setconsists of one inlet set, one outlet set,one empty vial, and two interlockingunits consisting of a positioning nee-dle guide, a priming needle guide, allcontained behind a Lucite shield. AMonarch 30-mL syringe (Merit Medi-cal, South Jordan, UT) is used to infusesaline solution through the system.

The saline solution containing theTheraSphere 90Y microspheres is in-fused through a catheter placed in thehepatic vasculature. When the catheteris positioned at the treatment site andthe authorized user verifies the integ-rity of the delivery system, the cathe-ter is connected to the outlet tubing.Delivery of TheraSphere is accom-plished by pressurizing the Monarchsyringe. For 3-F catheter systems, theinfusion pressure should range from20 to 40 psi, whereas for 5-F systems,the infusion pressure usually shouldnot exceed 20 psi (use of 5-F systemsfor infusion is discussed later). It isessential that, irrespective of the infu-sion pressure used, the flow of micro-spheres must closely mimic that ob-served angiographically. This can beaccomplished with a gentle hand in-jection of contrast medium followedby a column of saline solution. At thispoint, it is important that the autho-rized user become familiar with theactual flow dynamics of the vessel be-ing infused and use a correspondinglylowered TheraSphere infusion pres-sure where necessary, such as mightbe seen in patients with decreased car-diac output. Given the small volumeof microspheres contained in a givendose of TheraSphere (typically 27–90mg), the volume of saline solution re-quired to infuse a vial of TheraSphereis low; the majority of the micro-spheres have been infused after thefirst 20 mL of saline solution. In addi-tion, given the low number of micro-spheres infused with TheraSphere(typically 1.2–4.0 million), the entirevascular bed is never completely satu-rated. Hence, although the infusionmust be performed in the interventionalradiology catheterization laboratory,continuous and live fluoroscopic guid-ance while the infusion is occurring isnot necessary. A complete infusion usu-ally requires 20–60 mL and usually re-quires 5 minutes to complete.

Pros and cons of TheraSphere dosime-try and physical characteristics.—Thereare several advantages to the Thera-Sphere empiric dosimetry model andthe physical characteristics of the de-livery system. First, the microspheresare shipped in a v-bottom vial thatcontains a fixed volume of sterile wa-ter. The v-bottom vial containing themicrospheres is sealed within anacrylic shield with a removable capfor septum access. This configuration

Figure 7. Left lobe volume in a patient being treated with SIR-Spheres.


ensures that most of the micro-spheres are confined to a fixed vol-ume within the sealed vial. Becausethe activity of the microspheres is de-termined with use of a dose calibra-tor that measures the Bremsstrahlungradiation produced by the �-interac-tion with the shielding, the accuracyof the measurement depends on thevolume of the microspheres. For theTheraSphere dose vials, the shieldingis fixed and the volume of the sus-pended microspheres is limited toless than 0.5 mL. Both these condi-tions ensure that the accuracy indose calibrator calibration and thedose assay will be reproducible fromone facility to another. Moreover,this configuration requires no physi-cal manipulation, such as decanting,

thereby reducing the potential of adosing error caused by staff transfer-ring the dose from one vial to an-other. The fixed shielding and lack ofmanipulation of the microspheresalso reduces the radiation exposureto staff who assay the dose vials.

Second, because the target radia-tion absorbed dose is dependent onthe target vascular volume, which iscalculated with three-dimensional soft-ware, dosimetry calculations among dif-ferent users for any given volume arereproducible. That is, this dosimetry isindependent of tumor burden, furthersimplifying the determination of therequired activity. The relatively lownumber of microspheres results inmore than 95% delivery in nearly allcases without reaching stasis or com-

plete vessel occlusion, allowing theadministration of extremely high radi-ation dose or radioactivity (2,500 Bqper sphere) with (relatively) few mi-crospheres. Because embolic effectsare limited to only the smallest capil-laries within the target tissue, high ox-ygenation is maintained and radiationdose responses are improved (50,68).Moreover, this lack of significant em-bolic effect suggests potentially safeapplications in patients with compro-mised portal venous flow or portalvein thrombosis, and also correlateswith low clinical toxicities comparedwith embolic therapies (15,69).

The shelf life of a TheraSphere dosevial is 7 days, further adding to theclinical flexibility of patient treatmentand scheduling. Given this shelf life,the same activity may be administeredwith different numbers of micro-spheres infused. Therefore, this devicemay be used and altered with variousdegrees of desired embolic effect (seeprevious discussion of decay curve).Finally, the manufacturer has appro-priately recommended that, in cases ofvery high tumor burden (�70%), treat-ment should likely be withheld(70,71). This guidance is consistentwith other types of liver-directedtreatments.

There are also several disadvan-tages of TheraSphere glass micro-spheres. The dosimetry model lacks avariable for tumor burden. By consid-ering only the target volume, one ig-nores the fact that more activity (andhence microspheres) may be requiredto treat greater tumor burden (thislimitation is overcome with strategiesdescribed later in this article). More-over, the high specific activity per mi-crosphere may result in radiation dos-ages that prevent dose fractionation ifmaximum hepatic radiation toleranceis reached. The specific gravity ofTheraSphere microspheres is highcompared with SIR-Spheres and maytheoretically limit microsphere distri-bution. TheraSphere doses have a rela-tively low number of microspherescompared with SIR-Spheres that mayresult in inadequate tumor coveragefor large tumors. Also, although this isclinically inconsequential, the manu-facturing process may result in long-lived radioactive contaminants such aseuropium 152Eu and 154Eu. These con-taminants are found in other reactor-produced products used in medicine,

Figure 8. Three-dimensional reconstruction outlining volume of tumor burden.


such as strontium Sr 89, used for bonepain, or 90Y antiB1, used for lym-phoma. However, the glass micro-spheres are placed directly in the reac-tor to activate the 90Y, which increasesthe amount of contaminant to lessthan 30 uCi per dose vial. Finally,there is room for improvement in theadministration set.

SIR-Spheres.—As described in theproduct insert, SIR-Spheres consist ofbiocompatible resin-based micro-spheres containing 90Y with a sizebetween 20 �m and 60 �m in diame-ter. SIR-Spheres constitute a perma-nent implant and are provided in avial with water for injection. Eachvial contains 3 GBq of 90Y (at thetime of calibration) in a total of 5 mLwater for injection. Each vial contains40–80 million microspheres (3). Con-sequently, the activity per micro-sphere for SIR-Spheres is much lowerthan that of TheraSphere (50 Bq vs2,500 Bq) (66). SIR-Spheres are dis-pensed three times per week by themanufacturer (Sirtex) and are cali-brated for 6:00 PM (Eastern StandardTime) on the date of treatment. Theshelf-life is 24 hours after the calibra-tion date and time.

Just as with TheraSphere, assumingSIR-Spheres 90Y microspheres distrib-ute in a uniform manner throughoutthe liver and undergo complete decayin situ, radioactivity delivered to theliver can be calculated by one of twoavailable methods:

The first method incorporates bodysurface area and estimate of tumorburden, as follows (3,72):

SIR-Spheres: A (GBq) �

body surface area (m2)

� 0.2 � (% tumor involvement/100)

The second method is based on abroad estimate of tumor burden as de-scribed in Table 4. The larger the tu-mor burden, the higher the recom-mended activity in increments of 0.5GBq per 25% tumor burden. For eitherSIR-Spheres dosimetry model, A (inGBq) is decreased depending on theextent of LSF (�10% LSF, no reduc-tion; 10%–15% LSF, 20% reduction;15%–20% LSF, 40% reduction; �20%LSF, no treatment).

As an example, an authorized userwishes to treat a patient with the fol-

lowing characteristics: total weight of91 kg, height of 1.83 m (6 feet), livervolume of 1,000 mL, tumor volume of300 mL, LSF of 5%.

According to the first method, theformula for BSA as described byDubois and Dubois (73,74) is asfollows:

Body surface area � 0.20247

� height0.725 (m) � weight0.425 (kg)

Therefore:

A (GBq) � 2.13 � 0.2

� (30/100) � 2.23 GBq

would be required according to thebody surface area formula.

Alternatively, given the tumor bur-den of 25%–50%, the patient could beprescribed 2.5 GBq in activity, on thebasis of the calculations described (3).Given the 5% LSF, no reduction in ac-tivity would be required. It should benoted that, for SIR-Spheres, the dosim-etry described in the product insert isbased on whole liver infusion. If a lo-bar infusion is intended, the infusedactivity should be calculated assumingwhole liver volume and then “cor-rected” to the proportional volume ofthe target lobe that is to be infused. Forexample, if the right lobe is the targetand represents 70% of the entire livervolume, the calculated activity (in thiscase 2.23 GBq) to be delivered shouldbe multiplied by 0.7. Correcting SIR-Spheres dosimetry that is based onwhole liver infusion and applying it tolobar therapy requires an understand-ing of percentage mass of the right andleft lobes as well as the differing tumorburdens in each lobe.

The lung dose calculation assuming2.23 GBq activity infused with no re-sidual (see sections on Calculation ofLung Dose and Calculation of Resid-

ual Activity and Proportional ActivityDelivered) is as follows:

D (Gy) � 50 � (2.23 � 0.05) � 6 Gy

Note that SIR-Spheres body surfacearea dosimetry is based on tumor bur-den and liver mass (larger liver masswith higher body surface area). Em-piric SIR-Spheres dosimetry is depen-dent only on tumor burden and is in-dependent of liver mass.

SIR-Spheres administration and phys-ics.—Before administration, the SIR-Spheres dose is assayed in a dosecalibrator but should be verified witha different technique. The dose cali-brator must be calibrated for the vialconfiguration and the volume of thedosing vial as well as the shippingvial. The standard method for obtain-ing the calibration factors for 90Y re-quires obtaining a National Instituteof Standards and Technology–cali-brated source in the smallest volumeof sterile water that will be used forpatient dosing. This vial will be as-sayed and the reading adjusted untilit agrees within 10% of the decayedactivity. To obtain the calibration fac-tor for greater volumes, simply addthe desired amount of sterile waterand repeat the measurements. Be-cause this technique is based on theBremsstrahlung radiation, it is impor-tant to use a vial for the calibrationsource that is constructed of thesame material used for the shippingvial and the v-bottom vial. It is alsoimportant to use a vial of the sameshape, such as a v-bottom vial or aflat surface bottom vial. The manu-facturer recommends that the activityof the dose vial be determined fromthe difference between the assays ofthe shipping vial before and after de-canting. The dose vial activity shouldbe confirmed by assaying the acrylic-shielded dose vial in the dose cali-brator with the calibration factors fora shielded dose. This reading shouldbe within 10% of the reading ob-tained according to the manufactur-er’s method. Readings that differ bymore than 10% indicate that a signifi-cant amount of the SIR-Spheres maybe in the transfer needle. The doseshould also be verified by measuringthe dose rate with a portable ioniza-tion chamber that is placed at a cali-brated distance in such a mannerthat 1 mrem/h is equal to 1 GBq.

Table 4SIR-Spheres Dosimetry Based onTumor Burden

Tumor Involvementin the Liver (%)

RecommendedActivity (GBq)

�50 3.025–50 2.5�25 2.0


The center of the portable ionizationchamber is positioned approximately30 cm from the center of the acrylic-shielded vial. Again, the reading isonly accurate at best to within 10%.If the facility has multiple dose vialsat hand, it is crucial to use bothmethods just before infusion to avoida potential misadministration or med-ical event caused by confusion withdose vials. Because human error ispossible, an individual other than theone using the ionization chamber forthe dose verification should measurethe dose in the dose calibrator. Be-cause the treatment vial does nothave a label, a label describing thecontents, radioactive sign, activity,and procedure or patient identifica-tion should be placed on the shieldor vial. This label must not obscurethe view of the contents.

The SIR-Spheres administration setconsists of a Perspex shield, the dosevial, and inlet and outlet tubing withneedles. Standard 10-mL or 20-mL in-jection syringes preloaded with sterilewater are required to infuse the micro-spheres into the delivery catheter.Pressure gauges are not availablewhen SIR-Spheres are infused. Usu-ally, slow and deliberate hand-injec-tion of the SIR-Spheres through the 3-For 5-F catheter systems is adequate(use of 5-F systems for infusion is dis-cussed later). Care should be takenwhen performing the infusion becausethe infusion setup is composed of sev-eral connection tubings in series,thereby increasing the pressure re-quired to infuse the microspheres.Care should be taken not to allow toovigorous an injection rate, because thismay result in leaks at points of poten-tial weakness (eg, septum, tubing con-nections). Just as with TheraSphere,the infusion rate used must closelymimic the blood flow rate observedangiographically. The blood flow dy-namics should dictate the pace atwhich infusion occurs. If the autho-rized user is not an interventional ra-diologist, the proper communicationof this information to the authorizeduser becomes crucial. Depending onthe activity infused, an amount of20–40 mL is usually sufficient for in-fusion of the intended dose.

Pros and cons of SIR-Spheres dosime-try and physical characteristics.—Thereare advantages to the SIR-Sphere do-simetry model and the characteristics

of the resin microspheres. The depen-dence of activity required for treat-ment on tumor burden is inherentlylogical. The activity vial may be ma-nipulated and the specific activitydecanted in the nuclear medicinepharmacy tailored for the patient.The infusion may be done with alter-nating injections of sterile water andcontrast medium, thereby allowingmonitoring specifically with the useof fluoroscopy to ensure that stasis isnot reached. The lower specific grav-ity theoretically makes for better sus-pension. The lower specific activityper microsphere (50 Bq per micro-sphere) allows for the possibility ofdose fractionation. Dose fractionationis a well-accepted principle in radia-tion oncology (75). Finally, the ad-ministration kit is simple.

There are limitations to the dosim-etry and physical characteristics ofresin microspheres. The shelf-life ofthe device is 24 hours, restrictingclinical flexibility and patient sched-uling. The need to physically decantalthough, on the surface, potentiallyadvantageous introduces further hu-man technical manipulation that mayresult in spills requiring radiationcontainment or, more critically, resultin the wrong activity administered.There are several formulas for dosecalculation, potentially resulting inseveral different prescribed doses forthe same patient. This lack of unifor-mity makes interuser variability veryhigh. If one were to use the tumorburden formula described earlier inthis article, a patient with 1% tumorburden would receive the same pre-scribed activity as someone with 24%tumor burden. A patient with 51%tumor burden would receive thesame activity as someone with 99%tumor burden. Unlike the dosimetryfor glass microspheres, this dosime-try model does not account for pa-tients with the highest tumor bur-dens. Withholding treatment frompatients with extremely high tumorburdens should be considered (70).In addition, vessel stasis and prema-ture termination of the infusion hasbeen described in 35% of cases,bringing into question the reproduc-ibility of dosimetry and the ability toadminister sufficient radiation (29).Embolic effect may cause hypoxia,potentially limiting the effect of theradiation (76,77). Also, given the

large number of microspheres, it isdifficult to determine whether anytherapeutic effect is a result of radia-tion or embolic effects.

Comparison between TheraSphere andSIR-Spheres.—TheraSphere and SIR-Spheres are distinctly different prod-ucts. Although both are 90Y embolicbrachytherapy devices (ie, radioem-bolization devices), the differences farexceed the similarities. TheraSphere isa minimally embolic brachytherapydevice consisting of 20- to 30-�mparticles with specific activity of2,500 Bq, higher specific gravity, andlower number of microspheres (ap-proximately 2–4 million microspheresper treatment). The infusion can pro-ceed without concern for vascularstasis, given the lower embolic load(68). Because of these characteristics,the median percentage activity in-fused per vial is well above 95%.

This is in direct contradistinctionwith SIR-Spheres, which are a mod-erately embolic brachytherapy deviceconsisting of 20- to 60-�m particleswith specific activity of 50 Bq, lowerspecific gravity, and higher numberof microspheres (approximately 40–80million microspheres per treatment)(Table 5). Given these characteristics, itis not unusual for the vascular bed tobecome saturated and angiographicstasis to be reached (29). Hence, themedian percentage of microspheres in-fused is usually lower in comparisonwith TheraSphere.

Misconceptions about TheraSphereand SIR-Spheres.—With both micro-sphere devices, fluoroscopic guidanceis used, no blind infusions are per-formed, individually determined pa-tient-specific doses are used, and pa-tients are treated in a lobar fashion.The idea that resin spheres suspendbetter than glass, resulting in bettertumor coverage, has never been vali-dated. In fact, a previous report hasdescribed the opposite, in which nodifference in tumor distribution wasfound between glass and resin micro-spheres (66). Gravitational flow tomore dependent areas has been doc-umented (78,79). Also, studies havecorrelated distribution of micro-spheres with blood flow (80). Hence,flow in a dependent region is likelythe result of normal physiologic flowand gravity, not the specific gravityof the microsphere infused. It shouldtherefore be assumed that the distri-


butions of glass and resin micro-spheres are comparable until this isproved otherwise.

The other major misconception isthe idea little if any dosimetry workhas been completed on either agent.This is misleading, given that manyauthors have performed validation ofthe dosimetry models (37,50–63). Al-though dosimetry remains a work inprogress for both 90Y devices, suffi-cient data are available to allowwidespread use of 90Y with clinicalbenefit. It should be noted that, al-though the empiric nature of the do-simetry for 90Y microspheres is a faircriticism, empiric models are thestandard and are routinely used inmedicine, such as dosimetry forTACE, 90Y-labeled antibodies forlymphoma, and systemic chemother-apy. Empiric dosimetry is the stan-dard in medical practice rather thanthe exception.

Calculation of Lung Dose

Radiation pneumonitis is a theoret-ical concern with 90Y treatment. Previ-ous preclinical and clinical studieswith 90Y microspheres demonstratedthat as much as 30 Gy to the lungscould be tolerated with a single injec-tion, and as much as 50 Gy could betolerated for multiple injections (55).For this reason, patients with 99mTc-MAA evidence of potential pulmo-

nary shunting resulting in lung dosesgreater than 50 Gy should not betreated.

The absorbed lung radiation dose isthe total cumulative dose of all treat-ments (81):

Cumulative absorbed lung radiation

dose � 50 � lung mass �i � 1

n

� Ai * LSFi

where Ai represents activity infused(correcting for residual in vial), LSFirepresents LSF during infusion, n rep-resents number of infusions, and ap-proximate lung mass (for both lungs,including blood) is 1 kg (82).

This dose should not exceed thelimit of 30 Gy per single infusion and50 Gy cumulatively. In patients whorequire more than two treatments toachieve tumor coverage or in patientsbeing treated repeatedly in the sametarget volume after progression, re-peat 99mTc-MAA LSF assessment maybe necessary (67).

Variations in Activity in Relation toNumber of Microspheres

TheraSphere.—As described previ-ously, TheraSphere doses are dis-pensed weekly by the manufactureron Wednesdays and are calibratedfor Sunday, 4 days later, at 12:00noon Eastern Standard Time. All mi-crospheres carry the same specific ac-

tivity. The only difference betweenthe various activities at the momentthe different vials are dispensed isthe number of microspheres per vial.The numbers of microspheres pervial are 3 GBq (1.2 million), 5 GBq (2million), 7 GBq (2.8 million), 10 GBq(4 million), 15 GBq (6 million), and20 GBq (8 million). Knowledge of ra-diation physics, activities, number ofmicrospheres per vial, and requiredvolume for microspheres is impor-tant in the ordering and treatmentplanning process. The fact that themicrosphere number alone differenti-ates the various activities dispensedby the manufacturer must be takeninto consideration when dosimetrycalculations are performed. Figures 9and 10 illustrate the dose decaycurves of the manufactured activityvials to a certain activity.

An example best illustrates thisprinciple. An authorized user wishesto treat a patient with a 971-mL rightlobe volume with a dose of 120 Gy.Assuming zero lung shunt, thiswould imply the need for the follow-ing of activity to treat the patient:

(971 mL * 1.03 g/mL * 0.001 kg/gm

* 120 Gy)/50 � 2.4 GBq

This would translate into 2.4 GBq ofactivity generated by a 3-GBq vialMonday at 08:40 AM, a 5-GBq vialWednesday at 07:58 AM, a 7-GBq vial

Table 5Comparison of TheraSphere and SIR-Spheres

Characteristic TheraSphere SIR-Spheres

Isotope 90Y 90YHalf-life (h) 64.2 64.2Time to near-complete decay

(3% residual activity), days13 13

Particle size (�m) 20–30 20–60Range of spheres per vial 1.2–8.0 million 40–80 millionActivity per sphere (Bq) 2,500 50Specific gravity High LowActivities available (GBq) 3, 5, 7, 10, 15, 20 3Requires handling for dispensing No YesModern delivery route Transcatheter, intraarterial (hepatic) Transcatheter, intraarterial (hepatic),

hepatic ports (rare)Embolic effect Mild ModerateIndication for use Hepatocellular carcinoma with appropriately

positioned catheterColorectal metastases with intrahepatic

floxuridineSpecial radiation precautions

upon discharge*None Possible urine contamination

* Refer to package insert and to institutional, state, and federal regulations for radiation safety considerations.


Thursday at 15:08 PM, or a 10-GBq vialSaturday morning 12:10 AM (Fig 9).Hence, depending on the dose se-lected and patient treatment schedul-ing, different numbers of micro-spheres would be administered. Iftreatment is scheduled for Monday, a3-GBq vial decayed to 2.4 GBq Mon-day at 08:40 AM will be required, andthe patient will receive 1.2 million mi-crospheres. If the same patient re-ceives a 5-GBq vial decayed to 2.4 GBqon Wednesday at 07:58 AM, 2 millionmicrospheres will be administered. Atreatment with a 7-GBq vial decayedto 2.4 GBq will be performed onThursday at 3:08 PM, translating into2.8 million microspheres. The use of a10-GBq vial is unlikely in this exam-ple, because this would require a Sat-urday 12:10 AM injection time. Anunderstanding of this principle is ex-tremely important when dosimetrywith TheraSphere is considered, be-cause the authorized user may want totake into account tumor size and vas-cularity. For example, a tumor withdensely packed vascularity (eg, HCCor neuroendocrine tumor) may bestbe treated with a higher number ofspheres than one without such densevasculature. In addition, large bulkylesions may also be better treated withhigher-activity vials later in the week,translating into a higher number ofmicrospheres administered and en-hanced tumor coverage. Therefore, itis encouraged that this flexibility inmicrosphere number for any given ac-

tivity be recognized and factored intreatment planning.

In summary, it is important that au-thorized users recognize the signifi-cant difference in number of micro-spheres depending on the vialmanufactured. Larger, more vascularlesions may be treated later during theweek to enhance coverage, whereassmaller tumors may be treated withlower-activity vials early in the week.Figure 10 demonstrates another exam-ple of this concept of modifying thenumber of TheraSphere particles in-jected for a given activity.

SIR-Spheres.—Variation in micro-sphere number also exists with SIR-Spheres. SIR-Spheres are always de-livered in 3-GBq activity vials. Ac-cording to the package insert (3), a3-GBq vial might contain between 40and 80 million microspheres. Giventhis variation, a patient receiving theentire 3-GBq vial might receive 40million microspheres, whereas thefollowing week, an identical patientmight receive 80 million micro-spheres with less specific activity.Given this variation and the inabilityto reliably estimate the number ofmicrospheres being infused, the ne-cessity for slow and deliberate infu-sion of the microspheres with con-tinuous fluoroscopic observation isfurther reinforced. In addition, giventhe need for constant fluoroscopic ob-servation during the infusion, the as-sessment for an angiographic end-point (ie, stasis), the infusion of 90Y

(especially SIR-Spheres) represents atrue embolization procedure.

The Treatment Procedure

Room preparation.—Standard roompreparations for interventional proce-dures are used. A large 6-foot �6-foot drape should be placed on thefloor near the fluoroscopy unit insuch a manner that when the 90Y de-livery system is brought in proximityto the patient, any potential contami-nation or leak on the floor will becontained. For the administration of90Y microspheres, at least two differ-ent radiation detectors should beavailable in the room. A surveymeter with a thin window Geiger-Muller detector capable of detectingradiation levels less than 0.1 mR/hshould be available near the exitfrom the room. This meter will beused to detect possible radioactivecontamination of the staff, in thewaste, and on the fixed equipment inthe room. A portable ionizationchamber capable of detecting radia-tion doses as low as 1 mrem/hshould be available. This surveymeter is used to measure the radia-tion doses emanating from theBremsstrahlung within the patient. Asurface measurement with the ioniza-tion chamber will give a crude esti-mate of the dose delivery site. Theionization chamber is also used tomeasure residual activity that may bein the dose vial. To locate TheraSphere

Figures 9, 10. 9. Decay curve of 3-, 5-, 7-, and 10-GBq vials of TheraSphere. The day of the week and time at which each vial hasdecayed to 2.4 GBq is illustrated. 10. Decay curve of 7-, 10-, 15-, and 20-GBq vials of TheraSphere. The day of the week and time at whicheach vial has decayed to 5.6 GBq is illustrated.


microspheres along the catheter andtubing connected to the dose vial, ahigh count rate ionization chamberwith a thin-window �-detector isvery useful (eg, Eberline R07). Be-cause the TheraSphere delivery sys-tem has two solid-state radiation de-tectors to determine radiation levelsat the dose vial and the catheter con-nection point, the high count rateionization chamber is not necessaryto deliver 90% of the dose.

To ensure delivery of the SIR-Sphere,a visual indicator should be confirmedwith a radiation level measurement ofthe dose vial. Use of the exposure rateat a fixed distance from the devicebefore and after infusion provides apositive determination of the dose de-livered in addition to surface measure-ments of the patient.

In addition to the supplied TheraS-phere and SIR-Sphere administrationsets, a Nalgene container and anacrylic Nalgene shield are required.The dose vial, attached tubing, and thecatheter are placed in the Nalgene con-tainer immediately after infusion ofthe spheres. Residual activity is deter-mined with use of the ionizationchamber and the shielded Nalgene inthe same geometry as the ionizationchamber and dose vial before admin-istration.

Patient preparation and catheteriza-tion.—Access to the vascular systemis achieved in the standard fashion.Because planning visceral angiogra-phy and embolization of extrahepaticvessels has already been accom-plished, the choice of catheters shouldbe kept consistent and readily avail-able in the procedural report. Cathe-terization of the celiac trunk with 4-For 5-F systems is recommended (hy-drophilic catheters may be used).Subsequently, a 3-F coaxial system(�0.0325 inches) is recommended forinfusion rather than 4-F or 5-F cathe-ters. This will permit administrationof 90Y at low and consistent pres-sures. In addition, consistency inpressure and flow will maintain thespheres in suspension, allowing un-impeded passage through the cathe-ter to the intended target volume.The use of 0.018-inch systems is notencouraged for routine 90Y injectionbecause there may be excessive out-flow resistance to injection, prevent-ing adequate flow rates and particlesuspension. Exceptions may be made

if the vascular anatomy does not per-mit the placement of catheter sys-tems larger than 0.018 inches. If0.018-inch systems are used, higherpressures may be required to achieveproper microsphere suspension. Cau-tion should be exercised if such pres-sures are used, because points ofweakness (eg, septum, catheter con-nection point) may result in micro-sphere leakage.

Infusion of 90Y microspheresthrough 4-F or 5-F catheters is notrecommended. The rationale for thisis twofold: (i) larger catheters ad-vanced deep in the hepatic vascularbed, although technically feasible forthe experienced angiographer, maycause vessel spasm, limiting the abil-ity to infuse microspheres that relyon flow dynamics for optimal deliv-ery; and (ii) the resistance to injectionof the microspheres through 4-F or5-F catheters may be too low, therebyincreasing the risk of reflux of micro-spheres to nontarget areas, particu-larly in the case of an overly vigor-ous injection. By virtue of theirsmaller overall caliber, luminal size,and length, use of 3-F catheters isless likely to result in vessel spasmand microsphere reflux.

Administration set preparation (Thera-Sphere, SIR-Sphere).—It is crucial toassemble the microsphere deliverykit carefully because any error dur-ing the setup may lead to incorrectadministration. Significant errorsduring critical steps in the setup anddelivery of microspheres should beaddressed at dose termination points.To increase consistency and limit er-rors, three individual team membersshould check all steps. For Thera-Sphere infusions, it is very importantto ensure that the connections to thesaline solution bag, syringe, and inletand outlet catheters are tight to en-sure that the system is sealed and aconstant pressure can be maintained.For SIR-Spheres infusions, it is veryimportant that needles are within theseptum and that the septum has notbeen damaged during setup and ma-nipulation (eg, excessive punctures).The microsphere administration kitsfor TheraSphere and SIR-Spheresshould be assembled according toobjective, reproducible, and repeat-able checklists. The manufacturersroutinely provide these checklists. Ifany of the steps in the assembly pro-

cess fails, consideration should bemade of terminating the procedurewithout attempting to infuse the mi-crospheres.

Microsphere infusion technique.—The technical aspects of radioemboli-zation are quite complex and shouldnot be undertaken lightly. The infu-sion technique varies depending onthe 90Y agent being used, the sizeand location of the catheter, and thevessel undergoing infusion. In addi-tion, the deliveries of TheraSphereand SIR-Spheres are distinctly differ-ent. Given this, it is recommendedthat institutions beginning a 90Y pro-gram commit to having the same au-thorized user and/or interventionalradiologist perform the first 10–15cases. This allows for the learningcurve to be reached quickly, as wellas allowing individual institutionalinefficiencies to be identified andremedied rather than repeated withdifferent authorized users.

Unlike SIR-Spheres, the emboliceffects of 20- to 30-�m TheraSphere90Y particles are angiographically mini-mal (68). As previously publishedand further reinforced in this article,the presence of unrecognized collat-eral vessels with consequent infusionof radioactive microspheres is certainto result in clinical toxicities if properangiographic techniques are notadopted (1,3,29,50,57,58,83). Thesemight include gastrointestinal ulcer-ation, pancreatitis, skin irritation, andother nontarget radiation. For thisreason, aggressive prophylactic em-bolization of vessels before therapy,in such a manner that all hepaticoen-teric arterial communications arecompletely eliminated, is highly rec-ommended. These vessels include theGDA, right gastric artery, esophagealartery, accessory phrenic artery, falci-form artery, and variant arteries suchas the supra- or retroduodenal artery.At our institution, where over 900 ra-dioembolization procedures havebeen performed, we have found ourgastrointestinal toxicity rate to bewell below 1%. This is because of ourstandard practices of (i) aggressiveprophylactic embolization of GDA/right gastric artery and other variantvessels; (ii) use of minimally embolicTheraSphere in a lobar and segmen-tal fashion; (iii) use of SIR-Spheres ina lobar, segmental, and dose-fraction-ated method (ie, several small doses


rather than one larger dose) withoutreaching a completely static and em-bolic state; and (iv) routine prophy-lactic use of a 2-week course of anti-ulcer medications.

TheraSphere infusion technique.—Most infusions should be performedwith use of 3-F systems. This in-cludes standard lobar infusions aswell as cases with difficult anatomy,small vessels (eg, left hepatic artery),and subselective catheterizations. Inthese instances, the required pres-sures are greater, usually 20–40 psi.The delivery of TheraSphere is de-pendent on blood flow through thehepatic vasculature distal to the cath-eter tip. Therefore, it is necessary tomake certain that the catheter doesnot occlude the vessel in which it isplaced, because doing so will resultin vessel spasm and reflux as the in-fusion proceeds. Flushing should becontinued until optimal delivery ofTheraSphere is achieved. A mini-mum flush of 60 mL is recom-mended with 3-F systems.

Although discouraged, the use of4-F or 5-F systems should allow forconstant low-pressure (10–20 psi) in-fusion throughout the administration.The radioactivity is monitored as thespheres travel through the blue stop-cock. At times, it may be necessaryto slightly “tap” or “vibrate” the inletand outlet sides of the blue stopcockwhile the administration is underway. It is possible for a very smallpercentage of the microspheres to re-main lodged in the potential spacesthat exist at catheter connection sites(eg, in three-way connections), neces-sitating a tapping process that willresuspend the spheres as the injec-tion proceeds. Flushing should becontinued until optimal delivery ofTheraSphere is achieved. Just as with3-F systems, a minimum flush of 60mL is recommended with 4-F or 5-Fcatheter systems.

Irrespective of the size of the cath-eter system, it is necessary to ensurethat the rate of infusion mimics therate of hepatic arterial flow. This rateis assessed by visual inspection of atest dose of contrast material infusedbefore TheraSphere administration.In some patients (eg, elderly patientswith diminished cardiac function orceliac stenoses), slower hepatic arte-rial flow may be apparent. In such

cases, the infusion rate of Thera-Sphere must mimic this slower flow.

The process of placing a 0.035-inch(4-F or 5-F) catheter system into theartery designated for TheraSphere in-fusion is not always easy. Patientswho have recently received chemo-therapy have vessels prone to dissec-tion and spasm (50). In addition, asstated previously, diminished cardiacoutput in elderly patients may resultin slower than expected hepatic arte-rial flow. Although larger cathetersystems with low resistance may re-sult in easier infusion of Thera-Sphere, the combination with dimin-ished hepatic arterial flow may resultin the reflux of microspheres. There-fore, in cases in which a large 4-F or5-F catheter is placed into the hepaticartery and reflux is of concern, place-ment of a 3-F catheter coaxiallywithin the base catheter is recom-mended. This will create higher resis-tance to infusion and will minimizethe risk of reflux. A typical Thera-Sphere infusion requires less than 5minutes to complete.

Radiation monitoring of theTheraSphere administration set andcatheter must be used to establish thetiming of optimal delivery. It is im-portant to correlate the flow ofTheraSphere generated by the sy-ringe with the normal blood flowthrough the hepatic artery. The inter-ventional radiologist best estimatesflow through the vessel beingtreated. Therefore, if the authorizeduser is not an interventional radiolo-gist, it is essential that the interven-tional radiologist clearly communi-cate to the authorized user the vesselsize and flow dynamics. The individ-ual tailoring of the administrationprocess is dependent on the informa-tion obtained from the interventionalradiologist. Injecting TheraSphere atan infusion rate faster than the vesselwill tolerate can result in reflux andadministration to nontarget sites.

SIR-Sphere infusion technique.—Justas with TheraSphere, most SIR-Spheres infusions should be per-formed through 3-F catheter systems.Although the use of smaller cathetersystems results in the need to gener-ate higher pressures for microspheredelivery, they create a safety mecha-nism to prevent the forceful and un-impeded advancement of micro-spheres that might occur with larger

5-F catheters. There is no pressuregauge on the SIR-Spheres deliverykit, and hence pressure cannot bemonitored. The delivery of SIR-Spheres is also dependent on bloodflow through the hepatic vasculaturedistal to the catheter tip. Given theembolic load of SIR-Spheres, it iseven more necessary to make certainthat the catheter does not occlude thevessel in which it is placed to pre-vent reflux. Flushing should be con-tinued until optimal delivery isachieved. A minimum flush of 60 mLis recommended with 3-F systems,depending on the activity infused.

The technical aspects of SIR-Spheres infusion are quite complex.The objective is to percutaneouslyachieve microsphere delivery in amanner analogous to that achievedwith use of a surgically implantedpump. It is essential that the vascularbed providing blood flow to the liverbe altered in a manner comparable tothat achieved with surgical tech-niques. Because the surgical proce-dure would involve skeletonizationof the common hepatic artery and li-gation of the right gastric artery, aswell as identification and exposure ofhepatic vessels flowing in an extrahe-patic direction, the percutaneousequivalent must be accomplished.This includes prophylactic emboliza-tion of the GDA and right gastric ar-tery as well as the supraduodenal,falciform, accessory left gastric, andaccessory inferior phrenic (ie, extra-hepatic) arteries (50). When this isachieved, given the fact that meta-static cancer is usually multifocal andbilobar, treatment by a lobar ap-proach is favored, thereby minimiz-ing the risk of hepatic decompensa-tion.

Given the larger number of micro-spheres (40–80 million) and loweractivity of SIR-Spheres (50 Bq per mi-crosphere) compared with Thera-Sphere, the delivery of SIR-Spheres isdistinctly different from that ofTheraSphere. When the catheter is inplace and the authorized user isready for to perform delivery, thecatheter is connected to the outlettubing. Given the very large numberof SIR-Spheres required to deliverthe intended dose, if the dosimetryformulas are strictly followed, it isnot uncommon for the entire vascu-lar bed to become saturated with mi-


crospheres and an embolic state to bereached. For this reason, fluoroscopicguidance is essential during the infu-sion. The technique of SIR-Spheresinfusion involves the alternating in-fusion of sterile water and contrastmedium, never allowing direct con-tact between the SIR-Spheres andcontrast medium. This allows the au-thorized user to adequately monitorthe injection and ensure that vascularsaturation has not been reached. Incases in which unrecognized vascularsaturation occurs and microsphereinfusion continues, reflux of micro-spheres and nontarget radiation be-comes a distinct possibility. The infu-sion is complete if (i) the entire in-tended dose has been infused with-out reaching stasis or (ii) stasis hasbeen reached and only a portion ofthe dose has been infused. Given therisk of reflux and nontarget radiationafter stasis has been reached, the con-tinued infusion of SIR-Spheres is notrecommended.

Monitoring of the SIR-Spheres in-fusion, as well as estimation of thepercentage infused at any time, maybe performed with use of an ioniza-tion chamber (minimum detection, 1mrem/h) placed adjacent to the SIR-Spheres kit. Keeping the ionizationchamber at a fixed distance from thevial and measuring baseline preinfu-sion dose can provide a live assess-ment of percentage infused. The dosereading should decrease as the infu-sion percentage increases. This maybe helpful in cases in which an au-thorized user wishes to infuse a cer-tain portion of the microspheres inone vascular territory, move the cath-eter position, and continue the infu-sion. The use of the ionization cham-ber can provide this information. As-sessing percentage infused by visualestimates is fraught with error and isnot recommended.

The final 1%–2% of SIR-Spheresmay be expelled from the tubing byloading the 20-mL injection syringewith air. Pressuring the system witha column of air will result in theslow expulsion of the remaining mi-crospheres. When this air columnreaches the three-way valve, itshould be turned off to prevent theintraarterial injection of air. A typicalSIR-Sphere infusion requires 10–20minutes to complete.

The technical aspects of SIR-

Spheres infusion for HCC deservespecial mention. Unlike colorectalcancer, HCC has a propensity forunifocality and significant hypervas-cularity, despite being bilobar. Giventhis, the usual cirrhotic nature of pa-tients with HCC, and the high em-bolic load of SIR-Spheres, administra-tion of microspheres in these patientsrequires a delicate approach. In suchcases, segmental/subsegmental infu-sion and dose reduction is recom-mended to (i) minimize the risk ofreflux into nontarget organs, (ii) min-imize the risk of microsphere flowinto the normal parenchyma, (iii)minimize the risk of radiation-in-duced liver disease, and (iv) maxi-mize the concentration of micro-spheres into the tumor bed (83,84).

General Radiation SafetyConsiderations

Radiation safety is an importantconsideration in this procedure, giventhe potentially high exposure fromhandling therapeutic amounts of 90Y.90Y is a �-emitter; therefore, the pri-mary concern is exposure to the eyes,skin, and hands. �-emissions from 90Ycan travel more than a meter in air butare significantly reduced by less than 1centimeter of acrylic. Although thedose vial is shielded in acrylic, the in-let and outlet catheters are notshielded.

Because 90Y microspheres are notmetabolized, they are registered as asealed source. However, the micro-spheres are delivered with use of sa-line solution or sterile water andshould therefore be handled by thesame techniques as those for radio-pharmaceuticals. If the system be-comes compromised, radioactive con-tamination is a concern, and stepsshould be taken to prevent the spreadof contamination. 90Y microspheres(resin and glass) contain trace amountsof long-lived radioactive contaminantssuch as 152Eu. If significant amounts ofresidual dose are present (�1% of thedose), these contaminants may be de-tected with a GM meter after 30 daysof decay. Disposal of this radioactivematerial should be addressed accord-ing to the facilities governing regula-tions (ie, Agreement State, Environ-mental Protection Agency, NuclearRegulatory Commission).

Typical surface radiation dose rates

from the patient range between 4 and12 mrem/h. This dose range is wellwithin the accepted radiation levelsfor outpatient radiation treatments.Regardless of whether the patient is aninpatient or an outpatient, no specialprecautions are necessary.

In summary, to avoid inaccuratedose administration or termination,the following are recommended dur-ing treatment: (i) procedural checklist,(ii) two or three dose verification tech-niques, and (iii) team approach withdefined roles for individuals from ra-diation oncology, nuclear medicine,interventional radiology, and medialphysics or radiation safety depart-ments. Both manufacturers provideauthorized users with strict proce-dural checklists. This checklist allowsfor the deliberate setup and infusion ofmicrospheres, thereby minimizing hu-man error. Any portion of the checklistthat cannot be completed during thesetup should be considered an abortpoint. In such instances, rather thanrisk a misadministration, the proce-dure should be terminated.

When the infusion is complete, au-thorized users should be aware of thepossibility of contamination and mi-crosphere deposition in the base cath-eter while the microcatheter is beingremoved. Hence, if backbleeding is ob-served from the base catheter on mi-crocatheter removal, contamination isa possibility. Caution should thereforebe exercised during removal and han-dling of the base catheters. Althoughthis can occur with glass and resinmicrospheres, we have observed thismore commonly with the resin micro-spheres.

In accordance with basic radiationsafety precaution, all personnel in theroom at the time of 90Y infusion mustbe measured for any possible contam-ination on their exit.

Radiation Safety Considerations inPatients UndergoingTransplantation or SurgicalResection

In several instances, patients under-going 90Y microspheres therapy be-come candidates for surgical resectionor liver transplantation. Although in-vestigators should follow their owninstitutional guidelines for elapsedtime from 90Y treatment to time of sur-gery or transplantation, this should be


balanced against the medical needs ofthe patient. In our institution, we rec-ommend monitoring the patient sur-face dose rate to determine what pre-cautions should be followed at thetime of surgery. Generally, a patientskin surface dose rate of less than 20�Sv/h does not require special han-dling by the surgeon at the time ofoperation. That is, lead gloves, specialinstruments, and extremity radiationmonitors (eg, ring badge) are not nec-essary. Radiation safety personnelshould be notified for transportationand storage of the explanted speci-men. In our institution, most of ourpatients treated with 90Y microsphereshave surface dose rates of less than 20�Sv/h at 30 days regardless of admin-istered activity.

After resection or transplantationsurgery, the explanted liver should beplaced in a formaldehyde solution forstorage in a leak-proof container. Thecontainer should be refrigerated whilein storage. Because the explanted livermay contain radioactive microspheres,the container should be monitoredwith an energy-compensated Geiger-Muller detector or a portable ioniza-tion chamber. If the dose rate at thesurface of the container exceeds 50�Sv/h, the container should be placedbehind lead shielding for decay in stor-age. While in storage, the explantedliver container should be labeled asradioactive material per federal andstate radiation safety guidelines. Insti-tutions should also follow federal andstate guidelines for room posting ofareas containing radioactive materialor designated radiation areas. Afterthe container has decayed for 60 days,the surface exposure rate will usuallybe less than 5–10 �Sv/h with the useof a portable ionization chamber. Atthat time, the specimen may be han-dled by the pathologist in the labora-tory according to standard precau-tions and techniques. However, whenpathologic analysis is complete, all tis-sue should be placed in the originalstorage container and returned to ra-dioactive material storage. All areas inwhich the liver specimen was handledshould be surveyed with a radiationdetection instrument such as a Geiger-Muller thin-window detector. Read-ings should be less than 0.1 mR/h.

In our institution, explanted liversfrom patients who received severaltreatments with 90Y microspheres were

obtained. We measured the container,slice specimens, and the grossing labo-ratory equipment and workspaces (ie,surgical pathology department). Theliver containers had been stored for atleast 60 days. With use of a portableionization chamber, the surface doserate measured approximately 4 �Sv/h.The sliced specimens containing thetreated site were measured with a Gei-ger-Muller pancake probe and expo-sure rates of 3–5 mR/h were obtained.Slices of explanted tissue that did notcontain microspheres were at back-ground levels (0.02 mR/h). The speci-men-slicing work area, slicing equip-ment, and documentation area werealso surveyed with the Geiger-Mullerpancake probe. No contamination waspresent, and all readings were at back-ground levels (0.02 mR/h).

Calculation of Residual Activity andProportional Activity Delivered

After administration of 90Y micro-spheres, the dose vial, inlet and outletcatheters, and towels beneath the de-livery device are placed in the Nalgenecontainer. This container fits into a cy-lindrical acrylic shield provided in theaccessory kit. With the same configu-ration of the ionization chamber withthe dose vial, the residual activity inthe assembly can be determined. Fordetermination of the actual dose (inGy) delivered to the target liver afterinjection, the following formula isused:

Dose (Gy) � 50 [(measured activity

(GBq) � (1 � LSF) � (1 � R)]/

liver mass (kg)

where LSF is the fraction of injectedradioactivity localizing in the lungs, asmeasured by 99mTc-MAA scintigra-phy, and R is the fraction of injectedradioactivity remaining in the dosevial, outlet catheter, and catheter asmeasured by the ionization chamber.

Authorized User Status

The technical aspects of micro-sphere delivery constitute an interven-tional radiology procedure almost intheir entirety. Therefore, it is impor-tant that interventional radiologistsplay a leading role in the future evo-lution and development of this tech-nology, including angiographic deliv-ery, dosimetry, and overall technical

enhancements in radioembolization.Depending on institutional policiesand hospital radiation safety com-mittees, interventional radiologistsshould function in a collegial mannerwith authorized users of brachyther-apy devices such as radiation oncol-ogy or nuclear medicine physicians.This having been said, although localregulatory bodies may impose somehurdles that limit the ability of inter-ventional radiologists to perform ra-dioembolization independently, thismodel should not be discouraged.Many successful radioembolizationpractices involving the interventionalradiologists as authorized users havebeen established, supporting the prac-tice model of the interventional radi-ologist as the authorized user of 90Y.

Interventional radiologists are cer-tified by the American Board of Radi-ology. Their specialized training is cer-tified by the American Board ofMedical Specialties. During their radi-ology residency and specialty fellow-ship training, interventional radiolo-gists receive formal didactic trainingin radiation biology, radiation physics,and radiation safety, making themfully able to undertake roles and re-sponsibilities of authorized users.Along with radiation oncologists, in-terventional radiologists undergowritten examinations administered bythe American Board of Radiology. Inmost clinical or hospital settings, radi-ologists, radiation oncologists, and nu-clear medicine physicians are the mostknowledgeable in matters regardingradiation and radiation safety. Inter-ventional radiologists also provide thefull spectrum of patient clinical careservices, including consultation andinitial patient evaluation, actual per-formance of the procedure, postoper-ative care, and follow-up care. Finally,one of the most compelling argumentsin support of interventional radiolo-gists as authorized users is that as ofcompletion of this manuscript, all ven-dor training of physicians beginning90Y use was being performed by inter-ventional radiologists, with the excep-tion of one radiation oncologist. There-fore, interventional radiologists areideally suited for authorized user sta-tus for radioembolization. This havingbeen said, institutions are directed toindividual state rules regarding autho-rized user status. At the time of manu-script completion, the Nuclear Regula-


tory Commission was consideringchanging the guidelines, which wouldpermit interventional radiologists tobecome independent authorized us-ers.

Postprocedural Care and Follow-up

Postprocedural considerations anddischarge.—Because 90Y therapy has alow toxicity profile, the treatment canbe performed on an outpatient basis.Given the possibility of small unrec-ognized arterial vessels coursing tothe gastrointestinal system, our pro-tocol recommends the routine use ofprophylactic antiulcer medications inall patients at the time of dischargeto minimize the risks of gastrointesti-nal irritation (85). Gastric coatingagents may also be used. In somecases, unless contraindicated (eg, dia-betes), a tapering 5-day steroid dosepack is also given to counteract fa-tigue. After the procedure, patientsrecover and are discharged within 6hours (2 hours if an arterial closuredevice is used). There are selectedand idiosyncratic reactions that mayoccur in the immediate postproce-dural time period. Aside from thetransient temporary burning that pa-tients may experience during 90Y in-jection, they may also experiencetransient, sudden-onset chills, shak-ing, and fever. These symptoms areshort lived and respond to meperi-dine, diphenhydramine, and acet-aminophen. They may occur im-mediately after the procedure andthrough the first week. At our insti-tution, the first five patients in whomthese symptoms were observed wereadmitted for standard fever workupincluding blood cultures, chest radi-ography, and urinalysis. The resultswere noncontributory in all cases.Overall, we have seen this reaction in10 patients. All responded to the im-plementation of the standard proto-col described earlier. Most patientswho experienced this reaction hadarterioportal (ie, not arteriohepaticvein) shunting or portal vein throm-bosis and all were being treated withTheraSphere for HCC (86). We nowroutinely administer diphenhydra-mine and meperidine immediatelybefore infusion of 90Y in patientswho exhibit the aforementioned an-giographic findings.

Because 90Y is a �-emitter, most

patients will have surface readings ofless than 1 mrem/h after implanta-tion. Therefore, standard biohazardprecautions are sufficient to protectfrom exposure to others after theyhave been discharged. With resin mi-crospheres, trace amounts (25–50kBq/L per GBq) of urinary excretionare a possibility in the first 24 hoursafter implantation (83). Investigatorsshould refer to the product inserts,institutional radiation safety commit-tees, and state and federal regulatoryagencies for guidance on 90Y use andpatient discharge instructions.

Follow-up care and side effects.—Themost common side effect of treatmentis fatigue. The majority of patients willexperience transient fatigue with vagueflulike symptoms. This is likely berelated to the effects of short-lived,low-dose radiation on the normal he-patic parenchyma (16,42,57). It is alsonot unusual for shaking, chills, andfever to occur days after treatment.One possible explanation for delayedshaking, chills, and fever is the para-neoplastic fever syndrome, with therelease of various pyretic agents andacute phase reactants such as tumornecrosis factor and C-reactive pro-tein. This is not unlike what may beobserved after TACE, and it is com-mon in patients with neuroendocrineand other highly vascular tumorssuch as HCC. The majority of thesesymptoms resolve on their own. How-ever, if they persist, infection shouldbe a concern, and proper steps shouldbe undertaken to exclude an infectiouscause. Other possible side effects in-clude abdominal pain, nausea, vomit-ing, and radiation cholecystitis. If thepatient presents with chronic abdom-inal pain, nausea, vomiting, or bleed-ing, endoscopic evaluation may beindicated to exclude gastrointestinalulceration. Despite the use of pro-phylactic drug regimens, the patientmay experience radiation gastritisand ulceration, both of which mayrequire surgery for definitive treat-ment. Radiation-induced liver diseaseis a possibility, particularly in pa-tients with compromised liver func-tion at initial treatment. Response tosteroids in this condition is variable.There has been some success in treat-ing patients with radiation hepatitisby the creation of transjugular intra-hepatic portosystemic shunts (BilbaoI, personal communication).

Because the majority of micro-sphere radioactivity has decayed by12 days (ie, four half-lives), all pa-tients should be seen clinically at 14days. This clinic visit is highly rec-ommended, particularly for the treat-ing physician during the early phaseof the 90Y program, because it pro-vides the treating physician with aclinical assessment and rapid learn-ing curve of the patient’s tolerancefor the treatment. The clinician canalso evaluate the patient for a secondtreatment to the other lobe at 30–60days follow-up if required. In addi-tion, this clinic visit permits monitor-ing of adverse sequelae such as fa-tigue (most common); tumor lysis orparaneoplastic syndrome; or hepatic,gastrointestinal, or pulmonary toxic-ity. Ultimately, the visit is most im-portant to confirm that the patient’sECOG PS has not deteriorated. Sig-nificant worsening in PS should betaken seriously and warrants delayin further therapy. If patients toleratethe first treatment, they should bescheduled for the second administra-tion as required. Performing labora-tory tests in patients in clinically sta-ble condition during the 14-day visitis left to the discretion of the treatingphysician, as most patients will ex-hibit a transient increase in amino-transferase and tumor marker levels.The majority of patients do experi-ence a sustained lymphopenia that isnot associated with opportunistic in-fections (16). This phenomenon ap-pears to occur more frequently withglass than with resin microspheres.

Treatment of the Second Lobe

Follow-up Evaluation of the FirstLobe.—Thirty days after the firsttreatment, assessment of response tothe first infusion must be under-taken, and overall clinic status of thepatient must be assessed. Liver func-tion tests, complete blood count withdifferential, tumor marker analysis,and cross-sectional imaging are per-formed. In the majority of cases,overall liver function (ie, total biliru-bin) should be unchanged. Depend-ing on the presence or absence of ex-trahepatic disease, tumor marker levelsmay increase, remain unchanged, ordecrease compared with baseline, mak-ing interpretation difficult. If they areincreased, tumor lysis, tumor progres-


sion in the liver, or the presence ofextrahepatic disease sites may be im-plicated. If they are unchanged, an ar-gument could be made that there hasbeen interval stabilization and/or im-provement in tumor burden. As a re-sult of the uncertainty in interpretationof tumor markers at 30 days, their usein the assessment of clinical responseshould be reserved for long-term fol-low-up. Finally, because radiation maybe implicated in thrombocytopeniaand bone marrow suppression, it isimportant to obtain a complete bloodcount with differential. Lymphocytesuppression without clinical sequelaehas been observed in many patientsundergoing 90Y therapy (11). The fol-low-up imaging modality used to as-sess tumor response should be con-sistent with that used for baselineimaging. If MR imaging is used, dif-fusion-weighted imaging permits thedocumentation of necrosis and celldeath (37). CT may limit the ability todefinitively document tumor necrosis.However, other indirect criteria can beused, such as size of the lesion and rel-ative alterations in vascularity and en-hancement. Follow-up functional imag-ing such as MR imaging or PET maybe helpful (28,37,44,45).

The 4-week follow-up scan hasa different purpose depending onwhether primary or metastatic dis-ease is treated. For HCC, the 4-weekscan determines the degree of tumorshrinkage and necrosis. Progressionto the point of alteration of the clini-cal liver-directed therapy plan is un-usual in HCC at 30 days after thefirst treatment. For metastatic dis-ease, the 30-day follow-up scan as-sesses tumor response and the pres-ence of necrosis. Most importantly,given the greater likelihood of tumorprogression (as opposed to HCC) inpatients with liver metastases pa-tients being treated with liver-di-rected therapy, the purpose of the30-day scan is to search for failure ofresponse to 90Y treatment in thetreated lobe or extrahepatic progres-sion. Either of these scenarios mayresult in the discontinuation of 90Ytreatment. Tumor progression in theuntreated lobe does not represent areason for discontinuation of treat-ment with 90Y if a positive result ofthe first treatment has been achieved.Positive results that support contin-ued 90Y treatment to the other lobe at

day 30 include (i) stability in tumorsize, (ii) tumor shrinkage, (iii) necro-sis within the tumor with or withouttumor shrinkage, (iv) improvementon PET in the treated area, (v) im-provement in liver function test re-sults, (vi) improvement in PS or pain,and (vii) improvement in tumormarkers if applicable. Results at day30 that may result in the discontinua-tion of 90Y treatment include (i) pro-gression in the treated area implyingradiation resistance, (ii) developmentof significant extrahepatic disease,(iii) worsening in liver function testresults precluding further treatment,and (iv) worsening of PS.

Treatment of the second (or addi-tional) lobe(s).—From an angiographicand technical standpoint, the admin-istration to the second (or other) lobeshould be straightforward, because itis the patient’s third (or more) cathe-terization. Therefore, optimal cathe-ters and guide wires will have beenestablished during the planning vis-ceral arteriography and first treat-ment session. If the first treatmentwas performed with lobar 99mTc-MAA evaluation (as might be donefor HCC), a second 99mTc-MAA ad-ministration and calculation of shunt-ing may be required. If whole-liver99mTc-MAA was performed duringplanning mesenteric angiography, arepeat 99mTc-MAA scan is not neces-sary. The cumulative dose of 90Y ad-ministered must not exceed a totallung exposure of 50 Gy (55).

Clinically, patients tolerate the sec-ond treatment similarly as they dothe first, although there is usuallyless fatigue. This may simply be aphenomenon of the smaller left lobesbeing treated second. Therefore, thedegree of fatigue varies, most of thiseffect occurring during treatment ofthe larger lobe, usually the right.There may also be a relationship be-tween dose received and the mani-festation of fatigue. Interestingly, thesensation of burning, fever, and chillsis often reproduced in the same pa-tient.

Repeat treatment of a lobe/segment.—Repeat injection of 90Y may be neces-sary in a previously treated vascularbed (ie, lobe), such as recurrent dis-ease or incompletely treated disease.However, it is important to recognizethat the mode of action of 90Y is dis-tinctly different from that of embolic-

type therapy such as TACE or drug-eluting microspheres.

TACE involves the infusion ofhigh-dose chemotherapy, followedby ischemia-inducing particles at the300- to 700-�m level. When ischemiais induced, hepatocytes may still re-cruit blood flow from the portal veinor extrahepatic arterial vessels suchas the right inferior phrenic or inter-costal and thereby maximize thechance of cellular viability. Repeattreatment with TACE is thereforepossible because this mechanism ofhypoxia induction is independent oftumor presence and hypervascular-ity. Normal hepatocytes will continueto recruit vascularity with each em-bolization procedure.

The mechanism of 90Y involvesabsorption of the microspheres byhypervascular tumors. The greaterthe hypervascularity, the more mi-crospheres are absorbed in the tu-mor, and the lower the correspond-ing dose to normal parenchyma. The-oretically, if blood flow to a lobewere 100% to tumor and zero to nor-mal parenchyma, the entire 90Y dosewould be absorbed by the tumor andnormal parenchyma would receiveno exposure. Given this mechanismof action, hypervascular tumors be-come necrotic and obliterated froman angiographic standpoint aftertreatment (58). This mechanism of ac-tion breaks down if retreatment isundertaken and tumor hypervascu-larity is not present, given the signifi-cant reduction in tumor blood flowthat occurs after treatment with 90Y.Therefore, during subsequent treat-ments, if tumor hypervascularity isnot present, fewer microspheres areabsorbed by tumor, and the corre-sponding normal parenchymal doseis increased. The limitation describedherein lies in the inability to defini-tively calculate normal parenchymaldose with 90Y microspheres. There-fore, until this limitation is corrected,repeat treatment with 90Y should beconsidered only when tumor hyper-vascularity persists after initial treat-ment, and in such cases, segmentalinfusions should be performed if an-giographically feasible. Little phaseI/II work has been done with re-peated treatment of 90Y; cautionshould be exercised in such cases.

In cases in which repeat treatmentis initiated, several steps should be


undertaken to minimize the risks ofradiation hepatitis and pneumonitis.First, it is essential that repeat 99mTc-MAA shunting fraction be evaluated.This is necessitated by the principleof hyperaugmentation of pulmonaryshunting (67). Previously treated seg-ments or lobes of liver may have al-tered microvascularity flow dynamicsat the tumor level, requiring a repeat99mTc-MAA scan. If shunting is notre-evaluated, the actual lung shunt-ing may be underestimated, placingthe patient at undue risk for lung in-jury. Second, depending on the an-giographic findings of tumor vascu-larity, a lowered dose should be con-sidered until phase I/II dose-escala-tion studies are complete. Finally,unless contraindicated (eg, in cases ofdiabetes), we advocate the use of a14-day low-dose steroid regimen aswell as antiinflammatory drugs inpatients undergoing repeat treatmentfor HCC or metastases. This may the-oretically reduce the risk of radia-tion-induced liver disease. Cautionshould be exercised in the use ofnonsteroidal antiinflammatory drugsin patients with cirrhosis who haveesophageal varices (87).

In summary, as long as there issignificant tumor hypervascularityacting as a receptacle and sump formicrospheres, repeat treatment maybe considered if all other factors areappropriate (eg, repeat lung shuntfraction, as discussed earlier). If alarge avascular necrotic lesion wascreated after initial treatment, repeattreatment should be considered withcaution because microspheres will nolonger be absorbed by tumor but willrather result in the irradiation of nor-mal parenchyma. If significant tumorhypervascularity persists after an ini-tial course of 90Y, the cautionary noteregarding repeat treatment becomeless valid. As long as there is tumorvascularity to absorb the micro-spheres, repeat treatment may beconsidered, provided the precautionslisted earlier (eg, hyperaugmentationof lung shunt) are recognized.

SUMMARY

Radioembolization with 90Y micro-spheres represents an innovative ap-proach that has gained increasingawareness and clinical use during the

past 5–10 years. As described in thisarticle, two 90Y radioembolization de-vices are available today. Althoughboth use 90Y as the radioisotope, theirmodes of action are distinctly differ-ent. A thorough understanding ofthese differences should help decipherwhich device might best be applied tocertain patients.

The minimal toxicity of radioembo-lization and the ability to dischargethe patient on an outpatient basismake the therapy an attractive alterna-tive in the treatment of primary andmetastatic liver malignancies. Patientsare able to resume normal activitiesshortly after treatment, with minimalside effects, in contrast to the postem-bolization syndrome often associatedwith current chemoembolic tech-niques.

Treatment planning requires a mul-tidisciplinary team with clear leader-ship (ie the driver, as discussed ear-lier) and accountability to ensure thatthe screening, diagnostic, and treat-ment procedures are conducted in aseamless fashion. The essential stepsinclude (i) calculation of target livermass to be infused and tumor burden,(ii) visceral angiography to map tu-mor-perfusing vessels and embolizecollateral vessels, (iii) assessment ofpulmonary shunt, (iv) determinationof the optimal therapeutic dose, (v)room preparation, (vi) radiation mon-itoring and safety procedures, and(vii) calculation of residual activityand efficiency of 90Y delivery.

Careful patient selection and prep-aration for 90Y liver-directed therapywill result in an optimal risk/benefitratio for the patient. For patients withHCC, the treatment of advancing dis-ease must be balanced against the of-ten-compromised functional liver re-serve caused by underlying cirrhosis.Selection of patients with adequatehepatic reserve and good functionalstatus will maximize the beneficialtherapeutic effect of 90Y therapy withminimal risk to normal liver paren-chyma. 90Y therapy has also beenshown to be beneficial for patientswith metastatic disease who have in-trahepatic progression despite stan-dard-of-care chemotherapy.

When a patient is being consideredfor treatment with 90Y therapy, lesionpresentation (focal or diffuse), hepaticand renal function, previous therapies

(systemic and intrahepatic), and PSwill all affect the prognostic outcome.Careful imaging and angiographicevaluation to assess tumor distribu-tion, vascular anatomy, contributionto intra- and extrahepatic vessels, andpulmonary shunt is essential for (i) ac-curate delivery of the microspheres tothe intended target lesion(s) (ie, safety)and (ii) the efficacy of 90Y in selectivelytargeting high radiation dose to effecttumor kill while minimizing radiationexposure to normal parenchyma.

Depending on the cause, functionalstatus, hepatic anatomic vasculature,and presentation of liver lesions, sev-eral factors need to be considered intreatment planning. If repeat treat-ment of the same target area is antici-pated, cumulative radiation exposureto the target area should be deter-mined and the radiation dose adjustedaccordingly to minimize the risk of ra-diation-induced hepatitis. If vascular-ity permits, selective infusion of asegment or lesion (ie, “radiation seg-mentectomy”) will mitigate the risk ofradiation exposure to normal tissue.Dose reduction may also be required ifestimated pulmonary shunt exceeds50 Gy among cumulative infusions of90Y or if hyperaugmentation of lungshunt is identified. Prophylactic ad-ministration of antiulcer medication(for 2 weeks) and steroids (for 5–7days) after treatment will mitigate therisk of pain from nontarget radiationinto the gastrointestinal tract and pro-vide relief from fatigue, respectively.

Postprocedural follow-up of the pa-tient to assess any treatment-emergentside effects and tumor response is con-ducted at 30 days and then at 2- to3-month intervals thereafter. Otherthan mild to moderate constitutionalsymptoms, the side effects of radioem-bolization include nontarget radiation,radiation pneumonitis, and radiationhepatitis. Diligent vascular mappingduring the treatment planning angiog-raphy with embolization of the GDAand right gastric arteries, as well asother perforating vessels, will mini-mize the likelihood of inadvertentdeposition to gastric structures. Asmentioned earlier, dose reduction andcareful consideration of functionalliver reserve will mitigate the occur-rence of radiation pneumonitis and ra-diation hepatitis, respectively.


CONCLUSION

There are inherent advantages tothe use of radioactive microspheres forthe treatment of liver cancers. Deliveryto small target volumes, the ability toeffect much higher doses of radiationcompared with external-beam radia-tion, the relatively low toxicity profile,and the tumoricidal effect of radiationirrespective of tumor origin make thismode of therapy particularly more at-tractive in comparison with disease-specific targeted microspheres. Doxo-rubicin-coated microspheres may haveactivity on HCC but not on colorectalcancer. Slow-release irinotecan or ox-aliplatin microspheres may have an ef-fect on colorectal cancer, but not onmelanoma. Radioembolization, in theform of 90Y, rhenium, phosphorous P32, or holmium, eliminates the needfor specificity of a therapeutic device.If delivered at the correct activity tothe intended location, radioemboliza-tion will have a tumoricidal effect onall neoplastic tissue. This is the basis ofradiation oncology: radiation deliveredto tumor at scheduled and sufficientlyincreased levels will have a tumoricidaleffect, irrespective of cellular origin.Therefore, radiation microparticles areparticularly attractive as a therapeuticdevice.

The unique aspects of 90Y therapyare its minimal toxicity profile andhighly effective tumor kill with mini-mal exposure to normal liver tissuein properly selected patients. Theseunique characteristics in conjunctionwith the minimally invasive natureprovide an attractive option for pa-tients for whom there are few alterna-tives. The technical and clinical de-mands of patient selection, treatmentplanning, 90Y administration, and clin-ical follow-up require a dedicated in-terdisciplinary team willing to workcooperatively to achieve the best resultfor the patient. The clinical benefit andpotential for enhancing quality of lifefor the patient, given this commit-ment, present an exciting opportunityfor the field of interventional oncol-ogy.

This article is the first of a three-part series that represents the culmi-nation of years of effort in introducingand developing 90Y therapy into clini-cal practice and is meant to spark in-terest in the future development ofthis and other radioembolization tech-

niques. The authors sincerely hopethat this goal has been accomplishedand that the medical community cancontinue to scientifically advance thispromising therapy.

Acknowledgments: The authors thankthe following individuals for their contri-bution to this comprehensive review arti-cle: Beth Oman, Agnieszka Stanisz, KarenBarrett, Krystina Sajdak, Margaret Gilbert-sen, and Vanessa Gates. In particular, theauthors thank Angi Courtney for her con-tribution to this manuscript.

References1. TheraSphere Yttrium-90 microspheres

[package insert]. Kanata, ON, Cana-da: MDS Nordion, 2004.

2. Humanitarian Device Exemptions(HDE) Regulations: Questions and An-swers; Final Guidance for Industry.Available at http://www.fda.gov/cdrh/ode/guidance/1381.html.

3. SIR-Spheres Yttrium-90 microspheres[package insert]. Lane Cove, Austra-lia: Sirtex Medical. 2004.

4. American Cancer Society. CancerFacts & Figures 2004. Atlanta: Ameri-can Cancer Society, 2004.

5. Salem R, Hunter R. Yttrium-90 mi-crospheres for the treatment of hepato-cellular carcinoma: a review. Int J Ra-diat Oncol Biol Phys (in press).

6. Lau WY, Leung WT, Ho S, et al.Treatment of inoperable hepatocellularcarcinoma with intrahepatic arterialyttrium-90 microspheres: a phase I andII study. Br J Cancer 1994; 70:994–999.

7. Lau WY, Ho S, Leung TW, et al. Se-lective internal radiation therapy fornonresectable hepatocellular carci-noma with intraarterial infusion of90yttrium microspheres. Int J RadiatOncol Biol Phys 1998; 40:583–592.

8. Cao X, He N, Sun J, et al. Hepaticradioembolization with yttrium-90glass microspheres for treatment of pri-mary liver cancer. Chin Med J (Engl)1999; 112:430–432.

9. Llovet JM. Treatment of hepatocellu-lar carcinoma. Curr Treat Options Gas-troenterol 2004; 7:431–441.

10. Bruix J, Sherman M, Llovet JM, et al.Clinical management of hepatocellularcarcinoma: conclusions of the Barce-lona-2000 EASL conference. EuropeanAssociation for the Study of the Liver.J Hepatol 2001; 35:421–430.

11. Carr BI. Hepatic arterial 90yttriumglass microspheres (Therasphere) forunresectable hepatocellular carcinoma:interim safety and survival data on 65patients. Liver Transpl 2004; 10(2 suppl1):S107–S110.

12. Dancey JE, Shepherd FA, Paul K, et al.Treatment of nonresectable hepatocel-lular carcinoma with intrahepatic 90Y-

microspheres. J Nucl Med 2000;41:1673–1681.

13. Geschwind JF, Salem R, Carr BI, et al.Yttrium-90 microspheres for the treat-ment of hepatocellular carcinoma. Gas-troenterology 2004; 127(5 suppl 1):S194–S205.

14. Deleted in proof.15. Salem R, Lewandowski R, Roberts C, et

al. Use of yttrium-90 glass micro-spheres (TheraSphere) for the treat-ment of unresectable hepatocellularcarcinoma in patients with portal veinthrombosis. J Vasc Interv Radiol 2004;15:335–345.

16. Salem R, Lewandowski RJ, Atassi B, etal. Treatment of unresectable hepato-cellular carcinoma with use of 90Y mi-crospheres (TheraSphere): safety, tu-mor response, and survival. J VascInterv Radiol 2005; 16:1627–1639.

17. Salem R, Thurston KG, Carr BI, et al.Yttrium-90 microspheres: radiationtherapy for unresectable liver cancer. JVasc Interv Radiol 2002; 13(suppl):S223–S229.

18. Steel J, Baum A, Carr B. Quality oflife in patients diagnosed with primaryhepatocellular carcinoma: hepatic arte-rial infusion of cisplatin versus 90-yttrium microspheres (Therasphere).Psychooncology 2004; 13:73–79.

19. Liu MD, Uaje MB, Al-Ghazi MS, et al.Use of yttrium-90 TheraSphere for thetreatment of unresectable hepatocellu-lar carcinoma. Am Surg 2004; 70:947–953.

20. Messersmith W, Laheru D, Hidalgo M.Recent advances in the pharmacologi-cal treatment of colorectal cancer. Ex-pert Opin Investig Drugs 2003; 12:423–434.

21. Mulcahy MF, Benson AB III. Bevaci-zumab in the treatment of colorectalcancer. Expert Opin Biol Ther 2005;5:997–1005.

22. Hoff PM. Future directions in the useof antiangiogenic agents in patientswith colorectal cancer. Semin Oncol2004; 31(6 suppl 17):17–21.

23. Sasson AR, Sigurdson ER. Surgicaltreatment of liver metastases. SeminOncol 2002; 29:107–118.

24. Gray B, Van Hazel G, Hope M, et al.Randomised trial of SIR-Spheres pluschemotherapy vs. chemotherapy alonefor treating patients with liver metas-tases from primary large bowel cancer.Ann Oncol 2001; 12:1711–1720.

25. Stubbs RS, Cannan RJ, Mitchell AW.Selective internal radiation therapy(SIRT) with 90Yttrium microspheresfor extensive colorectal liver metasta-ses. Hepatogastroenterology 2001; 48:333–337.

26. Stubbs RS, Cannan RJ, Mitchell AW.Selective internal radiation therapywith 90yttrium microspheres for exten-


sive colorectal liver metastases. J Gas-trointest Surg 2001; 5:294–302.

27. Rubin D, Nutting C, Jones B. Meta-static breast cancer in a 54-year-oldwoman: integrative treatment withyttrium-90 radioembolization. IntegrCancer Ther 2004; 3:262–267.

28. Wong CY, Qing F, Savin M, et al. Re-duction of metastatic load to liver afterintraarterial hepatic yttrium-90 radio-embolization as evaluated by [18f]flu-orodeoxyglucose positron emission to-mographic imaging. J Vasc IntervRadiol 2005; 16:1101–1106.

29. Murthy R, Xiong H, Nunez R, et al.Yttrium 90 resin microspheres for thetreatment of unresectable colorectal he-patic metastases after failure of multi-ple chemotherapy regimens: prelimi-nary results. J Vasc Interv Radiol 2005;16:937–945.

30. Cortez-Pinto H, Camilo ME. Non-al-coholic fatty liver disease/non-alco-holic steatohepatitis (NAFLD/NASH):diagnosis and clinical course. BestPract Res Clin Gastroenterol 2004; 18:1089–1104.

31. Kim TK, Jang HJ, Wilson SR. Imagingdiagnosis of hepatocellular carcinomawith differentiation from other pathol-ogy. Clin Liver Dis 2005; 9:253–279.

32. Alobaidi M, Shirkhoda A. Benign fo-cal liver lesions: discrimination frommalignant mimickers. Curr Probl Di-agn Radiol 2004; 33:239–253.

33. Lim JH, Choi D, Kim SH, et al. Detec-tion of hepatocellular carcinoma: valueof adding delayed phase imaging todual-phase helical CT. AJR Am JRoentgenol 2002; 179:67–73.

34. Kim YK, Kim CS, Lee YH, et al. Com-parison of superparamagnetic iron ox-ide-enhanced and gadobenate dime-glumine-enhanced dynamic MRI fordetection of small hepatocellular carci-nomas. AJR Am J Roentgenol 2004;182:1217–1223.

35. Rhee TK, Larson AC, Prasad PV, et al.Feasibility of blood oxygenation level-dependent MR imaging to monitor he-patic transcatheter arterial emboliza-tion in rabbits. J Vasc Interv Radiol2005; 16:1523–1528.

36. Llovet JM, Vilana R, Bru C, et al. In-creased risk of tumor seeding after per-cutaneous radiofrequency ablation forsingle hepatocellular carcinoma. Hepa-tology 2001; 33:1124–1129.

37. Geschwind JF, Artemov D, Abraham S,et al. Chemoembolization of liver tu-mor in a rabbit model: assessment oftumor cell death with diffusion-weighted MR imaging and histologicanalysis. J Vasc Interv Radiol 2000; 11:1245–1255.

38. Okuda K, Ohtsuki T, Obata H, et al.Natural history of hepatocellular carci-noma and prognosis in relation to

treatment: study of 850 patients. Can-cer 1985; 56:918–928.

39. Rhee TK, Omary RA, Gates V, et al.The effect of catheter-directed ct an-giography on yttrium-90 radioemboli-zation treatment of hepatocellular car-cinoma. J Vasc Interv Radiol 2005; 16:1085–1091.

40. Toyoda H, Kumada T, Kiriyama S, etal. Comparison of the usefulness ofthree staging systems for hepatocellu-lar carcinoma (CLIP, BCLC, and JIS) inJapan. Am J Gastroenterol 2005; 100:1764–1771.

41. Kulik L, Atassi B, van Holsbeeck L, etal. Yttrium-90 microspheres (Thera-Sphere) treatment of unresectable hep-atocellular carcinoma: downstaging toresection, RFA and bridge to transplan-tation. J Surg Oncol 2006 (in press).

42. Lewandowski RJ, Thurston KG, GoinJE, et al. 90Y Microsphere (Thera-Sphere) treatment for unresectablecolorectal cancer metastases of the liv-er: response to treatment at targeteddoses of 135-150 Gy as measured by[18F]fluorodeoxyglucose positron emis-sion tomography and computed tomo-graphic imaging. J Vasc Interv Radiol2005; 16:1641–1651.

43. Lencioni R, Crocetti L, Cioni D, et al.Percutaneous radiofrequency ablationof hepatic colorectal metastases: tech-nique, indications, results, and newpromises. Invest Radiol 2004; 39:689–697.

44. Wong CY, Salem R, Qing F, et al.Metabolic response after intraarterial90Y-glass microsphere treatment forcolorectal liver metastases: comparisonof quantitative and visual analyses by18F-FDG PET. J Nucl Med 2004; 45:1892–1897.

45. Wong CY, Salem R, Raman S, et al.Evaluating 90Y-glass microspheretreatment response of unresectablecolorectal liver metastases by [18F]FDG PET: a comparison with CT orMRI. Eur J Nucl Med Mol Imaging2002; 29:815–820.

46. Jerusalem G, Hustinx R, Beguin Y, etal. Evaluation of therapy for lym-phoma. Semin Nucl Med 2005; 35:186–196.

47. Tellez C, Benson AB III, Lyster MT, etal. Phase II trial of chemoemboliza-tion for the treatment of metastaticcolorectal carcinoma to the liver andreview of the literature. Cancer 1998;82:1250–1259.

48. Yu AS, Keeffe EB. Management ofhepatocellular carcinoma. Rev Gastro-enterol Disord 2003; 3:8–24.

49. Hashimoto K, Ikeda Y, Korenaga D, etal. The impact of preoperative serumC-reactive protein on the prognosis ofpatients with hepatocellular carci-noma. Cancer 2005; 103:1856–1864.

50. Liu DM, Salem R, Bui JT, et al. An-giographic considerations in patientsundergoing liver-directed therapy. JVasc Interv Radiol 2005; 16:911–935.

51. Chen JH, Chai JW, Huang CL, et al.Proximal arterioportal shunting associ-ated with hepatocellular carcinoma:features revealed by dynamic helicalCT. AJR Am J Roentgenol 1999; 172:403–407.

52. Ho S, Lau WY, Leung TW, et al. Par-tition model for estimating radiationdoses from yttrium-90 microspheres intreating hepatic tumours. Eur J NuclMed 1996; 23:947–952.

53. Mounajjed T, Salem R, Rhee TK, et al.Multi-institutional comparison of99mTc-MAA lung shunt fraction fortranscatheter Y-90 radioembolization.Presented at the Annual Meeting of theSociety of Interventional Radiology,2005. New Orleans, LA.

54. Hung JC, Redfern MG, Mahoney DW,et al. Evaluation of macroaggregatedalbumin particle sizes for use in pul-monary shunt patient studies. J AmPharm Assoc (Wash) 2000; 40:46–51.

55. Ho S, Lau WY, Leung TW, et al. Clin-ical evaluation of the partition modelfor estimating radiation doses fromyttrium-90 microspheres in the treat-ment of hepatic cancer. Eur J Nucl Med1997; 24:293–298.

56. Kim HC, Chung JW, Lee W, et al.Recognizing extrahepatic collateralvessels that supply hepatocellular car-cinoma to avoid complications oftranscatheter arterial chemoemboliza-tion. Radiographics 2005; 25(suppl 1):S25–S39.

57. Murthy R, Nunez R, Szklaruk J, et al.Yttrium-90 microsphere therapy forhepatic malignancy: devices, indica-tions, technical considerations, and po-tential complications. Radiographics2005; 25(suppl 1):S41–S55.

58. Sato KT, Omary RA, Ryu R, et al. Ra-dioembolization with yttrium-90 mi-crospheres: angiographic and technicalconsiderations. Tech Vasc Interv Ra-diol 2006 (in press).

59. Nakamura H, Hashimoto T, Oi H, et al.Prevention of gastric complications inhepatic arterial chemoembolization:balloon catheter occlusion technique.Acta Radiol 1991; 32:81–82.

60. Song SY, Chung JW, Lim HG, et al.Nonhepatic arteries originating fromthe hepatic arteries: angiographic anal-ysis in 250 patients. J Vasc Interv Ra-diol 2006; 17:461–469.

61. Urbach DR, Herron DM, KhajancheeYS, et al. Laparoscopic hepatic arteryinfusion pump placement. Arch Surg2001; 136:700–704.

62. Kulik LM, Mulcahy MF, Hunter RD, etal. Use of yttrium-90 microspheres(TheraSphere) in a patient with unre-


sectable hepatocellular carcinoma lead-ing to liver transplantation: a case re-port. Liver Transpl 2005; 11:1127–1131.

63. Charnsangavej C. Chemoemboliza-tion of liver tumors. Sem Invest Radiol1993; 10:150–160.

64. Fischer L, Thorn M, Neumann JO, et al.The segments of the hepatic veins: isthere a spatial correlation to theCouinaud liver segments? Eur J Radiol2005; 53:245–255.

65. Erbe EM, Day DE. Chemical durabil-ity of Y2O3-Al2O3-SiO2 glasses for thein vivo delivery of beta radiation. JBiomed Mater Res 1993; 27:1301–1308.

66. Kennedy AS, Nutting C, Coldwell D, etal. Pathologic response and microdo-simetry of (90)Y microspheres in man:review of four explanted whole livers.Int J Radiat Oncol Biol Phys 2004; 60:1552–1563.

67. Murthy R, Kennedy A, Line B, et al.Augmentation of hepatopulmonaryshunting following trans-arterial he-patic brachytherapy with Y90 Thera-Sphere: therapeutic implications. Pre-sented at the Radiological Society ofNorth America, 2002. Chicago, IL.

68. Sato K, Lewandowski RJ, Bui JT, et al.Treatment of unresectable primary andmetastatic liver cancer with yttrium-90microspheres (TheraSphere): assess-ment of hepatic arterial embolization.Cardiovasc Intervent Radiol 2006; 29:522�529.

69. Goin JE, Dancey JE, Roberts CA, et al.Comparison of post-embolization syn-drome in the treatment of patients withunresectable hepatocellular carcinoma:trans-catheter arterial chemo-emboli-zation versus yttrium-90 glass micro-

spheres. World J Nucl Med 2004; 3:49–56.

70. Goin JE, Salem R, Carr BI, et al. Treat-ment of unresectable hepatocellularcarcinoma with intrahepatic yttrium 90microspheres: a risk-stratification anal-ysis. J Vasc Interv Radiol 2005; 16:195–203.

71. Goin JE, Salem R, Carr BI, et al. Treat-ment of unresectable hepatocellularcarcinoma with intrahepatic yttrium 90microspheres: factors associated withliver toxicities. J Vasc Interv Radiol2005; 16:205–213.

72. Van Hazel G, Blackwell A, Anderson J,et al. Randomised phase 2 trial ofSIR-Spheres plus fluorouracil/leucov-orin chemotherapy versus fluoroura-cil/leucovorin chemotherapy alone inadvanced colorectal cancer. J Surg On-col 2004; 88:78–85.

73. Body Surface Area Calculator. Avail-able at http://www.halls.md/body-surface-area/refs.htm.

74. DuBois D, DuBois E. A formula toestimate the approximate surface areaif height and weight be known. ArchInt Med 1916; 17:863–871.

75. Deutsch E, Huguet F. Perspectives inbiological modulation of radiotherapy.Bull Cancer 2005; 92:90–96.

76. Koukourakis MI. Tumour angiogene-sis and response to radiotherapy. An-ticancer Res 2001; 21(6B):4285–4300.

77. Das U. A radical approach to cancer.Med Sci Monit 2002; 8:RA79–RA92.

78. Hakim TS, Dean GW, Lisbona R. Ef-fect of body posture on spatial distri-bution of pulmonary blood flow.J Appl Physiol 1988; 64:1160–1170.

79. Kosuda S, Kobayashi H, Kusano S.

Change in regional pulmonary perfu-sion as a result of posture and lungvolume assessed using technetium-99m macroaggregated albumin SPET.Eur J Nucl Med 2000; 27:529–535.

80. Beck KC. Regional trapping of micro-spheres in the lung compares well withregional blood flow. J Appl Physiol1987; 63:883–889.

81. Berger MJ. Distribution of absorbeddose around point sources of electronsand beta particles in water and othermedia. J Nucl Med 1971: suppl 5:5–23.

82. Snyder W, Ford M, Warner G, et al.Absorbed dose per unit cumulated ac-tivity for selected radionuclides and or-gans. New York: Society of NuclearMedicine, 1975–1976.

83. Sirtex Medical Training Manual.Lane Cove, Australia: Sirtex Medical,2005.

84. Gray BN. Use of SIR-Spheres in pa-tients with impaired liver function:safety notice. Lane Cove, Australia: Sir-tex Medical, 2003.

85. Yip D, Allen R, Ashton C, et al. Radi-ation-induced ulceration of the stom-ach secondary to hepatic embolizationwith radioactive yttrium microspheresin the treatment of metastatic coloncancer. J Gastroenterol Hepatol 2004;19:347–349.

86. Salem R. Idiosyncratic reactions fol-lowing Y90 treatment. Chicago: North-western Memorial Hospital, 2006.

87. De Ledinghen V, Heresbach D, Four-dan O, et al. Anti-inflammatorydrugs and variceal bleeding: a case-control study. Gut 1999; 44:270–273.


Radioembolization with 90Yttrium Microspheres: A State-of ... · Radioembolization with 90Yttrium...

Documents

Transcript of Radioembolization with 90Yttrium Microspheres: A State-of ... · Radioembolization with 90Yttrium...