Current Approach to Planning Angiography and MAA Administration

Abstract

Transarterial radioembolization of primary and secondary hepatic malignancies utilizing yttrium-90 microspheres is a commonly performed treatment by interventional radiologists. Traditionally performed as a two-part procedure, a diagnostic angiography is performed 1 to 3 weeks prior to treatment with the injection of technetium-99m-macroaggregated albumin followed by planar scintigraphy in the nuclear medicine department. Careful attention must be paid to the details during the diagnostic angiography to ensure the delivery of a safe and optimal dose to the diseased liver and to minimize radiation-induced damage to both unaffected liver and adjacent structures. In this article, we will review the steps and considerations that must be made during the angiography planning and discuss current and future areas of research.

Radioembolization using yttrium-90 (Y-90)-impregnated glass or resin microspheres has become a staple of locoregional therapy for intermediate-stage, unresectable primary and secondary hepatic malignancies. ¹ While many of the procedure steps mimic transarterial bland embolization and chemoembolization, Y-90 has multiple additional layers of complexity. Unlike other transarterial embolization procedures, Y-90 includes two separate procedures, a planning (or mapping) angiography with subsequent therapeutic infusion classically 1 to 3 weeks later. Mapping procedure is performed to evaluate efficacy and safety of the subsequent radioembolization, including hepatopulmonary shunting/non-target embolization to the lungs or abdominal viscera from the point of injection, liver volume intended to treat, thorough investigation of hepatic arterial anatomy, and relative deposition in tumor versus normal parenchyma. Concomitant prophylactic embolization can be performed, if necessary. Planning procedure includes a diagnostic angiography of the liver, infusion of technetium-99m-macroaggregated albumin ( ^99m Tc-MAA), and postprocedure imaging (scintigraphy) of the ^99m Tc-MAA deposition. Radiolabeled MAA is used as a planning agent because the size of MAA particles is similar to that of the Y-90-impregnated particles and it is biodegradable and, therefore, will not permanently occlude arteries into which it has been injected.

Planning angiograms can be time consuming and technically difficult, and their importance is easily overlooked. A thorough and complete planning angiogram sets the stage for appropriate dosimetry and optimal treatment effect following Y-90 administration. This article will review the authors' institutional approach to the planning angiogram and ^99m Tc-MAA administration.

Procedural Goals

Both planning of angiography and subsequent ^99m Tc-MAA scintigraphy serve multiple purposes and the intricacies of each must not be underestimated. The mapping procedure sets the stage for subsequent success following the treatment.

Determination of Hepatic and Tumoral Vascular Flow

Hepatic arterial anatomy is highly variable. In 1966, Michels proposed the first classification scheme for hepatic artery variants based on a study of 200 cadavers. ² This was modified (and simplified) in 1994 by Hiatt et al ³ in a larger study of 1,000 donor livers used for orthotopic transplantation. Hiatt et al's classification examined only the extrahepatic arteries and therefore does not distinguish between replaced and accessory arteries, as the intrahepatic arteries were not dissected. Table 1 lists the variants according to both schema and their relative incidences in both studies. Cadaver studies have also described and attempted to quantify the incidence of variant hepatic vascular anatomy. ⁴ Major branch vessel anatomy is often able to be ascertained by preprocedure magnetic resonance imaging (MRI) or computed tomography (CT); however, catheter angiography is necessary to delineate segmental anatomy and complex vascular flow to tumors. ⁵ Accurate understanding of the vascular anatomy is necessary for optimal dosimetry. Angiography, in particular angiography combined with cone-beam CT, allows determination of which vascular beds are involved and require treatment, and subsequently what hepatic parenchymal volume will be within the treatment territory.

Table 1. Common hepatic artery variants.

Hepatic artery anatomy	Michel classification (%)	Hiatt classification (%)
Conventional anatomy	Type 1 (55%)	Type 1 (75.7%)
Replaced (or accessory a ) LHA from LGA	Type 2 (10%)	Type 2 (9.7%)
Replaced (or accessory a ) RHA from SMA	Type 3 (11%)	Type 3 (10.6%)
Replaced (or accessory a ) LHA and RHA	Type 4 (1%)	Type 4 (2.3%)
Accessory LHA from LGA	Type 5 (8%)
Accessory RHA from SMA	Type 6 (7%)
Accessory LHA and RHA	Type 7 (1%)
Replaced RHA and accessory LHA or replaced LHA and accessory RHA	Type 8 (2%)
Replaced CHA from SMA	Type 9 (3%)	Type 5 (1.5%)
Replaced CHA from LGA	Type 10 (3%)
CHA arising from aorta		Type 6 (0.2%)

Abbreviations: CHA, common hepatic artery; LHA, left hepatic artery; LGA, left gastric artery; RHA, right hepatic artery; SMA, superior mesenteric artery.

^a No distinction is made between replaced or accessory arteries in the Hiatt classification. ³

Determination of Safety: Extrahepatic Deposition

During planning angiography, attention should be paid to identifying vessels for potential nontarget radioembolization. Specifically, the right gastric artery, cystic artery, falciform artery, and gastroduodenal artery should be identified in all cases, and their positions noted relative to the location of the planned area of treatment. Prophylactic embolization can be considered, which will be discussed in detail later. Gastric uptake can be secondary to flow via the right gastric artery, gastroduodenal artery, an accessory left gastric artery from the left hepatic artery, or from branches of a replaced or accessory left hepatic artery arising from the left gastric artery. If gastric uptake is identified on the ^99m Tc-MAA, nontarget embolization needs to be differentiated from free technetium. ⁶ Free technetium can be determined if the uptake is seen symmetrically throughout the gastric mucosa and also within the thyroid glands and kidneys. ⁷ Nontarget embolization to the stomach presents with focal increased gastric uptake. Delay between infusion and postprocedure imaging leads to increased biodegradation and therefore increase free technetium.

Determination of Safety: Lung Shunt Fraction

Following the angiography and the ^99m Tc-MAA infusion, patients are transported to the nuclear medicine department for ^99m Tc-MAA scintigraphy to calculate the hepatopulmonary shunt fraction, also known as the lung shunt fraction (LSF). Historically, LSF should not exceed 20% of the delivered dose. Current recommendations for both resin and glass microspheres limit the delivered lung dose to less than 30 Gy per treatment and total lifetime lung dose to less than 50 Gy. ⁸ Lung doses in excess of these thresholds place the patient at an increased risk of radiation pneumonitis. ⁹ The LSF is used during dosimetry which will be described in subsequent articles.

Determination of Safety and Efficacy: Intrahepatic Deposition

^99m Tc-MAA scintigraphy also allows evaluation of where the radiotracer is deposited within the liver parenchyma to mimic the location for dose delivery. Using advanced dosimetry calculations based on MAA single-photon emission CT (SPECT/CT), optimal dose to the tumor and to the background parenchyma can be estimated. ¹⁰ Dosimetry models will be discussed in detail in subsequent articles.

Preprocedural Considerations

Potential candidates for locoregional therapy should be evaluated in interventional radiology clinic to determine the appropriateness of Y-90 and for consideration of other locoregional treatment options. ¹¹ Imaging (CT or MRI), baseline laboratory values (complete metabolic panel with focus on total bilirubin, albumin, and renal function, complete blood cell count, and Protime), patient history with functional status, and physical exam should be considered during patient evaluation. Imaging should be up-to-date, ideally within 1 month of the planning angiogram. Counseling should include procedural expectations, procedural risks, and likely outcomes and prognosis. Special attention should be paid to potential contraindications to Y-90, in particular, poor baseline functional status, advanced underlying liver disease, and advanced tumor status. Standard precautions should be followed for arterial access procedures, including the review of current medications such as anticoagulation and antiplatelet agents, potential coagulopathy and/or renal insufficiency. At the authors' institution, laboratory procedures are repeated on the day of the angiography planning so that treatments can be modified if there are acute changes in liver function.

Prior antitumoral medications and treatments should be thoroughly reviewed. Many chemotherapies are radiosensitizers, in particular gemcitabine, which can affect dosimetry with consideration of dose reduction. ¹² Additionally, certain medications, specifically vascular endothelial growth factor inhibitors, such as the monoclonal antibody bevacizumab (Avastin, Genetech, San Francisco, CA), have significant effects on both tumor vascularity and vascular fragility. Brown presented a case report of a patient taking bevacizumab who developed dissection and pseudoaneurysm during angiography planning. ¹³ It is the authors' practice to hold bevacizumab for 4 weeks prior to the angiography planning.

Procedure

Initial Steps

In the vast majority of patients, arteriogram planning can be performed under moderate sedation. Arterial access can be obtained through common femoral or radial artery access. A standard 5-Fr vascular sheath is sufficient for a complete hepatic arteriogram and potential embolization. The use of cone-beam CT during the procedure is highly encouraged and beneficial.

Diagnostic Angiogram

Upon the completion of the diagnostic angiography, the interventional radiologist should have a complete understanding of the hepatic arterial anatomy and specific tumoral arterial supply. Based on these findings, the operator should have identified treatment sites and have the information necessary to calculate the volume of tumor and background parenchyma that will be targeted. Additionally, any extrahepatic vessels with the potential for nontarget embolization should be identified.

While some operators begin the diagnostic arteriogram with an aortogram, ¹⁴ it is the authors' convention to begin with celiac artery angiography. Numerous 5-Fr catheters are available for celiac artery catheterization and the choice depends on the operator's preference and access site. The celiac angiogram should be continued through the portal venous phase to identify portal patency, direction of portal flow, and potential varices or shunts ( Fig. 1 ). In cases of aberrant vasculature, superior mesenteric catheterization and angiography may be necessary. Typically, aberrant vasculature is able to be identified on preprocedure MRI or CT, reducing the need to routinely perform aortic and superior mesenteric artery angiograms. Contrast saved by not performing aortography and superior mesenteric angiography can be better used on selective intrahepatic angiograms. ¹⁵

Fig. 1.

Arterioportal shunting. Right hepatic artery injection via a microcatheter demonstrating a prominent arterioportal shunt and brisk opacification of the portal system. Arrowhead—hepatic artery; arrow—portal vein.

Selective hepatic arteriograms are performed following selection with a coaxial microcatheter. Numerous options are available and can be chosen as per operators' preference. Considerations include the access site for microcatheter length and the pros and cons of microcatheter size. High-flow microcatheters offer the benefit of increased flow rates and improved imaging. Low-flow microcatheters may be necessary for certain embolic agents or super-selection into segmental arteries. Power injection–assisted angiograms are often necessary to obtain adequate injection flow and duration for optimal imaging, especially in cirrhotic patients with increased arterial flow. Care should be taken to obtain excellent diagnostic imaging, including attention to flow rate and contrast density to ensure that necessary vasculature is thoroughly investigated. Gaba provides a table of common power injector settings for adults. ¹⁵ Values commonly used at the authors' institution are listed in Table 2 .

Table 2. Common power injector settings.

Artery	Injection rate (mL/sec)	Total volume of contrast (mL)
Aorta	15–25	30–50
Celiac artery	5–8	25–50
Superior mesenteric artery	5–8	25–50
Common hepatic artery	3–5	12–20
Proper hepatic artery	3–5	12–20
Left hepatic artery	1–3	4–12
Right hepatic artery	1–4	4–16
Gastroduodenal artery	2–4	6–12
Left gastric artery	1–3	8–14
Right gastric artery	Hand injection	1–4
Segmental hepatic arteries	1–2	5–10

The level of catheterization and angiography depends on patient anatomy, location of the tumor(s) to be treated, and quality of imaging. At a minimum, imaging should be performed from the celiac artery, common and/or proper hepatic artery, and right and/or left hepatic arteries. Segmental and super selective angiography should be performed if the intended treatment location is more selective than a lobar infusion. Catheterization of extrahepatic arteries is necessary if there is suspected tumor supply.

Cone-beam CT provides invaluable additional information, especially important when using personalized dosimetry. ¹⁶ Cone-beam CT can be performed from the common/proper hepatic artery or right/left hepatic artery to aid in the identification of tumor-supplying vessels and guide subsequent catheterizations. Selective cone-beam CT from future infusion locations allows exact volumes to be calculated for each particular vascular distribution and to assess tumor coverage and flow dynamics ( Fig. 2 ). Particular attention to flow rates is important to prevent reflux into proximal vessels which would confound the imaging and not accurately represent the affected territory if Y-90 infusion were to occur at that location. Quality cone-beam CT imaging can ensure that the entirety of a tumor will be covered with future treatment, demonstrate branch vessels that would cause nontarget extrahepatic embolization, and hint that extrahepatic supply for the tumor bed exists.

Fig. 2.

Magnetic resonance imaging (MRI), angiography, cone-beam, SPECT/CT correlation. ( a ) Postcontrast T1 MRI with a LI-RADS 5 lesion in segment 5 demonstrating washout and pseudocapsule. ( b ) Right hepatic artery angiography showing angiographic correlate. Arrow—hypervascular lesion. ( c ) Selective cone-beam CT demonstrating coverage of the entire tumor and enhancing the surrounding background parenchyma that will fall in the treatment zone. ( d ) SPECT/CT confirming coverage of the lesion and treatment zone.

For lesions with multiple feeding vessels, treatment can occur proximally with a single dose covering the entire lesion, or with multiple selective infusion sites. Flow dynamics can be approximated using a hand injection with a 20-mL syringe of 50% contrast and 50% saline into the proximal artery. This can help identify whether an individual feeding artery will be a vascular sink and lead to nonuniform distribution of microparticles.

In addition to delineating anatomy and determining tumor vascular supply, angiograms should be examined for signs of a potential elevated hepatopulmonary lung shunt, such as arterioportal or arteriovenous shunting. If identified, there is potential for reduction of the shunt prior to ^99m Tc-MAA infusion. Techniques for elevated lung shunt management are beyond the scope of this article but include particle embolization, shunt embolization, modified dosing algorithms, and hepatic venous outflow balloon occlusion. ¹⁷

Prophylactic Embolization

Prophylactic embolization has fallen out of favor over time at high-volume centers ¹⁸ and most planning angiograms are completed without embolization. While previously performed in the majority of the planning procedures, embolization should be minimized and performed only when there is a risk of nontarget embolization. The sequence of angiography and embolization can be fluid. It is often helpful to perform all necessary angiograms prior to prophylactic embolization. This helps ensure that embolization will take place only when absolutely necessary. Additionally, there are billing and coding implications if embolization is performed prior to all diagnostic arteriograms.

Common arteries for potential nontarget embolization include the gastroduodenal artery, right gastric artery, and falciform artery ( Fig. 3 ). Embolization is commonly performed with micro-coils. The gastroduodenal artery is typically necessary only to embolize prior to lobar treatments when there is a short course of the proper hepatic artery which increases the risk of reflux from a right or left hepatic artery infusion back to the gastroduodenal artery. At the authors' institution, embolization is performed for any intrahepatic variant downstream of the injection site, or the gastroduodenal artery (rarely) or right gastric artery if the origin is in close proximity to the planned infusion site. On a technical note, the right gastric artery can be difficult to catheterize in some cases. An alternative option is to perform a retrograde catheterization via the left gastric artery ( Fig. 4 ).

Fig. 3.

Falciform artery embolization. ( a ) Selective middle hepatic angiography with prominent falciform artery coursing medially (arrows). ( b ) Selective left hepatic artery angiography following proximal coil embolization (arrow) of the falciform artery.

Fig. 4.

Right gastric artery embolization. ( a ) Following failure to catheterize the right gastric artery directly, the left gastric artery was catheterized (arrow) with subsequent microcatherization of the right gastric artery (arrowhead). ( b ) Successful coil embolization of the right gastric artery (arrow).

With regard to the cystic artery, attempts are made to choose an infusion site distal to its origin. However, if this is not possible, embolization of the cystic artery is not performed due to the risk of ischemic cholecystitis outweighing that of radiation-induced cholecystitis secondary to nontarget embolization. ¹⁹ If the cystic artery is likely to be within the treatment field, post-radioembolization antibiotic coverage should be strongly considered. At the authors' institution, all patients are prescribed a broad-spectrum antibiotic (e.g., levofloxacin) for a week following Y-90 radioembolization.

The risk of nontarget embolization secondary to reflux of particles is a function of the radioembolic delivered and the activity dose (a function of the number of particles delivered). Presently, there are two commercially available ⁹⁰ Y-impregnated microspheres. SIR-Spheres (SIR-Spheres; Sirtex Medical, North Sydney, Australia) are a biodegradable resin microsphere with ⁹⁰ Y adsorbed on to the surface. They range in size from 20 to 60 µm in diameter and have an activity of 50 Bq per sphere. TheraSphere (Boston Scientific Corporation, Boston, MA) are 20- to 30-µm glass particles with 2,500 Bq of ⁹⁰ Y activity integrated into the glass at calibration. ⁶ Given the lower density of activity per sphere, the risk of reflux is greater with resin microspheres given that more particles are required to deliver a similar dose. Likewise, resin microspheres have a greater embolic effect than glass for a given administration. ²⁰ In locations where there is a short segment between the infusion location and a possible source of nontarget embolization, antireflux devices can be employed as an alternative to permanent embolization. Multiple antireflux devices are available that are compatible with the radioembolic devices.

Unnecessary prophylactic embolization carries multiple risks. In addition to increased procedure time, contrast use, radiation exposure, and cost, there is also risk of technical failure or improper coil placement prohibiting the planned treatment. Additionally, there is risk of a change in the flow dynamics between the angiography planning and treatment procedure with the creation of new outflow patterns and unpredictable nontarget embolization ( Fig. 5 ).

Fig. 5.

Complication following gastroduodenal artery embolization. ( a ) Initial celiac imaging demonstrating conventional anatomy. ( b ) Celiac imaging following coil embolization with coil prolapse into the proper hepatic artery (arrowhead) and thrombosis of the proper and common hepatic artery (arrow).

^99m Tc-MAA Infusion Site

After the diagnostic angiograms are performed and any necessary embolization is performed, the catheter is repositioned for infusion of the ^99m Tc-MAA. ^99m Tc-MAA can be infused in one or multiple locations which can be lobar or segmental. The catheter tip should be distal to any vessel that could cause nontarget embolization; however, it should not be so selective that the entire tumor is not covered.

Attention should be paid to vascular branch points. If the catheter tip is near a branch point, the flow can be unpredictable secondary to small catheter movements from respirations and turbulent flow at bifurcations. Some operators prefer to infuse the ^99m Tc-MAA at the exact location of planned radioembolization, while others will choose a more proximal location to allow flexibility in dose administration location.

^99m Tc-MAA Injection

The final portion of the procedure planning is the administration of the ^99m Tc-MAA, typically between 1 and 5 mCi. ²¹ The ^99m Tc-MAA should be prepared as close to the infusion time as feasible to avoid radioactive degradation and dissociation which can lead to a spuriously elevated hepatopulmonary LSF. At the authors' institution, ^99m Tc-MAA is requested from nuclear medicine 30 minutes prior to the predicted infusion time. Aliquots of 2.5 mCi are delivered. A three-way stopcock is flushed in both directions using a 20-mL syringe of saline. The clean (non-sterile) syringe is given to the operator and attached to the stopcock. The stopcock is attached to the microcatheter and the catheter is flushed with saline. Next, the ^99m Tc-MAA is delivered with subsequent flushing with saline. The ^99m Tc-MAA should be infused at a slow and steady pace to prevent over-pressurization of the system and reflux and to mimic the infusion technique during the radioembolization. If there is a single infusion site, the microcatheter and diagnostic catheter are removed with stopcock attached and discarded into radioactive trash for decay. If a second infusion site is necessary, the stopcock is removed and discarded into the radioactive trash. Gloves are changed and the catheter is repositioned to the second infusion site where the ^99m Tc-MAA delivery is repeated. A second microcatheter is not necessary. Following the completion of ^99m Tc-MAA administration, the diagnostic catheter and sheath are removed, and hemostasis is achieved with manual compression or a vascular closure device.

^99m Tc-MAA Scintigraphy and Hepatopulmonary Shunt Calculation

At the authors' institution, ^99m Tc-MAA imaging is performed within 30 minutes of injection. Images obtained include planar images of the chest and upper abdomen as well as SPECT/CT of the upper abdomen. Images are obtained on a Symbia SPECT/CT (Siemens Medical Solutions USA, Malvern, PA). With regions of interest drawn around the lungs and liver, geometric means are calculated to determine lung and liver activity. The LSF is then calculated as follows:

Complications

There are few studies investigating complications arising during planning angiography for Y-90 treatment, with most studies focusing on posttreatment complications. ²² In 2017, Ahmed et al reported on 518 mapping or treatment angiograms performed in 180 patients. ²³ Thirteen complications were reported in 13 patients. Eleven of the 13 complications occurred during coil embolization of branch vessels to prevent nontarget radioembolization. Specific complications reported were coil migration ( n  = 6), arterial dissection ( n  = 2), focal perforation ( n  = 2), arterial thrombosis ( n  = 2), and vasospasm preventing further arterial selection ( n  = 1). Extrahepatic complications included groin hematoma ( n  = 1), external iliac artery dissection ( n  = 1), and femoral artery pseudoaneurysm ( n  = 1). A logistic regression analysis concluded that attempted arterial coiling was the most significant risk factor for complication during angiography planning (odds ratio [OR]: 7.8). This study further highlights the fact that coil embolization of enterohepatic vessels should be carefully considered and avoided if possible.

Alternative Strategies

There have been several reports of authors performing same-day mapping angiography and Y-90 treatment. ^{24 25 26} There are many reasons why this would be preferred, both by patients and physicians as the inconvenience and inefficiency of undergoing two procedures for a single is treatment is obvious. Patients are often anxious and weary of delaying their treatment by up to 3 weeks. Also, it is not uncommon for patients to travel great distances to undergo radioembolization thus adding a burden to both patients and their caregivers. A subset of patients undergoing Y-90 treatment have poor functional status, making it difficult to undergo two procedures. ²⁴ Patients with iodinated contrast dye allergy or chronic renal insufficiency would also be better served by having their planning and treatment on the same day. Undoubtedly, both patients and healthcare systems would benefit if both mapping and treatment could occur in a single session.

In 2016, Gates et al were the first to report on same-day mapping and radioembolization in a cohort of 14 patients. ²⁵ Of these, nine patients had hepatocellular carcinoma and five had liver metastases. Using preprocedure CT or MRI, volume analysis was performed to calculate the estimated volume of liver to be treated. Using a generic estimated LSF of 10% for patients with HCC and 5% for patients with metastases, dosimetry was performed and source vials ordered. Patients underwent angiography mapping using ^99m Tc-MAA and planar scintigraphy performed to calculate the true LSF which averaged 5.7% (range: 0.7–27.9%). Final dosimetry was performed and the patient was then brought back to the angiography suite for radioembolization. The mean total procedure time was 2.7 hours.

The same group reported on an expanded cohort which included 78 patients treated from 2008 to 2015 who underwent same-day mapping and radioembolization using glass microspheres (TheraSphere, Boston Scientific Corporation). ²⁴ As earlier, LSF was estimated at 10% for patients with HCC and 5% for patients with liver metastases and dosimetry was performed on prior cross-sectional imaging. After MAA scintigraphy, patients were brought back to the interventional suite and treated. In this study, 80% of patients had actual calculated LSF less than 10%. The remaining 20% of patients were treated after appropriate dose modifications. The median procedure time was 2.7 hours. The authors proposed a four-step algorithm for efficient same-day radioembolization.

Li et al described a same-day treatment paradigm using resin microspheres (SIR-Spheres; Sirtex Medical) in a cohort of 26 patients undergoing 34 procedures between February 2017 and January 2018. ²⁶ The key difference between the two microspheres from a dosimetry perspective is that vials of resin microspheres can be fractionated to adjust for dose change calculations unlike vials of glass microspheres. Predicted dosimetry was performed (either lobar or entire liver) prior to the procedure day using the body surface area method according to the manufacturer and as detailed by Padia et al. ⁶ Patients underwent diagnostic angiography and injection of ^99m Tc-MAA followed by planar scintigraphy to calculate the LSF and identify any extrahepatic deposition of the radiotracer. Next, the predicted treatment dose was modified as needed and prepared by the radiopharmacist while the patient was transported back to the angiography suite. Upon return, the hepatic artery was catheterized, and at the same position as the ^99m Tc-MAA injection Y-90 dose was administered. No posttreatment imaging was performed. The mean LSF was 4.7% with an average estimated lung dose of 2.4 Gy. No predicted doses were modified based on LSF. Total angiography time (planning and treatment) was 2.8 hours and total procedure time (including nuclear imaging) was 4.2 hours. Collectively, these studies demonstrate both the feasibility and practicality of performing same-day radioembolization for select patients.

Sancho et al retrospectively reviewed 532 consecutive patients evaluated for radioembolization of primary ( n  = 248) or metastatic ( n  = 284) liver tumors. ²⁷ A multivariate logistic regression model was created to compare patient-specific variables among those whose angiogram mapping led to the modification of the original treatment plan. Findings which would lead to changes in planned therapy included LSF greater than 20% and/or uptake of ^99m Tc-MAA in the gastrointestinal (GI) tract and mismatch between intrahepatic tumor distribution and ^99m Tc-MAA uptake. LSF greater than 20% was best predicted by the presence of a single tumor (OR: 2.4) and vascular invasion (OR: 5.5). Uptake in the GI tract was predicted by injection of tracer in the common or proper hepatic artery (OR: 4.7) and metastatic neoplasia (OR: 3.7). Tumor thrombus was the only variable associated with LSF greater than 20% and/or GI tract deposition (OR: 4.1). The authors concluded that pretreatment angiography planning using ^99m Tc-MAA is essential for safe and effective ⁹⁰ Y dosimetry. They also suggest that patients without tumor thrombus may represent a cohort of patients in whom same-day planning and treatment could be considered.

Conclusion

Radioembolization with ⁹⁰ Y microspheres is a mainstay of treatment for primary and secondary liver cancer. Unique among interventional radiology procedures, treatment consists of a two-stage approach, with arteriography planning following treatment (usually) at a later date. While much attention is rightly paid to the effects of the radioisotope delivery, optimal treatment relies heavily on the arteriogram planning. Additionally, personalized dosimetry relies heavily on ^99m Tc-MAA SPECT/CT imaging. There are multiple steps that must be performed and careful attention should be paid to the details for the identification of anatomic variants that could alter the dose delivered or lead to adverse events following treatment. As Benjamin Franklin said, “by failing to prepare, we prepare to fail.”