Regional radioactive treatments for hepatocellular carcinoma

The toxicities of chemotherapy and limited tolerance of the cirrhotic liver in patients with hepatocellular carcinoma (HCC) have led to the search for safer and more effective treatment modalities. Several types of hepatic arterial radioactive treatment have been described in recent years, of which the most encouraging appears to be ⁹⁰Yttrium (⁹⁰Y) microspheres. This is an outpatient treatment modality, requiring limited treatment cycles (one to two) and has minimal clinical toxicities compared with transarterial chemoembolization (TACE). Phase II study results have been encouraging, even in the presence of portal vein thrombosis. Future studies will be needed to compare this modality with TACE and to assess the usefulness of combing regional radioisotope therapy with systemic anti-angiogenics or cell cycle inhibitors, such as sorafenib. A predicted use for this regional therapy may be as adjuvant or neoadjuvant therapy with resection or liver transplantation.

Radionuclide therapy of HCC

Proliferation of transarterial and ablative techniques to treat HCC was driven by the lack of efficacy with traditional oncologic therapies in nonsurgical candidates. Radiation therapy utilizing external beam therapy has had poor outcomes, as the tolerance of normal liver is lower than the mean therapeutic dose required to achieve cytotoxicity within the targeted tumor(s). However, over the last 10 years, clinical interest in locoregional application of radioisotopes has accelerated. ⁹⁰Y has drawn the most interest. However, other radioisotopes are undergoing clinical evaluation.

Yttrium-90

Therasphere^® (MDS Nordion, Ontario, Canada) is a ⁹⁰Y impregnated microsphere that is US FDA approved for HCC. Therasphere is made of 20–30 µm glass microspheres embedded with ⁹⁰Y . ⁹⁰Y is a pure b emitter and decays to nonradioactive Zirconium-90 with a half-life of 64.1 h. The average β-energy emission is 0.9367 MeV with a mean tissue penetration of 2.5 mm. The activity to be infused is based on the mean target dose (approximately 120–150 Gy) and mass of liver to be treated. The latter is calculated from computed tomography or MRI volumetry. Prior to infusion of the therapeutic dose, patients undergo a mapping arteriogram where vessels to the GI tract are embolized to limit nontarget deposition of the microspheres, which is the principal toxicity of ⁹⁰Y radioembolization [1]. The final portion of the mapping arteriogram involves a simulation of therapy by injecting Technetium-99 into the target arteries. Patients undergo scintigraphy to determine if the embolization has prevented GI contamination and also to measure the fraction of the ⁹⁰Y dose that may end up shunting through the liver into the lungs. The lower radiation tolerance of the lungs makes this calculation crucial.

In an initial multicenter evaluation of 80 patients, individuals with Okuda I and Okuda II disease survived a median of 628 and 384 days, respectively [2]. One patient (1.2%) died of complications, possibly related to treatment. As this trial involved dose escalation, a proportion of the patients were treated with lower doses than the current standard of 120–150 Gy.

Kulik et al. described the outcomes of 35 patients with unresectable T3 HCC who underwent Therasphere infusion specifically to downstage to transplantation, resection or thermal ablation [3]. A total of 19 patients in this group (56%) were successfully downstaged to T2 status. Of these patients, eight underwent liver transplantation of which seven had finalized pathologic evaluation. Five of the seven had complete tumor necrosis. Of the two patients with residual tumor, one had only been treated 10 days previously, with tumor necrosis probably evolving. The other patient was listed due to downstaging despite there being a known untreated portion of the tumor remaining. An organ became available before the next cycle of therapy was administered. Median survival for the group was 800 days with 1-, 2- and 3-year survival of 84, 54 and 27%, respectively.

Patients with HCC and portal vein thrombosis are challenging to treat with other catheter-directed therapies, such as TACE. As the tumor is typically supplied by only the hepatic artery, whereas the underlying liver receives most of its oxygenated blood via the portal vein, any treatment involving embolization needs to be performed in a selective fashion via a hepatic artery branch, to avoid fulminant liver failure. By nature, radiotherapy with Therasphere is nonembolic. Recent evaluation of patients with HCC and portal vein thrombosis demonstrated a favorable toxicity profile and reasonable outcomes [4]. A total of 34 patients with portal vein thrombosis were reviewed; 12 patients had main portal vein involvement in this group. Median survival in the patients with branch portal vein thrombosis was 304 days. Survival dropped to 134 days in patients with main portal vein involvement. Given the limited toxicity, further evaluation in this patient population is warranted.

Recently, clinical trials have begun with the more embolic product, SIR-Spheres^® (Sirtex, North Ryde, NSW, Australia; selective internal radiation therapy [SIRT]) in HCC. No results are yet available but it also appears promising, as it combines radiation with embolization, provided the toxicity profile proves to be acceptable.

Iodine-131-lipiodol

Iodine-131 (¹³¹I)-lipiodol was one of the first radioisotopes utilized to treat HCC. Lipiodol is formed from the iodized ethyl esters of fatty acids from poppy seed oil. The iodine portion of lipiodol is changed to radioactive ¹³¹I via an atom exchange reaction. ¹³¹I-lipiodol emits γ-radiation of 364 MeV. The mean penetration is 0.4 mm and the physical half-life is 8.04 days. Given the aforementioned factors, patients need to be admitted for 10–14 days following therapy and need to have their thyroid ‘blocked’ prior to, and following, therapy to prevent uptake of the radioisotope by the thyroid gland.

Even though ¹³¹I-lipiodol has been more thoroughly investigated than other radioisotopes outside of ⁹⁰ Y, reported outcomes remain limited. Response rates range from 17 to 92%. An early cohort of patients with HCC had a median survival of 6 months. Outcomes significantly worsen with tumors greater than 5 cm in diameter. In a small, randomized trial (27 patients), ¹³¹I-Lipiodol was superior to systemic chemotherapy in Okuda I or II patients with portal vein thrombosis with 48 versus 0% 6-month survival. A randomized prospective trial of ¹³¹I-Lipiodol versus chemoembolization was also reviewed [5]. This trial demonstrated no difference in survival with ¹³¹I-Lipiodol versus oily chemoinfusion followed by embolization when both were performed from the proper hepatic artery. It is worth noting that in current practice, chemoembolization from the proper hepatic artery is in opposition to the standard of care [6].

Rhenium-188-Lipiodol

Based on the potential of ¹³¹I-lipiodol, another group of investigators studied the potential of rhenium-188 (¹⁸⁸Re)-lipiodol [7]. Suggested advantages of ¹⁸⁸Re include greater β-energy and a shorter half-life, which limits the required hospitalization time. Additionally, in comparison with ¹³¹I, the device is cheaper and simpler to produce. For these reasons, ¹⁸⁸Re may be preferable in regions with limited financial resources where sophisticated substances, such as ⁹⁰Y, are cost prohibitive. ¹⁸⁸Re is produced by ion exchange from tungsten-188. There is only one study with this agent, which is a multicenter trial focused on recruiting patients from different countries [7]. Unfortunately, data collection prior to, and following, treatment was of poor quality. This weakness was attributed to limited resources at some of the treating centers. In total, 185 patients were treated with a dose designed not to exceed 1.5 Gy to marrow, 30 Gy to liver or 12 Gy to lung. There were large variances by country in the percentage of patients with cirrhosis (0–80%) and degree of hepatic dysfunction. Of these patients, 28 (15%) died within 2 months of treatment; approximately two-thirds of these deaths were attributed to tumor progression. Post-treatment CT scans were obtained in less than 50% of patients, making accurate response measurement impossible. Estimated median survival was 256 days with a range among countries of 175–518 days. The potential benefits of this agent make it worth further evaluation with proper oversight and data tracking.

Holmium-166

Kim et al. recently reported outcomes following percutaneous injection of holmium-166 (¹⁶⁶Ho) bonded with a chitin byproduct (chitosan) into small HCCs [8]. Preclinical research using a melanoma model demonstrated tumor necrosis in the area adjacent to the needle, with growth arrest in more distant tissues. Chitosan transforms from a liquid to a gel at a pH greater than 6, limiting efflux of the agent following injection. ¹⁶⁶Ho results from neutron activation of holmium-165 (¹⁶⁵Ho) and emits predominantly β-radiation (maximum energy [E_max]: 1.84 MeV) with a mean penetration of 2.2 mm and a half-life of 26.8 h. It also emits some γ-ray photons (6.2%) which allows scintillation imaging after device injection. Potential advantages of this agent include its natural abundance and conversion from ¹⁶⁵Ho to ¹⁶⁶Ho is simple.

In their Phase IIb trial, Kim et al. treated 40 patients with smaller than 3 cm HCC. Target dosimetry was 20 mCi/cm tumor diameter [8]. The entire tumor was treated in one visit using a 21-gauge multiside needle similar to percutaneous ethanol ablation. The median tumor diameter was 2.45 cm and median injected dose was 50 mCi. At initial response assessment 2 months after treatment, 31 out of 40 (77.5%) patients had complete tumor necrosis. As with most ablative therapies, success rates were higher for smaller tumors, with complete necrosis in 11 out of 12 (91.7%) tumors smaller than 2 cm at initial follow-up. The 1- and 2-year accumulated local progression rates of treated patients were 18.5 and 34.9%, respectively. Toxicity was limited, with two patients requiring therapy for transient bone marrow depression. Although more research is clearly needed, this therapy is intriguing. With modification and evolution of the technique, it may be possible to treat small HCCs with a single injection using a small needle, overcoming limitations with ethanol injection and thermal ablation, respectively.

Phosphorus-32

A final reported radioisotope for the treatment of HCC is phosphorus-32 (³²P) glass microspheres. ³²P was formed from non-radioactive phosphorus-31 (³¹P) by a nuclear chemical reaction. The half-life of ³²P is 14.28 days. It is principally a β-emitter with E_max of 1.711 MeV/disintegration with an average soft tissue penetration of 3.2 mm.

In the only reported human case–controlled trial, the 46–76 µm ³²P spheres were soaked into an absorbable gelatin sponge and packed in the resection bed after hepatectomy for HCC in 29 patients, while 38 other resection patients served as controls [9]. The two groups of patients were similar in age, tumor size, microvenous invasion and other evaluated characteristics. At 6 months, 1, 2 and 3 years, recurrence in the ³²P-implanted group was significantly lower than the controls. Survival was similarly significantly longer in the 32P-implanted group at 1, 2 and 3 years. This treatment merits further prospective investigation with possible intra-arterial evaluation as well.

Our experience with Therasphere (⁹⁰Y) for unresectable HCC

We were the first group to evaluate Therasphere in the USA in HCC clinical trials, in a 65-patient biopsy-proven cohort [10,11]. In the absence of guiding literature on the effects of ⁹⁰Y on the cirrhotic liver (85% of HCC patients), we gave ourselves conservative guidelines of treating only one liver lobe in a session, as we had done previously for TACE (there were two patient exceptions, who developed HCC recurrence after prior resection), as well as the unusual rule of giving only one Therasphere treatment, even in the presence of bilobar tumors, unless later computerized axial tomography (CAT) scans showed tumor progression. Inclusion criteria also were: bilirubin less than 2.0 mg/dl, creatinine levels of less than 1.5, white blood cell count over 3500, platelet count over 60,000 and tumor unresectability.

Using these criteria, the tumor responses, using WHO criteria were: complete response in two patients, partial response in 19 patients (total response 21 out of 65 [32%]), minor response in 12 patients, stable disease in 15 patients and tumor progression in 18 (28%) patients. Two interesting observations made these responses different from those with TACE. First, they occurred slowly and continued for more than a year following a single treatment. Second, although these responses are for changes in tumor size, vascular responses (as judged by pixel intensity on the arterial phase of the triphasic CAT scan) occurred in 70% of patients in our series. This clearly begs the question as to how Therasphere works and whether its actual target is the vasculature into which it is directly injected, with tumor size changes being a secondary consequence of this. Overall survival for the Cancer of Liver Italian Program (CLIP) 0 patients (normal liver function: n = 42) was 649 days and for CLIP 1–2 patients (n = 23) was 302 days [12]; 70% of these patients had one treatment only. Interestingly, several patients developed late atrophy of the treated lobe, so far without clinical sequela. However, the significance of this for subsequent resection surgery is currently unknown.

Myelosuppression was insignificant; however, 83% of patients developed more than 50% decreased peripheral lymphocyte counts. This started within 24 h and lasted often up to 2 years. So far, there have been no immune or infectious clinical complications. Such patients who go on to subsequent liver transplant should therefore be treated with decreased doses of immunosuppressants. The main significant toxicity was elevation of bilirubin. A more than 100% increase occurred in 25 out of 65 patients within 3 months of treatment, of whom 13 had a subsequent return to baseline levels (transient) within 3 months of the peak. This left 12 patients out of 65 (18%) with prolonged bilirubin elevation post-therapy. Nausea, emesis and abdominal pain, which are frequent after TACE, were uncommon [12–14].

This therapy appears attractive for advanced, unresectable HCC, both for the relatively benign toxicity profile and the minimal number of treatments and thus hospital visits. A prospectively randomized, controlled clinical trial comparing single-dose Therasphere with repetitive cisplatin–TACE suggests equivalency of responses on preliminary analysis [15].

Five-year view

Future clinical trials are needed to establish the place of regional radioactive treatments in the therapeutic armamentarium. In particular, long-term follow-up of case cohort and trial series is needed to know the real value of these treatments on survival and on the natural history of the cirrhotic liver. In particular, it needs to be determined:

• • Whether these treatments are additive or synergistic with the new systemic anti-angiogenics and cell-cycle inhibitors. The latter appear to enhance survival (sorafenib) without significant responses, whereas ⁹⁰Y clearly results in tumor responses. It would seem logical to evaluate the combination of the two treatment modalities;

• • Are these regional therapies safer than and as effective as more traditional chemotherapies?

• • Will they have a role as neoadjuvant or adjuvant therapy alongside surgical resection of HCC, considering the high recurrence rates from resection in most series?

• • Will these regional therapies serve as a bridge to liver transplantation and decrease the drop-out rate in patients awaiting donor organs, as a result of tumor progression?

Key issues

• • Randomized controlled trials are needed to compare regional radioactive treatments with other treatments, especially transarterial chemoembolization.

• • The number of treatments required for individual patients is currently unclear.

• • Safety of the treated liver to subsequent surgical resection is not known.

• • Can survival be enhanced by combining regional radioactive treatments with the newer anti-angiogenics or kinase inhibitors?

• • What is the liver target of this therapy? Is it the tumor or the tumor vasculature into which it is actually injected?

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.