Holmium-166 radioembolization for the treatment of patients with liver metastases: design of the phase I HEPAR trial

The liver is a common site of metastatic disease. Hepatic metastases can originate from a wide range of primary tumours (e.g. colorectal-, breast- and neuroendocrine tumours) [1]. It is estimated that 50% of all patients with a primary colorectal tumour will in due course develop hepatic metastases [2]. Once a primary malignancy has spread to the liver, the prognosis of many of these patients deteriorates significantly. Potentially curative treatment options for hepatic metastases consist of subtotal hepatectomy or, in certain cases, radiofrequency ablation. Unfortunately, only 20-30% of patients are eligible for these potentially curative treatment options, mainly because hepatic metastases are often multiple and in an advanced stage at the time of presentation [3]. The majority of patients are therefore left with palliative treatment options.

Palliative therapy consists primarily of systemic chemotherapy. In spite of the many promising developments on cytostatic and targeted biological agents over the last ten years, there are still certain tumour types that do not respond adequately and the long-term survival rate for patients with unresectable metastatic liver disease remains low [4‐8]. Moreover, systemic chemotherapy can be associated with substantial side effects that lie in the non-specific nature of this treatment. Cytostatic agents are distributed over the entire body, destroying cells that divide rapidly, both tumour cells and healthy cells. For these reasons, a significant need for new treatment options is recognized.

A relatively recently developed therapy for primary and secondary liver cancer is radioembolization with yttrium-90 microspheres ( ⁹⁰Y-RE). ⁹⁰Y-RE is a minimally invasive procedure during which radioactive microspheres are instilled selectively into the hepatic artery using a catheter. The high-energy beta-radiation emitting microspheres subsequently strand in the arterioles (mainly) of the tumour, and a tumoricidal radiation absorbed dose is delivered. The clinical results of this form of internal radiation therapy are promising [9, 10]. The only currently clinically available microspheres for radioembolization loaded with ⁹⁰Y are made of either glass (TheraSphere ^®, MDS Nordion Inc., Kanata, Ontario Canada) or resin (SIR-Spheres ^®, SIRTeX Medical Ltd., Sydney, New South Wales, Australia).

Although ⁹⁰Y-RE is evermore used and considered a safe and effective treatment, ⁹⁰Y-MS have a drawback: following administration the actual biodistribution cannot be accurately visualized. For this reason, holmium-166 loaded poly(L-lactic acid) microspheres ( ¹⁶⁶Ho-PLLA-MS) have been developed at our centre [11, 12]. Like ⁹⁰Y, ¹⁶⁶Ho emits high-energy beta particles to eradicate tumour cells but ¹⁶⁶Ho also emits low-energy (81 keV) gamma photons which allows for nuclear imaging. As a consequence, visualization of the microspheres is feasible. This is very useful for three main reasons. Firstly, prior to administration of the treatment dose, a small scout dose of ¹⁶⁶Ho-PLLA-MS can be administered for prediction of the distribution of the treatment dose. This provides a theoretical advantage over ⁹⁰Y-RE, for which the distribution assessment depends on a scout dose of ^99mTc-MAA, with a disputable distribution correlation with the actual microspheres [13]. Secondly, quantitative analysis of the nuclear images would allow assessment of the radiation dose delivered on both the tumour and the normal liver (i.e. dosimetry) [14]. Thirdly, since holmium is highly paramagnetic, it can be visualized using magnetic resonance imaging (MRI). Quantitative analysis of these MRI images is also possible, which is especially useful for medium- and long-term monitoring of the intrahepatic behaviour of the microspheres [15, 16].

The pharmaceutical quality of ¹⁶⁶Ho-PLLA-MS has been thoroughly investigated and proven to be satisfactory [17‐19]. Multiple animal studies have been conducted in order to investigate the intrahepatic distribution (ratio tumour to normal liver), the toxicity profile/biocompatibility of the ¹⁶⁶Ho-PLLA-MS, safety of the administration procedure, and efficacy of these particles [20‐23].

Now that the preclinical phase of ¹⁶⁶Ho-RE has been successfully completed, we will start a clinical trial (the HEPAR study: H olmium E mbolization P articles for A rterial R adiotherapy) in order to evaluate ¹⁶⁶Ho-RE in patients with liver metastases. The main purpose of this trial is to assess the safety and toxicity profile of ¹⁶⁶Ho-RE. Secondary endpoints are tumour response, biodistribution prediction with ^99mTc-MAA versus a safety dose of ¹⁶⁶Ho-PLLA-MS, performance status, and quality of life.

The HEPAR trial is a phase I study to evaluate the safety and toxicity profile of ¹⁶⁶Ho radioembolization. Secondary endpoints are tumour response, biodistribution assessment, performance status, quality of life and comparison of the biodistributions of the ^99mTc-MAA scout dose and the ¹⁶⁶Ho-PLLA-MS safety dose.

With regard to the method of administration, viz. through a catheter placed in the hepatic artery, the in-vivo characteristics (no significant release of radionuclide), and the mechanism of action (local irradiation of the tumour), ¹⁶⁶Ho-PLLA-MS constitute a device analogous to the ⁹⁰Y microspheres, which are currently applied clinically. ¹⁶⁶Ho-PLLA-MS only differ in the radioisotope and the device matrix that are used. In a toxicity study in pigs on ¹⁶⁶Ho-RE, it has been demonstrated that (healthy) pigs can withstand extremely high liver absorbed doses, at least up to 160 Gy [23]. During these animal experiments, only very mild side effects were seen: slight and transitory inappetence and somnolence, which may well have been associated with the anaesthetic and analgesic agents that had been given and not necessarily with the microsphere administration. It is plausible that this low toxicity profile is caused by the inhomogeneous distribution of ¹⁶⁶Ho within the liver after intra-arterial injection, as was observed on MRI and SPECT images. The current study will investigate whether a similar distribution pattern can also be observed in human subjects and whether this inhomogeneous distribution is concentrated around the tumour sites.

Hepatic arterial injection with ^99mTc-MAA and subsequent scintigraphic imaging is widely used to predict the biodistribution of ⁹⁰Y microspheres, prior to the actual radioembolization procedure. Its accuracy can however be disputed. In our centre, we have observed that patients with a borderline lung shunt fraction of 10% to 19%, as calculated using the ^99mTc-MAA images (approximately 24% of all patients, all of whom were instilled a by 50% reduced amount of radioactivity), had no signs of lung shunting on post- ⁹⁰Y-RE Bremsstrahlung images. In these cases, it seems that the ^99mTc-MAA-scan had false-positively predicted extrahepatic spread. This may be explained by the fact that ^99mTc-MAA differs in many aspects from the microspheres that are used. Shape, size, density, in-vivo half-life, and number of ^99mTc-MAA particles do not resemble the microspheres in any way [13, 31]. In addition, free technetium that is released from the MAA particles can disturb the (correct) assessment of extrahepatic spread. We hypothesize that a small safety dose with low-activity ¹⁶⁶Ho-PLLA-MS will be a more accurate predictor of distribution than ^99mTc-MAA. The unique characteristics of ¹⁶⁶Ho-microspheres, in theory, allow a more accurate prediction of the distribution with the use of scintigraphy and MRI. In this study, we chose to perform both an injection with ^99mTc-MAA and administration of a safety dose of ¹⁶⁶Ho-PLLA-MS. The respective distributions of the ^99mTc-MAA and the ¹⁶⁶Ho-PLLA-MS safety dose will be compared with the distribution of the treatment dose of ¹⁶⁶Ho-PLLA-MS by quantitative analysis of the scintigraphic images.

Both commercially available ⁹⁰Y-MS products are approved by the Food and Drug Administration (FDA) and European Medicines Agency as a medical device and not as a drug. Radioactive microspheres are a medical device since these implants do not achieve any of their primary intended purposes through chemical action within or on the body and are not dependent upon being metabolized for the achievement of their primary intended purpose. In accordance with the definition of a medical device by the FDA and in analogy with the ⁹⁰Y-MS, we consider the ¹⁶⁶Ho-PLLA-MS to be a medical device [32]. The Dutch medicine evaluation board has discussed this issue (13 July 2007) and has concluded that the microspheres are indeed to be considered as a medical device.

One important issue concerning the resin-based SIR-Spheres ^® is the relatively high number of particles instilled (>1,000 mg), since this may sometimes be associated with macroscopic embolization as observed during the fluoroscopic guidance [28, 33]. Several authors have reported stasis of flow during administration of resin microspheres and were forced to end the procedure prematurely because of the risk of backflow, hence extrahepatic deposition of a part of the dosage [28, 34, 35]. The specific activity of the ¹⁶⁶Ho-PLLA-MS is considerably higher than that of the resin microspheres (≤450 and 50 Bq/microspheres, respectively). However, in order to obtain an equivalent absorbed dose, the total amount of radioactivity of the administered microspheres in ¹⁶⁶Ho radioembolization needs to be 3 times higher than in ⁹⁰Y radioembolization, due to the shorter physical half-life of ¹⁶⁶Ho. Even so, compared with the resin ⁹⁰Y microspheres, in ¹⁶⁶Ho radioembolization considerably less microspheres (≤600 mg) are used to obtain an equivalent radiation dose, resulting in a lower risk of stasis or backflow during administration [9, 29]. A further issue is that ⁹⁰Y microspheres can not be visualized under fluoroscopy during injection. Manufacturers of resin ⁹⁰Y microspheres state that their microspheres are to be administered with water for injection alternated with non-ionogenic contrast [36]. As a result, the operating physician cannot detect stasis or backflow of microspheres until he has switched from injecting microspheres to injecting the contrast agent. Holmium microspheres, on the contrary, are administered in a mixture of 50% saline and 50% non-ionogenic contrast under constant fluoroscopic imaging, which ensures constant control over the microspheres during injection [37]. However, continuous fluoroscopic imaging during microsphere administration may comprise an increased radiation dose delivered to the patient, specifically the abdominal skin, during the procedure.

If this phase I trial provides sufficient data to prove that ¹⁶⁶Ho-PLLA-RE has an acceptable safety and toxicity profile, further studies will be needed. The next step will be an efficacy study in a larger number of patients. The primary endpoints of that study will be tumour response and survival.