Additional hepatic ¹⁶⁶Ho-radioembolization in patients with neuroendocrine tumours treated with ¹⁷⁷Lu-DOTATATE; a single center, interventional, non-randomized, non-comparative, open label, phase II study (HEPAR PLUS trial)

In accordance with the most recent WHO/ENETS criteria, grade 1 and 2 neuroendocrine tumours (G1-/G2NET) are regarded as well- to moderately-differentiated tumours and grade 3 NET (G3NET) as poorly-differentiated NET or neuroendocrine carcinomas (NEC) [1, 2]. At diagnosis, 21% of all G1NET, 30% of all G2NET and 50% of all G3NET have distant metastases, of which the liver is most commonly affected [3, 4]. A correlation between the organ of origin and the likelihood of metastasis exists. For example, rectal NET has a slim chance of distant metastasis (5%) compared with pancreatic or colonic NET (respectively 64% and 53–86%) [3, 5]. Considering these numbers, many patients will be ineligible for curative treatment, which currently only includes surgical resection of the primary tumour.

Most G1-/G2NET have membrane receptors for somatostatin, allowing for targeted therapies, of which somatostatin-analogs are the most commonly used (e.g. octreotide). Treatment with somatostatin-analogs, chemotherapeutics and kinase inhibitors show only limited objective response rates in G1-/G2NET [6‐12]. In addition, systemic therapies give rise to systemic side effects. In the last decade, the treatment of G1-/G2NET with peptide receptor radionuclide therapies (PRRT) has increased. High objective imaging response rates (CR + PR 29–58%) [13‐17], clinical and biological response rates and a long median survival (95–128 months after diagnosis, 46 months after treatment) [13, 14] can be achieved after PRRT (Fig. 1).

All studies include high percentages of patients with liver metastases and show a dismal prognosis with increasing liver involvement (Table 1). As surgical resection techniques develop, some forms of hepatic involvement can be treated surgically. However, as three different patterns of hepatic metastases are described in NET, patients with the most common ‘diffuse pattern’ are not eligible for surgical resection (Table 2) [4]. Besides, most systemic therapies have a limited objective response rate. This indicates a need for improved treatment of extensive liver disease. In accordance with the ENETS guideline published in 2012, the treatment of choice in patients with NET liver metastases with a ‘diffuse’ or unresectable ‘complex pattern’, consists of systemic treatment followed by liver directed treatment [4]. Hepatic radioembolization (RE) is one of the liver directed treatments and it is an established minimal invasive treatment of patients with liver malignancies. RE has been demonstrated to be effective and well tolerated in primary, as well as secondary liver malignancies. A recent meta-analysis by Devcic et al. showed an average objective response rate (CR + PR) of 50% and an average disease control rate of 86% in a heterogeneous group of NET treated with RE [18]. ¹⁶⁶Ho-RE is quite similar to ⁹⁰Y-RE, but its distinct advantages will be discussed later on. Figure 2 shows an example of ¹⁶⁶Ho-RE in a NET patient.

In the current clinical setting, RE is used for liver dominant or liver isolated disease, often in a salvage setting. In this study, it is hypothesized that improved outcome can be obtained by escalating the treatment of liver metastases, the most significant prognostic factor for NET patients, additional to treatment of all extrahepatic disease in G1-/G2NET patients, by combining systemic PRRT with RE. In the presented study, patients with metastasized NET will receive PRRT in 4 cycles of 7.4 GBq with ¹⁷⁷Lu-DOTATATE, followed by ¹⁶⁶Ho-RE in the University Medical Centre Utrecht, the Netherlands, using ¹⁶⁶Ho-microspheres. The following paragraphs will address the details of the study.

Liver metastases significantly limit patient survival (Table 1). In the current study, the beneficial effect of additional ¹⁶⁶Ho-RE within 12 weeks after systemic ¹⁷⁷Lu-DOTATATE will be investigated. Combining these treatments may lead to an improved response rate for liver metastases, with acceptable and suppressible side effects, which may eventually lead to prolonged survival. Although the latter question is not an objective of the current phase 2 study, if significant efficacy and limited toxicity are shown, a subsequent phase 3 study might be initiated.

To date, Ezziddin et al. published the only report describing RE with ⁹⁰Y-microspheres after PRRT with ¹⁷⁷Lu-DOTATATE in a retrospective study [30]. They described a population of 23 patients in which RE was performed in a salvage setting. Patients had progressive or functionally uncontrolled disease after PRRT. Three months after RE, 30% had PR and 61% had stable disease without any serious toxicity, comparable with other reports on ⁹⁰Y-microspheres in NET patients. The authors concluded that salvage RE after PRRT shows a toxicity profile similar to RE alone, despite the high cumulative activity administrated. Less than 15% experienced a CTCAE grade 3 toxicity (abdominal pain, fatigue, fever, nausea and vomiting) and one patient developed a gastroduodenal ulcer. The interval between PRRT and RE was not mentioned, but patients were only referred for RE in case they had progressive disease (i.e. salvage setting). In their study, a cumulative liver dose of 2–12 Gy was described after PRRT. Due to the hypervascular nature of NET, the absorbed dose on healthy liver parenchyma due to RE is (far) below the presumed toxicity limit of healthy liver tissue (i.e. 70 Gy; and 50 Gy in cirrhotic livers with ⁹⁰Y resin microspheres) [31]. Thus, in theory, combining RE and PRRT can be safe. Nonetheless, concerns arise when implementing RE shortly after PRRT, due to the cumulative radiation dose and the short interval, potentially provoking REILD [28]. On the other hand in a recent case report, Filippi et al. described their treatment combination a patient with one hepatic metastasis, mesenteric metastasis and several bone metastases, diagnosed on a ⁶⁸Ga-DOTATATE-PET/CT [32]. The hepatic metastasis was downstaged with a lobar RE procedure, followed by 4 cycles of PRRT to treat extrahepatic lesions (a mesenteric metastasis and several bone metastases). Restaging after 3 months with a ⁶⁸Ga-DOTATATE-PET/CT showed a nearly complete remission of the extrahepatic metastases and an incomplete remission of the hepatic metastasis, thus another additional lobar RE procedure was performed to treat the hepatic lesion, with success [32]. They reported no significant adverse events of the combined treatments, complete symptomatic control and a survival of 42 months [32]. Additionally, they reported an absorbed dose on healthy liver parenchyma after the RE procedures of just 18 Gy and 20 Gy [32]. An example that the combination of RE and PRRT with short intervals can be safe.

In contrast to other studies, ^99mTc-macroaggregated albumin (^99mTc-MAA) will not be used as a scout dose for treatment planning. Instead, a small number of ¹⁶⁶Ho-microspheres will be used as a scout dose (approximately 250 MBq). This may overcome known limitations of ^99mTc-MAA: 1) differences in flow dynamics caused by the randomly shaped ^99mTc-MAA particles (90% between 10 and 90 μm) versus the spherically shaped ¹⁶⁶Ho-microspheres (30 ± 5 μm), 2) differences in scanning protocols of pre- and post-procedural imaging (i.e. ^99mTc-SPECT vs ⁹⁰Y-PET versus ¹⁶⁶Ho SPECT for both procedures), 3) employing a similar injection technique during both angiographies (i.e. bolus ^99mTc-MAA versus intermittent injection of ¹⁶⁶Ho microspheres), and 4) overestimation of the lung shunt using ^99mTc-MAA.

As shown by Elschot et al., in patients treated with ¹⁶⁶Ho-RE, using ^99mTc-MAA as well as ¹⁶⁶Ho-microspheres as pre-treatment imaging scout dose, ¹⁶⁶Ho-SPECT/CT was the most accurate in predicting the lung absorbed dose after ¹⁶⁶Ho-RE [33]. On ¹⁶⁶Ho-SPECT/CT a median lung shunt dose of 0.02 Gy was calculated. This was significantly overestimated in lung shunt dose calculations based on ^99mTc-MAA planar scintigraphy (5.5 Gy), ¹⁶⁶Ho planar scintigraphy (10.4 Gy) and ^99mTc-MAA SPECT/CT (2.5 Gy). An example of severe overestimation by ^99mTc-MAA compared to ¹⁶⁶Ho-microsperes is shown in Fig. 4. In the present study biodistribution / extrahepatic deposition assessment and lung shunt dose calculation will solely be evaluated by ¹⁶⁶Ho-SPECT/CT.

The safety of administrating 250 MBq of beta-emitting ¹⁶⁶Ho-microspheres as a scout dose has been studied by Prince et al. [25] They predicted the amount of extrahepatic activity and radiation absorbed dose, using ^99mTc-MAA SPECT/CT of 160 patients prior to ⁹⁰Y-RE. Based on a prior study by Kao et al., they defined a dose exceeding 49 Gy as clinically significant [25, 34, 35]. Simulating the use of 250 MBq ¹⁶⁶Ho-microspheres as a scout dose, only 1.3% of the patients had an extrahepatic deposition that could potentially be harmful (i.e. exceeded a mean dose of 49 Gy) [25]. Additionally, C-arm CT’s will be acquired at each injection position, prior to scout dose administration, to minimize the chance of extrahepatic depositions and partial tumour coverage, and to avoid extra angiography procedures [36, 37].

As mentioned in the ‘secondary objectives’ section, the application of the dual isotope SPECT/CT protocol will enable us to derive all relevant dosimetric parameters for treatment dose calculation. ^99mTc-phytacis, like other radiocolloids, is extracted from the blood pool by the reticuloendothelial cells of the liver [38]. Solely the functional liver parenchyma is depicted, due to absence of reticuloendothelial cells in tumours. Validation of this dual-isotope protocol will be performed in a side-study. Previous studies by Lam et al. have shown the prognostic value of the combination of ^99mTc-MAA and ^99mTc-sulphur colloid imaging, showing a tumour dose-response correlation and healthy liver dose-toxicity correlation [39, 40].

As an additional advantage, holmium is one of the 14 lanthanide elements, making ¹⁶⁶Ho-microspheres MRI-compatible for treatment imaging. In short, using estimated R₂* value changes from multi-gradient echo data, the holmium concentration per voxel could be determined [41, 42]. Conversion into units of activity enables dosimetric calculations. A prior study by Smits et al. has shown its feasibility in clinical practice and its comparability to SPECT-based dosimetry [21]. Using a similar MR-sequence, real-time imaging of the ¹⁶⁶Ho-microspheres during administration may become a future application [43]. However, the use of MRI is beyond the scope of this study, to minimize the study impact for patients.

In contrast to RE, PRRT’s main limitation is absorbed kidney dose. To reduce the dose, the most important preparatory measure is intravenous amino acid infusion prior, during and after ¹⁷⁷Lu-DOTATATE administration. To overcome the disadvantage of the unavoidable kidney dose, intra-arterial administration of ¹⁷⁷Lu-DOTATATE in patients with liver only disease could be a future application [44]. Several studies show a decreased absorbed kidney dose and an increased uptake in the liver metastases in most patients after intra-arterial administration. In a study by Pool et al., kidney absorbed dose decreased by 13% in addition to a 2.9-fold increase in liver metastases uptake [45]. In theory, a combination of intra-arterial PRRT and RE could be superior in patients with liver only disease.

Limitations of the study protocol are the small study cohort, non-comparative design, single center design and relatively short clinical follow-up for NET patients.

In conclusion, combining PRRT and RE could lead to improved treatment response and additional survival benefit: PRRT can be used to treat intra- and extrahepatic disease, whereas RE leads to an additional radiation boost on intrahepatic disease, the most incriminating factor in NET-patients’ survival. Based on this hypothesis, the HEPAR PLUS trial will include all patients treated with ¹⁷⁷Lu-DOTATATE with significant intrahepatic disease.