Radioembolization for neuroendocrine tumors: procedure, application and clinical outcomes

Abstract

Neuroendocrine liver metastases significantly affect patient prognosis and quality of life due to their symptomatic burden and challenging management. Besides conventional systemic therapies, liver-directed therapies improve patient outcomes in patients with liver-dominant disease. These liver-directed therapies have gained interest over the past decade, but their placement in the treatment algorithm of neuroendocrine liver metastases remains largely unclear. The purpose of this review is to evaluate the current role of selective internal radiation therapy (radioembolization) as a treatment for neuroendocrine liver metastases. This review examines the patient selection, procedural aspects, applications, and clinical outcomes. Radioembolization is effective as a standalone treatment. This treatment achieves disease control rates exceeding 90% and improves symptoms and quality of life. Moreover, combining radioembolization with systemic therapies may provide improved treatment response and additional benefits, but further investigation is required. The treatments effectiveness is influenced by appropriate patient selection, including consideration of liver function, tumor vascularity and previous interventions. A multidisciplinary approach is essential in assessing treatment eligibility. Patient management should be tailored on an individual level to optimize outcomes. The incidence of complications is rare (<1%), with radiation-induced liver disease being the most concerning. This review underscores the need for continued research to better understand the optimal use of radioembolization. Specifically, its placement within treatment, particularly in combination with other therapies, requires further exploration, ultimately to improve survival and quality of life for patients with neuroendocrine liver metastases.

Background

Neuroendocrine tumors (NETs) represent a large heterogeneous collection of neoplasms that arise from neuroendocrine cells. The most commonly affected organs are the gastroenteropancreatic tract or the lungs (Yao et al. 2008, Hallet et al. 2015). The incidence of NETs has increased with more than six-fold increase between 1973 and 2012 (Dasari et al. 2017). This increase is partly due to improved diagnostic techniques and heightened awareness (Dasari et al. 2017). Approximately 25% of patients have distant metastases at the time of initial diagnosis (Yao et al. 2008), often located in the liver (Pavel et al. 2012). Clinical presentation depends on the site of the primary tumor, the presence of metastases and whether a NET is ‘active’ (i.e., can secrete hormones). Most patients experience symptoms due to mass effect or hormone secretion, while others remain asymptomatic (Pavel et al. 2012, 2020). Effective management of neuroendocrine liver metastases (NELMs) is crucial as they significantly impact the patient’s overall survival (OS) (Dasari et al. 2017) and prognosis (Rindi et al. 2007, Yao et al. 2008, Frilling et al. 2010). The hormonal secretion caused by functional NELMs drastically affects morbidity and mortality of these patients. Despite advances in the detection of these tumors, treatment remains challenging, particularly for patients with unresectable or treatment-refractory (i.e., persisting hormone-related symptoms regardless of treatment with, e.g., high dose somatostatin analogs (SSA) or telotristat) liver disease.

Local treatment options

Liver-directed treatment for NELMs can be roughly divided into two categories: ablative localized treatments or trans-arterial treatments. Ablative localized treatments include radiofrequency ablation (RFA) or microwave ablation (MWA). Trans-arterial treatments encompass trans-arterial (bland) embolization (TAE), trans-arterial chemoembolization (TACE) and trans-arterial radioembolization (also known as selective internal radiation therapy; SIRT). Trans-arterial treatments use the dual blood supply of the liver to their advantage; hepatic tumors are predominantly vascularized by the hepatic artery (80–100%), while the normal liver parenchyma is predominantly vascularized by the portal vein (70%) (Braat et al. 2015). NELMs are often good candidates for trans-arterial treatment due to their hypervascularity. It is important to note that the therapeutic effect of radioembolization primarily results from the delivery of targeted radiation to tumor, rather than from an embolic effect, distinguishing it from other trans-arterial therapies. To date, retrospective comparative studies have not demonstrated the superiority of one trans-arterial treatment over another in terms of objective response rates (ORRs) or survival. However, differences in toxicity rates, patient selection and logistics have been observed.

Short-term toxicity is generally less frequent with radioembolization than with TAE or TACE. The incidence of post-embolization syndrome (a combination of nausea, abdominal discomfort and fatigue) was lower after radioembolization (Chen et al. 2017). Despite these findings, laboratory investigations did not show any significant differences between the three trans-arterial treatments. Regarding patient selection, patients who have undergone a Whipple resection along with biliodigestive anastomosis are less suitable for TACE or TAE (Braat et al. 2024). These patients are more prone to biliary ischemia, posing a risk for infection or abscess formation by gastrointestinal bacteria. This results from the absence of a physical barrier, resulting in retrograde biliary colonization (Kim et al. 2001, Woo et al. 2013, Braat et al. 2015). On the contrary, the occurrence of this complication is lower after radioembolization (Devulapalli et al. 2018). Cholapranee et al. (2015) reported liver abscess in more than 20% of patients in the TACE group despite aggressive antibiotic prophylaxis, compared to none in the TARE group. This lower incidence of severe adverse events was also observed in a systematic review comparing TACE (11.1%) vs TARE (5.9%) (Victory Srinivasan & Venugopal 2023). In addition, radioembolization offers logistical advantages. Radioembolization can typically be completed in a single treatment session, whereas TACE or TAE may require multiple treatments.

The aim of this review is to present an updated overview of the procedure, application and clinical outcomes of radioembolization as treatment for NELMs.

Patient selection

Proper patient selection is the key to improvement of outcomes. Radioembolization can be considered in patients with liver-only or liver-dominant disease. While the definition of liver-only disease is logical, there is no internationally accepted definition for liver-dominant disease. Given the multifactorial nature of treatment consideration, this decision rightfully resides in a multidisciplinary tumor board. Patients must have an acceptable clinical performance status (ECOG 0–2) and sufficient liver function to be considered eligible. More specific parameters can be found in Table 1. To date, no definitive evidence has been generated to select patients on NET grade or other NET-specific parameters (Braat et al. 2019). Besides these clinical parameters, several technical parameters need to be taken into account; these will be discussed in the next sections.

Table 1.

Work-up for radioembolization.

	Minimal work-up		Additional work-up
Clinical assessment	Patient level of functioning	ECOG performance score	Fibroscan/gastroscopy	Assess esophageal varices
	Signs of hepatic dysfunction	Child-Pugh or MELD score
	NET hormone-related symptoms	Assessment of clinical response
Blood work	Liver function	Bilirubin, ALP, AST, ALT, albumin	Hb, hematocrit, WBC, platelets	Assess anemia or bleeding disorders
	Kidney function	Creatinine, eGFR; for angiography
	Tumor markers	Assessment of biochemical response
	Coagulation assessment	E.g., prothrombin time or INR; for angiography
Imaging studies	GdMRI/contrast-enhanced CT	Assess intrahepatic tumor load	SSTR-PET-CT	Assess total body tumor load
	Early contrast-enhanced CT	Assess liver arterial vascularization	FDG-PET-CT	Assess tumor grade distinction and exclusion aggressive disease

Abbreviations: ECOG, Eastern Cooperative Oncology Group; NET, neuroendocrine tumor; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine; eGFR, estimated glomerular filtration rate; CgA, chromogranin A; INR, international normalized ratio; GdMRI, gadolinium-enhanced magnetic resonance imaging; CT, computed tomography; Hb, hemoglobin; WBC, white blood count; SSTR, somatostatin receptor; PET, positron emission tomography; FDG, fluorodeoxyglucose.

Procedure

Figure 1 shows a diagram illustrating the radioembolization process, including the work-up imaging, MAA/scout procedure and actual treatment.

Figure 1.

The step-by-step process of radioembolization. Abbreviations: SPECT, single-photon emission computerized tomography; CT, computed tomography.

Work-up and scout procedure

Radioembolization is a multidisciplinary procedure that involves the expertise of an interventional radiologist and nuclear physician. Work-up includes clinical assessment, blood tests and several imaging studies. In particular, an arterial phase contrast-enhanced CT is performed to gain information on the arterial vascularization of the liver and its tumors, e.g., to help identify anatomical variants. This knowledge is important for the interventional radiologist during the preparation phase as it reduces procedure time and thus limits radiation exposure to patients and medical personnel (Uliel et al. 2012, Kim et al. 2014).

A scout procedure (i.e., treatment simulation) will be performed before the actual treatment. A femoral or radial approach can be used for the angiography. The liver and tumor vasculature are evaluated using digital subtraction angiography. Additional periprocedural cone beam or angio CT scans are recommended as they allow the interventional radiologist to verify complete tumor coverage during the procedure and to identify any arteries supplying extrahepatic tissues, referred to as culprit vessels (Braat et al. 2015). Once these vessels are identified, the microcatheter is carefully adjusted to a position distal to the culprit vessels to prevent extrahepatic deposition of radioactive microspheres. Alternatively, coil embolization of the culprit vessel may be employed in certain cases, e.g., when immediate blockage of a vessel like the right gastric artery is necessary to prevent gastric deposition of activity. There has been a general trend toward preserving the original vascularization, rather than embolization, to avoid additional collateral formation (Powerski et al. 2012, Reinders et al. 2018, Keane et al. 2024). Scout particles are administered during the scout procedure once the interventional radiologist determined a safe injection position(s). Scout particles are either technetium-99m macroaggregated albumin (^99mTc-MAA) or a small dose of holmium-166 microspheres (¹⁶⁶Ho-scout dose, QuiremScout®, Terumo EU, Belgium). SPECT-CT imaging is performed after the particle administration and the closure of the femoral or radial access route to assess the distribution of particles and exclude extrahepatic depositions of activity in the abdomen (via culprit vessels) or ‘lung shunting’. The latter is a situation when particles pass through physiological or tumor-induced arteriovenous shunts and accumulate in the lungs. If a hepatopulmonary shunt is present, the estimated particle deposition based on the scout procedure must be calculated to ensure safety.

Radioembolization treatment

During a separate second angiography, therapeutic particles are administered from similar injection positions, as determined during the scout procedure. These particles are either yttrium-90 (⁹⁰Y)-loaded glass microspheres (Theraspheres®, Boston Scientific, USA), yttrium-90 (⁹⁰Y) resin microspheres (SirSpheres®, SIRTex Medical, Germany) or ¹⁶⁶Ho-microspheres (QuiremSpheres®, Terumo). Post-treatment imaging consisting of ⁹⁰Y-PET-CT (or ⁹⁰Y-Brehmstralung SPECT-CT) or ¹⁶⁶Ho-SPECT-CT (or ¹⁶⁶Ho-MRI) is performed to confirm the correct distribution of the particles in the liver and the absence of extrahepatic particles.

Prophylactic medication

The use of prophylactic medication for radioembolization in NELM is now considered outdated. A single-center retrospective review including NET patients treated with ⁹⁰Y between 2014 and 2020 found that combination treatment with prednisolone and ursodeoxycholic acid as prophylaxis did not reduce toxicity (Braat et al. 2023). Evidence supporting prophylactic medication in general or even for patients with a biliodigestive anastomosis is lacking (Devulapalli et al. 2018, Braat et al. 2023). Use of somatostatin analogs can be continued during the scout procedure and treatment (Kim et al. 2018) (NCT02859064). Additional use of periprocedural intravenous octreotide lacks evidence.

Dosimetry

Radioembolization requires meticulous preparation through its scout procedure, but it also holds the potential to become a patient-tailored treatment using dosimetry. Assessment of estimated absorbed dose in tumors and normal liver tissue allows for safely increasing the administered activity. This assessment improves the likelihood of inducing an objective response while maintaining safety. This concept was supported in patients with hepatocellular carcinoma (DOSISPHERE-1 study) (Garin et al. 2021) and applies to NELMs as well. Reported clinical data is hampered by the lack of proper dosimetry, thus outcomes are likely to improve nowadays due to the broader adoption of prospective dosimetry in clinical practice and future clinical studies (Keane et al. 2024). Ramdhani et al. (2024) analyzed dose–effect relationships of ¹⁶⁶Ho-radioembolization in NELMs, identifying a clear dose–response relationship on post-treatment ¹⁶⁶Ho-SPECT/CT. A tumor absorbed dose (D_t) of ≥120 Gy resulted in a >90% likelihood of obtaining an objective response according to the Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST 1.1). Ebbers et al. (2022a) were the first to establish the dose–survival and dose–response relationship in patients treated with glass ⁹⁰Y-radioembolization at D_t ≥150 Gy on post-treatment ⁹⁰Y-PET/CT. This relationship was recently confirmed by Watanabe et al. (2024) with a median estimated absorbed dose of 196 Gy per lesion on ^99mTc-MAA SPECT/CT. Chansanti et al. (2017) established this for resin ⁹⁰Y-radioembolization at an estimated D_t ≥191 Gy on ^99mTc-MAA SPECT/CT. Doyle et al. (2024) reported similar results with an estimated D_t ≥120 Gy on ^99mTc-MAA SPECT/CT, resulting in objective tumor response. All of the above studies support that a higher D_t results in an increased likelihood of obtaining an objective response after radioembolization.

None of these studies showed a dose–toxicity correlation. However, Ramdhani et al. observed a lower number of toxicities with normal liver absorbed dose (D_h) < 30 Gy versus D_h ≥30 Gy with ¹⁶⁶Ho. This finding aligns with the D_h limit previously reported in a post-hoc analysis of the phase 2 HEPAR PLuS study (Stella et al. 2023).

Radioembolization in a salvage setting

Three large multicentric studies retrospectively investigated radioembolization as salvage therapy for patients with NELMs, confirming its safety and efficacy in over 1,000 patients (Table 2) (Braat et al. 2019, Schaarschmidt et al. 2022, Wong et al. 2022). An international multicentric study including 244 patients with NELMs assessed disease control rate (radiologic response) at 3 and 6 months after treatment (Braat et al. 2019). Radioembolization achieved complete response in 2%, partial response (PR) in 14%, stable disease in 75% and progressive disease in only 9% according to RECIST 1.1 (Eisenhauer et al. 2009). This represents a disease control rate of more than 90%, independent of NET grade. In an older patient cohort, Kennedy et al. (2008) reported even higher rates of disease control of >95%. Figure 2 illustrates an example of disease control over the course of several months.

Table 2.

Overview of included studies with patient characteristics, complication rates, response assessment and survival (in months).

Author(s) (year)	Study design	Number of patients	Treatment	Population characteristics	Grade NET	Complication rate	Response
Soulen et al. (2024)	Single-center prospective/retrospective study	n = 37	Radioembolization with 90Y resin + CapTem	Primary tumor - Pancreas 43% - Gut 19% - Other 38%	- 2 (100%)	- REILD NR - Ulcera NR - Radiation pneumonitis NR - Radiation pancreatitis NR - Radiation cholecystitis NR	ORR 72% PFS 36 OS 41
Schaarschmidt et al. (2022)	Multicenter retrospective study	n = 297	Radioembolization with 90Y glass or resin	Primary tumor - Pancreas 24.9% - Small bowel 31.1% - Other 44%	- 1 (25.6%) - 2 (50.5%) - 3 (5.7%) - Unknown (9.1%)	- REILD NR - Ulcera NR - Radiation pneumonitis NR - Radiation pancreatitis NR - Radiation cholecystitis NR	ORR 41.3% PFS 15.9 OS 30.6
Wong et al. (2022)	Multicenter retrospective study	n = 170	Radioembolization with 90Y resin	Primary tumor - Pancreas 24% - Midgut 36% - Other 40%	- 1 (70%) - 2 (15%) - 3 (15%)	- REILD NR - Ulcera NR - Radiation pneumonitis NR - Radiation pancreatitis NR - Radiation cholecystitis NR	ORR 36% PFS 25 OS 33
Braat et al. (2020)	Single-center prospective study	n = 34	Radioembolization with 166Ho	Primary tumor - Pancreas (32%) - Ileum or jejunum (29%) - Other (39%)	- 1 (35%) - 2 (65%)	- REILD 3% - Ulcera 0% - Radiation pneumonitis 0% - Radiation pancreatitis 0% - Radiation cholecystitis 0%	ORR 60% PFS NR OS NR
Braat et al. (2019)	Multicenter retrospective study	n = 244	Radioembolization with 90Y resin	Primary tumor - Pancreas 31.2% - Small bowel 34.9% - Other 33.9%	- 1 (39.3%) - 2 (35.7%) - 3 (10.2%) - Unknown (14.8%)	- REILD 0.8% - Ulcera 2.8% - Radiation pneumonitis 0.4% - Radiation pancreatitis 0% - Radiation cholecystitis 0%	ORR 15.7% PFS NR OS 31.2
Chen et al. (2017)	Multicenter retrospective study	n = 155	Radioembolization with 90Y resin or glass	Primary tumor - Pancreas 40% - Gastrointestinal 47% - Other 13%	- 1 (50%) - 2 (39%) - 3 (11%)	- REILD 4.5% - Ulcera 0.006% - Radiation pneumonitis NR - Radiation pancreatitis NR - Radiation cholecystitis NR	ORR NR PFS 15.7 OS 48.2
Kennedy et al. (2008)	Multicenter retrospective study	n = 148	Radioembolization with 90Y resin	Primary tumor - Pancreas 19% - Small intestine 67% - Other 14%	NR	- REILD 0% - Ulcera NR - Radiation pneumonitis NR - Radiation pancreatitis NR - Radiation cholecystitis NR	ORR 63.2% PFS NR OS NR

Abbreviations: ⁹⁰Y, yttrium-90; CTCAE, Common Terminology Criteria for Adverse Events; ORR, objective response rate; PFS, progression-free survival; OS, overall survival; NR, not reported.

Figure 2.

¹⁶⁶Ho-radioembolization in a 70-year-old patient with liver metastases of a grade 2 insulinoma. Progressive liver-only disease (after pancreatic tail resection) with clinically multiple hypoglycemic episodes daily. Previous treated with on somatostatin analogs, followed by [¹⁷⁷Lu]Lu-DOTATATE (prematurely ceased after three cycles of 7.4 GBq due to grade 2–3 pancytopenia). (A) Pre-treatment arterial phase contrast-enhanced CT (CECT) showing bilobar hypervascular metastases (largest lesion is 40 mm in maximum diameter). (B) [^99mTc]Tc-MAA SPECT-CT and (C) maximum intensity projection (MIP) of [^99mTc]Tc-MAA showing good intrahepatic distribution of the particles without extrahepatic deposition. Following patient-tailored treatment (dosimetry: normal liver parenchymal absorbed dose 30 Gy; tumor absorbed dose >120 Gy; calculated activity 5.5 GBq ¹⁶⁶Ho-microspheres), (D) post-treatment ¹⁶⁶Ho-SPECT-CT and (E) MIP shows reproducible intrahepatic particle distribution, again in absence of extrahepatic depositions. During clinical follow-up, no further hypoglycemic episode at week 2 after treatment. (F) First radiological evaluation with arterial phase CECT 3 months after treatment, showing a PR according to RECIST 1.1 (largest lesion 31 mm; disappearance of multiple small metastases). (G) CECT 12 months after treatment showing continued regression of disease (largest lesion 24 mm). At that time, clinically no further hypoglycemic episodes, developed diabetes mellitus, and started metformin and additional insulin injections.

Studies consistently identify low tumor grade and limited intrahepatic tumor burden as positive prognostic factors for OS, resulting in disease control at first evaluation. Interestingly, one study discovered a negative correlation between the presence of extrahepatic disease and OS (Braat et al. 2019), which was not evident in other studies (Schaarschmidt et al. 2022, Wong et al. 2022). Unfortunately, the extent of extrahepatic disease was not comparable across these studies and the presence (or absence) of progression of extrahepatic disease before radioembolization was not reported.

Radioembolization in earlier lines

Recent European Neuroendocrine Tumor Society guidance papers recommend earlier use of locoregional therapies to prevent carcinoid crisis in active NETs and suggest locoregional treatments as alternatives to systematic therapies for inactive NETs in case of liver-only disease (Hofland et al. 2023, Rinke et al. 2023, Lamarca et al. 2024). The recent European Society for Medical Oncology guideline support liver-directed therapies to debulk liver disease or postpone systemic therapies, independent of treatment line (Pavel et al. 2020). To date, a single study investigated radioembolization in the second-line setting. Schaarschmidt et al. (2022) reported a median OS of 44.8 months and a median progression free survival (PFS) of 18.6 months (Table 2). This study found no clear differences in PFS between second-line and salvage setting (15.9 months).

Radioembolization combined with systemic therapy

Soulen et al. (2024) retrospectively analyzed 21 patients. The combination of capecitabine–temozolomide (CapTem) chemotherapy with resin ⁹⁰Y radioembolization provided longer disease control for grade 2 NELMs than expected for these therapies individually. The median OS since initial diagnosis was 130 months (95%, 67–172 months), with a median hepatic PFS of 35 months (95% CI, 21–45 months; Table 2). The authors suggest a synergistic effect of this combination.

The phase 2 HEPAR PLuS study combined ¹⁶⁶Ho-radioembolization with peptide receptor radionuclide therapy (PRRT) and found promising effects of this combination therapy with acceptable side effects (Braat et al. 2018, 2020). The authors suggest an improved treatment response and additional benefit, specifically for those with significant intrahepatic disease (Table 2). Follow-up analysis of laboratory markers reported limited (temporarily) toxicity after combined treatment (Ebbers et al. 2022b).

In hepatocellular carcinoma, a superior effect of combined immunotherapy with radioembolization compared to immunotherapy alone has been described (Yeo et al. 2023). In NETs, data for combining immunotherapy with radioembolization are lacking. Currently, a pilot phase 2 trial has been set up to determine the effectiveness of pembrolizumab and liver-directed therapy or PRRT (NCT03457948).

To date, no studies have analyzed the additional effect of radioembolization to SSA treatment. A study investigating this exact combination was terminated due to slow patient accrual (NCT02859064).

Clinical outcomes

Frilling et al. (2019) conducted a meta-analysis and found a median OS of 32 months (range 18–67 months) after radioembolization. The ORR was 51% (95%, CT 47–54%) and the disease control rate was 88% (95% CI 85–90%). These response rates are supported by a recent systematic review including 870 patients, concluding that radioembolization for NELMs results in a long survival rate (Jia & Wang 2018). This review also reported improvements of clinical symptoms in 69% of symptomatic patients (Jia & Wang 2018). Another study reported an improvement of symptoms in 50 and 84% of symptomatic patients with NELM treated with radioembolization (Braat et al. 2019, Zuckerman et al. 2019).

Cramer et al. (2016) conducted a prospective longitudinal study of 30 patients evaluating the quality of life of patients with NELM treated with radioembolization. A statistically significant improvement of mental health and social functioning was observed, lasting up to 24 months after treatment. In addition, baseline Mental Component Summary score over 50.0 was correlated with a higher mean survival than a score under 50.0 (37.50 vs 18.19 months, respectively (P = 0.0263)). Only one other study reported quality of life-related outcomes following radioembolization. The HEPAR PLuS study showed a temporary non-significant decrease in reported scales (mainly role functioning scale of the EORTC QLQ-C30) and complete resolution 3 months after treatment (Braat et al. 2020).

Complications

Complications of radioembolization are rare in general. Radioembolization-induced liver disease (REILD) is the most feared complication, which is a form of sinusoidal obstruction syndrome. The incidence of REILD is 0.5–4% in general and <1% in patients with NELM (Table 2) (Braat et al. 2017, 2019). REILD typically presents with jaundice and ascites, usually 4–8 weeks after treatment (Sangro et al. 2008, Braat et al. 2017). However, the severity of REILD varies widely, ranging from asymptomatic to fatal (Braat et al. 2017). The amount of activity administered relative to the targeted liver volume is an independent predictor of REILD (Sangro et al. 2008).

Hepatopulmonary shunting with an estimated lung absorbed dose exceeding 30 Gy is considered a contraindication for treatment and can result in radiation pneumonitis after treatment (Stella et al. 2022). Radiation pneumonitis often presents with a dry cough and dyspnea and occurs approximately 8 weeks after treatment. Due to its rare occurrence in NETs (<1%) (Braat et al. 2019), there are no clear correlations with clinical parameters and radiation absorbed dose. Other non-targeted embolization can result in radiation-induced cholecystitis, ulcer and gastritis (Sangro et al. 2017). Fortunately, these complications are very uncommon in the treatment of NELMs, and often self-limiting (Table 2) (Braat et al. 2015, 2019).

Long term toxicity remains unclear, to date, as structured follow-up studies are lacking and the definition of this complication is not well-established. Several studies have raised concerns, describing portal hypertension and signs of cirrhosis on imaging (e.g., ‘pseudo-cirrhotic morphology’), more frequently after bilobar treatment (Su et al. 2017, Tomozawa et al. 2018). Subsequently, patients who underwent bilobar treatment showed more signs of hepatic decompensation than those who underwent unilobar treatment. However, most patients with hepatic decompensation remain clinically asymptomatic, and in only 5.1% of the patients with bilobar treatment, this decompensation could be solely attributed to radioembolization (Su et al. 2017). A large systematic review of more than 800 patients reported a single death due to hepatic failure after radioembolization (Jia & Wang 2018). Based on currently available data, radioembolization does not seem to impede other subsequent therapies (Braat et al. 2019). Furthermore, similar findings are described following TACE in 22% (vs 29% with radioembolization; not significant) (Currie et al. 2019). Therefore, etiology and relevance of these ‘pseudo-cirrhotic morphology’ findings remain unclear.

The most common hematological complication is lymphocytopenia and thrombocytopenia, which occurs in 6.7 and 3%, respectively (Braat et al. 2019). Other possible complications are angiography-related complications, which are not radioembolization specific and are rare (<1%) (Braat et al. 2019).

Limitations

Radioembolization for the treatment of NELMs is currently not reimbursed in most countries and registries are difficult to design for this very heterogenous patient population. These factors limit its widespread availability for patients with NELMs. The costs of radioembolization and the complexities involved in its delivery should be acknowledged. However, dedicated cost-effectiveness analyses have not been performed to date. Consequently, the financial implications of radioembolization for NELMs remain inadequately assessed.

Besides, not all patients with NELMs are suitable candidates for radioembolization. The required liver function reserve may be hampered due to extensive intrahepatic disease, presences of hypovascular NELMs or progressive extrahepatic disease (Murthy et al. 2008, Anbari et al. 2023).

Another limitation is the retrospective nature of the published studies. Retrospective analyses may overestimate the efficacy of radioembolization and underestimate its risks compared to prospective trials, partly due to selection bias. The limited number of randomized-controlled trials including patients with NELM hinders the development of standardized guidelines defining the role and timing of radioembolization. Finally, both available retrospective and prospective data are hampered by the lack of prospective dosimetry. This limitation may lead to an underestimation of the efficacy and safety of radioembolization, as we use it nowadays.

Conclusions

Radioembolization offers a promising, highly effective and safe treatment modality for patients with NELMs. Emerging evidence suggests that combining radioembolization with systemic therapies and employing prospective dosimetry may further enhance clinical outcomes. Continued research is necessary to establish standardized protocols and to better define the role of radioembolization, both as a monotherapy and in combination with systemic treatments, to enhance the management of NELMs.

Declaration of interest

LSH is supported by a research grant from NWO (NWO P20-57). TB has acted as consultant for Boston Scientific and Sirtex Medical. MGEHL has acted as consultant for Boston Scientific and Terumo. AJATB has acted as consultant for Boston Scientific and Terumo.

Funding

LSH is supported by a research grant from NWO (NWO P20-57). This work did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

Author contributions statements

LSH collected all the data, analyzed the data, drafted the first version, revised and finalized the manuscript. TB analyzed the data and provided feedback on the manuscript. MGEHL read the manuscript and provided feedback on the manuscript. AJATB was the supervisor, provided feedback on the manuscript and finalized the manuscript. All authors approved the final version of the manuscript.