Radiation segmentectomy for hepatocellular carcinoma: redefining the limits of curative-intent therapy

Abstract

Radiation segmentectomy (RS) utilizing Yttrium-90 (Y90) microspheres has emerged as a highly effective, minimally invasive locoregional therapy for hepatocellular carcinoma (HCC), particularly in early-stage disease. This precision treatment delivers ablative radiation doses directly to the tumor and its immediate surrounding parenchyma, thereby minimizing systemic toxicity and preserving healthy liver tissue. RS consistently demonstrates high objective and complete response rates, which often correlate directly with complete pathological necrosis of the tumor, as rigorously assessed by the modified Response Evaluation Criteria in Solid Tumors and European Association for the Study of the Liver criteria. This robust imaging response translates into durable local tumor control and prolonged time to disease progression. The safety and feasibility of RS are generally well established. The procedure is typically well tolerated, with common adverse events usually being mild and transient. Severe complications are rare, emphasizing the procedure’s precision and minimal invasiveness. Optimal outcomes are contingent upon meticulous patient selection, strict adherence to technical guidelines, and precise dosimetry. Beyond its role as a definitive therapy, RS plays a critical function in bridging and downstaging patients awaiting liver transplantation. By effectively controlling tumor burden, it enhances patient eligibility for transplantation and improves outcomes on transplant waiting lists. Furthermore, RS offers distinct advantages over other conventional locoregional therapies in specific clinical scenarios, solidifying its integral position within the comprehensive management paradigm for HCC.

Introduction

Hepatocellular carcinoma (HCC) is the predominant form of primary liver cancer, responsible for nearly four-fifths of all cases worldwide. It remains among the leading global causes of cancer-related mortality, underscoring the substantial clinical and public health burden associated with this disease.¹ The formidable incidence and high mortality rates underscore the urgent and continuous need for effective treatment modalities to improve patient outcomes.

Curative treatment modalities, including liver transplantation (LT), surgical resection, and image-guided ablation, are options for patients with early-stage HCC. However, eligibility for these interventions remains limited by both tumor- and patient-related factors. Only a minority of HCC patients—generally fewer than 20%—qualify for hepatic resection because of constraints such as cirrhosis, inadequate liver function, or anatomically unfavorable tumor locations.¹ Unfavourable tumor location or tumor size >3 cm are also significant limitations for thermal ablation.² These constraints highlight the need to refine locoregional strategies capable of providing durable, potentially curative outcomes for patients who are not candidates for surgery or ablation. In this review, we summarize and critically appraise current evidence on transarterial radioembolization (TARE) performed with a radiation segmentectomy (RS) approach, emphasizing its role as a curative-intent modality for HCC, based on multidisciplinary experience at a tertiary referral center.

Definition and mechanism of radiation segmentectomy

RS is a specialized form of selective internal radiation therapy that entails targeted intra-arterial administration of Yttrium-90-labeled microspheres to the hepatic segment or subsegment harboring the tumor.³ The therapeutic effect of RS relies on Yttrium-90, a pure beta-emitter with a physical short half-life of 64.04 hours, capable of delivering localized high-energy radiation to tumor tissue while sparing adjacent parenchyma.⁴ Upon injection into the hepatic artery, these microspheres are preferentially concentrated within the hypervascular tumor tissue, largely sparing the surrounding nontumorous liver parenchyma. This selective targeting allows for the delivery of high-energy, low-penetration radiation (with an average penetration of approximately 2.5 mm) directly to the tumor cells, inducing localized tumoricidal effects. Nearly all Y-90 energy emission occurs within the first 10-12 days after administration, resulting in a swift and focused delivery of therapeutic radiation.³

The definition of RS has evolved over time. Historically, RS was defined as the selective administration of Y-90 to 1 or 2 hepatic segments according to Couinaud classification. The concept has since evolved toward microsegmental or angiosomal targeting, enabling subsegmental ablation within the liver. This refined approach aims to deliver an ablative dose to both the tumor and the normal tissue within that precisely perfused volume, enhancing precision and minimizing collateral damage.³

Current role and curative intent

In earlier clinical paradigms, radiation-based therapies for HCC were viewed largely as palliative interventions intended to mitigate symptoms and delay progression, rather than achieve tumor eradication.⁴ Recent refinements in RS methodology, individualized dosimetry, and rigorous patient selection have progressively redefined its place within the therapeutic framework of HCC. RS is now recognized as a guideline-supported, curative-intent treatment option for patients with earlier-stage HCC, particularly those presenting with solitary tumors who are not suitable candidates for conventional first-line treatments such as surgical resection or thermal ablation.^1,3,5

This transition from a palliative to a curative paradigm marks a significant conceptual evolution in the perceived therapeutic potential of RS. The reappraisal of RS stems from accumulating clinical data that confirm its high local efficacy, reproducible safety profile, and sustained oncologic outcomes. This evolution is underpinned by continuous improvements in technical delivery, refined dosimetry protocols, and the accumulation of compelling long-term outcome data. The reclassification of RS as a curative-intent therapy has significant implications for the development of HCC treatment algorithms, patient counseling, and potentially influences healthcare policy and reimbursement models. Consequently, RS broadens the spectrum of patients eligible for curative-intent therapy beyond those traditionally suited for surgery or ablation, thereby optimizing management strategies and extending potential survival benefits.

In addition to its definitive applications, RS serves an important role as a bridging or downstaging therapy for patients awaiting liver transplantation, facilitating eligibility under established selection criteria.^1,6 This dual role underscores its versatility and importance in managing the complex trajectory of HCC.

Technical considerations and patient selection for radiation segmentectomy

Technical considerations

The success of RS hinges on precise technical execution and advanced planning. RS is meticulously formulated to deliver high radiation doses selectively to the target tumor(s) and the adjacent hepatic parenchyma. This delivery is typically confined to a maximum of 2 hepatic segments or, with more contemporary approaches, superselectively to smaller subsegments known as angiosomes.^7,8 This segmental or angiosome-specific approach is critical as it minimizes radiation injury to non-cancerous hepatic parenchyma, thereby facilitating posttreatment regeneration of the non-targeted liver segments.⁹

Patient preparation for RS is comprehensive and typically involves planning angiography. Diagnostic angiography of the hepatic artery and its intrahepatic branches is performed first to map tumor-feeding vessels. Super-selective catheterization of these branches is achieved with a 2.1-2.8 F microcatheter introduced through a coaxial system (Figure 1A-D). When neoplastic portal vein thrombosis is present, complete tumoral thrombus targeting must also be ensured.¹⁰ At the same time, as much nontumoral parenchyma as possible should be preserved. Once the possible injection points are identified, cone-beam CT is performed to verify complete tumor targeting, provide volumetric analysis for dosimetry, and ensure complete coverage of the tumor margin. Cone-beam CT also allows exclusion of non-target perfusion. If extra-hepatic vessels are identified, they can be coiled, or, when feasible, the tip of the microcatheter can be positioned distally to their origin.⁴

Figure 1.

Pretreatment arterial (A) and venous-phase (B) CT scans show an HCC nodule, measuring 4 cm in S6, in a 62-year-old male patient with HCV-related cirrhosis. After multidisciplinary evaluation, the patient was proposed for ⁹⁰Y-TARE treatment. DSA of the hepatic artery (C) was acquired after catheterization of the hepatic artery via 5F Simmons catheter and 2.7F microcatheter (Terumo Europe, Leuven, Belgium). Once the feeding vessel arising from the right hepatic artery was identified, it was selectively catheterized through the microcatheter system (D) and Tc-99m-macroaggregated albumin was injected. Subsequent SPECT/CT evaluation (E and F) demonstrated tracer uptake in S6, encompassing the lesion to be treated, with absence of hepatopulmonary shunting and extrahepatic activity. After 7 days, the injection of SIR-Spheres (Sirtex, Massachusetts, USA) beads was performed from the fixed point, with administration of a total activity of 0.88 GBq with 1-day pre-calibration. After ⁹⁰Y microsphere administration, PET/CT reconstruction images (G) demonstrated intense activity in S6, including the lesion to be treated. Post-⁹⁰Y dose distribution map (H) showed the absorbed dose distribution, with a perfused tumor absorbed dose of 154 Gy and a whole-liver normal tissue absorbed dose of 31 Gy (I). 40-day follow-up arterial-phase CT images (J) depicts treatment response with a hypodense appearance of the treated nodule, measuring 3.5 cm, without any radiological signs of vital tissue. At 6-month follow-up arterial-phase CT images (K), the treated nodule continued to appear hypodense with a further reduction in size (3.0 cm).

After having chosen the injection point(s), usually 185 MBq of technetium-99m-macroaggregated albumin (MAA) are administered into selected arterial branches. Within 1-hour, planar scans of the lung and liver area in anterior and posterior views are acquired followed by SPECT/CT scan to detect the distribution of spheres, lung shunt, extrahepatic uptakes and for provisional dosimetry purposes⁴ (Figure 1E and F).

This meticulous mapping is paramount to ensuring the precise delivery of Y90 microspheres to the tumor while minimizing unintended radiation exposure to healthy tissues or other organs.

Innovative refinements such as flow-diversion techniques are being explored to improve treatment selectivity and reduce radiation exposure to uninvolved parenchyma, which is especially relevant in patients with marginal hepatic reserve.⁹

During the treatment phase, Y-90-labeled microspheres—either glass (TheraSphere) or resin (SIR-Spheres)—are infused. The 2 formulations vary in particle size, specific activity, and total number of spheres administered.⁴

The administration set consists of a sterile system of tubes and 3-way valves contained in a plexiglass shield. The dispensing devices allow activity to be administered by injecting the pushing liquid into the conical-bottom bottle containing the microspheres, without the need for special manipulation of the bottle by the operator. For resin microspheres, administration takes place by alternating small aliquots of microspheres with iodized contrast medium infusion in order to monitor microsphere flow under fluoroscopy; for glass microspheres, administration is usually performed as a single delivery. At the end of the administration process, which lasts approximately 10 to 30 minutes, the catheters are flushed. Posttreatment PET/CT is performed on the same day or the following day to assess microsphere distribution and mean dose to the target liver and lungs (Figure 1G-I). The treatment requires 24-hour hospitalization in our practice. In some centers, the same-day radioembolization paradigm is adopted, given the low risk of radiation pneumonitis associated with low lung shunt fractions and limited hepatic target volume.¹¹

Dosimetry

The principal dosimetric goal of RS is to achieve an ablative absorbed dose within the treated volume. Evidence suggests that mean doses exceeding approximately 400 Gy for glass microspheres, or ≥ 500 Gy for resin microspheres, are associated with complete pathologic necrosis and superior clinical outcomes.^3,12 Furthermore, some investigations have correlated doses exceeding 190-200 Gy with improved survival and local tumor control, and doses of 446 Gy or more with complete pathological necrosis (CPN).¹³

These are not merely theoretical dosimetric calculations; they are directly linked to observed clinical outcomes such as CPN and prolonged survival. The DOSISPHERE-01 trial, for instance, confirmed that patient-specific, high-dose Y-90 radioembolization yields higher objective response and overall survival than conventional activity prescriptions.¹⁴ Kappadath et al¹⁵ derived a model for single-compartment dose prescriptions required for CPN in terms of mean tumor dose, point dose at the tumor boundary, and point dose at 2 mm beyond the tumor edge for a range of tumor diameters (1-7 cm) and tumor-to-normal liver uptake ratios. However, available radiopathologic studies of ablative radioembolization have not consistently shown tumor size itself to correlate with CPN.¹⁶

Several methodologies are employed for dose estimation, broadly categorized into tissue-level and voxel-level approaches.^17,18

Tissue-level models

The simplest is the Medical Internal Radiation Dose (MIRD) model which assumes that the distribution of the activity is completely homogeneous over the target volume, which is the only source of activity. More in detail, in the MIRD formalism, the absorbed dose in a volume is expressed as the ratio of energy deposited in that volume over time and its mass.^19–22 Since, after TARE procedures, the microspheres are considered permanently fixed in the target volume, a single SPECT/CT is sufficient to estimate the mean absorbed dose to the target. Therefore, only a tissue-specific activity-to-dose conversion factor (S-factor) for Y90 and the target mass are required for target dose calculation. To consider differences in activity concentration between target and tumor, the MIRD formalism can be extended to multiple compartments.^17,18,22,23 In this case, called the partition model, it is possible to estimate the mean absorbed dose to different volumes, such as tumor and normal liver, by knowing the S-factors, the tumor mass, the liver mass, and the ratio between the activity concentration in the tumor and the normal liver (T/N ratio). In this case, each individual volume is considered to be both an independent source of activity and a target for dose calculation. However, within each volume, the activity concentration is still considered homogeneous. Both methods can be adopted in the same way to estimate the mean lung dose by considering the prescription activity and the lung shunt fraction. Other empirical approaches such as the body surface area (BSA) method have also been adopted. However, the degree of personalization of these methods is questionable, and the use of MIRD-based methods is considered more appropriate.²³

Voxel-level models

The limitations of the tissue-level dosimetry in terms of dose estimation accuracy due to inhomogeneous activity concentration and energy deposition have led to the implementation of voxel-based approaches.²³ Again, the MIRD method can be extended at a voxel level by considering the activity concentration of each voxel, which is treated as a source and/or a target depending on the dose deposition model.^17,18,22,23 In its simplest form, the voxel-based MIRD approach implements the local deposition (LD) model, in which all the emitted energy from a voxel is absorbed locally in the same voxel, without considering any contribution from or to neighboring voxels. Given that 90Y is mostly a beta emitter, the LD method gives good results, and many commercial software programs are based on this method.²⁴ However, better dosimetric accuracy could be achieved by considering the contribution of energy coming from neighboring voxels to the considered one. This approach is implemented by the dose kernel (DK) deposition model, in which the activity concentration map obtained from SPECT/CT is convolved with a kernel (a filter matrix) that describes how the energy is deposited around an activity source.²⁵ The final result of this process is an accurate estimation of the absorbed dose in a homogeneous medium. Additionally, kernels can also be scaled to different local tissue densities to give more accurate results (eg, for lung dose estimation).²⁶ The DKs are usually precomputed from Monte Carlo simulations and are voxel-size dependent.²⁵ Monte Carlo simulations can also be adopted to perform a direct dose calculation starting from the activity concentration: this approach allows the simulation of the radiation behavior in tissues, and it is considered the most accurate approach.^27,28 This method intrinsically considers activity and density inhomogeneities voxel by voxel.

After the treatment, the same method employed in the pretreatment dose estimation should be adopted in posttreatment dose evaluation. This is usually done by considering posttreatment PET/CT images.¹⁷ Differences in 99mTc-MAA and microspheres’ distribution can lead to discrepancies in pretreatment and posttreatment dose estimation.

Patient selection criteria

Eligibility for RS is established through multidisciplinary assessment involving hepatology, interventional radiology, surgery, nuclear medicine, radiation and medical oncology teams to ensure optimal individualized treatment planning.

Key criteria for patient selection include tumor characteristics, liver function, performance status, and vascular anatomy. Suitable candidates usually present with solitary or limited HCC confined to 1 or 2 hepatic segments, generally ≤8 cm in maximum diameter, and without extrahepatic spread.^1,3 The absence of macrovascular invasion is preferred, although RS has demonstrated effectiveness in carefully selected cases with portal venous invasion when the tumor thrombus can be completely targeted.¹⁰

Adequate hepatic reserve is essential to tolerate the high localized dose required for tumor ablation. Most candidates have Child-Pugh A liver function, while carefully selected Child-Pugh B7 patients may be considered, particularly when RS is used for bridging or downstaging to transplantation.¹ Patients should also have good functional status, typically ECOG 0-1, although ECOG 2 may be acceptable for TARE in selected cases.^1,3 A favorable hepatic arterial anatomy is required to permit safe and effective delivery of Y90 microspheres; prophylactic embolization of distal branches leading to the gastrointestinal tract is rarely necessary but can be performed to prevent nontarget embolization.⁴

Certain conditions preclude the safe use of RS or require particular caution. These include uncorrectable arteriovenous or portosystemic shunting, substantial pulmonary shunting that could result in excessive lung radiation dose (exceeding 30 Gy in a single treatment or 50 Gy in bilobar treatment), severe renal or hepatic insufficiency, significant biliary obstruction, and prior biliary interventions (eg, bilio-enteric anastomosis) or a dysfunctional sphincter of Oddi because of the increased risk of hepatic abscess.⁴ Patients with biliary abnormalities should receive antibiotic prophylaxis when treatment proceeds. Prior hepatectomy also warrants caution, as the remaining future liver remnant and potential toxicity require careful consideration.²⁹ The ability to deliver high, ablative doses is therefore intrinsically linked to hepatic reserve: healthier liver function allows safer administration of more potent doses, whereas compromised liver function may limit the dose and reduce the probability of a curative outcome.

Imaging response and efficacy of radiation segmentectomy

Assessment criteria for response

Accurate assessment of tumor response following RS for HCC is critical for guiding clinical management. For RS, imaging response is most precisely evaluated by assessing the degree of tumor enhancement, which serves as a reliable marker of tumor viability.³⁰ This approach is particularly important because, unlike cytotoxic chemotherapy, treated lesions may not immediately decrease in size.

The modified Response Evaluation Criteria in Solid Tumors (mRECIST) and the European Association for the Study of the Liver (EASL) criteria are the preferred assessment tools for RS.^3,31,32 These criteria focus on the disappearance of arterial enhancement within the tumor, which is indicative of complete necrosis, rather than solely on volumetric changes³⁰ (Figure 1J and K). According to mRECIST, a complete response (CR) is defined as the disappearance of any intratumoral arterial enhancement in all target lesions.³¹ The Liver Imaging Reporting and Data System (LI-RADS) also incorporates a specific algorithm for evaluating treatment response after Y90 radioembolization for HCC, with a 2024 update addressing the expected persistent enhancement often observed in responding tumors.³³

Objective response rates and complete response

RS consistently demonstrates high objective and complete response rates across studies, as summarized in Table 1. The multicenter study by Vouche et al³⁴ first established the strength of this response in unresectable HCC ≤ 5 cm, reporting an 87% objective response rate by mRECIST, with almost half of patients achieving complete response. These findings were subsequently reinforced in larger and more contemporary cohorts.

Table 1.

Key imaging response rates and survival outcomes for RS (across major trials).

Study name (year)	Number of patients	Tumor characteristics	Imaging criteria	Objective response rate (ORR)	Complete response (CR)	Median time to progression (TTP)	Median overall survival (OS)	Complete pathological necrosis (CPN) rate
Riaz et al (2011)41	84	N/A	EASL	81%	N/A	13.6 months	26.9 months	N/A
Vouche et al (2014)34	102	Unresectable HCC ≤ 5 cm	mRECIST	87%	47%	33.1 months	53.4 months	52%
Lewandowski et al (2018)42	70	Solitary HCC ≤ 5 cm	EASL, WHO	86% (EASL)	59% (EASL)	2.4 years (29 months)	6.7 years	N/A
LEGACY Study (2021)35	162	Solitary, unresectable HCC ≤ 8 cm	mRECIST	88.3%	72.2% (confirmed)	93.9% PFS at 24 months	86.6% at 3 years	N/A
RASER Trial (2022)36	29	Solitary HCC ≤ 3 cm (suboptimal location)	mRECIST	100%	83% (initial), 90% (sustained)	N/A	N/A	100% (explant, n = 8)
De la Garza-Ramos C (2025)29	58 (RS cohort)	Solitary HCC ≤ 8 cm	Radiologic CR	97% (rCR)	97% (rCR)	77 months (median PFS)	63 months (median censored OS)	76% (explant, n = 28)
Sarwar A (2024)37	67	HCC	mRECIST	94% (per-lesion ORR at 6 mo)	70% (per-lesion CR at 3 mo)	1 year (88% per-lesion DoR)	26 months	N/A

Abbreviations: CPN = Complete Pathological Necrosis; CR = Complete Response; DoR = Duration of Response; N/A = Not Available/Applicable; ORR = Objective Response Rate; OS = Overall Survival; PFS = Progression-Free Survival; rCR = Radiologic Complete Response; TTP = Time to Progression.

In the LEGACY study, which evaluated 162 patients with solitary unresectable HCC ≤8 cm, the best objective response rate by mRECIST was 88.3%, and most responders maintained response for at least 6 months.³⁵ The prospective RASER trial extended these observations to patients with HCC ≤3 cm in locations considered suboptimal for percutaneous ablation; all patients had an initial objective response, and most achieved sustained complete response after a single treatment.³⁶ Together, these studies support RS as an ablative option particularly suited to small or solitary tumors in technically challenging locations.

Evidence obtained with resin microspheres is also consistent with this pattern. A retrospective single-center series reported high per-lesion objective response and disease-control rates at 6 months, with many patients maintaining response for 1 year or longer.³⁷ Interim results from the DOORwaY90 study similarly reported very high best overall and complete response rates according to mRECIST at 9 months.¹² Additional comparative studies further support the depth of response after RS: Padia et al³⁸ observed higher complete response rates with RS than with segmental chemoembolization, while Arndt et al³⁹ and De la Garza-Ramos et al⁴⁰ also reported high mRECIST objective response rates.

Time to progression and overall survival outcomes

RS provides durable local control and prolonged time to progression (TTP), contributing to favorable overall survival (OS) outcomes (Table 1). In the earliest report by Riaz et al,⁴¹ median TTP was 13.6 months and median OS was 26.9 months. Subsequent studies reported longer disease-control intervals in more selected cohorts: Vouche et al³⁴ observed a median TTP of 33.1 months and a median OS of 53.4 months, while Lewandowski et al⁴² reported a median TTP of 2.4 years, no target lesion progression at 5 years in 72% of patients, and a median OS of 6.7 years.

The LEGACY study further demonstrated the durability of local control, with no local progression at 24 months and a 93.9% progression-free survival (PFS) rate within the treated area.³⁵ OS was also favorable, particularly among patients who subsequently underwent resection or liver transplantation. A multicenter comparison of RS with external beam radiation therapy similarly reported high 1-year PFS and OS after RS, with median censored OS of 63 months.²⁹ Although survival outcomes vary according to liver function, transplant eligibility, and competing risks from cirrhosis, these data consistently indicate that RS can provide durable local tumor control in appropriately selected patients.

Complete pathological necrosis after radiation segmentectomy

A critical aspect of RS efficacy is the strong correlation between radiological response, particularly complete response, and CPN observed on explant pathology.³ This correlation indicates that RS is not merely shrinking tumors but ablating them. In the study by Vouche et al,³⁴ 52% of patients achieved 100% necrosis on explant. Gabr et al⁴³ further showed that all patients receiving a dose greater than 400 Gy achieved CPN, supporting this threshold as a reproducible marker of ablative treatment.

The RASER trial provided additional prospective support, as all 8 target lesions from patients who underwent subsequent liver transplantation exhibited CPN.³⁶ Similarly, a validation study reported that 68% of tumors demonstrated ≥99% pathological necrosis; CPN was present in 53% of tumors treated with >190 Gy and in 75% of those treated with >500 Gy, and complete response by mRECIST was significantly associated with ≥99% tumor necrosis.³⁰ Importantly, the achievement of CPN is strongly associated with a lower incidence of post-transplant recurrence and superior survival outcomes.⁴⁴

The emphasis on tumor viability, rather than size reduction alone, represents an important evolution in oncologic assessment after RS. The consistent relationship between radiological complete response and CPN validates the use of mRECIST and EASL criteria in this setting and supports the concept that, for radiation-based treatments delivered with ablative intent, biological effect is more prognostically relevant than simple volumetric change.

Safety and feasibility profile of radiation segmentectomy

Radiation segmentectomy is widely regarded as a versatile, safe, and effective ablative therapy for early-stage hepatocellular carcinoma. It is generally well tolerated by patients, often demonstrating a more favorable tolerability profile compared with other locoregional therapies such as transarterial chemoembolization (TACE).⁴⁵

Clinical side effects associated with RS typically manifest within the first 4 to 6 weeks following treatment. Post-embolization syndrome, a constellation of fatigue, anorexia, abdominal pain, nausea, and/or vomiting, is the most commonly reported adverse event following TARE, with a total incidence of 20% to 55%. This syndrome is mild and self-limited in the majority of cases, due to the minor embolizing effect of TARE, with management of symptomatic patients with pain medications, antiemetics, and hydration, and glucocorticoids as needed in severe cases.⁴⁶

While RS is generally safe, severe adverse events (SAEs) and complications can occur, though their incidence is relatively low. In the LEGACY study, SAEs occurred in 9.9% of patients; 5.6% were at least possibly related to radioembolization.³⁵ However, other trials, such as the DOSISPHERE-01 trial, reported SAEs in less than 5% of patients.¹⁴ A study specifically evaluating Y90 resin microspheres noted only 1 grade 3 clinical adverse event (abdominal pain) in 67 patients (1.5%).³⁷

Potential complications include liver-related, gastrointestinal, pulmonary, and procedure-related events. Liver-related complications may include acute or chronic altered liver function, ascites, bile duct injury, hepatic encephalopathy, hepatorenal failure, portal hypertension, and decompensation of preexisting chronic liver disease. Radiation-induced liver disease (RILD) is rare but serious; notably, no patient in the LEGACY study experienced RILD.³⁵ Gastrointestinal complications, including cholecystitis and ulceration, are less common and are usually related to ectopic microsphere distribution, which can generally be prevented by meticulous technique.⁴⁶ Pulmonary complications such as radiation pneumonitis are rare and primarily result from unintended lung exposure due to arteriovenous shunting.⁴⁶ Procedure-related complications may include allergic reactions, arterial injury such as dissection, and access-site bruising or hematoma.⁴⁷

The feasibility of RS is enhanced by its outpatient nature in some healthcare systems.⁴⁸ The potential for a same-day radioembolization paradigm is also being considered, given the low risk of radiation pneumonitis and the limited hepatic volume targeted.¹¹ Compared with TACE, TARE has been associated with a better quality of life for patients.⁴⁹

The safety profile of RS presents an interesting observation: despite delivering a high, ablative radiation dose locally, it maintains a significantly lower incidence of major adverse events compared to more invasive systemic therapies or even surgical resection. This safety is attributed to the highly selective nature of Y90 microsphere delivery. By concentrating the radiation precisely within the tumor and a confined surrounding hepatic segment, RS effectively minimizes systemic exposure and limits damage to healthy, non-targeted liver parenchyma and other organs. This observation underscores that the method of delivery is as crucial as the radiation dose itself in determining the overall safety profile and patient tolerability, making RS a particularly advantageous option for patients with compromised liver function or other significant comorbidities.

Broader role of radiation segmentectomy in HCC management

Bridging and downstaging to liver transplantation

RS plays a crucial and multifaceted role in the management of HCC patients awaiting liver transplantation. It serves as an effective strategy for both bridging patients on the transplant waiting list and downstaging tumors to meet the eligibility criteria for LT, usually defined locally at each center.^1,50

The LEGACY study provided compelling evidence of RS’s impact on transplant eligibility, demonstrating that 85.3% of patients were able to attain and/or maintain Milan criteria after treatment at 24 months.³⁵ This is particularly significant given the often-prolonged wait times for liver transplantation, which can lead to tumor progression and patient dropout from transplant lists. TARE has been shown to be superior to TACE in effectively downstaging patients into Milan criteria and is associated with the need for fewer treatments to maintain tumor control in a bridge-to-transplant setting.⁴⁵ Furthermore, patients who achieve extensive partial necrosis or CPN following TARE may experience improved recurrence-free survival after LT and exhibit lower rates of microvascular invasion on explant pathology, contributing to better long-term outcomes.⁵¹

Advantages and disadvantages compared to other locoregional therapies

RS offers distinct advantages and disadvantages when compared with other established locoregional therapies for HCC.

RS has been compared with surgical resection in patients with localized HCC, providing useful insights into the balance between oncologic efficacy and procedural safety. In a multicenter retrospective analysis by De la Garza-Ramos and Toskich, patients with solitary HCC lesions ≤8 cm treated with either RS or resection had nearly identical local tumor control, with tumor progression occurring in only 5% of RS patients and 8% of resected patients.⁴⁰ Overall disease progression, including new or recurrent intrahepatic lesions, occurred in about one-third of patients in both groups. Although TTP was somewhat longer after surgery than after RS (29.4 months vs 21.9 months), this difference was no longer evident among patients with more advanced liver disease or lower platelet counts, and OS remained unreached in both cohorts. A separate study of larger tumors (≥5 cm) similarly suggested comparable overall survival and progression-free outcomes between TARE and surgical resection after adjustment for baseline differences, with a more favorable safety profile for TARE.⁵² Safety is one of the strongest advantages of RS: in the De la Garza-Ramos study, major adverse events occurred in none of the RS patients compared with 20% of resected patients, and 30-day readmission was 0% after RS vs 12% after surgery.⁴⁰ These comparisons must be interpreted cautiously because RS and resection are usually offered to different patient populations. Nevertheless, the available evidence supports RS as a viable curative-intent alternative for patients who are not surgical candidates, particularly when minimizing procedural risk is crucial.

The comparison with TACE further highlights the potential advantages of TARE in selected patients with HCC. The PREMIERE study, a prospective phase 2 randomized controlled trial, compared TARE with conventional TACE in patients with BCLC stage A or B HCC who were not candidates for surgical resection or thermal ablation. Patients treated with TARE had a significantly longer time to tumor progression, exceeding 26 months compared with 6.8 months in the TACE group, and experienced fewer side effects such as diarrhea and hypoalbuminemia.⁵³ Padia and colleagues similarly found that RS produced higher complete and overall response rates and longer progression-free survival than segmental TACE in localized unresectable HCC.³⁸ These findings were supported by Biederman et al,⁵⁴ who reported higher imaging response and a less frequent need for repeat treatment after TARE than after TACE in small solitary unresectable tumors. The randomized TRACE trial further strengthened this evidence, demonstrating longer median time to progression and overall survival with TARE than with drug-eluting bead TACE.⁵⁵ Taken together, these data indicate that TARE may offer more durable tumor control and better tolerability than TACE in appropriately selected patients with unresectable HCC.

RS has also been evaluated against percutaneous thermal ablation techniques, particularly microwave ablation (MWA), in small solitary HCC. A retrospective propensity score-matched study conducted between 2014 and 2017 directly compared RS with percutaneous MWA in 68 treatment-naive patients with solitary HCC ≤4 cm. Both techniques achieved similar results in terms of toxicity, tumor response, and overall survival. Although the objective tumor response rate was slightly higher for RS than for MWA (90.9% vs 82.6%), this difference did not reach statistical significance. Mean overall survival was also comparable, at 59 months for RS vs 44 months for MWA. However, RS showed a clear advantage in targeted-tumor PFS, with a mean of 57.8 months compared with 38.6 months for MWA.³⁹ Additional comparative evidence comes from a large retrospective study by Biederman and colleagues, which evaluated patients with unresectable solitary HCC measuring up to 3 cm and compared RS with combined TACE followed by MWA. The study, conducted between 2010 and 2015, initially included more than 650 patients, with a carefully matched subset of 121 treatment-naive individuals: 41 treated with RS and 80 treated with TACE-MWA. No significant differences were observed between the 2 groups in complete response rate, TTP, or OS. Both treatments also demonstrated excellent safety profiles, with no deaths reported within 90 days after treatment. The authors concluded that, for solitary HCCs up to 3 cm, RS and TACE-MWA offered comparable efficacy and safety.⁵⁶

Although current evidence is insufficient to establish clear noninferiority or superiority of RS compared with thermal ablation across all clinical settings, RS remains a valuable alternative for patients who are not ideal candidates for percutaneous ablation because of tumor location or other clinical factors. Moreover, while thermal ablation is validated for curative intent primarily for HCC lesions smaller than 3 cm, RS has shown favorable response rates for lesions up to 8 cm.

Finally, RS has been compared with external beam radiation therapy (EBRT) in patients with solitary HCC. In a multicenter retrospective review including 86 patients, 58 were treated with RS and 28 with EBRT. The EBRT cohort included older patients, larger tumors, and worse performance status.²⁹ Radiologic complete response was achieved in 97% of RS patients and 82% of EBRT patients, while 1-year local control was 97% and 93%, respectively. Subsequent LT was performed in 48% of RS patients and 11% of EBRT patients, with complete pathologic response rates of 76% (22/28) and 33% (1/3), respectively. Progression-free survival at 1 year was 87% after RS vs 80% after EBRT, and 1- and 2-year overall survival was 88% and 85% after RS vs 84% and 59% after EBRT.²⁹ From a technical perspective, RS delivers a high, localized radiation dose directly to the target tumor via the hepatic artery, sparing a larger proportion of normal parenchyma compared with EBRT, where a wider field is often irradiated. Unlike EBRT, RS is not constrained by external-beam dose limits to adjacent structures. Challenging lesion locations, such as the caudate lobe or hepatic dome, may also be more accessible via radioembolization, whereas EBRT in these areas may increase the risk of injury to adjacent structures such as the lungs and porta hepatis.⁵⁷

Conclusions

RS has emerged as a highly effective and well tolerated locoregional therapy for HCC, particularly for early-stage disease. Its ability to induce high rates of imaging response, often correlating with CPN, translates into durable local tumor control and prolonged TTP. This robust efficacy has led to its recognition as a curative-intent therapy for selected patients, marking a significant paradigm shift from its historical palliative role. The safety and feasibility of RS are well established, supported by its precise delivery mechanism that minimizes systemic toxicity while maximizing local tumoricidal effect. Common adverse events are typically mild and transient, with severe complications being rare. The success of RS is highly dependent on meticulous patient selection, precise dosimetry, and a multidisciplinary approach to care. Beyond its definitive role, RS is indispensable for bridging and downstaging patients to LT, improving their eligibility and outcomes on transplant waiting lists. Its distinct advantages over other locoregional therapies, such as TACE, thermal ablation, and EBRT, in terms of efficacy, safety, and accessibility for challenging lesions (lesions >3 cm and/or in difficult locations), further solidify its integral position in the comprehensive management of HCC. Future investigations should continue to refine dosimetry techniques to consistently achieve optimal ablative doses and explore their application in broader patient cohorts, including those with more limited hepatic reserve. Further prospective, comparative studies will continue to define the precise role of RS within the evolving landscape of HCC treatment.

Conflicts of interest

Laura Crocetti: Boston, Sirtex, Terumo, Histosonics, AstraZeneca, EISAI, GE Healthcare, Varian-Siemens

Elena Bozzi: Sirtex, Terumo, Varian-Siemens, GE Healthcare, AstraZeneca, EISAI

Funding

None declared.