Rethinking the Role of Radiation Therapy in the Treatment of Unresectable Hepatocellular Carcinoma: A Data Driven Treatment Algorithm for Optimizing Outcomes

Hepatocellular carcinoma (HCC) is the second most common cause of cancer death worldwide, with a majority of HCC patients not suitable for curative therapies. Approximately 70% of initially diagnosed patients cannot undergo surgical resection or transplantation due to locally advanced disease, poor liver function/underlying cirrhosis, or additional comorbidities. Local therapeutic options for patients with unresectable HCC, who are not suitable for thermal ablation, include transarterial embolization (bland, chemoembolization, radioembolization) and/or external beam radiation therapy (EBRT). Regarding EBRT specifically, technological advancements provide a means for safe and effective radiotherapy delivery in a wide spectrum of HCC patients. In multiple prospective studies, EBRT delivery in a variety of different fractionation schemes or in combination with transcatheter arterial chemoembolization (TACE) demonstrate improved outcomes, particularly with combination therapy. The Barcelona Clinic Liver Cancer classification provides a framework for treatment selection; however, given the growing complexity of treatment strategies, this classification system tends to simplify decision-making. In this review, we discuss the current literature regarding unresectable HCC and propose a modified treatment algorithm that emphasizes the role of radiation therapy for Child-Pugh score A or B patients with ≤3 nodules measuring >3 cm, multinodular disease or portal venous thrombosis.


INTRODUCTION
Hepatocellular carcinoma (HCC) is the second most common cause of cancer death worldwide (1). While the incidence of HCC is highest in Asia, the rate has been increasing significantly in North America (2). HCC can result in patients with liver cirrhosis, and known major risk factors for cirrhosis include viruses [chronic hepatitis B virus (HBV) and hepatitis C virus (HCV)], toxins (e.g., alcohol, tobacco, and aflatoxins), and metabolic disorders (e.g., nonalcoholic steatohepatitis, and diabetes) and other conditions, such as hereditary hemochromatosis (3). The American Association for the Study of Liver Diseases (AASLD) recommends surveillance of patients with cirrhosis using ultrasound, with or without alphafetoprotein (AFP), every 6 months (4). Patients with HCC are often asymptomatic at the time of diagnosis leading to a delay in diagnosis for patients not being screened for HCC in the setting of viral hepatitis infection. Classic imaging characteristics of arterial enhancement and venous or delayed-phase washout of lesions >1cm in patients with cirrhosis or chronic HBV are considered by some as diagnostic for HCC even in the absence of histologic confirmation (5).
Local control (LC) is the most important prognostic factor for HCC, because up to 92% of deaths can be directly correlated to local progression leading to liver failure rather than distant metastases (6,7). While liver transplantation or surgical resection remains the principal curative option for patients with HCC, only 30% are suitable for this therapy. Patients are often deemed nonsurgical candidates due to locally advanced disease, poor liver function, additional comorbidities, and/or poor performance status (8). For HCC patients who are non-surgical candidates, an alternative curative therapeutic option is radiofrequency ablation (RFA), which has optimal outcomes in tumors <3 cm that are primarily located away from major blood vessels, bile ducts, and abdominal organs (9). Some unresectable HCC patients are also not candidates for RFA due to location of the tumor, intrahepatic bile duct dilation, and volume of disease; for these patients other local therapies include transcatheter arterial bland embolization, chemoembolization (TACE), radioembolization (TARE) with Yttrium-90 ( 90 Y) microspheres, and external beam radiation therapy (EBRT). Inoperable lesions for which local ablation is not possible are treated with TACE since non-randomized studies suggest it may improve survival compared to best supportive care (10,11). Historically, liver EBRT was not employed due to risk of radiation-induced liver disease (12,13). However, modern imaging techniques, advances in EBRT planning and delivery, and improvements in biological understanding of radiation dose tolerances to liver parenchyma have led to reconsideration of EBRT in the context of other local treatment options. Moreover, with a better understanding of the dose-volume effects of partial liver radiation and utilization of advanced radiation technology, severe toxicity rates following EBRT are now less than 10% (14).
The Barcelona Clinic Liver Cancer (BCLC) staging classification combines tumor number, size, extent of spread, performance status, and Child Pugh (CP) score to provide treatment recommendations for patients with HCC (Table 1) (15,16). While the BCLC is the most commonly used treatment algorithm for newly diagnosed HCC, it may inadequately reflect therapeutic advances and simplify decision making, particularly in patients with BCLC Stage B or C disease. For example, the algorithm does not provide treatment recommendations for patients with ≤3 nodules measuring >3 cm, a common presentation in the era of advanced imaging techniques. Furthermore, in light of growing prospective data evaluating the role of modern radiation therapy (RT) for HCC, EBRT has emerged as a potential treatment option in select patients with BCLC Stage B and C disease and EBRT is not incorporated into this algorithm. Given this constantly evolving treatment paradigm, herein, we evaluate the published data on local therapeutic options for unresectable HCC and propose a functional treatment algorithm for CP-A or B patients with ≤3 nodules measuring >3 cm, multinodular disease, or portal venous thrombosis (PVT) (Figure 1).

TACE
TACE is a commonly used treatment modality in patients with HCC who are deemed poor candidates for curative treatment with surgery or RFA. According to a recent Surveillance, Epidemiology, and End Results (SEER) database analysis, TACE is the most common initial therapy utilized in the US (64%) among patients who receive treatment for HCC (17). This treatment involves the intra-arterial injection of chemotherapeutic agents followed by obstruction of selective hepatic arterial inflow with embolizing particles. The purpose of this procedure is to increase tumor cell exposure to cytotoxic agents and selectively tamponade the blood supply to the tumor or affected liver lobe. The resultant stasis of blood flow and hypoxia induces vascular endothelial growth factor (VEGF) secretion, which increases vessel permeability and results in higher intra-hepatic chemotherapy deposition (18). TACE with drug-eluting beads (DEB-TACE) utilizes embolizing particles to both embolize the hepatic artery and carry the cytotoxic agents. The most commonly employed chemotherapeutic agents are mitomycin-C and doxorubicin.
Patient selection for TACE is an important step to avoid treatment-related adverse events. Ideal candidates present with Eastern Cooperative Oncology Group (ECOG) performance status ≤2, preserved liver function (CP-A), and tumors measuring <10 cm without portal vein thrombosis (PVT). Select patients with impaired liver function (CP-B) and mildly impaired performance status can be treated with TACE; however, the incidence of treatment-related toxicities including abdominal pain, nausea, and vomiting increases in this patient population (19). Contraindications to TACE include an ECOG performance status >2, advanced cirrhosis (CP-C), >50% replacement of the liver by tumor, PVT, renal insufficiency (creatinine ≥2 mg/dl or creatinine clearance <30 ml/min), bilirubin levels >3 mg/dL, macroscopic vascular invasion, extrahepatic disease, bile duct occlusion, and comorbidities involving compromised organ function, such as active cardiovascular disease (20)(21)(22)(23)(24)(25).

TACE Plus Molecularly Targeted Therapy
Treatment with TACE can lead to the promotion of tumorigenesis and angiogenesis (41), which may partially explain the limited long-term benefit of this therapy. Thus, in an attempt to improve the efficacy of TACE, studies have combined this treatment with concurrent systemic targeted therapies ( Table 3) (31)(32)(33)(34)(35)(36)(37)(38)(39)(40). One such targeted therapy is sorafenib, a small molecule inhibitor of several tyrosine protein kinases (TKI) including vascular endothelial growth factor receptors (VEGFRs)-1, 2, and 3 and platelet-derived growth factor receptor β (PDGFR-β) (43,44). In preclinical studies, sorafenib demonstrated antiproliferative activity in malignant hepatic cell lines by decreasing tumor angiogenesis and tumor-cell signaling as well as increasing tumor-cell apoptosis (45). Given these findings, in a multicenter randomized phase III trial, sorafenib demonstrated improved median overall survival (OS) when compared to placebo (10.7 months vs. 7.9 months; P < 0.001) in patients with CP-A advanced HCC (46). Furthermore, in cell culture models, sorafenib reduced the susceptibility of hepatocytes to HCV infection via anti-VEGF activity (47,48) and directly inhibited HCV replication via non-structural HCV replicon protein NS5A interaction with C-Raf (49). In light of  this preclinical information, sorafenib is also thought to be more effective in hepatitis-related HCC (50). Given the mechanisms of action of both sorafenib and TACE, there is growing support for this therapeutic combination to take advantage of a possible synergistic effect. TACE increases the concentration of angiogenic growth factors such as VEGF and insulin-like growth factor-2 (IGF-2), which may contribute to disease progression (51), while sorafenib inhibits angiogenic growth factors to prevent progression. Based on this rationale, the combination of TACE with sorafenib has been investigated in a number of studies (32,42,52). Lencioni (31,33). Therefore, TACE has been combined with multiple targeted agents including sorafenib, but to date, this combination therapy has not led to a meaningful increase in survival ( Table 3) (35)(36)(37)(38)40). While sorafenib is considered first-line systemic therapy after failure of liver-directed therapies, a new multikinase inhibitor, lenvatinib, has emerged as a new alternative first-line treatment option (53). In the randomized phase III REFLECT trial, lenvatinib demonstrated a comparable OS (median, 13.6 vs. 12.3 months, HR 0.92, 95% CI 0.79-1.06) and improved TTP (median, 8.9 vs. 3.7 months, HR 0.63, 95% CI 0.53-0.73) over sorafenib. In the phase I/II CheckMate 040 trial, a PD-1 inhibitor, nivolumab, demonstrated a 20% objective response (HR 95% CI [15][16][17][18][19][20][21][22][23][24][25][26] in patients with advanced HCC and is now approved as second-line therapy following prior sorafenib (54). More recently, in the phase III CLESTIAL trial, cabozantinib has been shown to improve median OS compared with placebo after progression on sorafenib (10.2 vs. 8 months, P = 0.005) (55). Given the encouraging results of these studies, combination of these new agents with TACE should be investigated in prospective studies.
Radiosensitization with systemic therapy is the principle underlying many treatment regimens for solid malignancies and has gained interest for HCC in recent years. RT has multiple effects on the tumor microenvironment and the immune system, including cytokine and antigen release leading to increased immune cell infiltrate (56). The potential increase in toxicity with using combination therapy remains the main concern as several clinical studies (NCT03203304 and NCT03482102) are currently evaluating the optimal combination strategies with RT.

TACE Plus RFA
Radiofrequency ablation (RFA) plus TACE can also be utilized in HCC patients. RFA is considered an effective treatment for tumors <3 cm by conducting high-energy electrical current or microwaves into the target lesion which then leads to tumor tissue necrosis (57). Reported LC rates are as high as 90%; however, this decreases significantly with increasing tumor size as well as close proximity to major vessels due to a heat-sink effect (58,59). The heat-sink effect is a phenomenon that occurs when flowing hepatic blood causes a cooling effect, thereby reducing the ablation volume. Based on the theory that performing TACE before RFA may allow retention of thermal energy within the tumor environment by decreasing blood flow, studies compared RFA alone to TACE plus RFA and demonstrated improved survival for the latter in patients with HCC measuring <3 cm (60). However, the survival benefit decreases in tumors >3 cm. Lin et al. randomized 62 patients with HCC to either TACE plus RFA or RFA alone from 2006 to 2010 (61). Patients were CP-A or B and had ≤ 3 tumors measuring 3-5 cm with no evidence of extrahepatic tumor metastasis or macrovascular invasion. The 1-, 2-, and 3-year local tumor progression rates in the TACE plus RFA group (12.5%, 18.75%, and 18.75%) were significantly lower than in the RFA alone group (16.7%, 30%, and 36.6%, P = 0.047). However, 1-, 2-, and 3-year OS rates remained similar between the two treatment groups (90.6% vs. 83.3%, 72% vs. 56.75%, and 53.1% vs. 23.3%, P = 0.176). Given the improved prognosis with combination therapy in patients with small HCCs, TACE plus RFA may be considered in CP-A or B patients with ≤3 tumors measuring <3 cm.

Transarterial Radioembolization (TARE) With Yttrium-90 ( 90 Y)
TARE with 90 Y involves the injection of β-emitting 90 Y loaded glass matrices or resin microspheres into the hepatic artery which leads to delivery of concentrated radiation to the tumor. The radioisotope 90 Y is a pure β-radiation emitter with a half-life of 64.2 h, an average energy of 0.94 MeV, and an average penetration range in tissue of 2.5 mm (62)(63)(64). Absolute contraindications for 90 Y radioembolization include significant intractable clinical ascites, bleeding diathesis, severe portal hypertension with hepatofugal flow, or severe peripheral vascular disease that would preclude arterial catheterization (65). Moreno-Luna et al. compared unresectable HCC patients treated with TARE (n = 61, 87% CP-A, 69% multinodular, and mean tumor size 6 cm, range 2-9 cm) in a non-randomized study to those treated with TACE (n = 55, 80% CP-A, 42% multinodular, and mean tumor size 6 cm, range 2-10 cm) between 2005 and 2008 (66). While the complete tumor response rate was higher with TARE (12% vs. 4%, p = 0.17), there was no difference in median OS between the two groups (15.0 months for TARE vs. 14.4 months for TACE; p = 0.47). Furthermore, TARE was more likely to induce fatigue (p = 0.003) but less likely to cause fever (p = 0.02). Salem et al. also prospectively compared unresectable HCC patients treated with TARE (n = 123, 54% CP-A, 55% multinodular, and mean tumor size 5 cm, range 2-7 cm) to those treated with TACE (n = 122, 55% CP-A, 53% multinodular, and mean tumor size 3 cm, range 2-6 cm) (67). They found median TTP was longer following TARE (13.3 months vs. 8.4 months, p = 0.046) but median OS did not differ significantly between the two groups (17.4 months vs. 20.5 months, respectively, p = 0.232). Additionally, abdominal pain and increased transaminase activity were more common with TACE (p < 0.05). In the phase III SIRveNIB trial, Chow et al. randomized 360 HCC patients (90% CP-A and 24% with tumor size >50% of liver) to TARE or sorafenib (68). While there was no difference in OS (HR, 1.1; 95% CI, 0.9 to 1.4; p = 0.36), there was an improvement in response rate with TARE (16.5% vs. 1.7%, P < 0.001), as well as reduced grade ≥3 toxicity compared with sorafenib (27.7% vs. 50.6%, P < 0.001). In light of this data, 90 Y radioembolization is considered a viable treatment option for patients with multinodular HCC, but sorafenib remains a standard of care.

RADIATION THERAPY
Technological advances in EBRT such as CT-based treatment planning, management of respiratory motion, understanding of treatment dose distributions, delineation of organs at risk (OAR), and the transition from whole liver irradiation (WLI) to more conformal/dose-escalated treatment regimens have provided the opportunity to offer EBRT safely and effectively. As such, the use of EBRT in patients with HCC has increased in recent years (69). RT techniques, including 3D-conformal radiotherapy (3D-CRT), intensity modulated radiotherapy (IMRT), stereotacticbody radiotherapy (SBRT), and proton beam radiotherapy (PBT), have allowed for the delivery of higher RT doses to tumor volumes compared to historical WLI, and in turn, may have resulted in improved outcomes when compared to other local therapies for HCC (58,70,71).
Different RT techniques have been used for a wide range of patients and within various HCC subgroups. Nevertheless, we must note that the use of RT for patients with CP-C disease is very limited. Furthermore, the data on the use of RT for patients with CP-B disease is still unsettled, as a smaller proportion of patients with CP-B disease were enrolled in clinical trials (Tables 4-7). Therefore, use of RT in this subgroup of patients should be carried out after multi-disciplinary evaluation with individualized treatment for each patient.

Conformal Radiotherapy (CRT)
3D-CRT and IMRT improve the effectiveness of EBRT by increasing the RT dose to the tumor while simultaneously reducing the RT dose to the surrounding normal liver parenchyma when compared to WLI (  (142). The median OS in patients treated with < 40 Gy, 40-50 Gy, and >50 Gy were 6 months, 8 months, and 13 months, respectively. On multivariate analysis, greater RT dose to the tumor was the only significant factor associated with survival (p = 0.01). Other studies also demonstrated that a total dose of >40-50 Gy in standard fractionation led to a higher response or survival rate (73,143,144). Larger radiotherapy doses can often more readily be delivered to smaller tumors in locations where nearby organs are not abutting the tumor. However, the tolerance dose of the normal liver parenchyma, especially in the setting of poor baseline liver function, often limits the use of higher doses of EBRT in the setting of HCC.
IMRT is another conformal RT technique that allows delivery of a higher RT dose when compared to 3D-CRT which may further improve OS without increasing the risk of radiationinduced liver disease (RILD) in patients with ≤3 tumor nodules measuring >3 cm and/or PVT. IMRT uses inverse treatment planning which modulates the intensity of multiple beams to gain a desired target coverage while minimizing the dose to normal structures. Early dosimetric studies comparing IMRT to 3D-CRT suggested that IMRT improves planning target volume (PTV) coverage while maintaining normal tissue tolerances (145). Yoon et al. retrospectively reviewed 187 patients with HCC and CP-A treated with 3D-CRT (n = 122; median fractional and total dose: 1.8 Gy and 45 Gy, respectively) or IMRT (n = 65; median fractional and total dose: 2.5 Gy and 50 Gy, respectively) from 2006 to 2011. Median tumor size (9 cm vs. 10 cm, p = 0.779) and ECOG PS ≤1 (44% vs. 40%, p = 0.557) were similar in both groups and 74% had ≤3 tumor nodules. Patients treated with IMRT had significantly higher 3-year OS (33.4% vs. 13.5 %, P < 0.001), progression-free survival (PFS) (11.1% vs. 6.0%, P = 0.004), and in field-failure-free survival rates (46.8% vs. 28.2%, P = 0.007) when compared to patients treated with 3D-CRT; no difference in RILD was demonstrated (P = 0.716) (146). Similar results were reported by Hou et al. in 118 HCC patients with portal vein and/or inferior vena cava tumor thrombi (81% CP-A and 19% CP-B) (147). Higher RT doses were delivered when IMRT was utilized compared to 3D-CRT (average dose 57.86 ± 7.03 Gy vs. 50.88 ± 6.60 Gy, P ≤ 0.001). Additionally, median OS was significantly higher in patients treated with IMRT compared to 3D-CRT (15.47 months vs. 10.46 months, P = 0.005) while the overall toxicity was similar between the two groups (grade 3 toxicity 5% vs. 2%, P = 0.786). While robust prospective data are lacking, dose escalation with IMRT is considered as a treatment modality in HCC patients with CP-A/B and ≤3 tumor nodules measuring >3 cm and/or PVT.

SBRT
SBRT is a type of EBRT that delivers an highly conformal high dose of RT to a target in 1-5 fractions. One of the first series evaluating SBRT for HCC was described by Blomgren et al. in 1995 (148); since then, this technique has demonstrated excellent outcomes in numerous trials/studies despite often being utilized in patients unsuitable for other therapies believed to have a poor prognosis ( Table 5) .
Although there are no phase III data yet, growing retrospective as well as prospective evidence support promising outcomes with SBRT and its use as an alternative HCC therapy.  (58). One-year liver specific PFS after SBRT compared to RFA was 97.4% vs. 83.6%, and 2-year OS rates were 46% vs. 53%, respectively. For tumors ≥2 cm, LC with RFA was significantly lower compared to SBRT (HR, 3.35; 95% CI, 1.17 to 9.62; p = 0.025). There was no difference in acute grade ≥3 toxicities

Proton Beam Therapy
Protons, unlike photons, exhibit a sharp dose falloff known as the Bragg peak (151). The lack of exit dose with PBT becomes important in the management of HCC as it allows the sparing of large volumes of normal liver parenchyma and other surrounding OAR, which may potentially decrease the risk of toxicity while permitting possible escalation of radiation doses. Numerous single-institutional series have evaluated the efficacy and toxicity of PBT for HCC (   it remains unclear if this translates to a clinically relevant decrease in hepatotoxicity.
Non-randomized data suggests TACE plus RT improves outcomes for HCC patients with partial PVT (155,156  curative surgery due to downstaging. Therefore, the combination of TACE and RT can be considered a treatment option in patients with partial PVT. In the setting (4-6 nodules) of multinodular HCC, TACE plus RT also improves survival in CP-A and B disease. Peng et al. randomized 91 patients to TACE plus RT (23% multinodular disease) or TACE alone (17% multinodular disease) (115). The 1-and 3-year OS rates were significantly higher with TACE plus RT (73% and 36%) compared to TACE alone (52% and 12%) (p < 0.05). Yubing et al. randomized 54 patients to TACE plus RT (63% multinodular disease) or TACE alone (56% multinodular disease) (125). The 1-and 3-year OS rates were significantly higher with TACE plus RT (74% and 30%) compared to TACE alone (60% and 19%) (p < 0.05). Guo et al. randomized 114 patients (71% CP-A, 29% CP-B, and 37% multinodular disease) to TACE plus RT or TACE alone (126). The 1-and 3-year OS rates were significantly higher with TACE plus RT (72% and 45%) compared to TACE alone (52% and 17%) (p < 0.05). There were no differences in toxicities between treatment groups, other than a more frequent elevation in alanine aminotransferase levels with TACE plus RT (Grade ≤ 2, 25% with combination vs. 7% with TACE alone). In light of this prospective data, TACE plus RT should be considered in patients with multinodular HCC.

CONCLUSIONS
While the BCLC classification provides a framework for treatment selection of patients with HCC, it may simplify the decision-making process and may not uniformly take into consideration recent studies, and detailed tumor volumes, which may better guide decisions about local therapies, including radiation therapy and combination treatments. We critically reviewed the literature and devised a data-driven treatment algorithm for optimizing outcomes for patients with unresectable BCLC Stage B or C HCC (Figure 1). This treatment algorithm captures modern data to guide treatment options for those with CP-A or B and ≤3 nodules measuring >3 cm, multinodular disease, or PVT, incorporating tumor volume considerations.
For unresectable, localized HCC patients with either partial PVT, ≤3 nodules >3 cm or multinodular disease, prospective, randomized data suggest that TACE plus RT provides improved survival outcomes and response rates compared to TACE alone; however, toxicities appear more frequent with combination therapy (154). Therefore, appropriate patient selection is important to minimize toxicity. The treatment algorithm used at our institution is shown in Figure 1. For patients with PVT (either partial or complete), RT alone (delivered via IMRT, SBRT or PBT) appears to be a viable option for those unfit to undergo TACE plus RT. One-year survival rates of 44-69% and 30-51% have been observed for those with partial and complete PVT, respectively. For patients without PVT who have ≤3 nodules measuring >3 cm, TACE plus RT results in 1-year survival rates of 75-85%. Lastly, for patients without PVT who have multinodular disease, TACE alone or TARE with 90 Y is a reasonable option for those unfit for TACE plus RT, particularly those presenting with ≥7 nodules.
One-year survival rates of 46-70% have been observed after either modality.
We acknowledge that the best treatment approach is determined through a multidisciplinary management approach in experienced cancer centers with a dedicated HCC program. This offers robust access to all the modalities discussed plus systemic therapy. We hope that this data driven treatment algorithm will aid clinicians in managing localized HCC.