Locoregional Therapy for Hepatocellular Carcinoma: Transarterial Chemoembolization Versus Transarterial Radioembolization Versus Stereotactic Body Radiotherapy

Watch a video presentation of this article

Abbreviations

BCLC A

Barcelona Clinic Liver Cancer stage A

CP

Child‐Pugh

CR

complete response

DEB

drug eluting beads

HCC

hepatocellular carcinoma

LRT

locoregional therapy

LT

liver transplant

mRECIST

modified response evaluation criteria in solid tumors

MVI

macrovascular invasion

PFS

progression free survival

PVT

portal vein thrombosis

RCT

randomized controlled trial

RFA

radiofrequency ablation

RS

radiation segmentectomy

SBRT

stereotactic body radiotherapy

TACE

transarterial chemoembolization

TARE

transarterial radioembolization

TKI

tyrosine kinase inhibitor

TTP

time to progression

TTST

time to secondary therapy

The use of locoregional therapy (LRT) has various potential roles in the treatment of hepatocellular carcinoma (HCC.) It is important to determine what the intended goal of LRT is prior to receipt of LRT. The choice of one form of therapy over another is often dictated by center preference and more importantly experience. This article will review the various forms of LRT that are often used in the treatment of HCC.

Bridge/down‐staging To Liver Transplant

Transarterial chemoembolization (TACE) is the most common form of LRT employed as a bridge or down‐staging modality to LT. Factors that have been associated with dropout while awaiting LT include: one tumor of 3.1 to 5 cm (versus one tumor ≤3 cm), two or three tumors (versus a single tumor), a lack of a complete response to the first LRT, and a high alpha‐fetoprotein level (>20 ng/mL) after the first LRT.1 Meeting any of these criteria is associated with a 1‐ and 2‐year probability of dropout of 21.6% and 26.5%, respectively. The American Association for the Study of Liver Disease recommends that patients with HCC who are listed for liver transplant (LT) be treated with LRT to prevent tumor progression and hence wait‐list dropout.2 No form of LRT is recommended over another to bridge to LT.

TACE Versus Y90

A single‐center randomized controlled trial (RCT) comparing TACE with Y90 in unresectable HCC was conducted in 45 patients who were primarily Barcelona Clinic Liver Cancer stage A (BCLC A). The primary endpoint, time to progression (TTP), was significantly longer in the Y90 group, (not reached, >26 months) compared with 6.8 months in the TACE group (P = 0.0012).3 In contrast with prior smaller pilot RCTs comparing these two intra‐arterial therapies, which allowed for Y90 administration once with TACE performed every 6 weeks until complete response was achieved, both therapies were given on demand based on radiographic response. Overall survival (OS) censored to LT was similar between the two groups (TACE 17.7 versus Y90: 18.6 months). Although this trial was stopped early because of slow accrual, a post hoc conditional analysis using the 45 patients enrolled reported that the chance of an erroneous conclusion that Y90 significantly prolongs TTP compared with TACE was only 3.2%. Based on these and other results with radioembolization, Y90 has been adopted as the first‐line intra‐arterial therapy in some institutions.4

Radiation segmentectomy (RS) involves high doses of Y90 injected into one to two hepatic segments. A median OS of 53.4 months was reported in 102 patients (51% CP B) with a single lesion smaller than 5 cm not amenable to radiofrequency ablation (RFA).5 Among those who subsequently underwent resection after RS, a tumor dose greater than 190 Gy was associated with a higher rate of complete pathological necrosis. Recently, a median OS of CP A patients with a single lesion smaller than 5 cm was 6.7 years, offering results comparable with curative intent, including LT, resection, and RFA.6 Retrospective analysis with propensity matching in patients with a single lesion less than 3 cm who were treated with RS versus TACE + microwave ablation reported similar TTP, response rates, and OS.7 Although there was a greater portion of progression of the targeted lesion in the combination group compared with Y90, conclusions cannot be drawn because of differences in follow‐up.

Another study retrospectively compared RS with segmental TACE in patients with a solitary lesion ≤3 cm. After propensity matching, complete radiographic response (mRECIST) and time to secondary therapy (TTST) significantly favored RS compared with segmental TACE (CR: 92.1% versus 52.6%, TTST: 812 versus 162 days, respectively), whereas OS was not significantly different.8

SBRT

Stereotactic body radiotherapy (SBRT) is also being used a bridge to LT. In a single‐center study from Toronto, an intention‐to‐treat analysis among patients who received SBRT, TACE, or RFA based on recommendations of a multidisciplinary tumor board showed no significant difference in dropout rate, OS from listing, or LT in the SBRT group compared with TACE or RFA.9 Notably, SBRT was considered an alternative therapy when TACE or RFA was not deemed feasible or did not lead to tumor control, and it was the preferred therapy with borderline liver function.

Combination Therapy: SBRT + TACE

In retrospective studies, combination of TACE + SBRT has been reported to significantly decrease recurrence compared with TACE alone in lesions ≥3 cm (10.8% versus 25.8%; P = 0.04) and improved OS censored to LT (33 versus 20 months; P = 0.02).10 Similarly, improved median OS has been reported with combination therapy compared with SBRT alone in lesions greater than 5 cm (42 versus 21 months).11 A phase 2 trial of adjuvant SBRT after an incomplete response to TACE in lesions smaller than 10 cm without extrahepatic disease or macrovascular invasion (MVI) showed a 2‐year local control rate of 94.6%; however, there were noted grade 3/4 gastrointestinal toxicities in three patients.12 An ongoing RCT is aiming to determine whether repeat TACE versus SBRT is the preferred therapy in patients with residual or recurrent HCC after initial TACE.13

A meta‐analysis of 25 trials (11 RCTs, 18 included patients with MVI) all conducted in Asia showed significantly improved pool OS in TACE + radiation therapy (22.7 months) compared with TACE alone (13.5 months).14 There were, however, significantly higher rates of gastric ulcerations and abnormal transaminases/bilirubin. A subanalysis of those with PVT did not alter these results.

Combination Therapy: LRT + Systemic Therapy

There has been intense interest in combining LRT with systemic therapy. Most of these data come from the combination of TACE with tyrosine kinase inhibitors (TKIs) in a BCLC B population based on the biological plausibility of sorafenib blocking the angiogenic flare associated with embolization‐induced hypoxemia, and thus prolonging TTP. Large RCTs have failed to show a benefit of TACE/DEB + TKIs compared with TACE monotherapy.15 Several reasons potentially explain these results, including initiation of TKI relative to TACE/DEB, dosage of TKI, duration, and early terminations of trials. TATICS (Transcatheter Arterial Chemoembolization Therapy in Combination With Sorafenib), a phase II trial from Japan, did demonstrate a significantly improved PFS in TACE + sorafenib compared with sorafenib (25.2 versus 13.5 months; hazard ratio, 0.59; 95% confidence interval: 0.41‐0.87; P = 0.006).16 It should be noted that the study design in this trial was unique in that sorafenib was continued despite new intrahepatic lesions. Results of OS are pending. Trials are also being conducted comparing TKIs combined with Y90 (STOP HCC) and SBRT (RGOT 1112).

Similar to bridging to LT, the results for down‐staging to transplant have been predominately reported with TACE. A systemic review reported that the success rate of down‐staging to the Milan criteria is 48%.17 The use of Y90 versus TACE did not significantly impact the ability to down‐stage to the Milan criteria.

The choice of LRT is dependent on multiple variables including the tumor burden, degree of liver dysfunction, and goal of LRT. When attempting to bridge/down‐stage to LT, this decision is often determined by center‐specific experience and algorithms.