Potential curative therapies for hepatocellular carcinoma (HCC) include orthotopic liver transplantation, resection, and radiofrequency ablation (RFA). The intent of other therapies is to prolong survival by increasing the time to tumor progression (TTP) while not incurring significant hepatic injury. Transarterial chemoembolization (TACE) and sorafenib are the proposed treatments for intermediate and advanced HCC based on the positive results of randomized controlled trials (RCTs).1-3 Transarterial radioembolization (TARE) has gained popularity; however, enthusiasm for the procedure has been tempered by a lack of randomized controlled data.
Abbreviations
BCLC, Barcelona Clinic Liver Cancer; CP, Child-Pugh; HCC, hepatocellular carcinoma; OS, overall survival; PVT, portal vein thrombosis; RCT, randomized controlled trial; RFA, radiofrequency ablation; TACE, transarterial chemoembolization; TARE, transarterial radioembolization; TTP, time to tumor progression.
TARE is a form of intra-arterial therapy that is mechanistically distinct from TACE. TARE, in contrast to TACE, results in minimal postembolization syndrome due to its lack of a macroembolic effect, thereby enabling patients with portal vein thrombosis to be treated safely with TARE and eliminating the need for postprocedural hospitalization.4
Hilgard and colleagues have incorporated TARE into their institutional treatment algorithm based on the BCLC (Barcelona Clinic Liver Cancer) staging system.5 The authors demonstrated a median overall survival (OS) of 16.4 months and TTP of 10 months in this observational study, which consisted of 108 patients with advanced HCC. These results corroborate the experience of others with TARE, and in fact, encompass the most encouraging data that have been reported to date with TARE.6, 7 Moreover, the safety of TARE is validated in the current study.
A few points merit comment. First, patients were selected for TARE if they had unresectable HCC and BCLC C or BCLC A/B that were ineligible for selective TACE (generally ≤ 2 segments) which comprised 49% of the cohort. Interestingly, the proposed candidates for potential TACE for downstaging in the authors' treatment algorithm include tumor criteria (single nodule up to 8-10 cm, 2-3 nodules maximum 3-5 cm, or 4-5 nodules ≤ 3 cm) in which the ability to perform “selective” therapy in all such cases is questionable, because tumor size, number, and location determine the extent of selectivity of intra-arterial therapy. The assumption that lobar TARE is a “safer” alternative compared to lobar TACE given the same tumor characteristics still requires further investigation. It is recognized that the concept of “selective” TACE is the preferred mode; however, the reality is that this is ill-defined from center to center (2nd, 3rd order, etc.). In addition, most tumors are large and disease is bilobar, which often does not permit “selective infusion.”
Of particular interest, the median OS of BCLC C patients was not reached at the time of publication. In the largest reported single-center experience with TARE, the median OS in BCLC C patients was 7.3 months and varied according to Child-Pugh (CP) classification, in which 55% were non–CP A.6 In contrast, this European cohort was composed of 22% CP B, limited to CP-7. Additionally, the results of Hilgard et al. are favorable compared to the median OS of 8.9 months in the SHARP trial (95% CP A) among patients with portal vein thrombosis (PVT) and/or metastatic disease who received sorafenib. However, direct comparisons are limited across studies but provide a compulsion for future studies for CP A patients with PVT comparing these therapeutic modalities. Among the BCLC B patients, median survival was 16.4 months, which is comparable to earlier reports with TARE in this patient population.
Radiographic response rates were lower (World Health Organization [WHO] 15% and European Association for the Study of the Liver [EASL] WHO 40% at 3 months), despite excluding 32 patients lacking well-defined tumor margins (39% of cohort had >50% tumor burden), than a large U.S. cohort (WHO 42% and EASL WHO 57%). Longer radiographic follow-up may have resulted in increased response rate; median time to WHO partial response and EASL WHO with TARE has been reported to be 6.6 months and 2.1 months, respectively, with lower WHO response in lesions >10 cm. Nonetheless, Hilgard et al. reported a superior overall TTP 10 months compared to 7.9 months in the same U.S. cohort. The authors offer the explanation that treatment with lobar 90Y may have treated the known field defect appreciable in HCC (improving TTP by treating nontarget microscopic disease), but lacked the delivery of higher doses of radiation to the targeted lesion that is achieved with selective TARE, which leads to a greater tumoricidal effect and hence a superior radiographic response. Methodological discrepancies in the assessment of radiographic response likely also contributed to these differences. In Salem et al., any progression that would have clinically led to repeat therapy was adjudicated as disease progression including those with <25% progression by WHO criteria, and hence may have lowered TTP.8 Additionally, Hilgard used the more recently endorsed modified RECIST criteria for TTP whereas the earlier study employed WHO.9 Lastly, Hilgard did not deem the development of PVT as disease progression barring stability of the tumor lesion. Such differences underscore the need for standard methods across studies.
Treatment trials have traditionally excluded CP B patients due to the competing risk of death from hepatic decompensation which can obscure any potential treatment benefit. In the current study, the median OS for CP B (CP-7) was 6 months. Similarly, in the study by Salem et al., the median OS was 5.6 months in CP B patients with PVT, which questioned the utility of TARE in such a patient population. However, more granular data, showed a median OS of 14.8 months in CP B without vascular invasion. Moreover, TTP showed a comparable degree of benefit assessed by the hazards ratio in CP A and B patients among radiographic responders, supporting a potential therapeutic benefit despite compromised liver function. The use of lobar versus selective TARE and effect clinical endpoints of TTP and OS becomes of particular interest in CP B.
The safety and tolerability of any therapy is paramount in patients with underlying cirrhosis. Fatigue was confirmed to be the most prevalent adverse event post-TARE. In contrast to sorafenib, this symptom is short lived and generally abates within 1 week following TARE. With proper identification of nontarget sites and appropriate coiling of collaterals, no cases of radiation pneumonitis or gastric ulcers were reported in this cohort. The results of the MAA scan excluded 7.7% of patients as candidates for TARE; exceeding 1.7% in the U.S. cohort. Larger tumors are associated with a higher degree of intratumoral shunting and likely contributed to a higher screening failure.
Now that we have Hilgard's confirmatory study from a European cohort on the safety and efficacy of TARE, where does it fit relative to alternative therapies? To date, comparative data between TACE and TARE is limited to single center retrospective analyses, which have concluded therapeutic equivalency in terms of OS. (Table 1).10-13 TARE compared to TACE has been reported to be superior in the ability to downstage T3 to T2, shorter median time to radiographic response and associated with significantly prolonged TTP. The potential implications for patients listed for orthotopic liver transplantation (i.e., enabling patients to wait longer without drop out) are merely speculative. Moreover, data supports the prognostic role of the response to liver directed therapy acting as a biological stress test to provide insight into a tumor's aggressiveness.14 Any differences exerted in selection pressure by different forms of LDT remains to be seen and can only be addressed in well developed randomized controlled trials. Data comparing sorafenib to TARE in patients with PVT is even sparser, currently existing only across studies and therefore less clinically meaningful.
| TARE vs. TACE | Lewandowski et al. (7)N = 43 vs. 35 | Carr et al. (8)N = 99 vs. 691 | Kooby et al.‡ (9)N = 27 vs. 44 | Salem et al. (10)N = 123 vs. 122 |
|---|---|---|---|---|
| Median OS (months) | 35.7 vs. 18.7; P = 0.18 | 11.5 vs. 8.5;P < 0.05* | 6 vs. 6; P = 0.74 | 20.5 vs. 17.4;P = 0.23 |
| Radiographic response: | ||||
| WHO Response (%) | 61 vs. 37; P = 0.12 | 41 vs. 60† | 11 vs. 6; P = 0.73§ | 49 vs. 36; P = 0.10 |
| T3 to T2 (%) | 58 vs. 31; P = 0.023 | N/A | N/A | N/A |
| TTP (months) | 33.3 vs. 12.8; P = 0.005 | N/A | N/A | 13.1 vs. 8.4;P = 0.023 |
| Tolerability | N/A | |||
| Median hospitalization (days) | 0 vs. 2; P < 0.001 | 1.7 vs. 5.0; P = 0.05 | 0 vs. 1.8; P < 0.001 | |
| Any complication (%) | 44 vs. 70; P = 0.05 | |||
| Hyperbilirubinemia (%) | Grade 3/4: 26 vs. 7 | >3 mg/dL: 4 vs. 16;P = 0.1 |
- * OS between 90Y and TACE became nonsignificant after adjusting for baseline bilirubin, presence of PVT, and baseline AFP level.
- † Single dose of TARE to lobe with dominant disease burden; 43% bilobar in TARE. TACE q 8-10 weeks.
- ‡ Sir-Spheres
- § Radiographic response by RECIST at 3 months.
To this end, RCTs comparing standard of care (TACE, sorafenib) to TARE are warranted. Logistic concerns include the number of patients required; a power calculation performed to determine the sample size to demonstrate therapeutic equivalency between TACE and TARE in BCLC B patients showed that more than 1000 patients would be needed.13 The feasibility of a large trial due to cost and the number of centers with adequate expertise in both treatment modalities requires careful consideration; however, the number of centers utilizing TARE appears to be increasing making this less of a limitation for conducting such a trial. Lastly, stratification for lobar versus selective treatment and standardization of TACE methodologies would be required given differences in treatment practices. In BCLC C patients, the anticipated trial design would be sorafenib ± TARE with a primary endpoint of TTP.
There are several examples of accepted treatment practices for HCC that are based on cohort analyses (not RCTs) that have been accepted into treatment guidelines including RFA (<3 cm) versus hepatic resection, transplantation versus hepatic resection, and open versus laparoscopic hepatic resection. Such trials for TARE are unlikely to come to fruition. TARE is currently not recognized by the American Association for the Study of Liver Diseases or EASL in the management of HCC due to lack of randomized data. However the National Comprehensive Cancer Networks have endorsed TARE as one of the treatment options for HCC.15 At our institution on ongoing RCT (PREMIERE Trial) is comparing TARE to various liver directed therapies (RFA, TACE, or RFA+TACE) based on tumor size and number.16
While TARE has shown promising results in observational single center cohorts, which is further, supported by Hilgard's contribution and validates the results of a large US cohort, it is unlikely to gain acceptance into the BCLC staging system without being scrutinized in a head to head manner against the current standard of care therapies. Given that it is unlikely that RCT data of TARE versus TACE will emerge in the near future, given prohibitive statistical barriers to completion, center experience will likely continue to play a dominant role in the preference of therapy and treatment algorithm for HCC.
