The place for TACE combinations in hepatocellular carcinoma is still uncertain

A lack of clear overall survival (OS) benefit, the absence of mature OS data and high rates of toxicity highlight more questions than answers in phase III EMERALD trials in patients with embolisation-eligible disease according to mixed findings presented at the ESMO Gastrointestinal Cancers Congress 2026 (Munich, 1–4 July).

In a final analysis of the phase III EMERALD-1 trial, durvalumab with or without bevacizumab and transarterial chemoembolisation (TACE) did not improve OS compared with TACE alone in 616 patients with unresectable embolisation-eligible hepatocellular carcinoma (eeHCC) (Abstract 183O). More encouraging data were presented from the phase III EMERALD-3 study of tremelimumab plus durvalumab in the ‘STRIDE’ regimen (single-dose tremelimumab with regular-interval durvalumab) with or without lenvatinib plus TACE in 760 patients with eeHCC, with continued improvements observed in progression-free survival (PFS) at a second data cutoff, and a trend towards improved OS also reported (LBA2).

In EMERALD-1, median OS was 29.9 months versus 33.3 months (hazard ratio [HR] 1.10; 95% confidence interval [CI] 0.87–1.39; p=0.470) for durvalumab plus bevacizumab plus TACE versus TACE, and 33.6 months versus 33.3 months (HR 0.93; 95% CI 0.74–1.19; p=0.666) for durvalumab plus TACE versus TACE, with consistent results reported across pre-specified subgroups.

“The OS data from EMERALD-1 are disappointing since the previously reported benefit in PFS with the TACE combination (15.0 months with durvalumab plus bevacizumab versus 10.0 months with durvalumab and 8.2 months with TACE alone did not translate into prolonged OS (Lancet. 2025;405:216–232),” states Prof. Jens Ricke from University Hospital, LMU, Munich, Germany, suggesting that there may not be a role for the durvalumab plus bevacizumab plus TACE regimen in patients with unresectable eeHCC.

“The trade-off in terms of toxicity should also be considered when reviewing the efficacy data: without prolonged OS, I do not believe the PFS benefit is worth the high rate of adverse events (AEs) patients experienced,” he adds. Grade 3–4 AEs and grade 3–4 AEs possibly related to study treatment were reported in a greater proportion of patients receiving durvalumab plus bevacizumab plus TACE (47%/27%) than in those receiving durvalumab plus TACE (31%/8%) or TACE alone (25%/8%). Serious AEs were experienced by more than half (52%) of those in the durvalumab plus bevacizumab plus TACE arm and in at least one-third of patients in the durvalumab plus TACE (38%) and TACE (34%) arms.

In EMERALD-3, the PFS improvements previously reported at the first data cutoff (September 2025; HR 0.70 [95% CI, 0.57–0.86]; p=0.0007) continued to be observed at a second data cutoff (February 2026), with a clinically meaningful improvement for STRIDE plus TACE versus TACE (HR 0.71; 95% CI 0.56–0.91) per blinded independent central review. The objective response rate (ORR) was also improved with both STRIDE plus lenvatinib plus TACE (38.9%) and STRIDE plus TACE (40.8%) compared with TACE alone (27.0%) at the second data cutoff. Notably, a trend towards improved OS was recently reported for the STRIDE plus lenvatinib plus TACE (HR 0.84; 95% CI 0.65–1.09) and STRIDE plus TACE (HR 0.70; 95% CI 0.51–0.95) combinations compared with TACE alone (J Clin Oncol. 2026;44(Suppl 17):LBA4000).

“The STRIDE plus TACE regimen seems to provide greater benefit with fewer AEs than the lenvatinib combination,” Ricke emphasises. In previous research, treatment-related grade 3–4 AEs were reported in 63% of patients receiving STRIDE plus lenvatinib plus TACE compared with 49% of those receiving STRIDE plus TACE (J Clin Oncol. 2026;44(Suppl 17):LBA4000). “This combination may be worth the price patients are willing to pay in treatment-related toxicity, although we need to see mature data. An additional benefit for the patient of omitting lenvatinib at this stage is that it will be available for subsequent lines of treatment,” he adds.

Based on current evidence, Ricke believes so, “but only for patients in whom embolisation eligibility is clearly established and likely those outside ‘up-to-seven’ criteria – i.e. the sum of the largest tumour size in cm plus number of tumours. This was presented recently, with increased benefit found in more advanced versus localised tumours in subgroup analyses of EMERALD-3 (J Clin Oncol. 2026;44(Suppl 17):LBA4000). Disease that exceeds ‘up-to-seven’ is less likely to respond to TACE and may benefit from systemic therapy (World J Gastroenterol. 2022;28:2561–2568). Equally, patients with earlier-stage disease who meet the ‘up-to-seven’ criteria may require TACE – or any locoregional treatment such as microwave ablation or transarterial radioembolisation – only. Ultimately, mature OS data and patient-reported outcomes that help to weigh the benefits of treatment against treatment-related AEs, will be needed to determine the role of TACE combinations in the treatment of eeHCC.”

Further considerations emerge from these presentations in terms of clinical trial design. For many years, PFS has been used as a surrogate endpoint for OS to provide an early indication of survival benefit, but the EMERALD-1 data clearly highlight an issue with this. Currently there are a number of TACE with or without immunotherapy HCC trials (in addition to EMERALD-1) that have reported PFS benefits but questionable or immature OS data with TACE combinations: phase III TALENTACE (Ann Oncol. 2025;36(Suppl 1):LBA2), phase III LEAP-012 (Lancet. 2025;405:203–215), and a phase III trial of TACE plus camrelizumab plus rivoceranib (J Clin Oncol. 2026;44(Suppl 16):Abstract 4001). The difficult role of PFS as an outcome in patients with HCC has been a matter of earlier debates, including systemic trials (J Hepatol. 2019;70:1262–1277). Ricke also feels that while it is clearly desirable to demonstrate prolonged OS, evidence of quality-adjusted OS benefit through patient-reported outcomes is necessary to fully consider the significant toxicity associated with some of these regimens from the patient’s perspective. That EMERALD-3 included ORR as a secondary endpoint is also interesting as he concludes, “We are entering an era where downstaging for either transplantation or resection is becoming feasible, and the ORR data are extremely encouraging with the STRIDE plus TACE regimen.”