The ESMO Gastrointestinal Cancers Congress 2026 provided important new insights into the management of hepatobiliary malignancies. The most relevant discussions focused on the integration of systemic and locoregional treatments in hepatocellular carcinoma, the appropriate selection of patients for transarterial chemoembolisation, and the increasing centrality of molecular profiling in cholangiocarcinoma.
Although not all the studies presented were immediately practice-changing, the meeting clarified several unresolved clinical questions and highlighted the direction in which the treatment landscape is evolving.
The role of TACE-based combinations
One of the major themes of ESMO GI 2026 was the treatment of patients with unresectable, embolisation-eligible hepatocellular carcinoma. For several years, investigators have explored whether combining transarterial chemoembolisation (TACE) with immune checkpoint inhibition and antiangiogenic therapy could improve outcomes compared with TACE alone.
At ESMO GI 2026, the final overall survival results from EMERALD-1 and updated tumour-response analyses from EMERALD-3 provided two different but complementary perspectives.
EMERALD-1: Final Overall Survival
EMERALD-1 was a phase III trial enrolling 616 patients with unresectable, embolisation-eligible HCC. Patients received:
- durvalumab plus bevacizumab and TACE;
- durvalumab plus TACE;
- or TACE alone.
The durvalumab–bevacizumab combination had previously demonstrated an improvement in progression-free survival. The final analysis presented at ESMO GI 2026 showed no overall survival advantage.
Median overall survival was 29.9 months with durvalumab, bevacizumab and TACE; 33.6 months with durvalumab and TACE; and 33.3 months with TACE alone.
The hazard ratio for the triplet versus TACE was 1.10, while the hazard ratio for durvalumab plus TACE versus TACE was 0.93. Neither comparison was statistically significant. Grade 3–4 adverse events considered potentially treatment-related occurred in approximately 27% of patients receiving durvalumab, bevacizumab and TACE, compared with about 8% in each of the other two groups.
The PFS benefit in intermediate-stage HCC did not translate into improved overall survival. The experimental strategy added clinically relevant toxicity and may have affected subsequent treatment opportunities.
These results represent an important cautionary message
A progression-free survival benefit in intermediate-stage HCC did not translate into improved overall survival, particularly when the experimental strategy added clinically relevant toxicity. EMERALD-1 raises questions regarding the use of PFS as a surrogate endpoint in TACE-combination trials and underscores the importance of evaluating post-progression treatment, liver-function deterioration and early mortality.
EMERALD-3: STRIDE Plus TACE
The phase III EMERALD-3 study evaluated the STRIDE regimen — single-dose tremelimumab followed by durvalumab — with or without lenvatinib, in combination with TACE. A total of 760 patients with embolisation-eligible HCC were enrolled.
At the updated analysis, STRIDE plus TACE continued to show a clinically meaningful improvement in progression-free survival compared with TACE alone, with a hazard ratio of 0.71. Objective response rates at the second data cut-off were:
- 38.9% with STRIDE, lenvatinib and TACE;
- 40.8% with STRIDE and TACE;
- 27.0% with TACE alone.
A favourable overall survival trend was also reported, particularly for STRIDE plus TACE, although longer follow-up remains necessary before definitive conclusions can be drawn.
The addition of lenvatinib did not appear to produce a clear incremental improvement in response compared with STRIDE plus TACE and was associated with greater toxicity in earlier analyses. The simpler STRIDE–TACE combination may offer a more favourable balance between activity, tolerability and preservation of lenvatinib as a subsequent-line option.
Patient Selection
Taken together, EMERALD-1 and EMERALD-3 do not support the indiscriminate addition of systemic therapy to TACE. They reinforce the need to identify the patients most likely to benefit from combination treatment.
Patients with limited tumour burden within the up-to-seven criteria may still achieve adequate disease control with locoregional therapy alone. Patients with more extensive liver-confined disease, particularly those beyond up-to-seven criteria but still technically eligible for embolisation, may represent a more appropriate population for combined systemic and locoregional strategies.
“Embolisation eligible” should not be interpreted solely as a technical definition. Tumour burden, distribution of liver lesions, hepatic reserve, expected number of TACE procedures and the likelihood of maintaining eligibility for systemic treatment should all contribute to the therapeutic decision.
The ESMO GI 2026 data also highlighted the importance of multidisciplinary assessment involving medical oncologists, hepatologists, surgeons, interventional radiologists and transplant specialists.
Response and Downstaging
The response rates observed with STRIDE plus TACE are particularly relevant because systemic–locoregional combinations may increasingly be used not only to delay progression but also to facilitate downstaging.
For selected patients, a profound radiological response could potentially reopen discussions regarding liver resection, ablation or transplantation. However, this approach remains investigational. Prospective evidence is needed to define the most appropriate response criteria, the optimal timing of surgical reassessment, the duration of immunotherapy, and the safety of transplantation following immune checkpoint inhibition.
High objective response rates should not automatically be equated with curative conversion, but they offer an important basis for future studies.
Molecular Profiling in Cholangiocarcinoma
In cholangiocarcinoma, ESMO GI 2026 reinforced that comprehensive molecular profiling is no longer optional.
The therapeutic landscape of biliary tract cancer is progressively being divided into molecularly defined subgroups, including tumours harbouring IDH1 mutations, FGFR2 fusions or rearrangements, HER2 amplification or overexpression, BRAF V600E mutations, NTRK or RET fusions, KRAS G12C mutations, and MSI-high or mismatch-repair-deficient status.
This evolution requires high-quality and timely molecular testing, preferably performed at diagnosis of advanced disease rather than after failure of first-line treatment.
FGFR2 Testing
A recurrent point of discussion at ESMO GI 2026 was that not all molecular techniques have equivalent sensitivity for detecting FGFR2 alterations.
In intrahepatic cholangiocarcinoma, the use of limited DNA panels may fail to identify uncommon fusion partners or complex rearrangements. Hybrid-capture next-generation sequencing, ideally including RNA-based analysis when available, may improve detection.
An inadequate test could incorrectly classify a patient as FGFR2-negative and prevent access to an effective targeted treatment. The molecular platform used is as important as the decision to request molecular profiling.
FGFR Inhibitors and Resistance
Several studies and translational analyses presented at the meeting continued to investigate next-generation FGFR inhibitors, including compounds designed to overcome acquired resistance to earlier FGFR2-directed treatments.
The emerging clinical problem is no longer simply identifying an FGFR2 fusion. It is understanding how the tumour evolves under selective pressure from FGFR inhibition. Secondary FGFR2 kinase-domain mutations may cause resistance, often with heterogeneous resistant clones emerging in the same patient. Liquid biopsy and repeated molecular testing at progression may therefore become increasingly important for selecting subsequent FGFR inhibitors.
The development of irreversible or highly selective next-generation agents, including fanregratinib and other investigational FGFR inhibitors, reflects the effort to obtain more durable control and retain activity against resistance mutations. The optimal sequencing of FGFR-directed therapies remains undefined.
ProvIDHe: Ivosidenib in IDH1-Mutated Cholangiocarcinoma
Updated findings from ProvIDHe, an international phase IIIb, single-arm study, further supported the role of ivosidenib in previously treated, IDH1-mutated, locally advanced or metastatic cholangiocarcinoma. The analysis included 298 patients from 15 countries.
Median progression-free survival was 5.6 months, while median overall survival was 16.4 months. A particularly relevant subgroup consisted of 53 patients who remained on treatment for at least 12 months. Among these long-term responders, median PFS was 19.6 months, median OS had not been reached, and disease control was observed in all patients.
Although radiological tumour shrinkage may not be dramatic in every patient, IDH1 inhibition can produce prolonged disease stabilisation in a clinically meaningful subgroup. No definitive clinical or biochemical marker has yet been identified to predict long-term benefit.
Real-World Ivosidenib Outcomes
A separate retrospective analysis evaluated 55 US patients receiving second-line ivosidenib for advanced IDH1-mutated cholangiocarcinoma. According to physician-assessed responses, 27.3% achieved a partial response, 60.0% achieved stable disease, and the overall disease-control rate was 87.3%. Median overall survival was 18.5 months, and the estimated 12-month survival rate was 74.2%.
The treatment was well tolerated, with no adverse events resulting in permanent treatment discontinuation. Although retrospective data should be interpreted cautiously, these findings support the effectiveness and feasibility of ivosidenib outside a strictly controlled clinical-trial population.
Immunotherapy in Biliary Tract Cancer
Following the establishment of chemoimmunotherapy as a first-line standard for advanced biliary tract cancer, current research is moving beyond the question of whether immunotherapy is active.
The new questions include which patients derive the greatest benefit, how the benefit differs according to anatomical subtype, whether molecular alterations influence immune sensitivity, and how immunotherapy should be sequenced with targeted agents.
Real-world analyses presented at ESMO GI 2026 explored treatment outcomes and prognostic factors among patients receiving immunotherapy-based regimens. These studies support the feasibility of chemoimmunotherapy in routine practice but also underline the biological heterogeneity of intrahepatic, extrahepatic and gallbladder cancers.
At present, no routinely available biomarker other than MSI-high or mismatch-repair deficiency reliably identifies patients with markedly enhanced sensitivity to immune checkpoint inhibition.
Combined HCC–Cholangiocarcinoma
Combined hepatocellular–cholangiocarcinoma is a rare and biologically heterogeneous primary liver malignancy for which there is no universally accepted systemic-treatment standard.
A retrospective comparison presented at the congress evaluated durvalumab plus gemcitabine–cisplatin in combined HCC–cholangiocarcinoma and conventional cholangiocarcinoma. Although limited by small patient numbers and retrospective selection, the analysis supports the use of biliary tract cancer-oriented chemoimmunotherapy in patients whose tumours display a predominant cholangiocarcinoma phenotype.
Prospective, histologically and molecularly annotated studies are needed because combined tumours may contain highly variable proportions of hepatocellular and cholangiocellular differentiation.
Additional Research Directions
Beyond the major clinical studies, ESMO GI 2026 included investigations into epidemiological and clinical features of cholangiocarcinoma arising in patients with chronic liver disease, international patterns of treatment and molecular testing, outcomes according to IDH1, FGFR2 and other actionable alterations, real-world effectiveness of targeted therapies, mechanisms of acquired resistance, prognostic biomarkers in patients receiving chemoimmunotherapy, and improved pathological and molecular classification of primary liver cancers.
These studies address important gaps between the results of registration trials and treatment delivery in everyday practice.
You can also read about Highlights of ESMO GI 2026 in Colorectal Cancer with Davide Ciardiello on OncoDaily.
Key Takeaways from ESMO GI 2026
The hepatobiliary cancer programme delivered several practical messages.
First, PFS improvement alone may not be sufficient to establish a new treatment standard in embolisation-eligible HCC. EMERALD-1 demonstrated that better disease control does not inevitably translate into longer survival, particularly when treatment increases toxicity.
Second, STRIDE plus TACE remains promising but requires mature survival and patient-reported outcome data. The appropriate population is likely to be a biologically and clinically selected subgroup rather than all patients eligible for embolisation.
Third, the boundaries between intermediate and advanced HCC are becoming less rigid. Treatment decisions should reflect tumour burden, liver function, technical feasibility of locoregional treatment and the possibility of subsequent systemic therapy.
Fourth, molecular profiling should be performed early in advanced cholangiocarcinoma. Waiting until second-line treatment may result in clinical deterioration before an actionable alteration can be therapeutically exploited.
Fifth, testing methodology matters. Hybrid-capture DNA sequencing and, where appropriate, RNA-based testing can reduce the risk of missing clinically relevant FGFR2 fusions.
Finally, targeted therapy is producing a growing population of long-term responders, but resistance remains inevitable for many patients. Serial molecular profiling and rational sequencing strategies will be central to the next phase of precision treatment.
Conclusion
ESMO GI 2026 moved the field towards a more selective and individualised model of care. In hepatocellular carcinoma, the central challenge is to determine which patients truly benefit from adding systemic therapy to locoregional treatment and which patients may be adequately treated with TACE alone.
In cholangiocarcinoma, optimal care increasingly depends on identifying the molecular driver, using the correct diagnostic platform and delivering targeted treatment before clinical deterioration.
The future of hepatobiliary oncology will be defined not simply by the availability of new drugs, but by better patient selection, high-quality molecular diagnostics and genuinely multidisciplinary decision-making.
Find more videos on OncoDaily YouTube TV.
Written by Maria Maddalena Laterza, MD, PhD