For patients with advanced hepatocellular carcinoma (HCC), systemic therapy is generally continued until disease progression or unacceptable toxicity. With increasingly effective systemic treatments, however, some patients may achieve sufficient disease control to become candidates for liver resection, raising interest in conversion surgery as a potential treatment strategy.
The phase 3 TALENTOP trial evaluated this strategy in a selected group of patients with locally advanced HCC and macrovascular invasion who had no extrahepatic metastases. The study compared liver resection followed by atezolizumab plus bevacizumab with continued atezolizumab plus bevacizumab in patients whose disease remained controlled after initial systemic therapy and who were considered suitable for surgery.
The study, titled “Liver resection after atezolizumab and bevacizumab versus maintenance therapy for locally advanced hepatocellular carcinoma (TALENTOP): a multicentre, open-label, randomised, phase 3 trial,” was published on August 22, 2026, in The Lancet.
Authors: Hui-Chuan Sun, Xiao-Dong Zhu, Feng Shen, Tian-Qiang Song, Xue-Li Bai, Yong-Yi Zeng, Min-Shan Chen, Zhi-Yong Huang, Jian-Qiang Cai, Jie Gao, Ming Kuang, De-Yu Li, Bao-Cai Xing, Xiang-Cheng Li, Tian-Fu Wen, Tao Peng, Yao-Dong Wang, Lian-Xin Liu, Cheng Huang, Ying-Hong Shi, Xiao-Wu Huang, Kui Wang, Qiang Gao, Jian Zhou, and Jia Fan.
TALENTOP Trial Design
TALENTOP was a multicenter, open-label, randomized phase 3 trial conducted across 24 hospitals in China. Eligible patients were treatment-naive adults with HCC, macrovascular invasion, no extrahepatic spread, Child–Pugh class A liver function, and an ECOG performance status of 0 or 1. All patients first received induction treatment with three cycles of atezolizumab 1200 mg plus bevacizumab 15 mg/kg every 3 weeks, followed by one cycle of atezolizumab alone.
Among 489 patients who entered the induction phase, 201 patients were subsequently randomized. To proceed to randomization, patients had to complete induction treatment, have a partial response or stable disease according to RECIST 1.1, and be considered suitable for complete surgical resection.
Patients were assigned 1:1 to liver resection followed by 12 months of atezolizumab plus bevacizumab, beginning 4–6 weeks after surgery, or to continued atezolizumab plus bevacizumab until loss of clinical benefit or unacceptable toxicity. A total of 101 patients were assigned to the surgery group and 100 to maintenance therapy.
The primary endpoint was independently assessed time to treatment failure, defined as the time from randomization to local recurrence or disease progression, development of extrahepatic spread, or death.
Longer Time to Treatment Failure With Surgery
At the January 27, 2025 data cutoff, median follow-up was 18.4 months. The primary endpoint was met. Median time to treatment failure was 20.4 months with the surgical strategy compared with 11.8 months with maintenance atezolizumab plus bevacizumab alone. This corresponded to a hazard ratio of 0.60 (95% CI 0.39–0.91; p=0.015), meeting the prespecified statistical boundary for the interim analysis.
Treatment failure had occurred in 37% of patients assigned to surgery and 59% of those assigned to maintenance therapy. The benefit observed with surgery was generally consistent across prespecified clinically relevant subgroups.
Results were also similar when time to treatment failure was assessed by investigators rather than independent review. Investigator-assessed median time to treatment failure according to RECIST 1.1 was 18.0 months in the surgery group and 11.8 months in the maintenance group, with an HR of 0.60.
Surgical Outcomes After Systemic Therapy
Not all patients randomized to surgery ultimately underwent resection. Of the 101 patients assigned to this group, 83 underwent liver resection and subsequently received atezolizumab plus bevacizumab. Among these 83 patients, 77 achieved an R0 resection, corresponding to an R0 resection rate of 93%. Pathological complete response, defined as the absence of viable tumor cells in the resected specimen, was observed in 23 patients, or 28%.
Most procedures were performed through open surgery: 68 patients underwent open resection and 15 underwent laparoscopic surgery. Median operating time was 211 minutes, and the median postoperative hospital stay was 10 days. Among patients who underwent surgery, R0 resection was achieved in 93%, demonstrating the feasibility of complete resection in this carefully selected population.
Overall Survival Remains Immature
At the data cutoff, 15 patients in the surgery group and 23 patients in the maintenance group had died. Overall survival data were still immature, and no statistically significant overall survival advantage had been established. The HR for death was 0.67 (95% CI 0.35–1.29).
Estimated overall survival rates at 12, 18, and 24 months were 86%, 86%, and 83%, respectively, in the surgery group, compared with 91%, 78%, and 67% in the maintenance group. The investigators noted that longer follow-up will be required to determine whether the improvement in time to treatment failure eventually translates into an overall survival benefit. The final overall survival analysis is planned for a subsequent report.
Safety of the Surgical Strategy
The improvement in time to treatment failure was accompanied by additional treatment-related toxicity and surgical risk. Among the 83 patients included in the surgery-group safety population, grade 3 or 4 treatment-related adverse events occurred in 39%, compared with 21% of the 100 patients receiving maintenance therapy.
The most common grade 3 or 4 treatment-related events were increased alanine aminotransferase, reported in 8% versus 1%; decreased platelet count, in 8% versus 3%; and proteinuria, in 4% versus 7%, respectively. Serious treatment-related adverse events occurred in 23% of patients in the surgery group compared with 7% in the maintenance group.
Two treatment-related deaths occurred in the surgery group. One patient died from abnormal liver function considered related to atezolizumab, and another died from liver failure considered related to atezolizumab, bevacizumab, or surgery. No treatment-related deaths were reported in the maintenance group.
Surgery-related adverse events occurred in 35% of patients who underwent resection, while grade 3 or higher surgical complications according to the Clavien–Dindo classification occurred in 16%.
What TALENTOP Adds
Conversion surgery after systemic therapy has attracted increasing attention in HCC, but prospective randomized evidence comparing resection with continued systemic treatment has been lacking.
TALENTOP provides randomized phase 3 evidence that, among carefully selected patients with HCC characterized by macrovascular invasion and no extrahepatic metastases who achieve disease control with atezolizumab plus bevacizumab and are considered suitable for complete resection, the surgical strategy significantly prolongs time to treatment failure compared with continued systemic therapy.
The findings should nevertheless be interpreted within the specific trial population. Only 201 of the 489 patients who entered induction therapy proceeded to randomization, emphasizing the degree of selection required for this strategy. The trial was also conducted entirely in China, approximately 90% of randomized patients had hepatitis B virus-associated disease, and the study population was younger than that of IMbrave150, with median ages of 58 and 56 years in the two TALENTOP groups.
Importantly, TALENTOP has not yet demonstrated an overall survival benefit. The primary evidence at this stage is an improvement in time to treatment failure, accompanied by greater toxicity and the additional risks associated with major liver surgery.
Conclusion
The TALENTOP trial showed that liver resection followed by atezolizumab plus bevacizumab significantly prolonged time to treatment failure compared with continued atezolizumab plus bevacizumab in selected patients with locally advanced HCC, macrovascular invasion, and no extrahepatic spread who achieved disease control after initial systemic therapy and were considered resectable.
The results provide prospective randomized evidence supporting multidisciplinary reassessment for resectability after partial response or stable disease following induction therapy in this carefully selected setting. Whether this approach ultimately improves overall survival remains to be determined with longer follow-up.
The full article is available in The Lancet.