Summary
A South Korean phase 1b/2 trial found a 55% objective response rate when trastuzumab and nivolumab were added to gemcitabine and cisplatin for previously untreated HER2-positive advanced biliary tract cancer. The single-arm results provide early evidence for first-line HER2-targeted chemoimmunotherapy, while randomized confirmation is still needed.
A prospective phase 1b/2 trial in South Korea found that a four-drug regimen combining trastuzumab, nivolumab, gemcitabine and cisplatin produced tumour responses in 55% of people with previously untreated, HER2-positive advanced biliary tract cancer. The HERBOT study, published in Nature Medicine on 22 September 2026, included 40 participants and reported a median progression-free survival of 10.6 months after a median follow-up of 17.0 months.
The study was open-label, single-arm and nonrandomized, making it an early efficacy and safety assessment rather than a comparison with chemoimmunotherapy alone. Its results support further testing of upfront HER2-targeted treatment in this cancer group.
What the HERBOT trial tested
Biliary tract cancers include intrahepatic and extrahepatic cholangiocarcinoma and gallbladder cancer. The current first-line treatment approach for advanced disease generally combines an immune checkpoint inhibitor with gemcitabine and cisplatin. HER2, also known as ERBB2, is overexpressed or amplified in approximately 15% of biliary tract cancers and can provide a target for treatment.
The trial added trastuzumab, an antibody directed at HER2, and nivolumab, an immune checkpoint inhibitor that blocks PD-1, to the chemotherapy backbone. The study rationale was that HER2-directed treatment may promote immune activation and antigen presentation, while chemotherapy can release tumour antigens as cancer cells die. Nivolumab is intended to help restore antitumour T-cell activity.
The 40 participants were enrolled at 12 academic cancer centres in South Korea between 12 May 2023 and 3 December 2024. All had treatment-naive advanced or metastatic disease. HER2 positivity was centrally confirmed through either immunohistochemistry (IHC) 3+ or IHC 2+ combined with positive in-situ hybridisation. Among the participants, 62.5% had IHC 3+ tumours and 72.5% had gallbladder cancer. The median age was 66 years.
Participants received nivolumab 360 mg and trastuzumab 6 mg per kilogram after an 8 mg per kilogram loading dose on day 1 of each three-week cycle. Gemcitabine 1,000 mg per square metre and cisplatin 25 mg per square metre were given on days 1 and 8. Chemotherapy was planned for up to nine cycles, while trastuzumab and nivolumab could continue for up to 24 months.
Results, safety and next steps
The primary phase 2 endpoint was met. Twenty-two of the 40 participants had an objective response under RECIST 1.1: one complete response and 21 partial responses. This produced an objective response rate of 55% (95% confidence interval 38.5–70.7%). The disease control rate was 95%, including participants whose disease remained stable.
Among those who responded, the median duration of response was 12.6 months. Median progression-free survival was 10.6 months (95% confidence interval 7.8–17.4), with 80.0% of participants free from progression at six months. Overall survival data were not mature enough to calculate a median; the 12-month overall survival rate was 72.7%. Two participants underwent surgery with curative intent after their tumours shrank, at 4.54 and 6.97 months after treatment began.
Treatment-related adverse events occurred in 36 participants, or 90%. The most frequent grade 3 or higher treatment-related events were neutropenia in 57.5%, anaemia in 30.0% and thrombocytopenia in 22.5%. Most non-blood-related adverse events were grade 1 or 2. One participant developed a grade 2 reduction in ejection fraction after the sixth cycle, a relevant cardiac effect for trastuzumab treatment. Two people stopped treatment because of adverse events: pneumonia unrelated to treatment and nivolumab-related arthralgia.
The researchers also conducted a preplanned exploratory analysis using artificial-intelligence software to measure the proportion of HER2 3+ cells across whole tumour slides. Participants whose slides contained at least 10% HER2 3+ tumour cells had an objective response rate of 80%, compared with 40% among those below that threshold. Their median progression-free survival was 17.4 months versus 10.5 months. These analyses were exploratory, used a threshold adapted from gastric-cancer HER2 guidance and were not externally validated, so they are best viewed as a possible way to refine patient selection in future studies.
The findings are encouraging because HER2-targeted treatment has generally been studied after progression in biliary tract cancer, while aggressive disease can prevent some patients from reaching later-line therapy. However, the study had no control group, included only 40 people and was enriched for gallbladder cancer. The HER2 definition was also largely adapted from gastric cancer because a biliary-tract-specific consensus is not yet established. Ongoing randomized phase 3 studies, including HERIZON-BTC-302 and DESTINY-BTC01, will be important for determining whether this strategy improves outcomes over current chemoimmunotherapy.