Cross Currents: HARMONi-GI1 and First-Line Treatment in Advanced Biliary Tract Cancer

"It’s still a small sample size, and I was impressed to see the benefit in the PFS and response rate. So I personally, if I were presented with a patient [with] intrahepatic, extrahepatic, [or] gallbladder [cancer], I don’t think I would treat them differently. I think the approach is reasonable."

In a 3-part video series recorded at the 2026 ESMO Congress, James J. Harding, MD, an associate attending physician at Memorial Sloan Kettering Cancer Center and an associate professor of medicine at Weill Cornell Medical College, and Jian Zhou, MD, PhD, president of Zhongshan Hospital, Fudan University, discussed first-line treatment of advanced biliary tract cancer and the implications of the phase 3 HARMONi-GI1 trial (NCT06591520).

The discussion is part of Cross Currents: Where East Meets West in Cancer Care, a video series produced through a collaboration between OncLive and China-based digital health platform DXY. In Cross Currents, oncologists from the United States and China compare clinical practice in their regions and discuss key data from major meetings.

Zhou presented HARMONi-GI1 during a Presidential Symposium at the 2026 ESMO Congress. The randomized, double-blind trial conducted in China compared ivonescimab, a PD-1/VEGF bispecific antibody, with durvalumab (Imfinzi), each given with gemcitabine and cisplatin, in 682 patients with previously untreated advanced biliary tract cancer.

At a median follow-up of 14.3 months, the median overall survival (OS) was 14.9 months with ivonescimab versus 12.1 months with durvalumab (HR, 0.75; 95% CI, 0.61-0.92; 1-sided P = .0031). The median progression-free survival (PFS) was 8.4 versuss 5.8 months (HR, 0.61; 95% CI, 0.51-0.72; 1-sided P < .0001), and the objective response rate (ORR) was 53.8% versuss 32.5% (P < .0001).

Grade 3/4 treatment-emergent adverse effects occurred in 82.0% vs 72.3% of patients; hypertension occurred in 15.0% vs 2.4% and hemorrhage in 9.7% vs 6.5%.

In part 1, Harding said gemcitabine and cisplatin plus durvalumab or pembrolizumab (Keytruda) is the preferred frontline approach in the US for patients with unresectable, locally advanced, or metastatic disease. Molecular profiling, including targeted NGS, RNA sequencing to identify rare fusions, and in some cases immunohistochemistry, is performed routinely to guide later-line therapy, although targeted agents have not yet moved into the first line.

Zhou described a similar approach in China, where domestic PD-1 and PD-L1 antibodies are also used and some centers, including his own, favor gemcitabine plus oxaliplatin (GEMOX) for its lower rate of hepatic toxicity. Harding then outlined the rationale for ivonescimab: chemoimmunotherapy is more active than chemotherapy alone but does not yet help as many patients as clinicians would like, and VEGF blockade may enhance the immune microenvironment, supporting combined PD-1/VEGF targeting.

In part 2, Zhou said he expects the HARMONi-GI1 results to change clinical practice. With more than half of patients responding, he suggested that future work could combine systemic therapy with surgery to convert unresectable disease to resectable disease for curative-intent treatment, including earlier use in the neoadjuvant or adjuvant setting.

He acknowledged that the absolute OS gain was modest and said biomarkers are needed to identify the patients who benefit most. Harding said the safety profile, including rates of hypertension and hemorrhage, appeared tolerable enough to consider the regimen for older patients and those with comorbidities, and both noted that subgroup results by age were generally consistent.

In part 3, the pair examined outcomes by primary tumor site. In HARMONi-GI1, the OS HR was 0.63 (95% CI, 0.48-0.83) for intrahepatic cholangiocarcinoma, 0.94 (95% CI, 0.57-1.55) for extrahepatic cholangiocarcinoma, and 0.95 (95% CI, 0.63-1.42) for gallbladder cancer.

Zhou suggested that biliary obstruction and inflammation may affect survival in extrahepatic disease, and both noted the small size of these subgroups. They also discussed pairing ivonescimab with targeted therapy for molecular drivers such as FGFR2 fusions, IDH1 mutations, KRAS mutations, and HER2 alterations. Zhou said his group plans to study pemigatinib (Pemazyre) with ivonescimab and chemotherapy in the first line, but noted that the low frequency of FGFR alterations makes a phase 3 trial difficult to complete. He added that molecular profiling is routine at his center, both at biopsy and on resected specimens, to expand options at recurrence.

Ivonescimab will be further evaluated in the United States in the phase 2/3 EA2251 trial (NCT07765433), which will compare ivonescimab plus chemotherapy with pembrolizumab or durvalumab plus chemotherapy.