Compass Therapeutics (Nasdaq: CMPX) reported that tovecimig, its DLL4/VEGF-A bispecific antibody, met the key secondary endpoint of progression-free survival in the randomized COMPANION-002 trial, cutting the risk of disease progression by 56% compared with paclitaxel alone in previously treated biliary tract cancer — a disease where no FDA-approved second-line therapy exists for the majority of patients.
COMPANION-002 is a Phase II/III randomized, controlled study (NCT05506943) evaluating tovecimig plus paclitaxel versus paclitaxel alone in 168 adults with unresectable advanced, metastatic, or recurrent biliary tract cancer who had received one prior systemic chemotherapy regimen, randomized 2:1.
Median progression-free survival was 4.7 months in the tovecimig arm versus 2.6 months with paclitaxel alone (HR=0.44, p<0.0001). The overall response rate primary endpoint, previously disclosed in April 2025, was 17.1% versus 5.3% (p=0.031), with all responses confirmed by blinded independent central review. The overall survival analysis did not meet statistical significance, substantially confounded by a 54% crossover rate from the control arm; in a post-hoc subset analysis, crossover patients who subsequently received tovecimig had a median OS of 12.8 months compared with 6.1 months in those who did not cross over (HR=0.54, p=0.04).
Biliary tract cancers — encompassing intrahepatic and extrahepatic cholangiocarcinoma, gallbladder cancer, and ampullary carcinoma — affect an estimated 26,500 patients annually in the United States. For the roughly 80–85% of patients whose tumors lack an actionable mutation, second-line treatment options are limited to off-label chemotherapy regimens, principally FOLFOX, which produces response rates of approximately 5% or less and a median overall survival of around six months. The ABC-06 Phase III trial, published in The Lancet Oncology in 2021, established FOLFOX as the closest thing to a standard of care in this setting, though it carries no discrete FDA approval for BTC.
Tovecimig is engineered to simultaneously block two distinct pro-tumorigenic signaling axes: VEGF-A, which drives classical tumor angiogenesis, and DLL4, a Notch pathway ligand implicated in vascular remodeling and cancer stem cell maintenance. The rationale for dual blockade is that each pathway can partially compensate for inhibition of the other when targeted individually — a limitation observed with single-agent anti-VEGF therapies in several solid tumor types. Whether that mechanistic logic translates into durable clinical benefit in biliary tract cancer remains to be fully established, but the PFS data from COMPANION-002 are consistent with the hypothesis.
Cross-trial comparisons are limited by differences in patient populations, prior treatment histories, and study designs, but the 4.7-month median PFS compares with the approximately 4-month PFS reported with FOLFOX in the ABC-06 dataset, and the response rate of 17.1% is substantially higher than the roughly 5% seen with chemotherapy in unselected second-line BTC. The OS picture is harder to interpret. The 54% crossover rate — which Compass permitted on ethical grounds, allowing control-arm patients who progressed to receive tovecimig — materially diluted any OS signal in the intent-to-treat analysis. The ITT OS was 8.9 months in the tovecimig arm versus 9.4 months in the control arm (HR=1.05, p=0.78), a result the company acknowledges is uninterpretable as a measure of tovecimig's survival effect. The rank-preserving structural failure time analysis, intended to adjust for crossover, also failed to meet its underlying assumptions and produced similarly uninterpretable results.