The Trial That Worked Too Well for Its Own Good

More than half the patients in the control arm (54%, or 31 out of 57) crossed over to receive tovecimig after their cancer progressed. This is common in oncology trials; it's often considered the ethical thing to do when a drug shows early signs of working. But it creates a statistical nightmare.

The crossover patients did remarkably well. Their median overall survival was 12.8 months, compared to just 6.1 months for control patients who never got tovecimig. That's a striking gap (with a hazard ratio of 0.54), and it actually suggests the drug helped these patients live longer.

The irony is painful: because tovecimig appeared to extend life in patients who switched to it, the control group's survival numbers got artificially inflated. The comparison between the two arms blurred. The overall survival endpoint, which needed a clear difference between groups, washed out.

It's like running a taste test between Coke and Pepsi, but halfway through, you let most of the Pepsi drinkers switch to Coke. Then you conclude there's no meaningful difference between the two. Technically accurate; practically misleading.

The Pooled Data Tells a Different Story

Compass pointed to one number that arguably matters more than the failed endpoint. Because of the massive crossover, 85% of all patients in the trial (142 out of 168) eventually received tovecimig. The pooled median overall survival across this group was 8.9 months.

For context, historical chemotherapy benchmarks in second-line biliary tract cancer sit around 6 months. Nearly three extra months of median survival is meaningful in a cancer with very few treatment options.

But "meaningful" and "approvable" are two very different words at the FDA. Without a clean survival win from a randomized comparison, the regulatory path forward gets considerably murkier.

Wall Street's Verdict (and the Dissenters)

The market's reaction was swift and brutal, but not everyone agreed with the selloff's magnitude. HC Wainwright reiterated a Buy rating with a $24 price target on the same day the stock hit $1.79. Guggenheim had previously maintained a Buy at $12. Jefferies reportedly called the selloff an overreaction.

On the other side, William Blair raised questions about whether the drug could win FDA approval given the muddied survival data, and Wall Street Zen had already downgraded to Sell the day before results dropped.

That gap between analyst targets and market price tells you something: this is now a story where conviction runs in opposite directions.

Compass Still Has Cards to Play

Tovecimig isn't the company's only program. Compass has two other bispecific antibodies in clinical development. CTX-8371 targets PD-1 and PD-L1 (two immune checkpoint proteins) simultaneously and is in Phase 1 expansion across non-small cell lung cancer, triple-negative breast cancer, and Hodgkin lymphoma. CTX-10726, a PD-1 and VEGF-A bispecific, cleared an FDA IND (permission to begin human testing) in early 2026 and is now in Phase 1 dose escalation.

The company's proprietary StitchMabs platform, which engineers bispecific antibodies using a common light chain approach, has reportedly been used to drug over 40 immune targets. That's a deep toolkit, even if the lead program just hit a wall.

What This Means for the Bispecific Antibody Boom

Bispecific antibodies are one of the hottest modalities in oncology right now. The global market hit roughly $16.9 billion in 2025. The broader immune checkpoint space has been littered with failures: Roche's tiragolumab stumbled in late-stage trials despite promising early signals.

Compass's experience adds a different wrinkle. This wasn't a drug that failed to work; it was a trial design that couldn't capture the drug's effect cleanly. That's a more nuanced problem, and arguably a more frustrating one.

The Bottom Line

Compass Therapeutics is now a $1.79 stock with a drug that shrank tumors and slowed cancer but couldn't prove a survival benefit in a trial contaminated by its own success, and two earlier-stage programs that could either rescue the story or fade into irrelevance.

For investors, the question isn't whether tovecimig works. There's reasonable evidence it does. The question is whether "reasonable evidence" is enough to get a drug approved, keep the lights on, and convince a market that just priced in catastrophe. That's a much harder question, and nobody has a clean answer yet.