Phase 3 interim shows 3x CR advantage in R/R AML, but cash runs out before Part A unblinding — survival to data is the entire bet

Bull

Unblinded Part A data confirms CR rates of 36–43% across Annamycin dose arms versus ~12% for HiDAC control, with statistical significance on the primary endpoint now achievable given the full 90-subject dataset and O'Brien-Fleming budget preserved for this analysis. A clean safety profile — consistent with the pooled cardiotoxicity data showing no LVEF change — combined with strong remission depth in venetoclax-failed patients would validate the accelerated approval pathway, catalyze a partnership or licensing deal that bridges the cash gap, and support seamless advancement into Part B enrollment.

Bear

The most likely failure mode is statistical: blinded CRc of 37% across the mixed population could decompose on unblinding into heterogeneous arm-level outcomes where neither dose achieves significance against the 29% control CRc, especially if randomization imbalances or site effects inflate the control arm. A secondary risk is that cash exhaustion forces dilutive financing or a halt to operations before the data lock is complete, preventing the readout entirely — the thesis explicitly frames survival to this event as the core risk.