A bottleneck-first evaluation of antigen delivery platforms for therapeutic cancer vaccines.

Therapeutic cancer vaccines require tumor-relevant antigens and delivery systems that preserve antigen integrity, control tissue and cellular exposure, support productive processing, and sustain antigen-specific immunity. Yet platform performance cannot be inferred from material properties, cellular uptake, or tumor control alone, because these endpoints do not identify which delivery step changed or whether that change contributed to an antigen-specific response. Seven material matrices and a separate carrier-free molecular assembly mode are evaluated across four functional domains. A carrier advantage is supported only when a defined variable changes a prespecified delivery function under matched conditions, a matched delivery endpoint captures that change, and the result is linked to an antigen-specific outcome; selected variable-to-outcome chains are appraised qualitatively using four possible categories: supported, partial, untested, or refuted. Repeat-dose behavior, carrier-directed immunity, routine lot release, manufacturing consistency, and comparability following manufacturing process changes are treated as additional constraints because they can modify or confound an initial advantage. Artificial intelligence-assisted design is considered only for hypothesis triage and validation planning. The reviewed evidence does not support a universal platform hierarchy. Platform selection should instead begin with the dominant biological or product-level bottleneck.