A biotech company decorates a nanoparticle with HER2 antibodies. In the lab, the targeted particle is taken up at many times the rate of an untargeted control. They raise money on the strength of active tumor targeting. In the animal study, the targeted and untargeted particles accumulate in the tumor at nearly the same total amount. The antibody, so decisive in a dish, made almost no difference to how much drug reached the tumor. Two things went wrong — and this chapter is about both of them.

The enhanced permeability and retention effect — EPR — was characterized in 1986 and is the mechanism behind every approved nanoparticle cancer therapy, including Doxil. A 2016 meta-analysis found that the median fraction of injected dose actually reaching the tumor across all published studies was about 0.7 percent. That number is contested, but the underlying message is not: EPR is variable and often weak in human tumors, and we cannot currently predict before treatment which patients have tumors permeable enough to benefit. That tension — foundational mechanism, uncertain reliability — is the chapter's central problem.

Passive targeting means accumulation without a targeting ligand — getting particles to the tumor through physics alone. EPR rests on two features. Tumor blood vessels are leaky, with gaps that normal capillaries do not have; nanoparticles in the right size range — roughly 10 to 200 nanometers — can escape through them. And tumors have impaired lymphatic drainage, so particles that enter do not clear. They accumulate — not because the particle is sticky, but because the exit is broken. PEGylation extends circulation half-life by shielding the particle from clearance, letting it drift out of leaky tumor vessels again and again. That is Doxil's logic.

EPR was characterized primarily in rodent tumor models — fast-growing, relatively uniform, highly permeable. Human tumors are slower-growing, more heterogeneous, and far more variable in their vascular permeability. Some human tumors have meaningful EPR-driven accumulation; many do not; and there is currently no reliable way to know which, before treatment. The honest position is not that EPR is fake, nor that it reliably targets drug. It is something more uncomfortable: EPR is real, it is the mechanism behind every approved passive-targeting nanomedicine, and its magnitude in any given patient cannot currently be predicted.

The moment a nanoparticle enters blood, plasma proteins begin adsorbing onto its surface. Within seconds it acquires a dense coating — albumin, immunoglobulins, fibrinogen, apolipoproteins, complement proteins — called the protein corona. The engineered surface you designed is no longer what the body sees; the corona is. Two consequences matter. The corona can mask targeting ligands — proteins physically covering the antibodies that were engineered to bind tumor receptors. This is precisely what happened in the opening case: the HER2 antibodies were buried before the particles ever encountered a tumor cell. And the corona can redirect particle fate — opsonins in the corona flag the particle for rapid removal by the liver and spleen, eliminating it from circulation before it can reach the tumor.

A particle travels through a sequence of environments — blood, tumor vessel wall, tumor interstitium, tumor cell surface, intracellular space — and targeting strategies act at specific points, not uniformly across all of them. Passive targeting operates at the vessel-wall step: getting particles from blood into tumor tissue, when tumor vessels are permeable enough. Active targeting operates at the cell-surface step: improving uptake into cells, after the particle has arrived. It does not improve accumulation at the tumor, and it can be sabotaged by the corona before it ever operates. Neither strategy overrides the dose-loss chain. Claiming otherwise requires data from the steps the strategy is supposed to affect — not inference from the step it actually affects.

Two findings would force revision in opposite directions. If a large, rigorously controlled human study showed that a targeting ligand increases total tumor accumulation — not just intracellular uptake — by a clinically meaningful margin across patients, the claim that ligands cannot overcome the circulation-and-clearance bottleneck would need revising. And if robust noninvasive imaging across many patients showed that EPR delivers a consistent, substantial fraction of injected dose to most human solid tumors, the variable and often weak characterization would be wrong. The current evidence — the 0.7 percent median, the variability across tumor types, the corona-masking mechanism, and the repeated failure of active targeting to improve accumulation over passive — points toward the claims as stated. But these are active experimental questions, not closed ones.

Here is the chapter's claim. EPR is real — it is the mechanism behind every approved passive-targeting nanomedicine — and it is variable, often weak in human tumors, and cannot currently be predicted patient by patient. Active targeting improves cellular uptake, not tumor accumulation, and can be sabotaged by the protein corona before it ever operates. These are not arguments against nanomedicine. They are arguments for measuring delivery rather than assuming it, for testing in conditions that allow corona formation, and for identifying which patients have tumors permeable enough to benefit before committing to a therapy built around that permeability. Still open: whether patients can be stratified by tumor permeability before treatment, and whether the corona can be engineered as a feature rather than blocked as interference.