Every EV product specification rests on a number of particles per millilitre. That number has to come from somewhere. MISEV2023 takes a clear position on where: a primary single-particle measurement, ideally orthogonal, with the method named and the limitations acknowledged. Nanoparticle Tracking Analysis ended up as the field default because, of the available options, it satisfies that brief best. It is also not infallible, and the guideline is explicit about why.
Why NTA
NTA tracks the Brownian motion of individual nanoparticles in suspension under a focused laser. The instrument records a video of each particle's diffusion, fits the trajectories to the Stokes–Einstein equation, and outputs a per-particle hydrodynamic diameter and a concentration estimate. The measurement is single-particle, label-free, and operates in solution at physiologically relevant concentrations.
Three properties earned it the default status:
- Single-particle resolution. Bulk methods, like dynamic light scattering, fit a population to an aggregate signal and lose the distribution. NTA keeps it.
- Wide dynamic range across EV-relevant sizes. NTA can resolve particles roughly from 50 nm to 1,000 nm reliably on well-calibrated instruments, which covers the bulk of the EV size space, including both exosomes and most microvesicles.
- Workflow compatibility. Sample prep is minimal, runs are fast, and the method tolerates buffers that occur in EV preparations.
For a CoA-grade particle concentration on a per-batch basis, NTA is currently the path of least resistance among methods that meet MISEV2023's reporting bar.
Where NTA falls short
Anyone who has run NTA on real EV preparations has seen its limits. MISEV2023 names them.
NTA does not distinguish vesicles from non-vesicular particles of similar size. A lipoprotein, a protein aggregate, or a polymer micelle in the size range looks identical to a small EV in the trajectory data. This is why MISEV2023 pairs particle counting with marker enrichment and contamination screening — the count alone says "there are particles," not "those particles are EVs."
NTA is sensitive to operator parameters and instrument-specific differences. Camera gain, detection threshold, and analysis settings all influence the resulting number. Cross-instrument comparison without standardised parameters can produce reportable differences from the same preparation. The literature has direct head-to-head comparisons that quantify this.
NTA underestimates the smallest EVs. Particles below approximately 50 nm produce weaker scattering signal and may fall below the instrument's detection threshold, depending on refractive index. For samples enriched for very small exosomes, NTA can systematically undercount the lower tail.
NTA does not measure refractive index directly. Concentration estimates assume reasonable optical properties for the particle population. EVs and non-EV nanoparticles of similar size can differ in refractive index in ways that bias the absolute count.
What MISEV2023 expects alongside NTA
MISEV2023 frames characterization as an orthogonal-methods exercise. No single measurement is sufficient. The guideline names the alternatives most useful in combination with NTA.
Tunable Resistive Pulse Sensing (TRPS). Counts and sizes particles individually as they pass through a nanopore. It is orthogonal to NTA because the physics is different — a current-blockade measurement rather than an optical-scattering one — and is less sensitive to refractive index. TRPS struggles at higher throughput and with the smallest particles, but it makes a strong cross-check against an NTA result.
High-resolution flow cytometry with a fluorescent label. When the EVs are labelled for a specific marker, flow cytometry can count vesicles that carry the marker, not just particles of the right size. This is what makes flow cytometry a useful follow-up after NTA: it converts a particle count into a specifically-marker-positive particle count.
Dynamic light scattering (DLS). Bulk-measurement method that delivers a hydrodynamic size distribution and is fast to run as a sanity check. DLS does not give per-particle data and is biased toward larger particles in heterogeneous samples. Useful for quick QC, not for primary concentration claims.
Transmission electron microscopy (TEM). Visualises individual EVs, confirms morphology, and is the gold standard for "do these look like vesicles." Not a counting method at scale, but indispensable for batch-release confirmation that the particles in the preparation have the expected bilayer-membrane appearance.
What this looks like in a real CoA
A defensible Certificate of Analysis reports an NTA-measured particle concentration with its standard deviation across replicate measurements, the NTA instrument and parameter set used, and at least one orthogonal confirmation. The orthogonal confirmation does not have to be exhaustive on every batch; it has to be performed often enough that the manufacturer can show the relationship between methods has been characterized.
mPDEV's batch-specific CoA reports NTA-measured particle concentration, marker enrichment on a defined panel, an absence-of-contaminants screen, and TEM confirmation of morphology. The number on the label is the number that came out of that workflow, with the workflow declared. The orthogonal methods are not decorative; they exist because NTA on its own is not enough to satisfy MISEV2023.
What clinicians and partners can ask
Three questions cut through marketing copy on this point:
- Which instrument and parameter set were used for the NTA measurement on this batch, and how reproducible is the number across replicate runs?
- Has the relationship between the NTA-reported count and at least one orthogonal method (TRPS, flow cytometry, TEM) been characterized for this manufacturing workflow?
- Is the reported count corrected for known biases, particularly the lower-tail under-counting that affects very-small-EV-enriched samples?
A manufacturer that has thought about NTA the way MISEV2023 asks can answer all three. A manufacturer that has not, cannot.