Overview of different types of quantification methods and how these can be achieved using nanoscopy and confocal microscopy, including information on throughput and disadvantages of the methods.

Technique	Quantification	Quantification process	Throughput	Disadvantages
aConfocal	Co-localization	NP endosomal vesicles are tagged with different fluorophores and average fluorescence intensity is calculated (e.g. using ImageJ – Color2/JACoP). Co-localization calculated using Mander's/Pearson's correlation coefficients56,148	Fairly good throughput – (i.e. a few cells, tens of lysosomes and tens of NP clusters)	-Limited resolution
				-Cannot resolve individual NPs
				-Localization precision is affected by resolution
	Particle tracking	The total number of particles and endosome co-localized NPs is tracked and counted with particle tracking software. As NPs cannot be individually detected, particle events are calculated instead, whereby a single NP event likely corresponds to one vesicle containing NPs75,154		-Invisible particles (i.e. due to bleaching of fluorophores or de-coupling of fluorescent dyes)
				-Number of particles is underestimated
				-Choice of fluorophore can influence results (e.g. pH sensitive dyes can have reduced signals in acidic vesicles)
				-Fluorescence must be quantified in relation to a control to account for fluorophore instability
EM	Direct visualization and quantification	The ratio between NPs found in the cytosol and endosomes is calculated148,155,156	Low throughput – (i.e. one cell, tens of NPs and a few endosomes per field of view)	-Complicated sample preparation
		Can distinguish between intracellular/extracellular/intramembranous nanoparticles		-Generally, samples are fixed and sectioned (i.e. no living cells)
	Serial sectioning or electron tomography	Imaging in 3D of sequential sample sections. Location, size, and the number of vesicles as well as NPs can be calculated in whole 3D cells74,156		-Difficulty in distinguishing different intracellular vesicles
	Stereological image analysis	Using the relative particle distribution within cells (RDI). Tests if NPs are localizing randomly or specifically within cellular compartments. The particle density of each compartment is calculated by relating the number of particle events in the specific compartment to the fractional volume of the compartment157		-Particles must be smaller than the section thickness (∼150 nm)
		Correlating the total number of intracellular particles of a sample with the total cell number of that sample. Using the fractionator principle54,158		-Quantification from 3D reconstructions is difficult
		The density of intracellular particles is multiplied by the average cell volume to calculate the average number of NPs per cell75		-Restricted to samples with adequate atomic contrast
SMLM	Spatial analysis and clustering	Single-molecule localization microscopy techniques produce point cloud data as a result of multiple localizations in time. These data can be analyzed to identify objects, and determine densities or spatial correlations93	Intermediate – low throughput. The field of view may vary from one to few cells. The imaging time would greatly depend on the specific technique used (seconds to minutes)	-In some cases, there are undesired non-specific interactions or background noise
	Molecule counting	Single-molecule localization microscopy techniques are based on the identification of individual molecules. Therefore, it is possible to quantify the exact number of molecules on a specific area.159,160 For example, the ligands or proteins on the surface of a nanoparticle161,162		-High amount of data that can make the quantification process slow
	Stability of NPs and vesicles	The increased resolution and precise molecule counting of super-resolution microscopy allow the determination of the stability of small objects such as nanoparticles and vesicles. It is possible to establish their shape96 and observe the degradation in time99
STED/SIM	Size and shape of NPs	The improved resolution of these techniques allows the measurements of the size and shape of smaller objects compared to confocal microscopy109,113,114	Good throughput – imaging times in the millisecond-second range	-No single-molecule quantification
	Co-localization	Standard colocalization coefficient calculations are also applied for these techniques, although better resolution yields more precise results109
CLEM	Combination of FM and EM techniques	Generally, fluorescence microscopy is carried out prior to EM. Images can be manually aligned using plugins such as eC-CLEM. Quantification can be achieved either via EM or FM, or both.74,163 Detection of ‘invisible particles’ in light microscopy is possible with CLEM, as well as compartment-specific quantification	Low throughput – still limited by EM	-Complex and time-consuming sample preparation
				-NPs must be detectable using both light and electron microscopes
				-Alignment mismatch can affect correlation

^aQuantification is not absolute (not at a single particle level) due to the resolution of the microscope.