Measuring regional blood flow with stable isotope-labeled microspheres

The measurement principle

Spheres are introduced into the arterial circulation upstream of the tissue of interest, typically into the left heart so that they distribute systemically, and mix with the blood leaving it. Being carried passively, they partition among vascular beds in the same proportions as the blood itself.

What separates them from a diffusible tracer is what happens next. Each sphere travels until it reaches a vessel narrower than itself, and stops. The trapping is mechanical, not chemical: no binding, no uptake, no metabolism. A sphere sized above the calibre of the terminal arterioles has nowhere to go once it arrives.

Two things follow. Extraction is effectively complete on first pass, so what is deposited records the flow a region received during the brief window while spheres were in circulation, rather than an average over the minutes that follow. And because deposition tracks flow, the number of spheres recovered from a piece of tissue is proportional to the share of cardiac output that piece received. Dissect the organ, count the spheres in each piece, and the distribution of perfusion across it falls out.

The technique is therefore a snapshot with unusual spatial resolution and no temporal resolution at all. It resolves space as finely as you are willing to dissect. It resolves time not at all. Each injection yields one instant, and the tissue must be collected to be read, so the measurement is terminal in the region sampled.

Why the label must stay put

The whole inference rests on the count in a segment being a count of trapped spheres. A label that leached into surrounding tissue or back into blood would break that correspondence, and would do so in a direction that is difficult to detect after the fact. In stable isotope-labeled spheres the element is held within the polystyrene matrix rather than adsorbed onto the surface, which is what makes the count durable through hours of circulation and subsequent handling.

From trapped spheres to perfusion values

Counting spheres across dissected segments gives the relative distribution of flow. Expressing the result in absolute terms (volume per unit time per unit mass) requires a calibration, and that is the job of the reference sample.

While spheres are circulating, blood is withdrawn from a systemic artery at a constant, known rate, beginning before the injection and continuing until the spheres have cleared. That withdrawal acts as an artificial organ perfused at a flow you have set yourself. The spheres it captures, set against the pump rate, fix the relationship between sphere count and flow for that particular injection. Apply the same relationship to a tissue sample's count, divide by the sample's mass, and the result is perfusion per unit mass of tissue.

This is why every segment is weighed, and why a failed or interrupted reference withdrawal costs the absolute scale of the entire injection even when the tissue counts are perfectly good.

Counting statistics set the dose

The number of spheres in a segment is a counting problem, and the relative error grows as counts fall. Enough spheres must be injected that even the least perfused segment retains a number large enough to measure with acceptable precision. The dose therefore depends on the mass of the animal, the share of cardiac output the target organ receives, and how finely the tissue will be divided, since more segments means fewer spheres in each.

Low-flow regions are the binding constraint, and the reason to think about dose before the experiment rather than after. A dose that is comfortable for normally perfused tissue can leave an ischemic segment with too few spheres to support a reliable number, which is precisely the segment the experiment usually exists to characterise.

The general protocol puts a floor on it: a minimum of 400 to 600 microspheres per tissue segment for a 95% confidence interval on a blood flow measurement. To reach that, it estimates the minimum total injected as

Y = 1.2 × 10⁶ + 1.9 × 10⁵ X

where Y is the minimum number of microspheres per injection and X is the mass of the subject in kilograms. The protocol gives these as working figures:

The equation is fitted to an average myocardial study, so other organ systems may need the dose adjusting. Anticipated low-flow zones raise it further: the protocol's own worked example takes a 20 kg myocardial study from 5 million to 10 million once a 75% flow reduction is expected in some segments.

Preparing tissue for assay

At the end of the experiment the organ is divided along whatever scheme the study calls for, and each piece is weighed and sealed into its own vial. The reference blood sample is handled alongside it. From this point the sample is not opened again.

Two forms of background matter, and both are dealt with before the sample is sealed. The first is physiological. Sodium and chloride are abundant in tissue and in residual surface blood, and both activate readily under neutron irradiation, contributing signal that the label must then be measured against. Rinsing the tissue, or retro-perfusing the organ, with an isotonic buffer that contains neither ion in place of saline or phosphate-buffered saline lowers that background rather than adding to it. The same substitution applies to any histochemical staining done in the same preparation: a stain delivered in ordinary saline loads the sample with exactly the ions you were trying to remove.

The second is contamination from the labware. A container that carries trace quantities of sodium, or of any element being used as a label, activates along with the sample and is indistinguishable from it. Vials must be cleaned to a standard set by the elements being measured, not by ordinary laboratory cleanliness.

Samples are then dried before shipment, which both stabilises them in transit and removes water that would otherwise be irradiated to no purpose.

Quantifying the labels by neutron activation

The sealed samples are placed in a neutron flux. Nuclei in the sample capture neutrons, and a fraction of the label is converted into radioactive isotopes of the same element. Those isotopes decay, emitting gamma rays, and a high-resolution detector records the resulting spectrum. Nothing radioactive existed before this step, and what is created decays away after it.

Identification is by energy. Each activated nuclide emits at energies characteristic of that element, and decays on its own characteristic timescale. Because those emission lines are distinct from one another, several different labels present in the same piece of tissue can be separated in a single measurement. That is what makes serial injections practical: inject one element at baseline, a second during intervention, a third in recovery, and one irradiation of one sample yields a flow value for each timepoint, in the same tissue, with no dissection error between them.

The number of timepoints a study can resolve is set by how many labels are available that remain distinguishable when measured together. The reported quantity is a decay rate for each label in each sample, directly analogous to the counts a gamma well counter returns from radiolabeled spheres, and it enters the reference-sample arithmetic the same way.

Two properties make this workable on intact tissue. Neutrons penetrate, so the sample is assayed as it sits in its sealed vial: no digestion, no extraction, no filtration, and therefore no opportunity to lose material between collection and measurement. And the measurement is elemental, so it does not depend on the optical or chemical character of the surrounding tissue: muscle and fat read alike, and a stained sample reads the same as an unstained one.

The sample also survives. Because it is neither digested nor consumed, and because the induced activity decays, tissue can be archived and returned to later, to repeat a measurement that looks anomalous or to ask a question of the material that was not anticipated when the study was designed.

Why stable isotopes

Radiolabeled microspheres have been the reference standard for regional perfusion measurement since the late 1960s, and the case for a stable isotope approach is not that it measures something different. It is that it reproduces the same measurement while moving the radioactivity out of the laboratory. The two have been compared directly, in a rabbit model of myocardial ischemia and reperfusion, by co-injecting stable- and radiolabeled spheres and measuring both in the same tissue.

Reinhardt CP, Dalhberg S, Tries MA, Marcel R, Leppo JA.Stable labeled microspheres to measure perfusion: validation of a neutron activation assay technique. Am J Physiol Heart Circ Physiol, 2001.

Against radiolabeled spheres

With radiolabeled spheres the tracer is radioactive from the moment it arrives. Its useful life is bounded by decay, which puts a clock on the experiment before the animal work begins, and everything the tracer touches inherits its status. Gloves, glassware, bedding, the carcass and the ordinary consumables of a surgical preparation all become low-level radioactive waste requiring licensed disposal, and personnel are working near an active source for the duration.

With stable labels none of that applies, because nothing in the laboratory is radioactive at any point. Activity is induced only in the submitted samples, inside the irradiation facility, after the biology is complete. The practical consequence is that the animal work needs no radioactive materials licence, and it is the reason the technique can be run in facilities where the radioactive method is not an option.

Against fluorescent spheres

Fluorescent spheres remove the radioactivity too, but they pay for it at the readout. Light does not travel through tissue, so the spheres have to be brought to the detector: the sample is digested, the spheres recovered by filtration or centrifugation, and the dye extracted into solvent before it can be read. The tissue does not survive the process.

Every transfer in that sequence is an opportunity to lose spheres, and the losses are real enough that a known quantity of process-control spheres is normally added to each sample so that recovery can be estimated and corrected for. That is a correction, not an elimination; it bounds the error rather than removing it. The digestion itself calls for aggressive reagents that carry their own handling and disposal burden, which erodes part of the reason for leaving radioactivity behind in the first place. And because the readout is optical, it is vulnerable to anything else in the sample that absorbs or emits light, which puts histochemical stains and the microsphere assay in competition for the same tissue.

Neutron activation inverts each of those constraints. The sample is measured intact and unopened, so there is nothing to recover and nothing to correct for; the tissue remains available afterwards; and the elemental readout is indifferent to stains, dyes and tissue composition. What it requires in exchange is access to a neutron source, which means samples are measured in batches rather than at the bench.

Validation and published work

The stable isotope method was compared directly against radiolabeled microspheres by co-injecting both and measuring each in the same tissue. That study, and a bibliography of published work using the technique, are below.

• Stable labeled microspheres to measure perfusion: validation of a neutron activation assay technique

  The published validation study comparing stable isotope-labeled microspheres assayed by neutron activation against co-injected radiolabeled spheres, in a rabbit model of myocardial ischemia and reperfusion.

  Download PDF, Stable labeled microspheres to measure perfusion: validation of a neutron activation assay technique182 kB

• Selected published papers

  Bibliography of published work using the technique, organised by species for blood flow and microsphere studies, with validation studies listed separately.

  Download PDF, Selected published papers204 kB