Fluorescence Imaging

Fluorescence imaging uses the Stokes shift to separate signal from background. A light source (LED, laser or arc lamp) delivers light at the fluorophore's absorption wavelength, and the excitation filter selects the right band. The dichroic mirror, set at 45°, reflects that light toward the specimen.

The fluorophores absorb excitation photons and emit photons at a longer wavelength and lower energy, radiating in all directions. The objective collects the emitted light, and the dichroic mirror transmits it, since it passes the longer wavelengths. The emission filter blocks any leftover excitation light, so only the fluorescence reaches the detector, a camera or PMT.

In the resulting image, labeled structures glow against a dark background, which gives the specific molecular targets maximum contrast.

The firefly comparison comes from Dobrucki, and it reaches further than it first seems. A firefly flickers (photon emission is random), slowly dims (photobleaching), and in a crowded field its light merges with its neighbors' until they can't be told apart (the diffraction limit). Every limitation of fluorescence microscopy has its firefly analog.

Every fluorescence image is built from individual photons. Detect n photons at a pixel and the measurement uncertainty is √n, from Poisson statistics: 100 photons give 10% uncertainty, and 10,000 give 1%. No image processing can create information the photons don't carry, so every photon lost in the optical path, to filter absorption, dichroic reflection or detector inefficiency, degrades the final measurement.

Dobrucki puts the quantitative problem plainly: "In all honesty, one has to admit that a standard widefield fluorescence microscope is not made to be an analytical device capable of straightforward measurements of the quantities of fluorescently labeled molecules." The confounding factors include non-uniform illumination (vignetting), focal plane effects, photobleaching, quenching, autofluorescence, spectral overlap and detector nonlinearity. Intensity is related to concentration, but far from simply, so quantitative fluorescence requires calibration and correction.

Every fluorescence technique balances four competing variables, resolution, speed, signal and phototoxicity, and none can optimize all four at once. Widefield is the fastest and gentlest but lacks optical sectioning. Confocal adds sectioning but is slower and more phototoxic, while light sheet illuminates only the imaged plane, keeping toxicity minimal, but needs special sample preparation. Matching the technique to the biological question deserves as much thought as choosing the fluorophores.

Here is a 6-color multiplex immunofluorescence panel for immuno-oncology:

All six targets are visualized at once on the same tissue section. That is what makes single-cell phenotyping and spatial analysis of tumor-immune interactions possible, and it rests on the specificity and multiplexing capacity of fluorescence.