Quantitation of In Vivo Fluorescense Signals
The quantitation of fluorescence signals in vivo can be a difficult task. There are many biological reasons why in vivo quantitation is difficult—developing and validating good calibration samples can take some effort; however, compounding this difficulty are many technical imaging issues, the primary of which is the interference caused by skin autofluorescence. Skin is extremely fluorescent, so much so that it can be observed with the naked eye using a black light (or Wood's lamp) in a well-lit room, and because skin autofluorescence originates from a complex mixture of hundreds if not thousands of fluorophores, skin fluoresces at all excitation wavelengths. This skin autofluorescence and the contrast problem it causes for in vivo fluorescence measurements, although less of a problem in the NIR, is the limiting factor for the limit of detection of fluorophores at all wavelengths.
Click on images to enlarge
Figure 1: Images and data acquired from 4 mice, each with a known amount of doxorubicin (DOX): 1000, 500, 250, 125, 62.5, 31.25, 15.63, and 7.18 ng. Monochrome images were acquired at the peak emission of DOX. Upper left image shows a monochrome image of the mouse with the 250 and 125 ng amounts of DOX, and the lower left images shows the same animal imaged on the Maestro. Both the increase in sensitivity and the improvement in quantitative accuracy can be seen in the two plots, with the upper right being monochrome and the lower left being Maestro.
Autofluorescence, regardless of the excitation wavelength, emits at the same wavelengths that fluorophores of interest emit, which causes the contrast problem seen in the upper right image in Fig 1. Although the DOX signal is bright enough to be seen above the autofluorescence, the mouse can be seen clearly. Buy utilizing a patented spectral imaging methodology, the Maestro system is able to eliminate the autofluorescence by unmixing it into a separate channel (autofluorescence image not shown). The remaining DOX signal seen in the lower left image in Fig 1 has much higher contrast and is free from the interference of autofluorescence.
In order to quantitate these in vivo signals, one takes a region of interest (ROI) and finds the sum or integrated intensity of all of the fluorescence within that region. Since the monochrome image contains a significant amount of autofluorescence, the integrated intensities for ROIs in such images are a measure of the fluorophore of interest plus an unknown but large amount of autofluorescence. The Maestro images, on the other hand, contain only fluorescence from the fluorophore of interest, and therefore have a much higher degree of correlation with the amount of fluorophore present. Comparisons of the integrated intensities for both monochrome and Maestro ROIs from a group of 4 mice each with a known amount of DOX in subcutaneous spots can be seen in the right panel of Fig 1, with monochrome (upper right) having an R2 value of 0.657 and Maestro having an R2 value of 0.947. It can be clearly seen from this example that removing the autofluorescence from in vivo signals dramatically increases contrast and improves quantitation.
