Morpho-spectral imaging in the biosciences

Abstract

Three-dimensional (3D) imaging is a current challenge in biology and medicine for investigation of the smallest details of living systems, from cells to organisms, and with the expectation of micro to nanoscale resolution in real time. This is not a new objective, but rather a long quest toward excellence; to achieve such analytical performance, however, imaging technology must integrate the latest innovations in light sources, optics, detectors, supercomputers, and automation, among others. Spectroscopic imaging has already been proved to achieve 3D rendering, either directly with a confocal microscope (e.g., Raman microscopy) or with post-acquisition reconstruction (e.g., infrared (IR), mass spectrometry (MS), and X-ray fluorescence (XRF) microscopy), which work in transmission mode or on the sample surface. The main advantage of the development of 3D spectroscopic imaging is to provide multiple chemical information extracted from raw spectra. The real challenge seems to be the ability to normalize the signalto-noise (SNR) ratio between 3D images stacked for 3D rendering. The duration of acquisition, the stability of light sources, and signal quantification are the three main components of SNR instability or fluctuations. It seems that the development of new sources is crucial to the development of 3D imaging technology for the achievement of micro and nanoscale resolution. At the very least, this source development is mandatory for introduction of 3D spectral imaging techniques into the routine environment of biological laboratories or hospitals for diagnostic purposes. Saving time in data acquisition and treatment to obtain a 3D image of a sample is the current major limitation for merging these 3D imaging techniques with commercially available industrial products. This is frustrating the expectations of the spectroscopist community wishing to exploit these new methods of spectral analysis. Nevertheless, Raman microscopy has recently benefited from low-power lasers and ultra-fast detectors, thus limiting the destructive effect of lasers on organic samples and leading to in vivo analysis of small biological specimens. IR imaging is now being proposed for this application with tunable lasers for stable SNRs with ultra-fast acquisition (ms range). The first attempts at 3D rendering with these techniques have led to a glimpse of the tremendous analytical potential these approaches will bring to biosciences, notably by achieving microscopic rendering of large samples. Each technique benefits from its own physical properties, leading to different developments for specific uses in the biosciences. Raman microscopy is currently more successful for cell imaging, because of its confocal geometry and nanoscopic resolution. IR imaging has greater potential for analyzing larger samples (typically tissues) because of focal plane array detectors and rapid acquisitionwith high SNRs. XRF is sensitive enough to achieve trace analysis of multiple inorganic elements in cells and tissues. Laser ablation-based inductively coupled plasma mass spectrometry (ICP–MS) has also been used to achieve nanoscopic resolution for detection of trace concentrations of metals; it is, thus, also compatible with cell imaging dimensions. All of these spectroscopic imaging techniques will soon be available for fully quantitative 3D analysis, leading to tomographic reconstruction when they are coupled to morphological imaging, notably throughX-raymicroscopy, which is able to reveal the microscopic structure of hard materials (e.g., Published in the topical collection Morpho-Spectral Imaging with guest editors Cyril Petibois and Yeukuang Hwu.