1. May/June 2009 BioOpticsWorld.com
Advances in lasers, optics, and
imaging for the life sciences
Also:
Deep, hi-res
Hyperspectral
mesoscopic
imaging in vivo
The right stuff
for Raman-based
microscopy
cancer diagnosis
Super-resolution
optical
microscopy
Bioimaging
pioneer
more info per view
Sunney Xie
Stimulus funding
for biophotonics
Deeper cancer
detection with
SORS
Unmasking
kidney-stone
development
Quantum OCT
2. hyperspectral imaging
By James Beach
A richer view of
bio structures
Hyperspectral microscopy combines disparate
methodologies to produce a data‑rich
view of biological structures useful in
research and clinical applications.
T
he value of hyperspectral
microscopy for life scientists
is the ability to acquire the
optical spectrum of all points
in a microscope image, coupled with spe-
cialized spectral analysis. The approach
produces uniquely rich views of biolog-
ical tissue, yielding revelations for both
research and clinical applications. For
instance, hyperspectral imaging (HSI)
can distinguish normal, precancerous,
and cancerous cervical cells on Pap-test
slides based on the combination of their
morphological and spectral characteris-
tics, as a prelude to development of pre-
screening tests for more efficient cervi- FIGURE 1. The HSI microscope system includes a smooth spectrum halogen light source, spectral
cal-cancer diagnosis. imager (Headwall Photonics), color camera (Dage-MTI, Michigan City, IN) and automated stage (Prior
Scientific). Dual monitors allow the image and spectral data to appear on separate screens.
methodology and instrumentation
Hyperspectral microscopy grew out of surfaces of distant targets like the Earth’s liquid-crystal and acousto-optic filters
two unrelated disciplines born in the surface below an aircraft. increased the speed and number of
1970s and 1980s: microspectrophotom- The disciplines began to merge when wavelengths, and added programming
etry and spectral remote sensing. The cell scientists turned to digital cameras flexibility to spectral sequences. As far
former is well known to cell researchers and image-analysis software designed back as the early 1980s, flat-field spectro-
for following photochemical reactions for image arrays. It became possible, graphs using holographic gratings could
and revealing properties of components with technologies like the filter wheel, faithfully reproduce the spectrum of all
within cells’ interior. The latter was a to rapidly change wavelengths while points along the spectrograph slit, cre-
NASA creation for capturing and inter- taking a series of pictures to create ating one-dimensional spectral images.
preting spectral information from the spectral images. Electronically tuned When the scene is scanned across the
spectrograph slit, a two-dimensional
James Beach is president of Willis Optics and associate professor at Louisiana State hyperspectral image (a hypercube with
University Health Sciences Center (New Orleans, LA). He was the lead developer for the two spatial dimensions and a third spec-
CytoViva hyperspectral microscope. Contact him at eadeae@gmail.com; www.cytoviva.com. tral dimension) is produced.
Reprinted with revisions to format, from the May/June 2009 edition of BioOPTICS WORLD
Copyright 2009 by PennWell Corporation
3. hyperspectral imaging
For microscopists, scanning was already in place with motor- ing researchers
ized stages. The remaining problem was to integrate all of the a much brighter
hardware and software components into a system for hyper- view of nanoscale
spectral microscopy, and aim this at a market that would ben- str uctures than
efit from the power of the technology. is available with
other methods.
hsi for bio It provides annu-
In early 2008, CytoViva (Auburn, AL) saw the opportunity for lar structured illu-
commercial hyperspectral microscopy to serve the growing mination at a low
research in nanomedicine. The company worked with Head- angle of incidence
wall Photonics (Fitchburg, MA) to incorporate its VIS-to-near- onto the sample
infrared (IR) spectro- located just above
Intensity
graphic camera to enable the condenser. The
sensitivity over the visi- TiO2 spectral profilesfield
2500 resulting dark
ble and near-IR spectral of illuminat ion
2000
ranges that have been is approximately
1500
reported for nanoparti- 150 times brighter
1000
cle applications. ITT-VIS t han is possi-
500
(Visual Information Solu- ble with conven-
tions of Boulder, CO), tional dark field,
500 600 700 800 900 1000
developer of remote sens- and can effectively
Wavelength (nm)
Intensity ing software, scatter light from
TiO2 spectral profiles
supplied the nanoparticles with
2500 sophist icated enough brightness 1000
2000 spectral-anal- to enable the cap-
1500 ysis features ture of spectral 800
1000 of its ENVI information. Vir- 600
500 software. And tually all the light
Bruxton (Seat- that is collected has 400
500 600 700 800 900 1000 tle, WA) pro- interacted multiple
Wavelength (nm) 200
vided its SIDX times with sample
FIGURE 2. The presence of TiO nanoparticles device inter- components and
2 500 600 700
becomes clear following intradermal injection face product to carries the unique Wavelength
into mouse skin; bright spots indicate clusters of add instrument spect r a l sig na -
FIGURE 3. Hyperspectral microscopy has proven
the highly reflective particles (top). The spectral controls, con- tures of the sam-
helpful for investigating Anthrax spores, (top)
signature of TiO extracted from pixel groups in verting the file- ple constituents.
2, shown at 100X; pixels matching the Anthrax
the HSI image with ENVI software, resembles shark input orienta- The smooth spec- library spectra are pseudo colored in red (center).
fins, and is distinct from that of other particles tion of ENVI tral output from a
Anthrax reference spectra, obtained from
(bottom). The spectral peak wavelength of TiO into an inte- halogen source is
2 different regions of the spore, can help to identify
nanoparticles appears to depend on the number of gral part of the used to avoid prob- strains and possible sources (bottom).
particles in clusters. product. lems with spectral
The group analysis when line structure is present, as it is with commonly
designed the hyperspectral imagery (HSI) microscope sys- used mercury, xenon, and metal halide sources.
tem to work with bright- and dark-field transmission modes, Visible and near-infrared wavelengths between 400 and
and with incident light for reflectance and epifluorescence. 1000 nm are resolved with an imaging spectrograph contain-
It includes the spectral detector, a second color camera, an ing an original holographic grating, and recorded at 12-bit
automated stage, a halogen light source and dual monitors, depth. Hyperspectral images are produced by moving the tar-
which are integrated with a research microscope (Fig. 1). An get under the microscope objective to the position sampled by
optional live chamber mounts to the microscope stage so the spectrograph slit, and recording the spectra of all points
researchers can examine living cells in real time at high res- along the line onto the digital camera. The target is then
olution. By November 2008, two CytoViva HSI microscopes moved a very short distance using the automated stage (from
were delivered to government research facilities at the FDA Prior Scientific, Boston) to bring the adjacent region of the
and USDA. target to the recording position. The process repeats until the
The hyperspectral microscope is designed to take full area of interest surrounding the target has been recorded.
advantage of its high-intensity dark-field illuminator, giv- The resulting HSI data are represented as a three-dimen-
4. hyperspectral imaging
sional structure that holds a stack of spectral imaging can also serve as a tool it is possible to quantify particle abun-
conventional two-dimensional images, for “data mining” to determine optimal dance and cluster size.
each containing a narrow band of wave- wavelengths and recording conditions Similar methods have been demon-
lengths that collectively cover the spec- for specific applications. strated for using captured spectra to
tral range. identify strains and possible sources of
The result is the ability to acquire sample applications Anthrax spores (Fig. 3). Since there is a
spectral information from samples in Applications for the HSI microscopy growing need to increase biofuels pro-
a way that allows one to separate the system include nanotoxicology, drug duction from nonfood sources such as
unique spectral features of a target mol- delivery, and biomass analytics. At the forest products, microbial activity dur-
ecule or other object from background FDA’s National Center for Toxicological ing the fermentation process is being
spectra in a sample, or to unmix the Research, Drs. Neera Gopee and Paul investigated with HSI methods. Cellu-
spectral information from single pixels. Howard realized the approach could losic, hemicellulosic, and lignin com-
Objects such as gold, silver, and TiO provide benefits as a primary detection ponents can be identified spectrally in
2
nanoparticles, carbon nanotubes, flu- tool in research on dermal penetration thin slices of plant materials.
orescent probes, quantum dots, and of topically applied formulations includ- Additional uses now being investi-
many endogenous components of cells ing nanoparticles. They found HSI help- gated include targeted drug delivery
and biomass have their own unique ful for quantifying nanomaterials in tis- with nanoparticles and quantum dots,
spectral features that can be identi- sue samples based on their unique spec- and tumor-cell differentiation. In diag-
fied. The hyperspectral image contains tral signatures (see Fig. 2) The approach nostic imaging, HSI is being combined
these signatures within pixels associ- provided a relatively simple and quick with colonoscopy, ocular funduscopy,
ated with distinct objects in the image. quantitative method for screening sam- and body scanning for detection of can-
With the image viewer, individual pix- ples prior to using more time-consum- cer and eye disease, and examination of
els of objects can be selected and the ing methods such as inductively cou- skin ulcers. We expect HSI microscopy
spectra of those pixels saved to a data- pled plasma mass spectrometry or elec- will begin to play a significant role in
base. The spectra are used with spectral tron microscopy. By collecting spectra new contributions to basic and applied
classification methods in the software to of particles dispersed in liquid media, research, in a wide range of life-science
determine the number and locations of
similar objects in other samples. Hyper-
and then looking for these spectra in
hyperspectral images of treated tissues,
disciplines.«
For More Information: Kelly Marino · 888-737-3130 · kelly.marino@cytoviva.com
300 North Dean Road · Suite 5 – PMB 157 · Auburn, AL 36830
www.cytoviva.com