A Hybrid Multi-Distance Phase and Broadband Spatially
Resolved Algorithm for Resolving Absolute Concentrations
of Chromophores in the Near-Infrared Light Spectrum:
Results from Studies in Dynamic Phantoms
Ilias Tachtsidis, Terence S. Leung, Bilal Tahir, Clare E. Elwell
Department of Medical Physics & Bioengineering, Malet Place Eng. Bldg., University College London, Gower St., London WC1E 6BT, UK
t.leung@medphys.ucl.ac.uk
Matthias Kohl-Bareis, Markus Gramer
University of Applied Sciences Koblenz, RheinAhrCampus Remagen,, Südallee 2, Remagen 53424, Germany
Chris E. Cooper
Department of Biological Sciences, University of Essex, Wivenhoe Park, Colchester Essex, CO4 3SQ, UK
Abstract: For resolving absolute concentration of tissue chromophores in the human adult brain
with near-infrared spectroscopy it is necessary to calculate the light scattering and absorption, at
multiple wavelengths with some depth resolution. We report a novel methodology that combines
multi-distance frequency and broadband spectrometers to quantify chromophores in turbid media
by using a spatially resolved algorithm. Preliminary results in a dynamic phantom showed good
agreement between measured and theoretical calculated absolute concentrations of known dyes.
©2008 Optical Society of America
OCIS codes: (170.6510) Spectroscopy, tissue diagnostics; (300.6190) Spectrometers
1. Introduction
In cranial near-infrared spectroscopy (NIRS) the quantification of brain tissue chromophores requires high depth
sensitivity and the ability to separate the effects of absorption from those of scattering. Haemoglobin, water and
cytochrome-c-oxidase are known tissue chromophores in the near-infrared (NIR) region, but their absolute
concentrations vary between individuals. It is also likely that other important chromophores, such as fats, may make
a significant contribution to the non-haemoglobin absorption. Most of the current spectroscopic methods involve a
minimum of two wavelengths, usually above and below the isosbestic wavelength of oxygenated (HbO2) and
deoxygenated (HHb) haemoglobin (near 800nm). With multi-spectral data, more chromophores can be resolved
improving spectroscopic resolution. In addition to the spectral bandwidth effect, NIR brain investigations would be
greatly improved if the spectral information could be enhanced by improving depth resolution. This can be resolved
by using a multi-distance (MD) approach and applying the technique of spatially resolved spectroscopy (SRS)[1].
This allows the measurement of the slope of light attenuation versus distance, which provides a high signal-to-noise
ratio, and reduces sensitivity to the optical coupling and changes in superficial tissue layers. However in order to
calculate the absolute concentration of the tissue chromophores one needs to measure the coefficients of absorption
(µa) and of reduced scattering (µs’), which can be determined with time domain or frequency domain spectrometers.
Here we propose a method that (i) uses a multi-distance frequency domain (MDFD) system to measure the
absorption coefficient (µa) and the reduced scattering coefficient (µs’) of tissue at a selection of discrete
wavelengths, (ii) assumes a wavelength dependence of scattering and calculate µs’ over all wavelengths in the
spectral window of interest, (iii) uses a multi-distance broadband spectrometer (MDBBS) to measure the attenuation
slope over a wide spectral range and at a number of source detector spacings, (iv) corrects the attenuation slope for
the previously calculated µs’ and hence determines a depth resolved µa, (v) scales the calculated depth resolved µa
with the MDFD measured µa to obtain an absolute measurement of absorption over a wide spectral range.
This is an improvement on a previous concept described by Bevilacqua et al.[2], where they used single sourcedetector separation distances. This papers reports on a novel alternative method to calculate absolute concentration
of chromophores in turbid media such as tissue by fusing multi-distance frequency resolved and broadband nearinfrared spectroscopy with a hybrid spatially resolved algorithm.
2. Experimental Methods
Figure 1 shows the experimental and instrumentation setup. The instrumentation configuration consisted of two
spectrometers. The first is a four wavelength (690, 750, 790, 850nm) multi-distance frequency domain (110MHz)
spectrometer (MDFD), which is a modification of the commercially available OxiplexTSTM from ISS Inc
(Champaign, IL, USA). This instrument allows for the determination of the average value (DC), amplitude (AC) and
phase (Φ) of the modulated intensity at two different source-detector distances (3 and 3.5cm) at each wavelength.
This multi-distance method affords the quantitative assessment of the absorption and reduced scattering coefficients
by using the AC, Φ pair of data as described by Hueber et al.[3]. The second spectrometer is a multi-distance
broadband instrument (MDBBS). It consists of a white light source that utilises a 50W halogen bulb and a short pass
filter to minimise temperature effects. The spectrograph is based on lenses, with f/#=1:1.2 resulting in a throughput
about one order of magnitude larger from conventional spectrometers, and is comprised of a 50×50 mm grating
with 300 grooves per mm that is blazed at 1000nm (to optimize the reflection in NIR). The light spectrum is
detected by a CCD camera (PIXIS:512f, Princeton Instruments), with chip dimensions of 12.3 × 12.3 mm that
corresponds to 512 × 512 pixels with a pixel size of 24 × 24 µm. In order to use most of CCD chip area the current
design involves four detector fibres with different diameters, where the diameter of the furthest detector at 3.5cm is
3mm, the diameter for the second detector at 3cm is 2mm and 1mm for the two closest detectors at 2.5cm and 2cm.
In addition the geometry of the fibre bundles on the distal end was changed from round to approximate oval in order
to decrease the area blocked by the slit. The MDBBS is an upgrade of an early design of a Lens Imaging
Spectrograph (or LImS) [4].
MDFD Detector
MDFD
PC
MDBBS
Detectors
Channel 1
Source
Detector
MDBBS
Channel 1
MDFD Source
MDFD
White
0.0
Light
Source
1.0 1.5 2.0 2.5 3.0 3.5 cm
White Light
Source
Solution of
Intralipid and
ICG
MD
BBS
PC
Magnetic
Stirrer
Multiple
Optical Fibre
Fig. 1. Experimental set up showing also the configuration of the optodes from both systems.
The phantom contained 2850ml of deionised water mixed with 150ml of an aqueous scattering suspension
(Intralipid-20%); to that mix an absorbing compound, a green dye (Indocyanine green or ICG) was added in four
steps of 0.5ml from 25mg of ICG diluted at 200ml of deionised water. Optical measurements were performed with
sources and detectors submerged slightly below the surface of the sample.
3. Results
The MDBBS using the SRS methodology can measure at several wavelengths the slope dA/dρ (light attenuation
versus distance) from the algorithm showing in Eq. 1 (see Fig. 2(a)). This equation is derived from the diffusion
approximation which has been described elsewhere[1].
(1)
1 ⎛
2⎞
∂A
=
⋅ ⎜⎜ 3 ⋅ µ a ⋅ µ s , + ⎟⎟
ρ⎠
∂ρ ln 10 ⎝
(2)
µa =
⎛
∂A(λ ) 2 ⎞
⋅ ⎜ ln 10 ⋅
− ⎟⎟
, ⎜
∂ρ
ρ⎠
3⋅ µs ⎝
1
2
One can then calculate µa, from Eq. 2 using the MDFD derived µs’ values for the four wavelengths and
assuming µs’ is following a power law (µs’=a-λb where a and b are free parameters in the fit)[5], we can extrapolate
the µs’ and derive estimates at all wavelengths needed. We can then calculate µa by substituting µs’ in Eq. 2 (see
Fig. 2(b)). For obtaining the absolute µa we finally use the MDFD derived µa for the four wavelengths to scale the
calculated MDBBS µa by minimising the mean squared difference between the MDFD µa and the MDBBS µa at the
four wavelengths (see Fig. 2(c)).
(a)
0.014
0.012
0.012
0.01
0.008
0.006
0.01
0.008
0.006
0.004
0.004
0.002
0.002
0.01
0.009
µs’ (mm-1)
µa (mm-1)
0.014
0
690
0.008
0.007
0.006
710
730
750
770
790
810
830
850
Wavelengths (nm)
(c)
µa MDFD
2 ml
1.5 ml
1 ml
0.5 ml
0 ml
870
890
0.003
0.002
0.001
730
750
770
790
810
830
850
0 ml
730
750
870
890
770
790
810
830
850
Wavelengths (nm)
(d)
870
890
910
0.120
0.100
0.080
Measured
0.060
0.040
Calculated
0.020
0.000
0
910
0.5
1
1.5
2
2.5
Diluted ICG in Solution (ml)
Wavelengths (nm)
Diluted
ICG in
Solution
(ml)
710
0.140
0.004
710
0
690
910
0.005
0
690
(b)
0.016
Concentration ICG (µM)
dA/dρ (OD/mm)
0.016
0.5 ml
1 ml
1.5 ml
2 ml
Fig. 2. Measurements on the phantom with the different ICG volumes; (a) the light slope as measured by MDBBS; (b) the calculate MDBBS
µa using the derived µs’ from MDFD; (c) the absolute µa after scaling with MDFD µa; (d) comparison between measured and calculated ICG
concentrations.
From the absolute µa we calculate the absolute concentration of ICG and water using their measured specific
extinction coefficients (fitting from 740 to 915nm) and then compared them with the theoretical values (see Fig
2.(d)). The average error between the measured and the theoretical calculated concentration of ICG was 2.3% with a
standard deviation of 2.3%.
4. Conclusion
The central novelty and innovation of this methodology is that by using the MDFD derived µa and µs’ values we can
convert the MDBBS slope measurements of attenuation to measurements of absolute µa across the NIR wavelength
range. This method is especially promising for NIRS of tissue as it can obtain broad wavelength coverage with
increased penetration depth. For application to the adult head this method permits rapid data acquisition, some depth
resolution and robust resolution of the contributions from the major NIR tissue absorbers. This technology will soon
be used for studies in human adult volunteers and patients in neurocritical care unit.
5. References
[1] S. Suzuki, S. Takasaki, T. Ozaki, and Y. Kobayashi, "A tissue oxygenation monitor using NIR spatially resolved spectroscopy.," Proc. SPIE
3597, 582-592 (1999).
[2] F. Bevilacqua, A. J. Berger, A. E. Cerussi, D. Jakubowski, and B. J. Tromberg, "Broadband absorption spectroscopy in turbid media by
combined frequency-domain and steady-state methods," Applied Optics 39, 6498-6507 (2000).
[3] D. M. Hueber, M. A. Franceschini, H. Y. Ma, Q. Zhang, J. R. Ballesteros, S. Fantini, D. Wallace, V. Ntziachristos, and B. Chance, "Noninvasive and quantitative near-infrared haemoglobin spectrometry in the piglet brain during hypoxic stress, using a frequency-domain
multidistance instrument," Phys. Med. Biol. 46, 41-62 (2001).
[4] J. Soschinski, D. Geraskin, B. Milosavljevic, U. Mehlhorn, U. Fischer, G. Bennink, and M. Kohl-Bareis, "Cerebral oxygenation monitoring
during cardiac bypass surgery in babies with broad band spatially resolved spectroscopy.," Advances in Medical Engineering 114, (2007).
[5] J. M. Schmitt and G. Kumar, "Optical scattering properties of soft tissue: a discrete particle model," Applied Optics 37, 2788-2797 (1998).