10.1109/ULTSYM.2014.0412
Assessing storage-induced red blood cell lesions
using photoacoustic measurements of oxygen
saturation and the frequency content of photoacoustic
signals
Eno Hysi, Eric M. Strohm, Elizabeth S. L. Berndl
Michael C. Kolios
Jason P. Acker
Center for Innovation
Canadian Blood Services
Edmonton, Canada
[email protected]
Department of Physics
Ryerson University
Toronto, Canada
[email protected]
Abstract—Determining the quality of stored blood remains a
persistent challenge. It is well documented that morphological
and physiological changes occur during routine storage of whole
blood units impacting the ability of red blood cells (RBCs) to
deliver oxygen to tissues. We propose the use of photoacoustic
(PA) imaging and microscopy for assessing the storage-induced
changes that occur to RBCs. Whole blood units were measured
for a period of 40 days using a PA imaging system (VevoLAZR,
Fujifilm VisualSonics) and a PA microscope (SASAM, Kibero
GmbH). The oxygen carrying capacity of stored blood was
assessed by a two-wavelength approach by computing the oxygen
saturation as a function of time. Morphological changes were
assessed by PA microscopy through measuring the PA signals
and the respective frequency spectra from individual RBCs. As
RBCs aged, their oxygen carrying capacity significantly
diminished and distinct features in the frequency spectra from
the PA signals formed. This suggests that PA techniques are
sensitive to storage-induced RBC lesions and can potentially be
used to assess the quality of stored blood.
Keywords—photoacoustics; red blood cell storage lesion;
oxygen saturation; photoacoustic microscopy;
I. INTRODUCTION
The ex vivo storage of donated blood plays a paramount
role in modern medical care. Blood transfusions support a wide
variety of medical procedures spanning all surgical specialties
and to provide comfort during palliative care. As such, the
need for safe, easy and cheap access to blood is crucial [1].
Current biopreservation techniques involve the collection of
donor blood in polyvinyl chloride plastic containers with
solutions of anticoagulants such as citrate-phosphate-dextrose
and nutrients such as saline-adenine-glucose. The blood bag
mechanism allows for separation of red blood cells (RBCs),
plasma and platelets into satellite bags while storing the
components hypothermically (1-6 °C) in an additive solution
[2].
Concerns have been raised about the quality of the stored
RBCs and their safety for transfusions [3]–[5]. The term
“storage lesion” is introduced to summarize the physical and
978-1-4799-7049-0/14/$31.00 ©2014 IEEE
metabolic changes that occur to RBCs during hypothermic
storage. Several biochemical hallmarks of storage-induced
lesions have been identified [6]. These include the depletion of
ATP, decrease in pH and glycolysis, loss in membrane
carbohydrates and accumulating extracellular potassium. The
morphological changes cause RBCs to evolve from the smooth
biconcave shape to spheres and later irreversibly lose
membrane integrity and form echinocytic spines [1]. Although
there is debate as to whether the RBC age has observable
adverse effects on patient morbidity, pre-clinical models
suggest that there is diminished oxygen delivery of stored
RBCs [7]. Therefore, there is a need for a technology that is
capable of assessing the functional capacity of stored RBCs in
delivering oxygen to surrounding tissues.
We propose the use of photoacoustic (PA) imaging and
sensing for identifying the morphological and physiological
hallmarks of storage-induced RBC lesions. PA is an opticallybased imaging modality that has garnered significant attention
in the past decade for its ability to provide high resolution
structural, functional, metabolic and molecular imaging of the
vasculature [8]. Recent advances in PA microscopy (PAM)
have enabled the monitoring of the oxygen saturation (SO2) of
individual, flowing, single RBCs in the vascular network of
live mice brain [9]. Moreover, Zharov and colleagues have
pioneered PA flow cytometry, a family of sensing techniques
that have been used to monitor circulating tumor cells within
vasculature [10]. By adopting techniques derived from tissue
characterization approaches widely used in quantitative
ultrasound, our group has developed expertise in using the
frequency content of the PA signals to characterize the
underlying PA chromophore structure. These techniques have
been successfully applied to characterizing RBC morphologies
using PAM [11] and the aggregation of RBCs [12].
In this paper we present the application of PAM and PA
imaging of SO2 for the characterization of the morphological
and metabolic changes RBCs undergo during storage. The
techniques developed here can potentially be used to make
informed decisions about the safety of donated blood as it
pertains to the RBCs ability to oxygenate tissues.
1662
2014 IEEE International Ultrasonics Symposium Proceedings
II. MATERIALS AND METHODS
A. Blood collection and sampling
Ethics approval was granted by both Canadian Blood
Services and the Ryerson University Research Ethics Boards
before the initiation of this study. Whole blood units were
obtained from healthy volunteers under standard phlebotomy
procedures at the Canadian Blood Services’ Network Center
for Applied Development (Vancouver, Canada). Units were
stored at 4 °C during shipment and storage (up to 40 days post
collection).
At each testing point, units were thoroughly mixed and
under aseptic conditions, the desired volume was drawn
depending on the type of measurement performed. For the
PAM measurements, plasma was extracted using the standard
buffy coat method (10 min @ 1300 g) in order to dilute whole
blood to sufficiently low concentrations. For PA imaging of
RBC oxygenation, whole blood was used. In order to assess
morphological changes of RBCs as a function of storage time,
a blood smear was taken at each measurement day and >2500
RBCs were imaged using an inverted optical microscope
(Olympus Canada Inc., Richmond Hill, Canada).
B. PA imaging of whole blood
Whole blood samples were loaded onto a 2-mm diameter
vessel-mimicking phantom prepared for each measurement day
using 15% (w/v) porcine skin gelatin (Sigma Aldrich, Toronto,
Canada) dissolved in degassed water. The phantom was
imaged using the VevoLAZR PA imaging device (Fujifilm
VisualSonics, Toronto, Canada) [13]. A 40 MHz linear-array
transducer (256 elements) coupled to optical fibers was used to
deliver laser illumination (750/850 nm, 20 Hz, 30 mJ/pulse) to
the vessel-mimicking blood phantom and a total of 5 PA Bmode frames were recorded at each wavelength.
For each PA signal recorded, the frequency-domain power
spectrum was computed by taking the Fourier Transform of the
radio-frequency PA signal [11]. Frequency analysis was
limited to 200-500 MHz (-6 dB transducer bandwidth). On
each measurement day, deviations from the biconcave shape of
normal RBCs were assessed by counting the power spectra that
contained minima in the usable frequency range.
III. RESULTS AND DISCUSSION
A. RBC morphology and SO2 as a function of storage time
Fig. 1a shows optical microscopy images of aging RBCs
by comparing the images recorded 7 and 40 days postcollection. RBC morphology at day 40 deviates from the
typical biconcave shape that the majority of RBCs had at day
7. By storage day 40, a significant number of cells appear to
have formed irreversible echinocytic protrusions. This is
consistent with the observations that ATP depletion during
storage leads to RBC membrane injury that includes
microvessiculation and reduced surface area to volume ratio
[15].
An advantage of PA imaging is the ability to provide
functional information in addition to the underlying structure.
As shown in Fig. 1b, the oxygenation of the stored RBCs
changes significantly with storage time. At day 7, the RBCs
appear to be fully oxygenated as shown by the representative
SO2 map. By day 40, the oxygenation level has dropped to
almost 0%. This significant decrease in oxygen carrying
capacity occurs gradually after the first two weeks of storage
as shown by histograms of SO2 (Fig. 1d). By day 23, the
oxygenation level of the blood had dropped to ~40% and
further down to 20% by day 33.
The decrease in the oxygen carrying capacity of RBCs
correlates to the increase in the number of abnormal (i.e. not
biconcave) RBCs found as the blood aged. As measured by
optical microscopy, over 50% of RBCs deviate from their
biconcave shape by day 23 and this count remains high for the
remainder of the storage period (Fig. 1c). The storage-induced
morphological changes apparent from optical microscopy are
not the only element of interest during blood preservation. A
significant loss of 2,3-diphosphoglycerate (2,3-DPG) as a
function of storage time has been correlated to a drop in the
observed oxygen carrying capacity of RBCs [1]. This has
important clinical implications for transfusion medicine.
These results suggest that a non-invasive, high-resolution
imaging modality such as PA can assess the oxygen carrying
capacity of RBCs and perhaps assist in making clinicallyrelevant decisions about the safety of stored blood.
The SO2 of each vessel was computed using a standard two
wavelength approach [14]. The PA signal amplitude at each
wavelength was calculated on a pixel-by-pixel basis for the
region-of-interest (ROI) defined by the strongest PA signal. An
oxygenation map was constructed by computing the SO2 at
each pixel within the ROI and the corresponding histogram
distribution was extracted. The histograms were then averaged
across all frames for blood sampled on day 7, 12, 16, 23, 33
and 40 post-collection.
C. PA signal detection from single RBCs
All PAM measurements were completed using a PA
microscope consisting of an inverted Olympus IX81 optical
microscope fitted with a 375 MHz (f-number 1) transducer
(Kibero GmbH, Saarbrücken, Germany). A dilute suspension
of whole blood in native plasma was loaded onto an agarcoated glass bottom dish and was placed between the
transducer and the optical objective. A 532 nm laser (Teem
Photonics, Meylan, France) was focused through a 10× optical
objective to a ~10 μm illumination spot (150 nJ/pulse),
sufficient to irradiate an entire RBC. The PA signal from 100
individual RBCs was recorded for each measurement day (3,
10, 14, and 22 days post-collection).
B. Frequency content of PA signals and RBC morphology
Analysis of the RF data from PA imaging can be used to
infer the morphology of an absorbing structure. Since the RBC
structure changes during the formation of the storage lesion,
PA imaging can be used to monitor the quality of stored blood
as assessed by changes in RBC morphology. Our group has
pioneered PAM and PA imaging techniques based on the
power spectra of PA signals to characterize various
morphologies of RBCs [11], [12]. Normal, biconcave shaped
1663
2014 IEEE International Ultrasonics Symposium Proceedings
RBCs have a flat, featureless spectrum in the 200-500 MHz
frequency regime. As RBCs undergo morphological changes
such as echinocyte or spherocyte formation due to changes in
the suspending medium, frequency minima became apparent
in the same frequency band [11].
As ageing RBCs were measured using PAM, the number
of frequency spectra from individual RBCs that began to
exhibit features (spectral minima, shown in red) increased
with increasing storage time as shown in Fig. 2. By day 22,
the majority of RBCs measured contained features in their
power spectra that were not present in the early storage days.
This approach designed to monitor the features of frequency
spectra from individual RBCs shows potential in identifying
storage induced lesions. For example, microfluidic devices
can be designed to interrogate one individual cell at a time in
flowing blood. Along with the functional information obtained
by the SO2 approach discussed above, PA imaging appears to
be a reliable method for characterizing the storage-induced
changes in blood.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
IV. CONCLUSIONS
This paper reported a practical application of PA imaging,
namely the characterization of stored blood with the aim of
assessing its suitability for transfusions. Functional estimates of
the oxygen carrying capacity of blood along with the
information encoded in the frequency of PA signals provide a
reliable assessment of the changes occurring during the
prescribed storage of blood.
[9]
[10]
[11]
[12]
ACKNOWLEDGMENT
This research was undertaken, in part, thanks to funding
from the Canada Research Chairs Program awarded to M.
Kolios and a Collaborative Health Research Projects grant
awarded to M. Kolios and J. Acker. Funding to purchase the
equipment was provided by the Canada Foundation for
Innovation, the Ontario Ministry of Research and Innovation,
and Ryerson University. The authors would like to
acknowledge the assistance of Kirsten Cardinell during blood
measurements.
1664
[13]
[14]
[15]
J. R. Hess, ‘Red cell storage’, J. Proteomics, vol. 73, no. 3, pp. 368–
373, Jan. 2010.
R. Almizraq, J. D. R. Tchir, J. L. Holovati, and J. P. Acker, ‘Storage of
red blood cells affects membrane composition, microvesiculation, and
in vitro quality: RMPs and RBC quality measures’, Transfusion (Paris),
p. n/a–n/a, Feb. 2013.
C. R. Valeri, ‘Status Report on the Quality of Liquid and Frozen Red
Blood Cells’, Vox Sang., vol. 83, pp. 193–196, Aug. 2002.
M. Sayers and J. Centilli, ‘What if shelf life becomes a consideration in
ordering red blood cells?’, Transfusion (Paris), vol. 52, no. 1, pp. 201–
206, Jan. 2012.
P. M. Ness, ‘Does transfusion of stored red blood cells cause clinically
important adverse effects? A critical question in search of an answer
and a plan’, Transfusion (Paris), vol. 51, no. 4, pp. 666–667, Apr. 2011.
J. D. R. Tchir, J. P. Acker, and J. L. Holovati, ‘Rejuvenation of ATP
during storage does not reverse effects of the hypothermic storage
lesion: Rejuvenation and RBC Storage Lesion’, Transfusion (Paris),
vol. 53, no. 12, pp. 3184–3191, Dec. 2013.
N. J. H. Raat, P. M. Hilarius, T. Johannes, D. de Korte, C. Ince, and A.
J. Verhoeven, ‘Rejuvenation of stored human red blood cells reverses
the renal microvascular oxygenation deficit in an isovolemic transfusion
model in rats’, Transfusion (Paris), vol. 49, no. 3, pp. 427–424, Mar.
2009.
L. V. Wang and S. Hu, ‘Photoacoustic Tomography: In Vivo Imaging
from Organelles to Organs’, Science, vol. 335, no. 6075, pp. 1458–
1462, Mar. 2012.
L. Wang, K. Maslov, and L. V. Wang, ‘Single-cell label-free
photoacoustic flowoxigraphy in vivo’, Proc. Natl. Acad. Sci., vol. 110,
no. 15, pp. 5759–5764, Apr. 2013.
E. I. Galanzha and V. P. Zharov, ‘Photoacoustic flow cytometry’,
Methods, vol. 57, no. 3, pp. 280–296, Jul. 2012.
E. M. Strohm, E. S. L. Berndl, and M. C. Kolios, ‘Probing Red Blood
Cell Morphology Using High-Frequency Photoacoustics’, Biophys. J.,
vol. 105, no. 1, pp. 59–67, Jul. 2013.
E. Hysi, R. K. Saha, and M. C. Kolios, ‘Photoacoustic ultrasound
spectroscopy for assessing red blood cell aggregation and oxygenation’,
J. Biomed. Opt., vol. 17, no. 12, pp. 125006–125006, 2012.
A. Needles, A. Heinmiller, J. Sun, C. Theodoropoulos, D. Bates, D.
Hirson, M. Yin, and F. S. Foster, ‘Development and initial application
of a fully integrated photoacoustic micro-ultrasound system’, IEEE
Trans. Ultrason. Ferroelectr. Freq. Control, vol. 60, no. 5, pp. 888–
897, May 2013.
X. Wang, X. Xie, G. Ku, L. V. Wang, and G. Stoica, ‘Noninvasive
imaging of hemoglobin concentration and oxygenation in the rat brain
using high-resolution photoacoustic tomography’, J. Biomed. Opt., vol.
11, no. 2, p. 024015, 2006.
K. L. Scott, J. Lecak, and J. P. Acker, ‘Biopreservation of Red Blood
Cells: Past, Present, and Future’, Transfus. Med. Rev., vol. 19, no. 2, pp.
127–142, Apr. 2005.
2014 IEEE International Ultrasonics Symposium Proceedings
Fig. 1. (a) Representative optical microscopy images of RBCs recorded 7 and 40 days post-collection. Scale bar represents 15 μm. (b) Oxygenation maps of whole
blood displaying the SO2 values for blood at day 7 and 40. (c) Fraction of abnormal RBCs as assessed by optical microscopy as a function of storage time. (d)
Histogram distributions of the SO2 of RBCs measured over 40 days post-collection.
Fig. 2. Measured frequency power spectra from PAM of 100 single RBCs as a function of storage time for (a) 3, (b) 10, (c) 14) and (d) 22 days post-collection.
The blue represents the spectra for normal, biconcave shaped RBCs while the red denotes spectra from abnormal (spherocytic or echinocytic) RBCs that exhibit
deviations in the spectrum in the 200-500 MHz range.
1665
2014 IEEE International Ultrasonics Symposium Proceedings