Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
Phil. Trans. R. Soc. A (2011) 369, 4407–4424
doi:10.1098/rsta.2011.0250
REVIEW
Haemoglobin oxygen saturation as a biomarker:
the problem and a solution
B Y D AVID A. B OAS*
AND
M ARIA A NGELA F RANCESCHINI
Optics Division of the Martinos Center for Biomedical Imaging, Harvard
Medical School, Massachusetts General Hospital, 149 13th St rm 2301,
Charlestown, MA 02129, USA
Near-infrared spectroscopy measures of haemoglobin oxygen saturation are often used
as an indicator of sufficient oxygen delivery to assess injury susceptibility and tissue
damage. They have also often been used as a surrogate measure of oxygen metabolism.
Unfortunately, these measures have generally failed to provide robust indicators of injury
and metabolism. In this paper, we first review when haemoglobin oxygen saturation
does work as a robust indicator, and then detail when and why it fails for assessing
brain injury and breast cancer. Finally, we discuss the solution to obtain more robust
measures of tissue injury and cancer by combining oxygen saturation measurements
with measures of blood flow and volume to more accurately estimate oxygen
metabolism.
Keywords: near-infrared spectroscopy; haemoglobin oxygen saturation; oxygen metabolism;
muscle; breast; brain
1. Introduction
Oxidative metabolism provides more than 90 per cent of ATP production in
aerobic organisms. A sensor of cellular oxygen and oxygen utilization has long
been sought because of its importance in studying cellular energetics. A noninvasive oxygen sensor would have significant clinical utility for assessing tissue
injury and disease that give rise to impaired oxygen delivery and/or utilization,
and would be important for following tissue response to therapy. In 1977, Jobsis
[1] first exploited the weak absorption of tissue in the near infrared to measure
cytochrome c oxidase (CCO) in an intact organ, the feline brain, thus initiating
the development of near-infrared spectroscopy (NIRS). Cytochrome oxidase is an
important oxygen sensor because it reacts directly with oxygen at the terminus
of the cellular respiratory chain. An insufficient supply of oxygen would lead to a
*Author for correspondence ([email protected]).
One contribution of 20 to a Theo Murphy Meeting Issue ‘Illuminating the future of biomedical
optics’.
4407
This journal is © 2011 The Royal Society
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4408
D. A. Boas and M. A. Franceschini
build-up of reduced CCO. Oxidized CCO has an absorption band around 840 nm
that decreases when CCO is reduced. Thus, it appeared that NIRS would provide
the long-sought non-invasive oxygen sensor.
The technique for measuring CCO was quickly translated to the clinical
environment to assess whether neonates in the intensive care unit had a
sufficient supply of oxygen to the brain [2]. It was observed that hypoxia in
these infants led to a more reduced state of CCO, as expected. Following this
work, several other groups started using NIRS in the clinic [3,4]. In order
to estimate the oxidized or reduced state of CCO, it was also necessary to
untangle the contributing effects of oxygenated and deoxygenated haemoglobin
(HbO and HbR, respectively) to the near-infrared absorption spectrum. This
is a challenging task because of the low in vivo concentration of CCO relative
to haemoglobin. But by measuring the light absorption at several wavelengths
and correcting for the path length of light through the tissue at the different
wavelengths, it is in principle possible to separate the spectroscopic signatures of
the chromophores and estimate their in vivo concentrations [5]. By 1995, several
different algorithms had been developed for estimating the concentrations with
path length correction factors that varied with the particular application. This
raised concerns that using the wrong algorithm could give rise to artefacts in
the estimated CCO concentration. This was evidenced by Skov & Greisen [6],
who found that the switch to cardiopulmonary bypass in hypoxaemic children
with congenital heart disease led to a decrease in oxidized CCO, despite an
improvement in cerebral oxygen delivery as indicated by an increase in the
cerebral oxygenation index (i.e. HbO – HbR). They suggested that this nonphysiological finding was probably a result of using an incorrect algorithm
that did not accurately separate the confounding effects of HbO and HbR and
total haemoglobin concentration (HbT = HbO + HbR) when estimating CCO.
Matcher et al. [7] quantified the errors that arise from using the wrong algorithm,
which supported the growing sense that in vivo quantification of CCO was
not reliable.
While a few researchers continue the effort to improve the reliability of in
vivo CCO quantification through multi-spectral measurements that robustly
estimate and correct for wavelength-specific path length factors [8,9], NIRS use
shifted to measuring haemoglobin oxygen saturation (SO2 = HbO/HbT) as a
surrogate for tissue oxygenation and utilization for applications in the brain,
muscle and breast. The rationale was clear. Oxygen utilization with insufficient
oxygen delivery would certainly result in reduced oxygenation of haemoglobin.
Further, since the haemoglobin chromophores dominated the absorption spectrum
in the near infrared, the estimate of SO2 was less susceptible to algorithm
errors, although partial volume effects [10], myoglobin [11] and the presence
of other chromophores [12] can still confound the estimation of SO2 . Finally,
a decrease in SO2 was generally associated with an increase in the reduced form
of CCO [1,2,4].
It is becoming apparent that SO2 alone is not a robust indicator of sufficient
oxygen delivery to the tissue [13]. In many cases, HbT and blood flow (BF) are
proving to be better indicators of tissue status [14,15], and even better when
combined with SO2 to estimate oxygen consumption (OC) [16–18]. Below, we
review when SO2 works and when it should be replaced by more robust measures
of OC. While, historically, the first and major focus of NIRS was to study the
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
Review. Hb oxygenation as a biomarker
4409
brain haemoglobin oxygen saturation (see [19] and [20] as recent reviews), we
begin with the NIRS work in muscle, proceed to the application to breast cancer,
and finish with efforts in the brain.
2. Muscle
Skeletal muscle is strongly dependent on oxidative metabolism. Muscle OC can
increase 50-fold during exercise [21]. Any disruption of normal oxygen delivery or
consumption is likely to adversely affect exercise tolerance. Training is known to
improve muscle tolerance to increased energetic demands from exercise, and NIRS
is being used to assess muscle ability to adequately support metabolic demands
with the potential of using NIRS to guide and optimize individual training [22,23].
A number of chronic health conditions impair oxygen delivery to the muscle,
including peripheral vascular disease [24] and chronic heart failure [25].
Resting SO2 alone fails to detect disease-induced differences in oxygen delivery
or utilization, probably owing to the large differences in fat and muscle
distribution within and across individuals [26]. An advantage with peripheral
muscle is the possibility of performing manipulations that allow a direct
measure of OC. In fact, estimating OC from NIRS measurements of oxygenated
haemoglobin is a common practice in muscle applications. Pressure cuff occlusion
of BF in an arm or leg is an easy means of measuring tissue OC. The method is
quite simple, as described in Cheatle et al. [27], De Blasi et al. [28] and Casavola
et al. [29]. Rapidly increasing cuff pressure above arterial blood pressure will
occlude BF in the extremity with no associated change in total haemoglobin
concentration. A NIRS measurement downstream of the occlusion will reveal a
resultant increase in HbR and a decrease in HbO owing to OC. If the pressure was
increased sufficiently rapidly above arterial pressure, then there will be no change
in total haemoglobin concentration HbT = HbO + HbR. The rate of decrease of
HbO will thus be directly related to OC,
v
v
HbR = −4 HbO,
(2.1)
vt
vt
where the 4 accounts for the haemoglobin to oxygen molar ratio. An example
of this is shown in figure 1. Similar to the experiment described in Casavola
et al. [29], an ISS Inc. frequency-domain instrument was used to quantify the
HbO and HbR concentrations in leg muscle during a pressure cuff occlusion.
The cuff pressure was rapidly raised to 240 mm Hg, sufficient to occlude venous
and arterial BF. This pressure was maintained for approximately 8 min, as
indicated by the shaded area in figure 1. Very little change in HbT was
observed, confirming that arterial flow was rapidly occluded. During the first few
minutes, the concentration of HbO/HbR was seen to linearly decrease/increase.
Using equation (2.1), the OC in the leg muscle was calculated to be
OC = 3.2 mmol 100 ml−1 min−1 .
A partial pressure cuff occlusion will result in changes in HbT if the arteries
are not fully occluded, thus allowing blood to flow in, but veins are occluded
preventing blood from flowing out. In this case, De Blasi et al. [28] and Casavola
et al. [29] have indicated that it is still possible to estimate OC provided that a
correction is made for HbT changes. Specifically, any increase in HbT will result
OC = 4
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4410
D. A. Boas and M. A. Franceschini
haemoglobin concentration (µM)
110
HbT
100
90
ischaemia
80
HbO
70
60
50
40
HbR
30
20
0
2
4
6
8 10 12 14 16 18
time (min)
OC = 3.2 mmol 100 ml–1 min–1
Figure 1. Changes in the haemoglobin concentrations in leg muscle during an 8 min pressure cuff
occlusion of arterial and venous blood flow. The occlusion occurred during the period indicated by
the shaded box. During this period, deoxyhaemoglobin increased and oxy-haemoglobin decreased,
while total haemoglobin remained constant. The slope of the changes indicates a muscle oxygen
consumption (OC) of 3.2 mmol 100 ml−1 min−1 . (Online version in colour.)
from inflowing arterial blood. Since this haemoglobin is not fully oxygenated, it
will increase HbR. If we first subtract this contribution to the change in HbR,
then OC can be accurately estimated
100 − SaO2
v
HbR −
HbT ,
(2.2)
OC = 4
vt
100
where SaO2 is the arterial haemoglobin oxygen saturation.
Previous to this, De Blasi et al. [28] have tested the possibility of estimating
OC during muscle exercise without using occlusion. The results are encouraging,
showing similar results with and without occlusion. While there was no occlusion
and thus generally the inflowing blood cannot be neglected, they still estimated
OC using equation (2.1). We revisit this later in §5 below, but briefly this works
because the increase in OC is sufficiently large that the effects of inflowing and
outflowing blood can be neglected during the early transient.
3. Breast
Tumour physiology is marked by increased vascularization and hypoxia.
Historically, blood volume (BV) by itself, as an indicator of increased
vascularization, has not robustly detected and distinguished malignant and
benign lesions [30,31]. Several papers suggest the importance of haemoglobin
oxygenation for making the distinction, with the expectation that the hypoxic
tissue environment of the tumour would be reflected by reduced haemoglobin
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
Review. Hb oxygenation as a biomarker
4411
oxygen saturation [32–35]. One of the first papers to report tumour SO2 versus
healthy tissue found a reduction in SO2 [32]. In a follow-up report of 58 patients,
this reduction in SO2 was not found to be significant [36].
Imaging studies reported by four other groups over the past 10 years have
confirmed that SO2 is not significantly different in malignant lesions versus
healthy tissues or benign lesions. These results are summarized in figure 2.
The results from 93 patients reported by Grosenick et al. [37] and from 49
patients reported by Intes [38], both using time-domain optical transmission
imaging systems, are shown in figure 2a,b. Figure 2c summarizes the results from
14 patients reported by Ntziachristos et al. [39] using a time-domain optical
transmission imaging measurement combined with magnetic resonance imaging.
Figure 2d presents the results from 116 patients reported by Chance et al. [40]
using a continuous-wave optical imaging system in reflectance mode. From these
plots, it is clear that there is significant overlap among healthy, benign and
malignant lesions, when considering only SO2 . Some of the studies reveal that
this overlap is diminished when considering only HbT.
What is striking when examining these four different imaging studies is the
significant separation of malignant lesions from benign lesions and healthy tissue
when considering SO2 and HbT simultaneously, a point also suggested by Cerussi
et al. [36]. Note, in figure 2, the trend showing that tumour SO2 is low when
tumour HbT is comparable to non-malignant tissue and that tumour SO2 is
higher and can exceed non-malignant tissue SO2 when tumour HbT is higher.
This is a result of hyperperfusion, reflected by an increase in HbT, increasing
oxygen delivery to the tissue beyond the increase in oxygen demand. This is
an important result that we will return to in §5 that SO2 can actually increase
despite an increase in tissue oxygen metabolism and a possible decrease in tissue
oxygenation. It is because of the confounding impact of BF on oxygen delivery
that baseline SO2 alone is not a robust measure of oxygen metabolism or sufficient
oxygen delivery.
While baseline SO2 alone does not differentiate breast disease, it may be
possible to manipulate BF to induce changes in SO2 that can reveal the metabolic
needs of breast lesions. This manipulation would be analogous to the pressure
cuff occlusion done in muscle studies. This possibility is suggested by Carp
et al. [41], who reported that mild breast compression partially occludes BF and
produces a transient decrease in SO2 . By fitting the temporal decay of SO2 with
an oxygen mass balance equation, it is then possible to estimate tissue OC and
BF. Future development of this approach in breast imaging will probably provide
new biomarkers, OC and BF, for differentiating lesions.
4. Brain
A continuous supply of oxygen is critical to the health of the brain. As the brain
has no store of oxygen, any disruption in supply would quickly result in tissue
oxygen depletion and metabolic stress as the brain under anaerobic glycolysis
would not be able to produce sufficient energy to maintain normal cell function.
Thus, maintaining oxygen supply and utilization in the brain is a major goal
when treating brain injury, during cardiac surgery and during general anaesthesia.
The first attempt to use NIRS to obtain a more direct measure of brain oxygen
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4412
D. A. Boas and M. A. Franceschini
(a)
(b)
20
90
10
80
D SO 2
SO 2 (%)
100
70
60
50
100
150
HbT (mm)
–30
200
(c)
0
0.5
1.0
1.5
HbT (mm)
2.0
(d)
90
40
80
20
D SO 2 (%)
SO 2 (%)
–10
–20
50
40
0
0
70
60
50
40
0
0
–20
–40
100
200
300
HbT (mm)
400
–60
–5
0
5
D HbT (mm)
10
Figure 2. The results from four different optical breast imaging studies of malignant lesions
compared with benign lesions and healthy tissue are presented in scattered plots of SO2 versus
HbT. Details of the studies are provided in the text. While SO2 alone generally fails to distinguish
malignant lesions from non-malignant tissue, we see a diagonal separation indicated by the grey
lines that indicates that malignant lesions with HbT comparable to non-malignant tissue have a
lower SO2 , but that this SO2 increases as malignant lesion HbT increases. These results are adapted
from the authors. (a) Grosenick et al. [37], (b) Intes [38], (c) Ntziachristos et al. [39] and Chance
et al. [40]. Circles, malignant tumour; crosses, healthy or benign.
delivery and utilization, rather than rely on measurements of oxygen content at
sites distant from the brain (e.g. pulse oximetry and blood gases), was in 1985
when Brazy et al. [2] used NIRS in the neonatal intensive care unit. They used
a three-wavelength NIRS device with 775, 815 and 904 nm to achieve sensitivity
to changes in HbO, HbR and CCO. In this way they could monitor changes in
oxygen delivery as reflected by the oxygen saturation SO2 and total haemoglobin,
as well as changes in oxygen utilization as indicated by the CCO concentration.
Their results during brief apnoea in infants showed expected reductions in SO2
and CCO, with a prolonged recovery of CCO relative to SO2 . As discussed in
§1, subsequent studies revealed that the estimate of CCO was unreliable [7] and
therefore clinical measures with NIRS focused on using only SO2 .
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4413
Review. Hb oxygenation as a biomarker
hyperaemia from
collateral perfusion
3
occlusion
onset
D concentration (mM)
2
1
HbR
0
–1
HbT
–2
–3
–4
200
250
300
seconds
350
400
Figure 3. The changes in HbT and HbR observed over the right frontal cortex are shown for 1 min
before and 2 min after total balloon occlusion of the patient’s right internal carotid artery. HbT
is seen to initially drop, and then overshoot during a hyperaemia phase after collateral perfusion
through the circle of Willis, before returning to baseline although the right internal carotid is still
occluded. This return to baseline in HbT and HbR indicates that the patient’s right hemisphere
is sufficiently oxygenated from collateral perfusion through the circle of Willis. The separation
between the source and the detector was 3.5 cm and a differential path length factor of 6 was used
to calculate the concentration changes. This figure is reproduced courtesy of Juliette Selb.
One example of the usefulness of NIRS measures of cerebral SO2 is in assessing
collateral perfusion through the circle of Willis during carotid endarterectomy
[42–44] or carotid balloon occlusion test [45–48]. This is an important test
before carotid surgery since insufficient collateral perfusion during the surgery
would result in ischaemic damage to the brain. Immediately following carotid
occlusion, NIRS reveals a decrease in SO2 in the ipsilateral cortex. In patients with
sufficient collateral perfusion, SO2 generally returns to the baseline value within
a minute during the occlusion. In patients with insufficient collateral perfusion,
the SO2 remains below the baseline value for the duration of the occlusion. An
experimental example is shown in figure 3. Changes in HbT and HbR are shown
for 1 min before and 2 min after total balloon occlusion of the patient’s right
internal carotid artery. At the onset of the occlusion, we observe a dramatic
drop in HbT with little change in HbR. As the occlusion reduces the perfusion
pressure, the compliant vessels contract reducing HbT. The corresponding little
change in HbR indicates that SO2 has dropped as expected owing to an increased
extraction fraction to maintain a constant oxygen delivery. In this patient,
however, collateral perfusion through the circle of Willis responds within 30 s,
producing a hyperaemic overshoot in HbT with a corresponding decrease in
HbR, with everything returning to baseline levels within 2 min, although the
right internal carotid is still occluded.
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4414
D. A. Boas and M. A. Franceschini
NIRS measures of cerebral SO2 have also demonstrated the ability to
assess delivery of oxygen to the brain during cardiac surgery [49–53]. During
cardiopulmonary bypass surgery, drops in SO2 serve as an indicator of reduced BF
to the brain that can be corrected by the surgical team. Following bypass surgery,
the SO2 then provides assurance of the success of the surgery to restore and
maintain sufficient oxygen delivery to the brain. While the changes in SO2 clearly
correspond with cardiac surgery, there is no clear conclusion as to the utility
of using these measures to predict the outcome of the patient. In infants, the
outcome of the patient probably depends more on the cerebral damage acquired
from insufficient cardiac output before the surgery than on maintaining cerebral
perfusion during the surgery [49]. In adults, a recent randomized, prospective
study has demonstrated that patients with a lower mean SO2 during the surgery
had a longer length of stay in the hospital after the surgery owing to a greater
incidence of major organ dysfunction [51]. For adults, further study is needed to
optimize neuro-monitoring feedback during cardiac surgery, particularly to better
understand the importance of decreased cerebral oxygenation versus cerebral
emboli as the cause of neurological injury [50], the latter of which requires
transcranial ultrasound to detect.
Neonates in the intensive care unit receive numerous interventions that can
have an impact on oxygen delivery to the brain, such as manipulation of
haematocrit and blood pressure, yet the impact on the brain is not directly
measured nor is it understood. Thus, the possibility exists that such interventions
could adversely affect the patient. Wardle et al. [54] measured cerebral SO2
in anaemic and hypotensive neonates and found no significant difference with
respect to normal controls, suggesting that cerebral autoregulation maintained
oxygen delivery. When infants are asphyxiated at birth, an increase in cerebral
SO2 relative to healthy controls has been observed to correlate with abnormal
outcome [55]. In a different study, a reduction in SO2 relative to healthy controls
on the first day of life correlated with severity of a subsequently developed
intraventricular haemorrhage [56]. While these results indicate the possibility
of using SO2 to identify those patients who need aggressive treatment, there
was significant overlap in the SO2 values with infants that had normal outcome.
To prevent over-treatment, it has been suggested that NIRS instruments need
to be improved to provide a more precise measure of SO2 in order to distinguish
subtle differences with more power [57]. An important relevant observation is that
the measured SO2 can vary by more than 10 per cent with repeated placement
of the optodes, presumably because of subtle differences in underlying tissue
heterogeneity [57,58], thus indicating the need to reduce this variation through
repeated probe placement or spatially resolved oximetry [56].
By introducing a blood tracer with optical contrast, NIRS measurements can
be made to estimate BF and BV. Oxygen is an ideal tracer to use with NIRS
to estimate BF and BV, as detailed by Reynolds et al. [59] and Wyatt et al.
[60], respectively. An early demonstration of the predictive power of BF and
BV from Meek et al. [14] showed that increases in BF and BV above normal
levels in infants with perinatal asphyxia correlated with adverse outcome of the
infants. They suggest that acute hypoxia–ischaemia releases several vasoactive
agents, resulting in the most severely affected infants having the greatest cerebral
hyperaemia. While these results demonstrate the promising potential of NIRS
for assessing brain injury in newborns, they state that BV measurements as an
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
Review. Hb oxygenation as a biomarker
4415
early marker of the severity of perinatal asphyxia are of limited value because of
a poor specificity of 38 per cent (sensitivity of 85%) and the poor reproducibility
in making such oxygen manipulations. They suggest that upcoming advances
in NIRS technology may make it possible in the future to improve specificity
by obtaining absolute haemoglobin concentrations and obtaining more direct
measures of oxygen extraction fraction and oxygen metabolism [61,62] without
the need for introducing a blood tracer.
Multi-distance frequency-domain technology provides a means to estimate
absolute absorption and scattering properties of the tissue, and thus estimate
absolute haemoglobin concentrations without the need for vascular tracers [63].
This technology offers a robust approach to estimate cerebral BV and oxygenation
in infants, as demonstrated by measures of the twofold increase in cerebral BV in
the first year of life by Franceschini et al. [64]. Grant et al. [15] used this approach
to measure alterations in cerebral BV and SO2 in newborns with a variety of brain
injuries compared with infants without brain injury. Consistent with the earlier
work of others, they found no significant difference in SO2 between the infants
with brain injury and the stable control infants without brain injury. Similar to
Meek et al. [14], they also found that cerebral BV was significantly elevated in
the brain-injured infants, but, because no vascular tracers are needed, it is now
possible to apply this approach to all patients.
5. More robust measures of oxygen consumption
From the above examples, we have seen that, when there are transient changes in
BF or OC, changes in haemoglobin oxygenation can reveal information about
oxygen delivery and consumption. It is important to emphasize that, here,
transient refers to the SO2 changes that occur during the first minutes of the
oxygen delivery or consumption change.
Generally, OC can be estimated by the difference between the quantity of
oxygen being delivered by the arteries and that drained by the veins in a tissue of
interest. In a steady state, OC can be calculated from the flow of blood multiplied
by the difference between the arterial oxygen saturation (SaO2 ) and the venous
oxygen saturation (SvO2 ), where the saturations are given as a fraction. That is,
OC = 4
HGB
BF (SaO2 − SvO2 ),
MWHb
(5.1)
where HGB is the haemoglobin concentration in the blood (g l−1 ) and MWHb
is the molecular weight of haemoglobin (g mol−1 ) and the 4 accounts for the
haemoglobin to oxygen molar ratio. An abrupt change in BF (l 100 g−1 min−1 )
or OC (mol O2 100 g−1 min−1 ), as can result from a pressure cuff occlusion or
exercise, will result in an imbalance of the left- and right-hand side of equation
(5.1), thus giving rise to temporal changes in HbO and HbR. The temporal
changes are described by Carp et al. [41]
HGB
dHbO
=
BF(SaO2 − SvO2 ) − 14 OC.
dt
MWHb
(5.2)
Subject to the conditions that BF and/or OC change abruptly at t = 0 and then
remain constant, we obtain the solution linearized for small changes in HbO and
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4416
D. A. Boas and M. A. Franceschini
SvO2 , as occurs at time scales that are short compared with oxygen delivery and
consumption time scales,
DHbO(t) =
1 HGB
BF (SaO2 − SO2o ) t − 14 OCt.
VFV MWHb
(5.3)
Here, for simplicity, we assume that the haemoglobin saturation measured by
NIRS is the volume fraction weighted average of SaO2 and SvO2 , i.e. SO2 =
VFA SaO2 + VFV SvO2 , where VFV (VFA ) is the venous (arteriole) volume fraction
and VFV + VFA = 1. The subscript ‘o’ refers to the initial value. During a pressure
cuff arterial occlusion, BF = 0 and we arrive at the same equation relating OC to
DHbO(t) as used in muscle studies (equation (2.1) above). We still arrive at the
same equation, if the rate of OC, i.e. 1/4 OC, is much greater than the rate of
oxygen delivery,
HGB
BF (SaO2 − SO2o ),
2
MWHb
as demonstrated by De Blasi et al. [28], when estimating OC during exercise with
and without cuff occlusion. These equations demonstrate formally when we can
use transient changes in SO2 (or HbR) to estimate OC.
For measurements of the brain or breast, it is generally not feasible to introduce
an abrupt change in BF or OC. As a result, we must rely on equation (5.1) and
use measurements of SO2 combined with measurements of BF to estimate OC.
The need for combining these measurements is nicely illustrated by the breast
imaging studies depicted in figure 2. While it was reasonable to expect to observe
a decreased SO2 because of increased metabolic activity in the tumour, this was
not observed because of the confounding effect of BF. Higher BF, as suggested by
an increased HbT, increases oxygen delivery and thus SO2 . By considering both
BF and SO2 , as we did with the diagonal line separating malignant breast lesions
from the healthy tissue and benign lesions in figure 2, we see that malignant
lesions are generally consuming more oxygen.
Recognizing the confounding effect of BF on estimating OC from SO2 , Grant
et al. [15] applied equation (5.1), using BV as a surrogate measure of BF, to
estimate changes in OC in brain-injured infants relative to healthy controls
and found that OC was significantly elevated. The summary of results from 14
brain-injured infants compared with 29 infants without brain disorders is shown
in figure 4. These infants were in the acute stage of brain injury and so the
observation of elevated BV and OC suggests that the brain disorders measured
shared a common mechanism of hyperperfusion, possibly arising from increased
neuronal activity owing to excitotoxic processes evolving in the tissue over days.
In the chronic stage after approximately 7 days, they have observed that OC
drops to zero, thus providing a time scale for cell death in these injuries.
As discussed by Leung et al. [65] and Boas & Payne [66], care must be taken
in estimating BF from BV. It is common to use the Grubb relation [65,67],
but this relation may not be valid under all conditions [68], particularly pathophysiological conditions. Thus, it is preferable to obtain independent measures of
BF as offered by diffuse correlation spectroscopy (DCS) when estimating OC
[18]. A recent example by Roche-Labarbe et al. [17] in neonates during the
first six weeks of life nicely demonstrates first the uncoupling that can arise
between BF and BV, and second the improved estimates of OC that can be
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4417
Review. Hb oxygenation as a biomarker
2.6
2.4
rOC with respect to normals
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
1.2
1.6
2.0
2.4
2.8
3.2
BV (ml 100 mg–1)
3.6
4.0
Figure 4. Measurements of cerebral OC versus blood volume (BV) are shown for acutely braininjured newborns compared with newborns without brain injury. OC is reported relative to healthy
age-matched controls. In general, acutely brain-injured newborns have increased OC and increased
BV with respect to non-brain-injured newborns. This is indicated by the diagonal black line
separating brain-injured from non-brain-injured infants with a sensitivity of 78.6% and a specificity
of 96.6%. Filled circles, brain injured; triangles, unstable; diamonds, stable; open circles, healthy.
obtained with independent measures of BF, HbT and SO2 . Fifty-six premature
infants were measured longitudinally during their hospital stay for a total of
230 measurements. Figure 5 reports average results as a function of corrected
gestational age (cGA) for BV measured with the frequency-domain multi-distance
method, an index of blood flow (BFi) measured with DCS, and the relative OC
calculated with respect to infants born at 35 weeks GA. OC was calculated using
either BFi measured by DCS or using Grubb’s relation to estimate BF from the
measured BV. Figure 5a,b shows the uncoupling between BV and BF, with BV
more or less constant with cGA and BF increasing with age. This uncoupling is
probably due to the large decreases in haematocrit during the first two months
of life, which alter haemoglobin content and blood viscosity. As a result, in the
newborn population, quantifying OC using BV as a surrogate for BF not only
gives an incorrect result, but also produces a larger variance in OC as shown in
figure 5c,d.
As another example of combining measures of BF with SO2 to estimate OC,
Zhou et al. [69] combined DCS with frequency-domain NIRS to measure breast
tumour response to chemotherapy. The results from a single patient are shown
in figure 6 following the tumour response from baseline on day 0 to 7 days after
chemotherapy. A significant decrease in HbT and SO2 is observed during the first
7 days following chemotherapy. Interestingly, BF first significantly increased on
day 3 and then decreased below baseline on days 4–7, revealing an uncoupling
between HbT and BF that could be a result of a reduced haematocrit following
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4418
D. A. Boas and M. A. Franceschini
rOC with volume
(c)
BFi (arb. units) (× 10–6)
(b)
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
9
8
7
6
5
4
3
2
1
0
(d)
2.5
2.5
2.0
2.0
1.5
1.0
0.5
0
24 26 28 30 32 34 36 38 40 42
cGA (weeks)
rOC with flow
BV (ml 100 g–1)
(a)
1.5
1.0
0.5
0
24 26 28 30 32 34 36 38 40 42
cGA (weeks)
Figure 5. Plotted are average results of (a) BV and (b) blood flow index (BFi) measured on
56 newborn premature infants and measured longitudinally over several weeks on each infant.
Standard errors are indicated except for the last 2 points, which are from single measurements.
After 28 weeks, corrected gestational age (cGA) and BV remain constants while BFi continually
increases. If we estimate OC assuming Grubb’s relation to approximate BF from BV, then OC is
seen to remain constant from 28 to 40 weeks cGA as shown in (c). Instead, using the experimentally
measured BF, we see that OC actually increases significantly over this period of time as shown in
(d). Further, we see that the variation across the population is greatly reduced when using BFi.
chemotherapy [70], changes in vascular morphology and resistance, changes in
intra-tumour pressure or a complex combination of these factors. The estimate of
OC revealed that it first increased on day 3 and then decreased below baseline,
suggesting a complex metabolic response to the therapy that is interesting for
better understanding tumour response to therapy and will potentially help to
better identify therapy responders from non-responders in the first week of
therapy.
As suggested by the above-mentioned studies, multi-modal estimates of OC will
enable new physiological studies in health and disease leading to new diagnostic
and therapeutic approaches. One interesting physiological finding is the constancy
of SO2 in developing human infants over the first year of life despite a doubling
of BV [64] and a presumed twofold to fourfold increase in BF. This observed
tight coupling of oxygen demand and oxygen supply to maintain a constant
SO2 suggests that SO2 is an important factor in the regulation of BF over a
long-term time scale [71]. This tight coupling has recently been observed at the
level of the microvasculature by Yaseen et al. [72], in which, in individual rats,
BF in distinct ascending venules draining the cortex can vary by 20-fold over a
distance of 500 mm, yet SO2 within those individual draining venules is constant.
These results indicate that oxygen demand in a volume of tissue drained by
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4419
Review. Hb oxygenation as a biomarker
1.4
HbT
BF
SO2
OC
1.3
relative change
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0
1
2
3
4
5
6
7
days
Figure 6. Breast tumour response to chemotherapy, as measured from a single patient, reveals an
increase in BF and OC on day 3 following therapy, which then reduces below baseline along with
HbT and SO2 . These results are a composite of the results presented in Zhou et al. [69], and are
presented here courtesy of the authors. Small filled circles, HbT; open circles, BF; crosses, SO2 ;
squares, OC. (Online version in colour.)
a given ascending venule is matched to the flow in that venule to maintain a
constant SO2 . These are provocative results that are likely to be relevant to
pathological conditions where chronic hypoxia may be resulting in angiogenesis
to normalize SO2 .
6. Concluding remarks
NIRS has tantalized us with the prospect of obtaining non-invasive measurements
of an important oxygen sensor, CCO. Work continues to realize a robust measure
of CCO that can be used routinely in humans. Until that is achieved, NIRS
efforts have instead used the far more robust measures of haemoglobin SO2 as a
surrogate tissue oxygen sensor. As we have discussed in this paper, when there is
an abrupt change in BF or OC, transient changes in SO2 can reveal information
about tissue OC. However, when abrupt changes in BF or OC are not feasible, or
the measurement cannot be made in the first minute, then inference of tissue
OC or sufficient oxygen delivery from SO2 measures alone is confounded by
BF. This is nicely illustrated by studies of SO2 in asphyxiated newborn infants
revealing a slight increase in SO2 relative to healthy controls, which has been
interpreted as a reduction in oxygen demand from cerebral injury leading to
cellular energy failure. Alternatively, the injury could elicit a strong BF response
to overcompensate for an increased metabolic demand. Thus, generally speaking,
in order to robustly assess tissue OC and sufficient oxygen delivery, it is necessary
to measure both SO2 and BF. More and more groups are now exploiting all
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4420
D. A. Boas and M. A. Franceschini
optical methods, and optical methods combined with other imaging modalities,
to obtain this important physiological measurement and use it in a broad range
of applications.
Britton Chance’s leadership in biomedical optics and mitochondrial respiration has inspired us
in our endeavour to translate optical methods to the clinic and to unravel the dynamic energetic
processes in health and disease. We will miss him but are comforted by the opportunity to follow in
his giant footsteps. In preparing this presentation and manuscript, we are grateful for the assistance
of and discussions from Juliette Selb, Nadege Roche-Labarbe, Ellen Grant, Abbas Yaseen, Bruce
Tromberg, Arjun Yodh and Marco Ferrari. We acknowledge support from the National Institute
of Health under R01-EB06385, P41-14075, R01-HD042908, R01-NS057476 and P01-NS055104.
References
1 Jobsis, F. F. 1977 Nonivasive, infrared monitoring of cerebral and myocardial oxygen sufficiency
and circulatory parameters. Science 198, 1264–1267. (doi:10.1126/science.929199)
2 Brazy, J. E., Lewis, D. V. & Mitnick, M. H. 1985 Noninvasive monitoring of cerebral oxygenation
in preterm infants: preliminary observations. Pediatrics 75, 217–225.
3 Ferrari, M., Giannini, I., Sideri, G. & Zanette, E. 1985 Continuous non invasive monitoring of
human brain by near infrared spectroscopy. Adv. Exp. Med. Biol. 191, 873–882.
4 Wyatt, J. S., Cope, M., Delpy, D. T., Wray, S. & Reynolds, E. O. R. 1986 Quantification
of cerebral oxygenation and haemodynamics in sick newborn infants by near infrared
spectrophotometry. Lancet 2, 1063–1066. (doi:10.1016/S0140-6736(86)90467-8)
5 Wray, S., Cope, M., Delpy, D. T., Wyatt, J. S. & Reynolds, E. O. 1988 Characterization of
the near infrared absorption spectra of cytochrome aa3 and haemoglobin for the non-invasive
monitoring of cerebral oxygenation. Biochim. Biophys. Acta. 933, 184–192. (doi:10.1016/00052728(88)90069-2)
6 Skov, L. & Greisen, G. 1994 Apparent cerebral cytochrome aa3 reduction during cardiopulmonary bypass in hypoxaemic children with congenital heart disease. A critical analysis
of in vivo near-infrared spectrophotometric data. Physiol. Meas. 15, 447–457. (doi:10.1088/
0967-3334/15/4/006)
7 Matcher, S. J., Elwell, C. E., Cooper, C. E., Cope, M. & Delpy, D. T. 1995 Performance
comparison of several published tissue near-infrared spectroscopy algorithms. Anal. Biochem.
227, 54–68. (doi:10.1006/abio.1995.1252)
8 Cooper, C. E. et al. 1999 Use of mitochondrial inhibitors to demonstrate that cytochrome
oxidase near-infrared spectroscopy can measure mitochondrial dysfunction noninvasively in the
brain. J. Cereb. Blood Flow. Metab. 19, 27–38. (doi:10.1097/00004647-199901000-00003)
9 Tachtsidis, I., Tisdall, M., Leung, T. S., Cooper, C. E., Delpy, D. T., Smith, M. & Elwell, C. E.
2007 Investigation of in vivo measurement of cerebral cytochrome-c-oxidase redox changes using
near-infrared spectroscopy in patients with orthostatic hypotension. Physiol. Meas. 28, 199–211.
(doi:10.1088/0967-3334/28/2/008)
10 Strangman, G., Franceschini, M. A. & Boas, D. A. 2003 Factors affecting the accuracy of nearinfrared spectroscopy concentration calculations for focal changes in oxygenation parameters.
Neuroimage 18, 865–879. (doi:10.1016/S1053-8119(03)00021-1)
11 Hoofd, L., Colier, W. & Oeseburg, B. 2003 A modeling investigation to the possible role of
myoglobin in human muscle in near infrared spectroscopy (NIRS) measurements. Adv. Exp.
Med. Biol. 530, 637–643.
12 Ferrari, M., Mottola, L. & Quaresima, V. 2004 Principles, techniques, and limitations of near
infrared spectroscopy. Can. J. Appl. Physiol. 29, 463–487. (doi:10.1139/h04-031)
13 Weiss, M., Dullenkopf, A., Kolarova, A., Schulz, G., Frey, B. & Baenziger, O. 2005 Nearinfrared spectroscopic cerebral oxygenation reading in neonates and infants is associated with
central venous oxygen saturation. Paediatr. Anaesth. 15, 102–109. (doi:10.1111/j.1460-9592.
2005.01404.x)
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
Review. Hb oxygenation as a biomarker
4421
14 Meek, J. H., Elwell, C. E., McCormick, D. C., Edwards, A. D., Townsend, J. P., Stewart,
A. L. & Wyatt, J. S. 1999 Abnormal cerebral haemodynamics in perinatally asphyxiated
neonates related to outcome. Arch. Dis. Child Fetal Neonatal Ed. 81, F110–F115. (doi:10.1136/
fn.81.2.F110)
15 Grant, P. E., Roche-Labarbe, N., Surova, A., Themelis, G., Selb, J., Warren, E. K.,
Krishnamoorthy, K. S., Boas, D. A. & Franceschini, M. A. 2009 Increased cerebral blood volume
and oxygen consumption in neonatal brain injury. J. Cereb. Blood Flow. Metab. 29, 1704–1713.
(doi:10.1038/jcbfm.2009.90)
16 Yoxall, C. W. & Weindling, A. M. 1998 Measurement of cerebral oxygen consumption in the
human neonate using near infrared spectroscopy: cerebral oxygen consumption increases with
advancing gestational age. Pediatr. Res. 44, 283–290. (doi:10.1203/00006450-199809000-00004)
17 Roche-Labarbe, N., Stefan, A., Carp, S. A., Surova, A., Patel, M., Boas, D. A., Grant, P. E. &
Franceschini, M. A. 2010 Noninvasive optical measures of CBV, St O(2), CBF index, and
rCMRO(2) in human premature neonates’ brains in the first six weeks of life. Hum. Brain
Mapp. 31, 341–352. (doi:10.1002/hbm.20868)
18 Durduran, T. et al. 2010 Optical measurement of cerebral hemodynamics and oxygen
metabolism in neonates with congenital heart defects. J. Biomed. Opt. 15, 037004. (doi:10.1117/
1.3425884)
19 Murkin, J. M. & Arango, M. 2009 Near-infrared spectroscopy as an index of brain and tissue
oxygenation. Br. J. Anaesth. 103(Suppl 1), i3–i13. (doi:10.1093/bja/aep299)
20 Highton, D., Elwell, C. & Smith, M. 2010 Noninvasive cerebral oximetry: is there light at the end
of the tunnel? Curr. Opin. Anaesthesiol. 23, 576–581. (doi:10.1097/ACO.0b013e32833e1536)
21 Hamaoka, T., McCully, K. K., Quaresima, V., Yamamoto, K. & Chance, B. 2007 Near-infrared
spectroscopy/imaging for monitoring muscle oxygenation and oxidative metabolism in healthy
and diseased humans. J. Biomed. Opt. 12, 062105. (doi:10.1117/1.2805437)
22 Kell, R. T., Farag, M. & Bhambhani, Y. 2004 Reliability of erector spinae oxygenation and blood
volume responses using near-infrared spectroscopy in healthy males. Eur. J. Appl. Physiol. 91,
499–507. (doi:10.1007/s00421-003-1014-0)
23 Austin, K. G., Daigle, K. A., Patterson, P., Cowman, J., Chelland, S. & Haymes, E. M.
2005 Reliability of near-infrared spectroscopy for determining muscle oxygen saturation during
exercise. Res. Q. Exerc. Sport 76, 440–449.
24 Yu, G., Durduran, T., Zhou, C., Yodh, A. G., Lech, G., Chance, B. & Mohler III, E. R.
2005 Time-dependent blood flow and oxygenation in human skeletal muscles measured with
noninvasive near-infrared diffuse optical spectroscopies. J. Biomed. Opt. 10, 024027.
(doi:10.1117/1.1884603)
25 Belardinelli, R. 1999 Monitoring skeletal muscle oxygenation during exercise by near infrared
spectroscopy in chronic heart failure. Congest. Heart Fail. 5, 116–119.
26 McCully, K. K., Landsberg, L., Suarez, M., Hofmann, M. & Posner, J. D. 1997 Identification of
peripheral vascular disease in elderly subjects using optical spectroscopy. J. Gerontol. A Biol.
Sci. Med. Sci. 52, B159–165. (doi:10.1093/gerona/52A.3.B159)
27 Cheatle, T. R., Potter, L. A., Cope, M., Delpy, D. T., Coleridge Smith, P. D. & Scurr, J.
H. 1991 Near-infrared spectroscopy in peripheral vascular disease. Br. J. Surg. 78, 405–408.
(doi:10.1002/bjs.1800780408)
28 De Blasi, R. A., Cope, M., Elwell, C., Safoue, F. & Ferrari, M. 1993 Noninvasive measurement
of human forearm oxygen consumption by near infrared spectroscopy. Eur. J. Appl. Physiol.
Occup. Physiol. 67, 20–25. (doi:10.1007/BF00377698)
29 Casavola, C., Paunescu, L. A., Fantini, S. & Gratton, E. 2000 Blood flow and oxygen
consumption with near-infrared spectroscopy and venous occlusion: spatial maps and the effect
of time and pressure of inflation. J. Biomed. Opt. 5, 269–276. (doi:10.1117/1.429995)
30 Brenner, R. J. 1982 X-ray mammography and diaphanography in screening for breast cancer.
J. Reprod. Med. 27, 679–684.
31 Drexler, B., Davis, J. L. & Schofield, G. 1985 Diaphanography in the diagnosis of breast cancer.
Radiology 157, 41–44.
32 Tromberg, B. J., Shah, N., Lanning, R., Cerussi, A., Esponoza, J., Pham, T., Svaasand, L. &
Butler, J. 2000 Non-invasive in vivo characterization of breast tumors using photon migration
spectroscopy. Neoplasia 2, 26–40. (doi:10.1038/sj.neo.7900082)
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4422
D. A. Boas and M. A. Franceschini
33 Ntziachristos, V. & Chance, B. 2001 Probing physiology and molecular function using optical
imaging: applications to breast cancer. Breast Cancer Res. 3, 41–46. (doi:10.1186/bcr269)
34 Heffer, E., Pera, V., Schutz, O., Siebold, H. & Fantini, S. 2004 Near-infrared imaging of
the human breast: complementing hemoglobin concentration maps with oxygenation images.
J. Biomed. Opt. 9, 1152–1160. (doi:10.1117/1.1805552)
35 Gibson, A. & Dehghani, H. 2009 Diffuse optical imaging. Phil. Trans. R. Soc. A 367, 3055–3072.
(doi:10.1098/rsta.2009.0080)
36 Cerussi, A., Shah, N., Hsiang, D., Durkin, A., Butler, J. & Tromberg, B. J. 2006 In vivo
absorption, scattering, and physiologic properties of 58 malignant breast tumors determined
by broadband diffuse optical spectroscopy. J. Biomed. Opt. 11, 044005. (doi:10.1117/1.2337546)
37 Grosenick, D. et al. 2004 Concentration and oxygen saturation of haemoglobin of 50 breast
tumours determined by time-domain optical mammography. Phys. Med. Biol. 49, 1165–1181.
(doi:10.1088/0031-9155/49/7/006)
38 Intes, X. 2005 Time-domain optical mammography SoftScan: initial results. Acad. Radiol.
12, 934–947. (doi:10.1016/j.acra.2005.05.006)
39 Ntziachristos, V., Yodh, A. G., Schnall, M. D. & Chance, B. 2002 MRI-guided diffuse optical
spectroscopy of malignant and benign breast lesions. Neoplasia 4, 347–354. (doi:10.1038/sj.neo.
7900244)
40 Chance, B., Nioka, S., Zhang, J., Conant, E. F., Hwang, E., Briest, S., Orel, S. G., Schnall,
M. D. & Czerniecki, B. J. 2005 Breast cancer detection based on incremental biochemical and
physiological properties of breast cancers: a six-year, two-site study. Acad. Radiol. 12, 925–933.
(doi:10.1016/j.acra.2005.04.016)
41 Carp, S. A., Kauffman, T., Fang, Q., Boas, D., Rafferty, E., Moore, R. & Kopans, D. 2006
Compression-induced changes in the physiological state of the breast as observed through
frequency domain photon migration measurements. J. Biomed. Opt. 11, 064016. (doi:10.1117/
1.2397572)
42 Kuroda, S., Houkin, K., Abe, H., Hoshi, Y. & Tamura, M. 1996 Near-infrared monitoring
of cerebral oxygenation state during carotid endarterectomy. Surg. Neurol. 45, 450–458.
(doi:10.1016/0090-3019(95)00463-7)
43 Samra, S. K., Dy, E. A., Welch, K., Dorje, P., Zelenock, G. B. & Stanley, J. C. 2000 Evaluation
of a cerebral oximeter as a monitor of cerebral ischemia during carotid endarterectomy.
Anesthesiology 93, 964–970. (doi:10.1097/00000542-200010000-00015)
44 Rigamonti, A., Scandroqlio, M., Minicucci, F., Magrin, S., Carozzo, A. & Casati, A. 2005
A clinical evaluation of near-infrared cerebral oximetry in the awake patient to monitor
cerebral perfusion during carotid endarterectomy. J. Clin. Anesth. 17, 426–430. (doi:10.1016/j.
jclinane.2004.09.007)
45 Kuroda, S., Houkin, K., Abe, H. & Tamura, M. 1996 Cerebral hemodynamic changes during
carotid artery balloon occlusion monitored by near-infrared spectroscopy. Neurol. Med. Chir.
(Tokyo) 36, 78–86. (doi:10.2176/nmc.36.78)
46 Kaminogo, M., Ochi, M., Onizuka, M., Takahata, H. & Shibata, S. 1999 An additional
monitoring of regional cerebral oxygen saturation to HMPAO SPECT study during balloon
test occlusion. Stroke 30, 407–413. (doi:10.1161/01.STR.30.2.407)
47 Calderon-Arnulphi, M., Alaraj, A., Amin-Hanjani, S., Mantulin, W. W., Polzonetti, C. M.,
Gratton, E. & Charbel, F. T. 2007 Detection of cerebral ischemia in neurovascular surgery
using quantitative frequency-domain near-infrared spectroscopy. J. Neurosurg. 106, 283–290.
(doi:10.3171/jns.2007.106.2.283)
48 Kakihana, Y., Matsunaga, A., Yasuda, T., Imabayashi, T., Kanmura, Y. & Tamura, M. 2008
Brain oxymetry in the operating room: current status and future directions with particular
regard to cytochrome oxidase. J. Biomed. Opt. 13, 033001. (doi:10.1117/1.2940583)
49 Toet, M. C., Flinterman, A., van de Laar, I., de Vries, J. W., Bennink, G. B. W. E., Uiterwaal,
C. S. P. M. & van Bel, F. 2005 Cerebral oxygen saturation and electrical brain activity before,
during, and up to 36 hours after arterial switch procedure in neonates without pre-existing
brain damage: its relationship to neurodevelopmental outcome. Exp. Brain Res. 165, 343–350.
(doi:10.1007/s00221-005-2300-3)
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
Review. Hb oxygenation as a biomarker
4423
50 Razumovsky, A. Y., Gugino, L. D. & Owen, J. H. 2006 Advanced neurologic monitoring for
cardiac surgery. Curr. Cardiol. Rep. 8, 17–22. (doi:10.1007/s11886-006-0005-2)
51 Murkin, J. M. et al. 2007 Monitoring brain oxygen saturation during coronary bypass
surgery: a randomized, prospective study. Anesth. Analg. 104, 51–58. (doi:10.1213/01.ane.
0000246814.29362.f4)
52 Ohmae, E., Oda, M., Suzuki, T., Yamashita, Y., Kakihana, Y., Matsunaga, A., Kanmura, Y. &
Tamura, M. 2007 Clinical evaluation of time-resolved spectroscopy by measuring cerebral hemodynamics during cardiopulmonary bypass surgery. J. Biomed. Opt. 12, 062112. (doi:10.
1117/1.2804931)
53 Fedorow, C. & Grocott, H. P. 2010 Cerebral monitoring to optimize outcomes after cardiac
surgery. Curr. Opin. Anaesthesiol. 23, 89–94. (doi:10.1097/ACO.0b013e3283346d10)
54 Wardle, S. P., Yoxall, C. W. & Weindling, A. M. 2000 Determinants of cerebral fractional
oxygen extraction using near infrared spectroscopy in preterm neonates. J. Cereb. Blood Flow.
Metab. 20, 272–279. (doi:10.1097/00004647-200002000-00008)
55 Zaramella, P., Saraceni, E., Freato, F., Faicon, E., Suppiei, A., Milan, A., Laverda, A. M. &
Chiandetti, L. 2007 Can tissue oxygenation index (TOI) and cotside neurophysiological
variables predict outcome in depressed/asphyxiated newborn infants? Early Hum. Dev. 83,
483–489. (doi:10.1016/j.earlhumdev.2006.09.003)
56 Sorensen, L. C., Maroun, L. L., Borch, K., Lou, H. C. & Greisen, G. 2008 Neonatal cerebral
oxygenation is not linked to foetal vasculitis and predicts intraventricular haemorrhage in
preterm infants. Acta Paediatr. 97, 1529–1534. (doi:10.1111/j.1651-2227.2008.00970.x)
57 Greisen, G. 2006 Is near-infrared spectroscopy living up to its promises? Semin. Fetal Neonatal
Med. 11, 498–502. (doi:10.1016/j.siny.2006.07.010)
58 Dullenkopf, A., Kolarova, A., Schulz, G., Frey, B., Baenziger, O. & Weiss, M. 2005
Reproducibility of cerebral oxygenation measurement in neonates and infants in the clinical
setting using the NIRO 300 oximeter. Pediatr. Crit. Care Med. 6, 344–347. (doi:10.1097/01.
PCC.0000161282.69283.75)
59 Reynolds, E. O., Wyatt, J. S., Azzopardi, D., Delpy, D. T., Cady, E. B., Cope, M. & Wray, S.
1988 New non-invasive methods for assessing brain oxygenation and haemodynamics. Br. Med.
Bull. 44, 1052–1075.
60 Wyatt, J. S., Edwards, A. D., Cope, M., Delpy, D. T., McCormick, D. C., Potter, A. & Reynolds,
E. O. 1991 Response of cerebral blood volume to changes in arterial carbon dioxide tension in
preterm and term infants. Pediatr. Res. 29, 553–557. (doi:10.1203/00006450-199106010-00007)
61 Yoxall, C. W., Weindling, A. M., Dawani, N. H. & Peart, I. 1995 Measurement of cerebral
venous oxyhemoglobin saturation in children by near-infrared spectroscopy and partial jugular
venous occlusion. Pediatr. Res. 38, 319–323.
62 Cooper, C. E., Elwell, C. E., Meek, J. H., Matcher, S. J., Wyatt, J. S., Cope, M. & Delpy, D.
T. 1996 The noninvasive measurement of absolute cerebral deoxyhemoglobin concentration and
mean optical path length in the neonatal brain by second derivative near infrared spectroscopy.
Pediatr. Res. 39, 32–38. (doi:10.1203/00006450-199601000-00005)
63 Fantini, S., Franceschini, M. A., Fishkin, J. B., Barbieri, B. & Gratton, E. 1994 Quantitative
determination of the absorption spectra of chromophores in strongly scattering media: a lightemitting-diode based technique. Appl. Opt. 33, 5204–5213. (doi:10.1364/AO.33.005204)
64 Franceschini, M. A., Thaker, S., Themelis, G., Krishnamoorthy, K. K., Bortfeld, H., Diamond,
S. G., Boas, D. A., Arvin, K. & Grant, P. E. 2007 Assessment of infant brain development with
frequency-domain near-infrared spectroscopy. Pediatr. Res. 61, 546–551.
65 Leung, T. S., Tachtsidis, I., Tisdall, M. M., Pritchard, C., Smith, M. & Elwell, C. E.
2009 Estimating a modified Grubb’s exponent in healthy human brains with near infrared
spectroscopy and transcranial Doppler. Physiol. Meas. 30, 1–12. (doi:10.1088/0967-3334/
30/1/001)
66 Boas, D. A. & Payne, S. J. 2009 Comment on ‘Estimating a modified Grubb’s exponent in
healthy human brains with near infrared spectroscopy and transcranial Doppler’. Physiol. Meas.
30, L9–L11; author reply L13–L14. (doi:10.1088/0967-3334/30/10/L01)
67 Grubb Jr, R. L. & Dehner, L. P. 1974 Congenital fibrosarcoma of the thoracolumbar region. J.
Pediatr. Surg. 9, 785–786. (doi:10.1016/0022-3468(74)90120-1)
Phil. Trans. R. Soc. A (2011)
Downloaded from http://rsta.royalsocietypublishing.org/ on July 31, 2017
4424
D. A. Boas and M. A. Franceschini
68 Chen, J. J. & Pike, G. B. 2010 MRI measurement of the BOLD-specific flow-volume relationship
during hypercapnia and hypocapnia in humans. Neuroimage 53, 383–391. (doi:10.1016/j.neuro
image.2010.07.003)
69 Zhou, C. et al. 2007 Diffuse optical monitoring of blood flow and oxygenation in human breast
cancer during early stages of neoadjuvant chemotherapy. J. Biomed. Opt. 12, 051903.
(doi:10.1117/1.2798595)
70 Jakubowski, D. B., Cerussi, A. E., Bevilacqua, F., Shah, N., Hsiang, D., Butler, J. & Tromberg,
B. J. 2004 Monitoring neoadjuvant chemotherapy in breast cancer using quantitative diffuse
optical spectroscopy: a case study. J. Biomed. Opt. 9, 230–238. (doi:10.1117/1.1629681)
71 Takahashi, T., Shirane, R., Sato, S. & Yoshimoto, T. 1999 Developmental changes of cerebral
blood flow and oxygen metabolism in children. AJNR Am. J. Neuroradiol. 20, 917–922.
72 Yaseen, M. A., Srinivasan, V. J., Sakadiæ, S., Radhakrishnan, H., Gorczynska, I., Wu, W.,
Fujimoto, J. G. & Boas, D. A. 2011 Microvascular oxygen tension and flow measurements in
rodent cerebral cortex during baseline conditions and functional activation. J. Cereb. Blood
Flow Metab. 31, 1051–1063. (doi:10.1038/jcbfm.2010.227)
Phil. Trans. R. Soc. A (2011)