University of Calgary PRISM: University of Calgary's Digital Repository
Graduate Studies Legacy Theses
2000 Measurement of blood flow in bone by laser doppler imaging
Shymkiw, Roxane Chia-Chi
Shymkiw, R. C. (2000). Measurement of blood flow in bone by laser doppler imaging (Unpublished master's thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/17412 http://hdl.handle.net/1880/40506 master thesis
University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca THE UNIVERSITY OF CALGARY
Measurement of Blood Flow in Bone by Laser Doppler Imaging
by
Roxane Chia-Chi Shymkiw
.A THESIS
SUBMITTED TO THE FACULTY OF GRADUATE STUDIES
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE BIOMEDICAL ENGWEERING
DEPARTMENT OF MECHANICAL AND MANUFACTURING ENGJNEERlNG
CALGARY, ALBERTA
MAY, 2000
O Roxane Chia-Chi S hymkiw 2000 National Library Bibliotheque nationale (*Iof Canada du Canada Acquisitions and Acquisitions et Bibliographic Services services bibliographiques 395 Wellington Street 395, rue Wellington OttawaON K1AON4 Ottawa ON K1A ON4 Canada Canada Your file Vorre reference
Our fi& Norre reterence
The author has granted a non- L'auteur a accorde une licence non exclusive licence allowing the exclusive pennettant a la National Library of Canada to Bibliotheque nationale du Canada de reproduce, loan, distribute or sell reproduire, preter, distribuer ou copies of thls thesis in microform. vendre des copies de cette these sous paper or electronic formats. la forme de microfiche/filrn, de reproduction sur papier ou sur format electronique.
The author retains ownership of the L'auteur conserve la propriete du copyright in this thesis. Neither the droit d'auteur qui protege cette these. thesis nor substantial extracts &om it Ni la these ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent Ctre imprimes reproduced without the author's ou autrement reproduits sans son permission. autorisation. ABSTRACT
Although the mechanisms are not clearly explained. blood flow may play an
important role in mediating skeletal adaptation. Most techniques currently available to measure blood flow in bone are time-consuming and require destruction of the tissue whereas laser Doppler technology offers a less invasive method. The current series of
studies investigated the effectiveness of laser Doppler perfusion imaging (LDI) for
measuring perksion in cortical bone. Preliminary pilot studies were conducted to
determine the feasibility of using LDI on bone, as well as assessment of bone optical
properties and flow characteristics.
Results indicated that LDI effectively measured blood flow in cortical bone and
detected physiologically induced changes in perfusion. A significant positive correlation
was found between rnicrosphere-determined flow and LDL output (r = 0.58. p < 0.05).
Repeatability of consecutive LDI perfusion measurements was within 5%. Red and near
infrared wavelengths were significantly correlated suggesting the two wavelengths
measured pefision in the same region. Red and near infrared lasers were measuring
blood flow up to 0.8 mm and 0.9 mm. respectively, into the tissue.
This study characterized LDt for measuring perhsion in bone and correlated tlow
measured by LD t to standardized microsphere-determined flow. The ability to profile
heterogeneous structure and sequential changes in blood flow in bone suggests this
method has the potential for investigating the role of blood flow in bone metabolism.
healing, and remodeling. To my dad,
whose footsteps I've always wanted to follow
and to my best friend. and husband.
Dave
My many thanks to:
All that assisted in completion of this work: Catherine Leonard, Brett McGuinness, Craig Sutherland, and Greg Wohl
Kevin Forrester, for his overwhelming contribution of time and knowledge
Ron Zernicke, for his guidance and opening the door to many opportunities over the past Years
Those who gave me strength through their friendship: Barb Kralovic, Christine Maurette, Tasha Reid (Itett), Maggie Erickson. Connie Sullivan, Chantelle Carley, Todd Leask, Carla Sciarretta, Gail Thornton, Shawna Yanke, Leonard Pianalto, Greg Wohl, Greg Kawchuk, and Brett McGuinness
And most of all, to my parents and family (Jocelyn, Andrew, Charlie, Alma, and Frank) for their understanding, encouragement and support. TABLE OF CONTENTS
.. Approval Page ...... II ... Abstract ...... III Dedication and Acknowledgments ...... iv
Table of Contents ...... v . . List of Figures ...... vti
List of Tables ...... x
List of Abbreviations ...... si
CHAPTER I INTRODUCTION ...... 1
CHAPTER 2 BACKGROUND ...... -I
Anatomy, Physiology and Pathology of Blood Flow in Bone ...... -+
Laser Doppler Principle ...... 16
2.2.1 Laser Speckle Imaging ...... IS
2.2.7 Laser Doppler Flowmetry ...... 19
3.2.3 LaserDopplerPerfbsion Irnagins ...... $0
Tissue Optics ...... 20
Current Measurement Techniques ...... 35
Study Objectives. Rationale, and Scope ...... 37
CHAPTER 3 MATERIALS AND METHODS ...... -I2
3 .O Laser Doppler Pefision Imaging Standard Protocol ...... 42
3.1 Laser Doppler PeAsion Imaging Pilot Studies...... -43
3.1.1 Rooster Model...... -43
3.1.2 Rabbit Model ...... -44
3.2 Bone Optical Properties Study ...... -IS
3.2.1 Thin Section Transmission ...... -45
3 2.2 Thick Section Transmission ...... -47 3.3 Flow Model Study ...... A8 3 .3.1 Detectable Range of Flow ...... 48 3.3.2 Specular Reflectance ...... 50
Laser Doppler Perfusion Imaging Study ...... 51
3.4.1 Animal Preparation ...... -51
3.4.2 Laser Doppler Perfusion Imaging ...... 52 -- 3 .4.3 Evaluation of Blood Flow ...... 33
3.4.4 Blood Flow Determination ...... 54 3 4.5 Statistical Methods ...... 56
RESULTS ...... 57
Laser Doppler Pedbsion Imaging Pilot Studies ...... 57
4.1. I Rooster Model ...... 57
4.1.2 Rabbit Model ...... -58
Bone Optical Properties Experiments ...... I9
4.2.1 Thin Section Transmission ...... 59
4 Thick Section Transmission ...... 61
Flow Model Study ...... 61
4.31 Detectable Range of Flow ...... 61
4.3.2 S pecular Reflectance ...... 62 Laser Doppler Pehsion Imaging Study ...... 6-1
4.41 Repeatability ...... 67
4.4.2 Red versus Near Infrared Wavelength ...... 68 4.4.3 Comparison of Laser Doppler Pehsion Irnay ing and
Coloured Microspheres ...... 70
CHAPTER 5 DISCUSSION ...... 73
CHAPTER 6 CONCLUSION ...... 94 6.1 Future Work ...... 95
REFERENCES LIST OF FIGUWS
Figure 2.1 Anatomy of long bone blood supply ...... 5
Figure 2.2 Blood supply of long bone ...... -6
Figure 2.3 Experimental arrangement of laser speckle imaging ...... 17
Figure 2.4 Experimental arrangement of laser Doppler flowmetry ...... IS
Figure 2.5 Experimental arrangement of laser Doppler perfbsion imaging . 19
Figure 2.6 Interaction of laser light on tissue ...... -21
Figure 2.7 Laser-tissue interaction according to Snell's law ...... --73
Figure 2.8 Mean free path of laser light in soft noncoloured tissue as a ...... 26
function of wavelength .
Figure 29 Sequence of esperimental studies involved in the analysis of .....39
the utility of laser Doppler pehsion imaging in bone .
Figure 3.1 Laser Doppler pefision imaging experimental setup ...... -I3
Figure 3.2 Apparatus for thin section transmission experiments ...... -I5
Figure 3 3 Schematic for the thick section transmission experiments...... 47
Figure 3.4 Schematic for the flow model used to determine the range of flow ...19
detectable by the laser Doppler perfbsion imager .
Figure 3.5 Schematic of the flow model used to determine the range of ...... 50
detectable flow by LDI in bone .
Figure 3.6 Schematic of the flow model used to study the effect of ...... -51
specular reflection on LDI output .
Figure 4.1 Mean (+ standard deviation) flux (perfksion units) of the ...... 57
mid-diaphyseal cortex of rooster tarsometatarsus. Figure 3.2 Comparison of perfusion values between pre- and post- ...... 58
application of Ach or Xdr (% change f standard deviation).
Change in pefision was not significantly different 60m normal
for either Ach or Adr.
Figure 4.3 Mean (+ standard deviation) tlux of the mid-diaphyseal cortex of ...... 59
rabbit tibiae. Occlusion pehsion values were significantly lower
than those for control or recovery (p < 0.05; denoted by an asterisk *).
Figure 1.4 Relation of light transmitted through thin sections of cortical .. ..60
bone at different thicknesses for near infrared and red wavelengths.
Figure 4 5 Relation of red and near infrared light transmitted through thick ...... 61
cortical bone samples at varying sample thicknesses.
Figure 4.6 Comparison of known flow rates (pl/min) to laser Doppler ...... , .62
pehsion output (perfusion units).
Figure 4.7 Comparison of LDI output (pefision units), with lobrabbit . 63
blood in a heparanized saline solution. in the flow model laying
normal to the laser source and on an incline.
Figure 4.8 Comparison of LDI flux and DC output for 1% blood in ...... -.. . -6-4
heparanized saline at known flow rates when the model was (a)
normal to the laser source and (b) at a 7" incline.
Figure 4.9 Sample LDI images at (left to right): normal red. normal ...... 65
infrared (NIR),post-drug infusion (NR), and post-mortem (NIR). Figure 4.10 Comparison of LDI image to bi viva bone sample. Tissue ...... 66
heterogeneity and surface blood vessels were detected by LDI as
depicted in the image.
Figure 4.1 I Relation of NIR versus red wavelength mean output. LDI output ...... 69
of the two wavelengths were significantly correlated
(r = 0.93; p < 0.05).
Figure 4. I?. Relation of NIR versus red wavelength standard deviation output...... 70
LDI output of the two wavelengths were significantly correlated
(r = 0.94; p < 0.05).
Figure 4. I3 Relation between mean LDI output (pehsion units) and . . . . , . . . . 72
standardized blood flow (measured with coloured microsp heres).
A significant correlation between LDI and CM measured Row
was found (r = 0.69; p < 0.05).
Figure 5.1 Laser-tissue interaction at different wavelengths...... 79 LIST OF TABLES
Table 7.1 Previous studies of bone blood flow by the microsphere method...... 32
Bone flow rates in ml/min/100 g.
Table 4.1 Consecutive laser Doppler pehsion imaging measurements at . . . , . .65
near infrared wavelength.
Table 4.2 Concurrent tibia1 blood flow measurements with microspheres ...... 71
and laser Dopplei pefision imaging at near infrared wavelength. LIST OF ABBREVMTIONS
Ach Acetylcholine
ACL Anterior cruciate ligament
Adr Adrenaline
CM Coloured microsp heres
DC Direct current
LD Inner diameter
LDF Laser Doppler tlowmetry
LDI Laser Doppler perfusion imaging
LEAD Lower extremity arterial disease bIFP hiean free path
NIR Near infrared
OD Outer diameter
PU Pehsion unit
TMT Tarsometatarsus INTRODUCTION
Bone is a highly complex tissue and to understand the behavior of bone one must examine interrelated mechanical and physiological responses. Bone can adapt to altered loading conditions. such as functional overloads (76) and disuse (39), and bone modeling and remodeling are related to strain magnitude (96). circumferential strain gradients (10). and strain rate (86). Mechanical integrity of tissue may be a major contributor to joint adaptations (5). but vascular alterations may also intluence changes in joint properties
(11 ). Although questions remain as to whether tluid flow from within blood vessels (74) or the microporosity of the bone matrix is responsible for bone adaptation (22.48.75). fluid tlow in bone has been linked to skeletal adaptation.
The effect of the vascular system on tile mechanical properties of bone remains relatively unknown. but blood tlow has been susgested to play a role in the maintenance of bone. not only in the delivery of nutrients and removal of waste products of tissue metabolism. but also in adaptive responses to different states of stress (74). .As changes in the vascularity of soft connective tissues may influence the mechanical properties of these tissues (26),it is possible that vascularity may lead to altered behavior of bone.
Changes in bone blood flow are possible mechanisms influencing bone remodeling
(45.74) Alteration in perfbsion intluences mineral dynamics, deposition. and resorption of bone. and. therefore, an increase in blood flow may stimulate osteoclastic activity or increase mineral deposition (22).
Vascular alterations in connective tissue after joint injury have also been linked to the onset and progression of osteoarthritis and rheumatoid arthritis (21). The combination of angiogenesis and an increased neovascularization (42), coupled with a compensatory remodeling. may in pan explain the mechanical adaptations of the injured joint. The blood supply to bone also appears to be an important factor in bone infection. loosening around artificial joint implants, non-union, Fracture healing. osteonecrosis (bone death), and osteomyelitis (infection).
Direct measurement of cancellous and cortical bone blood tlow remains a challenge due to the complex arrangement of arterial and venous vessels (I 10). Several techniques. including radioactive clearance and entrapment of microspheres. have been used to determine blood perfusion in bone. but the techniques require destruction of the tissue and are not well suited for dynamically monitoring changes in pefision. In contrast. repeatable quantitative measurements of regional blood flow in bone offer
important advantages in studyins normal and pathological conditions of bone.
Laser Doppler technologies ofir a minimally invasive method of measuriny
pehsion in soft and hard tissues. This technique has focused on laser Doppler
tlowmeters in which a fibre optic probe was placed in contact with a discrete location on or within the tissue surface. This method. however, requires contact between the probe and tissue. which may induce a motion anifact in the Doppler signal or change the underlying physiolosy. Because of the heterogeneity of small vessel distribution in vascularized tissue, measurements made by laser Doppler flowmeters are site specific.
These shortcomings have been countered by laser Doppler perfitsion imaging
(LDI), which generates multiple measurements sites by scanning a laser beam over the surface of a tissue. The measurements can produce an average perfUsion value for the
tissue, minimizing the effects of microvascular heterogeneity. The ability of LD1 to map
areal pefision suggests that it may be used in regions undergoing vascular adaptations in response to physiological stimuli. LDI has been validated for hz vivo measurement of blood flow in soft tissues, showing a high correlation with microsphere-determined blood flow (6). LDI has yet to be validated as a tool for measuring bone blood flow, but its similarities to laser Doppler flowmetry and use in measuring pehsion in soft tissues suggest it may be an applicable technique in bone.
Relevonce ufSt14[lies
.A1 terat ions in regional blood tlow occur due to growth, fracture, healing, disuse.
and aging. Tissue remodeling and alterations in joint mechanical behavior may be
mediated by vascular alterations in bone following joint injury. Development of techniques to quantify and interpret changes in blood flow in bone may clarify its role in
bone remodeling and thus provide insight into the mechanisms underlyin~bone
adaptation. By understanding the physiological mechanisms and biological factors behind bone remodeling, knowledge of the role of blood flow may help predict the physical
behavior of healing tissue. enhance normal response. or optimize tissue healing. and
provide insight into the prevention of bone loss due to aging, disease. or disuse. BACKGROUND
2.1 Anrtonly, Physiology and Pathology of the Blood Supply of Long Bone
Blood
Blood is not a simple Newtonian tluid; it differs from Newtonian fluids in that its viscosity decreases with increasing shear rates. In large vessels. blood viscosity is independent of vessel size but the apparent viscosity decreases in vessels less than I mm in diameter and increases in vessels smaller than 6 ILm. The evhrocytes in tlowing blood are not evenly dispersed and tend to be concentrated in the axial regions of highest velocity. The cells are more dispersed in concentric lamina closer to the vessel wall leaving only a thin plasma film in immediate contact with the wall. As tluid flows through a tube the velocity remains greatest at the axis of the tube and slowest at the walls. The velocity profile has a parabolic form and tlow occurs in a series of concentric rings of decreasing velocity from the central axis to the wall. The characteristics of the protile depend on the radius of the tube and velocity and viscosity of the fluid (3).
Sluotl Fio~virt Bortc.
Bone is composed of hydroxyapatite, collagen, water. and carbohydrate. Long bones are comprised of two types of bone: cortical (compact) and cancellous (Figure 2.1 ).
Cortical bone constitutes hur-fifths of the mammalian skeleton. and cancellous bone constitutes the remaining one fifth (57). Bone composition and anatomical location are important determinants of mean blood flow. In long bone, the lowest flow rates are in the periosteal and endosteal cortices. and the highest flow rates are in the metaphyseal/ epiphyseal cancellous bone (37). In a mature skeleton, cortical bone (femoral diaphysis) has very low levels of blood flow (approximately 2 ml per min per 100g) whereas epiphyseal and metaphyseal cancellous bone has flow 10 to 15 times greater
(approximately 18 to 30 ml per min per 100g) (37). Flow in the surface 2 rnm layer of epiphyseal cancellous bone can be 3 to 10 times greater than the remaining deeper bone in the femur and tibia (8 1). Bone cells in cortical bone are connected to each other through a system of canaliculi, which are connected to Haversian canals. The canalicular system allows bone cells to connect with a blood supply. The trabeculae in cancellous bone of less than 0.2 mm thickness are thought to have blood vessel canals located near each centre (4 1) (Figure 2.1).
Cancellous bone 0 T robeculo
Lamella
Figure 2.1 Anatomy of long bone blood supply. Adapted fiom (57).
The circulation of blood in bone is necessary for osteogenesis, maintenance of bone vitality, bone growth, and repair of fracture and other injuries (100). It is generally agreed that long bone is supplied blood from three sources: vessels at the ends of bone
(epiphyseo-metaphyseal), one or two main arteries entering the diaphysis (nutrient arteries), and periosteal vessels (88) (Figure 2.2). Their individual contributions, however, remain in question. The six groups of arteries that supply long bone include: proximal epiphyseal, proximal metaphyseal, diaphyseal nutrient, distal metaphyseal, distal epiphyseal, and periosteal arteries. All branches of arteries are profhsely anatornosed such that adjacent groups can substitute for each other and protect against infarction caused by failure of another system (3 5,89).
_... - ...... ,.;.: . Figure 2.2 Blood supply of long bone. Adapted from (18).
The nutrient artery provides an arterial supply to the diaphyseal cortex (-10%) and marrow (-90%). After entering the diaphysis, the nutrient artery divides into ascending and descending branches. The lateral arterial branches extend radially outwards from the endosteal surface towards the diaphyseal cortex, supplying the vessels of the Haversian canals that supply the cortex. Blood flows in a centrifbgal pattern entering the endosteal surface of the cortex from the nutrient medullary system and exiting through the periosteal surface (57). The terminal branches of the main ascending and descending branches of the nutrient artery supply the ends of long bones.
Vessels enter the epiphysis and metaphysis through small foramina. The arteries branch into the Hunter's "vascular circle" that penetrate the epiphyseal-metaphyseal area where the vessels are arranged likes spoke of a wheel (Figure 2.1). After enterins bone arteries. the vessels branch into capillary arcades. which loop beneath the articular canilage. Blood tlows from the outside to the inside and this system anastomoses with the nutrient artery system (57).Eighty percent of remodeling activity in mature bone occurs in cancellous bone. and. there fore. the vascular arrangement in cancellous bone is imponant. The spaces in cancellous bone are typically filled with marrow. Each individual strut is in contact with a rich vascular supply that plays an important role in calcium homeostasis.
Venous drainage from a long bone occurs at the ends of the bone. Small venules travel in the cortex and drain both periosteally and endosteally. Large veins also connect the central medullary sinus and periosteal venous system (57).
The periosteum, a fibroelastic membrane covering the surface of bone. is a major
contributor to the vascularity of bone. Although the cortex is the main beneficiary of the
anatomical and physiological hnct ions of the membrane. the behavior of the entire bone
is influenced by periosteal activity (14). The periosteal membrane plays a role in maintenance of bone shape, metabolic ion exchange, and physiologic distribution of electro-chemical potential differences across its structure ( 14). Periosteal veins contribute to venous drainage and represent the site where the prelyrnphatic system of bone discharges proteins and other substances from the interstitial fluid system (57). The periosteum's specific contribution to the cortical blood supply has been suggested to be
20% peripheral medullary arterial blood supplying up to the outer third of the cortex at places of facial attachments (93).
The endothelial tubes in bone originating from the arterioles of the periosteum are thought to be an important source when the nutrient medullary blood supply is damaged or suppressed. Researchers have demonstrated the ability of the periosteal circulation to act as a compensatory blood supply during disrupted blood flow by maintaining vascularity of the outer cortex and following reversal of the centrifugal direction. to revascularize the cortex when the medullary supply is compromised. it appears there is an arterial network similar to the nutrient artery network present in the layers of the periosteum that is continuous with the intracortical circulations (88). By microarteriographs. Oni and Gregg demonstrated that periosteal afferent vessels can penetrate the cortex and marrow cavity centripetally from the outer surfaces following creation of a diaphyseal segment. New subperiosteal bone formation and cartilage were observed around the circumference of the shaft at the location of the cut (88). With only a nutrient artery supply, the cortical vessels in the segment were not pehsed and no new bone formation was distinguishable suggesting periosteal circulation may be essential for normal functioning medullary circulation (92.93). Athough Oni and Gregg did not observe revascularization of the diaphysis when the epiphyseo-metaphyseal supply was isolated (88). others have demonstrated cortical and medullary revascularization via longitudinal cortical vascular channels (87).
Damage to the nutrient artery may not affect the overall blood supply of bone as the periosteal and medullary circulations can potentially supply blood to the diaphysis
(58). Following interruption ofthe periosteal supply the afferent vessels tiom the nutrient artery circulation were able to penetrate the cortex (58). The periosteal supply was rapidly restored after the damage, but it was not clear whether the increase in perfusion was due to new vessel formation (92.93) or increased perhsibility in pre-existing vessels
( 128). The periosteal and medullary circulations were also potentially able to supply blood to the diaphysis of the cortex (58).
The circulation of blood through bone sustains functions such as mineral metabolism, hernatopoiesis. and n~obilizationof bone cells in bone and bone marrow
Blood circulation is also required for the maintenance of bone vitality. osteogenesis. and bone growth and repair. Ions and molecules either pass through channels or cletls at junctions between endothelial cells to reach bone cells or by diffusion across endothelial cells. The movement of ions in blood from the capillary to matrix is thought to occur by diffusion and emux controlled by a concentration dependent mechanism in the extravascular space (78). Fluid fluxes in the interstitial fluid in bone may produce streaming potentials. which stimulate osteoblasts to form new bone (78).
Bloocl Flow A ~lturegulntiorl
Re~wlationof blood flow is an important mechanism in controlling the rate of intake and release of nutrients and waste products in tissues. Circulation of blood in bone may be regulated in part by neural, humeral, and metabolic mechanisms. Bone and marrow are highly innervated and were responsive to direct and reflex nerve stimulation. The finding of unmyelinated C fibres in the nerve supply to blood vessels in bone and marrow suggested an autonomic function for these nerves (25).
Previous studies indicated that bone and marrow vessels were influenced by neurogenic
Factors. Stimulation of sympathetic nerve fibres produced vasoconstriction in bone (24) and decreased intramedullary blood supply and bone blood outtlow. while sectioning of the sciatic nerve (sympathetic fibres) in the hindlirnb of rabbits increased flow to the tibia
( I0I). Perhion studies have also shown that there is a highly reactive vasomotor control mechanism in bone (23). The primary neurovascular receptors in bone may be a[-
adrenergic receptors. and ischemia may produce a functional loss of perianerial
sympathetic nerves in addit ion to increased a,-and uz-adrenergic receptor sensitivities to
circulating catecholamines ( 16). Blood outflow from the tibia decreased when nerve
fibres to the bone were electrically stimulated (24). Electrical stimulation of nerves also
resulted in a change in intramedullary blood pressure that was correlated to hemodynamic changes in bone.
Drinker and Drinker noted a decrease in bone blood outflow following the addition of adrenaline to pefising blood (24). Considerable evidence for humeral control
has since been accumulated. Alteration in vascular resistance in response to vasoactive
drugs may have been a direct egect of the respective agent on vascular smooth muscle as.
the response of bone vessels was similar in magnitude to the response in skeletal muscle
(37). Bone vessels have the capacity for alteration in vascular smooth muscle tone by
humoral stimuli (37). Shim reported quantitative and qualitative studies indicating
vasopressor hormones caused vasoconstriction of blood vessels in bone, and because the hormones tested occur naturally in the body, these hormones likely influence bone circulation normally (100). Several studies have indicated that vasopressor peptidrs and catecholamines, such as adrenaline and noradrenaline, increased vascular resistance and caused vasoconstriction of bone Mood vessels and reduced bone blood tlow. Introduction of adenosine to the circulatory system resulted in dilatation of bone and marrow vessels. indicating dilator capacity and bone and marrow vasomotor responsiveness (37). Neural and hurneral regulatory mec hanisrns were likely linked since noradrenaline was produced by the sympathetic nerve endings and adrenaline was produced by the neural glands
( 100).
Noradrenaline is a vasopressor drug such that despite an increase in arterial blood pressure. it decreases bone blood flow by constricting bone vessels (37). The vasoconstrictor effect of noradrenaline was lar~eenouph to mask the increase in blood tlow during increased petfbsion pressure (3 7). Nitroprusside caused arteriolar vasodilatation as well as venodilatation. Nitroprusside caused vascular smooth muscle relaxation via the liberation of nitric oxide. which in turn, increased coronary tlow and decreased coronary perhsion in the rabbit hean (47). Total coronary resistance and arterial pressure were also significantly decreased. Nitroprusside primarily increased the capacity of the systemic vascular bed allowing blood to pool peripherally and cardiac volume to decrease (103).
Evidence suggested that bone blood flow was affected by metabolic factors such
as acid metabolites, pH, and oxygen and carbon dioxide concentrations in the blood at the
local and systemic levels (100). Bone exhibited a reactive hyperemia following release of
the femoral artery after a period of occlusion, providing evidence for the metabolic control of bone blood flow (101). The increase in outflow volume after occlusion persisted after transection of the nerves. electrical stimulation, and administration of vasopressor exogenous hormones. Analysis of blood from reactive hyperemic bone showed an increase in H', bicarbonate, and carbon dioxide concentrations and a decrease in oxygen content. Increased carbon dioxide and decreased oxygen levels in blood increased blood outtlow in long bones. Variations in pCOz and pH were suggested as factors capable of altering calcification that could lead to changes in the material propenies of bone (94). Bone exhibited vasomotor responsiveness; bone blood vessels constricted despite dilatation in surrounding skeletal muscle during exercise indicating its participation in the redistribution of blood tlow during physiological stresses. ReHex stimuli. produced by alterations in arterial baroreceptor activity also dilated and constricted bone vessels when stimulated or denervated, respectively (37).
Increased venous pressure was associated with increased periosteal bone formation. but subtle changes in blood tlow. bone oxygen saturation. or partial pressure of oxygen were effects and not factors that stimulated the new bone formation (57).
Preliminary studies suggested that increased venous pressure will increase tluid tlux from the capillary to the tluid flow channels in the bone matrix (56.57).The increase in pressure was associated with an increase in the extracellular space of the tibia, which implied a tluid flux through the capillary pores or clefts to the extravascular fluid compartment of the bone matrix. This was also accompanied by a decrease in the vascular spaces (lo,%). The vascular space, however, increased by 3 wk likely due to a rise in blood flow to meet the metabolic demands of new bone formation (54). The tlux of fluid may have increased streaming potentials in bone that act as a si~nalto bone cells to increase bone formation (56).
Patlr ology
The pathology of bone blood supply is important in understanding the response of bone to aging. disease. and physiological stress. It was suggested that necrosis results from a vascular insutiiciency that occurs in one third of the cases of femoral neck
Fracture (4). A positive correlation. found between the mineralization of lacunae and ischemia. implied there was a relation between ischemia and an increased number of dead cells (62). Increasing age was also related to an increase in the number of empty lacuanae. Blood flow rates in the femoral and metatarsal compact bones were lower in 20 wk old compared to I0 wk old pigs (8 I),which suggested the decrease with age was due to higher rates of bone remodeling and appositional bone growth in younger animals (SO).
Evidence exists for an association between osteopenia and cardiovascular disease.
Severe symptomatic lower extremity arterial disease (LEAD) was accompanied by decreased bone mineral density in the affected limb ( 13). The bony tissue proximal to the arterial stenosis showed loss of osteocytes and infarction of the marrow or bone cortex. ln a study of men with LEAD. the bone mineral content of the femoral neck on the atfected side was 3.3% lower than the healthy side and attributed to the effects of ischemic atherosclerotic disease (66). Vogt and colleagues gave evidence from a communitv study that a decrease in blood flow to the lower extremities may be associated with an increase in the annual rate of bone loss at the hip and calcaneus ( 1 18). Patients in the early stages of oaeoarthritis had higher marrow pressures and impeded venous drainage compared ro normal controls (2). It was hypothesized that the decreased blood flow was an early event in the development of osteoarthritis. A direct relation was found between intraosseous
pressure. decreased blood flow. and decreased oxygen availability, suggesting that early
osteoathritic changes in bone were related to circulatory impairment (59).
iLlcInnis and colleagues investigated a correlation between regional blood tlow
and the amount of endosteal bone formation in a bilateral tibia1 defect model in the
canine. Using I-Ap ("I-labeled 4-iodoantipyrine) washout. it was suggested that
endosteal new bone was directly related to flow (77). The study implied that the increase
in blood tlow in new bone was secondary to the increased metabolic demands. The early
phase of uptake of bonr seeking isotopes was a result of the increased vascularity. .A
possible mechanism for the increased mineral deposition following increased blood tlow
was the increase in surface area available for ditfusion due to more capillaries (77).
Healing bone exhibited changes in blood tlow from normal bonr however injury
to the so%tissues in a joint may also alter vascularity in bone. Shymkiw and colleagues
showed a significant correlation between increased blood tlow and decreased bone
mineral density in the perianicular cancellous bone in the early stages following joint
ligament injury ( 107). The peak in blood flow in the condyles at 3 wk post-ACL
transection iikel y occurred to meet the increase in metabolic demands of bone following
injury. whereas the increased flow at later time periods may have been related to chronic joint instability. The increased tlow may have been a combined effect of increased tlow
though existing channels and flow through new channels created by angiogenesis (53).
Both have been associated with tissue remodeling (1 L.27). Heightened blood flow may
mediate alterations in mechanical bone propenies and adaptation in the early stages aHer
injury to the anterior cruciate ligament. Following fracture of bone, an initial decrease in blood flow was followed by a substantial increase (90) driven by the increased metabolic demand necessary for the formation of new bone and interstitial and surface remodeling. Investigators suggested that the blood supply from the periosteum and soft tissues was more important than from the marrow in the early peak period of healing long bone fractures ( 1 15). Flow to the cortices proximal and distal to the fracture site was significantly higher than in the contralateral limb, but tlow to the marrow was not. Thus. the increase in cortical tlow of the injured limb was mediated by an arterial supply parallel to the marrow in either the periosteum or soft tissue ( 1 15).
Controversy remains on whether exercise increases. decreases. or does not affect blood flow in bone and marrow. Similar to Gross and colleagues (37). Spodaryk and
Dabrowski found there was no chanse in bone marrow blood tlow following periods of exercise ( 105). There was marked vasodilatation in adjacent skeletal muscle in contrast to vasoconstriction and a two-fold increase in vascular resistance in bone (37). follow in^ exercise there was a decrease in blood tlow to the diaphyseal marrow resion due to increased vascular resistance and increased tlow to metaphyseal regions due to increased external pressure exerted on the metaphyses. There was. however. no overall change to the blood flow in bone (105). Although perfbsion increased after one hour of exercise. a significant increase (50%) in cortical and cancellous bone perhion was only detected after two hours of moderate exercise (1 13). The increase in perfision rate was maintained during the post-exercise resting period (I h) possibly due to repayment or redistribution of blood in the limbs after exercise (1 13). Continuous strenuous exercise significantly increased intramedullary pressures resulting in the bone tissue suffering from isc hernia and cell-stress-initiated remodeling which ultimately weakened the bone (89). Static loading of bone resulted in decreased blood flow on the compressive side and increased flow on the tensile side. Bone. however, was only deposited after intermittent loads (73).
Alteration in bone blood flow suggests that bone has the capability of vasodilatation during muscular work. Since the pefision response following exercise was slow. vasodilatation may be mediated by a metabolically induced stimulus ( 1 13).
2.2 Laser Doppler Principle
The laser Doppler principle is based on the Frequency shift of scattered incident light (106). A collimated beam of monochromatic laser light is directed on tissue in a raster pattern. and individual photons are either scattered outward or absorbed. Photons that interact with stationary tissue undergo multiple scatterings and keep their original frequency. The frequency of photons that interact with moving blood cells undergo a
Doppler shift. The number of frequency shifts is related to blood cell concentration. The constructive and destructive interference of the shifted and non-shifted light. on the p hotodiode, creates a "beat" Frequency. The retlected shifted and non-shifted lights generate a Doppler beat signal in the photodiode output current and from spectral analysis of the beat signal. blood flow level in the tissue was estimated (8). The laser
Doppler flow signal is proportional to the product of the number of erythrocytes moving within the illuminated volume and the average velocity of the cells ( 104). This signal is independent of flow direction and is recorded as a perfbsion unit. The protile of reflected light is dependent on tissue structure and laser wavelength. The perfusion indices are normalized to the total reflected light. Instruments utilizing the laser Doppler principle can compensate for changes in light intensity from instabilities in laser output but cannot normalize for comparison in different tissues or locations. The sampling depth of the laser is dependent on the optical properties of the tissue based on the wavelength of the
v source, the position of the source, and the sink of photons on the tissue surface (82). The exponential decay of light intensity with tissue penetration defines the sampling volume of laser Doppler measurement. This technique has been used to monitor pefision in bone in patients since 1985 (71).
Currently, three methods exist based on the Doppler shifting of light to measure relative blood flow in tissues: laser Doppler flowmetry, laser Doppler imaging, and laser speckle imaging. The principle behind laser Doppler technologies and laser speckle are effectively identical. The Doppler principle involves the superposition of two waves of slightly different frequencies and detection of the resulting beat frequency. Laser speckle is based on interference where two waves of the same frequency are superimposed, and the correlation is detected as one wave slides past the other.
Figure 2.3 Experimental arrangement of laser speckle imaging. Adapted from (33). 2.3.1 Laserspeckle Imaging
The laser speckle method is a new form of flowmetry to analyze the interference pattern that appears when tissue is illuminated with a laser beam (Figure 2.3). The
Doppler shifted and non-shifted reflected lights were mixed and a random intensity distribution (speckle) pattern was generated. The speckle field was detected by a CCD image sensor and blood flow levels were expressed by the averaged difference between pairs of output data for successive scans (32). The structure of the pattern depended on blood flow and the rate of variation of the pattern varied with flow velocity (34). Areas with high levels of blood flow had large variation in the speckle partem, whereas areas with low levels of flow had minimal change in the pattern. Fujii and colleagues developed a system to visualize the 2-dimensional distribution of tissue microcirculation by electro-optical analysis of the speckle panem (32).
Oiffcrczct tj~i~nal1 - cut;ut amplifier prcccsscr sisnzl
indicatar dc!ectar dclcctcr indcater znd filter and filter
Optical fiovs Q4&
Figure 2.4 Experimental arrangement of laser Doppler flowmetry. Adapted from
(1 10). 2.2.2 Laser Doppler FZo~vmetry
A standard laser Doppler flowmetry (LDF) system includes a metal sheath probe containing one efferent (transmits light to the tissue) and two afferent (transmits light back to the detector) optical fibres (Figure 2.4). To take measurements, the probe must be in contact with the tissue, which introduces the risk of infection or pain. The probe must also be held constant as slight changes in probe position result in unusually high perfhion measurements (49). LDF is capable of constantly measuring perfusion at a single point, but because of spatial variations in tissue blood flow, this method is better suited for stimuli response experiments than mapping blood distribution in a given area
(5 1). +
Photodiodes
Computer
*
Figure 2.5 Experimental arrangement of laser Doppler perfusion imaging. Adapted
&om (79). 2.2.3 Laser Doppler Perfusiorz Imaging
In a laser Doppler pefision imaging (LDI) system, the laser light source and
photodetector are positioned in a scanner head. which can be fixed at a user-determined
height above the tissue. The optical system involves a collimated laser beam. scanning
mirror. and photodiode (Figure 1.5). The scanning mirror directs the laser beam
successively across the tissue surface in an area as larse as 12 x I2 cm. At each
measurement point. the backscattered and Doppler-shifted light are sampled and a
corresponding blood pefision value is calculated (5 1 ). Each pixel. in the 2D image.
generated as the laser scans over the tissue is representative of a grid point in the scan. An
image is senerated from the perfusion values showing the spatial distribution of tissue
pehsion and displayed on a computer monitor ( l 19).
2.3 Tissue Optics
The understanding of laser light propagation is important in medical applications.
as optical measurements can be used to monitor structural. physiological. and metabolic
changes in tissue. The rapidly increasin~use of light in diagnostic and therapeutic
medicine has created the need for determining the optical properties of living tissue. The
measurement of biological media influences exerted upon light as it travels through the
media is key to several noninvasive measurements of physiological variables such as
blood oxygenation. microcirculation, and tissue metabolism ( 19). Knowledge of tissue
optical properties, such as photon penetration depth, will permit interpretation of the
fraction of tissue microcirculation being sampled (3 1). Optical parameters of tissue are
obtained by converting direct measurable quantities into parameters that characterize
light propagation in tissue (1 5). Laser light specular reflection I I I I I I
.ssue surface
internal tissue reflection Y A absorbed moving blood cell
Figure 1.6 Interaction of laser light on tissue.
The characteristics of photon propagation in tissue such as scattering and absorption events. and retlection and transmission at the boundaries govern the number of photons that will reach the target chrornophore (Figure 2.6). Photons incident upon tissue will either be reflected or transmitted at the air-tissue boundary and scattered or absorbed within the tissue. Photons within the tissue will then be remitted or internally retlected at the tissue-air boundary. Photon path is affected by physical properties such as scattering and absorption coefficients and anisotropy of the medium.
Theoretical study suzgests that internal optically differentiable structures can be assessed by photon flux measurements on the surface of the medium. The spatial distributions of laser irradiated tissue are defined by its absorption and scattering components. The amount of light absorbed is determined by endogenous (naturally occurring molecules, such as water or pigment of hemoglobin) and exogenous (lipht absorbing molecules added to the tissue, such as photosensitizing dyes) c hromop hores, which are tissue components that absorb light. Chrornophores are highly absorptive to wavelengths less than 600 nm and greater than 1300 nm. Inside of this range absorption of light is at a minimum, and scattering dominates light propagation. At shorter wavelengths. absorption in tissue is high due to natural chrornophores (e.g.. heme pigment in hemoglobin. melanin pigment. and pigments of the respiratory chain in mitochondria). and in the ultraviolet region. absorption is due to proteins and nucleic acids. At longer wavelengths. in the infrared spectrum. absorption by the water in tissue is high (124). Oxygenation also affects light absorption where an increase in oxygenation results in additional absorption at 880 nm and less absorption at 660 nm ( 19). Scattering in tissues is due to discontinuities in refractive indices on the microscopic level (e.2.. collagen fibril within the extracellular matrix) (124). It also occurs at the boundaries of tissues with differing refractive indices. The optical propenies of tissue are defined by corresponding coeficients; absorption and scatterins coefficients are defined as the probabilities per unit path length that a photon will be absorbed or scattered. respective1y
(117). The total attenuation coefticient. p,. is equal to the sum of the scattering and absorption coefficients. For tissue. the absorption coefficient. pa,is approximately 0.0I -
I mm-',the scattering coefticient. p,. I0 - 100 mm*'. and the anisotropy factor. g. 0.8 -
0.95 (124).
Figure 2.7 Laser-tissue interaction according to Snell's law. Laser light, when delivered to the tissue surface through air. is partially reflected at the surface, according to Snell's law (121) (Figure 2.7).
n, sin0, = n. sin&
where n I = index of refraction of air 81 = angle of incidence on tissue = 92 Ir,dex of ref:.attisn of tissue 02 = angle of refraction into tissue due to the difference in the indices of refraction. The indices of retiaction for soft tissues with their high water content (60 - 70 YO)range from 1.38 to I ill. For reference. 1.33 is the index of refraction of water and 1.45 for adipose tissue A guideline relation for calculating indices of refraction is
where w = water content in grams water per grams total weight
Bone is approsimately 12% water by weight corresponding to an index of refraction around 1.5 1 ( 124).
At mismatched boundaries such as the air-tissue interface, two surfke effects must be considered. Internal reflection is imponant to consider as it reduces the escape of photons From the tissue. Internal reflection usually reflects 507'0 of the total photons that strike the tissue surface. The fraction of light directly reflected from the surface depends on the angle of incidence of the laser beam and the two indices of refraction. The differing refractive indices between air and tissue give rise to specular reflection of the incident light even when directed perpendicularly to the surface. The reflected lisht yields photons that have not sampled the tissue interior thereby altering reflectance measurements. Although these photons provide information about surface roughness and tissue refractive index, they do not provide information regarding tissue absorption or scattering properties. Specular reflection depends on the tissue refractive index and is defined by Fresnel's relation ( 12 1 );
- -I I A -1 tan (3, t 3. j sin ' (0, T G1j 1 where r = specular retlectance 0 1 = angle of incidence measured with respect to the surface normal 0: = angle of refraction measured with respect to the surface normal
and when the beam is normal to the surface (€I1= 0) ( 13 1):
where r = specular reflection n I = index of refraction for air n:! = index of refraction for tissue
The amount of specular reflection may be slightly higher than predicted by the above equation since it assumes a smooth surface. Specular reflection strongly depends on the exact surface and illumination conditions since surface optical irregularities and roughness can decrease specular reflectance and increase diffUse retlectance. In the tissue, scattering and absorption attenuate the collimated beam and %rther de-collimate the incident flux as photons are scattered away from the laser beam ( 12 1). Some light scattered from the beam undergoes multiple reflections and propagates backwards. The light that reaches the tissue surface is either internally reflected or transmitted according to Fresnel's relation. All measurements of reflection therefore include the specularly reflected and transmitted diffuse portions of backscattered flux. Effects of specular reflection can be reduced or eliminated by placing a refractive medium with a similar retiactive index to tissue on the surface of the experimental tissue. By irradiating tissue through a water layer. specular loss can be less than 0.1% (1 21).
h photon can penetrate deeply into tissue and be remitted. despite multiple scatterings. due to a higher frequency of scattering than absorption interactions and a highly fonvard-directed scatter. The penetration of 600-1300 nm light can be characterized by the effective attenuation coefficient. pCf~;in mm-'.or by its inverse which equals the effective penetration depth. S. in mm (124)
cv here S = penetration depth I- = et'fective attenuation coeficient Pa = absorption coefticient CLs = scattering coefficient = anisotropy coefficient
The penetration depth of laser light is the depth required to reduce the fluence rate by a factor of i/e or to 37% of its original value (111). The penetration depth increases rapidly between 600 and 700 nm due to the decrease in the hemoglobin absorption and increases slowly or remains constant above this wavelength, except for a small decrease at 960 nm due to the presence of the water absorption peak. The optical spectrum of light propagating through and remitted from tissues, however, is highly influenced by scattering and therefore, the spectrum does not simply depend on the absorption spectrum of tissue chromophores.
A collimated laser beam directed normal to the surface has a small component of reflected light at the surface and the remaining light attenuated in the tissue by absorption and scattering according to Beer's law (121)
where E(z) = fluence rate of collimated light at position z E, = collimated irradiance
~SC = specular reflection of collimated light on external surface Pa = absorption coefficient PS = scattering coefficient
KrF XcCl Dye Argon nlode Nd:YAC Tm:YAC 1lo:YAG Er:YAG CO t18 308 465 5143 830 1064 20 10 2 100 2940 1060b
Figure 2.8 Mean f?ee path of laser light in soft noncoloured tissue as a function of
wavelength. Adapted &om (12 1). The average distance a photon will travel in the tissue before being scattered or absorbed is described as the mean free path (Figure 2.8). It is approximately 10- 100 pm and defined as (12 1):
where MFP = mean free path of a photon Clt = total attenuation coefficient
For tissues that do not scatter light (p, = 0). the absorption coetliciet can be determined
tiom the slope of In 7, (collimated transmission) versus sample thickness ( I 2 I ):
where E(z) = tluence rate of collimated light at sample thickness z Eo = col I imated irradiance T; = collimated transmittance ~SC = specular reflectance pa = absorption coefficient
When scattering is a component. the slope of lnTc versus thickness gives the attenuation
coefficient ( 123)
where Pt = total attenuation coefficient z = thickness of tissue sample Tc = fraction of normally incident collimated flux that is transmitted through the sample of thickness z without scattering Tissue optical propenies can be measured directly or indirectly. The direct method of measuring tissue optical properties involves illumination of thin tissue samples such that multiple photon scattering is negligible. The tissue sample thickness must be less than the average distance traveled by a photon before a scattering or absorption event
(mean free path). In soft tissues illuminated by 630 nrn wavelength light. the samples must be less than 10 pm. Indirect methods are based on ir~virro or ir~rpitro measurements of macroscopic parameters in bulk tissue from which microscopic parameters are deduced by applying one or more light propagation models ( 127).
2.1 Current kleasurement Techniques
Direct measurement ofcortical bone blood flow is dificult due to the complesity of the arterial and venous channels. Indirect bone hernodynamics have been studied using a variety of methods including cannulation of vessels (38). uptake or clearance techniques (clearance of bone seeking tracers) (1 7). fractionation of diffusible indicators
(indicator fractionation) (1 14). washout of diffusible tracers ( I 14). entrapment of coloured or radiolabeled microspheres ( l 17). and laser Doppler tlowrnetry ( 107).
Cannulation or pehsion of nutrient vessels permits continuous assessment of bone blood tlow. which allows for transient alterations in flow to be quantified. This method, however. will only measure blood flow to one segment of bone in each experiment. and blood tlow to compact, hemopoietic, and marrow components in a given bone cannot be assessed independently (35).
Clearance of bone seeking tracers, such as "sr, has been used to measure blood tlow in cortical bone (17). This technique is based on the principle that the amount of tracer accumulating in the tissue is the difference between the input and output tracer volume. The complex venous drainage of long bones, however, precludes the collection of all of the venous blood so it is difficult to measure accurately the effluent tracer ( 1 14).
Although this method allows for repeated measures in a bone, the measures are limited to the one bone, and determination of regional flows within the same bone require extrapolation of the washout curve (38). Validity of this method is based on the assumption that tracer extraction is complete or at least constant with a fixed proportion of input being accumulated by the bone. It is assumed that the time between injection and cessation of circulation is shorter that the minimal re-circulation time but longer than the maximal transit time through the bone of interest (1 14). The variable nature of tracer extraction precludes quantitative blood flows from clearance of bone seeking tracers since extraction of diffusible tracers depends on the tlow rate. Extraction is higher at lower flow rates since the transit time though the capillary is too shon at higher flow rates for the ion to ditfise completely through the capillary wall (57.1 14). Despite the assumptions underlying this technique it is the best method of blood flow assessment in noninvasive human studies.
Ditfusible indicators allow measurement of regional blood tlow in multiple bones simultaneously but only one measurement of blood flow can be made in each animal
( 1 14). Determination of inert gas wash-out by mass spectrometry can estimate bone blood flow. Bone tissue is loaded with an inert gas, and the wash-out curve (58) is measured in the bone compartment with a blood gas catheter connected to a mass spectrometer (72). The indicator wash-out technique assumes that the vascular pattern in bone tissue is organized in parallel such that supply from the nutrient artery, rnetaphyseaI/epiphyseal branches and periosteal vessels do not contribute to the other areas. This technique also assumes that there is no recirculation of the tracer although flow was only underestimated by 8% without correcting for recirculation (60). Indicator wash-out curves may be disturbed by irrelevant indicator dispersions, independent of flow, that occur during passage through the measuring system (67). Use of multiple indicators allowed for examinat ion of differences between indicator curves making systematic errors in the dispersion of the indicator less likely (60).
Although washout of diffusible tracers has been well established for measuring blood tlow in several organs. bone measurement remains problematic. This method is based on the principle that the rate of removal of the introduced tracer depends on the blood tlow such that the flow per unit weight equals the rate constant of the exponential fall of tracer activity multiplied by the solubility ofthe tracer in tissue relative to that in blood ( 1 1I). introduction of the tracer. recirculation. heterogeneous tlows leading to multiple esponentials. and estimates of capillary blood tlow within an ill defined volume ofbone in the region of injection remain problems associated with this technique (1 I-I).
Injection depth and differing tissue diffusion rates also influence the technique (I 10).
Current standards of measuring blood perfusion in bone indicate that entrapment of microspheres is the most accurate method for experimental assessment of bone blood tlow (1 10). Several studies have been performed in a variety of animals and tissues. especially to study flow in long bones (Table 2. I). This technique assumes homogenous arterial distribution of microspheres, the tracer particles will be trapped in the capillaries and arterioles, and all microspheres will be removed in one passage with no back diffusion or recirculation (80.1 14). The size of the tracer is a compromise between the size of erythrocytes and the avoidance of nonentrapment in the capillary bed (46). Microspheres less than 10 pm were more likely to be shunted, and thus 15 pm microspheres are the most suitable for measuring bone blood flow (37,117). Although 25 pm microspheres were suitable for measuring bone blood flow. the uniformity of mixing larger microspheres in blood was more likely to be compromised by axial streaming or skimming at junctions ( 1 14).
Radiolabeled and coloured microspheres have the advantage of allowing several repeated measurements of blood flow in one animal such that total and regional flows to any bone can be measured (63).The microspheres can be labeled with gamma emitting isotopes such as "SC. '"~b, "sr. or '"1 making repeated measurements in one animal possible. Multiple injections of microspheres did not alter resting blood tlow to bone or interfere with vasomotor responses in bone (37).
Disadvantages of using the microsphere method include the inability to measure transient changes that occur over a few seconds since tlow is not measured continuouslv, the number of measurements per animal is limited. and it cannot be used in humans (38).
The variability in flow measurements with use of microspheres also increased with increasing blood tlow due to plasma skimming, an axial accumulation of spheres in the central stream of the arterial vessels ( 1 12). Table 2.1 Previous studies of bone blood flow by the microsphere method. Blood
flow rates in mllmin/ 100%.
BONE ANIMAL METHOD FLOW RATE REF ml/min/ 1 00g Proximal cortex tibia Rabbit Microspheres 1.73 2 0.6 114 Distal cortex tibia I Rabbit I Microspheres I 2.46 k 0.75 1 111 (1 Tibial cortes I Rahhit I Microspheres 1 2 n7 f or;h I 1 I4 11 Proximal cortex tibia I Rabbit Microspheres 1.07 k 0.4 1 Proximal cortex tibia I Rabbit I Microspheres 1 0.72 2 0.14 1 1 13 11 Distal cortex tibia I Rabbit I Microspheres 1 2.10 k 0.6 1 1 1 1 3 (1 Distal cortex tibia Rabbit Microspheres 2.10k0.38 I13 Proximal cortex tibia Rabbit Microspheres 1.42 k 0.34 112 Proximal cortex tibia I Rabbit I Microspheres 1 2.01 k 0.54 / 112 11 Distal cortex tibia I Rabbit ( Microspheres I 1.83 2 0.j9 1 1 12 11 Distal cortex tibia I Rabbit I Microspheres 1 1.99 5 0.56 1 1 12 11 blarrow I Rabbit I Microspheres 1 26.53 k 3.33 Muscle 1 Rabbit I Microspheres 1 15.57 c 4.07 1 1 14 11 Ligament (MCL) I Rabbit I Microspheres 1 0.68 2 0.08 ( 6 1 Ligament (ACL) Rabbit Microspheres 1.30 f 0.39 6 Femoral diaphysis Do&! Microspheres 2 k 0.5 36 Distal epiphyses Dog Microspheres 26 2 6 36 Femoral head Dog Microsp heres 14.3 21.9 110 Diap hysis I Dog I Microspheres 1 1.2 f 0.1 P'o 11
, Distal epiphyses IMicrospheres 18.3k2.1
Although it has been suggested that a minimum of 400 microspheres per sample
is required for reasonable accuracy (l2), blood flow determined in samples with less than
400 microspheres were not significantly different than samples with greater than 100
microspheres (70). The relative error of flow in bone samples with less than 200
microspheres was also less than 10% (70). An adequate number of microspheres per bone sample to obtain an accurate blood flow estimate is between 150 to 750.The number of coloured microspheres per sample to obtain an accurate estimate, however. is lower as this method is more accurate in determining flow in low flow samples compared to radioactive microspheres (39).
The above techniques permit the assessment of bone blood flow. but they remain
highly invasive and. for the most part. are not clinically applicable. Laser Doppler technologies. such as tlo wmetry, pehsion imaging, and most recent 1y laser speckle.
offer a less invasive method of measuring and visualizing perfusion in soft and hard
tissues.
The laser speckle method is a new form of flowmetry that analyzes the
interference pattern that appears when tissue is illuminated. The principle of this method
is the same as Doppler technologies, and the output generated is a random intensity
distribution (speckle pattern) from the reflected photons. The structure of the speckle
pattern changes with blood flow. and the rate of variation depends on tlow velocity (33).
This technique was effective in measuring blood flow in skin and the ocular retina and
significantly correlated to flow determined by hydrogen washout methods in subchondral
bone. This method is a non-contact. rapid technique that although unable to continuously
monitor flow, measurements can be repeated suRiciently rapidly to assess real time
changes in bone blood flow such as reactive hyperemia (33).
LDF has been used extensively to measure and monitor changes in perfirsion in
hard and soft tissues. In bone specifically, LDF has been used in animal models and
humans to evaluate cortical bone perfusion in fractured tibiae (50,65,97), circulation in
corticocancellous bone grafts (6 1). perfusion in femoral head osteonecrosis and osteomyelitis (68.98.107), and the effect of femoral neck fracture on femoral head blood flow (109). The reproducibility of multiple measurements of the same position ofthe probe has been found to vary between 5% and 15% in cancellous bone (44.69.1 10).
LDF can continuously measure perfusion of a single point over time but because of large spatial variations in tissues. this method is generally better suited for stimuli- response experiments than for mapping blood flow distribution over a certain area of tissue (5 1 ). LDF has been primarily used for monitoring dynamic responses of the
microvascular perfusion to various physiological stimuli or for demonstrating regional distribution of perhion in a specific organ (85.104.1 10).
Studies have been completed that determined the maximal depth at which LDF
can evaluate flow in bone. LDF was reported to measure an area approximately 1.5-2
mm' with a maximal penetration depth of 3.5 i. 0.2 mrn in trabecular bone. 2.4 to 2.9 t
0.2 rnm in conical bone and 4.0 mm in cartilage covered trabecular bone (84.97). In the
experiments. cylindrical samples of conical and trabecular bone were placed in a flow
chamber. The tip of the LDF probe or fibre was placed on the surface of a sample as a 246
solution of Neoprene latex panicles was circulated. at 60 mm/s. underneath the samples.
The bone samples were reduced in thickness by 0.7 mm if the measured flow was not 5
rnV above the noise values from Brownian motion. The thickness at which the detected
flow was above the noise value was determined to be the threshold thickness or maximal
depth. The threshold thickness in cortical bone was found to be within a clinically
significant range for flow detection demonstrating the usehlness of this technique (97).
Similarly, the depth in bone to which laser speckle could be used in subchondral bone
was 2 rnm (33). Although these techniques determined the maximal depth at which laser Doppler technologies detected flow, optical properties of the tissue are still major factors in understanding the output.
LDF output. blood cell flux (in volts), is difficult to apply to experimental and clinical problems of normal and pathological bone pehsion, but correlation to a technique that outputs a rate of blood flow would make application of LDF easier.
Several investigators have evaluated the ability of LDF to detect changes in bone perfusion. X significant correlation (r = 0.62, p = 0.0 I) was detected between simultaneous LDF and microsphere determined blood flows in cancellous bone (69).
Although others have not found significant correlation between the two methods ( 1 10). their caiculated correlation coeflicient was from observations of several different animals that may have masked significance because results from LDF were relative values. Thus. it may not be appropriate to pool LDF output when different test animals are studied (69).
The lack of correlation may also have been due largely to bleeding artifact. The inability of LDF to distinguish bleeding from capillary flow, larse arteriovenous shunts. insufficient microsphere counts to accurately measure blood flow. and the time differential between LDF and microspheres may have been contributing factors.
Microsphere derived flow only reflects a one-time static estimate of bone pehsion whereas LDF represents real time continuous flow (1 10). LDF evaluates blood cell motion in small volumes of bone whereas the microsphere methods estimate flow in larger regions.
Another disadvantage of LDF is the requirement of the probe to have direct contact with the tissue surface, which may impose a risk of infection or pain. The probe must also be fixed or held constant to prevent motion artifacts and unusually high pefision measurements (50). LDF output can only provide a relative estimate of microvascular pefision and cannot be calibrated as absoiute values because of variation in the optical properties of various tissues and spatial variation in the microvascular bed
(68).
The LDF technique. however. is useful in monitoring the dynamic response to physiological stimuli at a specific location. but several consecutive measurements must be taken to obtain a representative estimate of relative microvascular blood tlow in different regions. Shepherd and colleagues demonstrated a linear relation between LD F output and controlled pehsion rates at one specific point. but the slope of this relation differed between animals precluding the reproducibility of LDF measurements in different animals (99). Studies cornparins the output of LDF. electromagnetic flowmeter. and radioactive microspheres showed the mean LDF output signal calculated from multiple surface tissue readings (6 - 8 points) was significantly correlated with the other techniques ( 104). The mean signal. therefore. was a good measure of tissue pefision within a discrete area of the same tissue or vascular bed provided a sufiicient number of points were scanned. There is not. however. a universal calibration factor for LDF output since the level of hematocrit. red blood cell velocity. vascular geometry. tissue optical properties. and point-to-point variations of blood flow within a tissue may vary between tissues ( 104). This study indicated the potential of scanning technology since LDF measurements at multiple sites on a given tissue resulted in a range of output values.
Scanning techniques may provide important information about changes in heterogeneity of tissue perfusion that afiects tissue hnction ( 104). Laser Doppler pefision imaging (LDI) has been used to measure perhsion in soft tissues (6). Based on the same principles as LDF. LDI has the ability to scan a designated area to generate multiple measurements. Correlation between coloured microsphere and LDI-determined blood flow has determined LDI can be a valid tool for measuring blood tlow in soft tissues. LDI, therefore. has the potential to be used to measure perfbsion in hard tissues.
The ability of LDI to generate pehsion maps is advantageous over LDF as the maps can be used to determine average tissue perfirsion or to study the two-dimensional structurins of tissue microcirculation (30). The average tissue pehsion value can be used to compensate for the tissue heterogeneity such that pefision measurements of a particular tissue can be compared between different animals (3I ).
2.5 Study Objectives. Rationale, and Scope
Objectives
The objective of this thesis was to assess the ability of laser Doppler pehsion imaging to detect changes in cortical bone perfusion. More specifically. the feasibility of using laser Doppler perfusion imaging in bone was studied in relation to its sensitivity. repeatability, and comparison to a standard measure of blood flow in bone.
The principal hypothesis was that laser Doppler pefision imaging accurately measured induced changes in cortical bone pefision. Through a series of experiments. we determined if the laser Doppler pefision imager detected perfbsion in bone at a particular depth, given LDI's ability to scan with different wavelengths. A range of detectable flows and sensitivity of the laser Doppler pefision imager were determined. Rationale
The limited availability of less invasive techniques to measure perhsion in bone has increased interest in laser Doppler technologies. Although laser Doppler flowmetry
(LDF) is a generally accepted technique for monitoring dynamic bone blood tlow responses to physiological stimuli. it is unable to show tissue heterogeneity unless multiple measurements are taken across the tissue. limiting the temporal resolution of
LDF. Laser Doppler perhsion imaging. in contrast. has the potential ability to map tissue perfusion over a broader area (e.g.. 12 x 12 cm). The potential for creating areal perfusion maps suggested that LDI might be a useful tool in investigating changes in tissue pehsion patterns that other techniques could not study. LDI has been used to study pertision changes in several tissues, including skin ( 120) and ligament (6.30). but the ability oELDI to measure perfusion in bone has not been investigated.
Scope
A combination of animal and flow models were used to ascertain the utility of laser Doppler perhsion imaging to detect changes in bone blood tlow (Figure 3.9).
Laser-tissue interaction experiments were conducted in conjunction with animal model experiments to aid interpretation of LDI output. Pilot Study to test LDI on bone
Studies to Test Hypothesis
I I CpScai Ptoprrty uExperirnents wExperiments
Transmission Flow Range
IRepeatability of LDI
Comparison of red and nir wavelengths
Comparison with coloured microspheres
Figure 2.9 Sequence of experimental studies involved in the analysis of the utility of
laser Doppler perhsion imaging in bone
Initially. a pilot study was performed to determine if LDI could detect flow in hard tissues. specifically cortical bone. and was capable of measuring changes in bone blood flow. That first study was pivotal to decide whether or not the LDI technique warranted hrther investigation. A second study involved a larger number of animals to test our hypothesis that
LDL could measure statistically significant changes in bone perfusion. Limitations of the
LDI technique were investigated as foundation for comparison to other methods.
Examining light propagation and distribution in tissue is essential for the effective application of optical diagnostic and therapeutic techniques (3 1 ). Laser Doppler
techniques measure perfusion in tissue but the uncertainty in photon sampling depth leads to ambiguous interpretations in the fraction of microcirculation that contributes to the
Doppler signal (30.5 1). Investigations into the optical properties of cortical bone have
been limited thus far to the porcine skull (28). and therefore, two separate transmission
experiments were performed to determine the optical properties of conical bone. These
experiments defined the depth from which the majority of the LDI signal was contributed
and the depth to which LDI was capable of detecting blood tlow at red and near infrared
wavelengths.
Two flow models were used to study the limitations of LDI's perfusion
measurement capabilities. The first tlow model was constructed to study the ability of
LDI to detect a range of flows and determine if the detectable flows were in the
physiological range. Specular reflection. a limitation of imaging techniques. was first
detected in our pilot study. The second model was created to study the effect of specular
reflection on the output of the LDI and devise a method to compensate for the artifact.
Understanding the limitations on LDI's ability to measure perfusion in cortical
bone provided preliminary information on how to interpret the output data. The final
study to evaluate LDI to measure blood flow in cortical bone involved three separate
experiments. The first study sought to ascertain the reproducibility of the LDI signal in conical bone. Each area was scanned four times over a period of 4.5 min to test for repeatability and temporal variation.
It has been shown in skin that longer wavelengths can penetrate deeper into tissue and thus. different wavelengths are capable of measuring flow in different regions. A comparison between LDI output of red and near infrared wavelengths was completed to determine if the two wavelengths were measuring tlow in different regions of bone. The results were also compared to the derived penetration depths.
Coloured microspheres were used as a "reference standard" for the validation of
LDI in the measurements of cortical bone blood flow. Simultaneous measurement of blood flow by the two techniques determined the accuracy of LDI to estimate tlow in
bone. The penetration depth derived from the thick section transmission experiments
permitted an approximate comparison of LDI output and coloured microsphere-
determined tlows since it was possible to analyze the same volume of bone for both
methods. METHODS
3.0 Laser Doppler Perfusion Imaging Standard Protocol
Experiments were undertaken to determine the utility of the laser Doppler perfi~sionimager to investigate changes in bone blood flow. A standard protocol tbr measurin~perksion (6). modified for its use in bone. with the laser Doppler pefision imager was used as bone blood tlow was measured using a modified dual-wavelengh laser Doppler perhion imager (Moor Instruments Ltd.. Devon, England. UK) with either a red or near-infrared (NIR) wavelength helium-neon laser source (634 nm or 8 10 nm and
1.5 mW power). The scanner head was placed approsimately 25 cm above the exposed bone (Figure 3 1 ). Scanning speed was set at 10 ms/pixel. which gave a frequency bandwidth of 0.1 - 15 kHz. Flux and DC gains were adjusted to maximize signal strength while preventing saturation of the photodiode.
After the skin and sot1 tissues surrounding the area of interest were removed. the unexposed regions were masked with black cloth to demarcate the area of interest and to provide a zero background at the exposure margins. Afier all testing, a scan was taken immediately post mortem to give a "biological zero". The mean pixel value of each scanned image was calculated using MoorLDI Image Processing software (version 2.0.
Moor Instruments). The mean of the post mortem scan was subtracted from each corresponding test scan to eliminate background noise produced by signal processor offset voltases or Brownian motion. The calculated mean flux, in perfision units (PU), represented the average blood flow in cortical bone. Figure 3.1 Laser Doppler pefision imaging experimental setup. Adapted from (3 1).
3.1 Laser Doppler Perfusion Imaging Pilot Studies
Five 1 yr old White Leghorn hens (1.41 f 0.03 kg) were anaesthetized (2 % halothane; 0.8 L/min 02;0.2 Llmin N20),and with approval of the University of Calgary
Animal Care Committee, skin, soft tissue, and the periosteum surrounding the left tarsometatarsus (TMT)were dissected to expose the anterior mid-diaphyseal cortex. The relatively flat anterior cortex of the TMT was chosen as the pefision measurement site to reduce distortion and specular reflection due to the curvature of the cortex, which would cause excessive sideways scatter of the photons. Bone blood flow was measured using the standard protocol described in section 3.0.
To assess the ability of LDI to measure changes in bone perfusion, a variety of methods were used to alter blood flow in the TMT. Xnitially, blood flow was altered by occluding the femoral artery for 2 min. Blood flow was then altered by pharmacological intervention involving adrenaline (Adr) (0.2 mi 1 o4 M) and acetylcholine (Ach) (0.2 ml
lo-' M) injected through cannulae inserted into the femoral artery. Scans were taken prior to (control scan), during, and 5 min following occlusion of the femoral artery and at 0. 2. and 5 min after drug administration. A scan was taken immediately post mortem.
Significant changes in perfusion levels were evaluated using a Student's paired t-test (p <
0.05).
3.1.2 Rahbitibfutiel
Five New Zealand White rabbits (body mass range: 5.1-6.2 kg; (mean 2 standard
deviation) 5.71-2 0.55 kg) were prernedicated with I .S ml atrovet (Ayerst Laboratories.
Division of Wyeth-Ayerst Canada. Montreal. Quebec. Canada) and anaesthetized with
2.5% halothane and I Umin 02.With approval from the University of Calgary Animal
Care Committee, skin and soft tissue surrounding the left and right tibiae (n = 10) were
dissected to expose a 3 x I cm area of the mid-diaphyseal medial cortex. We chose the
medial tibia1 cortex as the site of measurement to minimize distortion caused by mapping
a three-dimensional object possessing depth to a two-dimensional image of perhsion.
The relatively flat surface also made it possible to direct specular reflection away from
the detector. Specular reflection is an unwanted artifact from laser Doppler imaging and
may dominate the output signal in bone measurement. The periosteum was removed. and
blood flow at the cortical bone surface was measured with use of the near-infrared (S 10
nm) wavelength of the laser Doppler perfusion imager.
To assess the ability of laser Doppler perhsion imaging to measure changes in
pertision in bone, blood flow was altered by occlusion of the femoral artery for 5 min.
The laser Doppler perfusion imager was set at normal resolution so that the exposed bone (approximately 2.5 x 0.7 cm, occupying 104 x 37 pixels) was scanned in 45 s. Scans were taken before each occlusion. immediately after occlusion. and at 45-s and 3-min time intervals after release of the artery. A final scan was taken after death to subtract a biological zero from all images. Significant changes in pehsion levels were evaluated with use of a Student's paired t-test (p < 0.05).
3.2 Boue Optical Properties Experiments
3.2. l TIt in Section Tmnsrtti.~sion
Both tarsometatarsi of a 1 yr old White Leghorn rooster were dissected of all so% tissue. Three cross sections. approximately 8 mm wide. were cut from the mid-diaphysis using a low speed diamond saw (Buehler. Isornet. Illinois. USA). The blade and samples were constantly bathed in 7050 EtOH. From each cross section. 3 mm thick slices were cut parallel to the anterior side of each section and hand ground to a thickness of 55. 70.
85. 125. 1-40, or I55 pm. The thickness of each sample was determined using a micrometer. Samples were placed in saline and frozen at -30°C.
Collimated laser source
Slide and bone samole
Detector
Figure 3.2 Apparatus for thin section transmission experiments. On the day of testing. each sample was thawed and then mounted on a slide and
held normal to the laser light source on a translation stage (Figure 3.2). The samples were kept moist with saline. Red (633 nm) and near infrared (8 10 nm) collimated laser beams
were used to illuminate each cortical bone sample. Each sample was diffusely
illuminated. in tive locations on one side, and the transmitted light intensity was
measured on the opposite surface with a photodiode (Universal Optical Power Meter.
Melles Griot. Colorodo. USA) placed 19 mrn From the tissue surface. The detected tlux
I(t) For each sample thickness. t. for each wavelength. was measured for the tissue and
repeated on a clean slide to measure the incident light flux I&). This measurement
compensated for losses due to specular reflections at the air-slide and slide-tissue
interfaces. The ratio of transmitted light with and without the sample present was
calculated. Linear regression analysis was used to determine photon pathlength.
The total attenuation coeflicient was calculated from:
where - total attenuation coefficient (mm*') t - thickness of sample (mm) - 10 - incident light flux (V) I - transmitted light flux (V)
which was equivalent to the slope of lnT, versus thickness (122). The mean free path of a
photon of each wavelength was then determined by (124): where bEP = mean free path of a photon (pm) - Pt - total attenuation coefficient (mm")
3.22 Thic*k Section Trtrnsr~lissiun
Two 2 cm wide conical bone samples were taken from the diaphysis of a mature rabbit tibia using a low speed diamond blade saw (Buehler. Isomet. Illinois. USA). The medial surface was removed just above the marrow cavity (approximately 1.5 mm thick)
The blade and samples were constantly bathed in 70% EtOH. The uncut surface served as the entrance site for the laser light.
Tissue Collimated -
Figure 3.3 Schematic for the thick section transmission experiments
Each sample was placed on a holder with a 2.5 cm diameter hole to allow for interaction of the laser light only between air and the tissue (Figure 3.3). Attenuation of
red and near infrared collimated laser light (wavelength 633 and 8 10 nm, respectively) was determined by directing the beam of the laser at the centre of the photodiode. The holder was placed normal to the laser and between the laser and photodiode. such that the photodiode was 19 mm from the bottom of the bone samples. A reference value through air, incident light flux, was obtained. Light transmitted through various thicknesses of bone was then measured. The bottom of each sample was hand ground progressively in increments of 0. I mm (84). After each grinding step. transmittance through the new thickness was measured.
Similar to the thin section transmittance experiments, comparison of the collimated transmission to sample thickness was equated to the effective penetration depth by (121):
and
where S = effective penetration depth (mm) pc,y = effective attenuation coefficient (mm-')
3.3 Flow Model
3.3.1 Detectrrbk Range of Flo~v
Two tlow systems were designed to determine the range and sensitivity of the
laser Doppler pefision imager. The first model consisted of continuous polyethylene
plastic tubing (PE 90, OD = 1.27 mm, ID = 0.86 mm) (Intrarnedic-Clay Adarns, Becton Dickson, USA) overlying a piece of plastic (Figure 3.4). Black cloth and tape marked an area of interest (2 1 mm x 15 mm) and created a zero background at the margins of the test area. One end of the tubing was placed in a beaker where a 1% volume fraction of blood diluted with a physiological heparinized saline solution was being continuously stirred. The other end of the tubing was attached to a pump (Harvard 3'Syringe Pump.
Mass. USA) that was used to draw the solution continuously through the model at knowi~ flow rates.
Pump Fluid of interest Flow model
Figure 3-4 Schematic of the flow model used to determine the range of tlow
detectable by the laser Doppler perhsion imager
The tlow rate was started at 0 pL/min and incremented by 10-20 pL/min until it exceeded the noise region of the LDI. The rates were then increased by 50 pL/min to determine the linear region of sensitivity. Flow ceased when the LDI detector saturated as indicated by a plateau or decrease in flux values with a fbrther increase in flow rate. The laser Doppler perhsion irnaser was placed in single point mode (similar to laser Doppler flowmetry) and continuously measured flow at a single point. Measurement at no tlow conditions showed movement of the particles under Brownian motion. Once a linear flow region was determined, a value in the middle of the region was chosen, and the rate was set in the pump. When the measured pefision unit was steady, the flow rate was incremented by 1 pL/min until LDI output changed.
The second flow model used bone. instead of plastic. over plastic tubing with fluid passing through at known flow rates (Figure 3.5). The mid-diaphyseal region of a rabbit femur was removed and cleaned of all soft tissue. The cortical bone surrounding but not including the marrow on the medial side of the bone was removed by a low speed diamond saw (Buehler-tsomet, Illinois. USA).
1 To tluid of interest
To pump
Figure 3.5 Schematic of the tlow model used to determine the ranse of detectable
flow by LDI in bone
X groove (approximate diameter 1.2 mm. depth = 1.5 mm)was hand ground into the conical shell. Thin plastic tubing (OD = 1.27 mm. ID = 0.86 mm) was placed into the groove and 1% blood in heparinized saline, being continuously stirred. was circulated through the system. As previously, the solution was continuously drawn through the system by a pump at known flow rates to determine the linear output range.
3.3.2 Specular Reflectance
A tlow model consisting of thin plastic tubing (OD = 1.27 mm. ID = 0.86 mm) encased between two pieces of semi-transparent white plastic was used to determine the effect of specular reflectance on the output of the LDI (Figure 3.6). The model was covered in black cloth and an area (2.1 ?: 1.5 cm) in the centre of the model was left open as the area of interest.
To fluid of interest
To Pump
Figure 36 Schematic of the flow model used to study the effect of specular reflection
on LDI output.
Solutions of either 10% milk or 1% blood in heparanized saline were pumped through the system at known tlow rates varying between 0 and 800 pLimin. The model was placed either flat or angled at approximately 7' to the horizontal to direct specular retlection away From the photodiode. Differences between LDI flux values of the angled and flat surfaces were compared using a Student's paired t-test.
3.4 Laser Doppler Perfusion Imaging Study
3.4. I At1 itrml Preparation
Twelve mature New Zealand white rabbits (Riemens Fur Ranche, St. Agathe.
Ontario, Canada), weighing approximately 6.0 k 0.4 kg, were randomly assigned to one of three groups: normal control, constriction (noradrenaline), or dilatation (nitroprusside). Rabbits were premedicated with 0. 15 ml intravenous acepromazine maleate and subsequently anaesthetised with a mixture of halothane (2.5%) and 100% oxygen ( 1 milmin). Body temperature was maintained at 37 "C with the aid of a heating pad.
Both brachial arteries of each animal were cannulated with PE-50 polyethylene tubing (Intramedic-Clay Adams. Becton Dickson. USA). The first cannula was connected to a physiological pressure transducer (Pl3XL Spectramed, Oxnard. CA. UAS) to monitor arterial blood pressure. The pressure signal was amplified using a transducer
(PM- 1000 CWE. Ardmore, PA. USA) and recorded on a computer based oscil lographic data acquisition system (AT-CODAS System. Dataq Instruments. Akron. OH. USA).
Zero pressure was taken at a level 2.5 cm below the midpoint of the chest with the rabbit in a supine position. The second cannula was connected to a syringe pump (model 3 3.
Ealing Harvard Scientific. St. Laurent. Quebec. Canada) for withdrawal of a reference blood sample. The leH corotid artery was cannulated and connected to a second pressure transducer to confirm placement in the left ventricle. Knee joint position was standardized to 90" to the horizontal. using foot and leg supports, to align the rnid- diaphysis of the tibiae parallel to the surface. With approval from the University of
Calgary Animal Care Committee. all skin and soft tissue surrounding the left and right tibiae were dissected to expose the mid-diaphyseal medial cortex. The periosteum remained intact.
3. A2 Laser Doppler Perfusion Intqing
Blood flow at the cortical bone surface was measured with red (634 nm) and near infrared (8 10 nm) wavelengths of the laser Doppler pefision imager. Setup of the LDI and area of interest was according to the previously described standard protocol (section 3 -0). The LDI was set at a normal resolution so that the exposed bone (average area 5 1 x
6 rnm; 800 pixels) was scanned in approximately 30 s. The medial mid-diaphyseal surface was chosen as the perfUsion measurement site to reduce distortion and specuiar retlection due to the curvature of the cortex. which would cause excessive sideways scatter of the photons. The head of the LDI was also angled sli_ghtlyto direct specular retlection away from the photodiode. The exposed tibia1 sections were scanned incrementally with both wavelengths with two scans performed consecutively at each wavelength to test for LDI repeatability. The scans were then repeated incrementally to give a total of 4 normal scans per tibia per wavelength.
A vasoconstrictor and a vasodilator were chosen to alter pehsion in bone.
Although noradrenaline causes an increase in arterial blood pressure. it has been shown to
decrease bone blood flow by constricting bone vessels (37). Nitroprusside has been
shown to significantly decrease arterial pressure and cause aneriolar vasodilatation as
well as venodilatation (17). Rabbits in the vasoactive drug groups were administered
either nitroprusside (1 mg/kg diluted in 0.3 ml saline (47)) via intravenous bolus infusion
(maginal ear vein) or noradrenaline (8 ug/kg/min (95)) via intravenous infusion in the
lateral ear vein in the right ear. The amount of vasoactive drugs, nitroprusside or
noradrenaline, corresponded to a 40 to 50% decrease or 20 to 30 % (20-30 mmHg)
increase in arterial pressure. respectively. Following injection. 3 to 5 rnin were allowed
for vascular system stabilization that was confirmed by the arterial pressure trace (47).
3.4.3 Evclluntiun of Blood Flow
Coloured microsphere (CM) blood flow determination was performed according
to standardized protocols (63). with modifications for low flow tissues (7). A reference blood sample was withdrawn at a rate of 3 mllmin starting 5 s prior to infusion of the microspheres through the cannula inserted into the left corotid artery. Approximately
10.2 million 15.5 prn coloured microspheres (Titron Technology, San Diego, ChUSA) were diluted in a saline solution, vortex-mixed, and then infused into the left ventricle over a period of 30 s. The reference sample was continuously withdrawn for another 60 s after infusion of the microspheres. For a total collection period of 95 s. LDI measurements. at an 8 10 nm wavelength. were obtained concurrently with CM determinations. The animal was sacrificed with an overdose of pentobarbitol. Final LDI scans were performed following sacrifice to obtain a "biological zero" reference point that was subtracted from each LDI measurement.
3.4.J Bluurl Ffurv Dcterrrriniltion
Both kidneys and tibiae including the periosteum covering the scanned area were removed. weighed. and prepared for blood flow processing. The cortical bone of the medial mid-diaphysis was removed with a Drernel rotary cutting tool (Rilultipro Variable
Speed. Drernel. Racine WI, USA). A11 traces of marrow were removed. and the samples
were hand ground to approximately 1.5 mrn thickness. Samples of conical bone were
approximately 5 1 mm x 6 rnm and with a mass of 1.2 g (k 0.2 3). Reference blood
samples and kidneys were digested in 10 rnl of4 M KOH at 60°C for 24 h. Bone samples
were decalcified in 10 ml HN03 for 3 d at room temperature and then digested in 10 ml 4
M KOH at 60°C for 2 d. The digested tissues were vortexed and then filtered through 8
pm filters (25 mm, Nucleopore Track Etch Membranes, Whatman, USA) fitted to a
vacuum filtration chamber. The vials and filters were rinsed with 2% Tween solution,
100% alcohol and distilled water to avoid loss of microspheres. All filters were allowed to dry overnight before counting. Microsphere counts for the reference blood samples and kidneys were quantified by their dye content.
The dye was recovered from the microspheres by adding 500 p1 DMF (toluene- dimethylformamide) as a solvent. The filters and DMF were placed in conical centrifuge tubes and vortex mixed for 30 s. Afier the tubes were centrifuged the dye solution was transferred to 10 mm glass cuvettes where the number of microspheres per sample were determined by spectrophotometry using absorbance values provided by the manufacturer for the microsphere lot number. The amount of dye or the number of CMs in a given sample was adjusted to achieve absorbance values of no more than 1.3 AU (absorbance unit. 1 AU = - I g [ 10% light transmittance/ IOOYO])to ensure linearity between absorbance and dye concentration according to the Lambert-Beer law (63). If sample absorbance was greater than 1.3 AU. the samples were diluted with DMF and counted again. The lower limit of detection that can be reliably distinguished from background noise is 0.05 AU
(63). Counts for bone and periosteum were determined by directly counting the microspheres with an epifluorescent microscope. Blood flow of each sample was determined by (6):
where Qt = tissue blood flow Qr = reference blood sample withdrawal rate (3 ml/min) CMr = number of microspheres in reference sample CM, = number of microspheres in tissue sample normalized to 100 g of tissue
Standardized blood flow (mI/min/lOOg) values were determined by relating tissue to reference blood microsphere counts and then normalizing to the sample mass. In addition to the mean pixel value fir each scanned image, standard deviations of the pehsion images were recorded to represent the heterogeneity of bone blood flow.
3.4.5 Statistictd hfetlrotis
Normality of the data was determined by a one sample KoImogorov-Srnirnov test.
Differences between right and left perhion values. pre- and post-drug infusion perhsion and pre- and post-cannulation pehsion values were compared using Student's paired t- tests. Coeficietlts of variation were determined to estimate differences between repeated measures of LDI output signals. Correlation between red and NIR wavelength output was determined by linear regression. Correlation between LDI perfirsion values and bone blood tlow measured by the CM technique was analyzed using a nonparametric two- tailed Spearman's rank correlation. A level of p < 0.05 was used to detect significant ditl'erences. RESULTS
4.1 Laser Doppler Perfusion Imaging Feasibility Study
4. I. I Rooster hfootiel
LDI detected changes in bone blood perfusion. Occlusion of the femoral artery for
5 rnin signi ticantly decreased flux values by 93.1 + 0.0% (from 5.1 k 1 .S PU to 0.4 k 0.3
PU)(mean r standard deviation) (p < 0.05). Recovery flux values (4.4 ? 0.6 PU) were not signiticantly different but remained lower than normal control values (p > 0.05) (Figure
1.1). A reactive hyperaemia was not detected.
control occ i udcd rccovcry
Figure 4.1 Mean (+ standard deviation) flux (perfhion units) of the mid-diaphyseal
Cortex of rooster tarsometatarsus. Occlusion values were significantly
lower than normal control (p < 0.05; denoted by an asterisk *).
Pharmacological intervention was performed to alter blood flow in the TMT.
Following infusion of each drug into the femoral artery, scans were performed at 7 and 5
min after injection. The greatest change in flux readings for each application of acetylcholine (Ach) and adrenaline (Adr) occurred at 2 and 5 min, respectively after injection. Close intra-arterial infusion of Ach and Adr increased (Ach) and decreased
(Adr) bone perfusion compared to normal control (Ach 1 1.6 5 19.2 %, p > 0.05; Adr
-25.1 k 34.1 %. p > 0.05) (Figure 4.3).
Figure 1.3 Comparison of pehsion values between pre- and post-application of .Ach
or Adr (% change k standard deviation). Change in perfusion was not
significantly different from normal control for either Ach or Adr.
I Rflbbitmfiel
Laser Doppler perfusion imaging detected changes in blood pefision in conical bone. Occlusion (5 min) of the femoral artery significantly decreased flux values by 6996
(mean k standard deviation) from 29.8 k 11.4 to 10.4 + 8.0 pehsion units (p < 0.05. n =
10) (Figure 4.3). Following release of the femoral artery, flux values were substantially increased by 45 s post-occlusion. Flux values at 45 s (25.0 + 8.9 PU) and 3 min (25.5 f 9.4 PLI) post-occlusion were not significantly different. Flux values at 3 min of recovery were significantly higher than those during occlusion (p < 0.05, n = 10) but remained lower than control values (p < 0.05. n = 10). The time required for flux values to return to normal. following occlusion. was not determined.
Figure 4.3
4.2 Bone Optical Properties Experiments
2 1 Thin Section Transmittance
The mean ln(I/I,) (i standard deviation) for each thickness, 55. 70. 85. 125. 140. and 155 pm, for red wavelength, was -1.4 + 0.1. -1.8 k 0.1, -2.1 k 0.3, -3. i i 0.1. -3.6 ?
0.2, and -3.7 + 0. I. Sample variation was relatively small. and the coefficients of variation for the six thicknesses (5 5, 70, 85, 125, 140, and 155 pm) were 0.07, 0.06, 0.17,
0.02, 0.04. and 0.04.
NIR red
Thickness (pm)
Fipure 4.4 Relation of light transmitted through thin sections of cortical bone at
different thicknesses for near infrared and red wavelengths.
Linear regression analysis determined a slope of -0.02 and y-intercept of-0.06
(Figure 1.4). The slope of the line was equal to the total attenuation coefticient.
Following Beer's law (12 1). the pathlength of a typical photon. of red wavelength. in
cortical bone was 4 1.2 pm (r = 0.98).
The mean In(V1,) (2standard deviation) for each thickness, 55, 70, 85, 135. 140,
and 1 55 pm. for near infrared wavelength. was - 1.4 k 0.0, - 1.7 t 0.1. -2.2 i 0.1. -2.4 f
0.1, -2.8 t 0.0, and -3.0 iO. 1. The coefficients of variation for the six thicknesses (55, 70,
85, 125, 140, and 155 pm) were 0.03, 0.03, 0.07, 0.03, 0.01, and 0.03. The total
attenuation coefficient was determined from the slope of the relation (slope = -0.0; y- intercept = -0.75) between ln(V1,) and sample thickness (Figure 4.4). The mean free path of a photon of NLR wavelength, therefore. was 70.9 pm (r = 0.97).
The total effective attenuation coefficient was determined from the slope of the
relation between in(UI,) and sample thickness (Figure 4.5).The effective penetration depths of red and near infrared wavelengths. therefore. were 0.8 and 0.9 mm.
. .. NIR G red
Thickness (prn)
bone samples at varying sample thicknesses.
4.3 Flow Model Study
The range of which the LDI measured flow was approximately 250 to 750 pVmin.
indicated by the linear region on a plot of LDI output versus flow rate (Figure 4.6). 0 200 400 600 800 1000 Flow (pl/min)
Figure 4.6 Comparison of known flow rates (pl/min) to laser Doppler pehsion
imaging output (perfusion units)
The smallest change in flow through the model that was detected by the LDI was
15 ? 5 pYmin. In the bone flow model. LDI detected flow at a depth of 1.5 mm below the surt'ace of the bone.
4.3.2 Sprculnr Reflectance
A significant difference between pefision values, within the linear re~ionof the
data points, with the flow model normal to the laser source compared to on an incline. was detected whether the fluid of interest was milk or 1% heparanized blood (Figure 4.7). 0 200 400 600 800 1000 Flow (pl/min)
Figure 4.7 Comparison of LDI output (perfusion units). with 196 rabbit blood in a
heparanized saline solution. in the flow model laying normal to the laser
source and on an incline.
The slope of the linear region of the milk curve was 0.15 PU/pVmin with a y- intercept of 2.87 PU when the model was laying flat and 0.66 PU/pVmin and -3.74 PU when the model was inclined. Linear regression of the linear region of 1% blood in a heparanized saline solution with the model normal to the laser source gave a slope of 0.07
PUfpVmin and y-intercept of -5.74 PU. When the model was inclined. the slope and y- intercept were 0.19 PU/pVmin and -22.3 PU, respectively. Comparison between the tlux and DC values for the model normal to the LDI (for both milk and 1% blood) indicated that flux and DC were related whereas the values were independent when the model was inclined (Figure 4.8a and Figure 4.8b).
Flus
O DC
o ~oo oo 00 soo looo o ZOO 400 c,oo soo loor)
Flou (pVtrtin) Flow pl/mm 1
Figure 1.8 Comparison of LDI flux and DC output for 1% blood in heparanized
saline at known tlow rates when the model was (a) normal to the laser
source and (b) at a 7' incline.
4.4 Laser Doppler Perfusion Imaging Utility Study
Sample images generated by LDI using red and NIR wavelengths before drug injection, immediately following a 3 min infusion of the drug (MR wavelength), and post
monem (NIR wavelength) are depicted in Figure 4.9. LDI detected heterogeneity in the tissue (Figure 4.10). One rabbit was removed from the study after examining the tibia1
cross section due to inconsistencies in the bone. One sample of one rabbit was also lost during the filtering stage ofthe microsphere testing and thus left out of the analysis.
Perfusion values of the control animals prior to and following cannulation of the left ventricle were significantly different. Cannulation resulted in a 17 + 14 % change in pehsion values (range 4 to 32% change). Injection of nitroprusside and noradrenaline resulted in approximately a 30-10% decrease or 20-25% increase. respectively. in blood pressure. Both drugs significantly changed perhsion values when compared to baseline.
4 4 I Repeatability
Consecutive scans. in sets of two. were performed on each bone with a time difference of 30 s between scan 1-2 and 3-4 (time required to perform a scan) and approximately 110 s between scans 2 and 3 (time required to reposition the LDI over the other leg). The four LDI pehsion measurements of each bone resulted in an average coefficient of variation of 0.05 (range 0.0 1-0.14) (Table 4.1). No significant differences existed between mean pefision values of the right and let? limbs measured with red or
NIR wavelengths. Table 4.1 Consecutive laser Doppler perfusion imaging measurements at near
infrared wavelength.
Rabbit Side Scan 1 Scan 2 Scan 3 Scan 4 Mean SD Coeft'icient
of Variation
Right Let? Right Left Right Left Right Left Right Left Right Left Right Lett Right Left Right Left Right Lett Right Left Right Left
4.4.2 Reti versus Ncar Infareti Wavelengtlt
Mean perfusion values of each limb at red and NIR wavelengths were normally
distributed. The normal (pre-drug application) pehsion values for control. nitroprusside.
and noradrenaline at red wavelength were 16.4 k 5.6 PU, 22.7 + 9.8 PU. and 17.1 k 6.7
PU. At NIR wavelength the perhsion values were 28.3 + 10.1 PU,36.6 i 1 1.3 PU,and 30.3 i 11.3 PU (Tabie 4.2). Figure 4.11 compares the mean LDI output signal from the red and NRlasers. A slope of 1.31 PUPU and y-intercept of 7.28 PU were determined (r
= 0.93, p < 0.05).
i 0 0 10 20 30 40 50 Red mean output (PU)
Figure 4.1 1 Relation of NIR versus red wavelength mean output. LDI output of the
two wavelengths were significantly correlated (r = 0.93: p < 0.05).
Figure 4.12 compares the average standard deviation in the images, as measured by the
LDI, from the red and NIR lasers. A slope and y-intercept of 1.18 PU/PU and 5.71 PC;.
respectively, were determined (r = 0.94, p < 0.05). Red std output (PU)
Figure 4.12 Relation of NIR versus red wavelen~thstandard deviation output. LDI
output of the two wavelengths were significantly correlated (r = 0.94: p <
1.4.3 Comparison of Laser Doppler Perfusion Imaging md Cotoured
Micros p heres
Average blood flow measured by coloured microspheres for control (normal). nitroprusside (dilatation), and noradrenaline (constriction) groups were 0.68 I0.16 rnl/min/100 g, 0.5 1 k 0.36 ml/min/100 g, and 0.43 k 0.37 rnVminil00 g, respectively. .At
NIR wavelength the LDI perfusion values were 24.5 + 13.0 PU (control). 14.1 k 3.7 PU
(nitroprusside). and 16.9 k 4.4 PU (noradrenaline) (Table 4.2). There were no significant differences in CM determined biood flow between control, nitroprusside. and noradrenaline groups. The average number of microspheres per bone sample in control. nitroprusside, and noradrenaline groups were 70.8 k 23.2, 41.2 + 23.3. and 35.9 k 26.7. respectively.
Table 4.2. Concurrent tibia1 blood flow measurements with microspheres and laser
Doppler perfusion imaging at near infrared wavelength.
Rabbit Side Group Microspheres LDI (ml/min/ 1009) (pefision units)
Right Control Left Right Control Left Right Nitroprusside Left Right Nitroprusside Left Right Noradrenaline Left Right Noradrenaline Left Right Control Left Right Nitroprusside Left Right Nitroprusside Left Right Control Left Right Noradrenali ne Left Right Noradrenaline Left A significant correlation (r = 0.58, p < 0.05) was detected between LDI pefision values and CM determined blood tlow (Figure 4.13). All data points were linearly related
(p < 0.05).
0 7 I f I 0.0 0.5 I.O 1.5
Standardized blood flow ( ml/min/ 100g)
Figure 4.13 Relation between mean LDI output (pehsion units) and standardized
blood flow (measured with coloured microspheres). A significant
correlation between LDI and CM determined flow was found DISCUSSION
The current study assessed the utility of laser Doppler perfbsion imaging to detect changes in cortical bone perhsion. Comparison between LDI and a standard measure of blood flow in bone and its sensitivity and repeatability were studied to determine the feasibility of using LDI to measure blood flow in bone. The limitations of LDI were also investigated.
A series of studies were performed to understand the limitations and utility of LDI to measure blood flow in cortical bone (see Figure 2.9):
1 ) Pilot study to test LDI on bone
2) Study to test hypothesis
3) Study to determine the optical properties of bone
i Thin section transmission (mean free path)
ii) Thick section transmission (effective penetration depth)
4) Study of LD I limitations (flow models)
i) Detectable range of flows
ii) Effects of specular reflection
5) Study of LDI to measure perhsion in bone
i) Repeatability
ii) Red vs. near infiared wavelength
iii) Comparison with another technique
Initially two studies were performed; the first, a pilot study, determined if the preliminary results warranted further study, and the second study tested our hypothesis that LDI could measure statistical changes in bone perfision. Results of these two studies
suggested that LDI had potential as a technique to investigate changes in bone blood
flow.
Optical properties of tissue affect the output of optical techniques. however. and. therefore, experiments were performed to assist in interpretation of LDI output. The
laser-tissue interaction experiments helped to better define the ability of LDI to measure
perhion in conical bone and sive insight into the potential use of LDI on hard tissues.
From the optical properties of bone. the depth to which LDI measured blood flow in
conical bone and the contribution of the subsequent layers of bone to the output signal
were defined. The limitations of LDI were studied using flow models where LDI output
could be correlated to known tlow rates. The sensitivity of LDI. to determine if LDI was
capable of measuring physiological flows. and the etTect of specular retlection. an
unwanted artifact inherent with smooth tissue. were studied. With more understanding of
the ability and limitations of LDI. we studied its repeatability. dual wavelength
capabilities. and validated LDI in the measurement of cortical bone perfusion by
comparison with coloured microsphere-determined flow.
LDI detected large changes in bone blood perfusion. achieved by occlusion of the
femoral artery. as well as more physiological changes induced by the injection of
vasoactive drugs. The pilot studies suggested LDI was capable of detecting and
measuring a range of blood flows in cortical bone. It is difficult to measure accurately the
absolute blood flow in bone, but the flow model demonstrated the range of flow that was
detected by LDI was within the physiological range. Images produced by LDI software
demonstrated that LDI could display heterogeneous profiles of pehsion in cortical bone. The majority of the output signal was From contributions of the upper layers of bone
(approximately outer 5 % of diaphysis). Although near infrared wavelength light traveled
40% hrther into bone than red wavelength light before a scattering or absorption event.
the significant linear correlation between LDI output for the two wavelengths suggested
that the two wavelengths measured flow in the same region of bone. This finding was
supported by the similarity of the penetration depths for red (0.8 mm) and near infrared
(0.9 mm)wavelengths. Although specular reflection altered LDI output. our
compensation for the anifact resulted in consecutive LDI measurements of the same area
of bone repeatable within 5%. Determination of coloured microsphere flow required
processing a volume of bone. and knowledge of the penetration depth of near infrared
wavelength of the LDI permitted an approximate comparison of the two methods. LDI
output was si~nificantlyand positively correlated to coloured microsphere-determined
flow. Thus. the LDI output signal appeared to be a good estimator of microvascular
perfusion. Its ability to detect spatial variations in the microvascular bed. con-elation with
another method, and repeatability suggested that the technique may have utility for
cortical bone. The results of this study susgested LDI was sufficiently reproducible and
sensitive to detect physiological changes in bone perfusion and, thus. had promise as a
tool to measure changes in bone pehsion in the outer layers of cortical bone.
Through our pilot studies, we ascertained that LDI was capable of measuring
perfusion in hard tissues. In the rooster model, close intra-arterial infiision was an
effective method of altering tarsometatarsus (TMT) blood flow because it allowed for
local administration of drugs into the bone microcirculation. LDI detected large changes
in blood flow resulting from occlusion of the femoral artery acd smaller changes more physiological in magnitude from the infusion of acetylcholine and adrenaline, both. which exist naturally in the body. Pefision indices were above zero following femoral occlusion, likely due to tributary vessels of the Femoral artery supplying the TMT (1 10).
Bone blood flow can decrease over time in an anesthetized animal without surgical intervention (20). Nonetheless. similar changes were detected in animals subjected to a short experimental protocol (30 rnin) compared to longer protocols (90 rnin) that suggested that the changes observed were a result of the intervention.
Similar results in the rabbit model confirmed that LDI was capable of detecting
perfusion in cortical bone. During the three time intervals for measurement following
release of the femoral artery after 5 rnin occlusion (45.90 and 370 s) progressive
increases in tltix were measured. At 270 s, tlux was near normal values but remained
marginally depressed. A reactive hyperaemic response within 1 rnin following occlusion
of the femoral artery for 6 and 16 min. as seen by others (108). was not observed.
however. suggesting that the period to scan the tibia was too long to differentiate the
response. In the time required for LDI to scan the area of bone. the hyperaemic effect
may have also been averaged out. Perfusion values detected by LDF after occlusion
returned to baseline after 4 min. LDF has greater temporal resolution than LD I in that
LDF can continuously monitor blood flow at a single point. whereas LDI measurement of
an area over time requires subsequent rapidly repeated scans. Differences in output
between this study and a similar LDF study (108) may have been due to the difference in
temporal resolution of each technique, contribution of tributary vessels of the femoral
artery, and in part. response of the rabbit to the anesthetic reagents. Although our pilot studies used LDI to assess bone in an experimental setting, questions arose regarding its capabilities and limitations. Transmission experiments to obtain the optical properties of cortical bone described the contribution of photons to the output and the depth to which LDI evaluated flow. The regression line calculated from plotting -In 7, (collimated transmission) versus bone thickness in the thin section transmission experiment gave a straight line indicating that a single scattering condition was valid. If a deviation tiom linearity had occurred with increasing thicknesses. it would have been an indication of multiple scattering photons. Use of a stronger detector could have reduced the problem of deviation, but the problem can not be completely eliminated because there are always photons coavial after multiple scattering ( 123). Use of a collimated beam reduced the deviation but there was a limit to collimation because the signal decreased as collimation increased ( 123). The fitted line also did not pass through the origin. which is an experimental artifact from imperfect surface contact between the tissue and glass slide that causes additional loss of light from surface scattering by the sample (29.122). Similar to the thin sections. the scatter in the thick section data was likely due to the anisotropy of bone and tissue handling and preparation.
The output of LDI was dependent on fluctuations in laser output and tissue retlectivity (6). Specular reflection was an undesirable artifact that can result in saturation of the photodiode. The effect of specular reflection on the output of LDI, therefore. was studied using a flow model. MoorLDI image analysis software attempted to compensate for the changing output and reflection by normalizing the output signal with respect to the intensity of the total reflected light (mean of DC photo image). The flux output of LDI (in pehsion units) was therefore dependent on the intensity of DC.The difference in slopes of LDI output versus flow between the model normal to and on an incline relative to the laser was due to the different contributions of reflected light (DC). When the model was flat, the intensity of all measured light (DC) was considerably greater due to the measured re tlection. which resulted in a decrease in tlux output. There was less of a reflection component in the inclined model as seen in the larger perfhion indices. Large contributions from specular reflection otT the tissue surface. therefore. resulted in uncharacteristic low perfusion measurements.
With knowledge of LDI penetration depths. an appropriate comparison of LDI output to a standard measure of blood flow determination, coloured microspheres. was performed to determine the accuracy of LD I for measuring perfusion in bone. Studies have demonstrated LDl's capability of measuring perfusion in soft tissues. displaying heterogeneous tlow. and detecting different degrees of perksion. Although no previous studies had measured perfusion in bone with LDI. the current study established its etljcacy in measuring bone blood tlow by significant correlation with coloured microsphere-determined tlow (r = 0.58. p < 0.05). Similar correlation between blood flow measured with LDI and the microsphere technique was found in the rabbit knee li=cyament
(r = 0.76, p < 0.001) (6).
The periosteum was not removed in the current study as cortical bone perhsion was shown to be significantly reduced (up to 20%) after periosteai stripping of the entire length of the tibia (64). Removal of the soft tissue covering the tibia may have resulted in some damage to the periosteurn, but, grossly, the periosteum appeared intact. The periosteum was also used to reduce specular reflection of the exposed bone. Longer wavelengths. such as near infrared (NIR). are advantageous to use because they have better penetration than shorter wavelengths in highly pigmented
tissues. In stratified tissues such as skin. however, different wavelengths may be able to
measure different regions of flow ( 1.5 1 ). Differences in red and near infrared responses
can be attributed to differences in photon absorption at the two wavelengths associated
with changes in blood volume and oxygenation in the deeper tissue. Despite the shon
mean free path between interactions in bone. red and NIR light penetrated deeply in
tissue due to a larger number of scattering than absorption events and the highly fonvard-
directed scatter that allowed photons to continue to penetrate the tissue even after
multiple scatterings ( 124).
NIR
Figure 5.1 Laser-tissue interaction at different wavelengths. (MFP = mean free path)
LDI has the capacity to measure pefision with two wavelengths. Comparison
between red and N[R wavelength LDI output suggested that MR was capable of
detecting slightly deeper flows, and, thus, gave a higher weighting to blood flow deeper
in bone (Figure 5.1). The significant correlation between the means of the output images, which represented the average flow. suggested either the two wavelengths were measuring flow in similar regions of bone or that blood flow was similar in the region sampled by NIR compared to that sampled by red wavelength (30). The correlation between the standard deviations of the images. which represented the degree of heterogeneity of the tissue suggested that either the heterogeneous structure in the upper layer of bone measured by red wavelength was highly correlated to the lower layer measured by NIR or the two wavelengths were sampling flow in the same volume of tissue (30).
Although measurement by the photodiode was cumulative. perfusion in the upper
layers of bone was dominant in the output signal. as determined by thin section transmission experiments (Figure 5 1 ). The number of photons were attenuated exponentially as they penetrated deeper in the tissue. .As the photons penetrated deeper
into the tissue. the probability of being absorbed increased (attenuation). The chance that
a deeper penetrating photon would make it back to the tissue surface and strike the
photodiode was low and therefore it was likely that red and NIR wavelengths were sampling similar perfusion regions. The deviations between red and NIR LDI output
readings could have been due to random instrument noise. and, perhaps, a slightly larger
number of NIR photons penetrating deeper in the tissue.
Even after subtraction of the "biological zero" from the mean perhsion values,
the non-zero y intercept From the comparison of microsphere-determined flow and LDI
output was detected as in previous studies (30). The intercept was likely a combined
effect of the measurement process and the instrument's inherent offset (30). Laser
Doppler systems often have an inherent signal offset that would result in a positive value for a no-flow condition (6). Because of the photon direction insensitivity and diffisive
light scatterings in laser Doppler technologies, these instruments give an output signal higher than electronic zero when the net blood flow in tissue is apparently zero (129).
This "biological zero" is caused by the photoelectronic white noises of the instrument.
The Brownian motion of particles. random wandering movement in a no-flow situation.
although not related to organized blood tlow, may also be detected by LDL and affect
perfusion values. Due to the diffusive light scattering in tissue. the random movement of
blood cells still senerates an output signal ( 129). Subtraction of the "biological zero"
should have eliminated these two effects suggesting averaging LDI images may have
resulted in a statistical artifact. The signal processor of the LDI gave a linear output for a
limited range of tlows. but the Frequencies outside of this range were filtered out (6). The
curve for low flows may be nonlinear. and. thus. the true relation may contain higher
order terms that would force the curve to pass through the origin (30). The comparison
between LDI and CM may have also accounted for the non-zero y-intercept. because
neither method is a "true" measure of tlow nor accurate for measuring extremely low
flows.
Laser Doppler flowmetry is a recognized technique for measuring temporal
variations in blood flow and monitoring dynamic responses to physiological stimuli at a
specific site, but LDI offers the ability to study regional variations in tissue blood flow.
image tissue heterogeneity, and measure average tissue blood flow. LDI offers several
advantages over LDF; because of the heterogeneity of small vessel distribution in
vascularized tissue, measurements made by laser Doppler flowmeters are site-specific
(104). The necessity of contact between the probe and tissue may also introduce the risk of infection and induce a motion artifact in the LDF signal. Although a linear relation was attained between LDF output and controlled perfbsion rates to an isolated tissue. the
slope of the relation varied between animals (99) and that suggested the results could not
be compared between different animals. Smits and colleagues showed regional variation
in tissue blood flow, and that by averaging 6 to 8 points across the tissue surface. LDF
was linearly reiated to radioactive microsphere-determined blood flow ( 104). This
method, although a better assessment of tissue blood flow than a single point. was subject
to temporal variations in tissue blood flow over the period of time required to move the
probe to obtain sufficient data points to represent accurately the entire tissue (6).
Laser Doppler technologies have been compared to microsphere-determined
tlows. Swointkowski and colleagues did not tind significant correlation between the
estimation of bone blood flow in rabbits by the microsphere method and LDF (1 10).
Lausten and co-workers. however. were able to obtain significant correlation between
LDF output and flow measured by microspheres (69). The discrepancy in experimental
results comparing LDF and microsphere flows may have been due to the area in which
the average bone blood Row was measured by the microsphere technique was
considerably larger than the area measured by LDF (69). Flow determined by the
microsphere technique was also for one specific time point whereas LDF provided
continuous flow measurement. The significant correlation. in the current study, detected
between LDI and the microsphere technique may have been due to the ability of both
techniques to study the same area of interest. The significant correlation between LDI and
CM also suggested that the thickness of the tissue samples used for microsphere-
determined flow was similar to the volume scanned by the NIR laser of the LDI. The average thickness ofthe samples, 1.5 mm, was similar to the penetration depth of 2.3 mm
in conical bone, reported in previous studies (84). Similar to LDF where LDI measures
flow over a period of time, comparison to microsphere-derived flow at one specific time
point may have accounted for some for the differences between the results of bone blood
flow measurements From the two techniques.
'The reproducibility of multiple measurements of the same area. including
realignment of the LDI after scanning the opposite leg, was within 5% of the other scans.
Although the coefficient of variation varied between 0.0 1 and 0.14 (mean = 0.05). the
reproducibility of LD I scans was higher than multiple LDF measurements in the same
position in cancellous bone (mean coetxcient of variation between 0.05 and 0.15)
(69.1 10). Due to the spatial variation in cancellous bone. however. the coefficient of
variation at four different positions was between 0.09 and 0.2 1, and although not
significantly different. single point measurements were not representative of an entire
tissue (68). Since LDI did not require contact with the tissue. positioning of the LDI
remained exactly the same between measurements and animals that increased the
accuracy and reproducibility of LDI output.
From the current studies it was shown LDI had sufficient sensitivity to detect
relative physiological magnitude changes in blood flow. Sensitivity of the LDI was
evaluated by creating changes in perhsion of physiological magnitude by infusion of
naturally occurring hormones. Given hormones may elicit a different response
systemically and in soft tissues compared to the reaction in bone (43). The unusually stiff
extravascular space may explain why vasodilators are generally less effective than
vasoconstrictors in bone, and why some agents that normally act as vasodilators in other tissues increase the vascular fluid resistance in bone (9.43). Results indicated that vasodilators increased interstitial pressure in bone that potentially decreased flow through the organ even in the presence of arterial dilatation, through an increase of flow in the venous drainage system (43).
Laser Doppler techniques can only provide a relative estimate of microvascular perfusion. and although it may be difficult to calibrate the signal into absolute values because of the variation in optical properties of various tissues and spatial variation in the microvascular bed (69). previous studies suaested it may be possible to derive calibration Factors for a specific tissue from LDI output since the effect of tissue heterogeneity is reduced by the averaging process (6). The slopes derived from the tlow model normal to the laser source and on an incline were different due to the different contributions of specular retlection. Because tissues have varying reflective properties. the output of the LDI would be different thereby making it necessary for difierent calibration factors for different tissues. Although LDI detected temporal variations in bone blood flow. absolute blood flow measurement is not possible until there is a better understanding of laser-tissue interaction and spatial variation in the microvascular bed
(30.69). Knowledge of these cornpiex interactions will help determine the penetration depth and effect of laser wavelengths on LDI measurements (30).
To better quantify the output of laser Doppler technologies, it is necessary to know the pathlength traveled by light through tissue as it gives an indication of which layer in tissue contributes to the majority of the signal. A large body of work exists on human tissue transillumination, but few experiments have reported on the optical properties of bone. Firbank and colleagues determined the scattering and absorption coefficients of the porcine skull by measurement of diffuse reflectance and transmittance of 2 mm thick samples (28). Linear interpolation between the measured values of reflectance and transmittance and values generated by a Monte Carlo simulation gave a more accurate estimate of the coefficients. The mean free path at 650 nm and 950 nm were 28.5 pm and 41.6 pm, respectively. which was similar to our thin section transmission results of 4 1 prn at 633 nm and ? 1 pm at 8 10 nm. Our coefficients were larger than those For the skull. possibly as a result of measurement of small residual multiple scatterings that may cause underestimation of the total attenuation coeficient
(pl)(29) and the inability of our experimental setup to capture all the exiting photons.
Comparatively. 633 nm light penetrated rabbit muscle to a depth of 20.8 - 62.5 pm. and
29.4 prn in bovine adipose tissue ( 15). A broad range of values exist for any tissue due to variation in samples. detection apparatus. boundary conditions. and the governing light propa_ration model ( i 5).
X limited amount of data exists on the optical properties of bone: studies however. used different experimental setups. which make comparison of output data difficult. The penetration depth reported by Notzli and colleagues was comparable to our effective penetration depths from the thick transmission experiment (84). Although their penetration depth was 2.28 mm, compared to 0.8 1 mm for red wavelength light, their samples were fresh, soaked in a 0.9% NaCl solution except during experimentation. and both ends of their bone samples were ground that may influence more scattering into the tissue than an unground sample. The threshold thickness of their cortical bone samples suggested that flow could still be detected by LDF at a depth of 2.90 rnm (84). In cartilage-covered trabecular bone samples, however, mean threshold thickness values ranged between less than 2 mm (33) and 3.45 mm (84) suggesting that techniques to determine optical properties applicable to laser Doppler techniques must be further
refined. Experiments that determine the maximal depth at which flow can be detected are
important for comparison with techniques such as microspheres as it gives a limit to the
volume of tissue being studied.
The attenuation of light was also dependent on the mineral content. water partial density of bone (84). and bone mineral density (I I 1) No work has been reported that
distinguished trabecular and cortical bone. but Takeuchi and colleagues demonstrated
that the optical properties of the two types of bone display different characteristics ( 1 1 1 ).
Using time resolved spectroscopy systems with a near-infrared pulsed laser. time
response waveforms were recorded for bovine cortical and trabecular bone and skeletal
muscle. Trabecular bone exhibited high transmittance and greater scattering than muscle.
whereas cortical bone showed very little transmittance and extremely strong scattering.
Litrrittrtiutts
LDI has limitations inherent to all laser Doppler perfirsion instrumentation. These
instruments can only provide a relative index of blood flow. although, as discussed
previously, calibration factors may be possible for a specific tissue group. Absolute
measurement by these techniques will not be possible until accurate measurements of the
penetration depths of tissues iri ~ivoare obtained. The temporal resolution of LDI was
also dependent on the user set scan speed (fastest 4 ms/pixel) and size of the area to be
scanned such that physiological phenomena such as reactive hyperaernia may not be
detected. Specular reflection off the tissue surface affected imaging technologies. such as
LDI and laser speckle, by saturating the photodiode. It was detectable on the displayed
LDI images by loss of the colour-coded heterogeneity and uniform patches of white. The
LDI can eliminate specular reflection from the tissue by adjusting gain settings that control the amplification of non-Doppler shifted light (DC gain) and Doppler shifted light
(flux gain) (30). Bone. however, was highly reflective, and even at the lowest gain settings specular reflection was still a component of the output. To attempt to compensate for specular reflection. the head of the LDI was angled slightly during the experiments to direct specular retlection away from the photodiode, but it may have still influenced LDI output resulting in a reduced correlation between LDI output and microsphere- determined tlow.
Output of laser Doppler technologies are also dependent on the angle of incidence of the laser beam. Inconsistencies appeared when the laser beam contacted the tissue surface at an an~leother than 90" such that different angles resulted in different flux values. If the laser beam entered the tissue on an angle. a smaller portion of light was able to penetrate the tissue thereby lowering the accuracy. Similarly. if the angle of the LDF probe was greater than 45". all of the Doppler shifted light was not received by the etl'erent optical fibres that resulted in a falsely low output signal ( 1 10).
A further limitation of this study was the use of microspheres to determine tlow in a relatively low flow tissue. Although it has been susgested that to obtain an accurate estimate of bone blood flow each sample should contain 150-250 radioactive microspheres (70). the coloured microsphere-determined flows in this study were similar to previous studies of the tibia in rabbits (1 15.116). Li and colleagues also determined that there was a 14% error in microsphere-determined flow if the bone sample contained less than 50 microspheres but only a 7 to 9% error in samples that contained 100 to 150 microspheres (701. The radioactive microsphere technique determined the number of microspheres within a sample indirectly by counting gamma radioactivity (similar to our determination of flow in the reference blood samples). The lower radioactivity levels. associated with low flow tissues. were subject to more error from background activity. physical decay, and cross talk during the counting procedure (63). The coloured microsphere technique, however. more accurately determined the number of microspheres in a low blood tlow sample since the microspheres in each sample were directly counted. Gross and colleagues showed less than 256 of the microspheres trapped in the bone samples were lost during processing of the samples (39). Although some bone samples in the current study contained a smaii amount of microspheres. and. thus. inherent error in the microsphere-determined blood flow. significant correlation was still observed with LDI output. The lack of significant difference between microsphere determined flows for the control. nitroprusside. and noradrenaline rabbits again was likely attributed to the inaccuracy of the method for low flows.
The influence of tissue on light as it passes through the tissue is key to understanding non-invasive measurements of physiological variables such as microcirculation and tissue metabolism. Interaction between laser light and bone was complex, but basic knowledge of the characteristics of photon propagation allowed hrther interpretation of output from techniques such as LDI. The mean free paths of red and near infrared wavelengths were estimated by direct measurement of bone's optical properties using a thin section transmission experiment. The greatest limitation of this method was that the use of thin samples may have changed the optical properties from their ir~vivo values by use of post mortem samples and grinding and/or freezing of the tissues (127). This technique required that tissue sample thickness had to be considerably less than the inverse of the scattering coefficient so multiple scatterings were negligible.
For example, the range of optical scattering coefficieilts for soft tissues was in the range of 100 to lo00 cm", thus, the sample thickness had to be less than 10 pm to negate the contribution of multiple scatterings in the output (123). In bone, however, it was difficult to prepare samples thinner than 50 - 60 pm because the hardness and porosity of the bone
resulted in a tendency for the samples to crumble (28). Multiple scattering started to afFect measurements for tissues that were thicker than 10% of the scattering mean free
path. Ln chicken muscle, for sample thicknesses 500 to 1000 pm, multiple scattered photons were detected and resulted in an underestimation of the total attenuation coefficient (29). In samples less than 100 pm thick. however, the multiple scattered photons did not seriously affect the measurements (29).
In preparation of the samples, since the thicknesses were so small. the tissue
handling and preparation (grinding or cutting) potentially altered the optical properties of the sample (127). Evidence suggested that the use of post rnortem samples (125) and
freezing of the samples (9 1) may have altered the optical properties from irr vjvo values
although the changes were relatively small above 600 nm (127). The effective attenuation
coefficient changed by at least a factor of two I hour post mortem Irr sitrr (125). Freezing
and thawing altered the transmission of light through thick samples (91), but Flock and
colleagues suggested there was no substantial difference between their total attenuation
coefficient determined by thin section transmission experiments and in vivo results (29). Their direct measurements were within a factor of two of the results of indirect methods
where the tissues were not frozen or ground (126). The total attenuation coefficient of
fresh unground samples was also 213 the value of frozen. ground tissue. Samples OFthin
section transmission experiments required mounting making it difficult to ensure the
tissue surface was optically smooth that gave the opportunity of surface scattering
artifacts to contribute spuriously to the signal. Interaction probability in the sample was
also low that resulted in a very weakly transmitted signal that could have been masked by
unavoidable tluctuations in incident light or ambient light (123). In addition to the
preparation induced errors (freezing. grinding), thick sect ion transmissions were
subjected to multiple scattering and although the detector was placed close to the tissue.
photons likely escaped measurement that would result in an underestimation of the total
attenuation coefficient.
The difficulties associated with direct techniques (thin and thick section
transmission) of tissue optical propenies measurement has left investigators to derive
optical properties using indirect methods. In the indirect method. a model of light
propagation was used to derive the optical properties from measurements of bulk tissue
properties ( 127). Although indirect techniques for determining tissue optical properties
have the advantage of potentially being used in intact tissue and even irr ~~iw,direct
techniques are model independent. Models are limited because they generally predict
macroscopic conditions given the microscopic parameters making it necessary to iterate
between microscopic and macroscopic conditions to determine the best fit. Macroscopic
data are also restricted since each macroscopic parameter may be a combination of
microscopic factors. Direct and indirect techniques. however, are complementary techniques in determination of microscopic and macroscopic optical properties of tissue
(127).
Although use of LDI may be limited to measurements on and near the bone's surface. it could be beneficial for studying the influence of blood flow on bone remodeling where areal perfusion is of importance. For example. microvascular blood perfusion. measured by LDF, was highly variable in different regions of the osteonecrotic femoral head (68). The incorporation of bone grafts was largely dependent on the establishment of circulation, and LDF demonstrated a substantial increase in blood circulation in the grafts and lower flux values proximai and distal to the grafi (65). A measured increase in flux value in the callus confirmed that increased blood tlow was a part of graft incorporation and was closely related to the formation of new bone in the graft (44,6 1 ). The functional relation of bone blood tlow to fracture healing suggested increased bone forrnation was due to the increased perfusion of the extravascular space
(56). Temporal and spatial variations were documented in several tissues. including bone. and were thought to be caused by vasomotor regulation (44). Therefore. technologies that can measure flow over an area and demonstrate the regional distribution of pehsion will increase knowledge of bone vasculature and may be used to predict osteopathies and manipulation of bone flow to enhance bone healing and other pathologies (22).
It~~plic(~tiurrs
Fluid flow and blood flow, under the influence of external loading, have been considered as possible mechanisms involved in bone remodeling (45). The advantage of using laser Doppler techniques on cortical bone, therefore, is that a substantial component of bone remodeling occurs at the bone surface. In the rabbit, remodeling was mainly induced at the surface by cellular changes, and bone remodeling caused minimal intracortical changes so fluid control to deeper cells was not as significant as to surface cells (74). It has yet to be established, however. whether blood flow changes are a primary or secondary event to remodeling. as changes in flow due to external loading at a local level may be linked to material property and pathological changes (73).
It has been suggested that flow within vessels may determine the differentiation of bone cells by its control over environmental nutrition (52). Fluid filtration along capillary walls was produced by an increase in capillary blood pressure. but this was secondary 10 an increase in venous blood pressure (55). Kelly and colleagues postulated that the increase in venous pressure was a cause for flow of interstitial tluid (55). The movement of interstitial fluid in bone has been thought to result from the deformation of the stressed matrix of bone (45). The demonstration of interstitial tluid flow in cortical bone may lead to the mechanism that could explain the adaptive response of bone to mechanical loads and effects of venous hypertension on growth and fracture healins (75).
Increased interstitial fluid flow triggers new bone growth (83) and may be a major factor in remodeling and orientation of new tissue (36). Hyperemia preceded disuse-induced intraconical resorption. and thus. it was proposed that bone vasoreylation was potentially associated with the cellular events that lead to bone resorption (39). Dillaman and colleagues proposed a mechanism to describe the influence of bone pehsion on bone dynamics (22). The mechanism predicted that bone mass was limited by the ability of the vasculature to supply oxygen and nutrients within the mineralized matrix. The behavior of osteogenic cells may have been influenced by the chemical and physical nature of the medium with which they are pefised, and bone cells may have evolved mechanisms to respond to changes in composition of the medium and its rate of flow (22).
The findings of the current experiments suggested that laser Doppler perfitsion imasing is a promising tool for imaging 111 vivo changes in pehsion in the superficial layers of bone. Although this imaging method only quantified a relative index of flow. it may be suitable for profiling heterogeneous structure and sequential changes in blood tlow in bone adaptation situations such as Fracture healing, implant viability, and mechanically induced structural remodeling responses. CONCLUSION
Laser Doppler perfusion imaging effectively measured and detected physiological changes in cortical bone perfusion. LDI detected pefision in bone to an approximate depth of 9..5 mm (red wavelength laser) and 0.9 mm (NIR wavelength laser). Photons entering the tissue travelled on average 41 pm or 70 pm, depending on the wavelength. before a scattering or absorption event suggesting that the majority of the LDI output signal was from the contribution of tlow in the upper layers of bone. LDI output was
reproducible within 5% in two sets of two consecutive LDI scan measurements.
Significant positive correlation between output from the red and NIR lasers suggested the
two wavelengths measured perhsion in the same region of bone. LDI accurately
measured blood flow in cortical bone as it was significantly correlated to a standard
measure of blood tlow (coloured microspheres).
This study characterized laser Doppler perfision imaging for measuring pehsion
in bone. The ability of LDI to provide a two-dimensional areal image that can represent
average blood flow in tissue and depict a biologically relevant heterogeneous profile of
pehsion in tissue suggested that it may be an effective method for studying irr \?LW
dynamic changes in pehsion in bone. LDI is a promising tool for imaging changes in the
superficial layers of bone and may be suitable for studying sequential changes in blood
flow in bone adaptation situations such as fracture healing, implant viability, and
mechanically induced structural remodeling responses. Further development of methods
to measure perfusion in bone may increase the understanding of the role of blood flow in
mechanisms involved in bone remodeling. Testing of the optical properties of bone was only preliminary to try to determine the limitations of using LDI in bone. More rigorous and 111 vivo optical property testing must be completed to hlly determine the ability of LDI to measure perfusion in bone.
6.1 Future Work
The series of experiments in this study were the initial steps in quantitatively defining the ability of the laser Doppler perfusion imager to measure perfusion in bone.
The next step in probing LDI's utility will be the development of a quantitative method to define the optical properties of cancellous and conical bone irr vivo. Better characterization of the ill viva optical properties at red and NIR wavelengths would lead to knowledge of penetration and measurement depths in bone. More accurate estimates of the optical properties of bone may be assessed by using a combination of measurement of
thick section transmissions with two integrating spheres. bulk tissue measurements ill
vivo. and development of a model for bone. Based on the experiments presented. knowled~eof tri rti~*openetration depths will permit a better understanding and interpretation of LDI output.
Additional methods are needed to filter specular reflection from LD I
measurements. Since flux measurements are dependent on the contribution of the unscattered photons. a method that limits measurement to only photons that have entered the tissue will result in more accurate perfbsion measurements. Placement of a filter
before the photodiode may reduce the effect of the photons reflected off the surface of the tissue on the output of LDI.
Although LDI only quantifies a relative index of flow. its reproducibility suggests
that LDI output could eventually be calibrated into absolute units of flow for a given tissue. A single calibration factor is not possible, however, as the output is dependent on the optical properties of the individual tissue.
Knowledge is very limited of the optical properties of conical and cancellous bone. Further studies in this area would improve understanding of LDI output and potentially expand its use in research and clinical applications. REFERENCES
I. Arnfield MR, Mathew R, Tulip J, McPhee MS. Analysis of tissue optical
coefficients using an approximate equation valid for comparable absorption and
scattering. Phys hfrd Biol3 7(6): 12 1 9- 1230, 1992.
2. Arnoldi C. Vascular aspects of degenerative joint disorders. Acftr Ortl~up.(;Llt~~i
65 (Suppl. 26 1): 1-82. 1994.
3. Beme RM, Levy M. Cardiovascular Physiology, 7th ed. Mosby -Year Book. New
York, 1997.
4. Boyd H. Avascular necrosis of the head of the femur. In: Arbor .A. Edwards I
(eds.) Instructional Course Lectures, vol. 14. The American Academy of
Orthopaedic Surgeons, pp 196-204, 1957.
..F Brandt K. Insights into the natural history of osteoarthritis provided by the
cruciate-deficient dog. An animal model of osteoarthritis. Atar rWrlcntl Sci
732:199-205, 1994.
6. Bray R Forrester K. McDougall J. Damj i A. Ferrell W. Evaluation of laser
Doppler imaging to measure blood flow in knee ligaments of adult rabbits. dfed CC
Biol E11g K- C'umptrt 34227-23 1, 1996.
7. Bray RC, Doschak M. Tyberg J. Elevation of blood flow in an ACL-deficient
rabbit knee is associated with deterioration of MCL biornechanical properties.
Proc Cir,~Or~hop Rrs Soc (Wi~ulipeg,iVla~~ituba) :28-3 9, 1994.
8. Briers I. Laser Doppler and time-varying speckle: a reconciliation. J Opt Suc rltr~
13(2):345-350, 1996. Brinker MR, Lippton H, Cook SD, Hyman AL.Pharmacological regulation of the
circulation of bone. J Borre Jobil Surg 72A:964-982, 1990.
Bronk JT. Kelly P. The early effect of a tourniquet on bone fluid spaces. FASLB
ibkrtirrg, New Orleans, Louisiana, 1989.
Brown Rh, Weiss J. Neovascularization and its role in the osteoarthritic process.
.-lirir Rherrtrr Dis 47238 1-885. 1958.
Buckberg GD. Luck J, Payne BD, HotTmann JIE, Archie P.Fixler DE. Some
sources of error in measuring regional blood flow with radioactive microspheres.
./ Appl Plysiol3 l : 598-604. 197 1.
Byers P. Circulatory disorders. In: Salisbury JR WC. Byers PD (ed.) Diseases of
bones and joints. Chapman and Hall Medical. London UK. 1994.
Chanavaz M. Anatomy and histophysiology of the periosteum: quantification of
the periosteal blood supply to the adjacent bone with s5~rand gamma
spectrometry. ./ Oral itnplci~ir2 1 (3):2 14-2 19, 1 995.
IS. Cheong WF, Prahl S. Welch AJ. A review of the optical properties of biological
tissues. IEEE J Qirorir Electt* 36( 12):2 166-2 185. 1 990.
16. Clothiaux PL, Wood M. Vanhoutte PM. The effects of progressive ischemia on
vascular reactivity of the isolated canine tibia. Orthop Trails 1 1 : 197. 1986.
17. Copp DH. Shim S. Extraction ratio and bone clearance of 85~ras a measure of
effective bone blood flow. C'irc Rrs 16461-467, 1965.
18. Cowin S. Bone mechanics. In: Cowin S (ed.). CRC Press Inc. Boca Raton.
Florida, pp 3 13, 1989. Cui W, Ostrander L. The relationship of surface reflectance measurements to optical properties of layered biological media. IEEE 7iml.s Biomrd Eizg
39(2): 194-20 1. 1992.
Davis TR Holloway I. Pooley J. The effect of anaesthesia on the bone blood flow of the rabbit. J Orthop Res 8479-484. 1990.
Dedrick DK, Goldstein S. Brandt KD. O'Connor BL, Goulet RW, Albrecht >I. .A longitudinal study of subchondral plate and trabecular bone in cruciate-deficient dogs with osteoarthritis followed up for 54 months. Artl1riti.s i~hrrrn~36: 1160-
1467, 1993.
Dillaman RM. Roer R Gay DM. Fluid movement in bone: theoretical and empirical. J Biontecctz 24 Suppl 1 : 163- 177. 199 1.
Driessens M. Vanhoutte P. Vascular reactivity of the isolated tibia of the dog. .Am
./ t%j:viof 23 6 : H904. 1 979.
Drinker CK. Drinker K. A method for maintaining an artificial circulation through the tibia of the dog with a demonstration of the vasomotor control of the marrow vessels. Am ./ Physiof 40.5 14-52 I, 19 16.
Duncan C, Shim S. The autonomic nerve supply of bone. J Bo~rrJui~rt S~lrg
59B:323-333, 1977.
Ferguson D. Physiology for dental students. Churchill Livingstone Press. Bristol.
1988.
Fiorelli G, Orlando C, Benvenuti S, Franceschelli F, Bianchi S, Pioli P, Tanini A,
Serio M. Bartucci F, Brandi ML. Characterization, regulation. and function of specific cell membrane receptors for insulin-like growth factor I on bone
endothelial cells. Bo~e!t/irjer lies 9329-337, 1994.
28. Firbank M, Hiraoka M. Essenpries M, Delpy DT. Measurement of the optical
properties of the skull in the wavelength range of 650 - 950 nrn. Phy.s bled Biol
38:503-5 10, 1993.
19. Flock ST. Wilson B, Patterson MS. Total attenuation coefficients and scattering
phase function soft tissues and phantom materials at 633 nm. iLlrd Ph~s
14(5):535-54 1, 1987.
30. Forrester K. Doschak M, Bray R. In vivo comparison of scanning technique and
wavelength in laser Doppler perfusion imaging: measurement in knee ligaments
of adult rabbits. Met/ Biol fi~('o~npzit 3 5.581-556. 1997.
3 I. Forrester KR, Shymkiw R, Yeung G. Invin P. Tulip J. Bray R. Spatially resolved
diffuse reflectance for the determination of tissue optical properties and metabolic
state. ,\'PIE (submitted 2000).
32. Fujii H. Nohira K. Yamarnoto Y. Ikawa K Ohura T. Evaluation of blood flow by
laser speckle imaging sensing. dppl Optics 26(24):532 1-5325. 1987.
3. Fukuoka S. Hotokebuchi T. Terada K. Kobara N. Fujii H. Sugioka Y. lwarnoto Y.
Assessment of subchondral bone blood flow in the rabbit femoral condyle using
the laser speckle method. J Orthop Rrs 17368-3 75, 1999.
34. Fukuoka S. Hotokebuchi T. Jingushi S, Fujii H, Sugioka Y. Iwamoto Y.
Evaluation of blood flow within the subchondral bone of the femoral head: use of
the laser speckle method at surgery for osteonecrosis. J Orlhop Rcs 17:80-87.
1999. Glowacki I. Angiogenesis in fracture repair. Clitr Orthop Rei Rrs 355S:S82-S89,
1998.
Gross D. Williams W. Streaming potential and the electromechanical response of physiologically-moist bone. J Biornoci~15277-295. 1982.
Gross PM. Heistad D. Marcus ML. Neurohormonal regulation of blood flow to bones and marrow. .dm J Physiol237:8440-448, 1979.
Gross PM, Marcus M, Heistad DD. Measurement of blood flow to bone and marrow in experimental animals by means of the microsphere technique. .J Bow
.Joi~~tSurg 63A(6):1028- 103 1. 198 1.
Gross TS. Damji A. Judex S. Bray RC. Zernicke RF. Bone hyperemia precedes disuse-induced intracortical bone resorption. J dppl Pl~):sroi86( 1 ):230-135. 1999
Gross TS. iLlcLeod K. Rubin CT. Validation of surface strain gadients as a potent predictor of skeletal adaptation. Troti.s Orthop Xes Soc 40278. 1994.
Ham A. Some histological problems peculiar to calcitied tissues. .J Bwie .loitit
Sriq 34A:70 1-728. 1952.
Hansen ES. He SZ. Soballe K, Kjolseth D. Henriksen TB, Hjortdal VE. Bunger C.
[99mTc]diphosphonate uptake and hernodynamics in experimental arthritis: effect of naproxen in the canine carrageenan injection model. JOrrhop Rrs 10(5):647-
56, 1992.
Harrigan T.A nonlinear model for myogenic regulation of blood flow to bone: equilibrium states and stability characteristics. AIM Biornrd El~g242 1 1-22 1.
1996. Hellum S, Jacobsson S, Nilsson GE, Lewis DH. Measurement of microvascuIar blood flow in cancellous bone using laser Doppler flowmetry and 133~e- clearance. Irrt J Oral SII~12: 165- 177, 1983.
Hert I, Liskova M, Landa M. Reaction of bone to mechanical stimuli. Part 1:
Continuous and intermittent loading of tibia in the rabbit. Fofitr hlury/~oi(I'rc~qre)
19:290-300, 1971.
Heymann MA, Payne B. Hoffman JI, Rudolph AM. Blood flow measurement with radionuclide-labeled particles. P~*ug(iir.dio~rrsc Dis 10:5 5-79. 1977
Huang CH, Hong C. Lai ST. Tsai SK. Effects of dipyridamole, nitroglycerin. and nitroprusside on coronary vascular resistance in rabbits: measurement with pulsed
Doppler veloci metry. J fimn~osbfe!tl dssoc 98(7):480-485. 1999.
Iannacone W, KorostotY E. Pollack SR. Microelectrode study of stress-generated potentials obtained from uniform and nonuniform compression of human bone. .l
BiomedlLfater Res 13 : 753-763, 1979,
Jain R Podworny N. Anderson GI. Schemitsch EH. Assessment of the relationship between standard probe and implantable fiber measurements of cortical bone blood flow: a canine study. <'ol~*rfTissirr Irrt 59:64-69, 1996.
Jain R, Podworny N, Hearn T, Anderson GI, Schernitsch EH. Effect of stainless steel and titanium low-contact dynamic compression plate application on the vascularity and mechanical properties of cortical bone after fracture. J Orthop
Trarima 1 1 (7):490-495, 1997.
Jakobsson A, Nilsson G. Prediction of sampling depth and photon pathlength in laser Doppler flowmetry. Med Biof Dzg Comprrt 3 1 :30 1-307, 1993. Johnson M. Behavior of fluids in stressed bone and cellular stimulation. (irlcg
Tistie Ijzf 3 6:S72-S76, 1984.
Judex S, Gross T, Bray RC. Zernicke RF. Adaptation of bone to physiological stimuli. J Biomrcl~30(5):4 12-429, 1997.
Kelly P. The effect of unilateral increased venous pressure on bone remodeling in canine tibia. J Lab ('lirr /bled 72:4 10-4 18. 1968.
Kelly P. An K, Chao EYS, Rand JA. Fracture healing, biornechanical. tluid dynamic, and electrical considerations. In: Peck W (ed.) Bone and Mineral
Research. Elsevier Science Publishers. New York. pp 157- 169. 1985.
Kelly PJ. Bronk J. Venous pressure and bone formation. 1tfic*ro\*crsRev 39:364-
375, 1990.
Kelly PJ. illontgomery R Bronk JT. Reaction of the circulatory system to injury and regeneration- ('li~rOrtl~op Re/ Rus 254275-288. 1990.
Kety S. The theory and applications of the exchange of inen gas at the lungs and tissues. Pharr~icrcolRev 3: 1-41, 195 1.
Kivr T. Bone perksion and oxygenation: animal experiments and clinical observations. Actn Orthop Scntrcl Slippi 65257, 1994.
Kizr T. Dahl B, Lausten GS. The relationship between inert gas wash-out and radioactive tracer microspheres in measurement of bone blood flow: effect of decreased arterial supply and venous congestion on bone blood flow in an animal model. J Orthop Res 1 1 28-3 5. 1993. Kirkeby OI, Nordsletten L. Skjeldal S, Larsen TB. Circulation in corticocancellous bone grafts measured with laser Doppler flowmetry. Scmrtl J
Pirst Rrcorrstr Hmrcl Stirg 249-254. 1994.
Komblum SS, Kelly P. The lacunae and Haversian canals in tibia1 cortical bone from ischemic and non-ischemic limbs. A comparative micro radiographic study. .
Rotre Jobrt Szirg 46A:797-8 10. 1964.
Kowallik P. Schulz R Guth BD. Schade A, Pamausen W. Gross R Heusch G.
Measurement of regional myocardial blood tlow with multiple colored microspheres. C'irc 83974-981. 199 1.
Kowalski MJ. Schemitsch E, Kregor PJ. Senft D. Swiontkowski MF. Effect of periosteal stripping on conical bone pefision: a laser Doppler study in sheep.
(2rlcrf Tissue It~r59: 24-26, 1996.
Kregor PI. Sene D. Parvin D. Campbell C. Toomey S. Parker C. Gillespy T.
Swiontkowski MF. Conical bone perfusion in plated fractured sheep tibiae. .J
Oriltop Rev 13(5):7 1 5-724. 1995.
Laroche iM, Pouilles J. Ribot C. Comparison of the bone mineral content of the lower limbs in men with ischemic atherosclerotic disease. ('Iitr Rhrtmmni 136 1 1-
6 14, 1994.
Lassen N. Perl W. Tracer kinetic methods in medical physiology. Raven Press.
New York, 198 1.
Lausten GS, holdi C. Blood perfusion uneven in femoral head osteonecrosis:
Doppler flowmetry and intraosseous pressure in 12 cases. Acta Urthop SCatrd
64(5):533-536, 1993. Lausten GS. Kiar T, Dahl B. Laser Doppler flowmetry for estimation of bone blood tlow: studies of reproducibility and correlation with microsphere technique.
J Or~hopRrs 1 1 373-5 80, 1993.
Li G. Bronk J. Kelly PJ. Canine blood flow estimated with microspheies .J
Orthop Res 7:6 1-67. 1989.
Linden M. Sirsjo Lindbom L. Nilsson G. Gidlof A. Laser Doppler perhsion imaging of microvascular blood flow in rabbit tenuissimus muscle. ANZ.I Plysiol
238(38):H1496-H1500, 1995.
Lundsgaard JS. Jensen B. Gronlund J. Fast responding tlow independent blood gas catheter for oxysen measurement. ./dpp,/l/Hy.siol-I8:3 76-3 8 1. 1980.
MacDonald F. Pitt Ford T. Blood flow changes in the tibia during external loading. ./ Ortllup Re.$ I 1 136-4s. 1993.
MacDonald F. Pitt Ford T.The effect of blood tlow on the material properties of bone. .I Riomrd Eiig 72: 194-20 I. 1994.
Mal; .AFT. Huang D. Zhang JD. Tong P. Deformation-induced hierarchical tlows and drag forces in bone canaliculi and matrix microporosity. J Biomcci~30( 1 ): 1 1 -
18, 1997.
Matsuda JJ, Zernicke R, Vailas AC, Pedrini VA, Pefrini-Mille A Maynard JA.
Structural and mechanical adaptation of immature bone to strenuous exercise. J
Appl P!~~~sioi60:2028-34,1986.
McInnis JC. Robb R, Kelly PJ. The relationship of bone blood flow, bone tracer deposition, and endosteal new bone formation J Lab Cli~ibled 96(3):51 1-522.
1980. 78. Montgomery RJ, Sutker B. Bronk JT, Smith SR, Kelly PI. Interstitial fluid flow in
cortical bone. hIicroc~irscRes 3 5 295-307, 1988.
79. moorLD1 User Guide and Reference Manual - Software Version 2.0, Issue I. 1.
Moor instruments Ltd., Devon, UK, 1996.
SO. Morris MA, Kelly P Use of tracer microspheres to measure bone blood flow in
conscious dogs. ('alof Tisszir i11t 32x59-76. 1980.
81. Nakano T, Thompson J, Christopherson RJ, Aherne FX. Blood flow distribution
in hind limb bones and joint cartilage from young growing pigs. (iur J Ii.1 Nrs
50:96- 100, 1986.
Nilsson XMK. Berg R, Andersson-Engels S. Measurements of the optical
propenies of tissue in conjunction with photodynamic therapy .4pp/ Optics
34(2 1):46U9-4619, 1995.
Nilsson 0s. Urist M. Dawson EG. Schmalzried TP. Finerman GAM. Bone repair
induced by bone morphogenetic protein in ulnar defects in dogs. .l Bow .lorrrt
S'iirg 68B:635-642, 1986.
Notzli HP.Swiontkowski M, Thaxter ST. Carpenter III GK, Wyatt R. Laser
Doppler flowmetry for bone blood flow measurements: helium-neon laser light
attenuation and depth of pefision assessment. ./ Orthop Res 7(3):4 13-424, 1989.
Obeid AN. Barnett N. Dougherty G. Ward G. A critical review of laser Doppler
flowmetry. Jhied f51gTechtrol 14(5): 178- 18 1, 1990.
O'Connor JA, Lanyon L. The influence of strain rate on adaptive bone
remodeling. JBiomech 15:767-78 1, 1982. Olerud S, Danckwardt G. Fracture healing in compression osteosynthesis. An experimental study in dogs with avascular, diaphyseal, intermediate fragment.
A urn Orrhop Scnrrd 13 7(supp I), 1 97 1.
Oni 0, Gregg P. The relative contribution of individual osseous circulations to diaphyseal cortical blood supply. J Orlhop Trnrrmtr 1(4):44 1-448. 1990.
Otter MW, Qin Y. Rubin CT, McLeod KJ. Does bone perfbsion/reperfiision initiate bone remodeling and the stress fracture syndrome'? Mtl &pothr.st.-s
53(5):363-368, 1999.
Paradis GR, Kelly P. Blood flow and mineral deposition in canine tibia! fractures.
./ RO~IL'Joi)rt Slug 57A:120. 1975.
Preuss LE. Bolin F. Taylor RC. Environmental effects on light transport in tissues in vitro (100-1000 nm). Pl~oiucllr)~~PhoruhiolJ3S.32S, 1986.
Rhinelander F. The normal microcirculation of diaphyseal cortex and its response to fracture. JBot~rJobtt Sltrg 50A(4):784-800. 1968.
Rhinelander F. Tibial blood supply in relation to fracture healing. ('lit1 Ortlrop Ref
Rrs 105:34-8 1, 1974.
Richards DJ, Brookes M. Osteogenesis and the pH of osseous circulation. ('nlclf
Ti.sszie Rrs Suppl:93-93G 1968.
Rome CF, Maass-Moreno R. Hepatic venular resistance responses to norepinephrine, isoproterenol. adenosine, histamine, and AC h in rabbits. An1 ./
Phj9s10l274(3 Pt 2):H777-85, 1998.
Rubin CT, Lanyon L. Regulation of bone mass by mechanical strain magnitude.
CnIctf Tisslre It11 37:411-4 17, 1985. Schemitsch EH, Kowalski M, Swiontkowski MF, Harrington RM. Comparison of the effect of reamed and unreamed locked intramedullary nailing on blood tlow in the callus and strength of union following fracture of the sheep tibia. J Orthop Ne.s
l3:382-389. 1995.
Seiler I11 JG. Kregor P, Conrad EU, Swiontkowski MF. Posttraumatic
osteonecrosis in a swine model: correlation of blood cell flux. MRI, and
histology. dcra Orthot, Sccur~l67(3):249-254.1996.
Shepherd AP. Riedel G, Kiel JW. Haumschild DJ. Maxwell LC Evaluation of an
infrared laser Doppler flowmeter. Am J Physioi 6 1 :666-672. 1987
Shim S. Physiology of blood circulation of bone. ./ fif~lleJortrt Sirrg jOA(4): S 12-
824, 1968.
Shim SS. Patterson F. A direct method of qualitative study of bone blood
circulation. S'tirg (;~IIL'COhstef 125:261-268. 1967
Shymkiw RC. Bray R Boyd SK, Kantzas A, Zernicke RF. Physiological and
mechanical adaptations of periarticular cancellous bone following joint ligament
injury. .J ,4ppi Phwol (submitted 2000).
Srniseth 04 Manyari D. Lima Jr\. Scott-Douglas NW. Kingma I. Smith ER
Tyberg SV.Modulation of vascular capacitance by angiotensin and nitroprusside:
a mechanism of changes in pericardial pressure. ('irc 76(4):875-83. 1987.
Smits GI, Roman R, Lombard IH.Evaluation of laser Doppler flowmetry as a
measure of tissue blood flow. J Appl PIysiol6 1(2):666-672, 1986.
Spodaryk K Dabrowski 2. Blood flow in different regions of bone marrow after
short -ten exercise. Acta Physiol H~rrzg77( 1): 13- 17, 199 1. Stem M. In vivo evaluation of microcirculation by coherent light scattering.
NCI11ire 254:56-58, 1975.
Swiontkowski MF. Ganz R Schlegel U. Perren SM. Laser Doppler flowmetry for clinical evaluation of femoral head osteonecrosis: preliminary experience. ('li~r
Orthop Nel Res 218:181-185. 1987.
Swiontkowski MF, Sene D. Conical bone microperfusion: response to ischemia and changes in major arterial blood flow. J Orrllap Rrs 10(3):3 3 7-3 43. 1 992.
Swiontkowski MF. Tepic S. Rahn BA Cordey J. Perren SM. The effect of fracture on femoral head blood flow. rlct~~Orthup Sco,rtl64(2):196-202. 1993.
Swiontkowski MF. Tepic S. Perren Sbl. Moor R Ganz R Rahn BA. Laser
Doppler tlowmetry for bone blood flow measurement: correlation with microsphere estimates and evaluation of the effect of intracapsular pressure on femoral head blood flow. J Ortl~opRRL'S 4362-27 1. 1986.
Takeuchi A, Araki R, Proskurin Y. Yamada Y. Ishii J. Katayama S. Itabashi A. .A new method of bone tissue measurement based upon light scattering. J Botre 1bf1t1
Re.5. 132): 26 1-266, 1 997.
Tondevold E. Haemodynamics of long bones: an experimental study in dogs. rlcttr
Ortl~opSmtd Suppl 205:25-3 1. 1983.
Tondevold E. Bulow J. Bone blood flow in conscious dogs at rest and during exercise. Acfa Orrhop Scorzd 54: 53-57. 1983.
Tothill P. Bone blood flow measurement. J Biomrd Etzg 625 1-256. 1984.
Trifit PD. Cieslak C. Gregg PJ. A quantitative study of the routes of blood flow to the tibia1 diaphysis after an osteotomy. J Orthop Res 1 1149-57, 1993. 116. Triffit PD, Gregg P. Depression of bone blood flow after blunt trauma: A fracture
study in the adult rabbit. Acta Orthop Scartd65(2):195-195, 1994.
1 17. Triffit PD, Gregg P. Measurement of blood flow to the tibia1 diaphysis using I I
prn radioactive microspheres. A comparative study in the adult rabbit. J Ortlrop
Res 8:642-645, 1990.
I IS. Vo~qkMT. Cauley J. Kuller LH. Nevitt MC. Bone mineral density and blood flow
to the lower extremities: the study of osteoporotic fractures. J Bow iblhr Rrs
12(2):283-259, 1997.
1 19. Wardell K. Braverman I. Silverman DG. Nilsson GE. Spatial heterogeneity in
normal skin perfusion recorded with laser Doppler imaging and flowmetry.
iL1icrot.cr.s~Rrs 45:26-35. 1994.
120. Wardell K. Jakobsson A, Nilsson GE. Laser Doppler perfusion imaging by
dynamic light scattering. !EEL 'lktrra BME-40:309-316, 1 993.
12 1. Welch AJ, van Gemert M, Star WM. Wilson BC. Definitions and overview of
tissue optics. In: Welch k). van Gemert M (eds.) Optical-Thermal Response of
laser-irradiated tissue. Plenum Press. New York. pp 15-46? 1995.
122. Welch AJ, van Gemen M. Star WM, Wilson BC. Optical-Thermal Response of
Laser Irradiated Tissue. In: Welch AJ, van Gemert M (eds.). Plenum Press, New
Yoric, pp 15-46,233-303. 1995.
123. Wilson B. Measurement of tissue optical properties: methods and theories. In:
Welch N,van Gemert M (eds.) Optical-thermal response of laser-irradiated
tissue. Plenum Press, New York, pp 23 3-274, 1995. 124. Wilson BC. Jacques S. Optical reflectance and transmittance of tissues: principles
and applications. IEEJ Qlror~lztmElectrorrics 26( 12):2 186-2 199, 1990.
125. Wilson BC, Jeeves W, Lowe DM. In vivo and post mortem measurements of the
attenuation spectra of light in mammalian tissues. P/utuchrm Pl~otohioi42: 1 53 -
162, 1985.
126. Wilson BC. Patterson M. Bums DM. The effect of photosensitizer concentration
in tissue on the penetration depth of photoactivating light. Lrcsr 1Lfc.tlSci 1:Xj-
244, 1986.
127. Wilson BC, Patterson M. Flock ST. Indirect versus direct techniques for the
measurement of the optical properties of tissues. Pllotod~rmPltotohiol46(5):60 1 -
608, 1987.
128. Wray B. Lynch C. The vascular response to fracture of the tibia in the rat. ./ Hotre
.Joir~rS1rt-g 4 1 A: 1 143- 1 148. 1959.
129. Zhong J. Seifalian A. Salerud GE. Nilsson GE.A mathematical analysis on the
biological zero problem in laser Doppler flowmetry. IEEE Trtuis Riotnetl Etig
45(3):354-364, 1998.