RadloGraphlcs Indexterms: Mechanisms responsible forthe MR ... · Hemorrhage, MRstudies Brain,hemorrhage Volume 8,Number3#{149}May,1988#{149}RadioGraphics 427 RadloGraphlcs Indexterms:

Cumulative index terms:Hemorrhage, MR studiesBrain, hemorrhage

Volume 8, Number 3 #{149}May, 1988 #{149}RadioGraphics 427

RadloGraphlcs Index terms:Magnetic resonance imaging

TECHNICALNeurologic imaging

BRAINRadiation Physics

MAGNETIC RESONANCEIMAGING

THIS EXHIBIT WAS DISPLAYED ATTHE 72ND SCIENTIFIC ASSEMBLYAND ANNUAL MEETING OF THE RA-DIOLOGICAL SOCIETY OF NORTHAMERICA. NOVEMBER 30-DECEM-BER 5. 1986. CHICAGO, ILLINOIS. ITWAS RECOMMENDED BY THE NEU-RORADIOLOGY PANEL AND WASACCEPTED FOR PUBLICATION AF-TER PEER REVIEW AND REVISIONONJULY3I. 1987.

From the Department ofRadiology, Hospital of the Uni-vensity of Pennsylvania, Phila-delphia.

(#{149})Presently at the De-pantment of Radiology, Ha-dassah Hospital, Jerusalem, Is-rael.

Address reprint requeststo R. I. Grossman, M.D., Neuro-radiology Section, Depart-ment of Radiology, Hospital ofthe University of Pennsylvania,3400 Spruce Street, Philadel-phia, PA 19104.

Mechanisms responsible for theMR appearance and evolutionof intracranial hemorrhage

John M. Gomori, M.D.

Robert I. Grossman, M.D.

The sequential degradation products of hemoglobin Inan evolving hemorrhage are described here, and thephysical mechanisms underlying their MR imagingcharacteristics are discussed.

Introduction

The hemoglobin in extravasated blood (hemorrhage) undergoesa predictable sequence of chemical degradations. The degradationproducts are paramagnetic and for a yaniety of reasons theirrelaxation rates following exposure to a strong external magneticfield and an RF pulse differ from those of the normal surroundingbrain. Since MR signal intensity is a function of the relaxation rate ofthe substance imaged, hemoglobin degradation products and,hence, hemorrhage is distinguishable from normal brain on MRimages. Moreover, since the degradation product present at aparticular time and its spatial distribution are a function of the age ofthe hemorrhage, it is usually possible to distinguish between acute,subacute and chronic hemorrhage with MRI.

The purposes of this article are: (1) to define the sequentialchemical degradation of hemoglobin in an evolving hemorrhage, (2)to elucidate the physical mechanisms that affect the relaxationrates of the various degradation products, and (3) to specify thedependence of these physical mechanisms on one’s choice of fieldstrength and pulse sequence.

The clinical MR images of evolving hemorrhage, the biologicfactors that affect those images and the variety of vascular lesions inwhich they may be encountered are the subjects of the followingarticle.

MRI of hemorrhage: Mechanisms Gomori and Grossman

428 RadioGraphics #{149}May, 1988 #{149}Volume 8, Number 3

The Sequential Degradation of Hemoglobin

The acute hemorrhage (blood collection)consists of intact red cells and plasma. Theplasma is nesorbed and, depending on the pHand p02, the hemoglobin becomes desaturat-ed. With time, the hemoglobin is converted byoxidation to methemoglobin and the red cellslyse. Methemoglobin is eventually degradedand resonbed, and some of its iron is converted

to fennitin and hemosidenin in the adjacent re-active macrophages (Figure 1). These macro-phages persist for a long time if the blood brainbarrier is intact. The iron of some states of he-moglobin and its derivatives is paramagneticbecause it has unpaired electrons. This pro-foundly affects the MR image.

Figure 1Evolution of intrapanenchymal hemorrhage.(See page 442.)

Gomori and Grossman MRI of hemorrhage: Mechanisms

Volume 8, Number 3 #{149}May, 1988 #{149}RadloGraphics 429

Physical M#{149}chanisms that Aff#{149}ctthe Relaxation

Rat#{149}sof Hemoglobin Degradation Products

MAGNETIC SUSCEPTIBILITY

Both the proton and electron are changedparticles with spin 1� (the unit of quantized an-gulan momentum). The momentum and elec-tnic change of the proton and electron result ina magnetic moment; i.e., they are magneticdipoles (north and south magnetic poles sepa-rated by a distance). The electron’s lightenmass, about one-thousandth of the proton’smass, gives it a magnetic moment 1,000 timeslarger. In the absence of an external field, theelectron magnetic dipoles of a paramagneticsubstance point randomly in all directions, ne-suIting in zero net magnetization. When an ex-tennal field is applied, the strength of the fieldwill be augmented because some of the un-

paired electrons will align with the appliedmagnetic field. The ratio of the additional fieldstrength to the strength of the applied field iscalled the magnetic susceptibility of the para-magnetic substance. It is an indication of howeasily it can be magnetized. If a paramagnet-ic atom has N unpaired electrons, its magneticsusceptibility is proportional to (N)(N+2).

Eq. 1 x, magnetic susceptibility, � (N)(N+2)

The T2 relaxation rate of a paramagnetic sub-stance varies as the square of its magnetic sus-ceptibility.

PROTON-ELECTRON DIPOLE-DIPOLE PROTON RELAXATIONENHANCEMENT

In the absence of unpaired electrons, theprotons of water molecules relax (realign withthe magnetic field) via fluctuations in their 10-cal magnetic fields caused by the motion ofadjacent protons. Adjacent unpaired elec-trons, because of their larger magnetic mo-ment create fluctuations in local magneticfields I ,000 times larger than those producedby the motion of protons. These larger fluctua-tions enhance the relaxation of protons. This iscalled proton-electron dipole-dipole protonrelaxation enhancement (PRE). The strength of

the field of a dipole decreases as the thirdpower of the distance from the point of mea-sunement to the dipole, and the strength of amutual dipole-dipole interaction is proportion-al to the product of the two dipole fields.Hence, a dipole-dipole interaction decreasesas the sixth power of the distance betweenthe dipoles. This means that a water protonmust approach within 3 angstroms (A) of anunpaired electron for the dipole-dipole inter-action to take place.

SELECTIVE T2 RELAXATION ENHANCEMENT OWING TOHETEROGENEITY

In addition, if a panamagnetic substance isheterogeneously (nonunifonmly) distributed inthe volume of tissue being examined and wa-ten protons cannot come within 3 A of the un-pained electrons, selective T2 relaxation en-hancement is produced. This is in contrast toproton-electron dipole-dipole proton relax-ation enhancement (PRE) which causes bothTI and T2 PRE.

The precession nate (Lanmon frequency) ofwater protons is proportional to the local field

strength, which in turn, varies with the localmagnetic susceptibility. A heterogeneouslydistributed panamagnetic substance causesvariations in local magnetic susceptibility frompoint to point. Following the initial 900 pulse ofa spin-echo sequence phase differences de-velop among the precessing protons that arenormally refocussed by the ensuing 180#{176}pulses. The phase spread cannot be nefo-cussed by the spin-echo pulse sequence, how-ever, if the water protons diffuse to regions of

U,

EU,

Ca

‘CU0

aUU,>1

‘Ca.



different field strengths during the pulse se-quence. Thus, water protons diffusing throughregions of varying field strengths due to theheterogeneous distribution of a paramagneticsubstance will lose their transverse phase co-herence fasten; i.e. , have a shorten T2 thanthey would in a completely homogeneous me-dium.

With spin-echo techniques, the 12 de-creases with increase in the interecho interval(the time between 180#{176}pulses) because there

is more time for water protons to diffuse to ne-gions of greaten field difference between twoconsecutive echos. The effect of this T2 PREmechanism increases as the square of themagnitude of the applied magnetic field andas the square of the variation in magnetic sus-ceptibility. The effect of the proton-electrondipole-dipole PRE mechanism, on the otherhand, varies little or decreases with increasingfield strength.

Structural Characteristics of Hemoglobin Degradation ProductsThat Aff#{149}ctRelaxation Rates

Oxyhemoglobin contains iron in the fennous(F&2) state but is not panamagnetic. Deoxy-hemoglobin contains iron in the ferrous (Fe�2)state with four unpaired electrons and is para-magnetic. These unpaired electrons, however,are shielded from direct dipole-dipole intenac-tion with water protons.

The iron of methemoglobin is in the fernicstate (Fe�3), has five unpaired electrons and isparamagnetic. In this case, the unpaired elec-tnons cause proton relaxation enhancementby the dipole-dipole interaction (shorten TIand T2).

Fernitin and hemosidenin have a similarstructure with a cone containing, on the aver-age, 2,000 iron atoms in the F&3 state (5 un-

paired electrons pen iron atom). The iron conesare 50-70 A in diameter and are surroundedby a protein shell with an outer diameter ofabout 120 A. Most of the unpaired electrons ofthe iron in fennitin are shielded from direct di-pole-dipole interactions with water protons.The iron cores are single-domain crystals ofFeOOH (fennic hydnoxyoxide) and are super-paramagnetic; i.e., all the unpaired electronspins of the iron core stay parallel to (remainedaligned with) each other. In effect, each fern-tin and hemosidenin cone acts as a magnetwhich is always oriented in the direction of theapplied magnetic field, and whose strength in-creases with the field strength.

Effscts of Relaxation Rate Changes on Signal intensities of

Hemorrhages (3)

ACUTE HEMORRHAGE

In acute hemorrhage, depending on thepH and p02, the hemoglobin loses its oxygen,becoming deoxyhemoglobin (Hb) which ispanamagnetic because it has 4 unpaired elec-tnons. In the physiological range, a lowering ofthe pH shifts the oxyhemoglobin dissociationcurve to the night; i.e. more desatunated he-moglobin (Bohr effect). Increasing the CO2 (atconstant pH) also lowers the oxygen affinity. Itsunpaired electrons are inaccessible for dipole-

dipole interaction with water protons, but in-tracellulan deoxyhemoglobin (Hb IRBC) hasheterogeneous magnetic susceptibility. Whenan external magnetic field is applied, the high-en susceptibility inside the red cell (RBC) con-taming the deoxyhemoglobin results in fieldgradients inside and outside the cell (acrossthe cell membrane) (Figure 2). Although thespin-echo sequence will correct for static fieldinhomogeneities, it will not completely correct

ii

0

0.

Inside Cell:Susceptibility

S

S

Figure 3Diffusion of water across thecell membrane to regions ofdifferent magnetic suscepti-bility and therefore to regionsof different augmented localfield strengths.

0

00

0

000

S Deoxyhemogiobi�

0 Water 1

Gomorl and Grossman MRI of hemorrhage: Mechanisms


for the effects of water proton diffusion acrossfield gradients (Figure 3). On spin-echo Se-quences, the T2 relaxation rate (I/T2) due tointracellular deoxyhemoglobin increases asthe square of the magnetic field strength and

increases with lengthening of the interecho in-terval (Figures 4-8). The increased T2 relax-ation rate results in a decrease in the signal in-tensity of a hemorrhage on T2 weightedimages.

Figure 2Distortion of the magnetic fieldlines by the difference in magneticsusceptibility between intracellulardeoxyhemoglobin and extracellulan

. � plasma.

Outside Cell:

Susceptibility X0

0

0

0

1/T2oc B20

0

C

0=0303.a0

0a’a

‘C

0E0=0

4-3U

.300

.200

[HbLRBC �

#{174}-�

c’J

I-

.100

.300

.200

U)

c’.JI-

.100

IRBC

.300

.200

Ua,U)

C’4I-

.100

U

IRBC

In Field 5trength (tesla)



.19 .35 71 .94 1.18 1.41

Magnetic Field Strength (tesla)

Figure 4Variation of T2 of deoxyhemoglobin ly-sate as a function of magnetic fieldstrength with interecho intervals from 2msec to 32 msec. Hb LRBC = deoxyhe-moglobin, lysed red blood cells

.1U9 .35 3i .94 1.18 1.41Magnetic Field Strength (tesia)

Figure 6Variation of the T2 of intracellular deoxyhe-moglobin as a function of field strength withan intenecho interval of 32 msec. Hb IRBC =

deoxyhemoglobin, intact red blood cells

.19 .35 .71 .94 1.18 1.41


Figure 5Vaniation of the T2 of intracellular deoxyhe-moglobin as a function of field strength withan interecho interval of 2 msec. Hb IRBC =

deoxyhemoglobin, intact ned blood cells

Figure 7Variation with field strength of the T2 relaxation rate( 1/T2) of intracellular deoxyhemoglobin owing to in-tnacellulan deoxyhemoglobin’s panamagnetic suscep-tibility. The slope of the individual plot in this In-Ingraph indicates the exponent of the relationship be-tween the relaxation rate and field strength. The cm-cled numbers one the intenecho intervals in msecand the number under each of them is the averageslope of the curve. The average of all the curves isindicated at the bottom.

10.0

(sec-i)

5.0

0

C

0

=0303.a0

�5�3�25 6�5 I�5

1I�t (sec-’)cpmg

Figure 81.18T Variation of the T2 relaxation rate ( 1/12) of intnacel-

‘.94T lular deoxyhemoglobin as a function of the echo nate7ff (#{189}rcpmg); i.e., the inverse ofthe intenecho interval

-Tho (2rcprng). Note the marked increase in T2 relaxationrate with long intenecho intervals (short 1,4rcpmg).

SUBACUTE HEMORRHAGE

bin.



Methemoglobmn (MHb) is panamagneticbecause it has 5 unpaired electrons. Theseelectrons are accessible to water protons fordipole-dipole interactions which shorten bothTI and T2. This interaction is not significantlyaffected by field strength or cell integrity. Inthe MRI of dilute solutions of methemoglobin,the TI shortening is more important than the T2shortening. Figures 9-11 depict the variationsin the TI of plasma and various states of hemo-globin with field strength. Intracellular methe-moglobmn (MHb IRBC) also possesses heteroge-neous magnetic susceptibility. From equationI (page 429), its magnetic susceptibility is pro-portional to (5) x (5 + 2), and the magneticsusceptibility of deoxyhemoglobin is pnopon-tional to (4) x (4 + 2). Also from page 429, theT2 relaxation rate of a panamagnetic sub-stance varies as the square of its magnetic

susceptibility. Hence, on spin echo sequences,the T2 relaxation rate of methemoglobin is2.13 times that of intracellular deoxyhemoglo-

((5)(5 + 2)/(4)(4 + 2))2 (35/24)2 2.13

It also increases as the square of the magneticfield strength and with intenecho intervallengthening (Figures 12-15). On gradient echosequences, however, the I/T2 of methemo-globmn will be only I .46 times that of intracellulardeoxyhemoglobmn.

((5)(5 + 2)/(4)(4 + 2)) = (35/24) = I .4#{243}

Table I summarizes the relationships nespon-sible for the MR appearance of acute and sub-acute hematomas.

2

Plasma

2

,� ABC� LRBC

.1’g .35 :��i .94 1.18


Figure 9Variation of the Ti relaxation times of plas-ma, intracellular oxyhemoglobmn and oxyhe-moglobmn lysate as a function of fieldstrength. Hb02 IRBC = oxyhemoglobmn, in-tact red blood cells Hb02 LRBC = oxyhe-moglobmn, lysed ned blood cells

Hb

_#____#{149}�_#{149}________ Hb LRB�

.1�9 .35 :h .94 1.18 141


Figure 10Variation of the Ti relaxation times ofintracellular deoxyhemoglobmn anddeoxyhemoglobin lysate as a functionof field strength. Hb IRBC = deoxyhe-moglobmn, intact red blood cells HbLRBC = deoxyhemoglobin, lysed nedblood cells

2

Ua,

I-

- MHb IRBC

-MHb LRBC


RadioGraphics #{149}May, 1988 #{149}Volume 8, Number 3

0a’a

‘C

0E0=0

4-3Ua3

U)

434

Figure 1 1Variation of the Ti relaxation times of intracellularmethemoglobin and methemoglobmn lysate as afunction of field strength. MHb IRBC = methemoglo-bin, intact ned blood cells MHb LRBC = methemo-globin, lysed ned blood cells

.1� .:�s :�i .94 1.18Magnetic Field Strength (tesla)

.300

.200

C.)a)Cl)

.100

.300

.200

C)a,U)

csJI-

.100

[MHb IABC

.300

.200

Ua,U)

c’.JI-

.100

2J

oa,Cl)

�t87

IRBC



MHbLRBC

.1’g .35 ii .94 1.18 1.41Magnetic Field Strength (tesla)

Figure 12Variation of the T2 relaxation of methemo-globmn lysate as a function of field strength.There was no change in relaxation time as afunction of intenecho intervals from 2msec-32 msec. MHb LRBC = methemo-globmn, lysed ned blood cells

.1#{149}9.35 :�i .94 1.18 1.41Magnetic Field Strength (tesla)

Figure 14Variation of the T2 relaxation time of intra-cellular methemoglobin as a function offield strength at an interecho interval of 32msec. MHb IRBC = methemoglobin, intactned blood cells

.19 .35 .71 .94 1.18 1.41


Figure 13Variation of the T2 relaxation time of intna-cellular methemoglobmn as a function offield strength at an interecho interval of 2msec. MHb IRBC = methemoglobmn, intactred blood cells

,�

2.16

2.23

Average Slope=2.O1± .19

-.062In Field Strength (tesla)

.lel

Figure 15Ln-In graph of the T2 relaxation rate ( i /T2) of intra-cellular methemoglobin due to its panamagnetism asa function of field strength. The interecho interval ofeach curve in msec is circled. The number below theinterecho interval represents the slope of eachcurve. The quadratic relation is indicated by the aver-age slope of 2.


0a’a

‘C

0

E0=0

U 04-3 -�a �.a �3 eq

U) 0 -� 4) C)

0 �

�.% � � C) � � Q�”�’

-�.‘ ,_4._� :� � bO�4 4) �40%.-, C� a3 �0 Z5”

� E 14 �.I �4 4.)� � + �

�

� ,-� �0

� � � C� (� C�‘�, -� � � �4 I-. l� 4.)

�1� 4.) 4.) a“r c’�l � ‘_�4� �.4 �4 1’4� � � .-

-� Lt) � a,LI) C1) �Cl� I � �

I ,� �% :� b04.),� 0 00 � � ‘4

� � � � 0 0bOa,0� � � 2�

� �

II � � 9 a,�p�� ,c1.� S �� 1� I� � � +� �‘0

� � � �- - �� - C) 0 .�

� � � � a,U) .4.) C):� � 0 �

I :�: � a � ��;% � Z5’ + 2�o� �C’, � -� �4.) ‘� � To- bO

0 s-i ,�j 0 �4 0 � 0+ + � � 0 � � � E

,� & � � �4 ��#{247})

0 � - ;.4 O4.!� � p�1 �4’�’.-4

� � � ,.a � � � � 0

� � ‘� s- � �;�eq- - :� �I .�.) � C� E-4 � s.� �

� � + �1� � �

� � 6� 5% E,-� 4.) � .� � � a�4.) � �.&.4 � 0 0 C) V C) �

� � � -4 �4 eqeq,�4 � e3 � c’3 “�4 �4 14

I I � �1 � E-4E-4E4 �.�-..-.......-‘.- � � � 1.4

0 � � � � ‘-4 ‘-4 - - ‘-4 -4 -� z�- :�

� � ‘�‘1�

� � � � __ __‘-,, ‘� 0 0

� � � � �‘.-.- .U)

� I

� �c’� ‘�N cz� �E� E’ � �- d E4 E-4 � � ,..�

436 RadloGraphics #{149}May, 1988 #{149}Volume 8, Number 3

50

Ua)U)

Ec�J

75

50

C)a,U)

Ec’,JI-

25

Normal Spleen

Hemochromatotic�S<en�n

Ferritin LRB�107 mg/cc



CHRONIC HEMORRHAGE (1,2)

Hemosidenin and fernitin are very closely ne-lated chemically. They are the characteristiccomponents of chronic hematomas. The moreabundant hemosidenin is insoluble and is foundmostly in the lysosomes of reactive macno-phages in the brain parenchyma surrounding ahemorrhage. Fernitin is water soluble and ho-mogeneous solutions of fennitin have a signifi-cant I/T2 (rate of T2 relaxation) that increasesalmost as the square of the field strength butdoes not vary with spin-echo intenecho interval(between 2-32 msec) (Figure 16). This appearsto be due to the superparamagnetic suscepti-bility of fennitin resulting in field gradients nearthe fernitin macromolecule. The I/T2 of hemo-sidenin deposits containing an equivalentamount of iron is 5 times greaten. Not only doesthe T2 relaxation rate of hemosidenin increasequadratically with field strength, it also in-

creases with lengthening of the spin-echo in-terecho interval (Figures 17-19). The heteno-geneous distribution of hemosidenin wasmodeled with a heterogeneous distribution offernitin obtained by suspending intact oxygen-ated ned cells (nonparamagnetic) in an equalvolume of fernitin solution thus creating a fieldgradient across the red cell membrane. Theadditional I/T2 owing to the heterogeneousdistribution of fennitin was about 4 times that at-tnibutable to the panamagnetic properties offennitin itself, at I .4 T with long spin-echo inter-echo intervals (Figures 20-22). This confirmsthat the dominant contributor to the T2 relax-ation nate of hemosidenin deposits is the het-enogeneous distribution of hemosidenin . Figure23 summarizes the properties of various con-stituents of hemorrhage.

.19 .35 71 .94 1.18 iAi


Figure 16Variation of the T2 relaxation time of afennitin nod cell lysate as a function offield strength. The T2 relaxation timewas unaffected by the intencho interval atinterecho intervals between 2 msec and32 msec. LRBC = lysed red blood cells

:19 .35 .7�1 .94 1.18 1.41


Figure 17Variation of the T2 relaxation times of a normalspleen sample (0.52 mg Fe/g) and of a hemochno-matotic spleen sample (10.6 mg Fe/g) as functionsof field strength. The intenecho interval was 2 msec.

0a’a

‘C

0E0=UC0

‘C0

4.0

ln(sec’)

2.0

100

(sec�’)

50

Hemochromatotic

�n�n

Normal Spleen

0 62.5 125 500

Ua)U)

EC’JI-

Ferritin IRBC

107 mg/cc

2

.�9 .::�5 :�i .94 1.18 1.41Magnetic Field Strength (tesla)

438

MRI of hemorrhage: Mechanisms Gomorl and Grossman

RadloGraphics #{149}May, 1988 #{149}Volume 8, Number 3

I#{174}HemochromatoticSpleen

©

Normal Spleen

-1.66 -1.05 -34 -06 +17 +34In Field Strength (tesla)

Figure 18Ln-ln graph of the field strength dependent relax-ation rate versus field strength for normal and he-mochromatotic spleen samples. The interecho inter-val was 2 msec.

250

2T cpmg��

Figure 19Variation of the T2 relaxation rate ( 1/T2) of a normalspleen sample as a function of echo-rate (1hrcpmg)

at 1.4 T and of a hemochnomatotic spleen sample atO.7i T.

Figure 20Variation of the T2 relaxation time of a fernitin red cell(IRBC) suspension as a function of field strength.The circled numbers identify the plots at spin-echointercho intervals of 2 msec and 8 msec. IRBC = in-tact red blood cells

50

U

U

C)

EF-

0C

a,

U-

0

-j

C’jI-

0

C’j

C

03.

030

=0303.

a0

400(seci

300

200

100

C)

0

0)

EN-

0C

a,

U.

0

-J

c’,JI-

0

CsJ

Average Slope 1.57±48

ln(Field Strength) (In tesla)

62.5 1�5 250 5#{212}00

PEDO PrutouPIE Distrlbutiuu 12 PEE Deuslty

PrutsuDsusity

WI11W I 12W I

LRBC

- IRBC + -- -- -�

LRBC - --

ConcentratedLRBC

DiluteLRBC

+-�--,

-�-..-4--,-

Ferritin Fe�’ “�iO,OOO + - Intracellular + 10+4 -- �-

Hemosiderln Fe�3 “10,000 - - Intraiysosomai + to++ -- .-_ort !Arrows indicate increased (t) or decreased (�) signal intensity. Size of arrow indicates magnitude of the effect.



Figure 21Ln-In graph of the variation in the T2 relaxationrates of fennitin ( i fF2) owing to the hetenoge-neity in susceptibility caused by the suspen-sion of ned cells in the fennitin as a function ofthe field strength. The intenecho intervals inmsec are indicated as circled numbers. Theslope of each curve is indicated adjacent tothe plot. The average slope is consistent witha quadratic relationship between field strengthand relaxationrate(1/12). IRBC = intactnedblood cells LRBC = Iysed ned blood cells

liii # Uupslrsd WitirUSISCUIS Fire Elictrius EslubIs

Figure 22Variation of the T2 relaxation nate ( i /12) of fennitinowing to the heterogeneity of susceptibility causedby the suspension of red cells in fennitin as a functionof the echo-rate (#{189}rcpmg) at 0.35 T (lower plot) and1.4 T. IRBC = intact red blood cells LRBC = lysedned blood cells 2rcpmg spin-echo intenecho interval

Figure 23Summarizes the MR properties of the vanious con-stituents of hemorrhage. Ti WI, Proton Density WIand T2WI refer to , Proton Density and T2 weight-ed images respectively.

Oxyhemoglobin FV’ 1 + - IRBC

tSensitMty top0, & pH

Deoxyhemoglobin Fe’ 4 +

Methemoglobin FV3 5 + + IRBC

MRI of hemorrhage: Mechanisms Gomorl and Grossman

440 RadioGraphlcs #{149}May, 1988 #{149}Volume 8, Number 3

Gradient Echo Techniques in the MR Imaging of Hemorrhage

The spin-echo technique has been used toovercome static field and chemical shift inho-mogeneities; but the high field uniformitiesachieved in modern MR units permit us to dis-card the refocussing 180#{176}pulses in a spin-echosequence. The introduction of gradient-echotechniques has exciting implications for theMRI of hematomas (4). With these techniques,it is possible to obtain much greater sensitivityto field mnhomogeneities that are caused bythe heterogeneity of magnetic susceptibilityintrinsic to the tissues of the patient. Gradientecho techniques are sensitive to heterogene-ity oven the dimensions of an image voxel (ap-proximately I mm) rather than the distancediffused by water between echoes (approxi-mately 0.01 mm). This increased sensitivity ismanifested by one’s ability to observe mag-netic susceptibility differences at lower fieldstrengths and by one’s ability to detect het-erogeneous susceptibility oven much larger dis-tances; e.g., inhomogeneous clotting. There

will be no significant change in apparent T2with changes in intenecho interval, and bound-aries between regions of different magneticsusceptibility on chemical shift will appear tohave shortened T2. The boundary betweendeoxyhemoglobin or methemoglobin, whetherintracellular on extracellulan, and the surround-ing panenchyma will appear hypointense nela-tive to normal white matter on T2 images, mim-icking the hypointensity of hemosidenindeposits. This may cause some confusion whenevaluating hemorrhagic tumors, which usuallydo not have complete hypointense rims onspin-echo sequences. Of course, similar man-ginal hypointensity will occur between fat andother tissues and perhaps between melanoticmelanomas and surrounding tissues and be-tween air or dense bone and other tissues.Thus, the marginal hypomntensity on gradientecho imaging is not as specific as on spin echosequences.

Summary

The sequential degradation of hemoglobin in an evolving hemorrhage has been reviewed.Physical mechanisms of proton relaxation enhancement that contribute to the clinical MRappearances of hemorrhage have been described, and the dependence of relaxation rates onfield strength and interecho interval in spin-echo imaging techniques has been defined.

R#{149}ferences

I . Gomori JM. Grossman RI, Goldberg HI. Zimmerman RA.Bilaniuk LI: Intracranial hematomas: Imaging by high-field MR. Radiology 1985; 157:87-93.

2. Gomori JM. Grossman RI: Mechanisms responsible forthe MR appearance and evolution of intracranial hem-orrhage. (abstr.) Radiology 198#{243};1#{243}1(P):3#{243}4.

3. Gomori JM, Grossman RI. Vu Ip C. Asakura I: Depen-

dence of NMR relaxation times of blood on fieldstrength. oxidation state. and cell integrity. J ComputAssistlomogr 1987; 11:#{243}84-#{243}90.

4. Edelman RR. Johnson K. Buxton R. et al. MR of hemor-rhage: a new approach. AJNR 1986; 7:751-756.

Documents

RadloGraphlcs Indexterms: Mechanisms responsible forthe MR ... · Hemorrhage, MRstudies Brain,hemorrhage Volume 8,Number3#{149}May,1988#{149}RadioGraphics 427 RadloGraphlcs Indexterms: