MR Colonography Optimized Enema Composition1

Document Sample
MR Colonography Optimized Enema Composition1 Powered By Docstoc
					                                              Technical Developments

Wolfgang Luboldt, MD                            MR Colonography: Optimized
Johannes M. Frohlich, PhD
Nicole Schneider, MD                            Enema Composition1
Dominik Weishaupt, MD
Franziska Landolt, MSc
Jorg F. Debatin, MD                                                                          gadopentetate dimeglumine (Magnevist;
                                                  Manganese chloride, iron glycero-          Berlex Laboratories, Wayne, NJ) per liter
                                                  phosphate, and cellulose additive          of water is used for MR colonography.
                                                  were assessed as base materials for        Although it fulfills the SI requirements,
Index terms:
Colon, MR, 75.121412, 75.121419,                  use in a T1-shortening single con-         the low viscosity induces backflow into
   75.12143                                       trast enema for magnetic resonance         the small bowel, which complicates sub-
Gadolinium                                        (MR) colonography. Contrast-to-            sequent analysis and reduces colonic dis-
Iron                                              noise ratios (CNRs) were compared          tention. More important however, the
Magnetic resonance (MR), contrast                                                            enema has a volume of up to 2,000 mL
   agents, 75.121412, 75.121419,
                                                  to those with the standard 10
                                                  mmol/L gadolinium–based enema.             and thus requires use of as much as 40 mL
Magnetic resonance (MR), contrast                 On T1-weighted three-dimensional           of a 0.5 mol/L paramagnetic contrast
   enhancement, 75.12143                          gradient-recalled-echo images, CNRs        agent formulation. The associated cost is
Manganese                                                                                    high, which negatively effects consider-
                                                  with the iron glycerophosphate en-
                                                                                             ation of MR colonography as a screening
Radiology 1999; 212:265–269                       ema exceeded those with the man-
                                                                                             alternative for colonic polyps.
                                                  ganese- and gadolinium-based en-
                                                                                                The purpose of this study was to iden-
Abbreviations:                                    emas. Use of an additive of 0.8%
CNR contrast-to-noise ratio                                                                  tify a less costly, more viscous enema
                                                  wt/wt cellulose was found to be            with SI characteristics similar to those
GRE gradient-recalled echo
SE spin echo
                                                  practicable as it increased viscosity      with the 10 mmol/L gadopentetate dime-
SI signal intensity                               sufficiently without altering CNR. The     glumine and water enema currently em-
2D two-dimensional                                gadolinium-based enema can be re-          ployed.
3D three-dimensional                              placed with an iron glycerophos-
                                                  phate enema to render MR colon-
1 From the Institute of Diagnostic                ography less costly.                         Materials and Methods
Radiology, University Hospital Zurich,
Ramistrasse 100, CH-8091 Zurich,                                                             MR Imaging
Switzerland. Received May 27, 1998;
revision requested August 5; revision                                                           All imaging was performed on a 1.5-T
received September 10; accepted De-                                                          system (Signa Acuspot; GE Medical Sys-
cember 16. Supported in part by                                                              tems, Milwaukee, Wis). To ensure a suffi-
German Research Foundation stipend              Magnetic resonance (MR) colonography
Lu 687/1-1. Address reprint requests
                                                                                             ciently large field of view to cover the
to J.F.D. (e-mail:        combines the application of a rectal en-     entire colon, the body coil was used for
    RSNA, 1999
                                                ema containing paramagnetic contrast         signal transmission and reception. Two
                                                agent with breath-hold, T1-weighted,         sequences were employed: 3D spoiled
                                                high-spatial-resolution, three-dimensional   GRE (repetition time msec/echo time
                                                (3D) imaging and heavily T2-weighted,        msec      4.7/2, 45° flip angle, 0.5 signals
                                                two-dimensional (2D), single-shot, spin-     acquired, 62.5-kHz bandwidth, 382 192
                                                echo (SE) imaging (1). Beyond being bio-     matrix, 40 40-cm field of view, 2.4-mm
                                                compatible, the enema used at MR colo-       section thickness) and 2D single-shot fast
                                                nography needs to fulfill MR signal          SE (65-msec echo time, 31.3-kHz band-
                                                intensity (SI) and mechanical require-       width, 256       192 matrix, 40       40-cm
                                                ments. For reconstruction of endoscopic      field of view, 6-mm section thickness).
Author contributions:                           views, the enema should be homoge-
Guarantor of integrity of entire study,         neously bright on T1-weighted, 3D, gradi-
J.F.D.; study concepts, J.F.D., W.L.,                                                        Assessment of New Contrast Media
                                                ent-recalled-echo (GRE) images and homo-
J.M.F.; study design, W.L.; definition of       geneously dark on T2-weighted images            Two agents—manganese chloride and
intellectual content, J.F.D., W.L., J.M.F.;
literature research, W.L., N.S., F.L.,          for better depiction of pathologic condi-    iron glycerophosphate—were assessed as
D.W.; experimental studies, W.L., N.S.,         tions contained within the colonic wall.     possible replacements for gadopentetate
F.L.; data acquisition, W.L., N.S., D.W.;       In addition, the enema needs to homoge-      dimeglumine. Water-based dilution series
data analysis, D.W.; statistical analysis,      neously fill and distend the colon. These    with each agent and with gadopentetate
D.W., N.S.; manuscript preparation
and editing, W.L., J.F.D., D.W.; manu-          mechanical properties are mainly depen-      dimeglumine were imaged with the 3D
script review, J.F.D., D.W., J.M.F.             dent on the viscosity of the enema.          spoiled GRE and 2D single-shot fast SE
                                                   Presently, an enema with 10 mmol of       sequences. In addition, a water-dilution

series with iron glycerophosphate and
600 mg of manganese chloride was im-
aged to determine possible synergistic
effects with the two substances. On the
basis of SI data acquired on 3D spoiled
GRE images, the concentration with
which the maximum SI was achieved was
determined with a finely incremented
dilution series. SI-concentration curves
were obtained for each of the four evalu-
ated contrast media.
   The contrast media that provided the
maximum contrast-to-noise ratio (CNR)
on the 3D spoiled GRE images was used at
the optimum concentration to fill the
colon of three anesthetized pigs. The ex-
periments were conducted in accordance
with state regulations governing animal
experiments. The colon was imaged with                       Figure 1. For T1-weighted 3D spoiled GRE images, graph plots SIs
both the 3D spoiled GRE and 2D single-                       relative to dilution series of manganese chloride (Mn), iron glycero-
shot fast SE sequences. To determine CNR,                    phosphate (Fe), manganese chloride and iron glycerophosphate [Fe
SI characteristics of the contrast agent–                    Mn (600 mg/l)], or gadopentetate dimeglumine (Gd). In all cases, an
filled colonic lumen were evaluated rela-                    initial SI increase is followed by an eventual SI decrease. This
                                                             appearance reflects the rising T2-shortening effect of these substances
tive to the colonic wall and surrounding
                                                             as their concentration is increased. The highest maximum was
structures. Additionally, three colonic                      measured with iron glycerophosphate.
specimens exposed to the enema were
obtained for histologic examination of
iron uptake with Prussian blue staining.
                                                previously optimized concentration, the           Beyond 15 mmol/L, iron glycerophos-
                                                gadopentetate dimeglumine and iron                phate provided SI values higher than the
Optimization of Enema Viscosity                 glycerophosphate dilution series were re-         combination of iron glycerophosphate
                                                peated with and without the cellulose             and manganese chloride. Results with the
   Cellulose was added to enhance the
                                                additive, with the 3D spoiled GRE se-             gadopentetate dimeglumine dilution se-
mechanical properties of the enema by
                                                quence. The ratio of the resultant SIs with       ries revealed a maximum SI at a concentra-
increasing its viscosity. In an initial step,
                                                and SIs without cellulose (SIwith/SIwithout)      tion of 23 mmol/L. The maximum gado-
the maximum additive of cellulose was
                                                were plotted against the concentrations           pentetate dimeglumine value (SI, 1,932)
determined that allowed passage of the
                                                of gadopentetate dimeglumine and iron             exceeded values with manganese chlo-
mixture through the enema tip at 1.5 m
                                                glycerophosphate. The paired Student t            ride (SI, 640) but remained lower than the
of hydrostatic pressure. With this maxi-
                                                test was used to assess statistically signifi-    maximum value with iron glycerophos-
mum cellulose concentration as a base-
                                                cant differences in the collected data sets.      phate (SI, 2,078).
line, the enema viscosity was titrated for
MR colonography by means of progres-                                                                 In pigs, the 45 mmol/L iron glycero-
                                                  Results                                         phosphate solution provided homoge-
sive dilution with water until complete
colonic filling at 1.5-m hydrostatic pres-         On the T1-weighted 3D spoiled GRE              neously high SI within the colonic lumen
sure was possible.                              images, both manganese chloride and               on the T1-weighted 3D spoiled GRE im-
   Dilution experiments were performed          iron glycerophosphate resulted in an ini-         ages (Fig 2a), which permitted reconstruc-
in six patients (four men and two women;        tial concentration-dependent increase in          tion of virtual endoscopic images. The
age range, 53–72 years; mean age, 61            SI followed by an eventual decrease in SI         mean for CNRs of the colonic lumen
years) undergoing MR colonography in a          as T2-shortening effects began to domi-           relative to surrounding structures was
larger study designed to determine the          nate T1 effects with increased concentra-         48     5 (SD). On T2-weighted 2D single-
diagnostic performance of MR colonogra-         tions (Fig 1). The highest SI was achieved        shot fast SE images, the colonic lumen
phy in the detection of colonic masses.         with the iron glycerophosphate and wa-            was homogeneously dark, which permit-
This study had been approved by the             ter solution at a concentration of 45             ted easy delineation of the hyperintense
local ethics committee, and written in-         mmol/L, which corresponds to an iron              colonic wall (Fig 2b). The means for CNRs
formed consent was obtained from all            load of 28 g/L. Manganese chloride solu-          of the colonic wall relative to the colonic
patients. The patients had undergone co-        tions were associated with lower SI val-          lumen or surrounding structures were
lonic preparation for subsequent conven-        ues. Maximum SI was observed with a               21     3 and 10      4, respectively. Histo-
tional colonoscopy. Colonic filling was         concentration of 5 mmol/L, which corre-           logic analysis of the colonic specimen
monitored with a non–section-selective          sponds to a manganese load of 1 g/L. The          revealed no evidence of iron uptake into
2D spoiled GRE acquisition collecting           combination of manganese chloride and             the colonic mucosal cells.
one image every second. The acquired            iron glycerophosphate showed a synergis-             The cellulose additive beyond 2% wt/wt
images permitted assessment of colonic          tic effect as evidenced by SI values that         precluded passage through the enema
distention and reflux into the small bowel.     exceeded those with only iron glycero-            tip. To ensure easy filling of the entire
   To determine the influence on SI char-       phosphate up to a concentration of 15             colon under in vivo conditions, the cellu-
acteristics of the cellulose additive in its    mmol/L of iron glycerophosphate (Fig 1).          lose additive was lowered to 0.8% wt/wt

266 • Radiology • July 1999                                                                                                            Luboldt et al
a.                                                                         b.
Figure 2. Coronal (a) 3D spoiled GRE and (b) 2D single-shot fast SE images of the colon after rectal administration of a highly concentrated (28 g/L)
iron glycerophosphate enema to a pig. The colonic lumen is bright in a. Even at this high concentration, iron glycerophosphate caused no artifacts.
Since the colon of the pig was artificially perforated, the T1-shortening iron glycerophosphate can also be identified within the peritoneal cavity
(arrows in a). In b, the colonic lumen is dark. Having inherently high SI on T2-weighted images, the colonic wall (arrowheads) is clearly delineated
from the colonic lumen and surrounding structures.

(Fig 3). At this concentration, there was         enema can be optimized to dramatically             concentration (Fig 1). This cannot be
no reflux noted in any of the six patients.       lower the overall cost of an MR colon-             considered sufficient for MR colonogra-
The addition of cellulose at this concentra-      ographic examination.                              phy.
tion had no effect on the SI characteris-            Presently, the amount of gadopen-                  Subsequently, iron was tested in its
tics on 3D spoiled GRE images for either          tetate dimeglumine required for opti-              most tolerable form as iron glycerophos-
the gadopentetate dimeglumine or iron             mum contrast on both T1-weighted 3D                phate (5). SI values achieved with iron
glycerophosphate dilution series (Fig 4).         spoiled GRE and T2-weighted 2D single-             glycerophosphate were far higher than
The paired Student t test revealed no             shot fast SE images represents a cost fac-         those associated with manganese chlo-
statistically significant differences in SI       tor that would preclude use of MR colon-           ride and even exceeded those with gado-
(P .5).                                           ography for screening purposes. Since the          pentetate dimeglumine (Fig 1). Maximum
                                                  contrast agent is administered rectally            SI values were seen at a concentration of
     Discussion                                   and not intravenously, there is a consider-        45 mmol/L, which corresponds to an iron
   The potential benefits of colonic polyp        ably wider latitude regarding tolerance            load of 28 g/L. In an attempt to lower the
screening are well documented (2). Re-            and safety of agents. We evaluated two             concentration of iron glycerophosphate,
cently, we reported promising results with        considerably less expensive agents with            manganese chloride was added. The
MR colonography (3). Although the cost            known T1- and T2-shortening effects as             amount of manganese chloride added
of MR colonography depends on the con-            possible alternatives to gadopentetate             corresponded to the concentration associ-
tent of the enema, its performance de-            dimeglumine: manganese chloride and                ated with 87% (SI, 558) of the maximum
pends on the intraluminal SI and disten-          iron glycerophosphate. Manganese was               SI increase in the manganese chloride
tion of the colon. The cost of MR colon-          chosen as it is the substance responsible          dilution series (SI, 640). Indeed, the addi-
ography has been rather high, which               for the T1-shortening effects inherent in          tion of manganese chloride to iron glyc-
predominantly reflects the use of as much         blueberry juice (4). Manganese chloride            erophosphate resulted in higher SI values
as 40 mL of gadopentetate dimeglumine             provided a maximum SI (SI, 640) that was           at lower concentrations. At higher con-
as the contrast agent for the enema. Find-        41% of that achieved with gadopentetate            centrations, the combination reached a
ings in this study demonstrate that the           dimeglumine (SI, 1,560) at the 10 mmol/L           maximum SI value of 75% (SI, 1,560) of

Volume 212 • Number 1                                                                      MR Colonography: Optimized Enema Composition • 267
the maximum SI achieved with iron glyc-
                                                  Figure 3. MR image of
erophosphate alone (SI, 2,078) (Fig 1).           colonic filling permits
   The in vivo imaging characteristics of         assessment of colonic
iron glycerophosphate are favorable for           distention and reflux
MR colonography. At a concentration of            into the small bowel.
45 mmol/L, the enema was displayed as             The additive of cellulose
                                                  prevented reflux of con-
homogeneously bright on T1-weighted               trast agent into the small
3D spoiled GRE images and totally dark            bowel in all six patients
on T2-weighted 2D single-shot fast SE             and resulted in good co-
images (Fig 2). The visual impressions are        lonic distention.
corroborated by the high CNRs. In con-
trast to superparamagnetic iron oxide,
which requires matching with diamag-
netic barium sulfate particles to avoid
susceptibility artifacts (6,7), images ob-
tained with iron glycerophosphate de-
picted no artifacts up to the tested concen-
tration of 45 mmol/L (Fig 2).
   Iron glycerophosphate appears to be a
safe agent. In fact, it has been used in
high doses without relevant side effects
in pediatric populations (8). It is poorly
absorbed when administered orally (9,10).
Although conclusive data are not avail-
able, it is not likely to cause any side
effect if administered rectally. As was also
confirmed in our study, there is no mea-
surable resorption of this physiologic iron
compound in the colon (11). It is conceiv-
able, however, that iron could be resorbed if
reflux occurs far into the proximal small
bowel. Even if such an extensive reflux were
to occur, the transient nature of the iron
present in the small bowel would obviate
any serious imbalance in iron metabolism.         Figure 4. For T1-weight-
As an additional safeguard, iron could be         ed 3D spoiled GRE im-
                                                  ages, graph plots the ra-
embedded in macromolecules (iron-dex-             tio of SIs obtained with
tran, iron-cellulose) that are not resorbed. In   and SIs obtained with-
this manner, the potential toxicity would be      out the addition of 0.8%
reduced as the viscosity increased.               wt/wt cellulose depend-
   Similar to double contrast barium en-          ing on the concentra-
ema examination (12,13), accurate assess-         tion of gadopentetate
                                                  dimeglumine (Gd) or
ment of the colonic lumen with MR                 iron glycerophosphate
colonography mandates adequate co-                (Fe). Cellulose had no
lonic distention (3). Beyond the adminis-         measurable influence on
tration of antiperistaltic, distention is de-     SIs.
pendent on the viscosity of the enema.
Currently, the enema used at MR colon-
ography consists mainly of water, which
results in quick passage of enema fluid
through the ileocecal valve into the small
bowel in a vast majority of patients. In-
creases in viscosity must, however, be
balanced against the need for easy and            quently, distention was good (Fig 3). Re-     the radioactive isotope iron-59 instead of
fast filling of the colon, particularly in        sults of in vitro experiments showed that     iron-56 [14]). A relative price difference of
view of limitations regarding usable hydro-       the addition of cellulose to the enema did    a factor of approximately 25 in conjunc-
static pressure. Although results of in vitro     not have any measurable effect on SIs on      tion with slightly better SI characteristics
experiments identified a 2% wt/wt cellu-          3D spoiled GRE images (Fig 4).                makes this a worthwhile endeavor. Even
lose additive as maximum, results of in             Iron glycerophosphate seems to be a         though MR colonography may become
vivo evaluations in six patients demon-           feasible alternative to gadopentetate dime-   less expensive with the use of iron glycero-
strated a cellulose additive beyond 0.8%          glumine as a contrast enema in MR colon-      phosphate, a cost analysis of MR colonog-
wt/wt to be impracticable. At this concen-        ography. Before any clinical use, more        raphy compared with other modalities,
tration, there was no reflux of contrast          safety data need to be collected (eg, by      such as barium enema examination or
material into the small bowel and, conse-         means of radiolabeling techniques with        colonoscopy, will need to be performed.

268 • Radiology • July 1999                                                                                                     Luboldt et al
References                                              from skim milk enriched with iron glyc-              molecular weight iron complexes. Eur J
 1. Luboldt W, Bauerfeind P, Steiner P, Fried           erophosphate (letter). Arch Latinoam Nutr            Drug Metab Pharmacokinet 1991; 16:203–
     M, Krestin GP, Debatin JF. Preliminary             1973; 23:145–150.                                    206.
     assessment of three-dimensional mag-          6.   Liebig T, Stoupis C, Ros PR, Ballinger JR,     11.   Lombard M, Chua E, O’Toole P. Regula-
     netic resonance imaging for various co-            Briggs RW. A potentially artifact-free oral          tion of intestinal non-haem iron absorp-
     lonic disorders. Lancet 1997; 349:1288–            contrast agent for gastrointestinal MRI.             tion. Gut 1997; 40:435–439.
     1291.                                              Magn Reson Med 1993; 30:646–649.               12.   Goei R, Nix M, Kessels AH, Ten Tusscher
 2. Winawer SJ, Zauber AG, Ho MN, et al.           7.   Briggs RW, Wu Z, Mladinich CR, et al. In             MP. Use of antispasmodic drugs in double
     Prevention of colorectal cancer by colon-          vivo animal tests of an artifact-free con-           contrast barium enema examination: glu-
     oscopic polypectomy: the National Polyp            trast agent for gastrointestinal MRI. Magn           cagon or buscopan? Clin Radiol 1995;
     Study Workgroup. N Engl J Med 1993;                Reson Imaging 1997; 15:559–566.                      50:553–557.
     329:1977–1981.                                8.   Pobedinskaia VN. Use of large doses of         13.   Rubesin SE, Furth EE, Rose D, Levine MS,
 3. Luboldt W, Steiner P, Bauerfeind P, Pel-            iron glycerophosphate in the therapy and             Laufer I. The effects of distention of the
     konen P, Debatin JF. Detection of mass             prevention of allergic diseases in children.         colon during air-contrast barium enema
     lesions with MR colonography: prelimi-             Vopr Okhr Materin Det 1976; 21:61. [Rus-             on colonic morphology: anatomic correla-
     nary report. Radiology 1998; 207:59–65.            sian]                                                tion. AJR 1995; 164:1387–1389.
 4. Hiraishi K, Narabayashi I, Fujita O, et al.    9.   Hallberg L, Solvell L, Brise H. Search for     14.   Chouly C, Pouliquen D, Lucet I, Jeune JJ,
     Blueberry juice: preliminary evaluation as         substances promoting the absorption of               Jallet P. Development of superparamag-
     an oral contrast agent in gastrointestinal         iron: studies on absorption and side-                netic nanoparticles for MRI: effect of par-
     MR imaging. Radiology 1995; 194:119–               effects. Acta Med Scand Suppl 1966; 459:             ticle size, charge and surface nature on
     123.                                               11–21.                                               biodistribution. J Microencapsul 1996; 13:
 5. Layrisse M, Martinez Torres C, Divo MR,       10.   Anghileri LJ, Cordova Martinez A, Main-              245–255.
     Jaffe W, Torres Suarez JE. Iron absorption         cent P, Robert J. In vivo behaviour of low

Volume 212 • Number 1                                                                       MR Colonography: Optimized Enema Composition • 269