(CANCER RESEARCH 50. 2330-2336. April 15. 1990]

Biodistribution and Radiation Dose Estimates for Yttrium- and Iodine-labeled

Monoclonal Antibody IgG and Fragments in Nude Mice Bearing HumanColonie Tumor Xenografts1

Robert M. Sharkey, Cecilia Motta-Hennessy, David Pawlyk, Jeffry A. Siegel, and David M. Goldenberg2

Center for Molecular Medicine and Immunology [R. M. S., C. M-H., D. P., D. M. C], University of Medicine and Dentistry of New Jersey, Newark, New Jersey 0711)3,and Temple University Hospital (J. A. S.j, Philadelphia, Pennsylvania 19140

ABSTRACT

An anti-carcinoembryonic antigen murine monoclonal antibody designated N'P-4, and its F(ab')3 and Fab' fragments, were coupled to the 1/1

mixture of l-isothiocyanato-benzyl-3-methyl- and l-methyl-3-isothio-cyanato-bcn/yl-diethylenetriaminepentaacetic acid chelate and labeledwith '"In or **Y. Biodistribution studies in nude mice bearing a human

colonie tumor xenograft were performed with these labeled conjugates,and comparisons were made to unconjugated NP-4 IgG and fragmentslabeled with '"I. Regardless of the labeling method, higher tumor uptake

was found with the intact IgG than with the fragments, but due to fasterblood clearance, tumor/blood ratios were higher for the fragments thanfor the IgG. Tumor uptake for the radiometal-labeled NP-4 was generallyhigher than the '"I-labeled NP-4. Tumor/nontumor ratios for the liver,kidney, and spleen were higher for the "'In- and *8Y-labeled NP-4 IgG

than the respective radiometal-labeled fragments, but tumor/nontumorratios for the '"I-NP-4 fragments were higher than the '-"I-NP-4 IgG.

Radiometal uptake in the kidney was approximately 8 and 150 timeshigher than the '"I-NP-4 F(ab')2 and Fab', respectively, and the clear

ance of radiometal activity in the kidneys was approximately 10 timesslower than the radioiodine. Quantitation of **Y or '"In activity in thefemur showed 3-5%/g for the IgG and F(ab')2 and only 1-2%/g for theFab'. The amount of radioactivity in the femur remained constant overtime, and between 60 and 100% of the **Y activity remained after flushingthe core of the femur with saline, whereas 50-70% of the '"In and only25-30% of the '"I activity remained after washing. Radiation dose

estimates derived from these studies suggest that at the maximal tolerateddose '"I-NP-4 IgG would deliver 5.9 times the dose to the tumor as "Y-labeled N'P-4 IgG. wY-labeled fragments would not be useful due tohigher doses to the kidneys than to the tumor. However, with '"I-labeled

IgG and fragments there is greater flexibility to permit tumoricidal doseswithout excessive toxicitv to the normal tissues.

INTRODUCTION

Antitumor antibodies labeled with radionuclides have beenshown to inhibit, and in some cases even cure, human tumorxenografts grown in animals (1-5). RAIT-' has also been attempted in man with some reports of tumor regression (6-10).Although encouraging, these early RAIT studies in man havenot yet made a significant impact on the treatment of cancer.Since these clinical trials have been phase I/II studies involvingadvanced cases of radioresistant tumors, such as colorectal orhepatocellular carcinomas, and melanoma, it is not surprisingthat the efficacy seen in animal models has not been translateddirectly to humans. Indeed, we have shown that, even in ani-

Received 8/18/89; revised 1/8/90.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported in part by USPHS Grants CA 39841 and BRSG RR-05903 fromthe NIH. Presented in part at the Second Conference on Radioimmunodetectionand Radioimmunotherapy of Cancer. Princeton, NJ. 1988.

2To whom requests for reprints should be addressed, at Center for Molecular

Medicine and Immunology. 1 Bruce Street, Newark, NJ 07103.'The abbreviations used are: RAIT, radioimmunothcrapy: CA-DTPA, cyclic

anhydride-DTPA; CEA. carcinoembryonic antigen: DTPA. diethylenetriamine-pentaacetic acid; HPLC. high pressure liquid chromatography: ITC. isothio-cyanate; MAb. monoclonal antibody; M.X-DTPA. a 1:1 mixture of 1-isothiocyan-atoben/yl-3-mcthyl- and l-methyl-3-isothiocyanato-benzyl-DTPA; MTD. dose atwhich 100'V of animals survive with less than 20r; weight loss.

mais, it is more difficult to treat well-established colonie cancerxenografts than smaller tumors (2), suggesting that there maybe limitations to the amount of tumor burden that can betreated by RAIT. The limitation of RAIT will depend on manyfactors, of which the MAb, radionuclide, and tumor type are afew examples. Currently, the most promising therapeutic resultshave been reported in patients with more radiosensitive lym-phomas (9).

Although there are many factors that will contribute to thesuccessful application of RAIT, there is considerable interest inthe choice of radionuclides. The first generation isotope usedfor therapy has been '"I (11). This isotope is readily available,

inexpensive, and can be conjugated to antibodies easily, and itsdistribution in tumor and tissue can be monitored by externalscintigraphy. However, the maximal range of nonpenetrating/^-emissions from '"I is only 2.4 mm, whereas the range of 90Yis 12 mm. W'Yhas a 64-h half-life and can be produced from areusable ""'Sr generator, and antibodies can be labeled usingDTPA types of chelates. For these reasons, 90Yhas become the

second generation radionuclide for RAIT studies. Animal studies have shown that 90Y-labeled antibodies can inhibit tumor

growth, but there has been concern over the potential forexcessive myelotoxicity due to the deposition of''"Y in the bone

(12, 13). Bone marrow toxicity has lead some investigators toexplore localized therapy in the peritoneal cavity in an attemptto reduce the accessibility of the TOY-labeled antibody to the

bloodstream and thereby reduce its contact with the marrow(14). Despite this regional delivery approach, Stewart et al. (15)have predicted that no more than 20 mCi of ""'Ycan be admin

istered safely as a single i.p. injection.We previously examined the biodistribution and therapy of

an anti-CEA murine MAb labeled with TOYin nude mice bearing

a human colonie tumor xenograft (12). With a single i.v.injection of the 9()Y-labeled antibody, tumor growth could be

inhibited, but not prevented, because we were limited in thedose we could inject due to marrow toxicity. Whole bodyautoradiography showed substantial uptake in all bones. Recentreports by Brechbiel et ai. (16) and Esteban et al. (17) haveindicated that MAbs conjugated to DTPA using the CA-DTPAhad more liver uptake and less tumor uptake than the sameMAb conjugated to DTPA using an ITC-DTPA derivative. Thecoupling of DTPA to MAbs via the ITC-benzyl-DTPA allowsmore of the carboxyl groups to participate in the binding ofradiometals than the CA-DTPA, thereby providing greaterstability for the chelated metal. In addition, the monofunctionalbinding of the ITC-DTPA eliminates the possibility of intramolecular cross-linking to the MAb that may occur with thebifunctional CA-DTPA (16). We found that conjugation of ananti-CEA MAb to an ITC-DTPA type of chelate improvestumor uptake and reduces liver and bone deposition of TOY.4Inthis report, we have taken one of these chelates, the MX-DTPA,

' R. M. Sharkey, C. Motta-Hennessy. O. A. Gansow. M. W. Brechbiel, I.

Fand, and D. M. Goldenberg. Selection of a DTPA chelate conjugate for monoclonal antibody targeting to a human colonie tumor in nude mice. Int. J. Cancer(in press).

2330

BloniSTRIBUTION AND DOSIMETRI FOR YERSUS ""Y-IgG OR FRAGMENTS

and conjugated it to an anti-CEA MAb, NP-4 IgG, or its F(ab')2and Fab' fragments. Biodistribution studies were performedwith both '"In and **Yto determine how well '"In may predictthe uptake of'"'V'-labeled antibody. ""Y was used in place of *°Y

because, as a ^-emitting radionuclide. it is easier to quantitatein tissues than 9"Y. We also compared these results to thebiodistribution of '"I-labeled NP-4 IgG and its fragments. The

biodistribution data were used to provide radiation dose estimates to the tumor and normal tissues in an attempt to determine which radionuclide would be best for i.v. RAIT applications and which vehicle [IgG, F(ab')2, or Fab'] would be best

suited to target radioactivity to the tumor.

MATERIALS AND METHODS

Preparation and Conjugation of NP-4 and Its Fragments. NP-4 IgGwas isolated from mouse ascites by Protein A and ion exchange ehro-matography, and its purity was checked by immunoelectrophoresis.sodium dodecyl sulfate-polyacrylamide gel electrophoresis and isoelec-tric focusing. F(ab')2 was prepared by pepsin digestion and purified byProtein A chromatography. Fab' fragments were generated from theF(ab')2 by reduction with cysteine and alkylation with iodoacetamide

and purified by molecular sieve chromatography. The purity of thefragments was tested using the same procedures as the IgG.

The synthesis of the MX-DTPA and its conjugation and labeling toNP-4 IgG and fragments were all done in the laboratories of Dr. OttoGansow. The final number of DTPA molecules/MAb was determinedby using a trace quantity of uC-labeled MX-DTPA in the coupling

procedure, according to the method described by Brechbiel et al. ( 16).By this method, the conjugates were determined to have 1.0. 0.7. and0.3 chelates/NP-4 IgG, F(ab')2. and Fab', respectively.

Radiolabeling and Quality Assurance. Radioiodination with '"I purchased from New England Nuclear (Boston, MA) was by the chlora-mine-T reaction, as described previously (2). Radiolabeling with '"Inand **Ywas performed as described earlier (16). The labeled productswere purified by HPLC and shipped at 4°Cby overnight express mail

for injection into animals on the following day. Quality assuranceassays were conducted on each antibody preparation on the day ofinjection. These assays included immunoreactivity measurements, molecular size determination by HPLC, and measuring unbound radionuclide content by instant thin layer chromatography. The immunoreactivity of the '"I-, '"In-, and 88Y-labeled NP-4 were similar, with

binding percentages to the CEA immunoadsorbent between 83 and94% for the IgG and F(ab')2 and 75 and 85% for the Fab' fragments.

Greater than 95% of the radioactivity for NP-4 IgG, 90-95% of theF(ab')2, and 80-90% of the Fab' bound to an anti-mouse IgG immu

noadsorbent. There were no detectable amounts of aggregation byHPLC size exclusion chromatography and less than 1% unbound '"Inor "*Y was found in the antibodies prior to injection. '"I-NP-4 preparations had between 3 and 5% unbound '"I.

Animal Studies. Female athymic mice (Harían,Madison, WI) at 4-5 weeks of age were given s.c. 0.2 ml of a 10% suspension of the CEA-producing. human colonie tumor xenograft, GW-39 (18), which isserially propagated in nude mice. Two weeks later, separate groups ofanimals were given i.v. 25 ^Ci (2 jig of MAb protein) of '"I-labeledNP-4 or its fragments, 4-5 nCi (2-3 ^g) of '"In-labeled NP-4, or 2-3¿jCi(1-3 Mg)of **Y-labeled NP-4. Four to five animals for each group

were anesthetized, bled by cardiac puncture, and then killed by cervicaldislocation on Days 1, 3, and 7 for IgG and F(ab')2 or on Days 1, 2.and 3 for Fab' fragments. Tumor weights were between 0.15 and 0.40

g at these times. Radioactivity in the bone was determined by removalof both femurs from the animals. After removing all muscle, the jointson both ends of the femurs were removed. One femur from each animalwas washed by inserting a 25-gauge needle into the core of the boneand passing 0.5 ml of 0.15 M NaCl first through one side and then anadditional 0.5 ml through the other side. The washed bone was weighedafter it had been blotted dry. The amount of radioactivity in the washedbone was compared to the amount of activity in the unwashed bonetaken from the same animal. The tumors and tissues were counted in a•¿�y-scintillationcounter, using windows set for each isotope (200-450.

150-300. and 500-2000 keV for '"I, '"In, and ""Y, respectively).

Standards prepared from the injected materials were counted each timewith the tissues so that the percentage of injected dose/g of tissue couldbe calculated with correction for the physical decay.

Radiation Dosimetry. The biodistribution data were used to generatetime-activity curves for the calculation of radiation doses to the tumorand tissues. Biodistribution data for normal tissues were fit to anexponential curve and integrated over all time for dosimetry. Thederived exponential is assumed to hold for data before 24 h. Otherbiodistribution experiments with data for 1 and 6 h support thisassumption.5 Tumor activity was not exponential, and so the integral

was computed trapezoidally by assuming activity equals zero at timezero. This integral is computed only to the last experimental time point.Since we did not account for activity after the last time point, theresultant tumor dose is likely conservative. Indeed, if extrapolation ofactivity in the tumor was taken over an infinite period of time, isotopeswith longer half-lives would impart higher doses to the tumor thanthose with shorter half-lives. Since the conventional medical internalradiation dose tables only provide S factors for tissues more than 1 g,we generated S factors to account for tissues less than this size (19).For all tissues except the bone, the assumption of uniform distributionof activity in a spherical mass was maintained, and doses were calculatedby procedures described earlier (2). For the blood, .9 values wereobtained by extrapolation of the values reported by Hui and Poston(20). For the bone, a cylindrical shape with the dimensions of 2-mmdiameter x 10-mm length was used to generate an S value as follows:the apparent fi-absorption coefficient (i>)in cnr/g was first determinedaccording to (21):

18.6(20 - 0.036)"7

where 2,, is the maximum energy of the ß-rayspectrum in MeV. Thiswas multiplied by the diameter of the cylinder (d) and then multipliedby 1.9 g/cm', the density of bone. An absorbed fraction (</>)of 0.45 was

then obtained for this particular value of vd (22). The S value was thencalculated to be 14.9 rads/VCi-h according to the relationship S= A0/?i,where A is the mean energy emitted/unit of cumulated activity and mis the mass of the cylinder in g (23). Using a similar procedure, a boneS value of 4.9 rads/VCi-h was obtained for '"I.

Statistics. All comparisons were made by a one-way analysis ofvariance using a two-tailed /-"test (24).

RESULTS

Biodistribution of Radioiodinated and Radiometal-labeled NP-4 IgG and Its Fragments. Fig. 1 shows the comparison of thetumor uptake of NP-4 IgG and fragments labeled with 90Y,'"In, or '"I. Maximal tumor accretion occurred by Day 3 for

the radiolabeled NP-4 IgG. Unlike IgG, maximal tumor accretion of the fragments was seen by Day 1 for the radiometal-labeled and radioiodinated NP-4. F(ab')2 fragments maintainedtheir level in the tumor for 2-3 days, whereas the tumor uptakeof Fab' fragments declined continually. The tumor uptake ofthe *"Y- and '"I-labeled F(ab')2 was similar on Day 1, but theamount of '"I-NP-4 F(ab')2 progressively decreased at a fasterrate, so that by Day 7 there was almost 6 times more 88Y-NP-4 F(ab')2 in the tumor than the '"I-labeled antibody. The tumoruptake of the '"In-NP-4 F(ab')2 was slightly higher than the88Y-labeled antibody on Days 1 and 3, but they were identicalby Day 7. Tumor accretion of the radiometal-labeled Fab'fragments was between 2 and 8 times higher than the "'I-Fab',but the rate at which the radiolabeled Fab' was removed from

the tumor was about the same.No significant differences were seen in the percentage of

radioactivity in the blood over time between the radiometal-labeled or radioiodinated antibodies. However, the fragments

5 R. M. Sharkey. unpublished data.

2331

BIODISTRIBI'TION AND DOSIMETRI FOR VERSUS wY-lgC OR FRAGMENTS

80

60

40-

20

O

E

Oín5o

30

20-

10

O

« 9

6

3

NP-4 IgG

NP-4 F(ab')2

NP-4 Fab'

0 Y-88

«I-1311 A ln-1 11

02468

Days Post Antibody InjectionFig. 1. Percentage of uptake/g of tumor for '"I-, '"In-, or 8gY-labeled NP-4

IgG, F(ab')¡.and Fab' fragments. ID. injected dose: bars, SD.

Days Post Antibody InjectionFig. 2. Tumor/blood ratios for radiolabelcd NP-4 and fragments. Bars, SD.

were cleared more rapidly than the IgG. For example, on Day1, the percentage of injected dose/g of blood for the radiolabeledIgG was between 16 and 21% and by Day 7 was reduced to6.0-8.9%. For the F(ab')2 fragments, on Day 1 there was 2.5-5.8%/g in the blood, and by Day 7 this was reduced to 0.05-0.12%. The percentage of injected dose/g in the blood with theFab' fragments was between 0.2 and 0.4 on Day 1 and by Day3 had been reduced to 0.03-0.09%. The faster blood clearanceof the fragments resulted in improved tumor/blood ratios forthe fragments in comparison to the intact IgG (Fig. 2). Tumor/blood ratios for the "'In-NP-4 IgG were slightly higher thanfor "'I-NP-4 IgG on Day 7, reflecting the higher tumor uptake

of the "'In-NP-4 IgG on this day. Overall, the tumor/bloodratios for the radiometal-labeled fragments were higher thanfor '"I-labeled fragments. The higher tumor/blood ratios forthe 88Y-NP-4 fragments, in comparison to the "'In-NP-4 fragments on Days 3 and 7 [Fab' and F(ab')2, respectively], were

due to differences in blood concentration on these days [0.05 ±0.02 (SD) versus 0.12 ±0.01% of injected dose/g for 88Y and"'In-NP-4 F(ab')2 on Day 7 and 0.03 ±0.01 versus 0.09 ±0.03% of injected dose/g for 88Yand '"In-NP-4 Fab' on Day

3].Tumors/nontumor ratios for the kidney, liver, lung, and

spleen are presented in Fig. 3 for the NP-4 IgG labeled with'"I, 88Y, and "'In. Significantly improved tumor/liver andspleen ratios were found only on Day 7 for the "'I-NP-4 IgGin comparison to the radiometal-labeled IgG. "'I-labeled frag

ments had significantly higher tumor/nontumor ratios than the""Y-labeled NP-4 fragments on all days tested (Table 1). For

the most part, tumor/nontumor ratios for the liver, spleen, andlungs for the 8HY-labeled F(ab')2 and Fab' were similar, ranging

from about 2 to 8. Tumor/kidney ratios were significantlyhigher for the 88Y-NP-4 F(ab')2 in comparison to the ""Y-NP-4 Fab', but there was still 2-4-fold more activity in the kidney

o-t—»oL_

1_O

ED

CO

KIDNEY

LUNG

LIVER

02460246B

Days Post Antibody InjectionFig. 3. Tumor/nontumor ratios for radiolabeled NP-4 intact IgG in the kidney,

liver, lung, and spleen. Bars. SD.

Table 1 Tumor/nontumor ratios for **Y- and ' ' '¡-labeledNP-4 fragments

Days postin-jectionF(ab')2137Fab'123OrganLiverSpleenKidneyLungLiverSpleenKidneyLungLiverSpleenKidneyLungLiverSpleenKidneyLungLiverSpleenKidneyLungLiverSpleenKidneyLung,3,,

-7.1

±1.9°6.7

±2.73.7±1.23.0±0.726.6±6.627.4±7.614.2±3.614.5

±4.839.0±16.449.5±28.525.5±10.635.5±17.113.614.62.27.135.452.47.621.918.733.25.22.74.91.82.512.026.21.512.08.213.51.615.9

±6.1My1.5

±0.42.2±0.50.2

±0.13.1±1.12.1±0.53.6±1.50.4±0.18.2±2.62.2±0.22.8+0.50.6

±0.16.8±1.52.7

±1.23.5±1.90.04±0.014.9±2.42.1±0.82.6±1.20.03±0.024.2±1.62.3±0.33.0±0.30.06±0.014.6±1.0

1Mean ±SD (n = 4-8).

2332

BIODISTRIBUTION AND DOSIMETRY FOR '"I- IfRSl'S "Y-lgCi OR FRAGMENTS

than in the tumor. Tumor/kidney ratios for "'I-fragments were

never less than 1, increased over time, and were 40 to almost200 times higher than 88Y-labeled F(ab')2 and Fab', respec

tively.The increased uptake of 88Y- in comparison to '"I-labeled

NP-4 IgG and fragments in the normal tissues is shown in Fig.4, using the liver and kidneys as representative tissues. Forradioiodinated NP-4 IgG, 5% of the injected dose/g was in theliver on Day 1, which decreased to 1.5% by Day 7. For the 88Y-NP-4 IgG, the activity in the liver remained between 5 and 7%throughout this entire period. Liver uptake of the 88Y-NP-4F(ab')2 was not significantly less than intact IgG on Days 1-7,but for the "'I-NP-4 F(ab')2, activity in the liver was 1.5% on

Day 1 and decreased to 0.04% on Day 7. Liver uptake of the88Y-Fab' fragments was reduced only 2-3-fold in comparisonto the IgG and F(ab')2, but by Day 1, liver uptake of '"I NP-4Fab' was 10 to 40 times less than the amount seen with '"I-labeled F(ab')2 or IgG, respectively. The difference between the

uptake of radioiodinated and radiometal-labeled fragments innormal tissues was most striking in the kidneys. The amountof activity in the kidney for "'I-NP-4 IgG was similar to theamount of 88Y-labeled NP-4 IgG for the first 3 days, but byDay 7 the '"I activity decreased to a level 2-fold lower than the88Y-NP-4 IgG. This is in sharp contrast to the tremendousretention of radiometal-labeled fragments in the kidney. Thepercentage of injected dose/g in the kidney for the '"I-NP-4F(ab')2 was 3% on Day 1 and decreased to 0.5% by Day 3. For'"I-NP-4 Fab', the percentage decreased from 1.9 to 0.1% onDays 1 and 3. With the 88Y-NP-4 F(ab')2, the percentage of

uptake in the kidney started at 50% on Day 1 and reduced to28% by Day 3. The 88Y-NP-4 Fab' initially was 180%, reducingto 74% by Day 3. Thus, the level of 88Y activity in kidney forthe fragments was reduced only 2-fold from Days 1 to 3, butradioiodinated fragments were removed from the kidney at asignificantly faster rate.

Since 9"Y is known for its accretion in bone (25), we used 88Y

to quantitate the amount of activity in the femur. In addition,we compared the bone data for "'In and 88Y to determinewhether '"In may be used to predict 88Y uptake. Of the two

femurs removed from each animal, one was washed with salineto determine the proportion of activity in the bone cortex versusthe marrow. Approximately 3-5% of injected dose/g was foundin the unwashed bone for both the '"In- or 88Y-labeled IgG andF(ab')2, with 2-3-fold less activity in the unwashed bones fromanimals given the Fab'. These levels were maintained over the

1-131 Y-88

100.00

10.00

O) 1.00ZiOT 0.10n'— 0.01

(j) 100.00

f> 10.00

0.10

0.01

Liver

Kidney

Liver

Kidney

o IqG•¿�r(ab')2A Fob1

Days Post Antibody InjectionFig. 4. Percentage of uptake/g in the liver or kidney for "'I-labeled (Left) or

"Y-labeled (Ri/!/iI) NP-4 IgG or its fragments. ID. injected dose; bars, SD.

duration of each study. More than 50% of the activity remainedin the bone after washing, regardless of the isotope, but by thelast day tested, the amount of 88Y in the washed bones washigher than the '"In activity, indicating a higher retention of88Y. The percentage of injected dose/g in the unwashed bonetaken from animals given '"I-NP-4 IgG was similar to thepercentages found with "'In and 88Yon Days 1 and 3, but there

was a steady decline in the percentage over time (5.7 ±2.2, 2.8±1.0, and 1.3 ±0.3 on Days 1, 3, and 7, respectively) suchthat by Day 7 significantly less activity was in the bone ofanimals given '"I-NP-4 IgG. Furthermore, after washing thebone from animals given "'I-NP-4 IgG, only 25-32% of the

activity remained in the bone. Histológica! examination ofwashed and unwashed bone samples indicated that some residual marrow remained in the sinuses of the bone despite washingwith saline. Therefore, the estimates provided here clearly donot discriminate completely the amount of activity in the cortical portion of the bone versus the marrow but rather representa partial removal of activity in the red marrow.

Radiation Dose Estimates. Estimates of radiation doses aregiven in Table 2 with normalization of the tissue doses to aconstant tumor dose. In order to give 5000 rads to the GW-39tumor, 0.29 mCi of "'I-NP-4 IgG and 0.34 mCi of 90Y-NP-4IgG would be required, but 2.5-10 times less '"Y-NP-4 fragments would be required than '"I-NP-4 fragments. The bonemarrow is the dose-limiting organ with radiolabeled IgG, asshown by the high dose to the blood when compared to thetumor. At equal tumor doses, 90Y-NP-4 IgG or fragments woulddeliver a higher dose to the normal tissues than '"I-NP-4 IgG.For 90Y-labeled NP-4 F(ab')2, the kidney may be the dose-limiting organ, but for "'I-NP-4 F(ab')2, the blood, i.e., themarrow, remains the dose-limiting organ. With Fab' fragments,the kidneys become the dose-limiting organ for both '"I- and90Y-labeled NP-4 Fab'. However, if the higher dose allowance

to the kidney as compared to the marrow is considered (byexternal beam, sublethal dose to the marrow is 200 versus 1500rads for the kidney (26); marrow toxicity may remain the dose-limiting organ for the '"I-labeled Fab', but the kidneys wouldprobably remain the dose-limiting tissue for the '"'Y-labeledFab'.

DISCUSSION

Many factors need to be considered when developing radio-labeled antibodies for tumor therapy. The choice of radio-nuclides has only recently been added to these factors due tothe advances in radioimmunochemistry. '"I has been the radi-

oisotope principally used in these studies for almost 30 years,but antibodies have since been labeled with isotopes such as90Y, 1S8Re,""Re, 67Cu, 2l2Bi, 2("At, and '09Pd (12, 13, 27-33).

Although the chemistry involved in labeling antibodies withthese nuclides has advanced, the availability, purity, and costof some of these isotopes remains the most formidable obstacleto the further development of this research. Of the nuclideslisted above, 9<IYhas surfaced as the first to breach theseobstacles. We must now carefully examine whether 90Y is a

suitable radionuclide for RAIT and whether it will be superiorto '"I.

The studies described herein have compared the biodistribu-tion of an anti-CEA MAb as an intact IgG to its F(ab')2 andFab' fragments labeled with either 88Y or '"I. 88Y was usedinstead of 9<IYto facilitate the quantitation of radioactivity intissues. We also included '"In-labeled NP-4 IgG and fragmentsto determine whether '"In may be used to predict 90Y distribution. If 9"Y-labeled NP-4 were to be used clinically, '"In may

2333

BIODISTRIBl'TION AND DOSIMETRY FOR VERSl'S ""Y-IgG OR FRAGMENTS

Table 2 Radiation dose estimates to organs in nude mice given '"/- or mY-labeled NP-4 IgG or fragments normalised to a 5000-rad dose to the tumor"

Tissue

F(ab')2

* Dose estimates for "°Y-labeledantibody was based on "Y-labeled antibody biodistribution.* Values in parentheses are the mCi of each that would be required to give a 5000-rad dose to the GW-39 tumor.' ND, not determined.

Fab'

Tumor(mCi)"BoneBloodKidneyLiverLungSpleen5.000(0.29)2612,1781.1099071,3836495.000(0.34)6044,0361,2442,4611,6491.4105.000(1.37)ND<8547013748093865.000(0.54)1.14194714,5143,8859521.8495.000(13.1)ND7031,92644S9043805,000(1.3)1,0443241

23.8603.0361.1171.652

be used to estimate the distribution pattern of 90Y for calcula

tion of radiation doses. For example. Leichner et al. (34) haveused the distribution of '"In-labeled anti-ferritin antibody topredict radiation doses to normal tissues of 90Y-anti-ferritin

antibody, because they claim to have observed comparabledistribution patterns in animal and humans. However, Hnatow-ich et al. (35) showed that '"In and 9"Y-labeled antibodies do

not distribute identically in all tissues, especially in the bone.In this study, the percentage of 88Y- or '"In-labeled NP-4 in

the bone was similar, but there were differences in the totalamount of '"In and 88Y activity remaining in the bones after

the marrow was partially removed by washing. RadioiodinatedNP-4 IgG was shown to have a similar percentage of uptake inthe bone on Days 1 and 3, but there was a steady decline in the"'I activity, such that by Day 7 there was significantly less '•"!activity in the bone than that seen with 88Yor '"In. Although

the procedure used to wash the marrow from the bones wasincomplete, it was apparent that less 88Y activity could beremoved than '"In or '"I. It is this additional amount of

yttrium activity retained by the bone that will escalate the actualmarrow dose in comparison to estimates provided from "'In-

labeled antibody. Thus, because more precise measures of dosecalculations for *"Y in the bone are lacking, careful correlationof myelotoxicity to estimated doses derived from "'In will berequired for future clinical applications of'"'Y-labeled antibody.

We have shown that 90Y-labeled antibodies also have the

additional problem of bone accretion (12) that probably occursas a consequence of the metabolic processing of the 90Y-labeledantibody in the serum and tissues. Bone uptake of 9"Y acetate

was demonstrated recently by Anderson-Berg et al. (13), butthe molecular form of TOYincorporated in the bone fromanimals given ""'Y-labeled antibodies is unknown. It is uncertainwhether chelates with a higher affinity to bind '"'Y will solvethe problem of 9"Y uptake in the bone. Chelates that could bind""'Ymore tightly may prevent leaching of 9"Y from the antibody

chetate in the serum, but the radioantibody will still be processed by many tissues in the body. It is this added metabolicprocessing by the tissues and the subsequent sequestration orrelease of the 90Y that new chelation technology must also

address. Ward et al. (36) have shown that it is difficult toremove radiometals from the tissues by the administration ofchelates once they have been internalized. However, Stewart etal. (15) reported a higher percentage of 90Y excreted in theurine of patients given "'Y-labeled antibody with i.v. EDTA

than without EDTA. Thus, clinical interventions that may alterthe metabolic processing of the 90Y-Iabeled antibody or itsmetabolites may play an important role in the future of radi-ometal-labeled antibody applications.

One reason for the uptake of 88Y-NP-4 IgG in the bone may

have been the increased time spent in the circulation. Thus, itwas interesting that, despite the more rapid removal of theF(ab')2 from the blood, there was just as much accumulation of

88Y in the bone with the F(ab')2 as the IgG. This may suggestthat the binding of ""Y to the F(ab'); was not as strong as tothe IgG. Radionuclides without the bone-seeking ability of 90Y

may provide a more reasonable alternative for RAIT by reducing the impact of at least one obstacle, namely, bone deposition.In this regard, the principal candidate would be isotopes ofrhenium (186Reor 188Re).

Radiation dose estimates provided in Table 2 were normalized to an equal rad dose to the tumor for either 90Y- or '"I-NP-4 IgG or its fragments in order to evaluate normal tissuedoses at doses of potentially equal tumoricidal activity. Sincethe clearance rate of radioactivity in the blood has been suggested as a means of predicting marrow toxicity (37, 38), andin our clinical experience, marrow toxicity has correlated wellwith bone marrow doses determined from blood clearancestudies (38), further normalization based on an equal blooddose may reflect the expected tumor dose at equal myelotoxiclevels. Both blood and bone doses for 90Y-NP-4 IgG were about2-fold higher than that calculated for '"I-NP-4 IgG. Preliminary therapy trials with '"I- and 9"Y-labeled NP-4 IgG in this

same tumor model have indicated an MTD dose (dose at which100% of the animals survive with less than 20% weight loss) of0.25 mCi for "'I-labeled NP-4 IgG and between 0.05 and 0.1mCi for ""Y-NP-4 IgG.6 Normalizing the tumor doses at theMTD for each radiolabeled NP-4 IgG, we find that the dose tothe tumor for "°Y-and "'I-labeled NP-4 IgG would be 1470

rads (0.1 mCi) and 4310 rads, respectively. Thus, at the assumedMTD, '"I-NP-4 IgG would provide as much as a 5.9-foldhigher dose to the tumor than 9"Y-NP-4 IgG. It should beemphasized that these dose calculations are based on 0.1-gtumors. As the tumor size increases, the difference between90Y- and '"I-labeled M Ab decreases such that in a 10-g tumorthere is only a 1.8-fold difference. Thus, 9"Y-labeled MAbs mayhave an advantage over '"I-labeled MAb in patients with bulky

disease.In the nude mouse, doses at the MTD to the blood and bone,

respectively, for either 90Y (0.05-0.1 mCi) or '"I (0.25 mCi)would be 1878 and 225 rads for '"I-NP-4 IgG and 593-1186and 89-178 rads for 90Y-NP-4 IgG. The discrepancy between

dose estimates for the blood and bone and the observed toxicitywith '"I- and 90Y-labeled NP-4 may reflect inaccuracies in the

calculations using the current medical internal radiation doseapproximations. Although we have tried to account for severalvariables in our dose calculations (i.e., using newly calculatedS values for small tissues and assuming a cylindrical shape forthe bone rather than a sphere), it is clear that further evaluationof radiation dose estimates coupled with direct comparisons oftumoricidal activity of both U'I- and 90Y-NP-4 IgG and the

MTD for each must be pursued.Our biodistribution studies have confirmed several of the

>Unpublished.

2334

BIODISTRIBUTION AND DOSIMETRY FOR "'1- lEKSLS "°V-lgC¡OR FRAGMENTS

observations already reported by others, namely, higher uptakeof radiometal-labeled antibody than radioiodine immunocon-

jugates in normal tissues, especially in the kidney, when antibody fragments have been used (39-42). Of course, quantitativedifferences in tumor uptake and other distribution propertiesbetween our study and others can be related to many otherfactors, most notably different tumor models, antibodies, andlabeling procedures. Nevertheless, the increased uptake in normal tissues and the retention in the kidney have been observeduniversally. Above all, the kidney uptake is the most difficultproblem to resolve if radiometal-labeled fragments are to beused for RAIT. If this is a general problem for all radiometal-labeled antibodies, fragments may not be a practical approachfor RAIT with this type of radionuclide. Experience with external beam irradiation has shown that a tumor/kidney dose ratioof 2.5 is necessary to expect a therapeutic advantage (26).Although the dose relationship between the tumor and normalorgans may be different for radiolabeled antibodies in comparison to external beam therapy, either a reduction in kidneyuptake or an increase in tumor targeting may be necessarybefore considering ""'Y-labeled fragments. In contrast, the bonemarrow remains the dose-limiting organ for '"I-labeled F(ab')2.

This allows for the consideration of fractionated schedules for'"I-F(ab')2. We have shown that 3 injections of 2 mCi of '"I-NP-4 F(ab')2 given over 6 days was equally tumoricidal as asingle 2-mCi injection of '"I-NP-4 IgG, but there was signifi

cantly less toxicity to the animals given the fractionated doseof fragments (43). This indicates that the doses of the fragmentscould be escalated further and thereby increase the tumoricidalability of this treatment in comparison to a single injection of'•"I-NP-4IgG. Indeed, in hamsters we have escalated '"I-labeled F(ab')2 doses to 12 mCi given over 6 days withoutsignificant loss in body weight of the hamsters.5 Multiple injections of "'I-NP-4 IgG given below the MTD have not been

tested, but predictions from biodistribution studies indicate thatmultiple treatment of fractionated doses of fragments may havean advantage over multiple treatments of intact IgG. If multipleinjections of either an IgG or a fragment is considered inpatients, the need for developing ways to circumvent a humananti-mouse antibody response will be prominent.

In summary, despite the more therapeutically efficaciousenergy emissions of'"'Y, antibodies labeled with '"I may haveequal to or even better therapeutic prospects than 9"Y-labeledantibody, principally due to the increased toxicity of g"Y-labeled

antibody brought on by the prolonged and higher retention of90Y in the normal organs, especially the bone. It is becomingincreasingly apparent that fractionated dose schedules for '"I-

labeled fragments will improve the tumoricidal effects whileminimizing the toxicity to normal tissues, yet antibody fragments labeled with WY or other radiometals may increase the

risk of renal toxicity in addition to myelotoxicity. These conclusions are based on i.v. injected NP-4 and may not be truefor other MAbs or injections that may allow for a higherproportion of the injected antibody to concentrate in the tumorin comparison to the normal tissues. Intrahepatic or i.p. injections may serve this function. However, there have been reportsof dose-limiting toxicity for these applications (14, 15, 44), andfor the treatment of disseminated disease, '"I-labeled antibod

ies may be the isotope of choice. Improvements in chelationtechnology or the development of techniques to modify '"Y

accretion in the bone or other tissues may also improve theprospects for use of ""'Y in RAIT. Isotopes other than '"I or90Y, such as '""Re or 67Cu, are also worthy of evaluation in

RAIT studies.

ACKNOWLEDGMENTS

We thank Drs. O. A. Gansow and M. W. Brechbiel for the preparation and labeling of the MAb-DTPA conjugates, R. Aninipot fortechnical assistance, and C. Ballance for the preparation of the manuscript.

REFERENCES

1. Goldenberg, D. M.. Gaffar, S. A., Bennett, S. J., and Beach, J. L. Experimental radioimmunotherapy of a xenografted human colonie tumor (GW-39) producing CEA. Cancer Res.. 41: 4354-4360. 1981.

2. Sharkey, R. M., Pykctt, M. J.. Siegel, J. A., Alger. E. A., Primus, F. J.. andGoldenberg. D. M. Radioimmunotherapy of the GW-39 human colonietumor xenograft with "'I-labeled murine monoclonal antibody to carcinoembryonic antigen. Cancer Res., 47: 5672-5677. 1987.

3. Jones, D. H., Goldman. A., Gordon, I.. Pritchard. J.. Gregory. B. J.. andKemshcad. J. T. Therapeutic application of a radiolabeled monoclonal antibody in nude mice xenografted with human neuroblastoma: Tumoricidaleffects and distribution studies. Int. J. Cancer, 35: 715-720. 1985.

4. Badger. C. C.. Krohn. K. A.. Shulman. H.. Flournoy. N., and Berstein. I. D.Experimental radioimmunotherapy of murine lymphoma with "'I-labelcdanti-T-cell antibodies. Cancer Res.. 46:6223-6228. 1986.

5. Zalcberg, J. R.. Thompson. C. H.. Lichtenstein, M., and McKensie. I. F. C.Tumor immunotherapy in the mouse with the use of "'I-labeled monoclonalantibody. J. Nail. Cancer Insù 72: 697-704, 1984.

6. Ettinger. D. S.. Order, S. E.. Wharam. M. D.. Parker, M. K., Klein, J. L..and Leichner. P. K. Phase I-II study of isotopie immunoglobulin therapy forprimary liver cancer. Cancer Treat. Rep.. 66: 289-297, 1982.

7. Carrasquillo, J. A.. Krohn. K. A., Beaumier. P.. McGuffin. R. W.. Brown. J.P.. Hellstrom, K. E., Hellstrom, L. and Larson. S. M. Diagnosis of andtherapy for solid tumors with radiolabeled antibodies and immune fragments.Cancer Treat. Rep.. 68: 317-328, 1984.

8. Lashford. L., Jones. D., Pritchard. J., Gordon. I.. Breatnach. F.. and Kerns-head. J. T. Therapeutic application of radiolaboled monoclonal antibodyUJ13A in children with disseminated ncuroblastoma. Nail. Cancer Inst.Monogr., 3: 53-57. 1987.

9. DeNardo, S. J., DeNardo, G. L.. O'Grady, L. F., Hu. E.. Sytsma, V. M.,

Mills, S. L., Levy. N. B., Macey, D. J.. Miller, C. H.. and Epstein, A. L.Treatment of B cell malignancies with "'I Lym-1 monoclonal antibodies.Int. J. Cancer, 3 (Suppl): 96-101. 1988.

10. Horowitz, J. A., Goldenberg. D. M.. DeJager. R.. Alger. E., Stowe, S., Hall,T.. Sharkcy. R. M.. Ford, E.. Pawlyk, D.. Siegel. J.. Ballance. C., and Varga,D. Phase I/II trial of radioimmunotherapy (RAIT) with "'I-labcled anti-CEA and anti-AFP monoclonal antibodies. J. Nucí.Med.. 2V: 846. 1988.

11. Bale. W. F.. and Spar, I. L. Studies directed toward the use of antibodies ascarriers of radioactivity for therapy. Adv. Biol. Med. Phys.. 5: 285-356,1957.

12. Sharkey. R. M.. Kaltovich. F. A.. Shih. L. B.. Fand. I.. Govelitz, G., andGoldenberg, D. M. Radioimmunotherapy of human colonie cancer xeno-grafts with *°Y-labcledmonoclonal antibodies to carcinoembryonic antigen.Cancer Res.. 48: 3270-3275, 1988.

13. Anderson-Berg, W. T.. Squire, R. A., and Strand. M. Specific radioimmunotherapy using *°V-labeled monoclonal antibody in erythroleukemic mice.Cancer Res.. 47: 1905-1912, 1987.

14. Hnatowich, D. J.. Chinol, M.. Siebecker, D. A.. Gionet. M., Griffin, T.,Doherty, P. W.. Hunter, R.. and Käse,K. R. Patient biodistribution ofintraperitoneally administered yttrium-90-labeled antibody. J. Nucí.Med.,29: 1428-1434. 1988.

15. Stewart. J. S., Hird. V.. Snook, D.. Sullivan. M., Myers. M. J..and Epcnetos.A. A. Intraperitoncal "'I- and ""Y-labellcd monoclonal antibodies for ovarian

cancer: pharmacokinctics and normal tissue dosimetry. Int. J. Cancer, 3(Suppl.): 71-76, 1988.

16. Brechbiel, M. W., Gansow, O. A., Atcher. R. W., Schlom, R. W., Esteban,J.. Simpson. D. E.. and Colcher. D. Synthesis of l-(jMsothiocyanatobenzyl)derivatives of DTPA and EDTA: antibody labeling and tumor-imagingstudies. Inorg. Chem.. 25: 2772-2781. 1986.

17. Esteban. J. M.. Schlom. J.. Gansow, O. A., Atcher, R. W.. Brechbiel. M. W.,Simpson. D. E.. and Colcher. D. New method for the chelation of indium-111 to monoclonal antibodies: Biodistribution and imaging of nude micebearing human colonie carcinoma xenografts. J. Nucí.Med., 28: 861-870,1987.

18. Goldenberg, D. M.. and Hansen, H. J. Carcinoembryonic antigen present inhuman colonie neoplasia serially propagated in hamsters. Science (Wash.DC). 775: 1117-1118, 1972.

19. Siegel. J. A., and Stabin. M. G. Adsorbed fractions for electrons and betaparticles in small spheres. J. Nucí.Med.. 29: 803. 1988.

20. Hui. T. E.. and Poston, J. \V. A model of the circulating blood for use inradiation dose calculations. In: Proceedings of the International Conferenceon Radiation Dosimetry and Safety, pp. 151-168. Taiwan. 1987.

21. Loevinger. R.. Japha. E. M.. and Browne!!. G. L. Discrete Radioisotopesources. /n:G. L. Hiñeand G. L. Brownell (eds.). Radiation Dosimetry. pp.712. New York: Academic Press, 1956.

22. Cloutier, R. J.. and Watson. E. E. Radiation dose from radioisotopcs in theblood. In: Medical Radionuclidcs: Radiation Dose and Effects. AEC Symposium Series 20. Conference G912I2. p. 332. Oak Ridge. TN: Oak RidgeNational Laboratories, 1970.

2335

BIODISTRIBUTION AND DOSIMETRY FOR "'I- YERSL'S *>Y-IgG OR FRAGMENTS

23. MIRD pamphlet 10, New York Society of Nuclear Medicine, 1975.24. Sokal, R. R., and Rohlf, F. J. Biometry, pp. 226-235. San Francisco: W. H.

Freeman and Co., 1969.25. Catsch. A. In: I. N. Kugelmass (ed.). Radioactive Metal Mobilization in

Medicine. Springfield, IL: Charles C Thomas Publishing, 1964.26. Fawwaz, R. A., Wang, T. S. T., Srivastava, S. C.. and Hardy, M. A. The use

of radionuclides for tumor therapy. Nucí.Med. Biol., 13: 429-436, 1986.27. Deutsch, E., Libson. K., Vanderheyden, J-L., Ketring, A. R.. and Maxon, H.

R. The chemistry of rhenium and technetium as related to the use of isotopesof these elements in therapeutic and diagnostic nuclear medicine. Nucí.Med.Biol.. 13:465-477, 1986.

28. Cole, W. C, DeNardo. S. J.. Meares. C. F., McCall. M. J., DeNardo, G. L..Epstein, A. L., O'Brien, H. A., and Moi, M. K. Serum stability of Cu-67chelates: Comparison with '"In and Co-57. Nucí.Med. Biol., 13: 363-368.

1986.29. Kozak. R. W., Atcher, R. W., Gansow, O. A., Friedman, A. M., Hiñes,J. J.,

and Waldmann, T. A. Bismuth-212-labeled anti-Tac monoclonal antibody:Alpha-particle-emitting radionuclides as modalities for radioimmunotherapy.Proc. Nati. Acad. Sci. USA, 83: 474-478. 1986.

30. Vaughan, A. T. M., Bateman. W. J., and Fisher, D. R. The in vivo fate of aAt-211 labelled monoclonal antibody with known specificity in a murinesystem. Int. J. Radial. Oncol. Biol. Phys., S: 1943-1946, 1982.

31. Fawwaz, R. A., Wang, T. S. T., Srivastava, S. C.. Rosen, J. M., Ferrone, S.,Hardy, M. A., and Alderson, P. O. Potential of palladium-109-labeledantimelanoma monoclonal antibody for tumor therapy. J. Nucí.Med., 25:796-799, 1984.

32. Anderson. P.. Vaughan. A. T. M.. and Varley. N. R. Antibodies labeled withAu-199: Potential of Au-199 for radioimmunotherapy. Nucí.Med. Biol.. 15:293-297. 1988.

33. Washburn, L. C., Lee, Y-C. C., Sun, T. T. H., Byrd, B., Crook. J. E.. Stabin.M. G., and Steplewski, Z. Preclinical assessment of ""Y-labeled monoclonalantibody CO 17-1A, a potential agent for radioimmunotherapy of colorectalcarcinoma. Nucí.Med. Biol.. 15: 707-711, 1988.

34. Leichner. P. K.. Yang. N-C.. Frenkel, T. L., Loudenslager, D. M., Hawkins,W. G.. Klein, J. L., and Order, S. E. Dosimetry and treatment planning for"°Y-labeledantiferritin in hepatoma. Int. J. Radiât.Oncol. Biol. Phys., 14:1033-1042. 1988.

35. Hnatowich. D. J., Virzi, F., and Doherty, P. W. DTPA-coupled antibodieslabeled with yttrium-90. J. Nucí.Med.. 26: 503-509, 1985.

36. Ward, M. C.. Roberts, K. R., Westwood. J. H.. Coombes, R. C. C.. andMcCready. V. R. The effect of dictating agents on the distribution ofmonoclonal antibodies in mice. J. Nucí.Med., 27: 1746-1750, 1986.

37. Bigler, R. E., Zanzonico. P. B.. Primus, F. J., Alger, E., DeJager, R. Stowe,S., Ford, E. Goldenberg, D. M., Leonard, R., and Cosma. M. Bone marrowdosimetry for monoclonal antibody therapy. In: Proceedings of the FourthInternational Radiopharmaceutical Dosimetry Symposium, pp. 535-544.Oak Ridge. TN: Oak Ridge Associated Universities, 1986.

38. Siegel, J. A., Pawlyk, D. A.. Lee. R. E.. Sasso, N. L.. Horowitz, J. A..Sharkey, R. M., and Goldenberg. D. M. Tumor, red marrow, and organdosimetry for I3'l-labeled anticarcinoembryonic antigen monoclonal antibody. Cancer Res., 50: (Suppl.): 1039s-1042s, 1990.

39. Andrew, S. M., Perkins, A. C.. Pimm. M. V.. and Baldwin, R. A. Acomparison of iodine and indium labelled anti CEA intact antibody. F(ab'h.and Fab' fragments by imaging tumour xenografts. Eur. J. Nucí.Med., 13:598-504. 1988.

40. Otsuka, F. L.. Fleischman. J. B.. and Welch. M. J. Comparative studies usingC25I-and '"In-labeled monoclonal antibodies. Nucí.Med. Biol., 13: 325-334,

1986.41. Sakahara. H., Endo, K., Nakashima, T., Koizumi. M., Kunimatsu, M.,

Kawamura. Y.. Ohta, H., Nakamura, T.. Tanaka, H., Kotoura. Y., Yama-muro. T.. Hosoi, S., Toyama, S., and Torizuka. K. Localization of humanosteogenic sarcoma xenografts in nude mice by a monoclonal antibody labeledwith radioiodine and indium-111. J. Nucí.Med.. 28: 342-348. 1987.

42. Brown, B. A., Comeau, R. D., Jones, P. L.. Liberatore, F. A., Neacy, W. P..Sands, H.. and Gallagher, B. M. Pharmacokinetics of the monoclonal antibody B72.3 and its fragments labeled with either 1MI-or '"In. Cancer Res.,47: 1149-1154. 1987.

43. Blumenthal. R. D., Sharkey, R. M., Kashi. R., and Goldenberg, D. M.Comparison of therapeutic efficacy and host toxicity of two different 131I-labcled antibodies and their fragments in the GW-39 colonie cancer xenograftmodel. Int. J. Cancer, 44: 292-300, 1989.

44. Order. S. E.. Vriesendorp. H. M., Klein. J. L.. and Leichner, P. K. A phaseI study of yttrium-90 anti-ferritin: dose escalation and tumor dose. AntibodyImmunoconjugate Radiopharmacol.. /: 163-168, 1988.

2336