Phase I study of (131)I-anti-CD45 antibody plus cyclophosphamide and total body irradiation for advanced acute leukemia and myelodysplastic syndrome.

Delivery of targeted hematopoietic irradiation using radiolabeled monoclonal antibody may improve the outcome of marrow transplantation for advanced acute leukemia by decreasing relapse without increasing toxicity. We conducted a phase I study that examined the biodistribution of (131)I-labeled anti-CD45 antibody and determined the toxicity of escalating doses of targeted radiation combined with 120 mg/kg cyclophosphamide (CY) and 12 Gy total body irradiation (TBI) followed by HLA-matched related allogeneic or autologous transplant. Forty-four patients with advanced acute leukemia or myelodysplasia received a biodistribution dose of 0.5 mg/kg (131)I-BC8 (murine anti-CD45) antibody. The mean +/- SEM estimated radiation absorbed dose (centigray per millicurie of (131)I) delivered to bone marrow and spleen was 6.5 +/- 0.5 and 13.5 +/- 1.3, respectively, with liver, lung, kidney, and total body receiving lower amounts of 2.8 +/- 0.2, 1.8 +/- 0.1, 0.6 +/- 0.04, and 0.4 +/- 0.02, respectively. Thirty-seven patients (84%) had favorable biodistribution of antibody, with a higher estimated radiation absorbed dose to marrow and spleen than to normal organs. Thirty-four patients received a therapeutic dose of (131)I-antibody labeled with 76 to 612 mCi (131)I to deliver estimated radiation absorbed doses to liver (normal organ receiving the highest dose) of 3.5 Gy (level 1) to 12.25 Gy (level 6) in addition to CY and TBI. The maximum tolerated dose was level 5 (delivering 10.5 Gy to liver), with grade III/IV mucositis in 2 of 2 patients treated at level 6. Of 25 treated patients with acute myeloid leukemia (AML)/myelodysplastic syndrome (MDS), 7 survive disease-free 15 to 89 months (median, 65 months) posttransplant. Of 9 treated patients with acute lymphoblastic leukemia (ALL), 3 survive disease-free 19, 54, and 66 months posttransplant. We conclude that (131)I-anti-CD45 antibody can safely deliver substantial supplemental doses of radiation to bone marrow (approximately 24 Gy) and spleen (approximately 50 Gy) when combined with conventional CY/TBI.

Radiation quality affects the efficiency of induction and the molecular spectrum of HPRT mutations in human T cells.

Human T lymphocytes can be used to determine the frequency and molecular spectrum of somatic cell gene mutations induced by ionizing radiations both in vivo and in vitro. In vitro exposure of these G0 cells to low-LET 137Cs gamma rays results in the induction of HPRT mutations and a predominant molecular spectrum of DNA deletions and rearrangements, particularly total gene deletions (11-12%). Similar results are found in samples from humans exposed to low-LET radiation from 131I. The doubling dose for mutation induction is calculated to be 0.8 and 1.0 Gy from these exposures performed in vitro and in vivo, respectively. In vitro studies of the effects of high-LET radiation from exposure to 222Rn also showed an induction of HPRT mutations, with a doubling dose of approximately 0.2 Gy. With this radiation, the predominant mutations were small partial deletions, with less than 2% total gene deletions. Studies of humans exposed to high-LET radiation from 239Pu showed an increased HPRT mutant frequency for the group, although no significant dosimetry could be defined. In contrast to the humans exposed to 131I, no increase in the frequency of total gene deletions was found. This is consistent with the results for 222Rn in vitro. The available data show that radiation quality affects both the efficiency of induction and the molecular spectrum of HPRT mutations in human T lymphocytes both in vitro and in vivo. The mutational spectrum may be relatively specific for radiations of different quality and thus allow a more precise measurement of the induction of somatic gene mutations resulting from individual exposures to radiation, and thereby provide more sensitive assessments of health risks.

Factors affecting use of CR-39 surface monitor technology to estimate past exposure to indoor radon.

In an epidemiologic study investigating influences of life-style and environment on lung cancer risk, CR-39 alpha-particle detectors, identified here as surface monitors, were affixed to subjects' selected household glass, ceramic, or enameled objects to measure residual radioactivity form embedded radon (Rn) decay products. The purpose was to estimate past cumulative indoor Rn concentrations to which the object was exposed to infer past exposures of the subjects. This approach was used to supplement exposure information obtained by methods traditionally used in Rn epidemiologic studies. In addition, surface monitors were affixed to objects of selected study subjects with complete exposure information to evaluate whether surface monitors provided estimates of cumulative past residual Rn exposure comparable to estimates obtained from year-long, ambient alpha track-etch measurements in each present and previous residence. These ambient measurements were time-weighted to estimate integrated exposure of objects and were adjusted for decay and ingrowth so as to be comparable to surface monitor measurements. A regression relationship was estimated between the two estimates of cumulative RN exposure. Surface monitor measurements had a satisfactory correlation (0.63) with adjusted ambient Rn measurements for new, nonceramic objects. Although not included in the study design, factors that might affect use of the technology were also investigated. Regression relationships were compared in graduated smoking environment (as judged by the subjects) to investigate possible differential plate out of radon progeny. In addition, regression relationships for windows were compared with those for other objects to investigate whether there was a significant difference between windows and other objects. It has been suggested that windows may have a higher plate out rate because of locally increased air flow. Results suggested that surface monitor information was useful to fill time gaps in estimates of historical radon exposure data obtained by ambient measurements. Glass samples provided the best correlation. Ceramic materials sometimes provided excessively high radon estimates, probably due to glazes that contained significant uranium or thorium. Due to small sample sizes, investigations of other factors were inconclusive.

The biological effectiveness of radon-progeny alpha particles. V. Comparison of oncogenic transformation by accelerator-produced monoenergetic alpha particles and by polyenergetic alpha particles from radon progeny.

Generation of estimates of risk caused by exposure to radon in the home, either from miner data or from A-bomb data, requires several scaling factors such as for dose, dose rate and radiation quality, and possible synergisms. Such scaling factors are best developed from laboratory-based studies. Two possible sources of alpha particles for such studies are (1) a polyenergetic spectrum, generated directly by radon and its progeny, or (2) a series of monoenergetic alpha particles. We compare here the results of oncogenic transformation from studies using both systems. At the Columbia University Radiological Research Accelerator Facility (RARAF), C3H 10T1/2 cells were irradiated with alpha particles of various energies, with defined LETs from 70 to 200 keV/mum. At Pacific Northwest Laboratory, cells from the same stock were exposed to alpha particles from radon gas and its progeny, which were in equilibrium with the culture medium. There was good agreement between the results of oncogenic transformation experiments using the two different exposure systems. Apart from the experimental transformation frequencies themselves, such a comparison requires (1) reliable dosimetry at both facilities and (2) estimated LET distributions for the polyenergetic alpha-particle irradiator. Thus this good agreement gives some confirmation to the technique which is used to fold together oncogenic transformation rates from monoenergetic alpha particles to yield a predicted rate for a spectrum of alpha particles.

Development of a marrow transplant regimen for acute leukemia using targeted hematopoietic irradiation delivered by 131I-labeled anti-CD45 antibody, combined with cyclophosphamide and total body irradiation.

In an attempt to decrease the relapse rate after bone marrow transplantation (BMT) for advanced acute leukemia, we initiated studies using 131I-labeled anti-CD45 antibody (BC8) to deliver radiation specifically to hematopoietic tissues, followed by a standard transplant preparative regimen. Biodistribution studies were performed in 23 patients using 0.5 mg/kg trace 131I-labeled BC8 antibody. The BC8 antibody was cleared rapidly from plasma with an initial disappearance half-time of 1.5 +/- 0.2 hours, presumably reflecting rapid antigen-specific binding. The mean radiation absorbed doses (cGy/mCi131I administered) were as follows: marrow, 7.1 +/- 0.8; spleen, 10.8 +/- 1.4; liver, 2.7 +/- 0.2; lungs, 2.1 +/- 0.1; kidneys, 0.7 +/- 0.1; and total body, 0.4 +/- 0.03. Patients with acute myelogenous leukemia (AML) in relapse had a higher marrow dose (11.4 cGy/mCi) than those in remission (5.2 cGy/mCi; P = .001) because of higher uptake and longer retention of radionuclide in marrow. Twenty patients were treated with a dose of 131I estimated to deliver 3.5 Gy (level 1) to 7 Gy (level 3) to liver, with marrow doses of 4 to 30 Gy and spleen doses of 7 to 60 Gy, followed by 120 mg/kg cyclophosphamide (CY) and 12 Gy total body irradiation (TBI). Nine of 13 patients with AML or refractory anemia with excess blasts (RAEB) and two of seven with acute lymphocytic leukemia (ALL) are alive disease-free at 8 to 41 months (median, 17 months) after BMT. Toxicity has not been measurably greater than that of CY/TBI alone, and the maximum tolerated dose has not been reached. This study demonstrates that with the use of 131I-BC8 substantially greater doses of radiation can be delivered to hematopoietic tissues as compared with liver, lung, or kidney, which may improve the efficacy of marrow transplantation.

Molecular analysis of hypoxanthine phosphoribosyltransferase gene deletions induced by alpha- and X-radiation in human lymphoblastoid cells.

Mutations caused by exposure to X-radiation and to radon and its decay products were compared in the hprt gene of a human lymphoblastoid cell line. Thirty-one X-radiation-induced, 29 radon-induced, and 24 spontaneous mutants were recovered from cell cultures under identical conditions except for the exposure to radiation. Seven spontaneous point mutations were recovered and DNA sequenced. These mutations included three C:G-->T:A transitions. These spontaneous point mutations were located in the exon or splice donor regions of five of the nine hprt exons. Four X-radiation-induced and three radon-induced point mutations were also analyzed by DNA sequencing. The frequency of induced mutants at the D0 doses for radon and X-radiation respectively were 5 x 10(-6) and 4.5 x 10(-6). Deletions were the predominant mutations recovered from both radon- and X-irradiated cells. Eighty-one percent of the mutants from X-radiation-treated cultures, 86% of the radon-treated cultures, and 63% of the spontaneous mutants involved deletions. Deletions involving exon and intron DNA, as well as intron DNA alone, were found to inactivate the hprt gene and result in a selectable HPRT- phenotype. Among the deletion mutants, however, only 21% of the spontaneous mutants versus 55% of both the X-radiation- and radon-induced mutants exhibited loss of the entire hprt gene. More X-radiation-induced deletions than radon-induced deletions extended further than 800 bp in the telomeric direction from the hprt gene (six of 17 versus two of 17). The results show that at the human hprt locus of TK-6 cells the predominant kind of mutation indicative of exposure to both high LET alpha-radiation and low LET X-radiation is a large deletion, spanning the entire hemizygous hprt gene and extending into flanking sequences.

Calculated and TLD-based absorbed dose estimates for I-131-labeled 3F8 monoclonal antibody in a human neuroblastoma xenograft nude mouse model.

Preclinical evaluation of the therapeutic potential of radiolabeled antibodies is commonly performed in a xenografted nude mouse model. To assess therapeutic efficacy it is important to estimate the absorbed dose to the tumor and normal tissues of the nude mouse. The current study was designed to accurately measure radiation does to human neuroblastoma xenografts and normal organs in nude mice treated with I-131-labeled 3F8 monoclonal antibody (MoAb) against disialoganglioside GD2 antigen. Absorbed dose estimates were obtained using two different approaches: (1) measurement with teflon-imbedded CaSO4:Dy mini-thermoluminescent dosimeters (TLDs) and (2) calculations using mouse S-factors. The calculated total dose to tumor one week after i.v. injection of the 50 microCi I-131-3F8 MoAb was 604 cGy. The corresponding decay corrected and not corrected TLD measurements were 109 +/- 9 and 48.7 +/- 3.4 cGy respectively. The calculated to TLD-derived dose ratios for tumor ranged from 6.1 at 24 h to 5.5 at 1 week. The light output fading rate was found to depend upon the tissue type within which the TLDs were implanted. The decay rate in tumor, muscle, subcutaneous tissue and in vitro, were 9.5, 5.0, 3.7 and 0.67% per day, respectively. We have demonstrated that the type of tissue in which the TLD was implanted strongly influenced the in vivo decay of light output. Even with decay correction, a significant discrepancy was observed between MIRD-based calculated and CaSO4:Dy mini-TLD measured absorbed doses. Batch dependence, pH of the tumor or other variables associated with TLDs which are not as yet well known may account for this discrepancy.

An internal dosimetry intercomparison study.

Pacific Northwest Laboratory performed a study to evaluate the consistency of internal dosimetry assessments. A total of eleven laboratories, including DOE sites and NRC licensees, participated in this intercomparison study. Participants were asked to respond to five actual exposure scenarios, previously used in a similar European study. The participating dosimetrists assessed the data of the test scenarios and calculated results in terms of estimated radionuclide intake and the resulting internal doses. To maintain confidentiality, results are given without identifying any site. Except for one scenario, the results showed that the standard deviation of the final results on committed effective dose equivalent for each exposure scenario was about 30-50% of the mean value, giving a consistency slightly greater variant than that of the European study. The discrepancies can be attributed to variations in 1) the interpretation and statistical treatment of the bioassay data; 2) the biokinetic models applied; and 3) the computational tools used. This represents a preliminary study; further intercomparison testing is needed to fully evaluate the problem of dose-assessment inconsistency.

Application of the cross-organ beta dose method for tissue dosimetry in tumor-bearing mice treated with a 90Y-labeled immunoconjugate.

BACKGROUND: Radioimmunotherapy of nude mice bearing human tumor xenografts using 90Y-labeled monoclonal antibodies has resulted in slower tumor growth, decreased tumor burden, and increased survival times. Dosimetry estimates in the murine model usually were based on biodistribution data and standard Medical Internal Radiation Dose approaches. A new dosimetric model for the mouse that takes into consideration the small dimensions, mass, and proximity of murine organs has been developed based on self-organ absorbed and cross-organ doses. METHODS: Nude mice bearing carcinoembryonic antigen-expressing WiDr human colon cancer xenografts were injected with 240 microCi 90Y-anti-carcinoembryonic-antigen monoclonal antibodies and then killed at 12, 24, 72, 120, and 168 hours. Tumors and major organs were removed, weighed, and counted on a gamma counter. Using the resulting biodistribution data, the radiation doses to tumor and normal organs were calculated using the new dosimetric model for the mouse. RESULTS: Three organs (the liver, kidneys, and large bowel) directly received > 50% of the total absorbed beta dose from radioactivity. Lungs, stomach, and marrow received the highest percentage (70-75%) of the total absorbed dose from adjacent organs. Tumor absorbed dose, estimated with the new dosimetric model, was three times less than that obtained with a MIRD-style calculation without correction for self-absorbed and cross-organ doses. CONCLUSIONS: The new dosimetric model, which accounts more accurately for self-organ absorbed and cross-organ beta dose fraction, allows the calculation of tumor and organ doses in the murine model. Accurate estimation of radiation doses to tumor and critical organs, such as the marrow, spleen and kidneys, is important in determining the efficacy and toxicity of radioimmunotherapy regimens in animals and in subsequent human applications.

A mouse model for calculating cross-organ beta doses from yttrium-90-labeled immunoconjugates.

BACKGROUND: The organs of laboratory mice used in radioimmunotherapy experiments are relatively small compared to the ranges of high-energy yttrium-90 (Y-90) beta particles. Current Medical Internal Radiation Dose (MIRD) dosimetry methods do not account for beta energy that escapes an organ. A dosimetry model was developed to provide more realistic dose estimates for organs in mice who received Y-90-labeled antibodies by accounting for physical and geometric factors, loss of beta dose due to small organ sizes, and cross-organ doses. METHODS: The dimensions, masses, surface areas, and overlapping areas of different organs of 10 athymic nude mice, each weighing approximately 25 g, were measured to form a realistic geometric model. Major organs in this model include the liver, spleen, kidneys, lungs, heart, stomach, small intestine, large intestine, thyroid, pancreas, bone, marrow, and carcass. A subcutaneous tumor mass also was included in the model. By accounting for small organ absorbed fractions and cross-organ beta doses, the MIRD methodology was extended from humans to mice for beta dose calculations. RESULTS: Absorbed fractions of beta energy were calculated using the Berger's point kernels and the electron transport code EGS4. Except for the tumor and carcass, the self-organ absorbed fractions ranged from 15% to 20% in smaller organs (the marrow and thyroid) to 65%-70% in larger organs (the liver and small intestine). Cross-organ absorbed fractions also were calculated from estimates of the overlapping surface areas between organs. CONCLUSION: The mathematic mouse model presented here provides more realistic organ dosimetry of radiolabeled monoclonal antibodies in the nude mouse, which should, in turn, contribute to a better understanding of the correlation of biodistribution study results and organ-tumor toxicity information.

Interlaboratory comparison of the effects of radon on L5178Y cells: dose contribution of radon daughter association with cells.

The cytotoxic and mutagenic effects of radon and its progeny were compared in murine lymphoblast L5178Y-R16 cells after exposure at three institutions. The cells were exposed to 222Rn at Case Western Reserve University (CWRU) and Pacific Northwest Laboratories (PNL) and to 212Bi, a decay product of 220Rn, at the University of Chicago (UC). The dose to the cell nucleus was calculated using a dosimetric model which addressed both the contribution of the dose from the radioactivity in the medium and that associated with the cells. The dose-response curves for cell survival showed D0's of 0.30 Gy at CWRU, 0.20 Gy at PNL, 0.37 Gy for chelated 212Bi, and 0.13 Gy for unchelated 212Bi. Induced mutant frequencies at the thymidine kinase locus at the 37% survival level were 1470 x 10(-6) at CWRU, 1518 at PNL, and 2414 x 10(-6) at UC using combined results for chelated and unchelated 212Bi. The variation between institutions was greater than obtained in a previous interlaboratory comparison of the effects of radon on CHO cells. Since less radioactivity was associated with CHO cells than L5178Y cells, we have concluded that the variation between institutions in the case of L5178Y cells is caused by the differences in cell-associated radioactivity and errors related to the measurement of this parameter.

Cytotoxic and mutagenic effects of radon and radon daughters on murine L5178Y lines differing in DNA repair.

The effects of 222Rn were measured in mouse L5178Y (LY) lymphoblasts that differ in repair capabilities. Line LY-S1 is deficient in the repair of X-radiation-induced DNA doublestrand breaks, while lines LY-R16 and LY-R83 are presumed to be deficient in the excision of UV-radiation-induced pyrimidine dimers. Line LY-R83 is hemizygous while the other two lines are heterozygous at the thymidine kinase (tk) locus. After exposure to radon the D0's were found to be very similar for the three lines (0.30-0.31 Gy), whereas for X radiation the D0 for line LY-S1 is lower (0.7 Gy) than that for the two LY-R lines (1.3 Gy). Mutant frequencies at the tk locus were higher per gray after treatment with radon than X radiation, but at equitoxic doses the mutant frequencies were similar for X and alpha-particle radiation. A low radon-induced mutant frequency was observed for the hemizygous line, in agreement with the hypothesis that multilocus lesions were induced by the alpha-particle radiation and that mutants bearing intergenic lesions were not recovered in the TK+/- line. The entire active tk allele was lost by 81% of the TK-/- mutants of line LY-R16. In lines LY-S1 and LY-R16, 39-43% of the TK-/- mutants exhibited loss of galactokinase activity, indicating that the mutational lesion inactivating the tk gene frequently extended to the neighboring galactokinase gene.

Evaluation of an alpha probe detector for in vitro cellular dosimetry.

This paper describes the design and testing of an alpha probe detector for the continuous measurement of the activity concentrations of alpha emitters in the culture media of in vitro cell suspension irradiation systems. The probe detector consists of a pen-size body housing a small silicon surface-barrier detector with a Mylar window. Theoretical calculations were performed to study the dependence of the alpha-energy spectrum on 1) the thickness of the Mylar barrier; 2) the Mylar-detector distance; and 3) the size of the detector window. These design parameters were selected by taking a compromise between the counting efficiency, the integrity of the detector, and its required range of application. The probe detector was tested using both chelated and unchelated 212Bi and 212Pb standard solutions; plate-out of these radionuclides on the Mylar barrier was observed for unchelated solutions. Alpha energy spectra were analyzed using a total integration technique. The measured activity concentrations and the calibrated values agree to within 4% for the chelated 212Bi and to within 6% for the unchelated 212Bi. The alpha probe detector can be used throughout an entire exposure time period to determine the total dose received by suspended cells, or at different time intervals to determine the dose rate in real time.

Single-cell gel technique supports hit probability calculations.

We have used the alkaline single-cell gel technique to provide a biological estimate of the percentage of cell nuclei "hit" by alpha particles during in vitro radon exposure. The single-cell gel electrophoretic technique measures DNA strand breaks as increased migration of the DNA out of lysed cells embedded in the middle layer of a three-layer gel formed on a microscope slide. Two of the advantages of this system are that individual cells of an exposed population can be evaluated and that histograms can be constructed to estimate the population response. Chinese hamster ovary and AL cells were each exposed to 0.39 Gy of radon, a dose at which our dosimetry model predicts that 63 and 73%, respectively, of the cell nuclei will be traversed by an alpha particle. The difference in the percentages at similar doses is mainly due to the larger nucleus volume in AL cells. A 1.5-Gy x-ray response was also evaluated as a low-LET control. As expected, the x-ray profile of DNA damage was shifted from the nonirradiated profile in the direction of greater DNA migration and approximated a normal distribution. The profile of the radon-exposed cells was biphasic, with one distribution corresponding to the control (nonirradiated) response and the other profile showing increased DNA migration. We interpret the second profile in the biphasic profile as representing cell nuclei that had received an alpha "hit." The percentages of cell nuclei in the "hit" category (approximately 51 and 45% for CHO and AL, respectively), as judged by the single-cell gel technique, were 81 and 62% of the calculated values.

Measurement of activity in anthropomorphic organ phantoms.

This paper describes two nondestructive measurement techniques for determining the radioactivity in a homogeneous organ phantom with an acceptable error and the capability of being traceable to the National Institute of Standards and Technology. These two techniques are based on a method developed by Robley D. Evans in 1937 for measuring the amount of radium deposited in a living person. There are two significant improvements in the new techniques: 1) the radially-dependent error is eliminated, and 2) the effect of self absorption in the unknown body is measured and taken into account. The first assay method is a single-source technique involving four measurements and requiring only one standard source, the second one is a double-source technique involving six measurements and requiring two unequal standard sources. Three pairs of lung phantoms radiolabeled with 241Am, 137Cs, and 154Eu were measured with both the single- and double-source techniques. The results imply that the self-attenuation of photons within the organ phantom cannot be neglected, especially at low energies. Finally, potential applications (other than the calibration of bioassay phantoms of the single- and double-source techniques) and their limitations are discussed.

Microdistribution and microdosimetry of thorium deposited in the liver.

The distribution of thorium in the liver of a patient 36 y after injection with Thorotrast was examined with autoradiographic and scanning electron microscope backscatter image techniques. Autoradiographic examination of randomly selected histologic sections of the liver showed a total alpha activity calculated at 33.7 Bq g-1, with the highest concentration of alpha activity sequestered in subcapsular scare tissue. Subcapsular scare tissue received 4.8 cGy d-1 of alpha radiation, periportal areas were accumulating 1.4 cGy d-1, and the hepatic cord areas 0.09 cGy d-1 of alpha radiation at the time of death. The concentration of dose in periportal areas correlates with higher incidence of bile duct tumors (than hepatocellular carcinomas) found in patients exposed to Thorotrast. The backscatter technique was demonstrated as useful for identifying thorium in liver specimens.

Interlaboratory comparison of different alpha-particle and radon sources: cell survival and relative biological effectiveness.

Alpha radiation-induced cell killing was determined in four different laboratories in order to: 1) measure interlaboratory variability and 2) compare the effects of radon and radon daughter exposures with the effects of 238Pu (an often-used model for radon exposure). The results suggest that differences in handling from laboratory to laboratory can affect both low and high linear energy transfer responses and should be considered when comparing results from different laboratories.

Selective radiation of hematolymphoid tissue delivered by anti-CD45 antibody.

The ability to deliver radiation selectively to lymphohematopoietic tissues may have utility in conditions treated by myeloablative regimens followed by bone marrow transplantation. Since the CD45 antigen is the most broadly expressed of hematopoietic antigens, we examined the biodistribution of radiolabeled anti-CD45 monoclonal antibodies in normal mice. Trace 125I or 131I-labeled monoclonal antibodies 30G12 (rat IgG2a), 30F11 (rat IgG2b), and F(ab')2 fragments of 30F11 were injected i.v. at doses of 5 to 1000 micrograms. For both intact antibodies, a higher percentage of injected dose/g (% ID/g tissue) in blood was achieved with higher antibody doses. However, as the dose of antibody was increased, the % ID/g in the target organs of spleen, marrow, and lymph nodes decreased. At doses between 5 and 10-micrograms, % ID/g in these tissues exceeded that in lung, the normal organ with the highest concentration of radiolabel. In contrast, thymus was the only hematopoietic organ in which the % ID/g increased with increasing antibody dose, although at high dose the % ID/g was still far below that achieved in the other hematopoietic organs. Antibody 30F11 F(ab')2 fragments were cleared more quickly than intact antibody from blood and from both target and nontarget organs, although the relationship between increasing antibody dose and decreasing % ID/g in spleen, marrow, and lymph nodes was observed. The time-activity curves for each dose of antibody were used to calculate estimates of radiation absorbed dose to each organ. At the 10-micrograms dose of 30G12, the spleen was estimated to receive a radiation dose that was 13 times more than lung, the lymph nodes 3 to 4 times more, and the bone marrow 3 times more than lung. For each antibody fragment dose, the radiation absorbed dose per MBq 131I administered was lower because the residence times of the fragments were shorter than those of the intact antibody. Thus these estimates suggested that the best "therapeutic ratio" of radiation delivered to target organ as compared to lung was achieved with lower doses of intact antibody. We have demonstrated that radiolabeled anti-CD45 monoclonal antibodies can deliver radiation to lymphohematopoietic tissues with relative selectivity and that the relative uptake and retention in different hematolymphoid tissues change with increasing antibody dose.