Imaging organic and mineral phases in a biomineral using novel contrast techniques.

Chitons are marine molluscs that mineralize their teeth by the process of matrix-mediated biomineralization. The teeth develop in a continuous manner forming hard minerals, including magnetite, making analysis of the matrix within these mineralized regions difficult. This article describes the use of charge contrast imaging techniques, rarely applied to biological samples, to simultaneously image the organic and mineral phases within the teeth of these animals. Resulting evidence demonstrates the power of this technique in delivering architectural information concerning both the matrix and mineral phases, without the need for removal of the hard mineralized material.

Enhanced TAG-72 expression and tumor uptake of radiolabeled monoclonal antibody CC49 in metastatic breast cancer patients following alpha-interferon treatment.

The IFNs, alpha and gamma, have been shown to enhance the tumor-associated glycoprotein (TAG-72) on adenocarcinoma cells in vitro and in mice with human breast cancer xenografts, resulting in improved targeting of monoclonal antibody CC49. To determine the effect of IFN-alpha on biodistribution and tumor uptake of 131I-labeled CC49, patients with metastatic breast cancer were randomized to either receive or not receive IFN-alpha (3 million units daily for 14 days) by s.c. injection. Three days after beginning IFN-alpha, all patients received 10-20 mCi of 131I-CC49 (specific activity, 16.7 mCi/mg) i.v. Total-body Anger camera scans, along with total-body blood and plasma pharmacokinetics, were performed. Tumor biopsies were taken in all patients before and 48 h after IFN-alpha treatment. There were no significant differences in number of metastases imaged or whole-body, blood and plasma pharmacokinetics between IFN-alpha-treated and untreated patients. Quantitative immunohistochemistry on biopsy specimens from IFN-alpha-treated patients demonstrated a significant increase in mean +/- SEM TAG-72 expression (45.7 +/- 19.4%) compared to patients that were not given IFN-alpha (1.3 +/- 0.95%; P < 0.05). Although slight increases in the percent injected dose of 131I-CC49 in tumor occurred after IFN-alpha-treatment, the changes were not significant at the P < 0.05 level. These data suggest that IFN-alpha may be useful in enhancing TAG-72 antigen expression in vivo in humans, despite modest improvement in tumor uptake of CC49, possibly because of limited tumor access or other unknown factors.

The effect of histological processing on the form of iron in iron-loaded human tissues.

Iron-loaded human spleen tissue was immersed in neutral buffered formalin over a period of 200 days. Over the first 60 days, iron leached steadily from the tissue until 3% had been lost. Thereafter, no further iron leaching was detected. Comparisons of Mossbauer spectra of freeze-dried tissue and tissue freeze-dried after immersion in formalin for 200 days showed no evidence of chemical transformation of the iron remaining in the tissue. The spectra indicated a difference in the heme-iron to non-heme iron ratio between the two samples probably reflecting inhomogeneity of the ratio throughout the spleen as measured on the centimetre scale. Mossbauer spectra of freeze-dried samples of iron-loaded human liver and pancreas tissue were compared with those for samples from the same patient that had been processed by routine hospital procedures for histology and archival. These spectra showed no evidence for chemical transformation of the iron present in the tissues. These results demonstrate that it is feasible to use archived fixed and embedded human tissue samples for studies aimed at gauging the relative fraction of goethite-like hemosiderin present in the tissue.

The form of iron oxide deposits in thalassemic tissues varies between different groups of patients: a comparison between Thai beta-thalassemia/hemoglobin E patients and Australian beta-thalassemia patients.

Mössbauer spectra of 12 beta-thalassemia/hemoglobin E spleen samples from Thai patients who had not received multiple blood transfusions and chelation therapy and seven beta-thalassemia spleen samples from Australian patients who had received multiple blood transfusions and chelation therapy were recorded with sample temperatures of 78 K. Each spectrum was found to consist of a superposition of a relatively intense central doublet characteristic of high-spin Fe(III), a low intensity sextet of peaks due to magnetic hyperfine-field splitting, and occasionally a doublet that could be attributed to heme iron. A significant (P=0.01) difference (Kolmogorov-Smirnov statistic of 0.71) between the distributions of sextet signal intensity as a fraction (Fs) of the total non-heme iron Mössbauer spectral signal for the two groups of patients was detected. The distribution of Fs for the Thai beta-thalassemia/hemoglobin E spleens had a mean value of 0.128 (S.D. 0.035) while that for the Australian beta-thalassemia spleens had a mean of 0.27 (S.D. 0.12). No significant difference between the distributions of non-heme iron concentrations in the tissues for the two groups of patients was detected by atomic absorption spectrometry. This study shows that the Australian beta-thalassemia patients had a higher fraction of their non-heme spleen iron in a goethite-like form than the Thai beta-thalassemia/Hb E patients.

The effect of prolonged iron loading on the chemical form of iron oxide deposits in rat liver and spleen.

Female Porton rats were loaded with iron either by supplementing the diet with 2.5% carbonyl iron for up to 22 months (18 rats) or by regularly injecting rat blood cells intraperitoneally for up to 10 months (eight rats). 57Fe Mössbauer spectroscopy of freeze-dried samples of liver and spleen was used to analyse the chemical forms of iron deposited in these tissues over the period of iron loading. A sextet signal in the Mössbauer spectra was identified as being due to a form of haemosiderin based on the structure of the mineral goethite. The spectral parameters of the sextet signal in the rat tissues indicate that the goethite-like haemosiderin particles are less crystalline than those found in iron-loaded human tissues. For the dietary-iron-loaded rat livers, the fraction (Fs) of the Mössbauer signal in the form of this sextet was found to increase significantly (from approx 0.04 to 0.09) with the age of the rats (r=0.77, P<0.0005). This indicates that the fraction of liver iron in the form of the goethite-like haemosiderin increases with age of the rat and hence with the duration of iron loading. In addition, Fs for these livers was found to increase significantly with the fraction of iron in non-parenchymal cells as measured by computer-assisted morphometric analysis of histological sections (r=0.71, P<0.005).

A strategy to reduce red marrow dose for intraperitoneal radioimmunotherapy.

The aim of this study was to determine whether shorter-lived radionuclides can reduce red marrow (RM) toxicity for i.p. radioimmunotherapy (RIT). The potential radionuclides, which included Lu-177, I-131, Y-90, Re-186, Re-188, and Ho-166, were attached to antibody CC49. Each radiopharmaceutical was assumed to have identical in vivo pharmacokinetics. Blood and whole body retention data acquired from 26 patients who received i.p. RIT with Lu-177 CC49 were used as input. The average biological half-time of Lu-177 CC49 in the whole body was 280 h, and the average Lu-177 concentration in plasma increased to a maximum at 2 days postinfusion, followed by steady clearance. The residence time and RM doses were calculated for each radionuclide. In the current model, Re-188 was found to deliver the lowest RM dose, primarily because it had the shortest half-life, whereas Y-90 delivers the highest dose. Re-188 delivers 60% of the RM dose as compared with Lu-177 and can increase the dose to metastatic sites in the i.p. space by a similar factor. Based on limiting the RM dose to 200 cGy, the maximum administered activity of each radionuclide is as follows: (a) 106 mCi, Lu-177; (b) 58 mCi, I-131; (c) 34 mCi, Y-90; (d) 70 mCi, Re-186; (e) 169 mCi, Re-188; and (f) 110 mCi, Ho-166. Because of the delayed steady leakage of radiopharmaceuticals from the i.p. cavity to the plasma, short-lived radionuclides may offer special advantages for i.p. RIT.

Phase II study of interferon-enhanced 131I-labeled high affinity CC49 monoclonal antibody therapy in patients with metastatic prostate cancer.

Adjuvant Interferon (IFN) was given to increase tumor antigen expression and enhance localization with 131I-labeled CC49 radioimmunotherapy in a Phase II trial for hormone resistant metastatic prostate cancer. Patients received four doses of alpha-IFN (3 x 10(6) IU) s.c. on alternate days, from day -5 to day +1 of 75 mCi/m2 131I-CC49 treatment. Toxicity was well tolerated, with the majority of patients experiencing transient grade 3 or 4 neutropenia and/or thrombocytopenia (maximal at 4-6 weeks). The absorbed dose was >25 Gy in four of eight tumors visualized, which represents an increase of >20 fold over whole body radiation dose. Two patients had radiographic minor responses by 6 weeks post-therapy, whereas five of six patients experiencing pain had symptom relief without radiographic changes. The protocol provided modest antitumor effects (pain relief in five of six patients and two minor radiographic responses). This study suggests that the addition of IFN enhanced tumor uptake and antitumor effects as compared to a prior Phase II trial of 131I-CC49 alone.

Effect of recombinant alpha-interferon on pharmacokinetics, biodistribution, toxicity, and efficacy of 131I-labeled monoclonal antibody CC49 in breast cancer: a phase II trial.

Preclinical studies have demonstrated that recombinant IFN-alpha (rIFN-alpha) can enhance the tumor associated glycoprotein 72 (TAG-72) on tumors. To determine whether rIFN-alpha could enhance TAG-72 expression in vivo in patients, 15 women with breast cancer were randomized to receive daily injections of rIFN-alpha (3 x 10(6) units/m2 for 14 days) beginning on day 1 (group 1 = 7 patients) or on day 6 (group 2 = 8 patients). On day 3, all patients received a 10-20-mCi tracer dose of 131I-CC49, a high-affinity murine monoclonal antibody reactive against TAG-72, followed by a therapy dose of 60-75 mCi/m2 of 131I-CC49 on day 6. Whole body and single-photon emission computed tomography scans along with whole blood pharmacokinetics were performed following tracer and treatment phases. Hematological toxicity was considerable; reversible grade 3-4 neutropenia and thrombocytopenia was observed in 12 of 15 patients. Twelve of 14 patients tested developed human antimouse antibodies 3-6 weeks after treatment. For group 1 patients, whole blood residence time increased significantly between that predicted from the tracer doses and therapy doses (42.6 +/- 4.7 versus 51.5 +/- 4.8 h, respectively; P < 0.01). The calculated radiation absorbed dose to red marrow from therapy compared to tracer activity was also significantly higher for this group (1.25 +/- 0.35 versus 1. 07 +/- 0.26 cGy/mCi; P < 0.05). Treatment with rIFN-alpha was found to enhance TAG-72 expression in tumors from patients receiving rIFN-alpha (group 1) by 46 +/- 19% (P < 0.05) compared to only 1.3 +/- 0.95% in patients not initially receiving IFN (group 2). The uptake of CC49 in tumors was also significantly increased in rIFN-alpha-treated patients. One partial and two minor tumor responses were seen. In summary, rIFN-alpha treatment altered the pharmacokinetics and tumor uptake of 131I-CC49 in patients at the expense of increased toxicity.

Requirements for a treatment planning system for radioimmunotherapy.

Cancer-seeking antibodies carrying radionuclides can, in theory, be very powerful agents for the radiotherapy of cancer. However, as with all radiotherapy, the undesired dose to critical normal organs is the limiting factor that determines success or failure. The distribution of radiation dose in cancer and noncancer tissue is highly dependent on choices the therapist can make: choices of the antigens to be targeted, choices of the antibodies or antibody fragments to be used, choices of radionuclides, of amounts, of timing, and other electives. New technologies, especially of monoclonal antibody production, make the options myriad. Optimization of this therapy depends on a foreknowledge of the radiation dose distributions to be expected. The necessary data can be acquired by established tracer techniques, in individual patients, for particular treatment selections. These tracer techniques can now be implemented by advanced equipment for quantitative, tomographic radionuclide imaging and strengthened by dynamic modeling of the physiological parameters which govern radionuclide distribution, and hence radiation dose distribution.

Both GABA(B) receptor agonist and antagonists decreased brain stimulation reward in the rat.

The present experiments were designed to determine the role of GABA(B) receptor function on brain stimulation reward. Using a discrete-trial current-intensity threshold procedure, dose-effect functions were generated for the GABA(B) receptor agonist CGP 44532 (0-1.0 mg/kg, s.c.) and the GABA(B) receptor antagonists CGP 56433A (0-10.0 mg/kg, s.c.) and CGP 51176 (0-300.0 mg/kg, s.c.) on brain reward thresholds in rats. The GABA(B) receptor antagonists CGP 56433A and CGP 51176 were used also to examine interaction effects with the GABA(B) receptor agonist CGP 44532 on reward thresholds. Administration of the highest doses of both the GABA(B) receptor agonist and antagonists elevated reward thresholds. Thus, both the agonist and antagonists used induced a reward decrement when administered separately. In addition, the co-administration of either of the two receptor antagonists with the agonist induced an additive effect on thresholds, rather than blocking the agonist-induced threshold elevations. These results suggest that activation of GABA(B) receptors modulates intracranial self-stimulation behavior in a complex fashion, possibly through differential effects of GABA(B) agonists and antagonists on pre- and post-synaptic GABA(B) receptors.

Absolute quantitation of radiotracer uptake in the lungs using a gamma camera.

A transmission-emission method for the absolute quantitation of Tc-99m in the lungs with a computerized gamma camera is described. The method requires no measurements of the linear attenuation coefficients of the lung and chest wall, or of their thickness. It yields results of acceptable accuracy for everyday clinical use and offers the great advantage of avoiding the use of a phantom. The method could be extended for the absolute quantification of Tc-99m agents in soft tissue and bone with errors of less than 10%.

A scintigraphic method for the assessment of intraluminal volume and motility of isolated intestinal segments.

The isolated in vivo intestinal segment is a popular experimental preparation for the investigation of intestinal function, but its value has been limited because no method has been available for measuring changes in intraluminal volume under experimental conditions. We report a scintigraphic technique for measuring intraluminal volume and assessing intestinal motility. Between 30 and 180 ml, the volume of a 75-cm segment of canine jejunum, perfused with Tc-99m-labeled tin colloid, was found to be proportional to the recorded count rate. This method has been used to monitor the effects of the hormone vasopressin on intestinal function.

Comparison of three boundary detection methods for SPECT using Compton scattered photons.

Three simple methods of defining the boundary of a transverse section in single photon emission computed tomography (SPECT) were compared using Compton scattered photons from a small 99mTc source located either inside or outside a water-filled cylinder of 22 cm diameter. Scattered events were acquired with 360-degree rotation of the gamma camera and transverse section images were reconstructed using a filtered backprojection method. The boundary of section images obtained with three different geometric arrangements of the source and the camera were compared. The 90-degree Compton scatter method, using a source external to the cylinder and at 90 degrees to the front of the detector, was found to give the best boundary definition. Similarly, detection of scattered events using a radionuclide source placed outside a human body was capable of providing good boundary information for the large stack of contiguous section images produced by a rotating SPECT camera. Calculations confirmed the profound influence of boundary errors on SPECT quantitations.

Development of a SPECT-based three-dimensional treatment planning system for radioimmunotherapy.

UNLABELLED: Two major obstacles in the development of improved methods for more accurate dose estimates for radioimmunotherapy have been the difficulty in obtaining an accurate patient-specific three-dimensional activity map in vivo and calculating the resulting absorbed dose. We propose a method for three-dimensional internal dosimetry that integrates the three-dimensional activity map from SPECT with a dose-point kernel convolution technique to provide the three-dimensional distribution of absorbed dose. METHODS: Accurate activity quantitation was achieved with appropriate methods. The count density map from SPECT images was converted into an activity concentration map with a calibration phantom approach. This map was then convolved with an 131I dose-point kernel and three-dimensional fast Fourier transform to yield three-dimensional distribution of absorbed dose, which was then processed to provide the absorbed dose distribution in regions of interest. RESULTS: The accuracy of quantitative SPECT was validated to be within 16%. The calculated penetrating radiation absorbed dose was verified with thermoluminescent dosimeter measurements to be within 8%. With standard organs and configuration, the method calculated absorbed dose in good agreement with the MIRD formalism (less than 14%). CONCLUSION: This method overcomes the limitations of planar imaging techniques and the current routine implementation of the MIRD formalism. The results can be processed to provide the absorbed dose distribution in regions of interest and parameters for treatment optimization. Absorbed dose distribution from any plane can be graphically displayed in various ways.

Dosimetry and toxicity of samarium-153-EDTMP administered for bone pain due to skeletal metastases.

UNLABELLED: Palliation of bone pain in patients with cancer metastatic to bone is being evaluated in several cancer centers by the administration of the bone-seeking phosphonate ethylenediaminetetramethylenephosphonic acid (EDTMP) chelated with the beta particle-emitting radionuclide 153Sm. METHODS: In this study, 153Sm-EDTMP was intravenously injected into 19 patients over a 1-min period. Patients received up to four injections of 18.5 MBq (0.5 mCi) or 37 MBq (1.0 mCi) per kilogram of body weight. Skeletal retention was calculated from urinary excretion. RESULTS: No uptake of 153Sm-EDTMP in nonskeletal tissues was observed in whole-body gamma camera images. The mean skeletal uptake for all patients was 54% +/- 16% of the injected dose (%ID). This resulted in the bone marrow receiving 89 cGy/GBq +/- 27 cGy/GBq (3.28 cGy/mCi +/- 0.99 cGy/mCi), with calculated marrow doses ranging from 27 cGy to 338 cGy. For each patient, the estimated radiation absorbed dose to the marrow was correlated to the percent decrease in platelet number, ranging from 7.4% to 78.9%. CONCLUSION: Since the deviation of uptake between the four injections for a given patient (7.6% ID) was less than the deviation for all patients (16% ID), the initial dose may be used to estimate the skeletal uptake for the remaining doses. These radiation dose estimates permit patients at risk to be identified prior to reaching myelotoxicity and develop dose-response models. Thirteen patients (68%) reported significant pain relief from this radionuclide therapy. Bone pain appears to be alleviated by 153Sm-EDTMP with limited red marrow doses and no toxic effects in other organs.

Comparison of low- and medium-energy collimators for SPECT imaging with iodine-123-labeled antibodies.

The planar and single photon emission computed tomography (SPECT) imaging performance of two low-energy and one medium-energy collimators has been compared for 123I(p,5n). Septal penetration in the low-energy collimators affected the planar uniformity and sensitivity and made the usual uniformity correction methods inappropriate. SPECT uniformity and resolution were also distorted by these artifacts induced in the planar images obtained with the low-energy collimators. The planar and SPECT images of a SPECT phantom contradicted the spatial resolution measurements made with a line source in air. The medium-energy collimator is recommended for imaging 123I(p,5n)-labeled antibodies in patients, particularly when quantitation is required.

Factors affecting 131I-Lym-1 pharmacokinetics and radiation dosimetry in patients with non-Hodgkin's lymphoma and chronic lymphocytic leukemia.

UNLABELLED: Lym-1, a monoclonal antibody that preferentially targets malignant lymphocytes, has induced therapeutic responses in patients with non-Hodgkin's lymphoma (NHL) and chronic lymphocytic leukemia (CLL) when labeled with 131I. Responders had statistically significant prolongation of survival compared with nonresponders. The nonmyeloablative, maximum tolerated dose for each of two doses of 131I-Lym-1 was 3.7 GBq/m2 (total 7.4 GBq/m2 [100 mCi/m2, total 200 mCi/m2]) of body surface area. The purpose of this study was to determine the pharmacokinetics and radiation dosimetry for the initial 131I-Lym-1 therapy dose in patients with NHL and CLL and to compare tumor dosimetry with 131I-Lym-1 dosing and other patient parameters. METHODS: Fifty-one patients with stage 3 or 4 lymphoma were treated with 131I-Lym-1 (0.74-8.04 GBq [20-217 mCi]) in either a maximum tolerated dose (MTD) or low-dose (LD) trial. Total Lym-1 given to each patient was sufficient in all instances to exceed the threshold required for stable pharmacokinetics. Quantitative imaging and physical examination, including caliper and CT measurement of tumor size and analysis of blood, urine and feces, were performed for a period of 7 to 10 d after infusion to assess pharmacokinetics and radiation dosimetry. Clinical records were reviewed to obtain data required for comparative assessments. RESULTS: The concentration (%ID/g) and biologic half-time of 131-Lym-1 in tumor were about twice those in normal tissues, although tumor half-time was similar to that of the thyroid. Pharmacokinetics were similar for patients in the MTD and LD trials, and for NHL and CLL patients in the LD trial, except that the latter group had less tumor concentration of 131I. Mean tumor radiation dose per unit of administered 131I was 1.0 Gy/GBq (3.7 rad/mCi) for patients with NHL whether in MTD or LD trials, about nine times greater than that for body or marrow. Tumor radiation dose was less and liver radiation dose was more in patients with CLL. Otherwise, radiation dosimetry was, on average, remarkably similar among groups of patients and among individual patients. Pharmacokinetics and dosimetry did not appear to be influenced by the amount of 131I or Lym-1 within the ranges administered. Tumor concentration of 131I and radiation dose per gigabecquerel were inversely related to tumor size but did not seem to be related to histologic grade or type, tumor burden or therapeutic response. CONCLUSION: The therapeutic index of 131I-Lym-1 was favorable, although the index for patients with CLL was less than that for patients with NHL. Pharmacokinetics and radiation dosimetry were, on average, remarkably similar among patients and groups of patients in different trials.

Pharmacokinetics, dosimetry and toxicity of holmium-166-DOTMP for bone marrow ablation in multiple myeloma.

UNLABELLED: In this Phase I clinical trial, six multiple myeloma patients who had not responded to conventional therapy and were scheduled for bone marrow transplantation received a bone-seeking radiopharmaceutical for bone marrow ablation. The pharmacokinetics, dosimetry, and toxicity of this radiopharmaceutical were studied. METHODS: Patients received from 519 mCi to 2.1 Ci (19.2 GBq to 77.7 GBq) of holmium-166 (166Ho) complexed with a bone-seeking agent, DOTMP (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetramethylene-phosphonic acid). The reproducibility of pharmacokinetics from multiple injections of 166Ho-DOTMP administered to these myeloma patients was demonstrated from blood (r2 = 0.926) and whole-body retention (r2 = 0.983), which allowed therapeutic parameters to be determined from a diagnostic study. RESULTS: Over 50% of the 166Ho-DOTMP injected dose was excreted within 2-3 hr postinjection, increasing to 75%-85% over a 24-hr period. Rapid blood clearance minimized radiation dose to nontarget tissue: less than 10% of the injected activity was retained in the blood pool at 1 hr postinjection, and less than 2% remained after 5 hr. The total radiation absorbed dose delivered to the bone marrow for the six patients ranged from 7.9 Gy to 41.4 Gy. CONCLUSION: All patients demonstrated severe bone marrow toxicity with a white blood cell (WBC) count < 1,000 cells/microliters, two patients exhibited marrow ablation (WBC count < 100 cells/microliters), and no other toxicity > or = grade 2 was observed in any of the patients.

Multicompartmental analysis of the kinetics of radioiodinated monoclonal antibody in patients with cancer.

A conceptual biologic model was developed and used to analyze the behavior of 123I-Lym-1 monoclonal antibody against African human B cell lymphoma in patients with B cell lymphoma. Originally, the observed data could not be simulated with parameters for homologous immunoglobulins reported in the literature because of a major processor that was capable of distinguishing this murine immunoglobulin from the patient's own immunoglobulins. With a nonlinear parametric model, the data observed in patients could be fitted to the model. The nonlinear parameter determined the transfer of antibody from the intravascular to a processor compartment, primarily the liver. This transfer was a function of the number of free receptors in the processor. Model simulated curves for the time course of concentration of antibody in the blood for different amounts of injected antibody revealed that blood clearance of radiolabeled antibody was profoundly decreased by increased amount of injected antibody. This model provides an explanation for the observations that tumor imaging is improved with injection of larger amounts of antibody, and a basis for modifying the pharmacokinetic behavior of an antibody in order to optimize radioimmunodiagnosis and radioimmunotherapy.

Prediction of radiation doses from therapy using tracer studies with iodine-131-labeled antibodies.

UNLABELLED: Tracer pharmacokinetic studies are often used in treatment planning for radionuclide therapy including radioimmunotherapy. This study evaluates the validity of using tracer studies to predict radiation doses from therapy with the same radiolabeled antibody. METHODS: Quantitative imaging and blood radioactivity were used to obtain the pharmacokinetics and radiation doses that were delivered to the total body, blood, marrow, lungs, liver, kidneys, thyroid, spleen and tumors. Tracer and therapy data for eight patients with lymphoma and one patient with breast cancer were compared using linear regression statistics. Doses of 131I-labeled antibody for the tracer studies ranged from 0.1 to 0.4 GBq (2 to 10 mCi), and therapy doses ranged from 0.7 to 5.6 GBq (20 to 150 mCi). RESULTS: Radiation doses to tissues and, in particular, the bone marrow and tumors were reliably predicted from tracer studies. In this group of patients, median dose to marrow from marrow targeting, total body and blood was 9.2 cGy/GBq for tracer studies and 7.6 cGy/GBq for therapy studies with a median difference of 0.5 cGy/GBq. Median dose to tumors was 81.1 cGy/GBq for tracer studies and 70.3 cGy/GBq for therapy studies with a median difference of 5.9 cGy/GBq. CONCLUSION: In these patients, tracer studies were predictive of the radiation doses from therapy for total body, major organs and tumors. The radiation doses to marrow and tumors, which are the usual determinants of the therapeutic index, correlated well between tracer and therapy studies (r > or = 0.95).