High-dose gallium-67 therapy in patients with relapsed acute leukaemia: a feasibility study.

Gallium-67 (67Ga) accumulates in malignant tissues via the transferrin receptor without need for a monoclonal antibody and emits cytotoxic low-energy electrons. In this study we investigated the feasibility, pharmacokinetics, toxicity and preliminary efficiency of high-dose 67Ga injected intravenously (i.v.) in patients with acute leukaemia not responding to conventional therapy. Twelve doses of 36-105 mCi of Gallium67 citrate were administered as a push injection to eight patients with resistant leukaemia in a pilot study. All five patients with acute myeloid leukaemia (AML) and three patients with acute lymphoblastic leukaemia (ALL) had resistant disease or resistant relapse. No (sub)acute toxicity was observed. Independent of the administered dose, whole-blood radioactivity levels 10 min after administration measured only 1.25 +/- 1.39 microCi ml-1, indicating a large volume of distribution. Urine excretion in the first 24 h ranged from 18% to 51.5% (median 29.5%) of the administered dose. Cellular uptake of 67Ga was less than in previous in vitro studies. Whole-body radiation dose was estimated to be 0.25 +/- 0.03 cGy mCi-1. Red marrow dose was estimated to be between 0.18 +/- 0.02 and 0.97 +/- 0.12 cGy mCi-1. One definite response was observed in an ALL patient with disappearance of skin lesions, normalisation of the enlarged spleen and profound leucopenia. Three other patients showed transient reductions in white blood cell counts without disappearance of blasts from the peripheral blood. We conclude that high-dose i.v. 67Ga can be safely administered but that the uptake of 67Ga in blast cells must increase to make 67Ga therapeutically useful in patients with relapsed leukaemia. ImagesFigure 2

Although the initial remission rate of acute leukaemia in adults is high (70-80%), a considerable portion of patients eventually die of their disease (Rohatiner et al., 1990). Therefore, new treatment modalities have to be explored. The idea of therapy with a radionuclide that accumulates in the target tumour cell by itself is appealing. Recently, promising results were reported from radioimmunotherapy with a '31iodine-labelled anti-CD33 monoclonal antibody (MAb) in patients with relapsed or refractory myeloid leukaemia (AML) (Appelbaum et al., 1990;Schwartz et al., 1993).
However, problems related to MAb-mediated radiotherapy include adverse reactions caused by the administration of foreign protein and the forming of human anti-mouse antibodies (HAMA), precluding repeated cycles of therapy (Rosen and Kuzel, 1993).
In vitro blast cells of some AML patients accumulate 67Ga strongly, and after incubation with 80 IsCi ml-' 67Ga, clonogenic survival was reduced more than 90% compared Correspondence: Received 4 January 1995; revised 3 May 1995; accepted 5 July 1995.
with control cells. In some blast cells clonogenic growth was completely abolished after only 20 1sCi ml l 67Ga (Jonkhoff et al., 1995). Some of the relative ineffectiveness of 67Ga for cell kill might be compensated by a high in vivo cellular uptake of 67Ga and favourable pharmacokinetic data. In this study we report the first in vivo data concerning toxicity and pharmacokinetics and preliminary efficacy of 67Ga in eight patients with resistant acute leukaemia.

Patients
Patients with end stage acute leukaemia were entered into the study after giving informed consent according to the Declaration of Helsinki. The study was approved by the ethical committee of the Free University. Five patients with acute myelogenous leukaemia (AML) and three patients with acute lymphoblastic leukaemia (ALL) were included. All patients had a WHO performance status of 0 or 1. No patient had pre-existent cardiac, pulmonary or renal disease. Supportive care medication in most patients included ciprofloxacin, fluconazol, ranitidine, tranexamic acid, and transfusion of blood products. The only cytostatic co-medication allowed was prednisone or dexamethasone in ALL and hydroxyurea in AML patients, in order to control peripheral blast counts. 67Ga Carrier-free 67Ga was obtained from Mallinckrodt Diagnostics (Petten, The Netherlands) as 67Ga chloride. 67Ga citrate for intravenous injection, with a low citrate concentration, was prepared as described previously . 67Ga citrate was given in a volume of 10 ml as a rapid intravenous push, except in patient 5 who was given a second injection of 67Ga as a 1 h infusion in order to study urinary excretion and transferrin binding. Radiation safety precautions, including rules for hospitalisation on a nuclear medicine unit, were in accordance with accepted guidelines (National Council on Radiation Protection, 1970).
The lowest dose level was based on experience in lymphoma patients, who suffered no side-effects other than myelosuppression after administration of 40-60 mCi 67Ga i.v. . Before and biweekly after 67Ga administration, whole blood cell counts were determined. Liver enzymes and renal function were tested weekly.

Transferrin receptor
The percentage of transferrin receptor-positive leukaemia blasts (5000 events) were analysed on a FACScan flow cytometer (Becton Dickinson). A FITC-conjugated mouse anti-human monoclonal antibody was used (Dako-CD71, Ber-T9; Glostrup, Denmark). An irrelevant IgGI was used as isotype control. Cellular uptake of 67Ga Whole blood samples were drawn from the patient by venipuncture 10 min, 60 min and 24 h after 67Ga administration and put on ice immediately. After lysing erythrocytes using three drops of lysing solution the cellular 67Ga uptake and cellular 67Ga content was determined as described previously (Jonkhoff et al., 1995). All radioactivity values were corrected for physical decay (at t = 0).

Pharmacokinetics
Radioactivity of whole blood samples was measured 10 min, 60 min and 24 h after injection of 67Ga. The first 24 h after injection the urine was collected to measure the renal excretion of 67Ga. In one patient plasma samples on more time points were measured. In this patient (patient 8) the area under the curve (AUC) of 67Ga was analysed by Topfit 2.0, using a non-compartmental model (Tanswell et al., 1993).

Scintigraphy
Twenty-four hours after 67Ga administration a whole body scintigraph was performed with a dual-head gamma camera and medium-energy collimators (ADAC Laboratories, Milpitas, CA, USA).

Dosimetry
In six patients (seven complete measurements) the blood and whole body data were used for red marrow dose calculations. The residence times of 67Ga in the blood and whole body were derived from numerical integration of the blood and urinary time activity curves. After 24 h we assumed no biological clearance and only physical decay of 67Ga. With approach A the activity in the whole body was assumed to be homogeneously distributed (Plaizier et al., 1994a,b). With approach B the activity in the whole body was assumed to be equally divided between the remainder of the body and the skeleton. The latter approach was used to illustrate the theoretical 'maximal' effect of specific bone uptake on the red marrow dose.
The activity in the red marrow and blood was assumed to be equal (Siegel et al., 1990). Specific uptake in the red marrow is neglected because of lack of information on red marrow kinetics. Whole-body radiation dose was estimated from the whole body residence time.
Whole body and red marrow dose were calculated according to MIRDDOSE2 (Watson and Stabin, 1984). The kinetic and dosimetric data were compared with the ICRP-53 (International Commission on Radiation Protection, 1987) and MIRDDOSE2 standard.

Statistics
Statistical analysis was performed with Stat-Graphics 2.6 statistical computer program.
In total, 12 doses of 67Ga were delivered i.v. Doses ranged from 36-105 mCi. None of the patients was admitted for more than 24 h. No acute toxicity was observed. Two patients noted a slight fruity flavour. One patient noted increased bleeding tendency and muscle pains in the 4 days following 60 mCi 67Ga. However, no objective increase in bleeding tendency was observed and muscle pains did not occur after a second administration of 60 mCi 67Ga in the same patient.
No change in kidney or liver function was observed. Levels of lactate dehydrogenase (LDH) varied with disease activity, and were not correlated with 67Ga administration. Haematological effects were restricted to those on blast counts. 10.6 11.7 Pharmacokinetic data, including total administered dose of gallium-67 citrate (mCi), dose per square metre body surface (mCi m-2), whole-blood 67Ga radioactivity 10 min, 60 min and 24 h post injection (pCi ml-'), total 67Ga urine excretion in the first 24 h post injection (mCi) and urine excretion as percentage of the injected dose [dose (%)]. All radioactivity values are corrected for physical decay (at t = 0). Median value and s.d. are given for the urinary excretion. a67Ga administered as 1 h infusion. NA, not available. ml-', with two patients (5 and 6) reaching higher levels of 4.5g.Ci ml-'. One h post injection (p.i.) blood levels were approximately halved, compared with 10min p.i. The 24h blood levels were approximately 30% of the 1 h blood levels.

Pharmacokinetics
One patient (patient 8) was more extensively monitored for plasma levels (Figure 1). A steep decrease in plasma radioactivity level was noted in the first minutes after administration. Pharmacokinetic data calculated in this patient are given in the legend of Figure 1.
Of the 12 administrations of 67Ga, urine excretion was measured on nine occasions. The median urine excretion of 67Ga in the first 24 h p.i. was 29.5% (range 18 -51%) of the administered dose. In three patients the first 0-6 h urine portion was collected separately. These collections contained 22.1 mCi (27% of injected dose) in patient 4, 32.8 mCi (39.6% of injected dose) in patient 6 and 17.5 mCi (16.6% of injected dose) in patient 8.
The whole-body residence time was estimated to be 90.89 ± 12.40 h compared with 88.56 h calculated according to ICRP-53. We calculated a red marrow residence time of 0.41 ± 0.16 h, which differed considerably from the comparable ICRP-53 value of 4.78 h.
The whole body dose was estimated to be 0.25 ± 0.03 cGy mCi-', which compares well with MIRD/ICRP-53 calculations of 0.24 cGy mCi-'. The red marrow dose depended on whether a homogeneous distribution was expected; approach A, 0.18 ± 0.02 cGy mCi -' or 50% accumulation of the total activity in the skeleton was assumed; approach B, 0.97 ± 0.12 cGy mCi-'. The comparable MIRD/ICRP-53 value for the red marrow dose was 0.70 cGy mCi-'.  Objective judgement of response was hindered by the concomitant anti-leukaemic medication (hydroxyurea or corticosteroid). In Figure 3 white blood cell counts (WBC) in all patients are presented in relation to the medication. AML patients 3 and 5 showed a reduction in WBC following administration of 67Ga. Normal WBC values were only reached in patient 5, although the white blood cell differentiation still showed the presence of 52% blast cells. Furthermore, notable in this patient was the normalisation of the serum LDH from 1119 U1I (N=250U1V') to 155U 1-'. The response, however, was short-lived as the WBC increased 15 days after the second administration. Of the ALL patients, patient 6 had a definite response with WBC decreasing <0.1 x 1091-', 20 days p.i., normalisation of the enlarged spleen, which ranged 7 cm beneath the left costal margin and disappearance of skin lesions. Patient 8 had an initial rise in WBC up to over 200 x IO' 1`in 4 days. Dexamethasone was started followed by a steep decrease in WBC. However, it is unlikely that dexamethasone caused the ensuing leucopenia, which was probably an effect of the 67Ga administration.

Discussion
In the present study eight patients with relapsed and/or resistant acute leukaemia received in total 12 doses of highdose 67Ga. No acute side-effects were observed nor any extrahaematological effect. Myelotoxicity is considered to be the major and dose-limiting toxicity in most radionuclide therapies in lymphoma patients (Kaminski et al., 1993;Press et al., 1993). In our study myelotoxicity was apparent by the (transient) effect on blast cells but not in changes in erythrocyte and platelet requirements. Whole blood levels of radioactivity were unexpectedly low, with values in six patients between 0.80 and 1.86 pCimh-' 10min after administration. We expected higher levels as 67Ga is temporarily confined to the plasma compartment by prompt binding to transferrin (Vallabhajosula et al., 1980). For instance, in patient 8, with an estimated blood volume of 7.5 1 and who received 105 mCi, one would have expected an initial blood level of 14 ILCi ml-' instead of the 0.80 yICi ml1we found. Low plasma 67Ga levels are described following chemotherapy (Shephton and Martin, 1980) but this was not the case in patient 8. The observed steep initial decrease in plasma or blood 67Ga levels seem to disagree with pharmacokinetic data of 67Ga that describes a short-lived and a long-lived component with a biological half-life of approximately 30 h and 613 ± 83 h respectively (Cloutier et al., 1988). The urine excretion was unexpectedly high with a median value of 29.5% (range 18-51.5%) of the injected dose excreted in the first 24 h. The urinary excretion seems larger than the 26% of the injected dose of 67Ga in the first 7 days after injection that Nelson et al. (1972) observed in 23 patients. Pharmacokinetic data of non-radioactive gallium nitrate report 24 h urinary excretion of 15-72% after a bolus injection (Hall et al., 1979;Kelsen et al., 1980). No difference in urinary excretion between bolus injection or 1 h infusion was observed in patient 5, who served as her own control. The observed pharmacokinetic differences could be related to the concomitant medication or differences in patient group. The iron status of our patients, who received frequent blood transfusions, could have been of influence. Another possibility is that our low-citrate formula of 67Ga influenced our data, as gallium-67 citrate can form multinucleate polymeric forms, gallium hydroxides or bind to serum proteins other than transferrin (Larson et al., 1978). We tried to measure the portion of 67Ga-Trf in the plasma samples by highperformance liquid chromatography (HPLC), but could not validate sufficiently the stability of the 67Ga-Trf binding during the procedure, as free gallium-67 citrate complexes with the silica of the column.
The uptake of 67Ga in the blast cells was approximately 40 times less than in our in vitro experiments (Jonkhoff et al., , 1995Leeuwen-Stok et al., 1993). This rather low cellular 67Ga uptake might be explained by insufficient Trfreceptor expression on blast cells, the relatively high Trf concentration in the blood/bone marrow compartment, which inhibits the uptake of 67Ga in the cell (Leeuwen-Stok et al., 1993) or the low blood levels of 67Ga. No apparent correlation was found between Trf receptor (CD71) expression, or 67Ga uptake in blast cells and in vivo response.
Our dosimetric calculations show similar whole body retention times between this study and ICRP-53. The seem-ing contradiction between a larger than expected urinary excretion and similar residence times to ICRP values can be explained by the neglect in our study of biological clearance after 24 h. The calculated activity in red marrow however is considerably lower than the red marrow activity suggested by the ICRP-53. The red marrow absorbed dose based on the ICRP-53 lies between red marrow absorbed dose calculations with assumed different skeleton activities (approach A or B). For individual patients the total red marrow dose varied between 10-17cGy (approach A) and 52-1lOcGy (approach B) depending on the chosen approach (data not shown). As we observed at least one definite clinical response, approach B seems more realistic. It is also possible that the absorbed whole body and red marrow doses are underestimated because microdosimetry was not taken into account (van Dieren, 1993). Furthermore, we cannot exclude the possibility that bone marrow blast cells had a higher uptake of 67Ga than the peripheral blasts, as the minimal 67Ga content in the cells with only a few disintegrations in a million cells is not likely to result in the observed clinical response.
The body distribution measured by scintigraphy showed an abnormal pattern, compared with high-dose 67Ga administration in lymphoma patients . The skeleton was imaged more clearly and liver, spleen and bowel less intensively. We cannot exclude the possibility that the distribution in the skeleton is caused by the bone-seeking properties of 67Ga (Ando et al., 1989) or uptake in bone marrow blasts. More likely, however, the distribution in the skeleton is due to additional factors such as iron overload. Engelstad et al. (1982) described a similar 67Ga distribution in patients with multiple red cell transfusions.
Responses are difficult to interpret with concomitant antileukaemic medication. Two AML patients seemed to respond with decreasing WBC after 67Ga administration (patients 3 and 5), but these responses seem to be short lived. Blast cells remained in peripheral blood smears. Of the ALL patients, one had a definite response with disappearance of skin lesions and normalisation of the enlarged spleen. Profound leucopenia (WBC <0.1 x I0 1') was encountered from 15 days p.i. onwards, until death 2 months later. Another ALL patient (patient 8) had a very steep decrease in WBC, with disappearance of blast cells, following *dexamethasone medication. This decrease in WBC might be caused by the 67Ga administration, as radionuclide therapy is known to exert its effect only after several days. In total, we observed one definite response (13%) and three possible responses (38%) out of eight patients.
Our conclusion is that high-dose 67Ga therapy is well tolerated, but cellular 67Ga uptake is relatively low. Pharmacokinetic data suggest a large proportion of nontransferrin-bound 67Ga, influencing urine excretion, body distribution and possibly cellular uptake. Nevertheless, transient responses and one definite response were noted.
Therefore, we feel that if cellular 67Ga uptake can be enhanced by additional measures, such as desferrioxamine or iron-dextran administration (Shani et al., 1986), high-dose 67Ga therapy might be useful in leukaemia patients.