Radioimmunotherapy of Non-Hodgkin’s Lymphoma with ⁹⁰Y-DOTA Humanized Anti-CD22 IgG (⁹⁰Y-Epratuzumab): Do Tumor Targeting and Dosimetry Predict Therapeutic Response?

Participating Institutions

Four institutions participated in this study: the Garden State Cancer Center (GSCC, at the Center for Molecular Medicine and Immunology, Belleville, NJ); Hospital of the University of Pennsylvania (HUP, Philadelphia, PA); Roswell Park Cancer Institute (RPCI, Buffalo, NY); and University Hospital, René Gauducheau Cancer Center (Nantes, France). The institutional review board or ethics committee at each institution approved the protocol. Patients were required to sign informed consents for their respective institution before participating in the trial. Enrollment in the trial was coordinated by GSCC.

Eligibility Criteria and Demographics

Patients were required to have a diagnosis of NHL of any histologic type and had relapsed or were refractory to one or more prior chemotherapy regimens. Patients who had failed or relapsed after prior high-dose chemotherapy (HDC) with either bone marrow or peripheral blood stem cell support were included, but dose escalation was performed separately. Patients were required to be ≥18 y of age with a Karnofsky performance status of ≥70 and a minimal life expectancy of 3 mo. Radiologic evidence of disease was required within 4 wk of the ⁹⁰Y-epratuzumab treatment, and at least one confirmed tumor site had to be targeted by an ¹¹¹In-epratuzumab IgG pretherapy imaging study (see below) to proceed with treatment. Patients needed to be at least 4 wk beyond any major surgery, radiation to an index lesion, or chemotherapy, and they could not have >25% of their bone marrow involved with lymphoma. Peripheral blood counts for white blood cells (WBC) of ≥3,000 per mm³, granulocytes of ≥1,500 per mm³, and platelets of ≥100,000 per mm³ were required to be eligible to receive ⁹⁰Y-epratuzumab. Adequate renal and hepatic functions were required, with serum creatinine <1.5 times the institutional upper limit of normal (IULN), serum bilirubin ≤2 mg/dL with aspartate aminotransferase and alkaline phosphatase (in the absence of bony involvement) levels <2 times the IULN, and the patient could not have severe anorexia or have nausea or vomiting. Patients who had prior exposure to rituximab were prescreened for antibodies to epratuzumab. Patients were excluded if they had brain metastases or a known history of HIV, hepatitis B or C, or any other serious liver abnormality. Female subjects who were pregnant or lactating were excluded, while premenopausal women were cautioned to use appropriate birth control for at least 3 mo after ⁹⁰Y-epratuzumab. Prior radiation could not have exceeded the maximum tolerable levels for critical organs and could not have encompassed >25% of the red marrow.

Antibody Preparation and Administration

Epratuzumab IgG and epratuzumab IgG-DOTA were provided by Immunomedics, Inc. Vials containing 12.0 mg epratuzumab-DOTA (10 mg/mL) were used to prepare up to 0.740 GBq ¹¹¹In-epratuzumab or 2.22 GBq ⁹⁰Y-epratuzumab, according to procedures described by Griffiths et al. (17). Briefly, the entire contents of epratuzumab-DOTA vials were transferred to vials containing ¹¹¹In- or ⁹⁰YCl₃ that was purchased from IsoTex or Perkin-Elmer. After incubating for 5–10 min at 45°C, the product was cooled and DTPA was added to bind any remaining non–antibody-bound ¹¹¹In or ⁹⁰Y. Instant thin-layer chromatography showed the unbound fraction in the ¹¹¹In and ⁹⁰Y in the radiolabeled epratuzumab IgG preparations to be 3.4% ± 1.7% and 6.3% ± 3.2%, respectively (n = 24 and 21, respectively). Two patients also received 222 MBq (6.0 mCi) ¹³¹I-epratuzumab 1 wk before the ¹¹¹In-epratuzumab. ¹³¹I-epratuzumab was prepared using an IODO-GEN (Pierce) method, as described previously (16). In all cases, the prescribed amount of radioactivity based on the planned treatment time was transferred to a 20-mL serum vial containing 0.04 mol/L sodium phosphate-buffered saline, pH 7.4, and 1.0% human serum albumin. In addition, each unit dose was supplemented with unlabeled epratuzumab IgG to equal 0.75 mg/kg (radiolabeled, DOTA-conjugated epratuzumab + unconjugated epratuzumab). The radiolabeled products were used on the same day or within 1 d of preparation, since testing showed no appreciable change in the quality of the product over this time. Immunoreactivity testing was performed by mixing a dilution of the radiolabeled product with an excess amount of a murine anti-epratuzumab-idiotype monoclonal antibody, WN (18). The mixture was then passed over a size-exclusion high-pressure liquid chromatography (SE-HPLC) column, and the fraction of the radiolabeled product eluting to a higher molecular weight, which represented complexes of epratuzumab and WN anti-idiotype antibody, was considered the immunoreactive fraction. This fraction never was <95%.

All dose calibrators were calibrated against a ⁹⁰Y standard. Injected activity was based on reading the activity in the 20-mL vial before and after removal of the dose. Before initiation of the radioantibody infusion, an open intravenous line with saline for injection was established. Radiolabeled epratuzumab was placed in a buretrol containing 20–30 mL saline. Vital signs were taken and the infusion was started. After approximately 10% of the dose was administered, vital signs were again taken, and if no significant changes were observed, the infusion was continued and completed within 20 min to 1 h. Vital signs were again observed at the end of the infusion and 1 or 2 more times before the release of the patient after the final imaging study on day 0.

Study Plan

Eligible patients were first given a pretherapy ¹¹¹In-DOTA-epratuzumab IgG dose (222–370 MBq/0.75 mg epratuzumab IgG [6–10 mCi/0.75 mg epratuzumab IgG]). Six to 8 d later, ⁹⁰Y-epratuzumab was administered. After the end of each infusion, 5–9 blood samples were drawn for pharmacokinetic analyses usually from between 0.5 h to 7 d after infusion, with the last sample taken on the day of the ⁹⁰Y-epratuzumab infusion. Blood samples were collected in serum-separator tubes, and 0.1–0.2 mL serum was counted for determination of ¹¹¹In activity in a γ-scintillation counter (Minaxi Autogamma 5000; Packard), whereas Cerenkov radiation of ⁹⁰Y was determined by direct counting in a β-counter (i.e., 0.1 mL serum or up to 1.0 mL urine was placed in saline and counted in a β-counter [Tri-Carb 2100TR; Packard] with a window setting of 0–50 keV). Scintillation counters were precalibrated against known standards to determine counting efficiencies with automatic background subtraction. β-Counting was performed to comonitor quenching conditions with correction to the counting efficiency if quenching occurred. Since the amount of residual ¹¹¹In-epratuzmab in the serum at the time of the ⁹⁰Y-epratuzumab infusion was low and ¹¹¹In counting efficiency was <0.5% in the β-window, no correction was necessary for residual ¹¹¹In radioactivity in the ⁹⁰Y-serum samples. The radioactivity concentration in the serum was converted to radioactivity concentration in the blood using the hematocrit value at the start of the blood collection. Using these data, time–activity curves were fit mono- and biexponentially using COSAM (conversion SAAM version 3.0) software (19), and the effective half-life (t_1/2) and residence times were computed, from which corresponding biologic data were derived. Since there was no apparent difference in the pharmacokinetic and dosimetric parameters for the 2 individual groups of patients, the data were combined. Twenty-six patients received a total of 51 injections (including 2 ¹³¹I-epratuzumab injections). Sufficient data were collected from 49 injections to evaluate blood pharmacokinetics. Twenty-two patients received both the ¹¹¹In- and ⁹⁰Y-epratuzumab doses.

Imaging studies were initiated within 90 min of the end of the ¹¹¹In-epratuzumab infusion, starting with rapid anterior or posterior whole-body images (25–30 cm/min) that were used primarily to monitor the clearance of the radioactivity from the body. Imaging was performed on a single- or dual-head camera fitted with a medium-energy collimator. Most of the initial whole-body images were performed before voiding, with the remaining images repeated 1.5–4 h later, and then again on at least 3 or 4 occasions, typically days 1, 2, and 3 or 5, and then again on day 6 or 7 after the ¹¹¹In-epratuzumab infusion. Anterior or posterior planar views were also taken starting from 0.5 to 5 h after the end of infusion and then repeated in most patients 3–5 more times over 7 d at the same times that whole-body imaging was performed. SPECT was also performed in some patients on 1 or 2 occasions, usually on day 2 or 3. A nuclear medicine physician at each site who was aware of prior history and radiographic results initially confirmed targeting of at least 1 known lesion to justify the subsequent treatment with ⁹⁰Y-epratuzumab. All imaging studies were subsequently transferred digitally to GSCC, where a nuclear medicine physician independently interpreted the ¹¹¹In-epratuzumab targeting and correlated his findings to baseline CT readings, which were also made independently on films acquired 1–5 wk before the ⁹⁰Y-epratuzumab treatment. No imaging studies were performed after the ⁹⁰Y-epratuzumab injection.

Planar views were used for drawing regions of interest for the major organs, from which time–activity curves and absorbed radiation doses were derived. Activity quantitation was determined by a geometric mean method (20,21). Regions of interest were drawn for the kidneys, liver, lungs, and spleen. Regions of interest for red marrow were based primarily on the lumbar spine, but the sacrum was used in the event of tumor involvement in the lumbar region. In cases where tumor (measurable by CT) could be visualized in both the anterior and posterior images, regions were also drawn to derive tumor dosimetry. Each set of organ or tumor time–activity data was fit to a mono-, bi-, or triexponential function using COSAM, but most often a triexponential function was selected. For each method, dividing the area under the curve by the injected activity derived a residence time that was used in MIRDOSE3 (22) to calculate the radiation-absorbed dose. Absorbed doses were calculated for ⁹⁰Y from the ¹¹¹In-epratuzumab imaging study using the biologic clearance curves for ¹¹¹In-epratuzumab and substituting the physical t_1/2 of ⁹⁰Y to derive an effective clearance curve for ⁹⁰Y. Organ volumes were assumed to be those given by the International Commission on Radiological Protection (23). For tumor dosimetry, volumes were derived from bidimensional CT measurements, using the average of the 2 diameters to derive a radius, and from this, a spheric volume was calculated. The integral for the tumor time–activity curves was multiplied by the tumor S factor that was interpolated from S values of the module in COSAM (conversion SAAM version 3.0) software.

Dose Escalation and Safety/Efficacy Evaluations

Treatment with ⁹⁰Y-epratuzumab IgG started at 0.185 GBq/m² for patients who had prior HDC (group 2) and at 0.370 GBq/m² for those who did not (group 1). Groups of 3 patients were enrolled at each dose level unless dose-limiting toxicity (DLT) was observed, in which case up to 3 more assessable patients were included at that dose level. Incremental increases of 0.185 GBq/m² were permitted for each group if safety was observed in at least 3 of 3 or 5 or 6 patients at a given dose level. DLT was initially defined as grade 4 hematologic toxicity lasting >7 d or grade 3 for >15 d and any grade 3 nonhematologic toxicity lasting >7 d or grade 4 of any duration. The hematologic DLT definition was modified during the course of the trial to any grade 4 hematologic toxicity lasting >5 d (neutropenia without intervention or platelets without transfusion). Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria version 2.0. A minimum observation period of 6 wk was required before determining if escalation could proceed to the next higher dose level. With hematologic toxicity expected to be dose limiting, weekly blood counts were scheduled until the nadir was observed, with twice weekly monitoring if ≥grade 3 was observed. Nonhematologic toxicity was assessed by serum chemistries in all patients over a 4-wk period, with 10 of the 23 patients who received ⁹⁰Y-epratuzumab IgG having additional serum chemistry monitoring for 8–12 wk. Physical examinations were performed generally at 4 and 12 wk after ⁹⁰Y-epratuzumab. Although a formal evaluation of efficacy was not the goal of this study, several responders are worthy of review. In these instances, baseline and follow-up CT scans (slices of 7–8 mm) were obtained from the study sites and an independent radiologist reviewed the films and documented the size of suspicious lesions, first identifying the longest diameter (LD) and then making a second, perpendicular measurement across the LD in the other direction. The area of the lesion (cm²) was calculated from the product of these perpendicular diameters (PD). The mass of these lesions was estimated by assuming a spheric volume with the radius equal to the average of the 2 LD/2. The total tumor burden was calculated by summing the PD for all regions (SPD), with the percentage change in total tumor burden defined as ([T_n –T_b]/T_b) × 100, where T_n is the SPD from a CT study performed some time after ⁹⁰Y-epratuzumab treatment and T_b is the SPD for the baseline CT study performed within 1–5 wk of ⁹⁰Y-epratuzumab treatment.

Human Anti-Epratuzumab Antibody Response

Antibody responses to epratuzumab IgG were measured by an enzyme-linked immunosorbent assay (ELISA). Briefly, epratuzumab IgG-coated microtiter plates were incubated with a 1:10 dilution of serum collected before epratuzumab exposure and at several times after treatment (e.g., 2, 4, 8, and 12 wk). The diluted serum was coincubated in the wells with an epratuzumab IgG-horseradish peroxidase conjugate, and binding was revealed by the addition of OPD substrate (Sigma). Development of the substrate was stopped after 30 min with the addition of 4.0N H₂SO₄, and the microtiter plate was read in a spectrophotometer at 490 nm. An anti-idiotype antibody to epratuzumab, WN (18), served as a standard. The assay sensitivity cutoff was 5.0 ng/mL and, therefore, because serum samples used in the assay were diluted 1:10, the minimum sensitivity for detecting human anti-epratuzumab in the patient serum samples was 50 ng/mL.

Injection Information

Table 1 lists patient demographics with the histologic type, prior treatment history, and the treatment course on this protocol. Fourteen patients were enrolled in group 1, with 13 receiving treatment (6 men, 7 women; age range, 43–86 y; median age, 68 y). Seven of the ⁹⁰Y-epratuzumab–treated patients had indolent histologic types, and 6 had aggressive histologies including 1 transformed and 1 mantle cell lymphoma. Five of these patients had received only 1 chemotherapy regimen with or without external-beam radiation therapy (XRT), with the remaining patients having 2–5 separate chemotherapy regimens (median, 3) with or without XRT. Five patients had prior rituximab therapy. Two group 1 patients (1-1 and 1-2) also received a pretherapy ¹³¹I-epratuzumab dose 7 d before receiving the ¹¹¹In-epratuzumab. Only 1 of these 2 patients was subsequently treated with ⁹⁰Y-epratuzumab, 12 d after the ¹¹¹In-epratuzumab imaging study. All other patients in this group only had the ¹¹¹In-epratuzumab imaging study performed 6 or 7 d before the ⁹⁰Y-epratuzumab therapy.

Twelve patients (8 men, 4 women; age range, 31–77 y; median age, 57.5 y) had prior HDC (group 2), with 10 patients subsequently receiving ⁹⁰Y-epratuzumab. Two patients were withdrawn after undergoing the ¹¹¹In-epratuzumab imaging study. Nine of the ⁹⁰Y-epratuzumab–treated patients had aggressive types of NHL and 1 had an indolent form. Most of these patients had ≥2 chemotherapy regimens before receiving HDC as well as XRT. Seven had prior rituximab, alone or in combination with chemotherapy. Two of the treated patients had total-body irradiation (TBI) as part of their HDC regimen. ⁹⁰Y-Epratuzumab was given 6–8 d after the ¹¹¹In-epratuzumab injection (2 patients had the ⁹⁰Y-epratuzumab treatment 11 and 15 d after their ¹¹¹In-epratuzumab study).

Biodistribution and Tumor Targeting

Figure 1 shows the anterior view from the fast whole-body scans, used primarily for calculating whole-body clearance, to illustrate the biodistribution of ¹¹¹In-epratuzumab over 7 d. At 4 and 24 h after injection, blood-pool activity is the most prominent feature, with activity in the heart, large pulmonary vessels, major blood vessels, vertebrae, liver, spleen, and kidneys. The bladder was also usually visualized on the 4-h prevoiding view and then variably at 24 h. The heart, pulmonary vessels, liver, spleen, and kidney were seen through 48 h, but the vertebrae became less apparent in most patients. The liver and, in some instances, the spleen often continued to be the only organs of appreciable uptake after 72 h; however, the blood-pool activity of the heart was also seen up to 72 h in 16 of 26 patients (61.5%), and this often persisted for as long as 168 h. In those patients with persistent visualization of the heart blood pool, there was often more activity seen in the lumbar spine region (10/16; 62.5%), and in 8 of the 16 patients (50%), the kidneys persisted beyond 72 h. In addition, the typical biodistribution included activity in the nasopharynx and the anterior and posterior perineum at all time points, and there was variable activity seen in the bowel from 24 h and beyond. In the 23 patients who had not undergone splenectomy, the liver had more activity than the spleen in 10 patients (43%), was equal to the splenic activity in 4 patients (17%), and the splenic activity was greater than the hepatic activity in 9 patients (39%). These findings were unrelated to clearance of the heart blood pool. Of the 5 patients where prestudy biopsies had indicated bone marrow involvement, the 3 who cleared the heart blood pool by 72 h showed only mild visualization of the lumbar vertebrae, whereas the 2 patients with persisting heart blood-pool activity demonstrated more intense lumbar spine activity, with the individual vertebrae clearly delineated from each other.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Anterior fast (30 cm/min) whole-body images illustrate distribution of ¹¹¹In-epratuzumab (0.218 GBq) in 54-y-old man with follicular NHL. Images taken at prevoiding (i.e., 1 h after end of infusion), 4, 24, and 48 h demonstrate blood-pool activity in heart and large blood vessels. Liver and spleen are seen, and bladder is seen only on prevoiding image. Images at 120, 144, and 168 h after injection show heart blood pool fading with persistent liver and spleen activity. There is perineal activity on all images, and nasopharyngeal activity through 48 h, in this patient. Tumor uptake can be seen in left neck and right iliac regions at 24 h. Additional sites in left supraclavicular region, left axilla, and 1 more right iliac/inguinal site are seen at 48 h. Tumor-to-background ratio improves on later images.

As illustrated in the whole-body image (Fig. 1), most tumors were clearly visualized by 48 h, although delayed imaging was often necessary to confirm pathologic uptake. The lesion-to-background ratio, and therefore the confidence of the pathologic nature of the imaging findings, improved over time. A small number of tumors were seen within 24 h, as well as a small number that required 120 h before being visualized. Even with the normal distribution of ¹¹¹In-epratuzumab in the liver and spleen, it was possible to visualize metastatic lesions in these organs (Fig. 2). Overall, ¹¹¹In-epratuzumab identified 116 of the 165 lesions (70.3%) seen by CT or MRI, with an additional 20 unconfirmed positive lesions. On a regional (patient) basis, the sensitivity of targeting known lesions was between 75.0% (e.g., liver, 5/6 lesions detected in 3/4 patients) and 92.3% (e.g., pelvis, 16/23 lesions detected in 12/13 patients), with an overall regional sensitivity of 85.9%. On a regional analysis, there were 4 unconfirmed positive findings, 3 in the pelvis and 1 in the liver.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Posterior image was taken 144 h after injection of 0.215 GBq ¹¹¹In-epratuzumab in 44-y-old woman who had 3 sites (arrows) of metastatic NHL in right lobe of liver. Uptake was clearly identified with ¹¹¹In-epratuzumab.

Tumors were visualized in both patients given ¹³¹I- and ¹¹¹In-epratuzumab, but in 1 patient there were 6 known lesions in the neck, supraclavicular, and mediastinal regions targeted with ¹¹¹In-epratuzumab that were not seen with ¹³¹I-epratuzumab, while other lesions in the abdomen and pelvis were seen with both. In all instances, the ¹¹¹In-epratuzumab had more intense uptake, lesion-to-background ratios were visually better, and heterogeneity of uptake and nodular appearance were visualized better with the ¹¹¹In-epratuzumab.

Pharmacokinetics and Organ Dosimetry

Table 2 provides a summary of the effective and biologic clearance estimated from monoexponentially fit data, since this adequately defined the clearance in a majority of the cases. In addition to the measured values, estimates derived for ⁹⁰Y-epratuzumab blood clearance using the ¹¹¹In-epratuzumab clearance data are shown. A paired t test analysis indicated that there was no significant difference between the measured biologic blood clearances of the 2 injections (P = 0.458). The measured effective t_1/2 for ¹¹¹In-epratuzumab and ⁹⁰Y-epratuzumab in the blood was 36.1 ± 7.9 h compared with 35.2 ± 7.0 h. The measured residence time for the ¹¹¹In-epratuzumab injections was 10.5 ± 2.8 h/L and 8.8 ± 2.9 h/L for the ⁹⁰Y-epratuzumab injections, whereas the average predicted effective residence time for ⁹⁰Y-epratuzumab estimated from the ¹¹¹In-epratuzumab injection was 10.1 ± 2.9 h/L. An average bioequivalency approach (24), which involves the calculation of the 90% confidence interval for the ratio of the averages of the biologic t_1/2 and residence time measurements (both measured and extrapolated), was evaluated in the 22 patients where both the ¹¹¹In- and subsequent ⁹⁰Y-epratuzumab injections were given. Logarithmic transformation (log₁₀) of the data was used and a bioequivalence limit of 80%–125% for the ratio of the method averages was adopted for use of an average bioequivalence criterion. The anti-logarithm of the mean difference between 2 methods and a 90% confidence interval are listed is Table 2. This analysis indicates that there is bioequivalency between the 2 injections for each of the evaluated parameters. Thus, the first injection (¹¹¹In-epratuzumab) did not have a significant impact on the blood clearance of the second injection (⁹⁰Y-epratuzumab). Even in the 1 patient who received ¹³¹I-epratuzumab, ¹¹¹In-epratuzumab, and, finally, ⁹⁰Y-epratuzumab, the clearance data for ¹¹¹In- and ⁹⁰Y-epratuzumab were similar.

The whole-body clearance kinetics were defined from the whole-body images taken during the ¹¹¹In-epratuzumab IgG study in 20 patients (18/20 had a prevoid image). The effective t_1/2 of ¹¹¹In-epratuzumab was 61.4 ± 5.2 h, and the predicted effective t_1/2 for ⁹⁰Y-epratuzumab was 58.3 ± 4.7 h. Since the effective t_1/2 of ¹¹¹In-epratuzumab was nearly equal to the physical t_1/2 of ¹¹¹In (67.4 h), it suggests that there was very little excretion of radioactivity from the body. This was confirmed by scintigraphy, which failed to reveal any appreciable radioactivity passage in the gastrointestinal tract or excessive activity in the bladder, as well as by direct measurement of ¹¹¹In and ⁹⁰Y radioactivity in urine samples taken over 3–4 d. Similar effective half-lives were derived from the time–activity data for the ¹¹¹In clearance by either whole-body scans or urine collection (Table 3). Furthermore, a comparison of the predicted whole-body clearance for ⁹⁰Y-epratuzumab from the ¹¹¹In-epratuzumab urine data also correlated well with the measured t_1/2 by counting ⁹⁰Y-epratuzumab in the urine. The daily sampling of urine indicated that the largest fraction of radioactivity was found in the first 24-h collection, which amounted to between 1.0% and 12% of the injected activity. We suspect that the majority of this radioactivity was due to the rapid urinary excretion of the DTPA-bound ¹¹¹In or ⁹⁰Y (i.e., non-antibody bound). Over the entire 3- to 4-d urine collection period, 2.2%–15.9% of the total injected activity was excreted.

Organ dosimetry data derived for ¹¹¹In- and ⁹⁰Y-epratuzumab calculated from the ¹¹¹In-epratuzumab imaging study are presented in Table 4. Even though patients who had enlarged spleens were not included in this dataset, uptake in the spleen was variable. The radiation-absorbed dose estimates derived in this trial for ⁹⁰Y-DOTA-epratuzumab were similar to those reported previously for ⁹⁰Y-benzyl-DTPA-epratuzumab, which were derived from ¹¹¹In-benzyl-DTPA-epratuzumab imaging studies over a similar time period (16).

Adverse Events Associated with Epratuzumab Infusion

Seven of the 26 patients had adverse experiences after the radiolabeled epratuzumab injection. Most of these experiences occurred during the first infusion and were not observed during the subsequent infusion. The most severe occurred in 2 patients who had grade 2 fever, grade 1 tachycardia, and grade 1 rigors. One of these 2 patients also had grade 1 diaphoresis and grade 1 hypotension. These symptoms started ∼15 and 20 min after the end of the antibody infusion and lasted for 2 and >3 h, respectively. Five other patients had fever ± chills, with agitation (n = 1), right flank pain (n = 1), and minor facial flushing (n = 1). These symptoms subsided spontaneously (n = 2) or after giving the patient diphenhydramine ± corticosteroids and acetaminophen (n = 3) within 3–4 h.

Hematologic Toxicity

Table 5 provides an overview of the hematologic toxicity observed in all patients.

Hematologic Toxicity as a Function of Red Marrow Radiation-Absorbed Dose.

Although prior HDC and bone marrow involvement were identified as risk factors for developing more severe myelosuppression, as shown in Figure 3, there was no relationship between the total red marrow dose as determined by external scintigraphic imaging (total cGy) or from the blood clearance curves (i.e., effective residence times determined after the ⁹⁰Y-epratuzumab injection) and toxicity grade for neutropenia and thrombocytopenia irrespective of the these factors. Of the 4 patients who received red marrow doses of >250 cGy (i.e., treated at a higher dose level), 3 had grade 2 and 3 neutropenia and thrombocytopenia, indicating that as the red marrow dose increases (and greater radioactivity injected), more severe hematologic toxicity occurs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Relationship between total radiation-absorbed dose to red marrow dose (cGy) or blood clearance, as represented by effective residence time (RT-E; h/L) and hematologic toxicity (graded at its nadir by National Cancer Institute Common Toxicity Criteria version 2.0). Red marrow dose was determined in most patients by scintigraphic methods. ▴, Group 1; □, group 2.

Nonhematologic Adverse Events

Thirty-nine and 40 adverse events were noted in group 1 and group 2 patients, respectively. However, the majority of these events were judged by the investigators as not related to the treatment. Table 6 lists all nonhematologic adverse events that were considered to have a possible relationship to ⁹⁰Y-epratuzumab. All of these events were mild to moderate.

Human Anti-Humanized LL2 Antibody Response

Sixteen patients had anti-antibody responses monitored over a period of at least 4 wk, and, in most, up to 12 wk after ⁹⁰Y-epratuzumab using an ELISA. Two patients were found to have an anti-antibody response, one at a level of 70.7 ng/mL detected 12 wk after treatment and another that was first detected at 8 wk after treatment (126 ng/mL) and subsequently rose to 316 ng/mL at 12 wk. No additional follow-up was pursued. One patient had a false-positive (FP) reading of 244 ng/mL at baseline. This patient received both ¹¹¹In- and ⁹⁰Y-epratuzumab because, based on the anti-antibody SE-HPLC assay (24) that was used to evaluate patients at this time, there was no evidence of the formation of high-molecular-weight complexes when the patient’s serum was incubated with radiolabeled epratuzumab. Furthermore, the blood clearance and biodistribution of ¹¹¹In-epratuzumab, as well as the blood clearance of the ⁹⁰Y-epratuzumab, were not different from that of the other patients, suggesting that no complexation occurred. At this time, the cause of this FP reading by ELISA in this patient is unknown.

Tumor Targeting, Dosimetry, and Anti-Tumor Response

Fifteen of the 23 patients (6 indolent and 3 aggressive in group 1; 1 indolent and 5 aggressive in group 2) given ⁹⁰Y-epratuzumab were selected to have detailed measurements of tumor burden before and after treatment based on follow-up studies that suggested they might have had either stabilization of their disease or an objective response. Figure 4 shows a scattergram of the maximum percentage decrease in the PD for the individual lesions measured at either 1, 3, or 6 mo after ⁹⁰Y-epratuzumab compared with their baseline PD. It is important to emphasize that several lesions continued to decrease in size, with their maximum responses at 9–15 mo after ⁹⁰Y-epratuzumab. However, since the majority of the lesions showed near-maximum responses within 3–6 mo, these data reflect maximum decreases achieved within this time frame. A total of 88 lesions were measured in these 15 patients, of which 61 were identified in the ¹¹¹In-epratuzumab imaging study (i.e., true-positive [TP]) and 27 were not seen (i.e., false-negative [FN]). Fifty-three lesions were identified in the patients who had an indolent histology, with 23% of these lesions being FN. In patients with an aggressive histology, 46% of the 35 lesions were FN. The highest proportion of FN lesions occurred in the chest, with 10 FN findings. Rapid blood clearance could not explain the FN findings, since 56% of the FN lesions were found in patients who continued to have discernable heart uptake after 72 h. Assuming that lesions ≤1.5 cm in diameter might not be so easily detected by planar imaging, TP and FN groups were further subdivided based on whether the LD of these lesions were >1.5 cm or <1.5 cm. The average length of time to achieve the maximum percentage reduction in size was between 2.8 and 3.8 mo for the 4 groups. Of the 61 TP lesions, 56 were ≥1.5 cm in their LD (maximum, 15 cm; mean ± SD, 3.7 ± 2.8 cm; median, 2.8 cm). Only 1 of these 56 TP lesions (baseline LD, 2.3 cm; PD, 4.6 cm²) increased in size by 20% after treatment. The average maximum percentage decrease in the PD for the TP lesions was 71.6% ± 29.4% (n = 55; median, 77.5%). Five TP lesions ranged in size from 1.0 to 1.3 cm in their LD. One of these lesions increased in size, whereas 2 decreased by 40%–55% and the other 2 decreased by >90%.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Influence of tumor size and targeting on response to ⁹⁰Y-epratuzumab. Fifteen patients were selected to have measurements of lesions in their baseline and follow-up CT studies. PD was determined from product of LD and a diameter perpendicular to LD through next largest size of lesion. Maximum percentage decrease was determined from follow-up CT studies performed at 1, 3, or 6 mo after ⁹⁰Y-epratuzumab treatment. Each lesion was also scored against whether it was seen with ¹¹¹In-epratuzmab using baseline CT as standard comparator. ♦, TP; □, FN.

Twelve of the 27 FN lesions were <1.5 cm in their LD at baseline (average LD, 1.1 ± 0.3 cm), which could in part explain the failure to detect these lesions by ¹¹¹In-epratuzumab imaging. Only 1 of these lesions (LD, 0.9 cm) failed to respond after treatment, whereas the other 11 FN lesions had an average decrease in the PD of 54.5% ± 34.1%. Three of the 15 FN lesions that were >1.5 cm in the LD increased in size, while the average maximum percentage decrease in the PD for the remaining 12 FN lesions (average LD, 2.6 ± 1.0; range, 1.6–5.2 cm) was 70.0% ± 30.2%. By histologic type, irrespective of whether the lesion was TP or FN, in patients with indolent NHL, the average decrease in size was about 77%. For the aggressive histologies, the TP lesions decreased in size by an average of about 73% (n = 19), whereas the FN lesions decreased by an average of 45% (n = 16). Overall, these data suggest that anti-tumor responses of potentially equal magnitude can occur irrespective of tumor targeting and tumor size. Indeed, using a Kruskal–Wallis test, we found that, while the PD for the baseline tumors was significantly higher for the TP lesions (median, 5.3 cm²) than the FN lesions (median, 1.8 cm²; P < 0.0001), there was no significant difference of the maximum percentage decrease between TP (median, 77.5%) and FN (median, 68.0%; P = 0.354) lesions. The data also suggest that although nontargeted lesions in patients with aggressive NHL can respond to treatment, they may not respond to the same magnitude as lesions with an indolent histology.

Ten patients had sufficient targeting to allow regions to be drawn in both the anterior and posterior projections as well an ability to determine their volume unambiguously by CT so that dosimetry could be determined (Table 7). These data show a considerable variability in the normalized absorbed dose to the tumors (i.e., cGy/MBq) with no evidence that this value was related to tumor size (Fig. 5, Top). Even within the 4 patients where dosimetry was calculated for >1 tumor, there was evidence of considerable variability in the calculated dose to the 2 tumors. Most of these tumors received a total radiation-absorbed dose of <1,000 cGy (Fig. 5, Middle), but the majority of these lesions responded to treatment. There was no evidence of a critical threshold for the total radiation dose delivered to a tumor for a given lesion to decrease in size, since at least 4 lesions that received similar doses failed to respond, and even a single small lesion (3.9 g) that was calculated to have received nearly 5,600 cGy failed to respond. The radiation dose delivered to the tumor in most instances was >2 times that to critical organs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Relationships between tumor dosimetry, tumor size, and response. Tumor size was determined from baseline CT. Maximum percentage change in tumor size is based on comparison of baseline measurements of product of 2 longest diameters (PD) compared with follow-up measurements of same lesions at time of their maximum reduction in size. Tumors that progressed in size after treatment were designated as +10%.

There were several notable responses in this study. For example, patient 1-5 (indolent NHL) presented with extensive, bulky lymphadenopathy in all regions of the body (baseline PD, 421 cm² with 8/17 lesions having a LD of >4.0 cm) and had >70% reduction based on the SPD in the disease in the chest, abdomen, and pelvis within 3 mo of ⁹⁰Y-epratuzumab treatment (0.370 GBq/m²). Figure 6 shows an anterior planar image 3 d after the ¹¹¹In-epratuzumab injection, revealing a large mass in the central abdomen that represents coalescent mesenteric (15 × 12 cm) and retroperitoneal (13.5 × 8.5 cm) lymph nodes seen in the baseline CT. Sixteen weeks later, these lesions decreased to 6.0 × 1.5 cm and 9.0 × 3.0 cm, respectively. This case also had bilateral pelvic wall disease. ¹¹¹In-Epratuzumab only identified 1 of these lesions (Figs. 6D–6F), but both lesions had significant responses after ⁹⁰Y-epratuzumab. At about the time these responses were observed, the patient complained of back pain that resulted in a PET study that revealed the presence of a new, previously undocumented L1 lesion. XRT was given to this region to relieve the pain. Despite this local intervention, all other lesions continued to decrease in size, achieving >90% reduction in the SPD for the chest, abdomen, and pelvic lesions, with the LD of the 2 largest lesions in the abdomen (LDs of 15.0 and 13.5 cm) decreasing to 1.0 and 3.2 cm, respectively, 20 mo after treatment, when the last follow-up CT evaluation was measured. This patient did not require any additional treatment for 33 mo after ⁹⁰Y-epratuzumab. Two other patients (1-9 and 1-11, also indolent NHL) treated at 0.555 GBq/m² also had significant anti-tumor responses. One patient had extensive lymphadenopathy in the neck, chest, and pelvis with 6 of 8 lesions with an LD between 2 and 4.2 cm (SPD, 36 cm²) and, within 3 mo of treatment, had >90% reduction in the SPD with all lesions having an LD of <1.0. Figure 7 illustrates planar and SPECT views 2 d after the ¹¹¹In-epratuzumab targeting study of a right iliac lymph node, along with CT scans showing this same lesion at baseline (3.8 × 1.5 cm) and 12 wk later (measured 0.5 × 0.5 cm). This response is ongoing after 28 mo of monitoring. The other patient originally presented with 5 lesions in the region of the neck and chest, all about 2.0 cm in diameter. Within 1 mo after treatment, the SPD had decreased 64% and, at 9 mo after treatment, all lesions were <1.5 cm with an 83% reduction in the SPD. At the last monitoring 19 mo after treatment, the patient continued to have no evidence of disease recurrence. Two patients with aggressive NHL (2-1 and 2-3) who had prior HDC also experienced objective responses. Patient 2-1 had 4 measurable lesions, which were stable 5 mo after treatment, but when measured 1 y later, 2 of the initial lesions, which at baseline had LDs of 2.2 cm, were not seen, and there was an overall 81.1% decrease in the SPD with all nodes’ LD ≤1.3 cm. This patient has currently not received additional treatment nearly 4 y after ⁹⁰Y-epratuzumab therapy. Patient 2-3 presented with 6 lesions in the abdomen and pelvis, including one 2.7 × 2.7 cm and another 3.3 × 3.3 cm in the left and right kidneys, respectively. Within 1 mo after treatment, there was a 79% reduction in the SPD with disappearance of the right kidney lesion and a residual left kidney lesion of 1.0 × 0.4 cm (94.5% reduction in PD for this lesion). Figure 8 illustrates the ¹¹¹In-epratuzumab targeting of the mass in the left kidney without any evidence of targeting of the mass in the right kidney. However, the CT scans clearly show the complete resolution of the mass on the right side and significant shrinkage of the mass on the left. This patient’s disease remained relatively unchanged until nearly 1 y after treatment, when a new 0.5 × 0.4 lesion appeared in the right kidney.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Patient 1-5 is 46-y-old man with follicular NHL who was treated with 0.829 GBq (22.4 mCi) ⁹⁰Y-epratuzumab (0.37 GBq/m², 10 mCi/m²). (A) ¹¹¹In-Epratuzumab anterior planar image of abdomen at 48 h where heterogeneous tumor uptake in coalescing nodular masses is seen that corresponded with large mesenteric and retroperitoneal masses seen on baseline CT (B). (C) Sixteen weeks after treatment, CT shows significant anti-tumor effects. (D) Anterior planar image of pelvis at 48 h demonstrates focal uptake of ¹¹¹In-epratuzumab in right pelvic side-wall disease seen (arrow) on baseline CT (E; arrows show bilateral tumor involvement) representing TP finding, but no definite tumor uptake in left pelvic side-wall disease seen on baseline CT (FN). (F) Significant improvement of both TP and FN lesions is shown on 16-wk follow-up CT image.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Patient 1-9 is 54-y-old man with follicular NHL who was treated with 0.944 GBq (25.5 mCi) ⁹⁰Y-epratuzumab (0.55 GBq/m², 15 mCi/m²). Transaxial image of pelvic SPECT (A) and planar anterior ¹¹¹In-epratuzumab image of pelvis (B) performed at 48 h after injection demonstrate tumor uptake in a right iliac node (arrows) that corresponds with baseline CT lesion (C, arrow). (D) Note marked improvement on follow-up CT 12 wk after ⁹⁰Y-epratuzumab treatment.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Patient 2-3 is 56-y-old man with aggressive NHL (follicular mixed) who previously had HDC with peripheral blood stem cell transplant 11 mo before receiving ⁹⁰Y-epratuzumab. He was treated with 0.333 GBq (9 mCi) ⁹⁰Y-epratuzumab (0.185 GBq/m² or 5 mCi/m²). (A) Posterior planar image taken 1 h after ¹¹¹In-epratuzumab illustrates positioning of left (outlined) and right kidney. At 24 h (B) and 48 h (C), uptake in left mass is seen to develop (arrow), whereas uptake in right mass could not be demonstrated. Four weeks after treatment, both left (D [baseline] and E [4-wk follow-up] CT) and right (F [baseline] and G [4-wk follow-up] CT) lesions had significant reductions in size, with complete disappearance of right renal mass and only small residual mass remaining on left side.