A Phase I Study of ⁹⁰Yttrium-Ibritumomab-Tiuxetan in Children and Adolescents with Relapsed/Refractory CD20-Positive Non–Hodgkin's Lymphoma: A Children's Oncology Group Study Free

Mature B cell–non–Hodgkin's lymphoma (B-NHL) represents ∼60% of all NHL that occurs in children and adolescents (1, 2). The 5-year event-free survival of children and adolescents with localized and advanced mature B-NHL after short and intensive multiagent chemotherapy now approaches ≥98% and 80% to 90% respectively (3–5). The Children's Oncology Group reported the results of an international multicooperative group trial (French-American-British) using a lymphoma malignancy B type chemotherapy (3) that resulted in ≥99% 5-year event-free survival with 3 weeks of therapy in children with localized disease, ≥90% 5-year event-free survival in children with intermediate disease after 12 weeks of therapy, and ≥80% 5-year event-free survival in children with advanced disease after 24 weeks of multiagent chemotherapy (6–8). However, there is a high degree of grade III/grade IV toxicity (mucositis/infection) and prolonged hospitalization with this therapeutic approach (7, 8). Furthermore, the retrieval rate for children and adolescents with relapsed/refractory mature B-NHL is dismal with 2-year overall survival, ranging between 10% and 30% (9, 10). Targeted therapy approaches may be warranted to either reduce the morbidity in children and adolescents with newly diagnosed mature B-NHL and/or improve the retrieval survival rate in patients with relapsed/refractory disease.

CD20, a transmembrane protein, is expressed at high levels on most mature B cells and adult B-cell lymphomas (11–13). We identified that CD20 is highly expressed on ≥98% of all childhood and adolescent mature B-NHL (14). CD20 is an ideal tumor target as it does not seem to modulate and is neither secreted nor shed. Rituximab, a chimeric antibody that targets CD20-positive B cells, enhances antibody-dependent cellular cytotoxicity and complement-dependent cellular toxicity. Phase I/phase II trials with rituximab in adult relapsed/refractory low-grade and follicular NHL showed overall responses in >50% of patients (15–17). The combination of cyclophosphamide-Adriamycin-vincristine-prednisone therapy and rituximab was associated with a significantly improved event-free survival and overall survival in elderly adults with diffuse large B-cell lymphoma (18). The safety and potential efficacy of rituximab in combination with French-American-British/lymphoma malignancy B 96 chemotherapy in children and adolescents with newly diagnosed mature B-NHL is currently under investigation by the Children's Oncology Group (19).

Mechanisms of resistance to rituximab may in part be related to poor access of antibody to tumor cells because of bulk tumor and/or inadequate vascular supply or failure of host effector mechanisms to kill lymphoma cells despite antibody binding (20). One potential therapeutic approach to circumvent these limitations with naked monoclonal antibody is the utilization of radioimmunoconjugates, in which a radioisotope is conjugated to a monoclonal antibody (20). Mature B-NHL seems to be inherently radiosensitive (21). Therefore, radioimmunoconjugates provide an opportunity to administer targeted radioimmunotherapy that may circumvent the limitations of poor access of naked antibody and/or lack of host effector mechanisms by its bystander or crossfire effect compared with naked monoclonal antibodies (20).

Yttrium-90 (⁹⁰Y) is a pure emitting B radioisotope that delivers five times more energy compared with iodine-131 (¹³¹I) and has a path length of 5 mm versus 0.8 mm and a more favorable half-life of 2.5 days versus 8 days compared with ¹³¹I (20). Presumably, the crossfire from ⁹⁰Y isotope high-energy B particle not only induces cell death to the CD20⁺ B-NHL cells it binds but also to neighboring B-NHL cells inaccessible because of tumor bulk and/or poor vascular supply or expressing low levels of CD20 antigens. ⁹⁰yttrium-ibritumomab-tiuxetan (⁹⁰Y-IT) is a radioimmunoconjugate composed of the murine parent of rituximab, ibritumomab (Y2B8), an IgG1 antibody, with tiuxetan (MX-DTPA), a bifunctioning chelating reagent with a strong affinity for both ⁹⁰Y and indium-111 that can be used for therapy and imaging, respectively (20). In a randomized controlled study comparing ⁹⁰Y-IT versus rituximab in adults with relapsed or refractory low-grade, follicular or transformed B-NHL, the overall response rate and complete response rate was significantly improved with radioimmunoconjugate (80% versus 56%, P ≤ 0.002) and (30% versus 16%, P < 0.04), respectively (22). In this phase 1 study, we investigated the safety, toxicity, and dosimetry of ⁹⁰Y-IT in children and adolescents with relapsed/refractory CD20-positive NHL.

Patient eligibility. Patients with recurrent/refractory CD20⁺ NHL, Hodgkin's lymphoma, posttransplant lymphoproliferative lymphoma, HIV-associated NHL, or lymphoblastic lymphomas were eligible for Children's Oncology Group phase 1 and pilot consortium protocol (ADVL0013). Patients were required to have had histologic and immunophenotypic CD20⁺-positive verification of the lymphoma at original diagnosis and/or at progression or relapse. Central pathology review was required for all patients (see below). Selection criteria included age, performance level, life expectancy, prior therapy, and organ function (Table 1). All patients were required to have a minimum of 2 × 10⁶ CD34/kg peripheral blood stem cells collected and cryopreserved before therapy for back-up in case of delayed hematopoietic recovery. Patients who had a prior myeloablative stem cell transplant and/or received extensive radiotherapy were stratified as poor marrow reserve, whereas those patients who had no history of myeloablative stem cell transplant and/or extensive radiotherapy were considered good marrow reserve. Extensive radiotherapy was defined as ≥3,600 cGy total body irradiation to the craniospinal axis and/or ≥50% to the marrow space. Patients with documented active CD20⁺ central nervous system lymphoma and/or known or suspected hypersensitivity to rituximab, ibritumomab-tiuxetan, Yttrium chloride and indium chloride were not eligible. Each participating clinical site had received institutional review board approval, and all patients and/or their parents or legal guardians provided written informed consent.

Treatment schema and dosimetry. The original trial design called for a dose escalation schema to enroll cohorts of three to six patients. However, the design was subsequently modified to a single-dose level pilot study secondary to slow accrual (see Results). All patients received rituximab (250 mg/m² i.v.) on days 0 and 7 and indium-111 ibritumomab-tiuxetan (5 mCi i.v.) on day 0 immediately after the first dose of rituximab. Gamma camera imaging was conducted within 1 h of the indium-111 ibritumomab-tiuxetan infusion on day 0 before urination, within 4 to 6 h of the indium-111 ibritumomab-tiuxetan infusion on day 0 and on days 1, 3, and 6 after the indium111 ibritumomab-tiuxetan infusion. Whole-body scans involved both anterior and posterior images at a speed of 10 cm/min (20-min scan) on days 0 and 1, 7 cm/min (30-min scan) on day 3, and 5 cm/min (40-min scan) on day 6 using a medium energy collimator, a 256 × 1024 computer acquisition matrix, and acquisition photo peak settings of 172 and 247 keV with 15% windows. A standard of ∼50 uCi in 10 mL water was used with each planar image acquisition. The day 3 image was the last image used for dosimetry calculations to allow for the dosimetry to be calculated before the ⁹⁰Y-IT infusion. Blood samples for dosimetry calculations were additionally obtained within 1 h of the indium-111 ibritumomab-tiuxetan infusion on day 0, within 4 to 6 h of the indium-111 ibritumomab-tiuxetan infusion on day 0 and on days 1, 3, and 6 after the indium-111 ibritumomab-tiuxetan infusion. Blood samples were drawn from a peripheral i.v. site contralateral to the indium-111 ibritumomab-tiuxetan infusion site. The whole-blood samples were counted after the day 3 sample had been drawn to allow for the dosimetry to be calculated before the ⁹⁰Y-IT infusion. The exact activity in the whole blood was calculated from the volumetric dilution of the measured indium-111 ibritumomab-tiuxetan against a standard of ∼0.5 uCi indium.

Whole-body and organ activities at each imaging time for kidney, liver, lungs, red marrow, spleen, and whole-body remainder were determined using geometric mean (GM) counts based on both the imaging standard and a whole-body calibration factor. Assuming no excretion occurred before the first imaging session, the 1-h whole-body GM counts were equated with the decay corrected injected dose (A₁). Subsequent whole-body activity at time t, A_t, was calculated as:

in which GM_t is the whole-body GM counts at time t postinjection and GM₁ is the whole-body GM counts at 1 h. Organ activity O_t was calculated as:

in which GMO_t is the organ GM counts at time t. This was previously proved to be very accurate in the quantification of uptake in spleen, liver, and whole-body remainder in experimental baboons (23). The blood-derived method was used to estimate red marrow doses, whole-blood indium-111 (μCi/mL) concentrations were decay corrected to determine the ⁹⁰Y activities (μCi/mL) which were then plotted on the Y axis versus the sampling in hours on the X axis. A monoexponential or biexponential curve fit of the data was done, and the curve was integrated to obtain the area under the curve which was the whole blood residence time (τ_b). The red marrow residence time was then calculated as:

The red marrow mass used in the residence time calculation was 1,120 g for males and 1,050 g for females.

Immediately after the end of the rituximab infusion on day 7, patients received ⁹⁰Y-IT if whole-body scan and blood dosimetry studies showed ≤2,000 cGy for all solid organs and/or ≤300 cGy for bone marrow. The first dose level to be tested in good marrow reserve patients was 0.4 mCi/kg ⁹⁰Y-IT, and poor marrow reserve patients were treated with 0.1 mCi/kg ⁹⁰Y-IT. No additional dose levels were studied. Patients with CD20 Burkitt's, Burkitt-like, and lymphoblastic NHL received central nervous system prophylaxis with cytarabine in age-adjusted doses every 3 weeks for three doses beginning at day 0 of therapy.

Radiological evaluation and human antimurine antibody/human antichimeric antibody determination. A history, physical exam, vital signs, performance status, blood hematology and chemistries, urinalysis, and quantitative immunoglobulins (IgG, IgA, and IgM) were done on all patients at baseline and repeated at specified time points in the study. Human antimurine antibody and/or human antichimeric antibody analysis was done by Biogen Idec Pharmaceuticals on all patient samples taken at baseline and day 35. All patients underwent a chest X-ray, computed tomography scan, MRI, gallium scan, and/or positron emission tomography scan of areas of suspected disease at baseline and day 35. Response was evaluated according to the National Cancer Institute's Response Evaluation Criteria in Solid Tumors criteria (24). A unilateral bone marrow aspirate and biopsy with flow cytometry for CD20 was done on all patients within 1 week before day 0 and on day 35 if previously involved with tumor and/or if absolute neutrophil count is ≤1,000/mm³ and/or platelet count is ≤20,000/mm³.

Central CD20 immunophenotyping. Patient samples underwent central analysis for CD20 phenotype via flow cytometry on day 0 or baseline in the peripheral and bone marrow and on days 7 and 35 in the peripheral blood. One-color fluorescence was analyzed according to standard procedures. Briefly, cells were labeled with fluorochrome-conjugated monoclonal antibody, anti-CD20 PE (clone L27; BD Biosciences). Labeling was allowed for 30 min in the dark at 4°C. The cells were then washed twice in PBS + 1% bovine serum albumin. Analysis was done with a FACScalibur flow cytometer (BD Biosciences) using Cell Quest software (BD Biosciences). In each sample, a minimum of 10,000 events was acquired in the analysis region, using log-amplified fluorescence and linearly amplified side-scatter and forward-scatter signals. All samples were analyzed by setting appropriate side-scatter/forward-scatter gates around the lymphocyte population. Jurkat cultured cells (American Type Culture Collection) were used as a negative control, whereas Raji cultured cells (American Type Culture Collection) were used as a positive control. PE-conjugated mouse IgG1 antibody was used as an isotype control.

Central hematopathology review and CD20 immunohistochemistry. All cases entered on the study were reviewed for accuracy of diagnosis and immunohistochemical confirmation of CD20 expression in the tumor cells (SP). Morphologic review was done on materials submitted by the institution in which the original diagnosis was made via H&E sections and tumors were classified using the WHO lymphoma classification (25). In addition, all cases were tested for CD20 expression (Dako) by standard immunohistochemical staining techniques and automated staining on the Ventana ES platform (Ventana Medical Systems). CD20 staining was done after heat-induced epitope retrieval or citrate buffer for CD74 (pH 6.0) for 3 min in an electric pressure cooker (BioCare Medical). Signal was detected using the IView 3,3′-diaminobenzidine detection kit (Ventana), and slides were counterstained with hematoxylin. CD20 staining was evaluated for the percentage of tumor cells expressing the antigen and the intensity of staining (14).

Toxicity evaluation. Toxicity was evaluated using National Cancer Institute Common Toxicity Criteria, version 2.0. Hematologic dose-limiting toxicity was defined as grade IV neutropenia or grade IV thrombocytopenia of >7 days duration, and if the absolute neutrophil count does not reach ≥1,000/mm³ and/or platelet count did not recover to ≥20,000/mm³ by day 35. Nonhematologic dose-limiting toxicity was defined as any grade III or grade IV nonhematologic toxicity attributable to the investigation agent with the specific exclusion of grade III nausea and vomiting, grade III transaminase (aspartate aminotransferase/alanine aminotransferase) elevation that returned to grade I or baseline by day 35, and grade III fever or infection. Serious adverse events, which occurred during the treatment period and/or within 30 days of the last dose of treatment, were recorded and reported to CTEP via the National Cancer Institute Adverse Event Expedited Reporting System. All data was reported to a centralized study database within the Children's Oncology Group and verified by annual audit.

Demographics. Five patients (four male), median age is 12 years (range, 5-18 years), were registered and treated on ADVL0013 between September 2002 and January 2004 (Table 2). Of the five patients, four patients were fully evaluable for toxicity; one patient had disease progression (patient 5, refractory posttransplant lymphoproliferative disease) on day 23 before dose-limiting toxicity evaluation. Patient diagnoses and disease status were as follows: diffuse large B-cell lymphoma (DLBCL; first relapse in one patient, second relapse in two patients), refractory Burkitt's lymphoma (one patient), and refractory posttransplant lymphoproliferative disease (one patient).

Review of the submitted pathology slides showed four cases of DLBCL and one case of Burkitt lymphoma (see Fig. 1A and C). One case of DLBCL was seen in a patient with a previous history of transplantation and could also be classified as a monomorphic posttransplant lymphoproliferative disorder, DLBCL. One case of DLBCL arose in the mediastinum and was associated with extensive sclerosis. All cases studied showed strong CD20 expression in all of the tumor cells (Fig. 1B and D). Figure 1E represents a T-lymphoblastic lymphoma demonstrating negative staining for CD20.

Toxicity. All patients received rituximab (250 mg/m² i.v.) on days 0 and 7 and indium-111 ibritumomab-tiuxetan (5 mCi i.v.) on day 0. Three patients with good marrow reserve received 0.4 mCi/kg and two patients with poor marrow reserve received 0.1 mCi/kg ⁹⁰Y-IT. No patients experienced hematologic or nonhematologic dose-limiting toxicity. Two patients had grade III/grade IV thrombocytopenia (nadir 24 and 35 days), two patients had grade IV neutropenia (nadir 20 and 35 days), one patient had grade III anemia, and one had grade III infection related to study treatment. No patients required growth factor support. Two patients experienced grade I muscle pain and cramping, and one patient experienced grade II infusion–related chills associated with rituximab. Other hematologic and nonhematologic adverse events were either grade I or grade II and not listed. Quantitative immunoglobulins (mg/dL) at day 35 were as follows: mean serum IgA, 65; mean serum IgG, 394; and mean serum IgM, 32. No patients developed human antimurine antibody/human antichimeric antibody responses.

Tumor assessment. No patients achieved a complete response or partial response. However, the day 35 maximum and mean positron emission tomography standard uptake values of the FDG avid disease in the anterior-superior mediastinum of patient 3 were reduced by 27% and 35%, respectively, from baseline. This is suggestive of disease stabilization in the absence of a Response Evaluation Criteria in Solid Tumors documented response.

Central CD20 immunophenotyping. CD20 expression was evaluated at baseline, in both peripheral blood and bone marrow aspirate samples, and at days 7 and 35 in the peripheral blood. In evaluable patients with ≥2% CD20 circulating B cells at baseline, there was a decrease in the number of cells expressing CD20 when comparing the peripheral blood baseline to samples evaluated at day 35 (Fig. 2A-E).

Dosimetry results. All patients showed <2,000 cGy exposure to all solid organs and <300 cGy to marrow and received 0.4 mCi/kg ⁹⁰Y-IT (n = 3) or 0.1 mCi/kg ⁹⁰Y-IT (n = 2) immediately after rituximab on day 7 (n = 4) or ∼24 h post rituximab on day 8 (n = 1; Table 3A). The mean exposure was 46 cGy in the bone marrow, and mean of 309, 341, 345, and 565 cGy in the lung, kidney, liver, and spleen, respectively (Table 3B). A representative indium-111 ibritumomab-tiuxetan scan is illustrated at 4 h and 7 days, respectively (patient 4; Fig. 3A and B).

This phase I study showed that ⁹⁰Y-IT was well-tolerated in a small cohort of heavily pretreated children and adolescents with relapsed/refractory B-NHL. In comparison to adults with low-grade or follicular B-NHL, who are less heavily pretreated with prior myelosuppressive chemotherapy, children and adolescents usually have had significant myelosuppressive chemotherapy during their original treatment for their underlying mature B-NHL. In the aggregate safety data of ⁹⁰Y-IT in adults with low-grade, follicular, or transformed B-NHL, the major toxicity was hematologic (20, 26). The incidence of grade III and grade IV neutropenia was 30% and 30%, respectively, and grade III and grade IV thrombocytopenia was 53% and 10%, respectively (26). The median time to neutrophil and platelet nadir in adult recipients was 63 and 55 days, respectively, and the nadirs lasted on average of 3 weeks in duration (26). The median neutrophil and platelet nadirs in adult recipients were 800/mm³ and 37,500/mm³, respectively (20, 26). The hematologic toxicity data is very limited in our small group of patients in the current study. However, because of being more heavily pretreated with prior myelosuppressive chemotherapy, there seems to be an earlier hematologic nadir of ⁹⁰Y-IT in pediatric patients compared with the experience seen in adults. Because of the limited patients available for this study, we were unable to determine a maximal tolerated dose in either the good or poor marrow reserve group.

Other toxicities associated with ⁹⁰Y-IT in adults has been minimal. The incidence of febrile neutropenia has only been 29%, and more importantly, only a total of 5% have experienced grade III/grade IV infection (26). The most frequent adverse events in adults have been grade I/grade II asthenia (35%), nausea (25%), and chills (21%), particularly during and right after the infusion. Despite ibritumomab being a murine antibody, the incidence of human antimurine antibody formation is only 1% in adult recipients (26). In our small group, none developed human antimurine antibody formation. An initial major concern in this patient population who have received prior alkylator therapies and radiation therapy was the development of myelodysplastic syndrome (myelodysplastic syndrome/acute myeloid leukemia). However, long-term follow-up studies have only shown a 1% incidence in adults who have received ⁹⁰Y-IT (20, 27). Because of the short survival in our high-risk patient population, we were unable to assess the incidence of secondary myelodysplastic syndrome/acute myeloid leukemia in this study.

The current study in children and adolescents with ⁹⁰Y-IT determined organ dosimetry and set a limit of 2,000 cGy to any solid organ and/or 300 cGy to red marrow as a maximal exposure. All dosimetry was determined before the ⁹⁰Y-IT dosing on day 8. All patients were well under the maximal threshold allowed for study drug administration (Table 3A). The mean dose to the organs were as follows: red marrow 46 cGy, liver 345 cGy, kidney 341 cGy, spleen 565 cGy, and lung 309 cGy (Table 3B). These values compare very favorably with dosimetry results in adults receiving ⁹⁰Y-IT (28–30). In the most recent report by Wiseman et al. (30), combining data from four clinical trials, the median doses of radiation to adult recipients receiving this radioimmunoconjugate were as follows: red marrow 62 cGy (blood-derived method), spleen 742 cGy, liver 450 cGy, lung 211 cGy, and kidney 23 cGy (Table 4). It seems, from our limited study in children and adolescents, that ⁹⁰Y-IT radiation exposure is similar or even less than seen in adult recipients.

This was a limited study in very high-risk patients with progressive intermediate and high-grade B-NHL, and response was not a major primary end point. Four of these patients were in second or subsequent relapse. No patients achieved a complete remission or partial remission. In adults with relapsed or refractory low-grade follicular or transformed B-NHL, the response rate was significantly higher in first relapse patients compared with those with two or more prior therapies (P < 0.01; ref. 31). In the initial phase I/phase II study by Witzig et al. (32) in adults with low-grade or follicular B-NHL, the overall response rate and complete response rate was 82% and 26%, respectively (32). In two subsequent phase II studies in a similar adult population, the response rate and complete response rate ranged from 80% to 84% and 34% to 44%, respectively (33, 34). In the randomized study comparing ⁹⁰Y-IT to four weekly doses of rituximab, the overall response rate and complete response rate were 80% versus 56% (P < 0.002) and 30% versus 16% (P < 0.04), respectively (22). There is, however, no data regarding the efficacy of this radioimmunoconjugate in high-grade mature B-NHL, i.e., Burkitt's lymphoma, which is the most common form of childhood and adolescent B-NHL. Although the numbers are small in this study, the lack of response in patients with higher grade B-NHL and/or in patients having received two or more prior therapies was, therefore, not unexpected.

⁹⁰Y-IT has also been successfully incorporated in adult patients with B-NHL as part of the ablative conditioning regimen followed by autologous stem cell transplantation (35) and as part of targeted radioimmunotherapy in reduced intensity conditioning followed by allogeneic stem cell transplantation (36). These studies are in a preliminary stage in development, and there have been no studies done in children and adolescents to date of using this radioimmunoconjugate in the conditioning before stem cell transplantation. Lastly, Wiseman et al. recently reported the durability of responses after ⁹⁰Y-IT in adults with refractory/relapsed or transformed indolent CD20⁺ B-NHL. The median time to progression and duration of response in complete responders were 29.3 months and 28.1 months, respectively. In those patients achieving a complete response, the time to progression was 31 months (37).

In summary, this limited phase I study showed ⁹⁰Y-IT to be safe and well-tolerated in a small high-risk group of children and adolescents with progressive/refractory high-grade B-NHL. The dosimetry studies showed that the radioactive dose to normal organs was well below the maximal accepted doses of 2,000 cGy to solid organs and 300 cGy to red marrow. Recent studies in adults have shown that subsequent chemotherapy regimens are well-tolerated in adults after receiving ⁹⁰Y-IT for B-NHL (38). Recently, El Gnaoui et al. reported preliminary results of combination gemcitabine and oxaliplatin with systemic rituximab (39). They treated 24 patients with relapsed/refractory B-cell lymphoma with the most common histology being diffuse large B-cell lymphoma (16 of 24), followed by follicular (6 patients) and mantle cell (2 patients), and found the overall response rate after four cycles was 88% (39). Based on results in adults and our preliminary results in children, we are planning an investigator-initiated phase II trial using ⁹⁰Y-IT followed by gemcitabine and oxaliplatin after recovery of nadir in children and adolescents with relapsed/refractory CD20-positive diffuse large B-cell lymphoma to further assess the safety and overall response rate to radioimmunoconjugate alone and followed by chemotherapy (Fig. 4).