Halperin7e_CH29

C H A P T E R 2 9 Radioimmunotherapy and Unsealed Radionuclide Therapy

Tod W. Speer

INTRODUCTION The concept of a “magic bullet” or targeted therapy against a tumor was first proposed by Paul Ehrlich in 1898, ultimately allowing him to garner the Nobel Prize in 1908.As testimony to the complexities and challenges of his vision, it was not until a half a century later when radiolabeled antibodies were used in the clinic to target cancer. The modern exegesis for targeting agents does not limit the use of carrier molecules to mere anti- bodies. Successful targeting of tumor cells, with high affinity, can also be accomplished with antibody fragments, peptides, and affinity ligands. As the research and clinical arena ever so modestly disengage from intact antibodies as the carrier mol- ecule for the radionuclide, the impact of the immune system has been somewhat abrogated. Hence, the “immunotherapy contribution” of “radioimmunotherapy” (RIT) has become less pronounced. Perhaps, a more appropriate term for this tech- nology would be “targeted radionuclide therapy.” To date, RIT has made significant progress secondary to advances in cell biology, immunology, radiation oncology, nuclear physics, and chemical technology. There are greater than 8 million cancer deaths worldwide and nearly 600,000 cancer deaths in the United States annually. 1,2 Millions of can- cer patients each year exhaust available drug options and succumb to cancer. Current cancer drug development in the United States takes approximately 8 years and 1.3 billion dol- lars for FDA approval. 3 This process appears unsustainable. 4 RIT offers the potential to accelerate drug development. RIT exercises its cytotoxic action by delivering targeted molecular radiation to malignant tissue. Therefore, the success of RIT depends upon engineering a targeting construct that brings the radionuclide into close proximity of the target cell or tis- sue. The cytotoxic agent, the radionuclide, is always the same. Of course, there is a selected differential of energy deposition into tissue, depending upon the specific chosen radionuclide. Standard drug development relies upon identifying a differ- ent target receptor or pathway that is prevalent within a par- ticular type of cancer. Antibodies (or blocking agents) are then designed, over months or years, to interfere with the receptor or process. A cytotoxic evaluation is then performed. Not all antibodies are initially cytotoxic. Hence, the abovementioned timeline and cost prevail. Immunity refers to protection from disease or infectious agents. 5 Our immune system is composed of the cells and mol- ecules responsible for the immune response, which can be divided into an early (1- to 12-hour) reaction, termed innate immunity , and a late (1- to >7-day) reaction, termed adaptive immunity . The innate immune system comprises biochemi- cal and cellular mechanisms that exist prior to the introduc- tion of a “foreign” or infectious agent and results in a rapid response. The innate immune system consists of epithelial barriers, phagocytic cells (neutrophils, macrophages), natu- ral killer cells, the complement system, and cytokines. The adaptive immune system develops over time, becoming more effective with subsequent exposures of antigen. It exhibits the IMMUNOLOGY AND TARGETING CONSTRUCTS

ability to “remember” and to respond more quickly with con- tinued exposures to the same antigen. The adaptive immune system consists of lymphocytes and secreted antibodies. The adaptive immune system can be divided into humoral immu- nity and cell-mediated immunity. Concerning humoral immu- nity, B lymphocytes secrete antibodies for protection. With cell-mediated immunity, helper T lymphocytes either activate macrophages or cytotoxic T lymphocytes, which then directly destroy pathologic (infectious or malignant) cells. It is well known that the host’s immune system is impor- tant for preventing the growth and development of cancer. 6 A large body of literature exists supporting the concept that the host immune system interacts with tumorigenesis and tumor progression. It has been shown in animal models and in the clinic that cancer immune surveillance is exceedingly impor- tant. For example, mice with an impaired innate or adaptive immune system will be more susceptible to develop chemically induced or spontaneous cancers. Additionally, the malignant transformation of cells in animals and humans, caused by the accumulation of somatic mutations and/or the deregulation of oncogenes or tumor suppressor genes, results in the expression of tumor antigens (TAs). These TAs are often recognized by the immune system as documented byTA-specificT-cell precursors and natural killer cells, found in the peripheral blood of can- cer patients, capable of killing tumor cells. Further evidence of cancer immune surveillance exists in patients with genetic or drug-induced immunosuppression. Transplant patients exhibit a predisposition for certain malignancies (squamous cell car- cinoma, basal cell carcinoma, Kaposi sarcoma, melanoma, and lymphoma). Patients with Chédiak-Higashi and Wiskott- Aldrich syndrome demonstrate an increased rate of lympho- proliferative malignancies. Discontinuing immunosuppressive drugs in solid organ allograph patients with occult malignant melanoma has resulted in tumor regression. Despite the evidence of cancer genesis and progression in immune-compromised hosts, the majority of cancers develop in seemingly immune-competent individuals. The last decades of research have revealed that cancer cells have developed means to avoid immune detection and surveillance, either through the selection of nonimmunogenic tumor cells or the active suppression of the immune response. It has therefore been rightfully suggested that “tumor immune escape” be added to Hanahan and Weinberg’s six hallmarks of cancer (self-sufficiency in growth signals, insensitivity to antigrowth signals, tissue invasion and metastasis, limitless replicative potential, sustained angiogenesis, and evasion of apoptosis). Interestingly, recent progress has been achieved utilizing drugs for immune checkpoint blockade therapy by target- ing the programmed death protein (PD-1) with antibodies. Unfortunately, response rates have been limited. 7 Currently, the quest for the “Holy Grail” vaccine that turns the immune system against cancer remains elusive. 8 The targets for RIT typically consist of tumor-associated anti- gens (TAAs) expressed on the surface of tumor cells or in the abnormal extracellular matrix. The reason for this is because the cytotoxic radionuclide must be delivered preferentially to malignant tissue and should avoid normal tissue.To date, >2,000 TAAs have been identified (http://www.re3data.org/repository/ r3d100012052). One of the main methodologies used to identify TAAs is termed SEREX (serologic analysis of recombinant cDNA

651

C H A P T E R 2 9  Radioimmunotherapy and Unsealed Radionuclide Therapy

The ideal target for RIT-targeting constructs would be one that is overexpressed on cancer cells, is uniformly expressed, is not found to any significant level in normal tissue, is not shed into the circulation, and exhibits an important role in tumor growth and progression. 10 TAAs, as the name implies, are antigens “associated” with tumors but are also present in normal tissue. True tumor-specific antigens have not yet been identified and utilized. Overexpression is necessary because typical targeting constructs require antigen densities ≥10 5 receptors on each cell for adequate targeting. A homogenous antigen expression is desired so that a uniform activity dis- tribution of the radionuclide will result. Nonuniform activity distributions (heterogeneity of antigen in target tissue being one potential cause) will significantly lower the effectiveness of RIT by subsequently resulting in nonuniform or heteroge- neous dose distributions. 11 This is particularly important for radionuclides with short path lengths of the emitted particles (i.e.,Auger and α -particle emitters). Radionuclides with longer path lengths, such as high-energy β -emitters, can partly over- come the problem of nonuniform dose distributions through the crossfire effect. If the target antigen is significantly shed into the circulation, the targeting construct may bind and “complex” with the antigen. This will result in a more rapid clearance of the RIT agent and a much less effective treat- ment. If the TAA has an important signaling role, then subse- quent binding of the targeting construct will most likely add to the cytotoxicity of the radionuclide because of the blockade or promotion of intracellular signaling, potentially resulting in disruption of growth pathways important for tumor growth. Some TAAs (receptors) will internalize when bound by the tar- geting construct. In truth, most receptors internalize, although they do so at different rates. A rapid internalization process will have an impact on the type of radionuclide that is selected and potentially on the delivery strategy of the RIT agent. A multitude of agents have been used as carriers (target- ing constructs) for the targeted delivery of radiation to can- cer. These consist of antibodies, antibody fragments, peptides, affibodies, aptamers, and nanostructures (i.e., liposomes, nanoparticles, microparticles, nanoshells, and minicells). By an exceedingly large margin, intact monoclonal anti- bodies (mAbs) have dominated the field of RIT as targeting constructs 12 (Fig. 29.1). In humans, there are five classes or isotypes of antibodies (IgA, IgD, IgE, IgG, and IgM). IgG is the most commonly used mAb for RIT because it is the most prev- alent antibody in serum and has the longest serum half-life, typically measured in weeks (~23 days). IgG is further divided into four subtypes, IgG 1–4 . IgG antibodies are large glycopro- tein macromolecules, with an atomic mass of approximately 150,000 dalton (Da) or 150 kDa. The “y-shaped structure” (Fig. 29.1A) consists of two Fab fragments (antigen-binding fragment; ~50,000 Da each) and an Fc fragment (crystalliz- able fragment; ~50,000 Da). The “tip” of each Fab fragment has a variable amino acid sequence, from one mAb to another. Accordingly, each tip is an antigen-binding site (ABS) and is responsible for antigen recognition. Each ABS forms a noncovalent bond (electro- static forces, van der Waals forces, hydrophobic interactions, and hydrogen bonds) with the target or antigen. The specific region of an antigen, which binds to the ABS, is referred to as an epitope. It has been proposed that a million or more different antibodies exist in various individuals. Theoretically, >10 9 different antibodies can be produced. The outer core of the mAb consists of two identical light chains (outer por- tion of the Fab fragment) designated with an “L.” The inner core, consisting of the Fc region and the inner Fab region, is designated as heavy or “H.” Both the light and heavy chains contain homologous, 110 amino acid sequences that fold on one another and are connected by a disulfide bridge, result- ing in “globular” motif or loop, called an Ig domain. There are three constant heavy domains (C H 1-3) and only one constant

libraries). SEREX involves a bacteriophage recombinant cDNA expression library, prepared from various malignancies (iso- lated tumors or malignant cell lines) or testis tissue. 9 This cDNA expression library is transduced in Escherichia coli to produce a recombinant protein library.These various proteins (clones) are then tested against the serum from autologous cancer patients. Clones that react to IgG antibodies are identified and are then further characterized as TAAs. Many of these SEREX-identified TAAs have been elucidated by other processes and laboratories. This has led to the concept of a finite number of TAAs that are produced in cancer patients and are potentially identified by the immune system.These finite TAAs are collectively referred to as the cancer immunome. SEREX-defined antigens, representing broad categories, may be organized as follows: mutational anti- gens, amplified or overexpressed antigens, differentiation anti- gens, and cancer/testis antigens.Within these categories, only a limited number of solid tumor TAAs have been used as targets for RIT (Table 29.1). TABLE 29.1  SELECT TUMOR ANTIGEN TARGETS AND MONOCLONAL ANTIBODIES EVALUATED FOR SOLID TUMOR RADIOIMMUNOTHERAPY Malignancy Antigen Antibody Colorectal cancer CEA

Section II

cT84.66, hMN-14, A5B7, TFT, IMP-288, CC49 B72.3, CC49, A33 huA33 NR-LU-10, NR-LU-13 chTNT-1/B F19

TAG-72 A33 EpCAM DNA histone H1 FAP

Breast cancer

MUC1 L6 TAG-72 CEA Lewis

huBrE3, M170 chL6 CC49 cT84.66 B3 HMFG-1 cMov18 B72.3, CC49 Hu3S193 chTNT-1/B CC49 cT84.66 U36, BIWA4 425 81C6, BC4 chTNT-1/B

Ovarian cancer

MUC1 Folate receptor TAG-72 Lewis

Prostate cancer

PSMA TAG-72 MUC1

huJ591, 7E11-C5.3, PSMA-617 CC49 M170

Lung cancer

DNA histone H1 TAG-72 CEA EGFR Tenascin-C DNA histone H1 p97 Ganglioside GD2 Melanin CD44v6 Ganglioside GD2 NCAM Tenascin-C Ganglioside GD2 4Ig-B7-H3 DNA/histone Ganglioside GD2 A33 CAIX/MN

Head and neck cancer

Glioblastoma

Melanoma

96.5 3F8 PTI-6D2 cG250

Renal cancer

Medullary thyroid cancer CEA

cT84.66, hMN-14, NP-4, F6-734, Labetuzumab

Neuroblastoma

3F8 UJ13 A, ERIC-1 81C6

Gastric

huA33

Brain tumors

3F8 8H9 chTNT-1/B

Medulloblastoma

3F8

CEA-producing tumors

CEA

M5A

Pancreatic cancer

MUC1

hPAM4

Leptomeningeal cancer

Ganglioside GD2

3F8

Gastrointestinal CC49 CAIX/MN, carbonic anhydrase IX; CEA, carcinoembryonic antigen; EGFR, epidermal growth factor receptor; EpCAM, epithelial cell adhesion molecule; FAP, fibroblast-activating protein; MUC1, mucin 1; NCAM, neural cell adhesion molecule; PSMA, prostate-specific membrane antigen; TAG-72, tumor-associated glycoprotein. Modified from Wong YC, Williams LE, Yazaki PJ. Radioimmunotherapy of colorectal cancer. In: Speer TW, ed. Targeted radionuclide therapy . Philadelphia: Lippincott Williams & Wilkins, 2011:325. With permission. TAG-72

652

S E C T I O N I I  Techniques, Modalities, and Modifiers in Radiation Oncology

FIGURE 29.1.  Antibody configurations for radioimmunotherapy. A: Typical structure of a humanized IgG antibody. B: Antibody fragments and size. C: Attachment of a nuclear localization signal. D: Humanization: green represents murine portion of IgG and red represents human portion. CDR, complementarity determining regions; CH, constant domain heavy chain; CL, constant domain light chain; Fab, antigen-binding fragment; F(ab’) antigen-binding fragment (retaining a portion of the hinge region following enzymatic digestion); Fc, crystallizable fragment; Fv, variable fragment; PKKKRKV, amino acid sequence of nuclear localization signal; scFv, single-chain variable fragment.

cytotoxicity (CDC). 5 Concerning ADCC, interaction of the Fc region of the antibody with Fc receptors (located on immune effector cells) results in the subsequent phagocytosis or lysis of the antibody-bound cancer cell. CDC is initiated by the inter- action of soluble blood proteins and the Fc region. Epitope- dependent (Fc-independent) functions of the mAb may result in the inhibition of ligand binding, inhibition of ligand-induced dimerization, and inhibition of receptor shedding. These epi- tope-dependent functions are characteristic of modern-day biologics that target growth factor receptors, such as cetux- imab and trastuzumab. The original technology used to produce mAbs was first published by Kohler and Milstein in 1975 and is referred to as the hybridoma technique. The technique has propagated the use of murine mAbs for research and for therapy in the clinic. In fact, the two U.S. Food and Drug Administration (FDA) RIT agents used to treat non-Hodgkin lymphoma (NHL; ibritumomab tiuxetan and tositumomab) are murine mAbs. Although these agents are delivered as single instillations in patients typically with decreased immune recognition capa- bilities, there is a concern that human antiglobulin antibodies (HAGAs) will develop. If this phenomenon occurs in response to murine antibodies, then the resulting HAGAs will be called human antimouse antibodies (HAMAs). The formation of HAMAs will expedite blood clearance of the antibody and decrease targeting capabilities as well as potentially cause various adverse symptoms. Two main strategies, through the use of genetic engineering, have emerged 12 that reduce the immunogenicity of mAbs: (a) the production of antibody chimeras derived from both murine and human DNA and

light (C L

), one variable heavy (V H

), and one variable light (V L )

domain. The ABS consists of a V L region. Within each variable domain, there are three hypervariable regions (about 10 amino acid residues per hypervariable region) that form a three-dimensional surface that is “complementary” to the shape of the antigen surface; they are called complementarity determining regions (CDRs). A total of six CDRs come together to form the ABS. There are two ABS for each IgG mAb; hence, each IgG mAb is considered bivalent. Affinity refers to the strength of the bond between the ABS and the antigen.The strength of this bond is represented by the dissociation constant (K d ).Avidity refers to the overall strength of the ABS–antigen interaction, depending on both the affinity and the valency of the interaction. It should be noted that a high-affinity interaction can improve specific delivery of the RIT agent and reduce overall dosing requirements. Increasing the affinity indefinitely, however, may decrease tumor pen- etration. It has been demonstrated that an affinity of 10 − 7 to 10 − 8 M is needed for tumor retention, whereas affinities in the range of ≥10 − 10 to 10 − 11 M will result in retention in normal tissue and asymmetric binding in tumor tissue, termed the binding site barrier . 13 The binding site barrier phenomenon may be at least partially overcome by increasing the antibody mass, or the overall delivered quantity of antibody. Unconjugated antibodies—those not attached to a radio- nuclide or cytotoxic agent—will also mediate biologic activi- ties. These activities may be mediated by the Fc region of the mAb or may be Fc independent. Fc-mediated interactions are termed effector functions and consist of antibody-dependent cell-mediated cytotoxicity (ADCC) and complement-dependent and a V H

653

C H A P T E R 2 9  Radioimmunotherapy and Unsealed Radionuclide Therapy

diethylenetriamine penta- acetic acid [ 111 In-DTPA]) has been shown to be highly diagnostic for NETs. Affibody molecules 22 are classified as affinity ligands or scaffold proteins that are approxi- mately 7 to 9 kDa. These proteins are based on a 58 amino acid residue derived from staphylococcus protein A, which binds immunoglob- ulin. Various applications have been applied to affibody use, including radiolabeled targeting for therapy. Aptamers are single-stranded DND or RNA oligonucle- otides (8 to 12 kDa; 10 to 100 bases) that are selected in vitro from a random library by a process termed SELEX (systemic evolution of ligands by exponential enrichment). Aptamers are an attractive alternative to larger mAbs because they are chemically synthesized (do not require a biologic system such as mAbs), have a low cost of production, exhibit high affinities, have a small size, are rapidly cleared from the circulation, have an unlimited shelf life, exhibit rapid tissue penetration, and are nonimmunogenic. 23 Aptamers fold into unique sec- ondary and tertiary structures that not only exhibit high affin- ities for targets, but can also gain entry into target cells. The basic SELEX process consists of exposing a target to a random single-strand nucleic acid library, selecting candidates that bind to the target, repeating the process to further select can- didates with increasing affinity for the target, and sequencing the final candidates (Fig. 29.2). Variations of SELEX include selecting aptamers that bind cells, internalize into cells, or are delivered in vivo , respectively, termed cell-SELEX, cell inter- nalization SELEX, and in vivo SELEX. The basic configuration of an aptamer in a random library is a 5 ′ forward primer- binding site, a random region, and a 3 ′ reverse primer-bind- ing site. The selected candidates can therefore undergo PCR amplification. The typical complexity of random library is between 1 × 10 13–24 different random sequences. For example, if the random region consists of 30 possible nucleotides, then potentially 1 × 10 18 different aptamers will be available. In this particular example, 1 × 10 18 represents a different aptamer or “key” for every grain of sand in the entire world. This is an amazingly large and diverse screening tool. The major det- riment of using aptamers as targeting constructs for RIT is their short serum half-life (measured in minutes) secondary to nuclease degradation. Fortunately, research has shown that aptamers can be rendered nuclease resistant with the following modifications; 5 ′ -modified uracil, 4 ′ -thio, 2 ′ -fluoro, 5 ′ - α -P-borano, 2 ′ -amino, 2 ′ -deoxy-L-ribose, 5 ′ -phosphorothio, 2 ′ -methoxy, modification of bases. 29 Aptamers are amazingly versatile and can recognize nearly any type of target, from metal ions to whole cells and even entire organisms. Because of their chemical synthesis, aptamers seem to have a great chance of becoming a true tumor-specific and personalized delivery construct. Aptamers for cancer imaging or therapy are shown in Table 29.3. THE PHYSICS AND RADIOBIOLOGY OF RADIOIMMUNOTHERAPY RIT delivers radiation to the target tissue in a continuous, although declining, low–dose rate (LDR) fashion. Typical dose rates for RIT are in the range of 10 to 20 cGy per hour. The total dose delivered by RIT is low, in the range of 1,500 to 2,000 cGy, with an effective half-life of 24 to 72 hours. This can be compared to the high–dose rate (HDR) delivery of radiation by external beam radiation therapy (EBRT). EBRT typically will deliver radiation at a dose rate of 100 to 500

TABLE 29.2  TARGETING AND PHARMACOKINETICS OF INTACT IgG AND ANTIBODY FRAGMENTS IgG F(ab’)2 CH2 deletion Minibody Fab Diabody scFv MW 150 100 120 80 50 40–50 20–25 Serum half-life 2–3 d a 1 d Hours Hours Hours Hours 1 hr Metabolism Liver Liver Liver Liver Kidney Kidney Kidney Tumor uptake b ***** **** *** *** ** ** * Time to accretion Days 1 d Hours Hours Hours Hours 1 h

Section II

a Serum half-life for fully human IgG is approximately 23 days. b Tumor uptake values range from large (*****) to small (*). MW, molecular weight (kDa).

(b) the production of humanized or fully human antibod- ies (Fig. 29.1D). Chimeric antibodies retain murine V H and V L domains, whereas humanized antibodies retain murine CDRs. Fully human antibodies retain no murine components. Although the development of HAGA may not be important after a single dose of mAb in lymphoma patients, HAGA will have a greater detrimental impact for patients with solid tumors when treated with RIT. 10 It has been well proven that as the antibody changes from murine to humanized, the immunogenicity is lessened. This concept is important so that multiple doses or fractions of RIT can be delivered. Current technology allows for the production of fully human mAbs. The concept of adding a nuclear localizing signal to bring the mAb from the cell surface or cytoplasm into the cell nucleus is shown in Figure 29.1C. In this setting, Auger-emitting radio- nuclides will be effective. Another factor that is critical and influences antibody targeting and pharmacokinetics is antibody molecular size (Fig. 29.1B). As stated previously, RIT has been less success- ful for treating solid tumors than hematologic malignancies. This is largely because of the lack of radiosensitivity of epithe- lial tumors (compared to hematologic malignancies) and the poor penetration of mAbs into large tumors. The decreased penetration of 150-kDa antibodies into large tumors is a direct result of increased tumor interstitial pressure, an aber- rant tumor vasculature, and an abnormal tumor extracellu- lar matrix. 14–17 Additionally, 150-kDa antibodies need longer periods of time to accrete into tumors and have long serum half-lives. When radiolabeled, a long serum half-life of the targeting construct will increase exposure of the bone mar- row to radiation, which causes hematologic toxicity and limits the amount of antibody and radionuclide that can be given. To overcome some of these issues, methods have been used to generate antibody fragments of varying size and valency. These smaller fragments exhibit superior tumor penetration and clear more rapidly from the circulation. However, if clear- ance from the circulation is too rapid, this can further limit tumor penetration. Table 29.2 summarizes these general con- cepts for targeting constructs of various molecular weights. Although mAbs and their fragments represent the most commonly used targeting constructs for the delivery of a radionuclide to malignant tissue, other agents are either in use or are being investigated, consisting of peptides, 18–20 affi- body molecules, 21,22 and aptamers. 23,24 Nanostructures are also being investigated as carriers of radionuclides. 12,25,26 In their unmodified form, the targeting capabilities of nanostructures are rather nonspecific. 27 Peptides are amino acid sequences (typically 7 to 14 amino acids) that serve as opioids, hormones, sweeteners, protein substrate inhibitors, releasing factors, antibiotics, and cytoprotectors. 28 The overexpression of receptors that are specific for various peptides has led to the development of peptide-based radiopharmaceuticals. Somatostatin is one of the most common peptides and is overexpressed in a mul- titude of malignancies, including breast cancer, small cell lung cancer, medullary thyroid cancer, and neuroendocrine tumors (NETs). Somatostatin is rapidly degraded; however, its derivative, octreotide, is very stable. Octreoscan (indium-111

654

S E C T I O N I I  Techniques, Modalities, and Modifiers in Radiation Oncology

FIGURE 29.2.  SELEX: A random ssDNA (or ssRNA) library is synthesized and exposed to laboratory equipment and target matrix. The aptamers that bind are considered nonspecific and are removed. The remaining aptamers are exposed to the target protein. Aptamers that don’t bind are removed. Aptamers that bind the target are eluted and PCR amplified. Candidate “sense” strand aptamers are isolated and the SELEX process is repeated in order to refine the selection process. Final candi- dates are cloned and sequenced.

absorbed dose of ionizing radiation to cell death. When the log surviving fraction of irradiated cells is plotted on the ordi- nate and the dose (Gy) is plotted on the abscissa, a cell sur- vival curve is generated (Fig. 29.3). The “hit” that results in most lethal event is a double-strand break (DSB) of DNA. The mathematical term, α , represents the initial slope of the cell survival curve. It is a constant for a given tumor (or tissue) and can be thought of as the probability, per unit of absorbed dose, of creating a lethal DSB. 31 The target is the resulting

cGy per minute. This total dose range for RIT occurs despite overall very low percent injected doses (0.1% to 10.0%) that ultimately localize in target tissue. 30 Regardless, radiation- induced apoptosis still occurs. The most radiosensitive component of a cell is the DNA. Irradiation of tissue results in DNA damage. This damage may be either repaired or result in permanent damage. Permanent damage will cause cell death. By using a target-hit model, the tissue response end point of cell death may be used to relate

TABLE 29.3  APTAMERS FOR CANCER IMAGING AND THERAPY (PRECLINICAL AND CLINICAL) Aptamer Target Condition

Radionuclide

Application

AS1411

Nucleolin

Renal cell carcinoma, non–small cell lung carcinoma, leukemia

None

Therapy

AS1411

Nucleolin

Glial tumor Leukemia

67 Ga

Imaging Therapy Therapy Therapy Therapy Imaging Imaging Imaging Therapy Imaging Imaging Imaging Imaging Imaging Therapy Imaging Therapy Therapy Therapy Therapy Therapy Therapy

Sgc8 TD05 14-16 TTA1

Protein tyrosine kinase 7 (PTK-7) Immunoglobulin μ heavy chains (IGHM)

None None None None

Lymphoma, leukemia

NOX-A12

CXCL12/SDF-1

Glioblastoma, multiple myeloma, solid tumors

p68

Colon cancer Glioblastoma Breast cancer Prostate cancer Prostate cancer

Tenascin-C

111 In, 18 F

AptA, AptB

MUC1 PSMA PSMA

99m Tc

A9

89 Zr

A10

225 Ac 99m Tc 188 Re 111 In 99m Tc 99m Tc None 99m Tc None None None None None

F3 U2

hMMP-9 EGFRvIII

Various cancers, metastases

Glioblastoma

E07

EGFR ErbB2

EGFR-expressing cells HER2-expressing tumors CEA-expressing tumors

Mini 15-8

Apt3, Apt3–amine CEA

A30

HER3

Breast cancer

TTA1 5TR1

Tenascin-C

Breast, colon, lung, glioblastoma Breast, colon, lung, ovary, pancreatic

O-glycan-peptide

None (photodynamic therapy agent)

J18

EGRF

EGFR-expressing tumors

None (gold nanoparticles)

Clone5 CTLA-4 aptamer

Sialyl Lewis X

Sialyl Lewis X–expressing tumors

CTLA-4

CTLA-4 receptor

TGF- β

A07

Chinese hamster ovary

PDGF β -receptor

ST1571

Colon

III.1 Therapy CTLA-4, cytogenic T-cell antigen-4; CXCL2, C-X-C motif chemokine 12; EGFRvIII, epidermal growth factor receptor variant III; HER3, human epidermal growth factor-3; hMMP-9, human matrix metalloprote- ase-9; SDF-1, stromal cell–derived factor-1; TGF- β , transforming growth factor- β . Pigpen Glioblastoma

655

C H A P T E R 2 9  Radioimmunotherapy and Unsealed Radionuclide Therapy

RIT, the dose rate is 1,000-fold lower; therefore, the quadratic portion of the curve will have a much lower impact on survival because many SSBs, considered sublethal damage, will be repaired during the more lengthy delivery of LDR radiation. This will result in a “small” or absent observable shoulder and flattening of the cell survival curve. Thus, when estimating cell survival for RIT, α alone will define the radiosensitivity of the tumor (blue line in Fig. 29.3). Considering dose rate alone, RIT is approximately 20% less effective than HDR EBRT. Regardless, RIT does appear to be effective. This phenome- non can be attributed to many radiobiologic processes that appear to cause greater than predicted rates of apoptosis. These processes include low-dose/dose rate apoptosis, low- dose hyperradiosensitivity-increased radioresistance, inverse dose rate effect (G 2 synchronization), radiation-induced bio- logic bystander effect, and the crossfire effect. 30 The use of high-LET radiation, in the form of alpha particles or Auger electrons, will further increase cell kill. Various radionuclides have been proposed for the use in RIT (Table 29.4), and their physical properties have been extensively reviewed in the nuclear medicine literature. They can be grouped into three basic categories depending on the type of emitted particulate radiation. Radionuclides that emit high-energy electrons are referred to as β -emitters. These electrons have maximum path lengths in tissue from 0.6 to 12.0 mm. This translates into a range of approximately 60 to 1,100 cell diameters. The most commonly used β -emitters for RIT are yttrium 90 [ 90 Y], iodine 131 [ 131 I], and lutetium 177 [ 177 Lu]. The maximum range of electrons in tissue for 90 Y and 131 I is 12 and 2 mm, respectively. It should be noted, however, that 90% of the electron energy is deposited over 5.2 mm for 90 Y and 0.7 mm for 131 I. This range of 90% energy deposition is referred to as the R 90 . The most commonly used α -emitters for RIT are 211 At and 225 Ac. An α -particle is a helium nucleus that has a maximum range in tissue of 55 to 100 μ m (5 to 10

Section II

DSB, and the cell survival versus absorbed dose is a pure exponential function: S = e − α D where S is the surviving cell fraction and D is the mean absorbed dose. Ionizing irradiation may also cause nonlethal single-strand breaks (SSBs). If these events accumulate, they may become lethal. The constant, β , is used to describe this phenomenon and represents the more distant, “linear” portion of the cell survival curve. The linear- quadratic (LQ) model combines the two processes into a con- tinuously bending curve: S e D D = − − a b 2 The shoulder on the cell survival curve is typically observed when HDR radiation is employed (green line in Fig. 29.3). In FIGURE 29.3.  Cell survival curves following treatment with radiotherapy. The blue curve represents low–dose rate radiotherapy; the green curve represents high–dose rate radiotherapy. Increasing LET is represented by the red line (alpha particle radiation; RBE = 5) and gold line (Auger radiation; RBE = 7 to 9). (Adapted from Bernhardt P, Speer TW. Modeling the systemic cure with targeted radionuclide therapy. In: Speer TW, ed. Targeted radionuclide therapy . Philadelphia: Lippincott Williams & Wilkins, 2011:265. With permission.)

TABLE 29.4  POTENTIAL RADIONUCLIDES FOR RADIOIMMUNOTHERAPY Radionuclide Physical Half-Life E ave (MeV) a

Maximum Range in Tissue LET (keV/ μ m)

Approximate Cell Diameters

β -Particle

Beta-emitters Yttrium 90 Iodine 131 Lutetium 177 Rhenium 186 Rhenium 188 Phosphorus 32 Phosphorus 33 Holmium-166 α - Emitters Bismuth-213 Bismuth-212 Astatine-211 Actinium-225 Terbium-149 Copper 67

0.2

2.7 d 8.0 d 6.7 d 3.7 d

2.19 0.28 0.15 0.36 0.80 0.18 0.70 0.08 1.86 5.87 6.09 5.87 5.83 3.97

12.0 mm 2.0 mm 1.5 mm 3.6 mm 11.0 mm 2.8 mm 7.6 mm 0.6 mm 8.4 mm 55–60 μ m 60–70 μ m 55–60 μ m 60–90 μ m 30–60 μ m 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm 2–500 nm

400–1,100

10–230 4–180 15–360

17.0 h 2.6 d 14.3 d 25.3 d

200–1,000

5–210

760

60

1.1 d

840

α - Particle

80

45.7 min 60.6 min

5–6 6–7 5–6 5–8 3–6

7.2 h 9.92 d 4.12 h

Low-Energy Electron Emitters (Auger)

Low-Energy Electron

4–26

Iodine 125 Iodine 123 Gallium 67 Indium 111

60.1 d 0.55 d 3.26 d 2.80 d 6.01 h 4.33 d 4.02 d 2.80 d 38.9 h 72.9 h 57.0 h 4.4 h 27.7 d

0.030 0.030 0.009 0.026 0.018 0.053 0.063 0.072 0.028 0.078 0.012 0.013 0.005

<1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1

Technetium 99m Platinum-193m Platinum-195m Platinum-191 Antimony-119 Thallium 201 Bromine-77 Bromine-80m Chromium 51

a When appropriate. Data were obtained from Eckerman KF, Endo A. eds. MIRD Radionuclide Data and Decay Schemes . 2nd ed. SNM MIRD Committee, 2008. E max , maximum energy; LET, linear energy transfer.

656

S E C T I O N I I  Techniques, Modalities, and Modifiers in Radiation Oncology

cell diameters). Although it has a short range, the α -particle is very destructive and has a high linear energy transfer (LET). Low-energy electron emitters also emit radiation that is high LET and have path lengths between 2 and 500 nm (width of a double-strand helix). Auger emitters, such as 111 In or 125 I, are most effective if delivered to the nucleus of a cell or incor- porated into the DNA. It has been mathematically postulated that it will only take 60 decays of 125 I, coupled to DNA, to reduce cell survival to 50%. For a patient with 1-g circulating micrometastatic disease, 1,000 decays in the malignant cells can produce a probable cure. This amount of radiation cor- responds to 0.1-MBq injected activity. This represents 5 mSv per 1 year of background radiation for the average human. 32 To further place this in perspective, Auger emitters can be safely injected into humans with an activity between 100 and 350 mCi, perhaps even at higher activities. It should be under- stood, 100 mCi = 3700 MBq; 1 MBq = 1,000,000 dps. Because radionuclides have different energy spectra for their emitted particulate radiation, they will each interact with tissue and deposit their energy over varying distances. There is therefore a relation between the type of radionuclide, tumor size, absorbed dose, and ultimately tumor cure prob- ability (TCP). If it is assumed that a tumor has a spherical volume and contains a uniform and identical activity concen- tration of a radionuclide, then the TCP can be calculated for different radionuclides and tumor size. 33 Figure 29.4 illus- trates the relation between tumor mass and TCP for asta- tine-211 [ 211 At], lutetium 177 [ 177 Lu], 131 I, and 90 Y. As can be seen, there is an optimum tumor size for the different energy spectra for each radionuclide such that the TCP is maximized. If the tumor is small relative to the emission range, then much of the energy will be lost to the surrounding tissue and the absorbed dose will be low. As the tumor size increases, more energy is absorbed until the maximum TCP is reached. As the tumor further increases in size, the absorbed energy remains high, although fewer cells are affected by the radiation and TCP begins to decrease. 33 These observations move forward the concept of using multiple radionuclides, in the RIT pro- cess, that deposit their energies over different ranges in tis- sue. As a result, more energy could then be deposited into tumors of various sizes, potentially improving the therapeutic ratio. Labeling the targeting construct with the appropri- ate radionuclide (radiochemistry) is exceedingly important and equally complex. Radionuclides are attached to target- ing constructs by either using a “linker” molecule, termed a FIGURE 29.4.  Tumor control probability (TCP) for various radionuclides. TCP = 0.9 versus tumor mass. The optimal TCP for various tumor masses when treated with 211 At, 177 Lu, 131 I, and 90 Y. This corresponds to approximately 10 − 5 , 10 − 2 , 0.1, and 10 g, respectively. (From Bernhardt P, Speer TW. Modeling the systemic cure with targeted radionuclide therapy. In: Speer TW, ed. Targeted radionuclide therapy . Philadelphia: Lippincott Williams & Wilkins, 2011:266. With permission.)

bifunctional chelating agent (BCA), or by a chemical reaction that forms a covalent bond between the radionuclide and the targeting construct. Three basic scientific fields converged to make radiochemistry a reality: coordination chemistry, directed biologic targeting, and the medical application of radiopharmaceuticals. 34 In general, metallic radionuclides will require a BCA for labeling, and radiohalogens will require a chemical reaction (halogenation). The most prevalent thera- peutic radionuclides used in RIT are 90 Y (metallic radionuclide) and 131 I (radiohalogen). One of the most commonly used BCAs is DTPA—a polyaminopolycarboxylate straight chain ligand. Tiuxetan, a modified DTPA molecule, is used as a linker mole- cule to chelate 90 Y to ibritumomab ( 90 Y ibritumomab tiuxetan; Zevalin, Spectrum Pharmaceuticals, Henderson, NV).Tiuxetan forms a urea-type bond to the antibody (ibritumomab), and its five carboxyl groups interact with and chelate 90 Y to form a stable coordination sphere. The halogenation reaction that bonds 131 I to a protein-targeting construct ( 131 I tositumomab; Bexxar, GlaxoSmithKline, Philadelphia, PA; discontinued 2013) is called iodination. Although there are many permuta- tions of the iodination reaction, it basically inserts 131 I into a tyrosine group on the mAb without the need for a chelation molecule. Regardless of the required labeling technique, it is incumbent that a reasonably high labeling yield, unaltered biodistribution, stability of the radionuclide, and immunore- activity are preserved. Historically, a single instillation or fraction of the RIT agent is delivered systemically (i.e., Zevalin and Bexxar). It is well known that although relatively effective for hemato- logic malignancies, RIT is much less effective for treating solid tumors. Therefore, a number of strategies are being devel- oped that will potentially increase the effectiveness of RIT. These strategies include modulating the tumor microenviron- ment; using pretargeting techniques, extracorporeal delivery, combined modality therapy (CMT), fractionation, and multiple radionuclides (radionuclide cocktail); increasing antibody mass (the amount of antibody delivered systemically); altera- tion of the physical properties (size and affinity) of the target- ing construct; and employing different types of LET radiation (i.e., β -emitter vs. α -emitter). These strategies are designed to deliver more radiation to the tumor, make the radiation more cytotoxic, or decrease the exposure of radiation to bone mar- row. As a result, the tumor to blood ratio will increase, and ultimately, the therapeutic ratio will increase. The pretarget- ing strategy warrants further discussion. 35,36 Because radiolabeled mAbs take 2 to 3 days to localize or accrete into tumors, antibody-based RIT results in a prolonged exposure of the bone marrow to radiation, causing hemato- logic toxicity and rendering the bone marrow as the dose- limiting normal tissue. Accordingly, the tumor/blood ratios of mAb will only slightly favor the tumor. This situation can seriously limit the successful prospects of antibody-based RIT, especially for treating solid tumors. Smaller targeting con- structs (antibody fragments, peptides, aptamers) can be used for RIT, and they will exhibit pharmacokinetics that result in a more rapid blood clearance allowing for the administration of higher activities. Unfortunately, because of the lower over- all tumor accretion and retention of smaller constructs, the advantage of a more rapid blood clearance is usually offset. Therefore, the ideal delivery construct would manifest the high-affinity targeting properties of an intact mAb but exhibit the blood clearance pattern of a small molecular weight con- struct. This conventional wisdom is based upon using beta and alpha radionuclides, both of which will be toxic with long circulation times. If targeting constructs using Auger emitters can be engineered, the circulation time becomes rather imma- terial as Auger radionuclides are only cytotoxic if internalized into cells. Because no known construct manifesting all of these attributes exists today, pretargeting strategies have been developed.The basic premise of pretargeting is to separate the

657

C H A P T E R 2 9  Radioimmunotherapy and Unsealed Radionuclide Therapy

FIGURE 29.5.  Bispecific pretargeting proce- dure. The bsMAb is injected, and over several days, it will localize in the tumor and clear from the blood. The bsMAb shown in this example is based on the dock-and-lock method for pre- paring recombinant bsMAb that has two bind- ing arms for the tumor and one for the hapten. Once the molar concentration of the bsMAb is low enough, the radiolabeled hapten–peptide is given. The hapten–peptide has two haptens for more stable binding within the tumor, perhaps by cross-linking two adjacent bsMAb through a process known as the affinity enhancement sys- tem (AES). The peptide portion usually contains four to five D-amino acids with a single chelator bound to one of the amino acids that is used to capture the radionuclide. (From Sharkey RM, Goldenberg DM. Pretargeted radioimmunother- apy. In: Speer TW, ed. Targeted radionuclide therapy . Philadelphia: Lippincott Williams & Wilkins, 2011:194. With permission.)

Section II

its respective indications and there are other competing drugs that are not radioactive and don’t require close coordination with medical departments such as medical oncology, radiation oncology, and nuclear medicine.Table 29.5 lists FDA-approved and current phase III RIT drugs. Progress has been less san- guine for solid tumor malignancies, and phase III trials are lacking. 39,40 For the sake of clarity and brevity, this section will focus on clinically relevant phase II/III trials and U.S. FDA (or its international equivalent)-approved RIT therapeutics. Hematologic Trials and Approved Therapeutic Agents The National Comprehensive Cancer Network (NCCN) guide- lines have recommended RIT for the following follicular lym- phoma clinical situations: (1) first-line therapy for the elderly or infirm (Category 2B), (2) first-line consolidation (Category 2B), and (3) second-line (relapse/refractory) and subsequent therapy (Category 1). Initially, the NCCN guidelines rendered a Category 1 designation for first-line consolidation but down- graded this to Category 2B because of concerns about toxicity, although there is not a uniform consensus. 41 Zevalin contin- ues to show very promising results for follicular lymphoma first-line monotherapy, diffuse large B-cell lymphoma and mantle cell lymphoma consolidation and second-line therapy, and transplantation studies. To date, however, none of these approaches have reached clinical phase III status. Zevalin has the only U.S. FDA approval for first-line consolidation and second-line therapy. Currently, there is only one U.S. FDA-approved RIT agent in the United States: 90 Y ibritumomab tiuxetan (Zevalin; 2002). 131 I tositumomab (Bexxar; 2003) was removed from the mar- ket in 2013 by GSK. The demise of Bexxar was not because of lack of efficacy or toxicity, but because of unforeseen mar- ket pressure and financial decisions (http://www.xconomy. com/national/2013/08/26/why-good-drugs-sometimes-fail-in- the-market-the-bexxar-story/). Both will be briefly discussed as they represent common RIT paradigms. Zevalin has U.S. FDA approval for relapsed or refractory follicular NHL and as a frontline adjuvant agent for follicular NHL achieving a complete response (CR) or partial response (PR) to induc- tion chemotherapy. Bexxar had U.S. FDA approval for the relapse or refractory setting as well as transformed NHL. Both are murine IgG mAbs that target the CD20 surface antigen on follicular NHL. 42 90 Y ibritumomab tiuxetan utilizes 90 Y, a

delivery of a large, macromolecule-targeting construct (pro- longed circulation time) from the delivery of a much smaller cytotoxic radioconjugate (more rapid circulation time). Two main approaches have been employed: a bispecific monoclo- nal antibody (bsmAb) system and a streptavidin–biotin sys- tem. In the bsmAb system (Fig. 29.5), a portion of the antibody has affinity for the tumor (antitumor), and another portion has affinity for the radionuclide carrier ligand or hapten–pep- tide (antihapten). Initially (step 1), a large “saturation” dose of the unlabeled bsmAb is administered, and the antibody local- izes in the tumor over several days. Occasionally, a clearing step is used to facilitate the clearance of the bsmAb from the circulation. Subsequently (step 2), a radionuclide conjugated to a hapten–peptide is administered that has high affinity for the antihapten portion of the bsmAb. This step results in a rapid distribution of the radionuclide in the tumor owing to the high affinity of the hapten–peptide for the bsmAb. Because the hapten–peptide has a small molecular weight, it will clear rapidly from the body and result in a low–bone marrow expo- sure to radiation. In the streptavidin–biotin system, strepta- vidin is conjugated to the initial pretargeting macromolecule, and biotin is conjugated to the radionuclide. Streptavidin and biotin have a very high affinity for each other (10 15 M − 1 ).When either system is used, the tumor/blood ratios of the target- ing agent are significantly increased, but there may be some advantages to the bsmAB system. 37,38 CONJUGATED THERAPY The current state of RIT continues to improve. The basic premise has been the delivery of targeted cytotoxic radio- therapy that is low dose, LDR, sparsely ionizing and delivered in a single fraction. Initially, this reality seemed to be a natu- ral “fit” for hematologic malignancies that were sensitive to most types of radiation. However, for RIT to ultimately impact significantly upon the world of oncology, it is clear that cur- rent approaches need to be modified so that it can be applied to carcinomas. Zevalin and Bexxar were FDA-approved in 2002 and 2003, respectively, as RIT drugs to treat relapse follicular NHL for an end point of progression-free survival. Subsequently, Zevalin received an approved frontline indica- tion (FIT trial). Bexxar was withdrawn from production by GlaxoSmithKline (GSK) in 2013. Currently, Zevalin is greatly underutilized because it has not shown a survival benefit for

658

S E C T I O N I I  Techniques, Modalities, and Modifiers in Radiation Oncology

TABLE 29.5  FDA-APPROVED AND CURRENT PHASE III RIT DRUGS Drug Name Target

Targeting Construct

Radionuclide Disease

Pipeline

Company/Sponsor

90 Y 131 I

NHL (low-grade follicular) NHL (low grade follicular) Midgut neuroendocrine tumors

FDA 2002 FDA 2003

Biogen Idec

Zevalin (Ibritumomab Tiuxetan) IgG1 (Murine)

CD20 CD20

Bexxar a (Tositumomab) Lutathera ( 177 Lu–DOTATATE)

IgG2a (Murine)

GlaxoSmithKline

Octreotate peptide Somatostatin receptor

177 Lu

Phase III completed filing

Advanced Accelerator Applications

Zevalin

IgG1 (murine)

CD20

90 Y

Relapse DLBCL/ASC transplant

Phase III completed Sheba Medical Center

131I–chTNT

IgG1 chimeric murine

Histone H1/DNA

131 I

Non–small cell lung cancer (postoperative)

Phase III complete Guangxi Zhuang

Autonomous Region Self Financing Project, China

Licartin (131I–Metuximab)

F(ab’) 2

CD147

131 I

HCC with RFA

Phase III complete Biotechnology Chengdu, China

Zevalin

IgG1 (Murine)

CD20

90 Y

Relapsed follicular NHL/ consolidation Untreated follicular NHL

Phase III recruiting (NCT01827605) Phase III recruiting (NCT02320292) Phase III recruiting (NCT02665065) Phase III recruiting (NCT02465112) Phase III recruiting (NCT03049198)

Fondazione Italiana Linfomi

Zevalin

IgG1 (Murine)

CD20

90 Y

Mayo Clinic

Iomab-B (BC8-I-131)

IgG1 (Murine)

CD45

131 I

AML

Actinium Pharmaceuticals

111 In–Pentetreotide

Octreotate peptide Somatostatin receptor Octreotate peptide Somatostatin receptor

111 In

Resected GI neuroendocrine tumors

GERCOR

177Lu–Edotreotide

177 Lu

GEP-NET

ITM Solucin GmbH

a Withdrawn from production by GlaxoSmithKline in 2013.

pure β -particle emitter with a physical half-life of 2.7 days. The β -particle has an energy of 2.3 MeV and a maximum tis- sue penetration of approximately 12.0 mm (R 90 = 5.2 mm). Tiuxetan is a DTPA-type chelate that attaches 90 Y to the mAb, ibritumomab. Because there is no gamma emission in the spectrum of this isotope, it is not visualized by gamma cam- era scans. As a result, a biodistribution assessment cannot be performed. Therefore, a surrogate imaging radionuclide that emits gamma radiation 111 In is required. In contrast, 131 I tosi- tumomab is a mixed β - / γ -emitter. The gamma spikes at 364 keV, and the beta emission has energy of 0.6 MeV. The maxi- mum range in tissue of the β -particle is 2.3 mm (R 90 = 0.7 mm). This agent can be imaged on gamma camera to calculate total body clearance. For both agents, the treatment is delivered over 1 to 2 weeks. On day 1, both protocols deliver an infusion of nonradioactive (cold) anti-CD20 antibody (Zevalin employs rituximab; Bexxar employed tositumomab) designed to saturate the CD20 antigen sink (depletion of peripheral B cells and the binding of nonspecific sites in the liver and spleen) and provide antibody mass, which improves biodis- tribution and tumor targeting. 43,44 The administered activ- ity for Zevalin is based on weight (0.4 mCi/kg for a platelet count ≥150,000; 0.3 mCi/kg for a platelet count of 100,000 to 149,000; maximum of 32 mCi). A single gamma scan ( 111 In ibritumomab tiuxetan) is used to confirm a normal biodistri- bution on days 3 to 4. A review of the Zevalin imaging registry reveals that only 0.6% of scans exhibited an altered biodistri- bution. Subsequently, the delivery of Zevalin has been simpli- fied. Based upon the analysis of five trials, which revealed an altered biodistribution scan in only about 1% of patients, the FDA removed the requirement of the biodistribution scan.The administered activity for Bexxar was based on a calculated total body clearance (three scans over 1 week) that delivers a total-body (red bone marrow) dose of 75 cGy. This calcula- tion is reduced to a total-body dose of 65 cGy for a platelet count <150,000. Eligible patients for Zevalin are also required to have an absolute neutrophil count (ANC) ≥1,500 and a bone marrow biopsy that reveals <25% lymphoma involvement. Relapse Setting Multiple prospective clinical trials have provided evidence for the use of RIT for treating relapsed or refractory follicu- lar NHL. Together, they represent >200 patients treated with

either Zevalin or Bexxar. Both agents appear to suggest an overall response rate (ORR) of 60% to 80% and a CR rate of 20% to 50%. Zevalin trials have been extensively reviewed. 45,46 A phase III study comparing Zevalin versus rituximab for patients with relapsed or refractory low-grade follicular B-cell NHL or transformed NHL was performed. 47 Patients were ran- domized to either a single intravenous (IV) dose of Zevalin 0.4 mCi/kg ( n = 73) or IV rituximab 375 mg/m 2 weekly for four doses ( n = 70). The RIT group was pretreated with two ritux- imab doses (250 mg/m 2 ) to improve biodistribution and tumor targeting. After the first rituximab dose on day 1, 111 In ibritu- momab tiuxetan was administered to assess biodistribution and to aide in dosimetry. No patients received the therapeutic dose of 90 Y ibritumomab tiuxetan (Zevalin) if >20 or 3 Gy was calculated to any nontumor organ or the red marrow, respec- tively. Zevalin was administered after the second rituximab dose approximately 1 week (days 7 to 9) after the first dose of rituximab and ( 111 )In ibritumomab tiuxetan. The administered activity of Zevalin was capped at 32 mCi. Patients in both arms of the study received two prior chemotherapy regimens. The ORR was 80% for Zevalin and 56% for rituximab ( P = .002). The CR rates were 30% and 16% ( P = .04), respectively, in the Zevalin and rituximab group. Durable responses ≥6 months were 64% versus 47% ( P = .030) for Zevalin versus rituximab. The conclusion of the study was that RIT with Zevalin was well tolerated and resulted in statistically significant and clin- ically significant higher ORRs and CRs than rituximab alone. Frontline Therapy Considering the concerns about RIT for treating large bulky tumors (tumor penetration, overall required dose, nonuni- form dose distributions), bringing RIT into a frontline thera- peutic setting after induction chemotherapy and maximum cytoreduction would be the next logical direction. A phase III first-line indolent trial (FIT) of consolidation with Zevalin compared to no additional therapy after first remission was reported for follicular B-cell NHL. 48 Patients with CD20+ stage III/IV follicular B-cell NHL who achieved a PR or CR to induc- tion chemotherapy were randomized to Zevalin ( n = 208) or to the control arm, representing no further treatment ( n = 206). Prior to chemotherapy, patients had documented <25% bone marrow involvement. After induction chemotherapy, blood counts had to recover such that the ANC was ≥1.5, plate- lets were ≥150,000, and hemoglobin was ≥9. Patients in the