Purpose/Objective(s) Prior studies have demonstrated that radiolabeled anti-carcinoembryonic antigen (CEA) monoclonal antibody (M5A) can target CEA-expressing cancers. This ongoing pilot study (NCT05245786) evaluates the ability of Cu-64-DOTA-M5A PET imaging to define disease extent before and after neoadjuvant radiotherapy in patients with locally advanced rectal cancer. Materials/Methods Patients with biopsy proven cT3-4, N0 or N+ rectal cancer who were scheduled for neoadjuvant chemoradiation prior to TME or total neoadjuvant therapy underwent PET imaging with Cu-64-M5A within 4 weeks prior to start of neoadjuvant RT and 6-8 weeks after end of all neoadjuvant therapy and prior to planned surgery. Cu-64-M5A PET scans were read by the same nuclear medicine physician. A regional analysis (primary, pelvic nodes, and extra-pelvic) later compared CEA scans to standard of care (SOC) CT and MRI imaging, FDG PET if available, biopsies and surgical findings. Results Nine patients have been imaged to date with Cu-64-M5A PET (ages 46-70 years old). CEA PET imaging prior to RT identified the primary tumor in all 9 patients. For disease involving pelvic nodes, CEA PET imaging was in concordance with SOC imaging in 6 of 9 patients. In 5 patients CEA scans identified suspicious lesions outside the pelvis, which in 2 patients were confirmed by SOC imaging and biopsy (lung metastases; and a single liver metastasis), while in 3 patients PET avid nodes detected (para-aortic, left supraclavicular, mediastinal, porta hepatis, and portacaval) were outside the regions imaged as part of SOC. 4 of 9 patients to date have also been imaged post RT and prior to surgery. 2 patients had radiologic complete response (CR) on CEA scans which correlated with pathologic CR at surgery and with clinical CR on follow-up sigmoidoscopy. Two patients had positive scans; 1 patient at the primary site confirmed as residual disease at surgery; and 1 patient at the primary site, in mediastinal, para-aortic, and pelvic lymph nodes, and a new focus near the hepatic surface with biopsy confirmation of disease at the primary site and liver. Conclusion Results demonstrate that Cu-64-M5A anti-CEA M5A PET imaging in patients with advanced rectal cancer shows promise in identifying known sites of disease at the primary site and regional lymph nodes, as well as identifying extra pelvic sites of disease. In addition, it shows promise for assessing response to neoadjuvant therapy. Cu-64-M5A imaging warrants further evaluation as a way to complement SOC imaging at initial staging, to select patients for local regional therapy and to select patients for non-operative management after neoadjuvant therapy. Given its targeting capabilities, M5A is also being evaluated radiolabeled with Ac-225 as a therapeutic radiopharmaceutical (NCT05204147) and as an anti-CEA-IL2 immunotherapy fusion agent (NCT06130826) in phase 1 trials.

Dose escalation by adding Y-anti-CD25 RIT at 0.3 mCi/kg to 12 Gy TMLI was safe, including in older patients, with no dose-limiting toxicities, mean critical organ doses lower than conventional myeloablative TBI, and encouraging response rates. The toxicity profile and dose estimates at 0.3 mCi/kg predict that the planned higher dose levels will also be feasible with acceptable toxicities. RIT and TMLI are complementary and when combined address the limitations of each modality. Combining these targeted therapies may be a superior strategy to intensify dose to leukemia compared to dose escalation of either modality alone.

CEACAM1 (carcinoembryonic antigen-related cell adhesion molecule 1) is a type I transmembrane glycoprotein expressed in epithelial cells with three or four extracellular domains (ECDs) and either long or short cytoplasmic domain isoforms. We have previously shown that the four extracellular domains, short cytoplasmic domain isoform, CEACAM1-4S, plays an essential role in lumen formation in an in vitro model of mammary morphogenesis. In this study, we transfected MCF-7 cells with either the long or short cytoplasmic domain isoforms of CEACAM1, and grew the cells in humanized mammary mouse fat pads in NOD/SCID mice. In this in vivo model, only the long cytoplasmic domain isoform, CEACAM1-4L, formed glands with lumen. On the basis of other studies that revealed phosphorylation of key Thr and Ser residues in the short cytoplasmic domain, we introduced phosphorylation mimic (for example, Thr or Ser to Asp) or null (Thr or Ser to Ala) mutations into the cytoplasmic domain of CEACAM1-4S and tested them in the in vivo model. Mutation of either Thr or Ser to Asp or the double mutant Thr+Ser to Asp, but not the null mutants, induced gland formation with a central lumen-containing apoptotic cells. Moreover, the phosphorylation mimic mutants of CEACAM1-4S induced downregulation of β1-integrin, overexpression of β2-integrin, inhibited phosphorylation of focal adhesion kinase (pTyr-397) and resulted in myofibroblast differentiation as characterized by expression of vimentin, α-smooth muscle actin and β2-integrin, as well as the production of abundant extracellular matrix.

Diabodies are single chain antibody fragments (scFvs) that spontaneously form bivalent dimers of molecular size 50-55000. Radiolabeled diabodies are almost ideal tumor targeting agents due to their high avidity (bivalent) binding to tumor antigens and small size (50-55000) that leads to improved tumor-to-blood ratio compared to intact antibodies (150000). However, due to their high retention and metabolism in the kidney, radioiodine is the current radiolabel of choice for diabodies since radioiodine is rapidly excreted from the kidney once metabolized. We have previously shown that 111In-DOTA-diabody gives higher tumor uptake in nude mouse xenografts than 125I-diabody, but has extremely high kidney retention since its 111In-labeled metabolites are retained by and only slowly excreted from the kidney. When a diabody is conjugated to a bifunctional PEG-3400 derivative followed by reaction with cysteinyl-DOTA, the resulting product has an apparent molecular size of 75000 and a Stokes radius of 35 angstroms on size exclusion chromatography, compared to a Stokes radius of 25 angstroms for intact diabody. When radiolabeled, the conjugate gives high yields of 111In-labeled product, retains high immunoreactivity, and gives improved biodistributions (30-40%ID/g, 12-48 h) compared to 111In-DOTA-diabody (12-13%ID/g, 6-12 h). We show that the improved biodistribution is due to an increase in Stokes radius caused by the linear PEG-3400 since conjugation of diabody with multiple (PEG)12 linkers followed by reaction with cysteinyl-DOTA does not reduce kidney accumulation. We also show that 111In-cysteinyl-DOTA-PEG3400-diabody gives excellent tumor images in the nude mouse xenograft model and that 125I-PEG3400-diabody gives equivalent images to 125I-minibody (molecular size, 80000), but improved tumor-to-liver ratios, suggesting that this imaging agent can be used to image liver metastases.

Most women with epithelial ovarian cancer are diagnosed with advanced disease. Despite surgery and initial tumor reduction by standard chemotherapy, the tumors frequently recur and the patients eventually die of their disease. New drugs that inhibit tyrosine kinase receptors (TKRs) are being investigated for treatment and this study was undertaken to determine the expression and mutational state for 3 TKRs (c-kit, platelet-derived growth factor receptor [PDGFR] α, and PDGFRβ) in ovarian cancer. Tissue arrays containing 84 epithelial ovarian tumors were studied by immunohistochemistry with antibodies specific for c-kit, PDGFRα, and PDGFRβ. Immunoreactivity was detected in 78% of the tumor to at least one TKR. PDGFRα was expressed in the largest percentage of ovarian tumors (58%) whereas 29% expressed PDGFRβ. Two commercial antibodies against c-kit were studied and 33% of the tumors stained with one but only 6% were interpreted as positive with the second antibody. Activation of TKRs may occur through mutations but, by sequence analysis, no mutations were detected in 6 ovarian tumors with elevated immunoreactivity for each of the TKRs (c-kit, PDGFRα, and PDGFRβ). Tyrosine kinase receptors could also be activated through autocrine or paracrine stimulation of receptor by its ligand. Of 43 (35%) tumors tested for both c-kit receptor and its ligand (stem cell factor), 15 expressed both proteins indicating the possibility that this autocrine stimulation feedback loop is a factor in the growth of some ovarian cancers. This study demonstrates that PDGFRα, PDGFRβ, and c-kit are expressed in a high percentage of epithelial ovarian cancers suggesting that tyrosine kinase inhibitors may be useful in the treatment of these tumors.

Antibody fragments are recognized as promising vehicles for delivery of imaging and therapeutic agents to tumor sites in vivo. The serum persistence of IgG1 and fragments with intact Fc region is controlled by the protective neonatal Fc receptor (FcRn) receptor. To modulate the half-life of engineered antibodies, we have mutated the Fc-FcRn binding site of chimeric anti-carcinoembryonic antigen (CEA) antibodies produced in a single-chain Fv-Fc format. The anti-CEA T84.66 single-chain Fv-Fc format wild-type and five mutants (I253A, H310A, H435Q, H435R, and H310A/H435Q, Kabat numbering system) expressed well in mammalian cell culture. After purification and characterization, effective in vitro antigen binding was shown by competition ELISA. Biodistribution studies in BALB/c mice using (125)I- and (131)I-labeled fragments revealed blood clearance rates from slowest to fastest as follows: wild-type > H435R > H435Q > I253A > H310A > H310A/H435Q. The terminal half-lives of the mutants ranged from 83.4 to 7.96 hours, whereas that of the wild-type was approximately 12 days. Additionally, (124)I-labeled wild-type, H435Q, I253A, H310A, and H310A/H435Q variants were evaluated in LS174T xenografted athymic mice by small animal positron emission tomography imaging, revealing localization to the CEA-positive xenografts. The slow clearing wild-type and H435Q constructs required longer to localize to the tumor and clear from the circulation. The I253A and H310A fragments showed intermediate behavior, whereas the H310A/H435Q variant quickly localized to the tumor site, rapidly cleared from the animal circulation and produced clear images. Thus, attenuating the Fc-FcRn interaction provides a way of controlling the antibody fragment serum half-life without compromising expression and tumor targeting.

Abstract We have recently described the in vivo properties of an iodinated anti-p185HER2 engineered antibody fragment [minibody (scFv-CH3)2; 80 kDa], made from the internalizing 10H8 monoclonal antibody. Although the 10H8 minibody showed excellent binding to the target in vitro, only modest tumor uptake [5.6 ± 1.7% injected dose per gram (ID/g) of tissue] was achieved in nude mice bearing MCF7/HER2 breast cancer tumors. Here, in an attempt to improve targeting, the 10H8 minibody was conjugated to 1,4,7,10-tetraazacyclododecane-N, N′, N′′, N′′′-tetraacetic acid (DOTA), radiometal labeled, and evaluated in vivo. The tumor uptake of 111In-DOTA 10H8 minibody was 5.7 ± 0.1% ID/g, similar to the radioiodinated 10H8 minibody. However, in addition to the expected liver clearance, the kidneys had unexpectedly high activity (34.0 ± 4.0% ID/g). A minibody derived from a second anti-p185HER2 antibody (trastuzumab; hu4D5v8) was also made. Tumor uptakes, evaluated by quantitative microPET using 64Cu-DOTA hu4D5v8 minibody, were 4.2 ± 0.5% ID/g. Furthermore, in non-tumor-bearing mice, 111In-DOTA hu4D5v8 minibody exhibited similar elevated uptake in the kidneys (28.4 ± 6.5% ID/g). Immunohistochemical staining of kidneys from non-tumor-bearing mice showed strong specific staining of the proximal tubules, and Western blot analysis of kidney lysate confirmed the presence of cross-reactive antigen. To further improve tumor uptake and normal tissue distribution, a larger hu4D5v8 fragment [(scFv-CH2-CH3)2; 105 kDa] was made, engineered to exhibit rapid clearance kinetics. This fragment, when evaluated by microPET, exhibited improved tumor targeting (12.2 ± 2.4% ID/g) and reduced kidney uptake (13.1 ± 1.5% ID/g). Thus, by manipulating the size and format of anti-p185HER2 antibody fragments, the kidney activity was reduced and high or low expression of p185HER2 in xenografts could be distinguished by microPET imaging.

An engineered antibody fragment (minibody; scFv-C(H)3gamma(1) dimer, M(r) 80 000) specific for carcinoembryonic antigen (CEA) has previously demonstrated excellent tumor targeting coupled with rapid clearance in vivo. In this study, variable (V) genes from the anti- p185(HER-2) 10H8 antibody were similarly assembled and expressed. Four constructs were made: first, the V genes were assembled in both orientations (V(L)-linker-V(H) and V(H)-linker-V(L)) as single chain Fvs (scFvs). Then each scFv was fused to the human IgG1 C(H)3 domain, either by a two amino acid linker (ValGlu) that resulted in a non-covalent, hingeless minibody, or by IgG1 hinge and a GlySer linker peptide to produce a covalent, hinge-minibody. The constructs, expressed in NS0 mouse myeloma cells at levels of 20-60 mg/l, demonstrated binding to the human p185(HER-2) overexpressing breast cancer cell line, MCF7/HER2. Binding affinities (K(D) approximately 2-4 nM) were equivalent to that for the parental 10H8 mAb (K(D) approximately 1.6 nM). Radioiodinated 10H8 hinge-minibody was evaluated in athymic mice, bearing MCF7/HER2 xenografts. Maximum tumor uptake was 5.6 (+/-1.65)% injected dose/g (ID/g) at 12 h, which was lower than that of the anti-CEA minibody, whereas the blood clearance (beta-phase, 5.62 h) was similar. Thus, minibodies with different specificities display similar pharmacokinetics, while tumor uptake may vary depending on the antigen-antibody system.

Purpose: The chimeric T84.66 (cT84.66) minibody is a novel engineered antibody construct (VL-linker-VH-CH3; 80 kDa) that demonstrates bivalent and high affinity (4 × 1010 m−1) binding to carcinoembryonic antigen (CEA). The variable regions (VL and VH) assemble to form the antigen-combining sites, and the protein forms dimers through self-association of the CH3 domains. In animal models, the minibody demonstrated high tumor uptake, approaching that of some intact antibodies, substantially faster clearance than intact chimeric T84.66, and superior tumor-to-blood ratios compared with the cT84.66 F(ab′)2 fragment, making it attractive for further evaluation as an imaging and therapy agent. The purpose of this pilot clinical study was to determine whether 123I-cT84.66 minibody demonstrated tumor targeting and was well tolerated as well as to begin to evaluate its biodistribution, pharmacokinetics, and immunogenicity in patients with colorectal cancer. Experimental Design: Ten patients with biopsy-proven colorectal cancer (6 newly diagnosed, 1 pelvic recurrence, 3 limited metastatic disease) were entered on this study. Each received 5–10 mCi (1 mg) of 123I-labeled minibody i.v. followed by serial nuclear scans and blood and urine sampling over the next 48–72 h. Surgery was performed immediately after the last nuclear scan. Results: Tumor imaging was observed with 123I-labeled minibody in seven of the eight patients who did not receive neoadjuvant therapy before surgery. Two patients received neoadjuvant radiation and chemotherapy, which significantly reduced tumor size before surgery and minibody infusion. At surgery, no tumor was detected in one patient and only a 2-mm focus was seen in the second patient. 123I-labeled minibody tumor targeting was not seen in either of these pretreated patients. Mean serum residence time of the minibody was 29.8 h (range, 10.9–65.4 h). No drug-related adverse reactions were observed. All 10 patients were evaluated for immune responses to the minibody, with no significant responses observed. Conclusion: This pilot study represents one of the first clinical efforts to evaluate an engineered intermediate-molecular-mass radiolabeled antibody construct directed against CEA. cT84.66 minibody demonstrates tumor targeting to colorectal cancer and a faster clearance in comparison with intact antibodies, making it appropriate for further evaluation as an imaging and therapy agent. The mean residence time of the minibody in patients is longer than predicted from murine models. We therefore plan to further evaluate its biodistribution and pharmacokinetic properties with minibody labeled with a longer-lived radionuclide, such as 111In.

The cell adhesion molecule, carcinoembryonic Ag-related cellular adhesion molecule 1, shown by others to both activate and inhibit T cell proliferation, exhibits a reciprocal relationship to IL-2R expression over the time course of activation of PBMCs, and upon Ab ligation, inhibits both the production of IL-2 and cell proliferation. Carcinoembryonic Ag-related cellular adhesion molecule 1 associates with CD3 and is found in lipid rafts of PBMCs, is phosphorylated on the immunoreceptor tyrosine-based inhibitory motifs (ITIMs) of the -4L isoform, and associates with Src homology protein-1, providing an explanation for its inhibitory activity. When the ITIM-containing -4L and non-ITIM-containing -4S isoforms are transfected into Jurkat cells that produce, but do not depend on IL-2 for growth, both IL-2 production and cell proliferation are differentially inhibited, demonstrating that the two isoforms signal via different pathways. When the two isoforms are transfected into Kit-225 cells that depend on IL-2 for growth, IL-2Rbeta and gamma, but not alpha subunits are down-regulated, and the -4L, but not the -4S isoform inhibits cell proliferation by 6-fold in an IL-2 dose-response study.

Carcinoembryonic antigen cell adhesion molecule 1 (CEACAM1) is a type 1 transmembrane, homotypic cell adhesion protein expressed on epithelial and hematopoietic cells. CEACAM1 has four major isoforms with three or four immunoglobulin (Ig)-like ectodomains and either long or short cytoplasmic domains. In a 3D model of breast epithelial cell morphogenesis, CEACAM1 plays an essential role in lumen formation [J. Cell Sci. 112 (1999) 4193]. Two soluble ectodomain isoforms of CEACAM1 expressed in myeloma cells were immunologically active and highly glycosylated. The molecular weights of the 3-ecto- and 4-ectodomain isoforms were 90 and 110 kDa, respectively, and monomers by sedimentation equilibrium centrifugation. Both isoforms were prolate ellipsoids with axial ratios of 6 for the 3-ecto- and 8 for 4-ectodomain isoforms, respectively, by size exclusion chromatography and analytical ultracentrifugation. Both isoforms caused a significant reduction in lumen formation when tested in the 3D model culture system.

An engineered anti-carcinoembryonic antigen (CEA) diabody (scFv dimer, 55 kDa) was previously constructed from the murine anti-CEA T84.66 antibody. Tumor targeting, imaging and biodistribution studies in nude mice bearing LS174T xenografts with radiolabeled anti-CEA diabody demonstrated rapid tumor uptake and fast blood clearance, which are favorable properties for an imaging agent. Current radiolabeling approaches result in random modification of the protein surface, which may impair immunoreactivity especially for smaller antibody fragments. Site-specific conjugation approaches can direct modifications to reactive groups located away from the binding site. Here, cysteine residues were introduced into the anti-CEA diabody at three different locations, to provide specific thiol groups for chemical modification. One version (with a C-terminal Gly-Gly-Cys) existed exclusively as a disulfide-bonded dimer. This cysteine-modified diabody (Cys-diabody) retained high binding to CEA and demonstrated tumor targeting and biodistribution properties identical to the non-covalent diabody. Furthermore, following reduction of the disulfide bond, the Cys-diabody could be chemically modified using a thiol-specific bifunctional chelating agent, for radiometal labeling. Thus, the Cys-diabody provides a covalently linked alternative to conventional diabodies, which can be reduced and modified site-specifically. This format will provide a versatile platform for targeting a variety of agents to CEA-positive tumors.

Chimeric T84.66 (cT84.66) is a monoclonal antibody (mAb) of high specificity and affinity for the tumor-associated carcinoembryonic antigen (CEA). Radiolabeled cT84.66 has demonstrated utility in the clinic as a reagent for the radioimmunoscintigraphy and radioimmunotherapy of CEA-positive colorectal and breast malignancies. To extend the therapeutic efficacy of T84.66, humanization by complementary determining region (CDR) grafting was employed. CDR grafting is a well-established technique, though often a series of framework back-mutations is required to restore high affinity. Recently, the crystal structure of the T84.66 diabody (scFv dimer) derived from the murine T84.66 mAb was determined, facilitating the humanization process by the availability of crystal structure data for both the graft donor and graft acceptor. A search of the Protein Data Bank revealed close structural similarity (r.m.s.d. of 1.07 A) between the Fv of T84.66 and the Fv of 4D5v8, a humanized anti-p185HER2 antibody marketed as Herceptin (Trastuzumab). This resulted in two humanized versions of the T84.66 M5A and M5B mAbs that differed only in the number of murine residues present in the C-terminal half of CDR-H2. Biochemical analysis and animal biodistribution studies were conducted to evaluate the humanized mAbs. The M5A, M5B and cT84.66 mAbs showed sub-nanomolar affinity for CEA and as radiolabeled mAbs exhibited specific tumor localization in tumor bearing mice. The T84.66 M5A mAb was selected for clinical development due to a slightly higher tumor uptake and a larger content of human residues, and was renamed hT84.66. A limited-scale production and animal imaging study have demonstrated hT84.66's ability to support clinical trials. Planned clinical trials will determine the effective utilization of this structure-based approach in the development of a promising new therapeutic.

UNLABELLED Prolonged clearance kinetics have hampered the development of intact antibodies as imaging agents, despite their ability to effectively deliver radionuclides to tumor targets in vivo. Genetically engineered antibody fragments display rapid, high-level tumor uptake coupled with rapid clearance from the circulation in the athymic mouse/LS174T xenograft model. The anticarcinoembryonic antigen (CEA) T84.66 minibody (single-chain Fv fragment [scFv]-C(H)3 dimer, 80 kDa) and T84.66 diabody (noncovalent dimer of scFv, 55 kDa) exhibit pharmacokinetics favorable for radioimmunoimaging. The present work evaluated the minibody or diabody labeled with (124)I, for imaging tumor-bearing mice using a high-resolution small-animal PET system. METHODS Labeling was conducted with 0.2-0.3 mg of protein and 65-98 MBq (1.7-2.6 mCi) of (124)I using an iodination reagent. Radiolabeling efficiencies ranged from 33% to 88%, and immunoreactivity was 42% (diabody) or >90% (minibody). In vivo distribution was evaluated in athymic mice bearing paired LS174T human colon carcinoma (CEA-positive) and C6 rat glioma (CEA-negative) xenografts. Mice were injected via the tail vein with 1.9-3.1 MBq (53-85 microCi) of (124)I-minibody or with 3.1 MBq (85 microCi) of (124)I-diabody and imaged at 4 and 18 h by PET. Some mice were also imaged using (18)F-FDG 2 d before imaging with (124)I-minibody. RESULTS PET images using (124)I-labeled minibody or diabody showed specific localization to the CEA-positive xenografts and relatively low activity elsewhere in the mice, particularly by 18 h. Target-to-background ratios for the LS174T tumors versus soft tissues using (124)I-minibody were 3.05 at 4 h and 11.03 at 18 h. Similar values were obtained for the (124)I-diabody (3.95 at 4 h and 10.93 at 18 h). These results were confirmed by direct counting of tissues after the final imaging. Marked reduction of normal tissue activity, especially in the abdominal region, resulted in high-contrast images at 18 h for the (124)I-anti-CEA diabody. CEA-positive tumors as small as 11 mg (<3 mm in diameter) could be imaged, and (124)I-anti-CEA minibodies, compared with (18)F-FDG, demonstrated highly specific localization. CONCLUSION (124)I labeling of engineered antibody fragments provides a promising new class of tumor-specific probes for PET imaging of tumors and metastases.

The carboxyl terminal octapeptide of cholecystokinin (CCK-8) has been hypothesized to account for the bioactivity of all the molecular forms of cholecystokinin. However, the physiological relevance of CCK-58 has not been rigorously examined because of the lack of sufficient amounts of the peptide and concerns about inactivation of natural peptides during their purification. Therefore, canine-sulfated CCK-58 was synthesized and conditions determined for its unblocking and purification that preserved the sulfated tyrosine. Synthetic CCK-58 was indistinguishable from natural CCK-58 by amino acid analysis and by mass spectrometry. Synthetic CCK-58 and CCK-8 have different patterns of pancreatic stimulation: both caused a dose-related increase in amylase release, while only CCK-58 stimulated bile-pancreatic output volume. Thus, CCK-58 and CCK-8 are biased agonists at the CCK-A receptor (they have distinct patterns of action mediated by the same receptor). Previous work has demonstrated that the identical carboxyl termini of CCK-8 and CCK-58 have different solution conformations. Taken together, the physiological and structural results support the hypothesis that different carboxyl terminal conformations of CCK-58 and CCK-8 alter the expression of their biological activity.

Purpose: Targeted systemic radiation therapy using radiolabeled antibodies results in tumor doses sufficient to produce significant objective responses in the radiosensitive hematological malignancies. Although comparable doses to tumor are achieved with radioimmunotherapy (RIT) in solid tumors, results have been modest primarily because of their relative lack of radiosensitivity. For solid tumors, as with external beam radiotherapy, RIT should have a more important clinical role if combined with other systemic, potentially radiation-enhancing chemotherapy agents and if used as consolidative therapy in the minimal tumor burden setting. The primary objective of this trial was to evaluate the feasibility and toxicities of systemic Y-chimeric T84.66 (cT84.66) anti-carcinoembryonic antigen RIT in combination with continuous infusion 5-fluorouracil (5-FU). Experimental Design: Patients with chemotherapyrefractory metastatic colorectal cancer were entered. The study was designed for each patient to receive Y-cT84.66 anti-carcinoembryonic antigen at 16.6 mCi/m as an i.v. bolus infusion combined with 5-FU delivered as a 5-day continuous infusion initiated 4 h before antibody infusion. Cohorts of patients were entered at 5-FU dose levels of 700, 800, 900, and 1000 mg/m/day. Upon reaching the highest planned dose level of 5-FU, a final cohort received YcT84.66 at 20.6 mCi/m and 5-FU at 1000 mg/m/day. For all patients, Ca-diethylenetriaminepentaacetic acid at 125 mg/m every 12 h was administered for the first 72 h after Y-cT84.66. Patients were eligible to receive up to three cycles of Y-cT84.66/5-FU every 6 weeks. Results: Twenty-one patients were treated on this study. All had been heavily pretreated with 19 having previously received 5-FU and 16 having failed two to four chemotherapy regimens. A maximum-tolerated dose of 16.6 mCi/m 90Y-cT84.66 combined with 1000 mg/m/day 5-FU was reached. These dose levels are comparable with maximumtolerated dose levels of each agent alone. Thirteen patients received one cycle and 8 patients two cycles of therapy. Hematopoietic toxicity was dose-limiting and reversible. RIT did not appear to increase nonhematopoietic toxicities associated with 5-FU. Two of 19 patients assayed developed a human anti-chimeric antibody immune response after the first cycle of therapy, which is significantly less than that observed in a previous trial evaluating Y-cT84.66 alone. No objective responses were observed. However, 11 patients with progressive disease entering the study demonstrated radiological stable disease of 3–8 months duration and 1 patient demonstrated a mixed response. Conclusions: Results from this trial are encouraging and demonstrate the feasibility and possible advantages of combining continuous infusion 5-FU with Y-cT84.66 RIT. The addition of 5-FU does not appear to significantly enhance hematological toxicities of the radiolabeled antibody. In addition, 5-FU reduces the development of human anti-chimeric antibody response, permitting multicycle therapy in a larger number of patients. Future efforts should continue to focus on integrating radiation therapy delivered by radiolabeled antibodies into established 5-FU regimens.