Pulmonary 18F-FDG uptake helps refine current risk stratification in idiopathic pulmonary fibrosis (IPF).

PURPOSE: There is a lack of prognostic biomarkers in idiopathic pulmonary fibrosis (IPF) patients. The objective of this study is to investigate the potential of 18F-FDG-PET/ CT to predict mortality in IPF. METHODS: A total of 113 IPF patients (93 males, 20 females, mean age ± SD: 70 ± 9 years) were prospectively recruited for 18F-FDG-PET/CT. The overall maximum pulmonary uptake of 18F-FDG (SUVmax), the minimum pulmonary uptake or background lung activity (SUVmin), and target-to-background (SUVmax/ SUVmin) ratio (TBR) were quantified using routine region-of-interest analysis. Kaplan-Meier analysis was used to identify associations of PET measurements with mortality. We also compared PET associations with IPF mortality with the established GAP (gender age and physiology) scoring system. Cox analysis assessed the independence of the significant PET measurement(s) from GAP score. We investigated synergisms between pulmonary 18F-FDG-PET measurements and GAP score for risk stratification in IPF patients. RESULTS: During a mean follow-up of 29 months, there were 54 deaths. The mean TBR ± SD was 5.6 ± 2.7. Mortality was associated with high pulmonary TBR (p = 0.009), low forced vital capacity (FVC; p = 0.001), low transfer factor (TLCO; p < 0.001), high GAP index (p = 0.003), and high GAP stage (p = 0.003). Stepwise forward-Wald-Cox analysis revealed that the pulmonary TBR was independent of GAP classification (p = 0.010). The median survival in IPF patients with a TBR < 4.9 was 71 months, whilst in those with TBR > 4.9 was 24 months. Combining PET data with GAP data ("PET modified GAP score") refined the ability to predict mortality. CONCLUSIONS: A high pulmonary TBR is independently associated with increased risk of mortality in IPF patients.

An epithelial biomarker signature for idiopathic pulmonary fibrosis: an analysis from the multicentre PROFILE cohort study.

BACKGROUND: Idiopathic pulmonary fibrosis (IPF) is a progressive, fatal disorder with a variable disease trajectory. The aim of this study was to assess potential biomarkers to predict outcomes for people with IPF. METHODS: PROFILE is a large prospective longitudinal cohort of treatment-naive patients with IPF. We adopted a two-stage discovery and validation design using patients from the PROFILE cohort. For the discovery analysis, we examined 106 patients and 50 age and sex matched healthy controls from Nottingham University Hospitals NHS Trust and the Royal Brompton Hospital. We did an unbiased, multiplex immunoassay assessment of 123 biomarkers. We further investigated promising novel markers by immunohistochemical assessment of IPF lung tissue. In the validation analysis, we examined samples from 206 people with IPF from among the remaining 212 patients recruited to PROFILE Central England. We used the samples to attempt to replicate the biomarkers identified from the discovery analysis by use of independent immunoassays for each biomarker. We investigated the predictive power of the selected biomarkers to identify individuals with IPF who were at risk of progression or death. The PROFILE studies are registered on ClinicalTrials.gov, numbers NCT01134822 (PROFILE Central England) and NCT01110694 (PROFILE Royal Brompton Hospital). FINDINGS: In the discovery analysis, we identified four serum biomarkers (surfactant protein D, matrix metalloproteinase 7, CA19-9, and CA-125) that were suitable for replication. Histological assessment of CA19-9 and CA-125 suggested that these proteins were markers of epithelial damage. Replication analysis showed that baseline values of surfactant protein D (46·6 ng/mL vs 34·6 ng/mL, p=0·0018) and CA19-9 (53·7 U/mL vs 22·2 U/mL; p<0·0001) were significantly higher in patients with progressive disease than in patients with stable disease, and rising concentrations of CA-125 over 3 months were associated with increased risk of mortality (HR 2·542, 95% CI 1·493-4·328, p=0·00059). INTERPRETATION: We have identified serum proteins secreted from metaplastic epithelium that can be used to predict disease progression and death in IPF. FUNDING: GlaxoSmithKline R&D and the UK Medical Research Council.

Functional capacity, health status, and inflammatory biomarker profile in a cohort of patients with chronic obstructive pulmonary disease.

PURPOSE: Prior research has shown a significant relationship between 6-minute walking distance (6MWD) and health-related quality of life (HRQOL) in patients with chronic obstructive pulmonary disease (COPD). However, few studies have examined this relationship above and below the 350-m threshold that prognosticates survival and whether serum biomarkers could provide insight into the causes of quality-of-life differences above and below this threshold. METHODS: Measures of lung function, 6MWD, and HRQOL were compared in patients with COPD. Differences in HRQOL domains and serum biomarkers were compared in patients whose 6MWD were > or < 350 m. RESULTS: In patients walking < 350 m, scores in the physical domains of the SF-36 and the St. George's Respiratory Questionnaire (SGRQ) were significantly different from scores of their counterparts with greater 6MWD. However, there was no association between any biomarkers and the physical domains of the SF-36 and the SGRQ. In patients walking < 350 m, only the IL-8 levels were associated with lower scores in SF-36 domains of emotional role, pain, vitality, and mental health (average r = -0.702; P = .01). In contrast, in patients walking > 350 m, surfactant protein D levels were associated with higher SF-36 scores in general pain, vitality, and social functioning (average r = 0.42; P = .04). CONCLUSIONS: In COPD, there is an association between 6MWD and the physical domains of the SF-36 and SGRQ in those patients walking < 350 m. The physical differences between patients walking < or > 350 m are not related to systemic inflammation. The association between interleukin 8 with nonphysical domains in patients with 6MWD < 350 m suggests that inflammation may play a larger role in the perceptive domain than previously recognized.

Longitudinal change in collagen degradation biomarkers in idiopathic pulmonary fibrosis: an analysis from the prospective, multicentre PROFILE study.

BACKGROUND: Idiopathic pulmonary fibrosis, a progressive and inevitably fatal disorder, has a highly variable clinical course. Biomarkers that reflect disease activity are urgently needed to inform patient management and for use as biomarkers of therapeutic response (theragnostic biomarkers) in clinical trials. We aimed to determine whether dynamic change in markers of extracellular matrix (ECM) turnover predicts progression of idiopathic pulmonary fibrosis as determined by change in forced vital capacity and death. METHODS: In this ongoing prospective, multicentre, observational cohort study (PROFILE), participants with idiopathic pulmonary fibrosis or idiopathic non-specific interstitial pneumonia diagnosed within the preceding 6 months were recruited from two coordinating centres (Nottingham, UK, and, Royal Brompton Hospital, London, UK). Serum samples were prospectively collected at baseline, 1 month, 3 months, and 6 months and were analysed for a panel of novel matrix metalloprotease (MMP)-degraded ECM proteins, by ELISA-based, neoepitope assay. 11 neoepitopes were tested in a discovery cohort of 55 patients to identify biomarkers of sufficient rigour for more detailed analyses. Eight were then further assessed in a validation cohort of 134 patients with 50 age-matched and sex-matched controls. Changes in biomarker concentrations were related to subsequent progression of idiopathic pulmonary fibrosis (defined as death or decline in forced vital capacity >10% at 12 months after study enrolment) using a repeated measures model. The PROFILE study is registered on ClinicalTrials.gov, numbers NCT01134822 and NCT01110694. FINDINGS: Of 214 eligible participants recruited between Sept 1, 2010, and March 31, 2012, 189 had a confirmed diagnosis of idiopathic pulmonary fibrosis and were included in subsequent analyses. In the discovery cohort, mean concentrations of seven neoepitopes (BGM, p=0·009; C1M, p=0·009; C3M, p=0·046; C6M, p=0·032; CRPM, p=0·008; ELM2, p=0·02; and VICM, p=0·0007) differed significantly between healthy controls and participants with idiopathic pulmonary fibrosis. Baseline concentrations of six neoepitopes (C1M, p=0·012; C3A, p=0·012; C3M, p=0·0005; C6M, p=0·0003; CRPM, p=0·021; and VICM, p=0·046) were significantly higher in patients with progressive idiopathic pulmonary fibrosis (n=32) than in those with stable disease (n=23). In the validation cohort, mean concentrations of C1M (p=0·001), C3M (p=0·044), C6M (p=0·003), and CRPM (p=0·024) at baseline were higher in patients with idiopathic pulmonary fibrosis than in healthy controls. When assessed longitudinally, concentrations of six neoepitopes (BGM, C1M, C3A, C3M, C6M, and CRPM) were significantly higher in patients with progressive idiopathic pulmonary fibrosis (n=71) than in patients with stable idiopathic pulmonary fibrosis (n=60) by 6 months. Baseline concentrations of two neoepitopes were associated with increased mortality (C1M: HR 1·62 [95% CI 1·14-2·31], p=0·0069; C3A: 1·91 [1·06-3·46], p=0·032). The rate of change between baseline and 3 months of six neoepitopes (BGM: HR 1·084 [95% CI 1·03-1·14], p=0·0019; C1M: 1·01 [1·003-1·017], p=0·0039; C3M: 1·106 [1·045-1·170], p=0·0005; C5M: 1·003 [1·001-1·005], p=0·0011; C6M: 1·042 [1·007-1·078], p=0·017; and CRPM: 1·38 [1·16-1·63], p=0·0002) was strongly predictive of overall survival, and the increased risk was proportional to the magnitude of change in neoepitope concentrations. The strongest association with 3-month rate of biomarker change was recorded for CRPM; greater than 0 ng/mL per month conferred a HR of 2·16 (95% CI 1·15-4·07), whereas a rate greater than 1 ng/mL per month resulted in an HR 4·08 (2·14-7·8), and a rate greater than 1·7 ng/mL per month was associated with an HR 6·61 (2·74-15·94). INTERPRETATION: Concentrations of protein fragments generated by MMP activity are increased in the serum of individuals with idiopathic pulmonary fibrosis compared with healthy controls. Increased neoepitope concentrations were associated with disease progression, and the rate of this increase predicted survival. Serial measurements of neoepitopes have potential to be used as theragnostic biomarkers in clinical trials and to guide management of idiopathic pulmonary fibrosis. FUNDING: GlaxoSmithKline R&D and the Medical Research Council.

Exploristics.

Exploristics provides world-class analytics and specialized software for life sciences, with specific expertise in clinical trials, biomarkers, personalized medicine research, observational studies, medical devices, diagnostics, real-world data analytics and risk stratification. We combine robust processes with scientific principles to create streamlined, flexible solutions that ensure quality while enabling new innovative approaches. The company provides comprehensive statistical support, offering researchers high quality services and software products in statistical modeling, data analysis, visualization and study design. Exploristics has just launched KERUS™, a novel in silico simulation platform for the prospective optimization of research studies involving complex, correlated data. This platform has a user-friendly interface, enabling researchers from a range of backgrounds to improve the efficiency of biomarker discovery and development.

A statistician's perspective on biomarkers in drug development.

Biomarkers play an increasingly important role in many aspects of pharmaceutical discovery and development, including personalized medicine and the assessment of safety data, with heavy reliance being placed on their delivery. Statisticians have a fundamental role to play in ensuring that biomarkers and the data they generate are used appropriately and to address relevant objectives such as the estimation of biological effects or the forecast of outcomes so that claims of predictivity or surrogacy are only made based upon sound scientific arguments. This includes ensuring that studies are designed to answer specific and pertinent questions, that the analyses performed account for all levels and sources of variability and that the conclusions drawn are robust in the presence of multiplicity and confounding factors, especially as many biomarkers are multidimensional or may be an indirect measure of the clinical outcome. In all of these areas, as in any area of drug development, statistical best practice incorporating both scientific rigor and a practical understanding of the situation should be followed. This article is intended as an introduction for statisticians embarking upon biomarker-based work and discusses these issues from a practising statistician's perspective with reference to examples.

Quantifying factors for the success of stratified medicine.

Co-developing a drug with a diagnostic to create a stratified medicine - a therapy that is targeted to a specific patient population on the basis of a clinical characteristic such as a biomarker that predicts treatment response - presents challenges for product developers, regulators, payers and physicians. With the aim of developing a shared framework and tools for addressing these challenges, here we present an analysis using data from case studies in oncology and Alzheimer's disease, coupled with integrated computational modelling of clinical outcomes and developer economic value, to quantify the effects of decisions related to key issues such as the design of clinical trials. This illustrates how such analyses can aid the coordination of diagnostic and drug development, and the selection of optimal development and commercialization strategies. It also illustrates the impact of the interplay of these factors on the economic feasibility of stratified medicine, which has important implications for public policy makers.

Pharmacogenetics: practices and opportunities for study design and data analysis.

Pharmacogenetics (PGx) is increasingly used as a way to target treatment to patients who are most likely to benefit. To date, PGx has shown clinical significance across a few applications but widespread use has been limited by the need for further technical, methodological and practical advances and for educating clinical researchers on the value of PGx. Here, I describe the current scope of PGx research, including recent contributions to prospective study design. A case study is included to demonstrate the limitations of current practice and to describe some practical steps for improving the chances of identifying genetic effects. Additionally, I describe some opportunities for the integration and application of disparate data sources in exploratory PGx research.

Localization of radiolabeled anti-CEA antibody in subcutaneous and intrahepatic colorectal xenografts: influence of tumor size and location within host organ on antibody uptake.

INTRODUCTION: Radioimmunotherapy (RIT) has been shown to be more effective against solid tumor micrometastases, possibly due to an inverse relationship between tumor size and radiolabeled antibody uptake. In this study, the accretion of radiolabeled antibody in intrahepatic micrometastases in an experimental model was investigated using quantitative digital autoradiography, enabling the analysis of antibody uptake in microscopic tumors. METHODS: Mice bearing subcutaneous or intrahepatic metastatic models of LS174T colorectal cancer were injected with radiolabeled anti-carcinoembryonic antigen antibody ([(125)I]A5B7). Tissues were taken to investigate distribution of radionuclide and tumor uptake. In a therapy study, mice bearing intrahepatic metastatic tumors were injected with [(131)I]A5B7. RESULTS: Subcutaneous tumors and large metastatic deposits had similar uptake (e.g., approximately 15%ID/g at 24 h). Small metastatic deposits had higher uptake (e.g., approximately 80%ID/g at 24 h) and prolonged retention at later time points. Small deposit uptake was significantly reduced by accompanying large deposits in the same liver. RIT resulted in increased survival time (untreated mean survival of 21.6+/-12.9 vs. treated mean survival of 39.1+/-30.8 days), but there was a large range of response within groups, presumably due to variation in pattern and extent of tumor as observed in the biodistribution study. Liver function tests and body weight did not change with tumor growth or therapy response, strongly supporting the use of in vivo imaging in metastatic tumor therapy studies. CONCLUSIONS: Radioimmunodetection and therapy might be greatly influenced by the size and distribution of intrahepatic tumor deposits.

Nonuniform absorbed dose distribution in the kidney: the influence of organ architecture.

The development of novel, systemically administered radionuclide therapies (such as radioimmunotherapy) relies on the ability to predict dose-limiting toxicity to normal tissue. Where the kidney is the principal route of excretion of the radionuclide preparation and/or breakdown of products, nephrotoxicity may be the dose-limiting factor. Until recently, conventional (MIRD) dosimetry assumed the distribution in the kidney to be uniform. A new MIRD phantom of the kidney models it as a set of uniform suborgans. In the work described here, the assumption of uniformity of distribution and of heterogeneity of dose rate (and, thus, absorbed dose) was tested in the mouse model. In this paper, we examine the nonuniformity of distribution and the subsequent dose rate for 4 antibody preparations (IgG (150 kD), F(ab)'(2) (100 kD), Fab (50 kD) and sFv (27 kD)) labeled with 4 radionuclides ((125)I, (131)I, (186)Re, and (90)Y) of interest in radioimmunotherapy (RIT). The kidney was considered as a whole and as two suborgans (cortex and medulla), and the nonuniformity of the dose-rate distribution was measured by a correlation of modeled dose-rate distribution with the dose-rate distribution obtained for an equivalent uniform radionuclide distribution. The heterogeneity of distribution, the inter- and intra-suborgan, was seen to increase as the molecular weight of the antibodies decreased. The assumption of uniform activity distribution for the whole kidney gives a poor estimation of the distribution of the dose rate. In the cortex, the longer-range emitters smooth out the effect of heterogeneous distribution and, in mice, an assumption of uniform cortex self-dose distribution may be sufficient for simple calculations. It is unclear how much this smoothing would be relevant in the human kidney.

Analysis of the regional uptake of radiolabeled deoxyglucose analogs in human tumor xenografts.

UNLABELLED: It has been shown in vitro that the cell uptake of (18)F-FDG, a tracer of glucose metabolism, increases under hypoxia. This is consistent with increased glycolytic metabolism. We have previously shown that in ischemic heart ex vivo the rates of uptake of (18)F-FDG and 2-(14)C-deoxy-D-glucose ((14)C-2DG) are both reduced. In this study, we investigated this effect in tumors by comparing the microdistribution of (18)F-FDG and (14)C-2DG in hypoxic and normoxic regions. METHODS: Mice (MF1) bearing LS174T human tumor xenografts were injected with premixed (18)F-FDG (100 MBq), (14)C-2DG (0.37 MBq), and pimonidazole hydrochloride (60 mg/kg). After 30, 60, and 120 min, tissues (n = 4) were taken and counted for whole-body biodistribution. Tumors were frozen, sectioned, and exposed to phosphor image plates to obtain a quantitative digital image of radionuclide distribution. Sections were then stained to reveal tumor pathophysiology: Hematoxylin and eosin staining demonstrated viable and necrotic regions, and immunohistochemical staining detected pimonidazole metabolism in hypoxic cells. The images of radionuclide microdistribution and histology were then coregistered and analyzed to assess radionuclide trapping throughout the tumor on a pixel-by-pixel basis. The Pearson correlation coefficients between the 2 radionuclides were calculated. The relative amounts of nuclide were then analyzed in viable and necrotic regions and in normoxic and hypoxic regions. RESULTS: Whole-body biodistributions for the 2 radiotracers were similar. A high Pearson correlation coefficient was obtained for the 2 radionuclides throughout the tumors (r = 0.85 +/- 0.10, P < 0.0001), indicating a highly similar microdistribution. When the tumors were divided into viable and necrotic regions, the ratio of mean counts per pixel was 1.96 (P < 0.0001), whereas for hypoxic versus normoxic regions it was 1.26 (P < 0.0001). There was no significant difference in selectivity for hypoxia between the 2 radiotracers (P = 0.86). CONCLUSION: The tumor microdistribution of deoxyglucose in viable, hypoxic, and necrotic regions show that there was little change in the microdistribution of deoxyglucose throughout this time course. This study extends previous in vitro work and confirms the selectivity of deoxyglucose for viable cells over necrotic regions and for hypoxic cells over normoxic regions in vivo.

The nonuniformity of antibody distribution in the kidney and its influence on dosimetry.

The therapeutic efficacy of radiolabeled antibody fragments can be limited by nephrotoxicity, particularly when the kidney is the major route of extraction from the circulation. Conventional dose estimates in kidney assume uniform dose deposition, but we have shown increased antibody localization in the cortex after glomerular filtration. The purpose of this study was to measure the radioactivity in cortex relative to medulla for a range of antibodies and to assess the validity of the assumption of uniformity of dose deposition in the whole kidney and in the cortex for these antibodies with a range of radionuclides. Storage phosphor plate technology (radioluminography) was used to acquire images of the distributions of a range of antibodies of various sizes, labeled with 125I, in kidney sections. This allowed the calculation of the antibody concentration in the cortex relative to the medulla. Beta-particle point dose kernels were then used to generate the dose-rate distributions from 14C, 131I, 186Re, 32P and 90Y. The correlation between the actual dose-rate distribution and the corresponding distribution calculated assuming uniform antibody distribution throughout the kidney was used to test the validity of estimating dose by assuming uniformity in the kidney and in the cortex. There was a strong inverse relationship between the ratio of the radioactivity in the cortex relative to that in the medulla and the antibody size. The nonuniformity of dose deposition was greatest with the smallest antibody fragments but became more uniform as the range of the emissions from the radionuclide increased. Furthermore, there was a strong correlation between the actual dose-rate distribution and the distribution when assuming a uniform source in the kidney for intact antibodies along with medium- to long-range radionuclides, but there was no correlation for small antibody fragments with any radioisotope or for short-range radionuclides with any antibody. However, when the cortex was separated from the whole kidney, the correlation between the actual dose-rate distribution and the assumed dose-rate distribution, if the source was uniform, increased significantly. During radioimmunotherapy, the extent of nonuniformity of dose deposition in the kidney depends on the properties of the antibody and radionuclide. For dosimetry estimates, the cortex should be taken as a separate source region when the radiopharmaceutical is small enough to be filtered by the glomerulus.

Synergy between vascular targeting agents and antibody-directed therapy.

PURPOSE: Tumor heterogeneity necessitates the use of combined therapies. We have shown that combining antibody-directed therapy with antivascular agents converts a subcurative to a curative treatment. The purpose of this study was to investigate, by radioluminographic and microscopic techniques, the regional effects of the two complementary therapies. METHODS AND MATERIALS: Nude mice bearing colorectal tumors were injected with 125I-labeled anti-carcinoembryonic antigen antibody, and images were obtained for antibody distribution and modeling studies using radioluminography. For therapy studies, the mice were given radioimmunotherapy alone (131I-A5B7 anti-carcinoembryonic antigen antibody), the antivascular agent combretastatin A-4 3-0-phosphate (200 mg/kg), or both. Extra mice were used to study the regional tumor effects of these therapies over time: relevant histochemical procedures were performed on tissue sections to obtain composite digital microscopic images of apoptosis, blood vessels, perfusion, hypoxia, and morphology. RESULTS: Antibody distribution, modeling, and immunohistochemistry showed how radioimmunotherapy (7.4 MBq/40 microg antibody) effectively treated the outer, well-oxygenated tumor region only. Combretastatin A-4 3-0-phosphate treated the more hypoxic center, and in doing so altered the relationship between tumor parameters. CONCLUSION: The combined complementary therapies produced cures by destroying tumor regions with different pathophysiologies. Relating these regional therapeutic effects to the relevant tumor parameters microscopically allows optimization of therapy and improved translation to clinical trials.

Dissociation of glucose tracer uptake and glucose transporter distribution in the regionally ischaemic isolated rat heart: application of a new autoradiographic technique.

Fluorine-18 fluoro-2-deoxyglucose ((18)FDG) and carbon-14 2-deoxyglucose ((14)C-2-DG) are both widely used tracers of myocardial glucose uptake and phosphorylation. We have recently shown, using positron emission tomography (PET) and nuclear magnetic resonance, that ischaemia-reperfusion (I-R) causes differential changes in their uptake. We describe here the novel application of an autoradiographic technique allowing the investigation of this phenomenon at high resolution, using tracer concentrations of both analogues in the dual-perfused isolated rat heart. We also investigate the importance of glucose transporter (GLUT 1 and GLUT 4) distribution in governing the observed phosphorylated analogue accumulation. Hearts ( n=5) were perfused with Krebs buffer for 40 min, made regionally zero-flow ischaemic for 40 min and reperfused for 60 min with Krebs containing tracer (18)FDG (200 MBq) and tracer (14)C-2-DG (0.37 MBq). Hearts were then frozen and five sections (10 micro m) were cut per heart, fixed and exposed on phosphor storage plates for 18 h (for (18)FDG) and then for a further 9 days (for (14)C-2-DG). Quantitative digital images of tracer accumulation were obtained using a phosphor plate reader. The protocol was repeated in a second group of hearts and GLUT 1 and GLUT 4 distribution analysed. Post-ischaemic accumulation of (18)FDG-6-P was inhibited by 38.2%+/-1.7% and (14)C-DG-6-P by 19.0%+/-2.2%, compared with control ( P<0.05). After placing seven "lines of interrogation" across each heart section and analysing the phosphorylated tracer accumulation along them, a transmural gradient of both tracers was observed; this was highest at the endocardium and lowest at the epicardium. GLUT 4 translocated to the sarcolemma in the ischaemic/reperfused region (from 24%+/-3% to 59%+/-5%), while there was no cellular redistribution of GLUT 1. We conclude that since decreased phosphorylated tracer accumulation occurs after ischaemia-reperfusion, despite greater externalisation of GLUT 4, hexokinase or the affinities of the GLUT transporters are changed under these conditions.

A model-based approach for the optimization of radioimmunotherapy through antibody design and radionuclide selection.

BACKGROUND: The effectiveness of radioimmunotherapy (RIT) is known to depend on at least six factors: total absorbed dose and pattern of delivery, radiosensitivity, rate of repair of sublethal damage, ongoing proliferation during treatment, tumor heterogeneity, and tumor size. The purpose of this study was to develop a mathematic model that would relate the absorbed dose and its pattern of delivery to tumor response by incorporating information on each factor. This model was used to optimize therapeutic efficacy in mice by matching the antibody and radionuclide characteristics while ensuring recoverable marrow toxicity. METHODS: Pharmacokinetic data were acquired in mice for a range of antibodies that varied in molecular weight, specificity, affinity, and avidity, and for a range of tumor sizes. This information was combined with the properties of iodine-131, rhenium-86, and yttrium-90 to determine the pattern of dose delivery. Tumor response was characterized in terms of radiosensitivity, rate of repair, and proliferation. Values for these parameters were obtained from in vitro assays and were incorporated into a response model based on the linear-quadratic model. Storage phosphor plate technology was used to acquire images of antibody distribution in tumor sections. These were registered with corresponding images showing tumor morphology, which were subsequently used to delineate regions that were distinct in terms of their response to radiation: oxygenated, radiosensitive areas that contained viable cells and hypoxic areas containing resistant viable cells and necrotic cells. Beta point dose kernels were then used to estimate the absorbed dose distribution in these regions. RESULTS: Therapy in normoxic areas was more effective than in hypoxic areas. The multivalent, tumor-specific antibodies, with intermediate clearance rates, delivered the highest absorbed dose to viable tumor cells. Antibody affinity and avidity facilitated the prolonged retention in radiosensitive areas of tumor, where most of the dose was deposited. The effectiveness of therapy could be enhanced further by matching the radionuclide with the antibody and tumor size. CONCLUSIONS: The model presented in this article allows the interaction between important radiobiologic parameters to be assessed and provides a tool for optimizing therapy in animal models and in patients.