Imaging of non-small cell lung cancers with a monoclonal antibody, KC-4G3, which recognizes a human milk fat globule antigen.

To determine the role of lung cancer tumor imaging with monoclonal antibodies directed against high molecular weight human milk fat globule antigens, we administered i.v. 111In-KC-4G3 to 24 patients with advanced non-small cell lung cancer. One mg of 111In-KC-4G3 was mixed with 0, 9, 49, 99, or 499 mg of unlabeled KC-4G3 and infused i.v. over 1 to 5 h. The mean 111In-KC-4G3 radiochemical purity was greater than 97% and the resultant immunoreactivity averaged 62%. Successful imaging of cancer sites was accomplished in 92% of 24 patients, and 57% of 91 total lesions were visualized. Successful localization of tumor sites related to size (P less than 0.001), with 81% of lesions greater than 3.0 cm in diameter, 50% of lesions 1.5 to 3 cm, and 6% of lesions less than 1.5 cm successfully imaging, and to location (P less than 0.05), with 69% of pulmonary lesions, 80% of soft tissue lesions, and only 32% of bone metastases being visualized. Nonspecific reticulo-endothelial uptake of radioactivity was a major problem. Approximately 35% of 111In was chelated to serum transferrin by 24 and 48 h after infusion. The mean t 1/2 beta for plasma radioisotope and immunoreactive KC-4G3 was 29 and 27 h, respectively. There was no correlation between total infused antibody dose and imaging success or between total dose and effect on 111In and KC-4G3 kinetics. Circulating free KC-4 antigen was measurable in all but one patient before study. Tumor biopsy following infusion could demonstrate antibody presence but not saturable antigen binding. We conclude that (a) 111In-KC-4G3 demonstrates successful tumor localization in non-small cell lung cancers bearing generally high expression of its antigen and (b) further investigations to diminish nonspecific radioactivity for imaging and utilization of high dose radiolabeled antibody for therapeutic intent are warranted.

Initial cerebral HM-PAO distribution compared to LCBF: use of a model which considers cerebral HM-PAO trapping kinetics.

The cerebral uptake of [99mTc]-d,l-hexamethylpropyleneamine oxime complex (HM-PAO) was compared to LCBF determined simultaneously with [14C]iodoantipyrine (IAP) using double radionuclide quantitative digital autoradiography. Awake male rats were given intravenous injections of a mixture of 50 microCi IAP and 15 mCi of HM-PAO and killed 20 s after tracer activity had first reached the brain. Two separate autoradiograms were produced from each 20 microns brain section. The autoradiograms were digitized, corrected for cross-contamination, and then converted into images of individual tracer concentration. The diffusible tracer model was used to convert the IAP concentration images into LCBF images. Regional HM-PAO concentration was found not to be linearly related to LCBF as determined with the IAP, and therefore a simple microsphere type model was inadequate in relating HM-PAO uptake to LCBF. A better HM-PAO uptake--LCBF correlation was obtained when the HM-PAO arterial input function was corrected for very rapidly produced, non-cerebrally extracted, metabolites and a kinetic model was used that considered the rate of intracerebral metabolism of HM-PAO to a retained metabolite. Even using this model, however, some differences between HM-PAO uptake and LCBF occurred in certain brain regions. Because these differences were small and the HM-PAO uptake pattern has been shown to be constant for many minutes, HM-PAO can probably be used to estimate LCBF in patients with single positron emission computed tomography (SPECT) imaging.

Maximizing precision in quantitative autoradiographic determination of tissue tracer concentration using exposure optimization.

The relationships between tissue tracer concentrations, length of time of tissue contact with film, and the darkness of resulting autoradiographic images were explored. Operational equations were then developed relating ranges of film darkening to the ranges of tracer concentration contained within the sections. These equations were solved and used to define ranges of optical density that are optimal for precise determination of tracer concentration using digital image analyzers. The solutions indicated that optimal optical densities are a function of the range of tracer concentrations in the sections. For autoradiograms of typical cerebral blood flow and metabolism tracers, exposure should be controlled to produce images that are significantly less dark than what is generally considered pleasing to the eye.

Autoradiographic measurement of cerebral lactate transport rate constants in normal and activated conditions.

We used quantitative autoradiography to measure the regional rate constants of blood-to-brain transport of lactate in normal rats and rats treated with kainic acid. Mean cerebral values of lactate transport rate constants were not significantly different between the normal and treated rats, being 0.13 and 0.14 min-1 (ml/g), respectively. Regional values were also generally similar between the groups, but structures that are known to be activated by kainic acid showed increased values in the treated rats compared with rates in the controls. Our measured values of lactate transport rate constants are approximately 50% as great as those published for glucose, indicating that blood-brain transfer of lactate can be significant. This observation supports the hypothesis that radiolabel derived from glucose can leave the brain as radiolabeled lactate in conditions in which intracerebral lactate concentration rises, a hypothesis that has previously been presented to explain differences between rates of accumulation of radiolabel derived from deoxyglucose and glucose in such conditions.

Comparison of cerebral glucose metabolic rates measured with fluorodeoxyglucose and glucose labeled in the 1, 2, 3-4, and 6 positions using double label quantitative digital autoradiography.

We compared local cerebral glucose metabolic rates (LCMRglu) that were determined with [18F]fluorodeoxyglucose (FDG) and [14C]glucose labeled in the 1, 2, 3-4, and 6 positions. Double label digital autoradiography was used with published kinetic models to determine LCMRglu for FDG and glucose in the same animals. Glucose showed metabolic rate dependent underestimation of LCMRglu compared to FDG, which worsened with increasing experimental times. The least underestimation occurred with glucose labeled in the 6 position at 6 min, reaching 10% in areas of high metabolism. Labeling in the 1 position, the 2 position and the 3-4 position caused progressively worse underestimation at all times. In addition, some structures showed differences not directly related to metabolic rate, indicating regional variations in relationships between individual kinetic constants of FDG and glucose.

Glycolysis-induced discordance between glucose metabolic rates measured with radiolabeled fluorodeoxyglucose and glucose.

We have developed an autoradiographic method for estimating the oxidative and glycolytic components of local CMRglc (LCMRglc), using sequentially administered [18F]fluorodeoxyglucose (FDG) and [14C]-6-glucose (GLC). FDG-6-phosphate accumulation is proportional to the rate of glucose phosphorylation, which occurs before the divergence of glycolytic (GMg) and oxidative (GMo) glucose metabolism and is therefore related to total cerebral glucose metabolism GMt: GMg + GMo = GMt. With oxidative metabolism, the 14C label of GLC is temporarily retained in Krebs cycle-related substrate pools. We hypothesize that with glycolytic metabolism, however, a significant fraction of the 14C label is lost from the brain via lactate production and efflux from the brain. Thus, cerebral GLC metabolite concentration may be more closely related to GMo than to GMt. If true, the glycolytic metabolic rate will be related to the difference between FDG- and GLC-derived LCMRglc. Thus far, we have studied normal awake rats, rats with limbic activation induced by kainic acid (KA), and rats visually stimulated with 16-Hz flashes. In KA-treated rats, significant discordance between FDG and GLC accumulation, which we attribute to glycolysis, occurred only in activated limbic structures. In visually stimulated rats, significant discordance occurred only in the optic tectum.

High-performance digital image analyzer for quantitative autoradiography.

A digital image-processing system was developed for high-spatial resolution analysis of autoradiograms. The system uses a linear array of charge-coupled devices operating under computer control to scan and digitize autoradiograms into 512 X 512 matrices with 256 gray levels. Software was developed to facilitate quantitative analysis of autoradiograms produced in single- and multiple-tracer studies. Because of the high output, linearity, and accuracy of the solid-state detectors, the system was found to digitize autoradiograms significantly more precisely and accurately than previously described video camera- or photomultiplier tube-based scanning densitometers.

Multiple-radionuclide autoradiography in evaluation of cerebral function.

We have developed the mathematical model and experimental technique for quantitative simultaneous multiple-radionuclide autoradiography. The technique is an extension of previously described dual-tracer methods and offers the advantage of a general approach in which any two or more radionuclides with different half-lives can be used to measure multiple parameters of cerebral function. The model allows the selection of nuclide doses and exposure times to balance cross-contamination of the multiple images with nuclide requirements. The technique was validated with homogeneous tissue sections containing combinations of 123I, 124I , 131I, 18F, 99mTc , 68Ga, and 14C, and then used in qualitative and quantitative investigations of interrelationships of local cerebral blood flow, glucose metabolism, and protein synthesis in normal and several pathological conditions in rats.

Evaluation of [123I]isopropyliodoamphetamine as a tracer for local cerebral blood flow using direct autoradiographic comparison.

We investigated [123I]isopropyliodoamphetamine (IMP) for potential use in the autoradiographic determination of local cerebral blood flow (LCBF) in animals. The technique of direct autoradiographic comparison, derived from double radionuclide autoradiography, was used to compare the simultaneous uptakes of IMP and [14C]iodoantipyrine (IAP), a reference tracer, in awake and anesthetized rats. This new technique offers several advantages over the previously developed methods of comparing tracers, brain uptake index and first pass extraction ratio. These include the avoidance of disrupting normal cerebral blood-brain tracer exchange and the ability to compare uptakes at substructural levels, whereas the other methods are limited to larger areas. Mean values of LCBF obtained with IMP agreed closely with those using IAP, from 20 to 300 ml/100 g/min. Because IMP was found to have an extremely high effective brain:blood partition coefficient, approximately 25:1, a linear uptake tracer model could be used for IMP yielding more precise values than could IAP for LCBF values above 150. IMP was found to measure choroid plexus flows much more accurately than IAP, values being greater than 500 for IMP compared to approximately 200 for IAP. Because the mechanism of the extremely high partition coefficient of IMP is not yet defined, however, care must be used in measuring LCBF with IMP where the trapping mechanisms of normal vessels may be disrupted.

Quantitative local cerebral blood flow measurements with technetium-99m HM-PAO: evaluation using multiple radionuclide digital quantitative autoradiography.

We investigated d,1 [99mTc]hexamethylpropyleneamine oxime complex (HM-PAO) as a tracer for quantitative measurement of local cerebral blood flow (LCBF) in a series of awake male rats. LCBF measurements with HM-PAO were compared to those of two other tracers, [14C] iodoantipyrine (IAP) and [201Tl]diethyldithiocarbamate (DDC), using quantitative double and triple tracer digital autoradiography. LCBF values with HM-PAO averaged 64% those of IAP and were generally linearly related. Detailed analysis suggested that the underestimation of LCBF by HM-PAO was related to blood constituent binding and/or rapid conversion to a noncerebrophilic compound, as well as noninstantaneous cerebral trapping, rather than to diffusion limitation.

Effect of exercise position during stress testing on cardiac and pulmonary thallium kinetics and accuracy in evaluating coronary artery disease.

We compared the effects of symptom-limited upright and supine exercise on 201Tl distribution and kinetics in the heart and lungs of 100 consecutive patients. Our analysis was based on data obtained with a digital gamma camera in the 45 degrees left anterior oblique position at 5, 40, 240, and 275 min postadministration of [201Tl]chloride. We found significant differences in the results at the 5- and 40-min intervals; viz, higher cardiac and lower pulmonary thallium activity after upright exercise in 94 subjects at both intervals, and greater variability in total and regional cardiac thallium kinetics after supine exercise. With supine exercise, the relatively low initial cardiac activity, relatively high lung activity, and the greater variability in thallium kinetics combined to make interpretation of quantitative data and cardiac images difficult and less accurate with respect to detection of coronary artery disease. These observations have important implications for the interpreting physician when thallium stress tests are performed in the supine position.

Comparison of regional blood-brain transport kinetics between glucose and fluorodeoxyglucose.

The fluorodeoxyglucose (FDG) method for estimating regional cerebral glucose metabolic rate (LCMRglc) requires that a fixed relationship (the "lumped constant") exists between net FDG and glucose (GLC) extraction throughout the brain. In addition to the relative rate of metabolism between FDG and GLC, this assumed constant is affected by the relative rate of blood-to-brain FDG transport compared to that of glucose. However, little data is available regarding the regional stability of the FDG versus GLC transport-rate relationship. We therefore used high resolution, quantitative dual-tracer digital autoradiography to directly compare the blood-to-brain transport rate constants (K1) of radiolabeled GLC and FDG in normal and pharmacologically-stimulated rats. The rats were given 45 sec terminal intravenous infusions of a mixture of 18F-FDG and 14C-GLC. Autoradiograms of the brain representing the FDG and GLC tracer concentrations were produced, digitized, and converted into digital images of K1. We found that the global K1 values of FDG and GLC were not significantly different from each other. However, detailed analysis revealed that some structures in the normal animals, such as the hippocampus and cerebellum, had different quantitative patterns of FDG transport compared to GLC transport. Thus, our results indicate that the relationship between GLC and FDG transport is not uniform throughout the brain as has previously been assumed. This observation suggests that regional variations in the type and distribution of glucose transporters may exist and that the fluorodeoxyglucose "lumped constant" may vary somewhat among different brain regions.

Autoradiographic comparison of thallium-201 diethyldithiocarbamate, isopropyliodoamphetamine and iodoantipyrine as cerebral blood flow tracers.

We investigated [201Tl]diethyldithiocarbamate (DDC) as a tracer for local cerebral blood flow (LCBF). Awake male rats were given intravenous infusions of a mixture of DDC and reference tracer(s): [123I]isopropyliodoamphetamine (IMP) and/or [14C]iodoantipyrine (IAP). LCBF values for DDC, IMP, and IAP were measured using simultaneous multiple radionuclide autoradiography and quantitative digital image analysis. Patterns of local cerebral blood flow (LCBF) obtained with DDC were intermediate compared to IMP and IAP, although they were more similar to those of IMP. DDC and IMP underestimated LCBF in some white matter and adjacent structures, while IAP underestimated LCBF values in high flow regions. We conclude that DDC uptake generally reflects local cerebral blood flow and that it can therefore be used as a cerebral perfusion tracer in humans using SPECT imaging.

Digital autoradiography: design, development, and evaluation of a solid-state image analyzer.

A digital image analyzer was developed for high spatial resolution analysis of autoradiograms. The system uses a linear array of charge-coupled devices operating under microcomputer control to scan and digitize autoradiograms into matrices of up to 1,500 X 2,000 pixels with 256 gray levels. The digitized images can be converted from gray scale to pseudo-color and displayed on a high resolution color monitor. Software was developed to facilitate quantitative analysis of autoradiograms produced in single and multiple tracer studies. Because of the high output linearity and accuracy of the solid-state detectors, the system was found to digitize autoradiograms significantly more precisely and accurately than previously described autoradiographic analyzers.

Quantitative measurement of renal perfusion following transplant surgery.

We developed an easily implemented clinical procedure for quantitative perfusion measurements in transplanted kidneys using intravenously administered [99mTc]DTPA and the tracer fractionation technique. F = Ak(T)/0 integral of T [Aa(t)/Va] dt, where F = renal blood flow, Ak(T) = DTPA activity in kidney at time = T, Va = ultrasonographically measured femoral artery segment volume, T = time postinjection of F determination, and Aa(t) = time course of DTPA activity in femoral artery segment. The technique was applied to a group of 80 studies in 35 patients in whom an independent clinical determination of transplant function was available. Blood flow (units of ml/min) measured 439 +/- 83 in normally functioning transplants, 248 +/- 63 in transplants with acute tubular necrosis, 128 +/- 62 in transplants with rejection, and 284 +/- 97 in transplants with cyclosporine toxicity. These preliminary results indicate potential usefulness of this method in the evaluation of renal function following transplant surgery.

DTPA aerosol in ventilation/perfusion scintigraphy for diagnosing pulmonary embolism.

UNLABELLED: The use of lung scintigraphy in evaluating suspected pulmonary embolism (PE) is controversial. Several diagnostic methods have been described for lung scans, of which the most widely applied uses 99mTc-MAA for perfusion, 133Xe for ventilation and PIOPED diagnostic criteria. This study evaluates the accuracy of lung scintigraphy using an alternative ventilation agent, 99mTc-diethylenetriamine pentacetic acid (DTPA) aerosol, and specific criteria. METHODS: Diagnostic criteria for DTPA aerosol ventilation were prospectively applied to 5017 patients over a 9-yr period. Lung scan interpretations were analyzed for frequency of occurrence, and results were compared to those of angiography in 455 patients. RESULTS: Scans were interpreted as normal, low or high probability in 79% of patients and as either indeterminate or medium probability in 21% of patients. Three patients had normal scans and negative angiography. In patients with low-probability scans, 111 angiograms were performed: 103 (93%) were negative, and 8 (7%) were positive. In patients with indeterminate scans, 114 angiograms were performed: 85 (75%) were negative, and 29 (25%) were positive. In patients with medium-probability scans, 149 angiograms were performed: 86 (58%) were negative, and 63 (42%) were positive. In patients with high-probability scans, 78 angiograms were performed: 6 (8%) were negative, and 72 (92%) were positive. CONCLUSION: These results indicate that lung scintigraphy using DTPA aerosol and our criteria is accurate in diagnosing and stratifying risk of pulmonary embolic disease. Compared with 133Xe and PIOPED criteria, DTPA ventilation and our criteria reduced the false-negative rate in low-probability scans (7% versus 16%, p < 0.005) and decreased the fraction of intermediate-probability scans (21 % versus 39%, p < 0.01).

Quantifying local cerebral blood flow by N-isopropyl-p-[123I]iodoamphetamine (IMP) tomography.

A model was validated wherein local cerebral blood flow (LCBF) in humans was quantified by single-photon emission computed tomography (SPECT) with intravenously injected N-isopropyl-p-[123I]iodoamphetamine (IMP) combined with a modification of the classic method of arterial input sampling. After intravenous injection of IMP in rat, autoradiograms of the brain showed activity distributions in the pattern of LCBF. IMP was nearly completely removed on first pass through monkey brain after intracarotid injection (CBF=33 ml/100 g/min) and washed out with a half-time of approximately 1 hr. When the modified method of arterial input and tissue-sample counting applied to dog brain, there was good correspondence between LCBF based on IMP and on that by microsphere injection over a wide flow range. In applying the method to human subjects using SPECT, whole-brain CBF measured 47.2 +/- 5.4 ml/100 g/min (mean +/- s.d., N=5), stable gray-white distinction persisted for over 1 hr, and the half-time for brain washout was approximately 1 hr. Perfusion deficits in patients were clearly demonstrated and quantified, comparing well with results now available from positron ECT.

Reverse and pseudo redistribution of thallium-201 in healed myocardial infarction and normal and negative thallium-201 washout in ischemia due to background oversubtraction.

While the interpolative background subtraction used in quantitative planar thallium scanning can significantly overestimate the background overlying the heart, the effects of background oversubtraction on quantitative analysis have not been well defined. A mathematical model that relates myocardial washout determined using interpolative background subtraction to true myocardial washout is presented. The model was validated using phantoms and applied to myocardial and pulmonary thallium kinetic data in 100 patients, 85 with and 15 without coronary artery disease. The model showed that when using interpolative background subtraction, measured washout equals true washout in normally perfused myocardium; however, depending on the relation between myocardial and pulmonary thallium clearance, myocardial washout in ischemic regions and areas of infarction can be substantially over- or underestimated. Based on generally accepted quantitative criteria, this incorrect washout determination can at times lead to misdiagnosis of infarction as ischemia and ischemia as normally perfused tissue. It can also cause both "reverse redistribution" and "pseudo redistribution" of thallium in myocardial infarction in the absence of a physiologic basis.