Safety, Biodistribution, and Dosimetry of ¹²³I-IMPY: A Novel Amyloid Plaque–Imaging Agent for the Diagnosis of Alzheimer's Disease

Subject Accrual

The design was similar to several others that our laboratory has previously reported (19–23). The protocol was approved by the local Committee on Research Involving Humans and the U.S. Food and Drug Administration. Healthy volunteers were recruited through advertisements in local newspapers and by word of mouth from other volunteers. AD patients were referred from the Alzheimer's Disease Center and met the probable-AD criteria of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer Disease and Related Disorders Association (24). Medical histories were taken and physical examinations performed before inclusion. At the time of the study, none of the subjects had a known history of a health problem that could have significantly affected the biodistribution or elimination of the radioligand. All AD patients were taking donepezil and multivitamins. One patient was taking finasteride for the prostate, 1 patient was taking L-citalopram, 2 patients were taking 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors, and 1 patient was taking a low dose of furosemide for mild hypertension. All urine drug screens and serum laboratory tests were performed to confirm that the individuals were healthy. The final sample included 9 subjects: 5 men and 4 women.

Clinical Assessment

The baseline clinical laboratory tests included a complete blood cell count with differential and tests of serum electrolytes and liver enzymes. Levels of creatinine, blood urea nitrogen, glucose, cholesterol, triglycerides, albumin, and total protein were also assayed. Routine urinalyses and urine toxicology screens were performed. The blood tests were repeated 4, 24, and, when possible, 48 h after administration of the radiopharmaceutical. The urine tests were repeated once after 24 h.

Electrocardiography was performed 20 min before injection of the radiopharmaceutical; 5, 10, 30, and 60 min after injection; and 24 h after injection. Vital signs (blood pressure, heart rate, and pulse oximetry) were taken about every 5 min from 20 min before injection to 60 min after injection.

Subject Preparation

On the morning of the first day of the study, intravenous access was obtained with the subjects lying supine on the imaging table. To block thyroid uptake, we gave Lugol's solution to the subjects 45 min before injection of ¹²³I-IMPY. Serum was obtained intravenously after the subjects had been lying supine for approximately 15 min. Electrocardiogram leads and a blood pressure cuff were placed on the subjects. Subjects were monitored for 20 min before injection of ¹²³I-IMPY and for up to 60 min after injection as already described.

Imaging Techniques

Linear attenuation was measured using a noncollimated transmission source prepared by dissolving about 1,000 MBq of ^99mTc in a 1,600-mL sheet flood made of clear acrylic. Transmission scans were obtained and attenuation calculated using a previously described technique (19). For acquisition of emission images, 185 MBq (5 mCi) of the radiopharmaceutical were injected rapidly as a bolus through the indwelling intravenous catheter (the dose was determined by subtracting the activity remaining in the syringe after injection from the activity before injection). Whole-body images were acquired using a dual-head scanner (Prism 2000; Picker) previously calibrated for the ¹²³I isotope. Imaging was concentrated over the initial 6 h after injection (typically 6–10 scans), and the remaining images were acquired at approximately 24, 30, and 48 h after injection (Fig. 2). A mean of 9.0 ± 3.2 whole-body images were acquired. Each whole-body scan was acquired in a 256 × 1,024 matrix for either 10 or 20 min over a total excursion length of 214 cm.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Representative anterior (left) and posterior (right) whole-body scans at 20 min (A), 26 h (B), and 48 h (C) after intravenous injection of ¹²³I-IMPY.

Image Analysis

An operator drew regions of interest (ROIs) around 12 different organs or tissues and around the whole body. The regions were drawn on whichever scan showed the organ most clearly. Most organ boundaries were placed on the first whole-body scan. An abdominal ROI was drawn to include the entire abdomen except the liver. Thus, the abdominal ROI obtained the maximum activity in the entire gut, and it was assumed the activity found its way there via hepatobiliary excretion. Because of the blocking with Lugol's solution, the thyroid was difficult to clearly visualize in most subjects, and therefore, the ROI representing the thyroid was large and stylized to reflect nonspecific activity in the region of the thyroid fossa. Because the kidneys were not visualized well in the anterior projection, the regions for them were placed by flipping the posterior ROIs.

Regardless of which scan the ROI originated from, it was cut and pasted into a single master set. Once the set was complete, the ROIs were transposed onto all the other images from the same patient, including the transmission scans. Occasionally, an operator had to move the entire set of ROIs to correct for repositioning errors between scans. In addition, the size of the ROIs for the urinary bladder sometimes had to be adjusted to account for normal changes in volume, with the regions always drawn to include all observable activity (the bladder was always clearly separated from any other areas of activity). The ROI for the body almost always required minor revisions to correct for subject movement, differences in pelvic tilt, and the position of the feet. Otherwise, the individual ROIs were rarely manipulated independently of the other regions in the set. An automated subroutine measured the number of counts in these ROIs.

Organ Activity Calculations

Attenuation corrections were applied using the experimentally measured ratio of counts in the transmission scans of the subjects on the imaging table and the forward-decay–corrected counts in the nonattenuated transmission scans through air.

Geometric means for each pair of decay- and attenuation-corrected conjugate ROIs were calculated by multiplying the net anterior counts by the net posterior counts and taking the square root of the product. Decay correction was necessary to compare each time point with the original whole-body activity level. The fraction of injected dose at each time point was then estimated by dividing the corrected geometric mean number of counts in each ROI by the net geometric mean number of counts in the initial whole-body image.

Organ Retention Functions

Time–activity curves were generated directly from the experimental data for the brain, gallbladder, heart, left kidney, liver, right lung, salivary glands, spleen, thyroid, urinary bladder, and the whole abdominal compartment excluding the urinary bladder. For this dataset, 1 or 2 exponential functions were iteratively fit to the time–activity curves using a nonlinear least-squares regression algorithm. These curves were numerically integrated to infinity to yield source organ cumulated activity values. Whole-body retention was estimated from the ROI encompassing the entire body. Urine was collected from each subject, and scintigraphic images of the collection jugs were obtained. However, when we discovered that difficulties in image acquisition or processing prevented the use of these data, urinary excretion was subsequently estimated from total-body clearance. The residence times for the urinary bladder, and thus the dose estimates that followed, were based on a theoretic bladder-voiding interval of 4.8 h, or 5 times a day (25).

Total fecal excretion was calculated as the maximum percentage injected activity in the gut at any image time during the study. The estimated gut radioactivity levels reached a maximum typically by about 24 h. Gallbladder residence times were calculated directly from the time–activity curves, with no accounting for routine voiding of the contents. Thus, gallbladder residence times (and radiation doses) are probably overestimates.

The percentages in the human organs were fit to 1- or 2-compartment exponential models using the SAAM II software (University of Washington). Time integrals of activity were calculated and converted to the number of disintegrations occurring in the source organs (26); these values were entered into the OLINDA/EXM 1.0 software (Vanderbilt University) using the adult male model, which calculates the values for other organs not measured directly using the regions. Excretion of activity was modeled by assuming that activity not accounted for in the measurements had been eliminated through the urine (a dynamic bladder model with a 4.8-h voiding interval was used to estimate the dose to the urinary bladder wall) and by assuming that the excretion was 75% urinary and 25% intestinal. This assumption was based on observations that the average percentage activity measured in the abdominal area, when activity in the kidneys, liver, and gallbladder was removed, was 23.4%, with an SD of 9.9%. The dose to the salivary glands was calculated separately (because this region is not defined in currently available phantoms and so is not in the OLINDA/EXM code) using the unit density sphere approximation in the OLINDA/EXM software.

Serum Metabolite Analysis

Heparinized blood samples (1–3 mL) were obtained from 3 healthy control subjects and 3 AD patients at various time points (2, 5, 10, 30, and 60 min) after an intravenous injection of ¹²³I-IMPY. The blood samples were centrifuged, and the obtained plasma samples were counted to determine total radioactivity. An equal volume (500 μL) of ethyl acetate was added to the plasma fraction in the presence of 20 μg of IMPY carrier to maximize extraction yield in the microcentrifuge tubes (1.3 mL). Extraction was repeated 3 times, and the combined EtOAc extraction and aqueous layer were counted separately to obtain the percentage of EtOAc extraction for each sample. To quantify the parent ¹²³I-IMPY nonmetabolized in the samples, EtOAc extracts were condensed and reconstituted in a small volume of MeOH for thin-layer chromatography analysis (silica gel, mobile phase: chloroform/EtOH/concentrated NH₃ = 8/1.5/drop). Under this thin-layer chromatography condition, ¹²³I-IMPY gave an R_f value of 0.8–0.9, whereas the metabolites showed lower R_f values, from which nonmetabolized ¹²³I-IMPY in EtOAc extract was quantified.