Epsilon Open Archive

Abstract

Renal scintigraphy has been used widely to measure the individual kidney glomerular filtration rate (IKGFR). In dogs, the estimated GFR is calculated by a regression equation relating the percent of the injected dose of radiopharmaceutical, technetium-99m DTPA (dietylene-triaminepentaacetic acid), taken up by the kidneys to the known GFR by standard (plasma) clearance normalized to bodyweight (BW). Uptake is calculated by the integral (area under the curve) of time activity curve (TAC) of the renogram, which is corrected for attenuation of radiation in the body from kidneys to body surface of the back and background activity. A more physiologically correct method is to normalize GFR to plasma volume (GFR/PV), which requires a region of interest (ROI) of the left ventricle (LV). The first aim of the thesis was to determine the variations within the standard method and minimize them to improve the accuracy of the methods. Physiological variability of GFR within and among dogs on different days was also studied. The integral method was more stable than a slope method and recommended to used for calculating IKGFR/BW. The best methods for measuring were: semi-automatic drawing of kidney ROI with 20% threshold, a perirenal background ROI at one pixel wide and one pixel out from kidney ROI, a threshold color scale rather than continuous for measuring kidney depth to calculate attenuation, and a time interval between 30 – 120 seconds from the start of the TAC. Measurement variation was also caused by significant observer variability, which indicates that methods must be standardized and the same person should measure all compared results. Physiologic day-to-day variability of GFR in normal dogs was mostly found between rather than within dogs, and accounted for most of the variability. GFR/PV was not affected by LV ROI sizes. Subtracting extravascular activity from the LV ROI did not improve precision, but increased variability due to different LV ROI sizes and time intervals chosen for LV plot. Manual LV ROI, without extravascular subtraction and a time interval for LV input between 1 and 4 minutes are recommended.