Validity and Reliability of Renal Cortical Interstitial Fibrosis Assessment by Pathologists on Needle Core Biopsy
LD Cornell, DV Miller, AD Rule, ME Fidler, FG Cosio, MD Stegall, H Amer. Mayo Clinic, Rochester, MN; Intermountain Medical Center, Murray, UT
Background: Interstitial fibrosis (IF) is a histological finding that estimates severity of irreversible chronic kidney disease. Pathologists routinely estimate the percentage of IF in kidney biopsies. The accuracy and precision of this determination is unknown. The amount of IF guides clinical decision making between therapy and anticipation of end-stage renal disease. Thus, it is important to know the accuracy and precision of IF assessment of a standard needle core biopsy.
Design: An autopsy kidney from a 66 yo man with chronic kidney disease(eGFR=36) was obtained. Multiple 18-gauge needle core biopsy samples were taken from upper, middle, and lower regions of the kidney. These biopsies (3-4 cores per slide, 3 slides per region) were processed in the usual manner for clinical practice. Three pathologists estimated the percentage (5% intervals) of cortical IF on each slide in a blinded manner. As the gold standard, the entire renal cortex was examined by morphometric methods for IF. Sections of the entire kidney at 0.3 cm intervals were stained for type III collagen. Slides were digitalized using an Aperio scanscope. Quantitation of % IF, with manual delineation of the cortex, was performed by ImageScope pixel count analysis (pixels of type III collagen divided by pixels of cortex). The mean % IF across 27 readings (3 pathologists, 9 slides each) was compared to the % IF by the morphometric method with a t-test. The standard deviation (SD) and coefficient of variation (CV) were used to assess reliability between slides assessed by the same pathologist and between pathologists assessing the same slide.
Results: The mean %IF across the 9 slides was 7%, 12%, and 13% for each of the 3 pathologists, with %IF ranging from 5-25% on each biopsy sample (mean + SD 11% + 6% overall). This was an underestimate of IF compared to the 14% by the morphometric method (p=.02). The mean SD for multiple pathologists interpreting the same core sample was 4% (CV 41%) and the mean SD for multiple core samples by the same pathologist was 6% (CV 51%).
Conclusions: There is considerable variability in the assessment of IF due to both differences in core samples and differences between pathologists. IF assessment from a single needle core biopsy has reasonable validity but inadequate reliability. These findings have major implications in studies of the progression of fibrosis in renal allografts.
Category: Kidney (does not include tumors)
Wednesday, March 24, 2010 1:00 PM
Poster Session VI # 253, Wednesday Afternoon