Novel Subtraction Morphometry Technique for Assessing Fibrosis in Renal Biopsies
AB Farris, CD Adams, N Brousaides, PA Della Pelle, AB Collins, RB Colvin. Massachusetts General Hospital, Boston, MA; Harvard Medical School, Boston, MA
Background: Quantitation of interstitial fibrosis has become of increasing importance for assessment of efficacy in clinical trials of drugs in transplantation and native kidney disease. Visual assessment is the standard practice using trichrome stained slides, but is poorly reproducible. The alternative, morphometry with special stains such as sirius red with polarization or anti-type III collagen IHC is cumbersome, requiring special processing. Therefore, we sought a new method that could be applied to routine renal biopsy stains.
Design: Trichrome stains all collagens, including basement membranes, which should not be included in measurements of interstitial fibrosis but are stained strongly by PAS. We reasoned that subtraction of the area stained by PAS from that stained blue by trichrome might be a robust measure of interstitial collagen. To test this, routine histology trichrome and PAS stained slides from 77 donor biopsies and 18 late transplant biopsies or nephrectomies were scanned with an Aperio ScanScope and analyzed using the ImageScope Positive Pixel Count algorithm with appropriate parameters that selected blue (trichrome) and pink (PAS). The % cortical interstitial fibrosis (T-P fibrosis) was calculated as (Trichrome area-PAS area)/total area and compared with a visual assessment.
Results: In 77 donor biopsies T-P fibrosis averaged 2016 % (-6% to 64%). This reflects the finding that most donor biopsies had some fibrosis by visual scoring. Negative T-P fibrosis values (n=6) occurred when there was essentially no fibrosis and PAS area exceeded trichrome area. In late transplant samples, T-P fibrosis values were 3121 % (4 to 75%). T-P fibrosis correlated moderately well with visual scores (R = 0.41, P < 0.0001). When analysis was performed at a different occasion on the same biopsy, essentially identical results were obtained. Subtraction removed the contribution of casts and sclerotic glomeruli and did not include tubular brush borders. The disparities between visual and morphometry seemed largely attributable to variability in staining intensity of trichrome and freeze artifact in donor biopsies.
Conclusions: The subtraction morphometry method for fibrosis provides the potential of routine histologic stains that are prepared on renal biopsies and widely available in clinical trials. The quantitative method is less subject to observer variability than visual assessment, but does require satisfactory tissue processing and staining.
Category: Kidney (does not include tumors)
Tuesday, March 10, 2009 1:00 PM
Poster Session IV # 195, Tuesday Afternoon