[1679] Reproducibility of the Columbia Classification of Lesions of Focal Segmental Glomerulosclerosis

Shane M Meehan, Anthony Chang, Ian Gibson, Lisa Kim, Neeraja Kambham, Zoltan Laszik. University of Chicago, Chicago, IL; University of Manitoba, Winnipeg, Canada; Stanford University, Stanford, CA; Univeristy of California San Francisco, San Francisco, CA

Background: The Columbia working classification of focal segmental glomerulosclerosis (FSGS) identifies 5 types of lesions, designated collapsing (COLL), cellular (CELL), glomerular tip lesion (GTL), perihilar (PH) and not otherwise specified (NOS) types. This study examined interobserver reproducibility of criteria for classification of glomerular morphologic lesions of FSGS.
Design: Sixty-one 400x digital images of FSGS lesions, stained by PAS or Jones methenamine silver methods, were examined by six renal pathologists. Each image was classified independently by each observer using Columbia criteria. Criteria used for classification of lesions were tabulated by each observer. Lesions were considered diagnostic when a majority of opinion favored a single diagnosis.
Results: The percentage of overall agreement was 75.2% with a kappa value of 0.69, indicative of substantial reproducibility. Complete agreement (6 of 6 observers) occurred in 31 of 61 (50.8%) of cases. Consensus by 4 or more observers was attained in 86.9% of instances. Diagnostic disagreement by category is shown in the Table.

Discrepant Diagnoses

	Other Diagnoses Rendered
Majority Diagnosis (# cases)	COLL	CELL	GTL	NOS	PH
COLL (n=5)	-	4	1	3	0
CELL (n=7)	3	-	2	5	0
GTL (n=2)	0	2	-	0	0
NOS (n=12)	0	7	4	-	2
PH (n=0)	0	0	0	0	-

Interpretation of capillary retraction and endocapillary hypercellularity accounted for disagreement in COLL v CELL and COLL v NOS. Different interpretation of endocapillary hypercellularity accounted for disagreement in CELL v NOS. Disagreement about location of lesions explained discrepancies in interpretation of GTL and PH lesions.
Conclusions: Columbia criteria for morphologic lesions of FSGS are substantially reproducible between experienced observers and provide a robust method for assessment of the varied morphologic manifestations of this lesion.
Category: Kidney (does not include tumors)

Wednesday, March 21, 2012 1:00 PM

Poster Session VI # 276, Wednesday Afternoon