The Value of Whole Slide Imaging for Enumeration of Plasma Cells in Plasma Cell Dyscrasia
J Kohan, S L Perkins, A Wilson, S Tripp, M E Salama. University of Utah, Salt Lake City; ARUP Institute for Clinical & Experimental Pathology, Salt Lake City
Background: Plasma cell (PC) quantification in bone marrow (BM) specimens is an important criterion for diagnosing and monitoring patients with plasma cell dyscrasia (PCD). Subjectivity and interobserver variability are unavoidable when enumerating PCs on aspirate smears or core biopsies. Whole slide scanning and image analysis (WSI) has emerged as a potentially field-changing technology, and can provide a more objective tool for assessing PCs. In this study we evaluated the utility of WSI as a tool for enumeration of PCs, and compared its performance to the traditional methods of PC enumeration.
Design: Samples from patients with diagnosis of PCD were selected from the University of Utah, Department of Pathology archives. Cases were included in the study if 3 or more BM biopsy procedures with adequate biopsy materials and laboratory data were available. PC% was determined by morphologic counts in an adequate aspirate smear (AS), by light microscopic estimation on a CD138 stained core biopsy (LM) and by WSI of the CD138-stained core slides that were digitally scanned (Aperio CS Scanscope), and analyzed with Aperio's Membrane IHC algorithm (Aperio Technologies, CA). Repeated-measures ANOVA was used to compare the different methods' ability to predict serum light chains (K/L) based on PC% in consecutive sample form a given patient.
Results: 52 samples collected from 16 patients (6 females) with PCD were included in the analysis. Patient age ranged from 43 to 80 years, median (m) = 59. PC % ranges were (<1- 90, m=5), (0.13-89.1, m=8.2) and (0-7.4, m=4) by LM, WSI and AS, respectively. Classification of samples based on the numeric value of %PC as having >10% or <10% by IA and LM were discrepant in 7/52 (13.5%) samples. PC% difference of ≥5% and ≥10% between values obtained by WSI and LM were noted in 17/52 (32.7%) and 9/52 (21.1%) samples.
WSI and LM showed statistically significant correlation with serum K/L ratio (p <0.005). Repeated-measures ANOVA indicated that WSI correlates more tightly with serum K/L ratio [standard errors (ST)= 0.7904] than LM [ST= 1.3509). Both WSI and LM performed significantly better than AS [ST= 3.3674].
Conclusions: PC count by WSI correlated more tightly with serum light chain levels and implied better disease burden than PC count by LM and both WSI and LM were superior to AS counts. WSI is a superior tool in enumerating plasma cells in CD138-stained BM core biopsies than LM with high degree of reproducibility and could be utilized for better patient stratification in clinical and research settings.
Wednesday, March 21, 2012 1:00 PM
Poster Session VI # 223, Wednesday Afternoon