A Ratio of miR-30b/miR-205 Is Accurate in Distinguishing Spitz Nevus from Primary Cutaneous Melanoma
Lars Margolis, Jiamin Chen, Zhang Nick, Duane Barber, Harriet Feilotter, Tron Victor. Queen's University, Kingston, ON, Canada; University of Calgary, Calgary, AB, Canada
Background: Distinguishing Spitz nevus (SN) from Primary Melanoma (PM) remains a difficult histopathologic diagnosis. Recently, microRNA (miRNA) expression has been successfully used to classify tumors into appropriate lineages. We speculated that miRNA profiles differ in Spitz nevus as compared to primary melanoma, and thus could be used as a diagnostic aid.
Design: To address this, we compared the miRNA expression profiles of 14 classic Spitz nevi and 47 primary melanoma tumors using archival formalin fixed, paraffin embedded (FFPE) tissue. We have previously confirmed the validity of using FFPE tissue to profile miRNAs using an array platform. An Agilent miRNA microarray platform was used to generate miRNA expression profiles.
Results: We found that unsupervised hierarchical clustering clearly separated Spitz nevi from primary melanoma samples. Furthermore, Significance Analysis of Microarray (SAM) was applied to identify differentially expressed miRNA. We found both up-regulated and down-regulated miRNAs in primary melanoma relative to Spitz nevi. Two of the most significantly dysregulated miRNAs were miR-30b and miR-205. When used as a ratio, miR-30b/miR-205, we were able to accurately classify Spitz Nevi. The mean value for the miR-30b/miR-205 ratio in Primary Melanoma was 13.3 (range 0.036-105.7; n=47). The mean value for the miR-30b/miR-205 ratio in Spitz nevi was 0.004 (range 0.001-0.011; n=14). The p-value for the ratio of miR-30b/miR-205 between PM and SN found using an independent t-test is 0.003. To technically validate the microarray data, we conducted qRT-PCR on 15 PM and 6 SN samples from the original cohort profiled by microarray. The mean CT difference between miR-205 and miR-30b from qRT-PCR data for PM was 6.3 (standard deviation 4.1), and for SN was 0.39 (standard deviation 0.93). The p-value from an independent t-test using the qRT-PCR cohort was lower than 0.001. Our data showed a significant correlation between matched samples from microarray and qRT-PCR data (Kendall tau, 0.800; Spearman rho, 0.929).
Conclusions: In conclusion, we have found that the ratio of two miRNAs is highly accurate as a diagnostic test for Spitz Nevus.
Wednesday, March 21, 2012 1:00 PM
Poster Session VI # 104, Wednesday Afternoon