Comparison of FNA and Core Biopsy Versus Complete Excision in Monitoring Tumor Response to EGFR Blockade in Murine Xenograft Models
Vamsi Parini, Pawel Mroz, William Grizzle, Peter Kulesza. Northwestern University, Chicago, IL; University of Alabama at Birmingham, Birmingham
Background: Fine Needle Aspiration (FNA) is a routine diagnostic tool, however core biopsy (cbx) is more widely used in settings that require ancillary studies. There are very few reports comparing their ability to adequately sample heterogeneous lesional tissue. To test this, a system is necessary where the extent of heterogeneity can be known, then sampled using FNA or cbx, and referenced to a gold standard i.e. complete excision. The objective of this study was to determine whether FNA is an appropriate tool to measure selective response to erlotinib in murine heterogeneous tumor model.
Design: Two cell lines were used in vivo: A375 melanoma (resistant to erlotinib) and H292 lung carcinoma (sensitive). Three sets of tumors were made: pure A375, pure H292, and a mixture of H292 and A375. Tumor-bearing animals received erlotinib or vehicle. After the therapy, FNA and cbx were procured; the tumors excised in toto for complete cross-sections. All specimens were formalin-fixed and paraffin embedded, the slides stained for cytokeratin 7 and S100 protein by immunohistochemistry; a fraction positive was assigned (0-100%).
Results: The average scores for CK7 for pure melanoma were low for all three sampling techniques: excision 1%, core bx 6% and FNA 0%; the S100 scores were: excisions 87%, core bx 72%, and FNA 82%. For pure carcinoma CK7 scores were: excision 99%, core bx100% and FNA 98%; S100 scores were: excision 0%, core bx 2.5% and FNA 0%. Treatment had no effect. The scores of vehicle group, mixed tumors for S100 were: excision 7%, core bx 15%, FNA 6%. For the same tumors, CK7 scores were: excision 83%, core bx 91%, and FNA 94%. For the tarceva group the S100 scores were: excision 29%, core bx 19% and FNA 29%; CK7 scores were: excision 53%, core bx 66% and FNA 33%.
Conclusions: The scores of pure tumors showed that CK7 is more accurate than S100 in discriminating between carcinoma and melanoma. In the untreated mixed tumors, the scores for FNA samples were similar to the reference standard, while the cbx overestimated the S100-positve fraction. In the treated mixed tumors, FNA undercounted the proportion of CK7-positive cells compared to reference, but it was accurate in S100 determination. In contrast, the cbx overcounted the CK7-positve cells, and undercounted the S100 fraction. These data suggest that the sampling accuracy of FNA may be superior to that of core biopsy, and that FNA can find application in following tumor response in the context of human clinical trials.
Wednesday, March 21, 2012 1:00 PM
Poster Session VI # 90, Wednesday Afternoon