Development of a Multi-Gene qPCR Assay for Adrenocortical Tumor Diagnosis and Prognosis
Scott A Tomlins, Michelle Vinco, Rork Kuick, Thomas J Giordano. University of Michigan, Ann Arbor, MI
Background: As incidentally detected adrenal masses are increasing, the correct diagnosis of adrenocortical tumors, particularly differentiating adrenocortical carcinomas (ACC) from adrenocortical adenomas (ACA), is of growing importance. Differentiating ACAs and ACCs can be challenging, and although overall survival of patients with ACC is poor, the course can be variable. Hence, there is a need for novel biomarkers to classify ambiguous ACTs and predict ACC behavior.
Design: We developed a multi-gene qPCR panel optimized for RNA from formalin fixed paraffin embedded (FFPE) tissues, consisting of 33 assays targeting 28 genes of interest (two assays targeted IGF2) and 4 housekeeping genes. Total RNA was isolated from 3x10um FFPE sections from 96 cases (5 normal adrenal cortex [NAC], 24 ACAs and 67 ACCs). Reverse transcription and qPCR was applied to these cases, with two cases assessed in duplicate (total qPCR n=98). Target gene expression was normalized to three housekeeping genes showing robust amplification.
Results: A median of 7.0 ug RNA was isolated per case (range: 1.0-37.5ug). Eighty two of 98 (84%) samples had sufficient amplifiable RNA for analysis. Samples with sufficient amplifiable RNA were from significantly more recent blocks than samples with insufficient RNA (mean 7 years vs. 12 years, p=0.0002). Of blocks less than 10 years old, 59/64 (92%) had sufficient amplifiable RNA for analysis. To assess the reproducibility of the reverse transcription and qPCR, RNA from two samples (ACC_64 and ACA_18) was subjected to duplicate reverse transcription and qPCR; for each sample, the duplicate sample showed the highest correlation of normalized target gene expression when compared to the 81 other samples (ACC_64_1 and ACC_64_2, r=0.992; ACA_18_1 and ACA_18_2, r=0.997). Normalized expression of IGF2 by the two assays was highly correlated across the 82 samples (r=0.963), supporting the robustness of the qPCR assay design. Unsupervised centroid linkage hierarchical clustering using normalized expression of the target gene assays robustly separated NAC, ACAs and ACCs (one low grade ACC clustered with ACAs and NACs). Expression of cell cycle genes was highly correlated across samples, and a cluster of ACCs with low proliferation and long over-all survival was present.
Conclusions: We have developed a robust, highly reproducible multi-gene qPCR panel for FFPE tissue which robustly discriminates ACAs and ACCs. Optimized diagnostic and prognostic models will be presented, including comparisons to standard clinicopathological parameters and clinical outcome.
Monday, March 4, 2013 2:30 PM
Proffered Papers: Section H, Monday Afternoon