Strategies for H-score Normalization of Preanalytical Technical Variables with Potential Utility to Immunohistochemical-Based Biomarker Quantitation in Therapeutic Reponse Diagnostics

Abstract

Digital quantitative immunohistochemical analysis of protein biomarker expression offers a broad dynamic range against which clinical outcomes may be measured. Semi-quantitative expression data represented as an H-score is produced by computer generated average intensity of positive staining given weight by the percentage of cells showing positive staining. While patient H-scores vary for biological reasons, variation may also arise from preanalytic technical issues, such as differences in fixation protocols. In this study, we present data on two candidate calibrator nuclear-localized proteins, SNRPA and SnRNP70, with robust and consistent expression levels across breast cancers. Quantitative expression measurement of these two candidate biomarkers may potentially be used to eliminate the effect of differences in preanalytic processing of specimens by normalizing H-scores derived from test biomarkers of interest. To examine the effects of preanalytical fixation variation on biomarker quantitation and potential utility of candidate calibrators to address such issues, 6 surgically-resected human breast cancer patient specimens were divided into 6 portions and fixed under distinct conditions (fixation following resection in formalin for 2 hr, 8 hr or 48 hr, or held overnight at 4°C in buffered saline prior to formalin fixation for 2 hr, 8 hr, or 48 hr). We find H-score variation between fixation conditions within individual patient's tumors that were stained for XPF, ATM, BRCA1, pMK2 and PARP1. Most interestingly, detectable expression of SNRPA and SnRNP70 is covariant to test biomarkers under distinct fixation conditions and so these hold the potential for serving as calibration standards for general antigen preservation and reactivity.