211 Evaluation of a New Semiquantitative Method for Scoring Immunohistochemical Stains

Abstract Objectives Routine scoring of immunohistochemical stains (IHC) accounts for only a single staining intensity, ignoring the staining heterogeneity displayed by most tissues. While quantitative visual scores and digital image analysis platforms are available, their clinical utility is limited by their cost and cumbersomeness. Here we describe a simple semiquantitative method for scoring IHC that acknowledges the heterogeneity of IHC staining, and evaluate its reproducibility. Methods We devised a summative IHC score (SSc) where, akin to the Gleason score, the first and second most frequent IHC intensities (range 0–3) are added. Each possible SSc is assigned an ordinal value (OV; range 0–15) such that for each SSc, the combination with a higher primary intensity always corresponds to a higher OV (eg, 1 + 0 > 0 + 1). Four pathologists independently analyzed a transferrin receptor 1 IHC on a subset of 320 cores from a prostate cancer TMA. All participants completed a training session on a separate set of practice cores. Intensity references were established by consensus and imaged for reference. Interobserver agreement on the primary intensity (PI), secondary intensity (SI), and SSc for each individual core was assessed using kappa statistics (ƙ), and internal consistency (IC) was measured using Cronbach α. Results ƙ for PI ranged from 0.58 to 0.98 (70% of values between 0.79 and 0.98). ƙ for the SI ranged from 0.50 to 0.89 (70% of values between 0.69 and 0.89). Interobserver agreement for the SSc ranged from ƙ = 0.41 to ƙ = 0.89 (70% of values between 0.63 and 0.89). Cronbach α for PI, SI, and SSc was 0.9, 0.86, and 0.91, respectively (0.90 > α ≥ 0.80 = good IC, α ≥ 0.90 = excellent IC). Conclusion We present an IHC score that is simple and reproducible, and accounts for staining heterogeneity. Comparison with established scores warrants further investigation.