Quantification of Hematoxylin Staining Variability in Esophageal Carcinoma from 5 Institutions. Erem AS, Woo ESJ, Cecchini MJ

Summary: Context: The diagnosis of dysplasia and malignancy relies upon assessing subtle differences in nuclear staining and morphology. The nature ofreagents utilized in hematoxylin and eosin (H&E) staining causes substantial nuclear hematoxylin staining variation between institutions. Given theimportance of accurate nuclear staining, a representative H&E stain is critical for review. Variation in appearance often requires local recuts inconsultation cases, utilizing additional tissue and increasing costs. However, this may not be needed in some cases that align with local norms ofstaining appearance. There are currently limited options to assess H&E variation beyond subjective visual inspection. Design: Esophageal squamous cell carcinoma cases were selected from the TCGA data set from five institutions. Digital slides were rapidlyassessed on QuPath for the intensity of hematoxylin staining and optical density in lymphocytes and tumor nuclei. Prism was then used to conductANOVA analyses and create violin plots to represent the data. (Figure 1, A-D). Results: Significant variation (p<0.05) was observed in hematoxylin staining across institutions. Three demonstrated comparable intensities forhematoxylin staining and optical density. These would likely be acceptable for interchangeable diagnoses. There were two outliers, one site withvery strong hematoxylin staining while the other demonstrated very light staining. Conclusions: Using this system, the variation of the hematoxylin staining can be rapidly quantified. This system could be utilized to identify casesthat fall outside normal, acceptable ranges of hematoxylin staining and identify cases that need to be re-stained at the local institution.