Whole-Slide Digital Imaging Versus Optical Microscopy for Primary Diagnosis of Hematoxylin-and-Eosin-Stained Breast Tissue Sections
Savitri Krishnamurthy, Kurt Mathews, Steven McClure, Melissa Murray, Daniel Visscher. Md Anderson Cancer Center, Houston; Scripps Memorial Hospital, LaJolla; Presbyterian Hospital, Charlotte; Memorial Sloan Kettering Cancer Center, New York; University of Michigan, Ann Arbor
Background: Whole slide imaging (WSI) is now used for educational purposes, archiving and quantitation of immunostains. However, WSI is not routinely used for primary diagnosis of hematoxylin and eosin (H&E)-stained tissue sections. We conducted a multi-institutional study to compare whole-slide digital images using the Aperio digital pathology system versus optical microscopy (OM) for primary diagnosis of H&E-stained tissue sections of breast lesions.
Design: The study was conducted at 3 clinical sites with 3 breast pathologists who interpreted 150 H&E slides, 3 times by WSI and 3 times by OM. For WSI, slides were scanned using the Aperio ScanScope® and interpreted on a computer monitor using Aperio ImageScope software and Aperio SpectrumTM data management software. Diagnoses was recorded using the College of American Pathologists breast checklist. WSI was compared with OM for accuracy, precision (interpathologist variation), and reproducibility (intrapathologist variation). Results were considered accurate only if the diagnosis matched exactly between the 2 platforms. The proportion of accurate results reported by each pathologist was expressed as a percentage for the comparison of WSI with OM.
Results: The accuracy of WSI with OM for classifying lesions as no carcinoma, noninvasive (ductal, lobular), or invasive (ductal, lobular, other) was 90.5%. Interpretations made using OM compared to themselves resulted in accuracy of 92.1%. The precision and reproducibility of WSI in diagnosis of the breast lesions in comparison to OM, obtained by performing pairwise comparisons, included 3 comparisons for each slide, resulting in 36 possible comparisons. Overall, 12,164/13,447 and 3174/3465 of the pairwise interpretations by WSI were the same, compared to 14,395/15,628 and 3703/3919 of the pairwise comparisons by OM, resulting in precision of 90.5% and 92.1%, respectively, and reproducibility of 91.6% and 94.5%, respectively.
Conclusions: The study demonstrated substantial equivalence between WSI and OM with similar accuracy, precision, and reproducibility for making a primary diagnosis of H&E-stained breast tissue sections. 2. Prospective clinical studies using routine surgical pathology specimens can confirm equivalence between WSI and conventional OM and facilitate the incorporation of WSI for making primary histologic diagnosis.
Monday, March 19, 2012 11:15 AM
Platform Session: Section B, Monday Morning