Pathological Evaluation of Endomyocardial Biopsies from Transplanted Hearts – Utilization of Automated Whole Slide Imaging for Quality Assurance.
Lise AM Matzke, Avi J Ostry, Bruce M McManus, Michael F Allard. University of British Columbia, Vancouver, Canada; St Paul's Hospital, Vancouver, BC, Canada
Background: Modern automated whole slide imaging (WSI) systems have great potential as clinical, educational and research tools in pathology. One potential benefit is the use of such systems in telepathology. Prior to widespread usage of WSI systems for telepathology, the impact of this technology must be evaluated with respect to quality assurance (QA) of diagnoses rendered from samples evaluated. In this project, we examined the use of automated WSI systems for surgical pathology with regard to QA for the pathological evaluation of endomyocardial biopsies (EMB) from heart allografts.
Design: Sixty EMB from transplanted hearts that were previously examined by a pathologist using traditional light microscopy (LM) were selected randomly from all EMB accessioned between 2009 and early 2010. A diagnosis based on the 2006 ISHLT scoring system was previously rendered on each case by one of three pathologists (AO, BM, MA) with training and expertise in cardiovascular and transplantation pathology. The sixty cases (20 for each pathologist) were blinded and automated WSI capture was performed on an Aperio ScanScope XT Scanner (Aperio Technologies, Vista, CA) located at St. Paul's Hospital. WSI for each case were produced and each pathologist re-evaluated their 20 cases and rendered a new diagnosis. These results were compared to an earlier intra-observer evaluation study comparing score concordance based solely on LM (2 pathologists; evaluating 4 and 5 cases respectively).
Results: Concordance rates based on LM versus WSI methods ranged from 14/20, 18/20 to 20/20 cases. Weighted Kappa score evaluation was applied to calculate intra-observer variation between the two diagnostic methods, the results of which correspond to moderate to very good agreement. Concordance rates based solely on LM ranged from 3/4 to 4/5 cases, both corresponding to good agreement. Following the case analyses, all pathologists were asked to critically comment on the benefits, drawbacks and challenges in using the two methodologies.
Conclusions: This study evaluated quantitative and qualitative QA aspects using WSI technology for pathological evaluation of EMB suggesting that, with further consideration, the WSI approach has the potential in enabling telepathology as an important tool in the remote pathological evaluation of cardiac specimens as well as for QA and consultation purposes. Such utility will ultimately lead to more effective, quickly available patient care.
Wednesday, March 2, 2011 1:00 PM
Poster Session VI # 45, Wednesday Afternoon