Background: Endomyocardial biopsy is used to guide therapy after heart transplantation. An accurate and reliable diagnosis of rejection is critical for proper patient management.
Methods: A sub-set of 827 biopsies from 273 patients were identified from 8 centers participating in the Cardiac Allograft Gene Expression Observational Study. These included all biopsies graded by local center pathologists as International Society for Heart and Lung Transplantation (ISHLT) Grade 1B or higher and also randomly chosen Grade 0 and 1A biopsies. Each of these cases was reviewed in a blinded manner by 3 study pathologists in the absence of clinical data. The study pathologists were assigned an ISHLT grade and noted nodular endocardial infiltrates (Quilty lesions).
Results: The study pathologists were significantly more likely than local pathologists to diagnose ISHLT Grade 0, 1A and 3B rejection and significantly less likely to diagnose ISHLT Grade 1B, 2 and 3A rejection. Concordance between local and study pathologists was lowest for Grade 2 (17% agreement). Quilty lesions were noted in 3.3% of local Grade 0 cases and in 31% and 37% of local Grade 2 and 3A cases, respectively. Quilty lesions were recognized by study pathologists in 35% of local Grade 2 cases "downgraded" to Grade 0 or 1, but in only 10% of local Grade 2 cases confirmed by study pathologists.
Conclusions: The greatest variability between pathologists in application of the ISHLT grading system is in Grade 2 biopsies, and Quilty lesions are a major contributing factor to the lack of concordance. Accurate application of the ISHLT grading system requires improved recognition and understanding of Quilty lesions.