Concordance Among Pathologists in the Second Cardiac Allograft Rejection Gene Expression Observational Study (CARGO II)

Background. There has been no large evaluation of the ISHLT 2004 acute cellular rejection grading scheme for heart graft endomyocardial biopsy specimens (EMBs). Methods. We evaluated agreement within the CARGO II pathology panel and between the panel (acting by majority) and the collaborating centers (treated as a single entity), regarding the ISHLT grades of 937 EMBs (with all grades >= 2R merged because of small numbers). Results. Overall all-grade agreement was almost 71% both within the panel and between the panel and the collaborating centers but, in both cases, was largely because of agreement on grade 0: for the average pair of pathologists, fewer than a third of the EMBs assigned grade >= 2R by at least one were assigned this grade by both. Conclusion. The 2004 revision has done little to improve agreement on the higher ISHLT grades. An EMB grade >= 2R is not by itself sufficient as a basis for clinical decisions or as a research criterion. Steps should be taken toward greater uniformity in EMB grading, and efforts should be made to replace the ISHLT classification with diagnostic criteria-EMB based or otherwise-that correspond better with the pathophysiology of the transplanted heart.