Interobserver variability in grading transbronchial lung biopsy specimens after lung transplantation

Background: Acute rejection remains a major source of morbidity after lung transplantation. Given the importance of this diagnosis, an international grading system was developed to standardize the diagnosis of acute lung-allograft rejection. The reliability of this grading system has not been adequately assessed by previous studies. Methods: We examined the level of agreement in grading transbronchial biopsy specimens obtained from a large multicenter study (AIRSAC [Comparison of a Tacrolimus/Sirolimus/Prednisone Regimen vs Tacrolimus/Azathioprine/Prednisone Immunosuppressive Regimen in Lung Transplantation] trial). Biopsy specimens were initially graded for acute rejection and lymphocytic bronchiolitis by the site pathologist and subsequently graded by a central pathologist. Reliability of interobserver grading was evaluated using Cohen k coefficients. Results: A total of 481 transbronchial biopsy specimens were graded by both the site and central pathologists. The overall concordance rates were 74% and 89% for grade A and grade B biopsy specimens, respectively. When samples from biopsies performed at different time points after transplantation were assessed, there was a higher level of agreement early (≤ 6 weeks) after transplant compared with later time points for acute rejection. However, there was still only moderate agreement for both grade A (k score 0.479; 95% CI, 0.29-0.67) and grade B (k score 0.465; 95% CI, 0.08-0.85) rejection. Conclusions: These results expand upon previous reports of interobserver variability in grading transbronchial biopsy specimens after lung transplantation. Given the variability in grading these specimens, we advocate further education of the histopathologic findings in lung transplant biopsy specimens, as well as revisiting the current criteria for grading transbronchial biopsy specimens to improve concordance among lung transplant pathologists.