Reproducibility of 3 histologic classifications and 3 staging systems for thymic epithelial neoplasms and its effect...

Data regarding the prognostic significance of the histopathologic classifications of thymic epithelial neoplasms are contradictory, perhaps reflecting issues in reproducibility. We studied the effect of reproducibility of 3 histopathologic classifications on prognosis and investigated the interobserver agreement on invasion and its effect on staging and prognosis. A total of 456 patients who underwent surgery for thymic epithelial neoplasm at Mayo Clinic Rochester (1942 to 2008) were staged (modified Masaoka, proposed Moran, proposed IASLC/ITMIG) and independently classified by 3 thoracic pathologists (World Health Organization, proposed Suster & Moran [S&M], and Bernatz). Interobserver agreement was moderate to substantial for all histopathologic classifications (κ values: 0.65, 0.52, 0.74 for World Health Organization, Bernatz, and S&M, respectively). All histopathologic classifications were significant for overall survival (OS) and disease-free survival (DFS) (all reviewers). If adjusted for Masaoka, only Bernatz classification for one reviewer and all histopathologic classifications for another reviewer were significant for OS. Interobserver agreement for invasion was substantial (κ=0.61) and almost perfect for Masaoka, Moran, and IASLC/ITMIG stage (κ values: 0.85, 0.81, and 0.92, respectively). The correlation coefficient for Masaoka and Moran staging was 0.93. Masaoka and IASLC/ITMIG staging were significant for OS and DFS (all reviewers). If adjusted for any histopathologic classification, Masaoka was significant for OS and DFS (all reviewers). In conclusion, reproducibility of histopathologic classifications has some effect on outcome. S&M is the most reproducible classification. Reproducibility of invasion has no effect on the prognostic value of staging. Masaoka, Moran, and IASLC/ITMIG staging are almost perfectly reproducible. The strong correlation between Masaoka and Moran staging suggests similar prognostic strength.